Structural Studies on Bovine Pancreatic Phospholipase · PDF fileStructural Studies on Bovine Pancreatic Phospholipase A 2 and Proteins involved in Molybdenum Cofactor Biosynthesis

Structural Studies on Bovine Pancreatic Phospholipase A2 and Proteins involved in

Molybdenum Cofactor Biosynthesis

A Thesis

Submitted for the Degree of

Doctor of Philosophy

in the Faculty of Engineering

by

Shankar Prasad Kanaujia

Bioinformatics Centre

(Centre of Excellence in Structural Biology and Bio-computing)

Supercomputer Education and Research Centre

INDIAN INSTITUTE OF SCIENCE

Bangalore-560 012

October 2010

Dedicated to

My Parents, Brothers, Sisters, Wife and Daughter

DECLARATION I hereby declare that the work reported in this thesis is entirely original and was carried

out by me under the general supervision of Professor K. Sekar, Bioinformatics Centre,

Supercomputer Education and Research Centre, Indian Institute of Science, Bangalore,

India.

I further declare that the contents of this thesis have not been the basis for the

award of any degree, diploma, fellowship, associateship or any other similar title of any

University or Institution.

Date: Shankar Prasad Kanaujia


Supercomputer Education and Research Centre

Indian Institute of Science

Bangalore – 560 012, India

CERTIFICATE This is to certify that the work described in the thesis entitled "Structural Studies on

Bovine Pancreatic Phospholipase A2 and Proteins involved in Molybdenum

Cofactor Biosynthesis" is the result of investigations carried out by Shankar Prasad

Kanaujia at the Bioinformatics Centre, Supercomputer Education and Research Centre,

Indian Institute of Science, Bangalore, India under my supervision and the results

presented in this thesis have not previously formed the basis for the award of any other

diploma, degree or fellowship.

Date: Professor K. Sekar


Indian Institute of Science

Bangalore – 560 012, India

ACKNOWLEDGEMENTS First of all I thank ever-loving God for the countless blessings HE showered upon me over the

years.

I take this opportunity to thank my thesis supervisor Prof. K. Sekar for all his support during

my stay here. I would always be indebted to him for introducing me in the field of structural biology. In

addition to science, I have learnt several things from him in my life. Also, I am privileged to

acknowledge my sincere thanks towards his family for their love for my daughter.

Most of the work reported in this thesis has been done in collaboration with Dr. Jeyakanthan

and RIKEN Structural Genomics Initiative groups, JAPAN. I am grateful to Dr. Jeyakanthan for his

keen interest in my work, support and supply of the materials required for my crystallization studies. I

am very thankful to him for collecting the Synchrotron data. I also thank Prof. M.-D. Tsai (Ohio State

University) for kindly gifting the bovine pancreatic phospholipase A2 mutants.

I thank Prof. R. Govindarajan, chairman SERC, for his encouragements and his generous

support by allowing me to stay in departmental family apartment for almost two years. My special

thanks are due to Prof. S. Ramakumar and Prof. S. Vishveshwara past and present chairman of the

centre, for their support and wonderful lectures. I thank Dr. Nagasuma Chandra and Dr. Debnath Pal

for their encouragements throughout my stay here. I thank all the students of their research groups.

Though I belong to the SERC department, most of my work has been carried out at the

Molecular Biophysics Unit (MBU) of the institute. Thus, I would like to take this opportunity to thank

all the faculties and students without whom most of my work would have never been possible. First of

all, I would thank all the respected teachers: Prof. M. Vijayan, Prof. M. R. N. Murthy, Prof. K.

Suguna, Prof. N. Srinivasan, Prof. M. Bansal and Prof. S. Vishweswara for their excellent teachings in

their respective fields. I would specially like to thank Prof. M. R. N. Murthy for all his efforts he took

to make sure that students understand the concepts in protein crystallography. I thank Prof. B. Gopal

for his help during X-ray diffraction data collection at MBU. I thank Prof. R. Varadarajan and his

student Tariq for helping me in ITC experiments, which I have used in my work.

It is always great feeling to have good friends and colleagues around. My heartfelt thanks to

Kavya and Ansuman for their help at various occasions. Our lab is known for recruiting a large number

of project assistants. I thank one and all for their cooperation and keeping the lab lively. My special

regards are due to Bala, Roshan, Selva, Shaahul, Jaikumar, Ramesh, Brinda, Daliah, Sumathi, Senthil,

Saravanan S, Uday, Venketesh, Gopal, Prabhakar, Sarani, Sowmiya, Vasuki, Kalaivani, Sivsankari,

Karthik, Praveen, Satiyamoorthy, Sanjeev, Chetan, Hema, Kokila, Satish, Saravanan SE, Sabari,

Gnanasankaran, Bhagya, Sangeetha, Vikas, Udayaprakash, Kesavan, Nivetha, Sathyaramya,

Sathyapriya, Archana, Shankar, Uthay, Sivabalan, Sherlin, Vaishnavi.

I thank all the members of MBU specially the crystallography labs. My special thanks to Dr.

Satyabrata Dr. Thamotharan, Dr. Krishna, Dr. Kalaivani and their families. I thank Dr. Kunthavai,

Dr. Rajan, Dr. Prem & Family, Dr. Kalaivani, Bhaskar & Family, Patra, Selvaraj, Anu, Tyageshwar,

Arif and Abhinav.

I am fortunate to have seniors and friends like Siddhartha and Alok and their families. Alok

has been like a guide and guardian for me. He has supported me both in personally as well as

professionally. Most of the work has been done in close interaction with him. I also thank Bharat (RV)

for his valuable help in my computational work. I thank all the members of Dcryst specially Dr.

Swarnamukhi, Dr. Simanshu, Dr. Rajaram, Dr. Gayathri, Kris, Rajavel, Sagar, Bharath, Kartika,

Koustav, Girish. I thank all my IISc friends: Sanjay & family (MCBL), Vijayabhaskar (MBU),

Susanta & family (SSCU), Dr. Subhash & family (IHBT, Palampur), Dr. Kalyan. My special thanks to

Dr. Ashima for her kind support by providing me the dialysis membrane whenever required. I thank

Senthil (Prof. M. Bansal's Lab, MBU) for his efforts to make the MBU cluster running all the time. I

thank the X-ray facility at MBU. I would like to extend my sincere thanks to Mr. James and Mr. Babu

(MBU) for the X-ray lab maintenance. I should acknowledge the kind of co-operation and help I

received from our office bearers at the centre Mrs. Vyjayanthi, Mrs. Gayathri, Mr. Krishnappa. I thank

all the SERC office staff. In particular Ms. Mallika for making sure that I get my fellowship at the

right time and Mr. Shekar for his timely help on different occasions.

I always feel fortunate for the lifetime opportunity the institute provided to me to work with a

group of great eminent scientists, for the wonderful campus and for the scholarship. I also thank the

institute for supporting my visit to Taiwan to attend the AsCA-2007 meeting. My sincere thanks are

due to IUCr (International Union for Crystallography) for the travel support to Taiwan.

I thank people of my village for their encouragements and supports, though they tease me

asking for how many more years I would study.

I thank all my family members and in-laws. First of all I thank my parents for their love and

patience they have. They always try to make sure that I am unaware of family conditions. I am

fortunate to have two brothers who have supported me and have taken over the responsibilities of

running the family smoothly, making me worry-free. I thank all my three sisters and their husbands. My

special thanks are also due to my Father-in-law for supporting me financially during my post graduation

at IIT-Bombay. Here, I would like to thank IITB for making my food cost-free. My hearty thanks are

due to my wife (Anita) and daughter (Anamika) for all their love, support and patience.

Contents Abbreviations ..................................................................................................................i Abstract ..........................................................................................................................ii CHAPTER 1 Structural Biology of Bovine Pancreatic Phospholipase A2 and Proteins involved in Molybdenum Cofactor Biosynthesis ................................................ 1

1.1 INTRODUCTION ...................................................................................................... 2 1.2 BOVINE PANCREATIC PHOSPHOLIPASE A2 .......................................................... 2

1.2.1 Introduction ............................................................................................................2 1.2.2 Physiological Roles of PLA2 .................................................................................5 1.2.3 Interfacial Catalysis ...............................................................................................6 1.2.4 Catalytic Mechanism .............................................................................................6 1.2.5 Types of PLA2 ........................................................................................................9 1.2.6 Classification of Secretory PLA2s ..........................................................................9 1.2.7 Structural Biology of Secretory PLA2 .................................................................11

1.2.7.1 Group IA sPLA2 ...............................................................................................12 1.2.7.2 Group II sPLA2 ................................................................................................13 1.2.7.3 Group III sPLA2 ...............................................................................................15 1.2.7.4 Group V sPLA2 ................................................................................................16 1.2.7.5 Group X sPLA2 ................................................................................................17 1.2.7.6 Group XI sPLA2 ...............................................................................................18 1.2.7.7 Group XII sPLA2 .............................................................................................19 1.2.7.8 Group XIII sPLA2 ............................................................................................19 1.2.7.9 Group XIV sPLA2 ............................................................................................19

1.2.8 Quaternary Structure ............................................................................................20 1.2.9 Structural Biology of Bovine Pancreatic PLA2 ....................................................21

1.2.9.1 Conserved Substructures of BPLA2 .................................................................23 1.2.9.2 Site-directed Mutagenetic Studies ...................................................................25 1.2.9.3 Role of His48 and Asp49 .................................................................................27 1.2.9.4 Role of Divalent Ca2+ ion .................................................................................28 1.2.9.5 Role of Water Molecules in BPLA2 .................................................................29

1.3 PROTEINS INVOLED IN MOLYBDENUM COFACTOR BIOSYNTHESIS .................. 31 1.3.1 Introduction ..........................................................................................................31 1.3.2 Molybdenum ........................................................................................................32 1.3.3 Molybdenum Cofactor .........................................................................................32 1.3.4 Molybdenum Cofactor Biosynthesis ....................................................................34 1.3.5 Operons involved in Molybdenum Cofactor Biosynthesis ..................................36 1.3.6 Molybdoenzymes .................................................................................................37 1.3.7 Physiological Roles of Molybdenum and Molybdenum Cofactor .......................39 1.3.8 Structures and Functions of Proteins involved in Moco Biosynthesis Pathway ..40

1.3.8.1 Conversion of GTP to cPMP ...........................................................................40 1.3.8.2 Synthesis of Molybdopterin .............................................................................45 1.3.8.3 Adenylation of Molybdopterin ........................................................................46 1.3.8.4 Transport of Molybdenum ...............................................................................48 1.3.8.5 Insertion of Molybdenum ................................................................................49 1.3.8.6 Maturation of Molybdenum Cofactor ..............................................................50 1.3.8.7 Storage of Molybdenum Cofactor ....................................................................51 1.3.8.8 Transfer of Molybdenum Cofactor to Molybdoenzymes .................................51

1.4 PLAN OF THE WORK ........................................................................................... 52

CHAPTER 2 Materials and Methods ............................................................................................ 54

2.1 INTRODUCTION .................................................................................................. 55 2.2 PROTEIN CRYSTALLOGRAPHY .......................................................................... 55

2.2.1 Crystallization ......................................................................................................55 2.2.2 Intensity Data Collection and Processing ............................................................56

2.2.2.1 Data Collection Strategy ..................................................................................56 2.2.2.2 Data Processing ................................................................................................56

2.2.3 Calculation of Structure Factor Amplitudes ........................................................58 2.2.4 Structure Solution ................................................................................................58

2.2.4.1 Molecular Replacement ...................................................................................58 2.2.4.2 Phaser ...............................................................................................................59

2.2.5 Structure Refinement ...........................................................................................60 2.2.5.1 Cross-Validation ..............................................................................................61 2.2.5.2 Target Functions ..............................................................................................61 2.2.5.3 Maximum-Likelihood Refinement Targets ......................................................62 2.2.5.4 Rigid-Body Refinement ...................................................................................63 2.2.5.5 Positional Refinement ......................................................................................63 2.2.5.6 Simulated Annealing ........................................................................................63 2.2.5.7 Atomic-Displacement (B-Factor) Refinement .................................................64 2.2.5.8 Torsional-Angle Dynamics ..............................................................................64 2.2.5.9 Constraints and Restraints ................................................................................64 2.2.5.10 Bulk-Solvent Scattering .................................................................................65

2.2.6 Electron-Density Maps and Interpretation ...........................................................66 2.2.6.1 Identification of Solvent Sites ..........................................................................66 2.2.6.2 Reducing Model Bias with Omit Maps ............................................................67

2.2.7 Structure Validation and Deposition ....................................................................67 2.2.7.1 PROCHECK ....................................................................................................67 2.2.7.2 MolProbity .......................................................................................................67 2.2.7.3 ADIT ................................................................................................................68

2.2.8 Analysis of Sequences and Structures .................................................................68 2.2.8.1 Sequence Analysis ...........................................................................................68 2.2.8.2 Phylogenetic Tree ............................................................................................69 2.2.8.3 Secondary-Structure Elements .........................................................................69 2.2.8.4 Structural Comparison .....................................................................................69 2.2.8.5 Structural Rigidity ............................................................................................70 2.2.8.6 Hydrogen Bonds ..............................................................................................70 2.2.8.7 Electrostatic Potentials and Surfaces ...............................................................70 2.2.8.8 Identification of Functional Sites .....................................................................70 2.2.8.9 Protein-Protein Docking ..................................................................................71 2.2.8.10 Others .............................................................................................................71

2.2.9 Structure Visualization .........................................................................................71 2.3 MOLECULAR DYNAMICS SIMULATIONS ............................................................. 71

2.3.1 Introduction ..........................................................................................................71 2.3.2 General Theory of Molecular Dynamics ..............................................................72 2.3.3 Protocols and Parameters of Molecular Dynamics Simulation.............................73

2.3.3.1 System Representation, Input and Parameters .................................................73 2.3.3.2 Computation of Forces .....................................................................................74 2.3.3.3 Configuration Update ......................................................................................75 2.3.3.4 Output ..............................................................................................................76

2.3.4 Force Fields ..........................................................................................................76 2.3.4.1 Non-Bonded Interaction Terms .......................................................................76 2.3.4.2 Long-Range Electrostatics ...............................................................................77 2.3.4.3 Bonded Interaction Terms ................................................................................78 2.3.4.4 Restraints .........................................................................................................80

2.3.5 Force-Fields Used ................................................................................................80

2.3.5.1 OPLS-AA ........................................................................................................81 2.3.5.2 AMBER03 .......................................................................................................82

2.3.6 Water Model ........................................................................................................82 2.3.7 Ligand Parameters ...............................................................................................83 2.3.8 Energy Minimization Methods .............................................................................83

2.3.8.1 Steepest Descent ..............................................................................................84 2.3.8.2 Conjugate Gradient ..........................................................................................84 2.3.8.3 L-BFGS ............................................................................................................84

2.3.9 Periodic Boundary Condition ..............................................................................85 2.3.10 Visualization ......................................................................................................85 2.3.11 Analysis .............................................................................................................85

2.4 OTHER TECHNIQUES USED ................................................................................. 86 2.4.1 Isothermal Titration Calorimetry .........................................................................86

CHAPTER 3 Structure and Molecular Dynamics Studies of Three Active-Site Mutants of Bovine Pancreatic Phospholipase A2 ……………...…..........................................87

3.1 INTRODUCTION .................................................................................................... 88 3.2 RESULTS AND DISCUSSION .................................................................................. 90

3.2.1 H48N Mutant .......................................................................................................90 3.2.2 D49N Mutant .......................................................................................................93 3.2.3 D49K Mutant .......................................................................................................95 3.2.4 Active-Site and Surface-Loop Residues ..............................................................98 3.2.5 Invariant Water Molecules .................................................................................101

3.3 CONCLUSION ..................................................................................................... 102 3.4 MATERIALS AND METHODS .............................................................................. 103

3.4.1 Protein Purification and Crystallization .............................................................103 3.4.2 Data Collection and Processing .........................................................................104 3.4.3 Structure Refinement, Validation and Analysis .................................................105

3.4.3.1 Refinement of H48N Mutant .........................................................................105 3.4.3.2 Refinement of D49N and D49K Mutants ......................................................107

3.4.4 Molecular Dynamics Simulation .......................................................................108 CHAPTER 4 Structural and Functional Role of Water Molecules in Bovine Pancreatic Phospholipase A2: a Data-mining Approach ……….……………………........ 109

4.1 INTRODUCTION .................................................................................................. 110 4.2 RESULTS AND DISCUSSION ............................................................................... 112

4.2.1 All 24 Invariant Water Molecules ......................................................................112 4.2.2 Invariant Water Molecules in Cluster-1 .............................................................116 4.2.3 Invariant Water Molecules in Cluster-2..............................................................121 4.2.4 Invariant Water Molecules in Cluster-3..............................................................125


4.4.1 Data Set...............................................................................................................130 4.4.2 Molecular Dynamics Simulation ........................................................................131

CHAPTER 5 Crystal Structures of Apo and GTP-Bound Molybdenum Cofactor Biosynthesis Protein MoaC from Thermus thermophilus HB8 ..…................ 133

5.1 INTRODUCTION .................................................................................................. 134

5.2 RESULTS AND DISCUSSION ................................................................................ 135 5.2.1 Crystallographic Results ....................................................................................135

5.2.1.1 Overall Structure ............................................................................................135 5.2.1.2 Active-Site Geometry ....................................................................................137 5.2.1.3 Phosphate and GTP Binding Site ...................................................................138 5.2.1.4 Other Molecules Bound in the Active Site ....................................................140 5.2.1.5 Changes due to Substrate Binding in the Active Site .....................................140 5.2.1.6 Invariant Water Molecules .............................................................................141 5.2.1.7 Plasticity of TtMoaC ......................................................................................142 5.2.1.8 Comparison with MoaC from Other Organisms ............................................144

5.2.2 Results from Isothermal Titration Calorimetry Experiments .............................146 5.2.3 Results from Molecular Dynamics Simulations ................................................148

5.2.3.1 General Features ............................................................................................148 5.2.3.2 Energetics ......................................................................................................150 5.2.3.3 Protein Dynamics ...........................................................................................152 5.2.3.4 Role of Invariant Water Molecules ................................................................154

5.2.4 A Possible Mechanisms for the First Step of Moco-Biosynthesis Pathway ......155 5.3 CONCLUSION ..................................................................................................... 156 5.4 MATERIALS AND METHODS .............................................................................. 157

5.4.1 Cloning, Expression and Protein Purification ....................................................157 5.4.2 Protein Crystallization .......................................................................................158 5.4.3 Data Collection and Processing .........................................................................159 5.4.4 Structure Solution, Refinement and Validation .................................................160 5.4.5 Isothermal Titration Calorimetry .......................................................................161 5.4.6 Molecular Dynamics Simulation .......................................................................161 5.4.7 Structural Analysis .............................................................................................163

CHAPTER 6 Structure, Dynamics and Functional Implications of Molybdenum Cofactor Biosynthesis Protein MogA from two Thermophilic Organisms …………..164

6.1 INTRODUCTION .................................................................................................. 165 6.2 RESULTS AND DISCUSSION ................................................................................ 166

6.2.1 Annotation of TTHA0341 as MogA ..................................................................166 6.2.2 Protein Activity ..................................................................................................167 6.2.3 Crystallographic Results ....................................................................................167

6.2.3.1 Overall Structure and Active Site of TtMogA and AaMogA .........................167 6.2.3.2 Sequence Comparison ....................................................................................170 6.2.3.3 Sequence Determinants of Quaternary Structure ...........................................172 6.2.3.4 Structure Comparison ....................................................................................175 6.2.3.5 Protein Surface Charge Distribution ..............................................................176 6.2.3.6 Oligomerization .............................................................................................181 6.2.3.7 Role of the N- and C-terminal Residues ........................................................183 6.2.3.8 MogA-MoeA Protein-Protein Complex .........................................................184 6.2.3.9 Invariant and Interfacial Water Molecules .....................................................187

6.2.4 Molecular Dynamics and Docking Results ........................................................192 6.2.4.1 General Features ............................................................................................192 6.2.4.2 Energetics ......................................................................................................193 6.2.4.3 Proteins Dynamics .........................................................................................195


6.4.1 Cloning, Expression and Protein Purification ....................................................198 6.4.2 Crystallization Experiments ...............................................................................200 6.4.3 Data Collection and Processing .........................................................................201 6.4.4 Structure Solution, Refinement and Validation .................................................202

6.4.5 Molecular Dynamics Simulation .......................................................................203 6.4.6 Molecular Docking ............................................................................................204 6.4.7 Structural Analysis .............................................................................................205

Summary and Future Perspectives .......................................................................207 References .................................................................................................................210

ABBREVIATIONS i

ABBREVIATIONS AaMogA Aquifex aeolicus MogA AMPBS Adenosine Monophosphate Binding Site ASC Active Site Channel AtCnx1G Arabidopsis thaliana Cnx1G BPLA2 Bovine Pancreatic Phospholipase A2 BsMoaB Bacillus subtilis MoaB cPLA2 cytosolic Phospholipase A2 cPMP cyclic Pyranopterin Monophosphate EcMoaB Escherichia coli MoaB EcMoaC Escherichia coli MoaC EcMoeA Escherichia coli MoeA EcMogA Escherichia coli MogA FeMoco Iron Molybdenum cofactor FLC Citrate FPT Formamidopyrimidine Type GkMoaC Geobacillus kaustophilus MoaC GTPFW Guanosine Triphosphate Without Citrate HiMogA Haemophilus influenzae MogA HpMogA Helicobacter pylori MogA HsGephG Homo sapiens GephG iPLA2 Ca2+-independent Phospholipase A2 ITC Isothermal Titration Calorimetry L-BFGS Low- memory Broyden-Fletcher-Goldfarb-Shanno LPA Lysophosphatidic Acid MCP Moco Carrier Protein MD Molecular Dynamics MGD Molybdopterin Guanine Dinucleotide Moco Molybdenum cofactor MPT Molybdopterin MPTBS Molybdopterin Binding Site MR Molecular Replacement OPLS Optimized Potentials for Liquid Simulations PBC Periodic Boundary Condition PDB Protein Data Bank PfMoaB Pyrococcus furiosus MoaB PhMoaC Pyrococcus horikoshii MoaC PLB Phospholipase B PLC Phospholipase C PLD Phospholipase D PPLA2 Porcine Pancreatic Phospholipase A2 Rmsd Root Mean Square Deviation Rmsf Root Mean Square Fluctuation RnGephG Rattus norvegicus GephG SaMoaB Staphylococcus aureus MoaB SoMogA Shewanella oneidensis MogA sPLA2 Secretory phospholipase A2 StMoaB Sulfolobus tokodaii MoaB StMoaC Sulfolobus tokodaii MoaC TtMoaC Thermus thermophilus MoaC TtMogA Thermus thermophilus MogA

ABSTRACT ii

ABSTRACT Phospholipase A2 (PLA2, EC 3.1.1.4) catalyzes the hydrolysis of

glycerophospholipids at the sn-2 ester bond to produce lysophospholipids and free fatty

acids. PLA2s were the first type of enzymes discovered to be involved in the interfacial

catalysis process wherein the enzyme first binds to the aggregates of the substrate

molecules and then perform its hydrolytic activity. PLA2s are suggested to be involved

in many biological processes such as inflammation, cell signaling and lipid digestion

and several diseases like arthritis, Alzheimer's, Parkinson's, etc. In addition, they are

recently proposed to be involved in the host defense against microbial pathogens,

fungal invasion and adenoviral infection. PLA2s are found in most living organisms and

in virtually all cell types. However, those found in snake venoms and pancreas is the

most thoroughly characterized and studied among the members of the family. PLA2s

are grouped into many classes based on localization (e.g. cytosolic and secretory) and

for their requirement of calcium ion (e.g. Ca2+-dependent and Ca2+-independent).

Bovine pancreatic phospholipase A2 (BPLA2) belongs to the secretory and calcium-

dependent group IB (PLA2GIB). BPLA2 is a monomer containing 123 amino acids of

which 14 cysteines form seven disulfide bonds, thus providing stability to the enzyme.

The mechanism of the catalytic activity of the enzyme PLA2 is similar to that of the

serine protease except a water molecule plays the role of nucleophile. The catalytic

dyad (His48-Asp99) along with a nucleophilic water molecule is responsible for

hydrolytic process of the enzyme. Furthermore, the residue Asp49 is essential for

controlling the binding of calcium ion and the catalytic activity of the enzyme.

Biochemical and NMR studies on His48 and Asp49 single mutants suggested that

H48N mutant is active though it is several folds weaker than the wild type enzyme.

Similarly, the study suggested that D49N and D49K mutants do not bind to the

functionally important calcium ion and shows structural perturbation, hence the

mutants D49N and D49K show no enzymatic activity. Thus, the present work was

started with the aim of understanding the structural basis of these three active-site

mutants.

It is well-established fact that water molecules are an integral part of

biomolecular systems and are crucial in the protein-folding process and their functions.

It is also known that protein hydration plays an important role in biological processes

ABSTRACT iii

and that hydration forces are responsible for the packing and stabilization of three-

dimensional protein structure. In addition, water molecules are found to be involved in

many hydrogen-bonding networks. The common hydrophilic nature of the interfaces of

protein-protein, protein-DNA and protein-ligand complexes and the abundance of water

molecules at the interface suggest that water molecules are an indispensable component

of biomolecular recognition and self-assembly. Thus, water molecules identified to be

invariant in all the crystal structures of BPLA2 were analyzed and their structural and/or

functional role was discussed.

In addition to the work on BPLA2, structural studies on the molybdenum

cofactor (Moco) biosynthesis proteins MoaC and MogA were also carried out. The

biosynthesis of Moco is an evolutionary conserved pathway among almost all

kingdoms of life including humans. It is required for the activity of several enzymes

known as molybdoenzymes, which contain molybdenum ligated into Moco and are

known to play major roles in nitrogen, carbon and sulfur cycles. The deficiency of

Moco in human causes the accumulation of toxic levels of sulfite and neurological

damages usually leading to death within months of birth. The biosynthesis of Moco is

generally divided into five steps. Out of which, the first step involves the conversion of

GTP to precursor Z by two proteins (MoaA and MoaC). The protein MoaC forms a

hexamer, belongs to the ferredoxin-like fold and has been suggested to catalyze the

release of pyrophosphate and the formation of the cyclic phosphate of precursor Z.

However, structural evidence showing the binding of a substrate-like molecule with

MoaC is not available. The third step of Moco biosynthesis involves the adenylation of

an intermediate compound MPT and is performed by protein MogA. The protein MogA

forms a trimer and belongs to the Rossmann fold. In the present work, the crystal

structures of these two proteins MoaC (both in apo and complex forms) and MogA (apo

forms) have been determined at high resolution.

The crystallization experiments were carried out using hanging-drop and sitting-

drop vapor-diffusion methods. The intensity data were collected from both home source

and Synchroton radiation. The data related to the enzyme PLA2 were collected on a

MAR research imaging plate mounted on Rigaku RU300 generator and the data related

to Moco biosynthesis proteins MoaC and MogA were collected on Synchroton

radiation, except for the data of the ligand-bound MoaC. The data were processed and

ABSTRACT iv

scaled using DENZO and SCALEPACK of the HKL suite. All the structure solutions

were obtained by molecular-replacement technique using the program Phaser. Structure

refinements were carried out using the package CNS. Model building was done using

the program COOT. Several programs like PROCHECK, MolProbity, ALIGN, ESCET,

NACCESS, HBPLUS and CONTACT were used for structure validation and analysis

of the refined structures. Furthermore, the program ClusPro was used to carry out

protein-protein interactions. The program GROMACS versions 3.3, 3.3.3 and 4.0.4

were used to perform molecular-dynamics simulations. OPLS-AA and AMBER03

force fields were used for different simulations. Simulations were performed in explicit

water system with SPC water model under NPT conditions with unit dielectric

constant.

The crystal structures of the three active-site mutants (H48N, D49N and D49K)

of the BPLA2 enzyme have been determined. The overall tertiary structures of all three

mutants are similar to that of the wild-type enzyme. However, the active site is

disturbed in the case of the Asp49 mutants, whereas it is intact in the H48N mutant.

Thus, the crystal structures and molecular-dynamics simulations of the three single

mutants confirm that residue Asp49 is important for both calcium binding and the

integrity of the active site. On the other hand, His48 is not crucial for the stability of the

active site. However, it is important for the catalytic activity of the enzyme.

Furthermore, the active site framework and the role of structural and functional water

molecules are verified using the MD simulations.

The water molecules in 25 (21 high-resolution and four atomic-resolution)

crystal structures of BPLA2 have been analyzed to identify the invariant water

molecules and their possible roles. In total, 24 water molecules are identified as

invariant. Of these, nine invariant water molecules (IW1, IW2, IW3, IW4, IW5, IW8,

IW9, IW10 and IW19) are located in the core of the enzyme and are likely to be

involved in the folding of the enzyme. Invariant water molecules IW1 and IW2 are also

involved in the catalytic activity of the enzyme. Two invariant water molecules IW5

and IW8 are structurally essential providing coordination to the functionally important

active-site calcium ion and to maintain the correct active site geometry. In addition,

some invariant water molecules are observed to be involved in mediating ion pairs that

play an important role in stabilizing the tertiary structure. A set of water molecules

ABSTRACT v

forms a water bridge that stabilizes the functionally important residues. In addition,

about half of the invariant water molecules play a role in stabilizing the surface residues

of the enzyme. Thus, it can be concluded that, in addition to the structurally and

functionally important water molecules, the present study helps to rationalize the water

molecules that are significant for the folding and stability of the enzyme PLA2.

The crystal structures of MoaC from Thermus thermophilus coupled with the

ITC experiments and the MD simulations provide insights into substrate binding,

structure dynamics and a possible mechanism. For the first time, the crystal structure of

MoaC bound with GTP was reported. GTP-bound crystal structure revealed that the

residues Lys49, His75, Asp126 and Lys129 are critical for the biological activity of the

protein molecule. ITC results along with the interaction energies calculated from the

MD simulations provide insights into the chemical nature of the possible substrate

molecules capable of binding to the protein molecule. These results reveal that the

molecules with triphosphate groups are more potent to bind to MoaC. A comparison of

the available subunits from the present study led to delineate the rigid and the flexible

regions of the protein molecule. These results show that all/most of α-helices are rigid,

whereas β-sheets are flexible. The identification of invariant water molecules led to the

assignment of their structural and functional roles. In addition, the MD simulations

were used to obtain the interactions energies for the protein-ligand complexes to

support the findings of the crystallographic and the ITC results.

Crystal structures of Moco biosynthesis protein MogA from two thermophilic

organisms Thermus thermophilus HB8 and Aquifex aeolicus VF5 have been determined

at high resolution. Comparative study of the present crystal structures and those

available in the literature has led to the identification of the residues Pro47, Pro48,

Lys52, Arg55, Asp59, Glu86, Gly115, Arg120 and Ser131 (MogA from T.

thermophilus), which could possibly be involved in the oligomerization of the protein

molecule. Furthermore, five invariant and two interfacial water molecules are also

believed to play a role in oligomerization. Similarly, another five invariant and an

interfacial water molecule are likely to play a role in anchoring the active-site residues.

Based on comparative analyses, a possible role of the N- and C-termini residues of

MoaB and MogA proteins, respectively, are proposed in the stabilization of the

substrate and/or product molecule in the active site of the protein molecule. A possible

ABSTRACT vi

protein-protein conformer between MogA and MoeA has been predicted. The results

show that the residues (Arg3, Asp11, Glu46, Arg77, Lys106, Ser131 and Thr154) are

involved in protein-protein interactions. Furthermore, results obtained from the MD

simulations and molecular-docking calculations of several ligands with protein

molecules support the experimental results reported in the literature. The results show

that MPT and MPT-AMP can bind strongly to MogA than to MoaB proteins. In

addition, in most of the cases, MPTBS is preferred to AMPBS except for ATP

molecule. Furthermore, results from the MD simulations show that the active-site loops

are stabilized upon substrate and/or product binding.

A part of the work presented in the thesis has been reported in the following

publications.

Kanaujia, S.P., Ranjani, C.V., Jeyakanthan, J., Baba, S., Chen, L., Liu, Z.-J., Wang,

B.-C., Nishida, M., Ebihara, A., Shinkai, A., Kuramistu, S., Shiro, Y., Sekar, K. and

Yokoyama, S. (2007). Crystallization and preliminary crystallographic analysis of

Molybdenum cofactor biosynthesis protein C from Thermus thermophilus. Acta Cryst.

F63, 27-29.

Kanaujia, S.P., Ranjani, C.V., Jeyakanthan, J., Ohmori, M., Agari, K., Kitamura, Y.,

Baba, S., Ebihara, A., Shinkai, A., Kuramitsu, S., Shiro, Y., Sekar, K. and Yokoyama,

S. (2007). Cloning, expression, purification, crystallization and preliminary X-ray

crystallographic study of molybdopterin synthase from Thermus thermophilus HB8.

Acta Cryst. F63, 324-326.

Kanaujia, S.P. and Sekar, K. (2008). Crystal Structures and Molecular Dynamics

Studies of Three Active Site Mutants of Bovine Pancreatic Phospholipase A2. Acta

Cryst. D64, 1003-1011.

Kanaujia, S.P. and Sekar, K. (2009). Structural and Functional Role of Water

Molecules in Bovine Pancreatic Phospholipase A2: A Data-Mining approach. Acta

Cryst. D65, 74-84.

ABSTRACT vii

Kanaujia, S.P., Jeyakanthan, J., Nakagawa, N., Sathyaramya, B., Shinkai, A.,

Kuramitsu, S., Yokoyama, S. and Sekar, K. (2010). Crystal structures of apo and GTP-

bound molybdenum cofactor biosynthesis protein MoaC from Thermus thermophilus

HB8. Acta Cryst. D66, 821-833.

Kanaujia, S.P., Jeyakanthan, J., Shinkai, A., Kuramitsu, S., Yokoyama, S. and Sekar,

K. (2010). Crystal structures, dynamics and functional implications of molybdenum

cofactor biosynthesis protein MogA from two thermophilic organisms. Acta Cryst. F66,

(In press).

CHAPTER 1 Structural Biology of Bovine Pancreatic Phospholipase A2

and Proteins involved in Molybdenum Cofactor Biosynthesis

CHAPTER 1: BOVINE PANCREATIC PHOSPHOLIPASE A2

2

1.1 INTRODUCTION The work reported in this thesis involves structural studies on (1) Bovine

pancreatic phospholipase A2 and (2) Two proteins required for the biosynthesis of

molybdenum cofactor. Both the systems have been introduced briefly in the following

sections.

1.2 BOVINE PANCREATIC PHOSPHOLIPASE A2 1.2.1 INTRODUCTION

The phospholipase A2 (PLA2) family of enzymes catalyzes the hydrolysis of the

sn-2 ester bond of glycerophospholipids to produce free fatty acids (FAs) and

lysophospholipids (LPs) and constitutes one of the largest families of lipid hydrolyzing

enzymes (van Deenen and de Haas, 1964). FAs and LPs have many important

downstream roles and represent the first step in generating important secondary

messengers that play several essential physiological roles (Dennis et al., 1991). FAs

such as arachidonic acid (AA) can be converted into eicosanoids (ESs) through the

action of a variety of prostaglandin synthases, lipoxygenases and cytochrome P450

proteins (Funk, 2001). The ES molecules act by binding to specific G-protein coupled

receptors (Tsuboi et al., 2002) and can exert a wide range of physiological and

pathological processes like pain, fever and inflammation (Funk, 2001). The LPs can be

converted into lysophosphatidic acid (LPA) or platelet activating factor (PAF), which

are involved in cell proliferation, survival and migration (Moolenaar et al., 2004).

PLA2s are widely distributed in nature and form a superfamily that contains 15 distinct

groups and several subgroups (Schaloske and Dennis, 2006). They are found in most of

the living organisms and in virtually all the cell types (Verheij et al., 1981).

Activity of PLA2 was first studied in phenomenological detail as early as 1890s

from the venom of cobras (Stephens and Myers, 1898). They belong to lipolytic

superfamily of enzymes such as lipases (LAs) and phospholipases (PLAs), which share

common structural features that are important for their association with substrates like

lipids, lipoproteins and phospholipid layers (Muller and Petry, 2005). Activities of

PLA2s are observed both in intra- and extra-cellular spaces and are controlled by a wide

variety of agonists including hormones, neurotransmitters, growth factors and


3

cytokines. As biological membranes are integral to living cells and are largely

composed of phospholipids, lipases play important roles in cell biology (Tjoelker et al.,

1995; Spiegel et al., 1996). In addition, they perform essential roles in the digestion,

transport and processing of dietary lipids such as triglycerides, fats and oils (Richmond

et al., 2001). Since many of these enzymes are water-soluble, while their substrates are

water-insoluble, they use unique strategies for regulating the catalysis at lipid-water

interface (Scott et al., 1990b) and show a large increase in activity toward a substrate

organized in an aggregate compared to a monomer in solution (Verheij et al., 1981).

This ‘interfacial activation’ complicates the analyses of enzyme kinetics by introducing

an additional surface-binding step that may be separate from the formation of a

Michaelis complex with the substrate (Gelb et al., 1995). PLAs are grouped into four

major classes (Dennis, 1994) depending upon the phospholipid ester bond being

hydrolyzed. (i) Phospholipase A1 (PLA1, EC 3.1.1.32), (ii) Phospholipase A2 (PLA2,

EC 3.1.1.4), (iii) Phospholipase C (PLC, EC 3.1.4.3) and (iv) Phospholipase D (PLD,

EC 3.1.4.4; Figure 1.2.1). In addition, another phospholipase, namely, Phospholipase B

(PLB, EC 3.1.1.5), which catalyzes the reaction of the sn-1 and sn-2 bond hydrolysis

simultaneously, is also found (Ghannoum, 2000). The representative tertiary structures

of all these PLAs are shown in Figure 1.2.2. Among them, PLA2s have been

extensively studied from snakes, porcine, bovine and human.

Figure 1.2.1 Mode of action of Phospholipases. A generalized depiction of a phospholipid where X=H, choline, ethanolamine, inositol, etc. The site of action of phospholipase types A1, A2, C and D are shown with arrows. Reaction depicts the catalysis by PLA2superfamily hydrolyzing the phospholipids at the sn-2 position to yield lysophospholipids and free fatty acid.


4

Figure 1.2.2 Overall tertiary structure and active-site residues of (a) PLA1 from Rhizomucor miehei (PDB-id: 3TGL; Brzozowski et al., 1992), (b) PLA2 from Apis mellifera (PDB-id: 1POC; Scott et al., 1990a), (c) PLB from Escherichia coli (PDB-id: 1JRL; Lo et al., 2003), (d) PLC from Bacillus cereus (PDB-id: 1AH7; Hough et al., 1989) and (e) PLD from Streptomyces sp. pmf strain (PDB-id: 1F0I; Leiros et al., 2000).


5

One of the major macromolecular crystallography projects in this laboratory is

concerned with the structural studies on bovine pancreatic phospholipase A2 (BPLA2).

A substantial part of the thesis deals with the structural studies on the active-site

mutants of BPLA2.

1.2.2 PHYSIOLOGICAL ROLES OF PLA2 A large variety of biological functions have been proposed for PLA2s, but

definitive evidence for an in vivo role is lacking in many cases. Once activated, PLA2s

can mediate a variety of pathophysiological reactions either through a direct action or

through subsequent transformations of its products like AAs and LPs (Burke and

Dennis, 2009). PLA2s are proposed to be involved in the following major diseases.

As a mediator in inflammation

The hydrolytic products of PLA2 such as free FAs are rate-limiting precursor for

the formation of prostaglandins, leukotrienes and PAFs, which play role in

inflammation (Nevalainen et al., 2000). Thus, mammalian PLA2 enzymes are

considered as anti-inflammatory targets (Schevitz et al., 1995).

In rheumatoid arthritis

High contents of PLA2, PLC, prostaglandins and related ESs have been

observed in the synovial fluid of patients suffering with rheumatoid arthritis (Robinson

et al., 1975; Seilhamer et al., 1989; Bomalaski and Clark, 1990) and osteoarthritis

(Pruzanski et al., 1991).

In Alzheimer's disease

The products like prostaglandins, FAs, LPs, ESs, PAFs and reactive oxygen

species (ROS), all generated by PLA2 activity, participate in cellular injury particulary

in neurodegeneration. Altered membrane associated PLA2 activities have been

correlated with acute and chronic brain injury including cerebral trauma, ischemic

damage, induced seizers in brain, epilepsy and Alzheimer's disease (Bazan et al., 2002).

In Parkinson's disease

Patients with Parkinson's disease show increased oxidative stress and low PLA2

activity in substantia nigra and hence decrease the activity of detoxification of oxidized

membrane phospholipids (Ross et al., 1998).


6

In addition, these enzymes have also been suggested to be associated in

pancreatic disorders (Miyamoto et al., 1993), ischemia (Muralikrishna and Hatcher,

2006), cell-migration (Gambero et al., 2002), apoptosis (Taketo and Sonoshita, 2002),

Krabbe disease (Giri et al., 2006), cancer (Cummings, 2007), atherosclerosis (Webb,

2005), exocytosis (Wei et al., 2003) and digestion of phospholipids in dietary food

(Richmond et al., 2001). Abundant evidence indicates that certain members of the

mammalian secretory PLA2 enzymes play important roles in host defense against

microbial pathogens (Nevalainen et al., 2008) and adenoviral infection (Mitsuishi et al.,

2006).

1.2.3 INTERFACIAL CATALYSIS As naturally occurring phospholipids are insoluble in water and form

aggregates, the enzyme must be able to bind to such aggregates for catalysis (Verheij et

al., 1981). The interfacial catalysis by PLA2 can be described by steps of

E E* E*S E*P E*+P, where E - enzyme, E* - surface bound enzyme, E*S -

substrate-bound enzyme, E*P - product bound enzyme and P - product. Two modes of

interfacial catalysis namely ‘Scooting mode’ and ‘Hopping mode’ have been proposed

in the literature (Scott et al., 1990b; Berg et al., 1991). In the Scooting mode of

catalysis, the bound enzyme (E*) remains at the interface between the catalytic turnover

cycles, whereas, in the Hopping mode of catalysis, the binding (E to E*) and desorption

of the bound enzyme (E* to E) occur during each catalytic turnover cycle. Most of the

PLA2s display a characteristic increase in activity when substrates are interchanged

from monomers to aggregates (Winget et al., 2006), except those from group III (Lin et

al., 1988).

1.2.4 CATALYTIC MECHANISM The biochemical data of sPLA2 enzymes and crystallographic studies with

transition state analogue confirmed the essential features of the enzyme catalysis for

groups I and II sPLA2 (Verheij et al., 1980; Scott et al., 1990a). The catalytic

mechanism of secretory PLA2 can be described into four steps (1) binding of a Ca2+ ion

and substrate molecule in the active site, (2) general base-mediated attack on the bound

substrate, (3) formation and collapse of the tetrahedral intermediate and (4) product


7

release. The formation, stabilization and collapse of the transition state are

schematically outlined in Figure 1.2.3. This proposal assumes that the phosphonate

emulates the tetrahedral intermediate of esterolysis whose formation, stability and

productive collapse are fostered by the catalytic surface of the enzyme (Scott et al.,

1990b). The active site of PLA2 contains catalytic dyad (His-Asp) along with a water

molecule, which acts as a nucleophile during the enzyme catalysis. The active-site

formation of PLA2 is similar to that of serine proteases (Kraut, 1977; Scott and Sigler,

1994a) except a water molecule, which replaces serine (the third residue of the catalytic

triad in serine protease). The catalytic water molecule is hydrogen bonded to the Nδ1

atom of the histidine residue. On the other side, Nε2 atom of the residue histidine is

hydrogen bonded to the carboxylate oxygen atom of the residue Asp99, which is

suggested to tautomerize the histidine residue (Li and Tsai, 1993; Annand et al., 1996;

Sekar et al., 1999). Mechanistic studies have demonstrated that the catalysis by

secretory PLA2s does not take place via the formation of classical acyl enzyme

intermediate of serine proteases (Dennis, 1994). Instead, secretory PLA2s use histidine

assisted by the aspartate residue to polarize the catalytic water molecule. The binding of

Ca2+ ion in the active site is required for the stabilization of the tetrahedral geometry of

the transition state analogue.


8

Figure 1.2.3 Schematic representation of the catalytic mechanism of PLA2. (a) Catalytic attack on substrate bound in a productive mode. (b) The tetrahedral intermediate as it collapse into products. (c) The products formed by productive release. Three water molecules move into the active site of (as indicated by the arrows) to replace the products. One will engage the Nδ1 of His48 and the remaining two will coordinate the calcium ion.


9

1.2.5 TYPES OF PLA2 PLA2s have been grouped into many types depending upon their localization

and the requirement for the Ca2+ ion. On the basis of localization, they have been

divided in two groups (i) secretory PLA2 (sPLA2, extracellular) and (ii) cytosolic PLA2

(cPLA2, intracellular; Six and Dennis, 2000). Furthermore, depending upon their

requirement for Ca2+ ion, these enzymes have been classified into two groups (i) Ca2+-

dependent PLA2s and (ii) Ca2+-independent (iPLA2s; Dennis, 1994). In the recent past,

several new members of the PLA2 superfamily have been discovered (Hiraoka et al.,

2002; Jenkins et al., 2004; Ohto et al., 2005). Generally, sPLA2s are of low molecular

weight (~13-19 kDa), whereas cPLA2s are of high molecular weight (~80 kDa).

Cytoplasmic PLA2s (cPLA2s) are often membrane associated and are involved in

phospholipid metabolism, signal transduction and other varied essential cellular

functions (Mukherjee et al., 1994). iPLA2s, like the cPLA2s, utilize a serine for

catalysis (Schaloske and Dennis, 2006). sPLA2s are found in numerous organisms

including mammalian tissues, plants, insects, mollusks, reptiles, fungi, bacteria and

parvovirus (Schaloske and Dennis, 2006). They are expressed and secreted into the

extracellular milieu of tissues by secretory processes including secretion into the

gastrointestinal tract.

1.2.6 CLASSIFICATION OF SECRETORY PLA2 Secretory PLA2s (sPLA2s) have further been divided into many groups based on

their amino-acid sequence, disulfide pattern and the catalytic dyads and various other

structural and functional features (Schaloske and Dennis, 2006; Table 1.2.1). sPLA2s

contain about 120-170 amino acids in a single polypeptide chain and among themselves

differ significantly in their sequence identity (~20-50%). Although the three-

dimensional structures have several features in common, they differ in their disulfide

architecture, chain deletion positions (60-70 loop) and insertions at the C-terminus

(Figure 1.2.4).


10

Table 1.2.1 Classification of sPLA2. Functional features and phospholipase activity are given only for mammalian sPLA2s (Boyanovsky and Webb, 2009; Burke and Dennis, 2009).

Group Source MW DB FF PA IA Cobras, Kraits 13-15 7 IB Human/Porcine/

Bovine pancreas 13-15 7 Digestion of dietary PLs, ES

formation, cell-contraction, migration

PG>PS>>PC

IIA Rattlesnakes, Human synovial

13-15 7 Acute phase, antibacterial, cell-proliferation

PG>PS>>PC

IIB Gaboon viper 13-15 6 IIC Rat/Murine

testis 15 8 N/D PG>>PC

IID Human/Murine pancreas/spleen

14-15 7 N/D PG~PC

IIE Human/Murine brain/heart/uterus

14-15 7 Antibacterial PG>PC

IIF Human/Murine Testis/embryo

16-17 6 Antibacterial PG>>PC

III Human/Murine/ Lizard/Bee

55 15-18

8

Antiviral

V Human/Murine 14 6 Antibacterial, antiviral PE>PC>PS X Human

spleen/thymus/ leukocyte

14 8 Secreted as proenzyme, antiviral, atherogenic, antibacterial

PC>PS

XIA Green rice shoots (PLA2-I)

12.4 6

XIB Green rice shoots (PLA2-II)

12.9 6

XII Human/Murine heart/kidney

19 7 Antibacterial PG>PS>>PC

XIII Parvovirus <10 0 XIV Symbiotic

Fungus/Bacteria 13-19 2

MW, Molecular weight (kDa); DB, Number of disulfide bonds; FF, Features and functions; PA, Phospholipase activity; ES, Eicosanoid; AA, Arachidonic acid; PG, Phosphatidylglycerol; PS, Phosphatidylserine; PC, Phosphatidylcholine; PE, Phosphatidylethanolamine; N/D, Not determined.


11

1.2.7 STRUCTURAL BIOLOGY OF SECRETORY PLA2 sPLA2s were the first type of PLA2 enzymes to be identified (Fairbairn, 1945).

The name sPLA2 was coined based on the high content of PLA2 in the synovial fluid of

patients with rheumatoid arthritis (Seilhamer et al., 1989). Over 150 sPLA2s have been

sequenced from diverse sources such as mammalian pancreas, synovial fluid, venoms

of reptiles, insects and mollusks (Heinrikson, 1991; Six and Dennis, 2000). sPLA2s are

characterized by their requirement for histidine in the active site, low molecular weight,

Ca2+-requirement (in mM range) for catalysis and the presence of six conserved

disulfide bonds with one or two variable additional disulfide bonds (Six and Dennis,

2000; Schaloske and Dennis, 2006). They contain a His-Asp catalytic dyad forming the

active center (DXCCXXHD) and a conserved Ca2+-binding loop (XCGXGG) that is

essential for the proper functioning of the enzyme (Dennis, 1994). Substrate hydrolysis

occurs due to the activation and orientation of a water molecule by hydrogen bonding

to the active-site histidine.

Figure 1.2.4 Schematic diagram of the conservation of the catalytic domain around the catalytic histidine and the Ca2+-binding loop in mammalian sPLA2s (Murakami and Kudo, 2002). Number of cysteines observed in the group and the chromosome number are given in Column1 and Column2, respectively.


12

1.2.7.1 Group IA sPLA2

One of the best-studied sPLA2 enzymes is the group IA (sPLA2GIA) from cobra

venoms. Several structures both in free and ligand-bound forms of sPLA2GIA have

been determined. These include sPLA2s from Naja naja atra (Scott et al., 1990b; White

et al., 1990), Naja naja naja (Fremont et al., 1993; Segelke et al., 1998; Dalm et al.,

2010), Bungarus caeruleus (Singh et al., 2001; Singh et al., 2005a,b; Le Trong and

Stenkamp, 2007), Ophiophagus hannah (Xu et al., 2003), Naja naja sagittifera (Singh

et al., 2003; Jabeen et al., 2005a,b,c,d; Jabeen et al., 2006) and Bothrops jararacussu

(Magro et al., 2005).

The overall fold of sPLA2GIA is of a typical sPLA2 structure with five α-

helices and two β-strands (Figure 1.2.5a). The active site residues and the hydrogen-

bonding network are also conserved (Figure 1.2.5b). These enzymes contain six

conserved disulfide bonds with an additional disulfide bridge. The catalytic mechanism

of these enzymes is pH dependent ranging around 7-9 (Burke and Dennis, 2009). In

addition to the primary (active-site) Ca2+ ion, some structures have shown the presence

of a secondary Ca2+ ion that may act as a supplementary electrophile (Scott et al.,

1990b). NMR studies of sPLA2GIA bound to inhibitor suggested a model for its

binding in the active site (Yu et al., 1990; Plesniak et al., 1995). sPLA2GIAs have been

found to be in dimeric and trimeric forms. In contrast to most of the sPLA2s,

sPLA2GIA enzymes hydrolyze zwitterionic substrate with equal preference to

negatively charged lipid surfaces (Adamich et al., 1979; Sumandea et al., 1999).


13

1.2.7.2 Group II sPLA2

The group II secreted phospholipase A2s (sPLA2GII) are 13-18 kDa protein

containing 120-125 amino acid residues and seven disulphide bridges including a group

II-specific disulfide linking Cys80 and a cysteine at the C-terminus (Kudo and

Murakami, 2002; Figure 1.2.4). The first crystal structure of sPLA2GII was solved from

Crotalus atrox (Keith et al., 1981). Subsequently, structures from Agkistrodon halys

blomhoffii (Tomoo et al., 1992; Wang et al., 1996; Zhao et al., 1998), Agkistrodon

piscivorus piscivorus (Han et al., 1997) and Homo sapiens (Scott et al., 1991; Schevitz

et al., 1995) have been determined.

Overall tertiary structure of sPLA2GII contains mainly three long helices, one β-

wing, an external loop and a calcium-binding loop (Figure 1.2.6a). The conformation of

Figure 1.2.5 (a) Overall tertiary structure of sPLA2GIA (PDB-id: 1POB; White et al., 1990). Secondary-structure elements and cysteines involved in disulfide bonds are labelled. Two calcium ions are shown as orange spheres. Transition state analogue (TSA) observed in the crystal structure of cobra-venom PLA2 is shown as ball-and-stick. (b) Active-site hydrogen-bonding network of sPLA2GIA. Water molecules (denoted as Ow) are shown as red spheres.


14

the Ca2+-binding loop is similar to that of sPLA2GIB (discussed later) rather than that

of sPLA2GIA (Wang et al., 1996).

Figure 1.2.6 (a) Overall three-dimensional structure of sPLA2GII (PDB-id: 1B4W; Zhao et al., 2000). Secondary-structure elements are labelled. The disulfide bonds are shown as ball-and-stick in yellow. Active-site residues along with the primary Ca2+ ion are shown as ball-and-stick and sphere, respectively. One ligand molecule β-OctylGlucose (BOG) observed in the crystal structure of basic sPLA2GII is shown as ball-and-stick. The quaternary structures (b) dimer and (c) tetramer of sPLA2GII.


15

The active-site residues His48, Asp49 and Asp99 are directly connected to the

channel (Figure 1.2.6a). Based on the structures of various complexes of sPLA2GII,

almost six ligand-binding subsites have been located (Singh et al., 2007). These

enzymes exist either as a monomer or dimer (Brunie et al., 1985; Tomoo et al., 1992).

The sPLA2GII enzymes are of three types (i) basic (ii) neutral and (iii) acidic. The basic

and neutral sPLA2GIIs are observed as homodimers (Zhao et al., 2000; Figure 1.2.6b),

whereas acidic ones are found as monomer in solution (Wang et al., 1996). In addition,

the basic sPLA2GII was observed as tetramer in the crystal structure (Figure 1.2.6c),

however, it was suggested to be a possible crystallization artifact (Zhao et al., 2000).

The observations that these isozymes have different oligomerization and association

properties imply their different behavior towards the aggregated substrates (Murkami

and Kudo, 2002).

1.2.7.3 Group III sPLA2

Class III sPLA2s (sPLA2GIII) include the evolutionary divergent venom

enzymes from the European honeybee (Apis mellifera, Kuchler et al., 1989), the Gila

monster (Heloderma suspectum, Gomez et al., 1989) and the Mexican beaded lizard

(Heloderm horridum horridum, Sosa et al., 1986). Most sPLA2GIII enzymes are of low

molecular weight, except for the enzyme from humans (55 kDa) which consists of three

domains, of those the central domain displays all the features of group III bee venom

sPLA2s including ten cysteines, the key residues of Ca2+-binding loop and catalytic site

(Valentin et al., 2000). The primary structure of sPLA2GIII from honeybee shows little

sequence identity to other small sPLA2s (~20%) except in the region of the catalytic

dyad residues, the calcium-binding loop and certain cysteine residues (Kuchler et al.,

1989). Remainder of the primary sequence as well as the overall tertiary structure is

distinct (Figure 1.2.7a) and contains five intramolecular disulfide bonds. However, the

active site geometry is similar to those of other groups of sPLA2 (Scott et al., 1990a;

Figure 1.2.7b). Compared to its class I/II relatives, sPLA2GIII enzymes are relatively

insensitive to the aggregation state of its substrate and hydrolyze dispersed and

aggregated substrates at similar high rates (Raykova and Blagoev, 1986; Lin et al.,

1988). The essential calcium ion probably has two functions (i) to stabilize the

oxyanion of the putative tetrahedral intermediate derived from the carbonyl oxygen of


16

the substrate and (ii) to control the stereochemistry of productive substrate binding

(Scott and Sigler, 1994a,b).

1.2.7.4 Group V sPLA2

The gene sPLA2GV and its product have been identified in human (Chen et al.,

1994a) and rat (Chen et al., 1994b). The human gene sPLA2GV is located on

chromosome 1, in close proximity to the gene of homologous sPLA2GII, implying

coordinated regulation of their expression and has been postulated that sPLA2GII and

sPLA2V have emerged from gene duplication events. The mature enzyme is distinct

from other sPLA2s in that it contains only 12 cysteines instead of 14 and lacks both the

elapid loops (residues 59-70) of group sPLA2GI and the carboxyl extension of

sPLA2GII. Structural and functional information on sPLA2GV has been scarce due to

the difficulty in obtaining a sufficient amount of the protein in pure form. So far,

sPLA2GV has not been purified from natural sources and no tertiary structure has been

Figure 1.2.7 (a) Overall tertiary structure and the active-site geometry of sPLA2GIII (PDB-id: 1POC; Scott et al., 1990a). The secondary-structure elements are labelled. Active-site calcium ion and transition state analogue (GEL; 1-o-octyl-2-Heptylphosphonyl-sn-glycero-3- phosphoethanolamine) are shown as sphere and as stick model, respectively. Disulfide bonds are also shown as stick model and cysteines are numbered. (b) The hydrogen-bonding network observed in the active site of bee venom PLA2. The structural water molecule (labelled as Sw) and other water molecule (labelled as HOH) are shown as red spheres.


17

reported. However, a theoretical (homology) model structure of sPLA2GV (PDB-id:

2GHN; Winget and Bahnson, unpublished work) is available in the Protein Data Bank.

The tertiary structure of sPLA2GV is identical to sPLA2GIB except a disulfide bond

between Cys11-Cys77 specific to sPLA2GIB.

1.2.7.5 Group X sPLA2

A new 13.6 kDa acidic sPLA2 was isolated from human fetal lung (Cupillard et

al., 1997). The tertiary structure and active site architecture of human sPLA2GX are

similar to other sPLA2s (Pan et al., 2002; Figure 1.2.8a). However, differences are seen

at the N- and C-termini and at the pancreatic and elapid loops of snake PLA2 enzymes.

Furthermore, the electrostatic surface potential of the interfacial-binding regions of

sPLA2GX and sPLA2GII are respectively highly neutral and cationic (Pan et al., 2002;

Figures 1.2.8b,c).

Figure 1.2.8 (a) Cartoon representation of overall tertiary structure of sPLA2GX (PDB-id: 1LE6; Pan et al., 2002). Active-site residues, disulfide bonds and the primary calcium ion are shown as stick and sphere, respectively. For clarity, cysteine residues are numbered. (b) and (c) Comparison of active-site electrostatics potentials of sPLA2GX and sPLA2GII (PDB-id: 1POD; Scott et al., 1991), respectively.


18

1.2.7.6 Group XI sPLA2

Group XI sPLA2 (sPLA2XI) differs in its sequence and structure when

compared to other classes of sPLA2s (Stahl et al., 1999; Guy et al., 2009; Figure

1.2.9a). The N-terminal half of the chain contains mainly loop structure, including the

conserved Ca2+-binding loop. The C-terminal half is folded into three anti-parallel α-

helices, of which the first two are present in other secreted PLA2s and contain the

conserved catalytic histidine and Ca2+-coordinating aspartate residues. These enzymes

contain six disulfide bonds. The water structure around the Ca2+-binding site suggests

the involvement of a second water molecule in hydrolysis, a water-assisted calcium-

coordinate oxyanion mechanism (Edwards et al., 2002). The crystal structure shows

that His61 is held in its proper orientation by interacting with the side-chain oxygen

atom of Asn78, which substitutes the aspartate residue in the catalytic dyad His-Asp of

other sPLA2s (Figure 1.2.9b). Asn78 is replaced by a serine residue in some of the other

plant sPLA2 enzymes (Guy et al., 2009). The substitution of serine to alanine or

aspartate in the Arabidopsis thaliana sPLA2 results in considerable loss of activity

(Mansfeld et al., 2006), confirming the importance of this interaction.

Figure 1.2.9 (a) Three-dimensional structure of sPLA2GXI (PDB-id: 2WG9; Guy et al., 2009). The disulfide bonds and active-site residues are shown as stick. The primary Ca2+

ion is shown as orange sphere. (b) The active-site hydrogen-bonding network of sPLA2GXI along with ligand OCA (octanoic acid) observed in the crystal structure.


19

1.2.7.7 Group XII sPLA2

sPLA2GXII is a 19 kDa enzyme containing a central catalytic domain with a

His-Asp catalytic dyad, yet the location of cysteines outside the catalytic domain is

distinct from that of other sPLA2s (Hattori et al., 1994). Furthermore, in the consensus

segment of the Ca2+-binding loop (X1CG1X2G2), the G2 residue is replaced by proline

in sPLA2GXII. Crystal structure of sPLA2GXII is not available.

1.2.7.8 Group XIII sPLA2

This group of enzymes is found in parvovirus and shows very low sequence

similarity to other sPLA2s and is mainly restricted to the catalytic site residues histidine

and aspartate and the calcium-binding motif CXG. The viral PLA2 motifs lack

cysteines, unlike all other previously characterized sPLA2s, and the long loops between

the α-helices that contain the active-site residues of classical sPLA2s (Zadori et al.,

2001). No crystal structure of sPLA2GXIII is available.

1.2.7.9 Group XIV sPLA2

The primary structure of the prokaryotic PLA2 (sPLA2GXIV), except for

residues Cys61 to Tyr68, is distinct from that of eukaryotic sPLA2s. In contrast to

eukaryotic sPLA2s, the bacterial enzyme contain two disulfide bonds and shows all

helical structure (Figure 1.2.10a). However, the geometry of the catalytic site is

conserved (Matoba and Sugiyama, 2003; Figure 1.2.10b). Although the C-terminal α-

helix of sPLA2GXIV corresponds to the N-terminal α-helix in the other sPLA2s, its

orientation is opposite. Though, the orientation of three long α-helices in the C-

terminal domain of the sPLA2GXIV is similar to that of sPLA2GIII (bee venom), the

overall structure is different (Figure 1.2.10a).


20

1.2.8 QUATERNARY STRUCTURE The kinetic steps of lipid hydrolysis by PLA2s of phospholipids aggregates are

preceded by an initial lag phase and several models have been proposed to explain this

phenomenon. This latency has been shown to be accompanied by dimerization due to

the autocatalytic transfer of a substrate derived acyl group (Tomasselli et al., 1989).

Several sPLA2s exist in solution as stable multichain complexes (Chang and Lee, 1963;

Fohlman et al., 1976; Ho and Lee, 1981; Su et al., 1983; McIntosh et al., 1995). Kinetic

studies suggest that optimal enzymatic interfacial activity requires sPLA2 to form dimer

or higher-order oligomers (Cho et al., 1988; Bell and Biltonen, 1989a,b; Tomasselli et

al., 1989). Experiments with immobilized enzymes support the above statement

(Ferreira et al., 1994). The first venom protein to be crystallized was the heterodimeric

protein crotoxin from Crotalus durissus terrificus (Slotta and Fraenkel-Conrat, 1938).

Since then, several dimeric (Brunie et al., 1985; Arni et al., 1995) and trimeric (Hazlett

and Dennis, 1985; Fremont et al., 1993) PLA2s have been observed (Figure 1.2.11).

Figure 1.2.10 (a) Overall tertiary structure of sPLA2GXIV (PDB-id: 1KP4; Matoba et al., 2002) along with disulfide bonds, active-site residues and the primary calcium ion are shown. For clarity, cysteine residues involved in disulfide bonds are numbered. (b) Active-site hydrogen-bonding networking of sPLA2GXIV. Water molecules are shown asred spheres.


21

Recently, crystal structures of sPLA2GIA suggest the presence of the dimers induced

by metal ions (Jabeen et al., 2005c,d; Jabeen et al., 2006) and carbohydrate molecules

(Singh et al., 2005b). However, the crystallographic data suggest that the dimer

interface actually impedes access to the hydrophobic pocket (Brunie et al., 1985; Jain et

al., 1991).

1.2.9 STRUCTURAL BIOLOGY OF BOVINE PANCREATIC PLA2 The first non-venom sPLA2 named GIB (Group IB) was isolated from the

pancreatic juices of cows and was also found in many other animals (Puijk et al., 1977;

Seilhamer et al., 1986). Several crystal structures of sPLA2GIB from porcine (PPLA2)

and bovine (BPLA2) pancreases are available. Both the PPLA2 and BPLA2 share more

than 85% sequence identity and contain almost identical fold. More than 50% of these

structures have been studied for BPLA2 and many of them have been determined in this

laboratory. Hence, the general features observed in the BPLA2 structures are described

in the following sections.

The first crystal structure of BPLA2 was solved in 1978 (Dijkstra et al., 1978).

Following that several structures of apo, holo and mutant-enzymes have been

determined at high and ultra-high resolution (Dijkstra et al., 1981a,b; Renetseder et al.,

Figure 1.2.11 Quaternary structures of (a) sPLA2GII from Crotalus atrox (PDB-id: 1PP2; Brunie et al., 1995) and (b) sPLA2GIA from Naja naja (PDB-id: 1A3D; Segelke et al., 1998).


22

1988; Dupureur et al., 1992a, Steiner et al., 2001; Sekar et al., 2005). BPLA2 enzyme is

a small (14 kDa) compact kidney-shaped protein measuring approximately 22x30x42

Å3 (Figure 1.2.12a).

Roughly 50% of the residues are incorporated into α-helix and 10% into β-

sheet. Fourteen conserved cysteines form seven disulfide bridges that stabilize much of

the tertiary structure. BPLA2 (in general sPLA2GIB enzymes) has a unique five amino

acid extension termed ‘the pancreatic loop’ in the center of the molecule and a

sPLA2GI-specific disulfide between Cys11 and Cys77 (Verheij et al., 1981). The

His48-Asp99 pair forms the catalytic dyad with Ca2+ ion as the cofactor. The Ca2+ ion

is bound to five oxygen atoms provided by the protein (two oxygens of the Asp49 side

chain and backbone carbonyl oxygen atoms of Tyr28, Gly30, and Gly32) (Figure

1.2.12b). In addition, there are two coordinated water molecules (W5 and W12) which

Figure 1.2.12 (a) Overall tertiary structure of bovine pancreatic phospholipase A2 (PDB-id: 1UNE; Sekar and Sundaralingam, 1999). All three calcium ions (taken from mutant structure of BPLA2, PDB-id: 2B96) are shown as orange spheres. Active-site residues are shown as ball-and-stick. Chloride ion, observed almost in all BPLA2 crystal structures, is shown as cyan sphere. (b) Active-site hydrogen-bonding network observed in sPLA2GIB enzyme (PDB-id: 1UNE).


23

are replaced by the oxygen atoms of the sn-2 and sn-3 substituents of substrate mimics.

The catalytic residues are located at one end of the active site slot with hydrophobic

walls lined with Leu2, Phe5, Ile9, Ala102, Ala103 and Phe106.

1.2.9.1 Conserved substructures of BPLA2

Substructures of BPLA2 enzyme have been described here in brief.

The N-terminus

The N-terminus nitrogen is involved in a highly conserved network of hydrogen

bonds via a conserved structural water molecule (W11) and the side chain carboxylate

of the active-site Asp99 (Figure 1.2.13a). Disruption of this network by modification or

removal of terminal group impairs interfacial catalysis but has little effect on the kcat for

dispersed substrates (Dijkstra et al., 1984; Renetseder et al., 1985).

The N-terminal helix (residues 1-12)

The N-terminal helix is crucial for accommodating the acyl chains of

productively bound substrates in the active site (Scott et al., 1990b, 1991; Thunnissen

et al., 1990; White et al., 1990). The side-chains arising from this helix creates a lipid-

water interface during interfacial adsorption (Jain and Vaz, 1987; Ludescher et al.,

1988; Figure 1.2.13b).

The calcium-binding loop (25-33)

The functionally essential active-site Ca2+ ion is coordinated by a conserved and

flexible loop of residues with the consensus sequence YGCYCGXGG along with

oxygen atoms donated by the carboxylate of Asp49 and two water molecules and form

a tight pentagonal bipyramidal coordination cage (Strynadka and James, 1989; Figure

1.2.13c).

The anti-parallel helices (37-54 and 90-109)

Two long anti-parallel helices are crucial for accommodating the substrate and

contain half of the cysteines involved disulfide bonds. The conserved side chains

(Asp49) arising from these helices assist in Ca2+ coordination, the residues Cys45,

Ala102, Ala103, Phe106 form the deeper contours of the hydrophobic channel and the

residues His48, Tyr52, Tyr73, Asp99 create the catalytic network (Figures 1.2.12a,b).


24

The surface loop (60-70)

In most of the BPLA2 structures studied so far, the surface loop region (60-70)

is disordered (Sekharudu et al., 1992; Kumar et al., 1994; Huang et al., 1996; Sekar et

al., 1997a; Sekar et al., 1998a; Sekar et al., 1999). However, the loop is ordered in the

orthorhombic form of the enzyme (Sekar and Sundaralingam, 1999), in the presence of

second calcium ion (Rajakannan et al., 2002) and in the inhibitor bound structures

(Sekar et al., 1997b; Sekar et al., 1998b; Sekar et al., 2003; Sekar et al., 2006a,b;

Figure 1.2.13d).

The β-wing (74-84)

BPLA2 contains one well-developed β-wing oriented towards bulk solvent and

is known to play a role in anticoagulation (Kini and Evans, 1987; Figure 1.2.12a).

Disulfide bonds

In total, seven disulfide bonds (Cys11-Cys77, Cys27-Cys123, Cys29-Cys45,

Cys44-Cys105, Cys51-Cys98, Cys61-Cys91 and Cys84-Cys96) are found in the

BPLA2 enzyme (Figure 1.2.12a). BPLA2 displays remarkable stability against

denaturation ( OHdG 2Δ of 9.6 kcal mol-1; Zhu et al., 1995). However, five out of the

seven double mutants (Cys-Cys to Ala-Ala) displayed relatively modest changes in the

conformational stability of the protein (Δ OHdG 2Δ of +1.9 to –2.9 kcal mol-1). Only one

mutant, C11A/C77A, exhibited a large decrease in Δ OHdG 2Δ (-6.2 kcal mol-1).


25

1.2.9.2 Site-directed mutagenetic studies

Many residues of BPLA2 that are highly conserved in the subfamily have been

investigated by site-directed mutagenesis in conjunction with structural and functional

analysis as summarized in Table 1.2.2.

Figure 1.2.13 Conserved substructures of sPLA2GIB. (a) The illustration of the hydrogen-bonding networks which stabilizes the N-terminus of class I/II PLA2 enzymes. A glutamine is almost invariant at position 4. (b) Role of the N-terminal helix in substrate stabilization and group I specific disulfide bond. (c) The calcium-binding loop in the crystal structure of BPLA2 (PDB-ids: 1G4I, warm pink; 2B96, lime green and 1KVY, light blue). Two water molecules coordinating the calcium ion in the ligand free structures are shown as spheres. The anisic acid (ANN) bound crystal structure of BPLA2(PDB-id: 2B96, lime green) is also shown. (d) The surface loop conformations observed in the native (PDB-ids: 1G4I, red; 1UNE, green; 1VL9, cyan and 1MKT, wheat), the presence of second calcium ion (PDB-ids: 1GH4, orange) and the ligand-bound (PDB-ids: 1MKV, blue; 1O2E, yellow and 1FDK) crystal structures of BPLA2.


26

Table 1.2.2 Site-directed mutagenetic studies of BPLA2.

Mutants Observations and conclusion Active site

H48(A, N, Q)1,2 Nδ1 of His48 plays catalytic role while Nε2 of His48 is structurally essential.

Calcium binding residues

D49(A, E, K, N, Q)1,3, G30S4

Carboxylate oxygen atoms of Asp49 are important for Ca2+ binding. The metal Ca2+ binding is obligatory for the substrate binding to the active site of the enzyme.

Hydrogen bonding network Y52(F, K, V)5,6, Y73(A, F, K, S)5,6, D99(A, N)5,7,8, Y52F/Y73F(double)5,6, Y52F/Y73F/D99N(triple)9

The hydrogen-bonding network involving Tyr52 and Tyr73 is crucial for any form of the interfacial activation. Asp99 is structurally essential for anchoring the His-Asp catalytic dyad.

Disulfide bond (C to A)10

C11-C77, C27-C123, C29-C45, C44-C105, C51-C98, C61-C91, C84-C96

Disulfide bond C11-C77 is very crucial for the conformational stability of the enzyme. Deletion of C27-C123 disulfide bond showed 2.4 kcal mol-1 increase in stability. C84-C96 bond is important for protein folding.

Hydrophobic channel L2(A, R, W)11, F5(A, V, W, Y)11, I9(A, F, S, V, Y)11, F5V/I9F (double)11, F22(A, I, Y)12, F106(A, I, Y)12

These residues play major role in substrate binding. In particular, substitutions of Leu2 showed an acyl chain length discrimination toward different substrates.

Interfacial binding site W3A, Q4(A, E, K, N)11, N6(A, D)11, K10E13, L19(K, S, W)14, L20(K, S, W)14, K53(E, M)13,15, K56(E, F, I, M, N, Q, R, T)14-16, K116 (E, K)13, K120A/K121A, Δ115-123/C27A17, K53M/K56M (double), K120A/K121A (double), K53M/K56M/K120M (triple), K53M/K56M/K121M (triple), K53M/K120M/K121M (triple), K56M/K120M/K121M (triple), K53M/K56M/K120M/K121M (quadruple)18-22,

Interfacial binding is likely to be governed by electrostatic as well as hydrophobic interactions. Since the interfacial-binding site includes a large number of residues, the effect of point mutations is very modest. Crystal structures of triple mutants (K53M/K120M/K121M and K53M/K56M/K121M) and quadruple mutants (K53M/K56M/K120M/K121M) revealed the binding of second calcium ion and dynamics of surface loop which correspond to the solution and membrane bound state of the enzyme.

Surface loop Δ59,62,64-67/S60G/V63Y23

The deletion of surface loop causes conformational change in the active site of BPLA2. Val63 aids in helix stabilization.

1. Sekar et al., 1999; 2. Li and Tsai, 1993; 3. Li et al., 1994; 4. Bekkers et al., 1991; 5. Sekharudu et al., 1992; 6. Dupureur et al., 1992a; 7. Kumar et al., 1994; 8. Sekar et al., 1999; 9. Sekar et al., 1997a; 10. Zhu et al., 1995; 11. Liu et al., 1995; 12. Dupureur et al., 1992b; 13. Dua et al., 1995; 14. Lee et al., 1996; 15. Rogers et al., 1998; 16. Noel et al., 1991; 17. Huang et al., 1996; 18. Yu et al., 2000; 19. Rajakannan et al., 2002; 20. Sekar et al., 2003; 21. Sekar et al., 2005; 22. Sekar et al., 2006a; 23. Kimura et al., 1990.


27

1.2.9.3 Role of His48 and Asp49

Residue His48 along with Asp99 plays a key role in the enzymatic activity and

is highly conserved among the sPLA2 enzyme (Figure 1.2.14). NMR and

conformational analyses of His48 mutants have shown that the catalytic dyad, which is

retained in H48Q mutant but not in H48N or H48A mutants, is important in

maintaining the conformational integrity of PLA2 (Li and Tsai, 1993). Site-directed

mutagenesis of His48 also shows the significantly reduced catalytic activity of the

enzyme (Janssen et al., 1999). The H48Q and H48N mutants of the BPLA2 enzyme

have less than 0.00014% and 0.006%, respectively, of the wild-type activity. These

studies also suggested that the H48N mutant shows relatively higher activity than

H48Q and H48A (Li and Tsai, 1993). However, both the mutants (H48N and H48A)

are as stable as the wild type enzyme ( OHdG 2Δ decreases from 9.6 kcal mol-1 for WT to

6.3 kcal mol-1; Yuan and Tsai, 1999). Furthermore, the amide nitrogen atoms of Asn

and Gln could mimic the Nδ1 and Nε2 of His48, respectively. On the other hand, two-

dimensional NOESY spectra suggests that the global conformation is largely retained

in H48Q but not in H48N or H48A (Li and Tsai, 1993). In human sPLA2GII, H48Q

shows considerable activity whereas H48N mutant has significant activity (Edwards et

al., 2002). These results suggest that the hydrogen bond owing to Asp99 to amide

nitrogen atom of His48 plays an important structural role.

Figure 1.2.14 Multiple sequence alignment of those sPLA2s which contain seven disulfide

bonds. The fully conserved residues are presented in red boxes.


28

The critical role for Asp49 in primary calcium binding, first suggested by

chemical modification studies (Fleer et al., 1981), is confirmed by mutagenesis studies

(van den Bergh et al., 1988; Li et al., 1994). Site-directed mutagenesis and structural

studies on Asp49 mutants provide insights into the structural and functional roles of the

highly conserved residue Asp49 and observed that the mutants D49N and D49K do not

bind to the calcium ion (Li et al., 1994). However, the D49E mutant binds to the

calcium with 12-fold weaker binding affinity and the specific catalytic activities of the

mutant enzymes decrease significantly ranging from 5.4 x 102 to 5.8 x 105 fold in

comparison with that of the wild-type enzyme (Li et al., 1994; Sekar et al., 1999). The

crystal structure of D49E mutant indicates that Ca2+ is coordinated to only one of the

carboxylate oxygen atoms of Glu, resulting in only four ligands (instead of five) from

the protein (Sekar et al., 1999). On the other hand, these mutants (D49N, D49E, D49Q,

D49K and D49A) fully retain the affinity for binding to the surface of zwitterionic

micelles and anionic vesicles, the conformation and stability. Together these

observations suggest that Ca2+ ion is required for the binding of the ligand to the active

site (Yu et al., 1993), but not for the stabilization of the overall structure of the enzyme

(Li et al., 1994).

1.2.9.4 Role of divalent Ca2+ ion

Calcium ion is essential for sPLA2 catalysis (Scott and Sigler, 1994b). Three

calcium ions, including the primary (active-site) Ca2+ ion, have been observed in the

crystal structure of BPLA2 determined so far (Dijkstra et al., 1978; Rajakannan et al.,

2002; Sekar et al., 2006b; Figure 1.2.12a). The primary calcium ion is an essential

cofactor for enzymatic actions (Scott et al., 1990a,b). It directs the stereospecific

positioning of the substrate molecule in the active site and serves as an electrophile

during general base-mediated catalysis. In the ligand-free enzyme, the primary calcium

ion is arranged in pentagonal bipyramidal geometry. It is coordinated by five protein

atoms (three backbone oxygen atoms of Tyr28, Gly30 and Gly32 and both carboxylate

oxygen atoms of Asp49) and two water (equatorial, Weq/W5 and axial, Wax/W12)

molecules (Figure 1.2.13c). Weq bridges the primary calcium ion and the catalytic water

(Wcat/W6) molecule (Figure 1.2.12a). Wax is exposed to the solvent channel (Figure

1.2.13c). There is a strong hydrogen-bonding network among the coordinating ligands


29

of the primary calcium ion and the conserved structural water molecule (W11) through

His48 and Asp99 (Figure 1.2.12b). During the catalysis, oxygen atoms of substrate

molecule replace both equatorial and axial water molecules.

In addition to the catalytic calcium ion, a second Ca2+ ion with 10-fold lower

affinity has also been observed in several sPLA2 enzymes (Slotboom et al., 1978;

Andersson et al., 1981; White et al., 1990; Scott et al., 1991; Rajakannan et al., 2002).

From NMR and UV spectroscopy, it was concluded that the second Ca2+-binding site

must be located close to the N-terminus of the polypeptide chain (Slotboom et al.,

1978). The role of the second calcium ion was inferred that the binding of calcium at

this site might be important in maintaining the surrounding residues in a position

favorable for substrate binding (Scott et al., 1991; Scott and Sigler, 1994b). In the

crystal structure of BPLA2, the second calcium ion is coordinated by three protein

atoms (Oδ1 of Asn71, backbone oxygen atom of Asn72 and Oε2 of Glu92) and three

water molecules (Rajakannan et al., 2002). The role of the secondary calcium ion is

proposed to interact with organized lipid/water interfaces (Sekar et al., 2006a).

Recently, third calcium ion has also been observed in the crystal structure of

inhibitor-bound BPLA2 enzyme (Sekar et al., 2006b). It is coordinated by four protein

atoms (two terminal oxygen atoms of Cys123 and two oxygen atoms of Tris molecule),

one water molecule and three ligands from crystallographic symmetry-related molecule

(Oδ1 of Asn112 and two water molecules; Sekar, 2007). However, the functional role of

the third calcium ion is not yet understood.

1.2.9.5 Role of water molecules in BPLA2

Water molecules in general play a pivotal role in governing biomolecules, aid in

stabilizing the three-dimensional architecture, dynamics and function (Cheung et al.,

2002; Halle, 2004; Papoian et al., 2004; Eisenmesser et al., 2005; Smolin et al., 2005).

A notable involvement of water molecules is their participation in many hydrogen-

bonding networks (Meyer, 1992). Most of the crystal structures of BPLA2 solved at

higher resolution (better than 2.0 Å) contain more than 100 water molecules per

subunit. Out of which, four water molecules have been shown to be structurally and/or

functionally important (Sekar and Sundaralingam, 1999). Water molecule W6 (Wcat)

plays the role of a nucleophile during the enzyme catalysis process (Steiner et al., 2001;


30

Sekar et al., 2005). Two water molecules W5 (Weq) and W12 (Wax) are crucial for

calcium positioning in the active site (Scott et al., 1990b). The water molecule W11 is

hydrogen bonded to the N-terminal residue Ala1 and one of the catalytic dyad residues

Asp99 is suggested to be structurally important (Kumar et al., 1994; Sekar et al.,

1997a; Sekar and Sundaralingam, 1999). Thus, it was interesting to investigate the role

of other water molecules found to be invariant in BPLA2 structures.

CHAPTER 1: Moco BIOSYNTHESIS PROTEINS

31

1.3 PROTEINS INVOLVED IN MOLYBDENUM COFACTOR

BIOSYNTHESIS Structural and functional studies on proteins obtained from thermophilic

organisms have been recently initiated in the laboratory. These include structural

studies on enzymes involved in molybdenum cofactor (Moco) biosynthesis. In the

present thesis, structural studies on two proteins MoaC and MogA involved in Moco

biosynthesis have been carried out.

1.3.1 INTRODUCTION Molybdenum cofactor (Moco) is required for the activity of several enzymes

(collectively known as molybdoenzymes) such as sulfite oxidase (SO), xanthine

oxidoreductase (XOR), aldehyde oxidase (AO) and nitrate reductase (NR; Schwarz et

al., 2009). Molybdoenzymes contain molybdenum ligated into Moco (Hille, 2002a,b)

and are known to play major roles in nitrogen, carbon and sulfur cycles (Stiefel, 2002).

Molybdoenzymes are found nearly in all organisms except Sachharomyces (Zhang and

Gladyshev, 2008). However, many anaerobic archaea and some bacteria use tungsten

(W) instead of molybdenum for their growth (Bevers et al., 2009). As a result, the term

Moco refers to the utilization of both the metals. In addition, the metal selenium (Se)

has also been observed to be utilized by some prokaryotes (Zhang and Gladyshev,

2008). The recent study on molybdoenzymes suggested that almost all organisms are

found to either possess both Moco biosynthesis proteins and known molybdoenzymes

or lack them (Zhang and Gladyshev, 2008). It was observed that ~72% bacteria, ~95%

archaea and almost all eukaryotes utilize Moco. In contrast, all parasites, yeasts and

free-living ciliates lack Moco biosynthesis proteins and molybdoenzymes (Zhang and

Gladyshev, 2008). Moco deficiency is a human disease and leads to accumulation of

toxic levels of sulfite and which causes neurological damage usually leading to death

within months of birth due to the lack of active sulfite oxidase. In addition, a mutational

block in Moco biosynthesis causes absence of enzyme activity of molybdoenzymes.

Atleast 130 cases of Moco deficiency have been reported (Ichida et al., 2006).


32

1.3.2 MOLYBDENUM Molybdenum is the only second-row transition metal that is required by most

living organisms and those species that do not require it, use tungsten (Hille, 2002a).

The metal was first isolated in 1781 by Peter Jacob Hjelm (Anke and Seifert, 2007).

Molybdenum does not occur as the free metal in nature, rather it exists in various

oxidation states (+4 to +6). According to the recent theories, molybdenum incorporated

into molybdoenzymes was used as a catalyst by the single-celled organisms to break

atmospheric molecular nitrogen into atoms allowing biological nitrogen fixation

(Mendel and Bittner, 2006). This, in turn, allowed biologically driven nitrogen-

fertilization of the oceans and thus the development of more complex organisms. In

certain bacteria, the nitrogenase enzyme (involved in nitrogen fixation) usually contains

molybdenum in the active site though replacement of molybdenum with iron or

vanadium is also seen (Robson et al., 1986; Eady, 1995). Molybdenum concentration

affects protein synthesis, metabolism and growth (Mitchell, 2003). The human body

contains about 0.07 mg of molybdenum per kilogram of weight (Holleman and Wiberg,

2001) and it occurs mostly in the liver, kidneys and tooth enamel (Curzon et al., 1971).

The dietary sources of molybdenum include pork, lamb, beef liver, green beans, eggs,

sunflower seeds, wheat flour, lentils and cereal grains. However, animal studies have

shown that chronic ingestion of more than 10 mg/day of molybdenum can cause

diarrhea, growth retardation, infertility, low birth weight, gout and affects the lungs,

kidneys and liver (Coughlan, 1983; Barceloux and Barceloux, 1999; Turnlund, 2002).

1.3.3 MOLYBDENUM COFACTOR Though molybdenum forms complexes with various organic molecules like

carbohydrates and amino acids, it is transported throughout the human body as

molybdate (MoO2−4). Once molybdate enters the cell, it is incorporated into metal

cofactors through complex biosynthetic processes (Schwarz et al., 2009). In nature, two

very different types of cofactors are known to control the redox state and catalytic

power of molybdenum. The first type is the iron-sulfur-cluster-based cofactor denoted

as iron-Moco or FeMoco (Allen et al., 1994; Figure 1.3.1a) and the second type is

pterin-based cofactor denoted as Moco (Rajagopalan and Johnson, 1992; Figure

1.3.1b). Moco is a complex of molybdopterin (MPT) and an oxide of molybdenum. The


33

molybdenum in both types of cofactors is coordinated by sulfur and oxygen atoms i.e.

cluster of iron-sulfur atoms in FeMoco and ene-dithiolate in Moco. Moco is found in all

molybdoenzymes except nitrogenases where FeMoco is utilized (Burgess and Lowe,

1996). With only the exception of carbon monoxide dehydrogenase, which is a

binuclear (Dobbek et al., 2002), all other known prokaryotic and eukaryotic

molybdoenzymes are mononuclear molybdenum.

The first biochemical evidence for the existence of a cofactor common to all

molybdoenzymes was provided using the crude protein extract of the Neurospora

crassa nit-1 mutant (Sorger and Giles, 1965; Nason et al., 1971). The elucidation of the

chemical nature of Moco is based on the work of Rajagopalan and Johnson (1992). Due

to the labile nature of Moco and its high sensitivity to oxidation, most of the works are

done by using degradation or oxidation products of the cofactor and thereby revealing

the nature of Moco (Johnson et al., 1984). The atom types of the metal ligands in Moco

were demonstrated based on urothione structural similarity, the presence of sulfhydryl

groups and the oxidized state and carboxamidomethylation of Moco (Kramer et al.,

1987). The redox state of the pterin was proposed to be tetrahydro (fully reduced;

Kramer et al., 1987; Rajagopalan and Johnson, 1992). Subsequently, the crystal

structures of different molybdoenzymes bound with Moco confirmed its core structure

and redox state (Nieter Burgmayer et al., 2004). The task of the cofactor is to position

the catalytic metal molybdenum correctly within the active center to control its redox

behaviour and to participate with its pterin ring system in the electron transfer to or

from the metal (Kisker et al., 1997). Once Moco is liberated from the holoenzyme, it

loses molybdenum and undergoes rapid and irreversible loss of function due to

oxidation (Rajagopalan, 1996).

Figure 1.3.1 Chemical structures of (a) FeMoco and (b) Moco.


34

1.3.4 MOLYBDENUM COFACTOR BIOSYNTHESIS The first study of molybdenum metabolism was carried out using the genetic

analysis of mutants of the filamentous fungus Aspergillus nidulans (Cove and Pateman,

1963; Pateman et al., 1964). Subsequently, similar studies were described for

Escherichia coli (Glaser and DeMoss, 1971), Neurospora crassa (Tomsett and Garrett,

1980), Homo sapiens (Johnson et al., 1980), Drosophila melanogaster (Warner and

Finnerty, 1981) and plants (Muller and Mendel, 1989). Based on these studies, six

different genetic complementation groups have been identified and have provided the

basis for an evolutionary old multi-step biosynthetic pathway proposal (Mendel, 1992).

The first model of Moco biosynthesis was derived from E. coli studies for which five

Moco-specific operons are known (Rajagopalan and Johnson, 1992). In all organisms

studied so far, Moco is synthesized by a conserved biosynthetic pathway that can be

divided into five steps i.e. (i) conversion to GTP to cyclic pyranopterin monophosphate

(cPMP), (ii) cPMP to molybdopetrin (MPT), (iii) MPT to adenlyated MPT (MPT-

AMP), (iv) MPT-AMP to Moco and (v) Moco to bis-MGD (Figure 1.3.2). The last step

is found only in prokaryotes. The gene products catalyzing Moco biosynthesis have

been identified in plants (Mendel and Schwarz, 2002), fungi (Millar et al., 2001) and

humans (Stallmeyer et al., 1999a,b). In plants, genes and their products are named

according to the cnx nomenclature [cofactor for NR and xanthine dehydrogenase

(XDH)] and the cDNAs are labeled by numbers (cnx1-3, cnx5-7). In fungi, they are

labeled by letters (cnxA-F). In humans, it is named as MOCS (molybdenum cofactor

synthesis; Reiss et al., 1998). These genes are homologous to their counterparts in

bacteria and some but not all of the eukaryotic Moco biosynthesis genes are able to

functionally complement the matching bacterial mutants (Hanzelmann et al., 2002;

Table 1.3.1). In general, first two steps have been extensively studied in bacteria and

humans, whereas, third and fourth steps have mostly been studied in plants.

The present work involves the structural studies on proteins involved in the first

and third step of this pathway.


35

Figure 1.3.2 Schematic diagram of Moco-biosynthesis pathway. Enzymes and the metabolites required for the reactions are indicated by arrows. Steps are numbered in parenthesis. The last step is observed only in the case of prokaryotic organisms.


36

Table 1.3.1 Comparison of proteins involved in bacterial and eukaryotic Moco biosynthesis.

Step Bacteria (E. coli)

Plants (A. thaliana)

Fungi (A. nidulans)

Humans (H. sapiens)

1a MoaA [1,2] Cnx2 [3] CnxA [4] MOCS1A [5]

1b MoaC [6,2] Cnx3 [3] CnxC [4] MOCS1B [5]

2 MoaD [7,8] Cnx7 [9] CnxG [10] MOCS2A [11]

MoaE [12,8] Cnx6 [13] CnxH [10] MOCS2B [11]

MoeB [7,14] Cnx5 [15] CnxF [16] MOCS3 [17]

3 MogA [18,19] Cnx1 (C) [20,21] CnxE (N) [22] Geph (N) [23,20]

MoaB [24,25]

4 MoeA [26,27] Cnx1 (N) [21] CnxE (C) [22] Geph (C) [23,28]

5 MobA [29,30,31] - - -

MobB [32] $N, N-terminus; C, C-terminus. 1. Hanzelmann and Schindelin, 2004; 2. Wuebbens and Rajagopalan, 1995; 3. Hoff et al., 1995; 4. Unkles et al., 1997; 5. Reiss et al., 1998; 6. Wuebbens et al., 2000; 7. Lake et al., 2001; 8. Pitterle et al., 1993; 9. GeneBank: Cnx7 Accession Number-AF208343; 10. Unkles et al., 1999; 11. Stallmeyer et al., 1999a; 12. Rudolph et al., 2001; 13. GeneBank: Cnx7 Accession Number-AJ133519; 14. Rajagopalan, 1996; 15. Nieder et al., 1997; 16. Appleyard et al., 1998 ; 17. GeneBank: AF102544; 18. Liu et al., 2000; 19. Joshi et al., 1996; 20. Schwarz et al., 2001; 21. Stallmeyer et al., 1995; 22. Millar et al., 2001; 23. Stallmeyer et al., 1999b; 24. Rivers et al., 1993; 25. Bader et al., 2004; 26. Xiang et al., 2001; 27. Hasona et al., 1998; 28. Sola et al., 2004; 29. Lake et al., 2000; 30. Palmer et al., 1994; 31. Stevenson et al., 2000; 32. Palmer et al., 1996.

1.3.5 OPERONS INVOLVED IN MOLYBDENUM COFACTOR

BIOSYNTHESIS In E. coli, at least five operons namely moa, mob, mod, moe and mog, each

encoding for one or more genes, are known to be involved in the biosynthesis of Moco

(Shanmugam et al., 1992; Mendel and Schwarz, 2002; Figure 1.3.3). The operon moa

encodes for five genes, namely, moaA, moaB, moaC, moaD and moaE (Johnson and

Rajagopalan, 1987a,b; Rivers et al., 1993). Of these, two proteins MoaA and MoaC are

involved in the conversion of GTP to cPMP (also known as precursor Z; Pitterle and

Rajagopalan, 1989; Rieder et al., 1998). Another two proteins MoaD and MoaE form

protein-protein complexes and synthesize MPT (Pitterle et al., 1993).

The functional role of gene product MoaB is suggested to be adenylation of

MPT based on the study of its homologous protein Cnx1G in plants (Bevers et al.,

2008). Operon mob encodes at least two co-transcribed genes mobA and mobB, which

are involved in the maturation of Moco (Iobbi-Nivol et al., 1995). Operon mod encodes


37

at least five genes modA, modB, modC, modE and modF and most of them are involved

in the transport of molybdenum into the cell (Maupin-Furlow et al., 1995; Grunden et

al., 1996; Grunden and Shanmugam, 1997). Operon moe codes for two genes moeA and

moeB. MoeA has been proposed to be involved in molybdenum insertion into MPT to

generate Moco. The gene product of moeB plays a role of adenylation in the second

step of Moco biosynthesis (Zhang et al., 2010). Mog operon encodes the gene mogA,

which is involved in the adenylation process of MPT in an ATP- and Mg2+-dependent

reaction (Nichols and Rajagopalan, 2002, 2005).

1.3.6 MOLYBDOENZYMES

More than 50 molybdoenzymes are known to occur in bacteria while in

eukaryotes, only six are found (Sigel and Sigel, 2002; Schwarz and Mendel, 2006). The

eukaryotic molybdoenzymes are subdivided into two classes (1) xanthine

oxidase/oxidoreductase (XO/OR) family containing xanthine dehydrogenase (XDH),

aldehyde oxidoreductase (AOR), pyridoxal oxidase and nicotinate hydroxylase and (2)

sulfite oxidase (SO) class formed by sulfite oxidase (SO), dimethyl sulfoxide reductase

(DMSOR) and nitrate reductase (NR; Hille, 1996; Kisker et al., 1997, 1998). While

Figure 1.3.3 Schematic representation of the organization of operons and their gene products found in E. coli K-12 substrain DH10B. The left and right arrows denote the complement and forward directions, respectively. The genes surrounding each operon are in red.


38

pyridoxal oxidase and nicotinate hydroxylase were exclusively found in Drosophila

melanogaster (Warner and Finnerty, 1981) and Aspergillus nidulans (Lewis et al.,

1978), respectively, XDH, AO, and SO are typical for all eukaryotes analyzed so far.

As NR is required for nitrate assimilation, it is present in autotrophic organisms like

plants, algae and fungi. In bacteria, except for the AOR family, other molybdoenzymes

are widespread (95%, 69% and 67% for DMSOR, SO and XO, respectively). In

archaea, members of the DMSOR family are found in all molybdenum-utilizing

organisms, whereas other families (AOR, SO and XO) were found only in half the

organisms. In contrast to prokaryotes, eukaryotes contain only two molybdoenzyme

families (SO and XO). However, no Moco utilization trait and molybdoenzymes are

found in yeast Saccharomycotina (Zhang and Gladyshev, 2008).

Molybdoenzymes hold key positions both in the biogeochemical redox cycles of

carbon, nitrogen and sulfur on Earth and in the metabolism of the individual organism

(Stiefel, 2002). In mammals, SO is the most important molybdoenzyme, which

catalyzes the last step in the degradation of sulfur-containing amino acids and sulfatides

(Kisker et al., 1997). Very similar to SO is eukaryotic NR found in autotrophic

organisms where it catalyzes the first and rate-limiting step in nitrate assimilation

(Campbell, 2001). XO catalyzes the oxidation of hypoxanthine to uric acid and the

catabolism of purines in some species, including humans (Harrison, 2002). In plants, it

plays a part in cellular processes like plant-pathogen interactions between

phytopathogenic fungi and legumes or cereals (Montalbini, 1992a,b), cell death

associated with hypersensitive response (Montalbini and Della Torre, 1996) and natural

senescence (Pastori and Rio, 1997) besides purine degradation. AO is a cytoplasmic

enzyme that catalyzes the process of carboxylic acids generation from aldehydes

(Koshiba et al., 1996). Remarkably, all the eukaryotic molydoenzymes are

homodimers, which depends on the presence of Moco (Figure 1.3.4). They harbor an

electron transport chain from or to the substrate involving different prosthetic groups

(FAD, heme, Fe-S clusters) (Kisker et al., 1997; Hille, 2002b).


39

1.3.7 PHYSIOLOGICAL ROLES OF MOLYBDENUM AND

MOLYBDENUM COFACTOR Biologically, molybdenum belongs to the group of trace elements and

organisms require it only in minimum amounts. In case of high intake of molybdenum,

toxicity symptoms are observed (Turnlund, 2002). On the other hand, unavailability of

molybdenum is lethal for the organism. For higher organisms like humans and plants, a

shortage of molybdenum in nutrition or a mutational block of the cellular ability to use

it i.e. to synthesize MPT, to take up molybdenum into the cell or to incorporate

molybdenum to MPT, leads to the loss of essential metabolic functions because all

molybdoenzymes lose their activity at the same time (Duran et al., 1978; Johnson and

Duran, 2001). Babies born with this defect show feeding difficulties, severe and

progressive neurologic abnormalities and dysmorphic features of the brain and head

(Reiss and Johnson, 2003). So far, disease-causing mutations have been identified in

three out of the four known Moco-synthetic human genes: mocs1, mocs2 and gephyrin

(Schwarz et al., 2009; Figure 1.3.5). The clinical symptoms may result from the

deficiency of SO that protects the organism, particularly in the brain, from elevated

Figure 1.3.4 Schematic diagram of domain structure of plant (Arabidopsis thaliana) molybdoenzymes. The Moco domains of NR/SO and XDH/AO are not significantly homologous. Additional redox active domains (FAD, heme b5, Fe-S cluster) involved in electron transfer are also shown. In NR and SO, ‘dimer.’ denotes the domain responsible for dimerization.


40

levels of toxic sulfite (Smolinsky et al., 2008). Moco deficiency cannot be treated by

supplementation with the cofactor Moco because it is extremely unstable outside the

protecting environment of an apo-molybdoenzymes (Kramer et al., 1984). In addition,

no chemical synthesis of Moco or any of its intermediates have been successful so far.

Recently, Mendel and Bittner group developed a model similar to the precursor Z

compound that could lead to the cure of Moco-deficiency (Mendel and Bittner, 2006).

In addition, molybdenum is known to be directly or indirectly involved in avoiding

bone and tooth decay (Adler and Straub, 1953), in Wilson's disease (Mendel and

Bittner, 2006) and in xanthinuria disease (Ichida et al., 2001). In plants, the deficiency

of Moco-sulfurase causes the reduction of abscisic acid levels due to the lack of AO

activities (Xiong et al., 2001).

1.3.8 STRUCTURES AND FUNCTIONS OF PROTEINS INVOLVED IN

Moco BIOSYNTHESIS PATHWAY 1.3.8.1 Conversion of GTP to cPMP

Several pteridines such as biopterin have three-carbon side chains and are

synthesized using two pathways that begin with the conversion of GTP by the enzymes

Figure 1.3.5 Classification of Moco-deficient patients according to the three distinguishable steps in human Moco biosynthesis. Type-A patients cannot form cPMP, whereas type-B patients accumulate cPMP, which is excreted in the urine. So far, only one type-C patient has been described with a deletion of gephyrin due to an early stop codon in the gephyrin gene, a protein also needed for the formation of inhibitory synapses.


41

cyclohydrolase I and II (Thony et al., 2000; Bacher et al., 2001; Rebelo et al., 2003).

On the other hand, MPT is the only four-carbon side chain substituted pterin known so

far which is synthesized using the third route that also starts with GTP (Hanzelmann

and Schindelin, 2004). Based on labelling studies in E. coli, the generation of cPMP

from GTP has been suggested (Santamaria-Araujo et al., 2004). cPMP is the most

stable intermediate generated in the Moco biosynthesis pathway (Wuebbens and

Rajagopalan, 1993). In this step, the C8 atom of the purine base is inserted between the

2' and 3' ribose carbon atoms (Figures 1.3.6a,b), thus forming the four-carbon atoms

(all derived from the ribose) of the pyrano ring (Wuebbens and Rajagopalan, 1995;

Rieder et al. 1998), whereas it is released in the reaction catalyzed by GTP

cyclohydrolases. Under experimental conditions, precursor Z is rapidly oxidized by air

or iodine to 2'-alkyl pterin, which is termed compound Z (Johnson et al., 1989b; Figure

1.3.6c). Functional characterization of proteins involved in the first step of Moco

biosynthesis started with human MOCS1A and MOCS1B (MoaA and MoaC in E. coli,

respectively; Hanzelmann et al., 2002).

Structure of MoaA

MoaA contains Fe-S clusters that are bound via highly conserved cysteine

residues and shows sequence similarities to a variety of proteins including biotin

synthase, pyruvate formate lyase and anaerobic ribonucleotide reductase (Menendez et

Figure 1.3.6 Schematic diagram of precursor Z according to (a) Wuebbens and Rajagopalan, 1993 and (b) Santamaria-Araujo et al., 2004. (c) Oxidation of precursor Z/cPMP to compound Z.


42

al., 1996). MoaA and its homologues belong to the family of S-adenosylmethionine

(SAM)-dependent radical enzymes. Members of this large family catalyze the

formation of protein and/or substrate radicals by reductive cleavage of SAM by a [4Fe-

4S] cluster (Sofia et al., 2001). MoaA and homologues contain a conserved double

glycine motif at the C-terminus. Deletions or mutations in this motif result in the loss of

function (Hanzelmann et al., 2002). In archaea, the synthesis of cPMP does not depend

on this motif (Hanzelmann and Schindelin, 2004).

Crystal structure of MoaA from Staphylococcus aureus is available both in

ligand-free and ligand bound forms (Hanzelmann and Schindelin, 2004, 2006). The

core of the protein is characterized by an incomplete (αβ)6 triosephosphate isomerase

barrel (TIM; Murzin et al., 1995), which binds to the N-terminal [4Fe-4S] cluster

typical for SAM-dependent radical enzymes (Frey and Magnusson, 2003; Jarrett, 2003;

Figure 1.3.7a). The lateral opening of the incomplete barrel is covered by the C-

terminal part containing a second, MoaA-specific [4Fe-4S] cluster (Hanzelmann et al.,

2004). The N-terminal Fe-S cluster is coordinated by SAM and the substrate GTP.

MoaA exists both as monomer and dimer in solution (Figure 1.3.7b), but the functional

significance of the dimer is not yet clear. However, helix swapping between monomers

of a dimer is known to be involved in oligomeric assembly (Liu and Eisenberg, 2002).

Figure 1.3.7 (a) Tertiary and (b) Quaternary structures of MoaA from Staphylococcus aureus. The bound GTP, radical 5′-adenosine and two iron-sulfur (Fe-S) clusters are also shown as stick.


43

Structure of MoaC

MoaC is suggested to be involved in the release of pyrophosphate of the

intermediate compound generated by MoaA (Hanzelmann and Schindelin, 2006;

Schwarz et al., 2009). However, the exact mechanism of pyrophosphate release is still

unknown. Crystal structures of MoaC, all in apo form, are available from E. coli

(Wuebbens et al., 2000), Pyrococcus horikoshii and Sulfolobus tokodaii (Yoshida et al.,

2008). The overall three-dimensional structure of MoaC has an α+β structure and is

composed of a four-stranded antiparallel β-sheet with two large and two small α-

helices, all located on the same side relative to the β-sheet (Figure 1.3.8a). Ignoring the

two shorter α-helices, the structure can be described as being composed of two

connected interlocked β-α-β units (Figure 1.3.8a). The fold of MoaC belongs to the

ferredoxin-like family. The highest structural similarity is observed with the NAD-

binding domain of 3-hydroxy-3-methylglutaryl-coenzyme A reductase (Friesen and

Rodwell, 2004). Active site is located at the dimer interface. Biological unit of MoaC is

a hexamer made up of three dimers (Figures 1.3.8b,c).


44

Site-directed mutagenesis of E. coli MoaC

Several site-directed single amino acid substitutions (K51A, G52A, K67A,

C76A, H77A, G110A, E112A, T117P, D128A and K131A) of E. coli MoaC have been

studied (Wuebbens et al., 2000). Out of which, the substitution T117P causes Moco

deficiency in humans (Reiss et al., 1998) and the mutant D128A shows no growth

(Wuebbens et al., 2000). The crystal structure of the D128A variant of MoaC revealed

surprisingly large conformational changes in the region 47–51 (active-site loop).

Figure 1.3.8 (a) Overall tertiary structure and quaternary (b) dimer and (c) hexamer arrangement of E. coli MoaC (PDB-ID: 1EKR; Wuebbens et al., 2000). MoaC forms hexamer for its catalytic activity with the active site located at the dimeric interface.


45

1.3.8.2 Synthesis of molybdopterin

In the second step of Moco biosynthesis, two sulfur atoms are incorporated in

precursor Z to generate MPT (Johnson and Rajagopalan, 1987a,b; Leimkuhler et al.,

2001; Leimkuhler and Rajagopalan, 2001) in a similar reaction to that of ubiquitin-

dependent protein degradation (Hershko and Ciechanover, 1998). This reaction is

catalyzed by the enzyme MPT synthase, a heterotetrameric complex of two small

(MoaD, 9 kDa) and two large subunits (MoaE, 17 kDa) that stoichiometrically converts

precursor Z into MPT (Pitterle et al., 1993; Pitterle and Rajagopalan, 1993). The

resulfuration of MPT synthase is catalyzed by another enzyme MoeB involving an

adenylation of MPT synthase followed by sulfur transfer (Matthies et al., 2004, 2005)

similar to the role of UbA1 in the adenylation of ubiquitin in thiamin biosynthesis

(Begley et al., 1999a,b).

This is one of the best-studied steps of Moco biosynthesis in terms of structural

biology. Several crystal structures of MoaD-MoaE and MoaD-MoeB complexes are

available in PDB (Rudolph et al., 2001, 2003; Lake et al., 2001; Daniels et al., 2008).

Both subunits (MoaD and MoaE) of MPT synthase have a mixed α+β architecture

(Figures 1.3.9a,b). The crystal structure of E. coli MPT synthase shows that the C-

terminus of MoaD is deeply inserted into the large subunit to form the active site

(Figure 1.3.9c) and its biological unit (heterotetramer) is formed by dimerization of two

large subunits resulting in an elongated protein complex with two clearly separated

active sites (Figure 1.3.9d). The structure of MoeB is also made up of α+β topology

with central β-sheet flanked by α-helices (Figure 1.3.9e). The MoeB-MoaD interface is

65% hydrophobic in character, a situation similar to that observed in molybdopterin

synthase (Rudolph et al., 2001; Figure 1.3.9f). The crystal structure of the complex

between E. coli MoaD-MoeB complex shows a MoaD2-MoeB2 heterotetramer (Figure

1.3.9g).


46

1.3.8.3 Adenylation of molybdopterin

Third step of Moco biosynthesis involves the adenylation of MPT in an Mg2+

and ATP dependent reaction. This step, in particular, is extensively studied in plants. It

was found that MPT is accumulated in molybdate-repairable mutants (Joshi et al.,

1996) and consequently binding of MPT to Cnx1G (MogA in E. coli) was also

demonstrated (Kuper et al., 2000). Recently, the crystal structures of the wild type and

S583A mutant Cnx1G in complex with MPT and an intermediate MPT-AMP,

Figure 1.3.9 Tertiary structures of (a) MoaD, (b) MoaE and (e) MoeB. (c) The dimeric form of MoaD (yellow) and MoaE (blue), (d) tetrameric form of MoaD (orange and magenta) and MoaE (green and violet), (f) dimeric form of MoaD (green) and MoeB (blue) and (g) tetrameric form of MoaD (yellow and cyan) and MoeB (green and orange).


47

respectively, confirmed the proposed binding of MPT (Kuper et al., 2004; Llamas et

al., 2004). However, unexpectedly, a copper instead of molybdenum was found to be

incorporated in the dithiolate group of MPT (Kuper et al., 2004; Figure 1.3.10a).

Although the exact functional role of copper is not known, it is speculated that it

might play a role in sulfur transfer to cPMP, in protecting the MPT dithiolate from

oxidation and/or in presenting a suitable leaving group for molybdenum insertion

(Kuper et al., 2004). In contrast to bacterial systems, archaeal species seem to possess

MoaB (MogA homologous protein) and contain tungsten instead of molybdenum in

Moco (Bevers et al., 2008). However, they contain two different MoeA orthologs,

whereas, bacteria contain only one (Bevers et al., 2008). It is suggested that MoaB, like

MogA, can carry out the adenylation process of MPT in archaeal species. In contrast,

bacterial MoaB protein is inactive though it can bind MPT (Bevers et al., 2008).

Furthermore, Cnx1G and MogA exclusively form trimer (Wuebbens et al., 2000; Kuper

et al., 2004; Figure 1.3.10b), whereas MoaB forms trimer as well as hexamer (Figures

1.3.10b.c).

Figure 1.3.10 (a) Tertiary structure of E. coli MogA along with the MPT-AMP bound (taken from A. thaliana Cnx1G, PDB-id: 1UUY). The quaternary structures of (b) E. coliMogA (trimer) and (c) E. coli MoaB (hexamer).


48

1.3.8.4 Transport of molybdenum

Moco biosynthesis depends on additional gene products that transport

molybdate anions into cells, synthesize and assemble Moco (Schwarz et al., 2009). In

bacteria, high-affinity molybdate ABC transporters (ModABC) have been described

that consist of ModA (molybdate-binding protein), ModB (membrane integral channel

protein) and ModC (cytoplasmic ATPase; Grunden and Shanmugam, 1997; Mendel

and Bittner, 2006; Figure 1.3.11a). In E. coli, the repressor protein ModE (Figure

1.3.11b) regulates the genes modABC and controls the transcription of molybdopterin

synthases (moaABCDE; Anderson et al., 1997; Hall et al., 1999; Tao et al., 2005). In

addition, other classes of the transport systems (WtpABC) and (TupABC) are also

known (Bevers et al., 2006). In contrast to bacteria, eukaryotic molybdate transport is

poorly understood, but recent studies on Arabidopsis thaliana suggested the occurrence

of a high-affinity molybdate transport system, MOT1 (Tomatsu et al., 2007).

Figure 1.3.11 (a) A view of the ModAB2C2 complex (PDB-id: 2ONK; Hollenstein et al., 2007) and (b) the dimeric form of ModE (PDB-id: 1O7L; Schuttelkopf et al., 2003).


49

1.3.8.5 Insertion of molybdenum

In the penultimate step of Moco biosynthesis, MPT-AMP has to be converted

into Moco (Mendel and Schwarz, 2002). It is proposed that MPT-AMP is transferred to

the N-terminal domain of Cnx1 i.e. Cnx1E (MoeA in E. coli) thereby building a

product-substrate channel (Nichols and Rajagopalan, 2005). It has been suggested that

Cnx1E cleaves the adenylate, releases copper and inserts molybdenum, thus yielding

active Moco (Llamas et al., 2006). However, the exact mechanism of molybdenum

insertion into MPT to make Moco is still an open area of research. The monomer of E.

coli MoeA is an extended L-shaped molecule and can be divided into four discrete

domains (Figure 1.3.12). Domain I forms the upper arm of the letter L. The two chains

in this domain run antiparallel to each other and are mostly dominated by β-strands.

Domain II is located at the end of the vertical arm of the L-shaped monomer and is

composed predominantly of β-strands. Domain III is the largest domain at the corner of

the letter L and its fold resembles to that of MogA. Domain IV contains five β-strands

twisted together into a barrel (Figure 1.3.12).

Figure 1.3.12 Three-dimensional structure of E. coli MoeA (PDB-id: 1G8L; Xiang et al.,

2001). MoeA contains four domains (I to IV), of these domain III adapts a similar fold to that of MogA (Figure 1.3.10a).


50

1.3.8.6 Maturation of molybdenum cofactor

It is found that all bacterial molybdoenzymes (DMSOR family) contain a bis-

MPT-based cofactor instead of a mono-MPT-based cofactor (Figure 1.3.2) observed in

all eukaryotes as well as bacterial enzymes of the XOR family (Moura et al., 2004). In

E. coli, the nucleotide attachment to Moco is carried out by MobA and MobB of the

mob locus (Johnson et al. 1991; Eaves et al., 1997; Palmer et al., 1998) in a

GTP+Mg2+-dependent process (Buchanan et al., 2001). Crystal structure of E. coli

MobA (Lake et al., 2000; Stevenson et al., 2000) shows an α/β architecture with a

nucleotide-binding Rossmann fold (Figure 1.3.13a) and forms a monomer in solution

(Stevenson et al., 2000). However, an octameric form has also been observed in the

crystal structure (Lake et al., 2000; Figure 1.3.13b). MobB adopts a dimeric quaternary

structure (Figures 1.3.13c,d) though not required for the conversion of Moco to MGD,

it increases the activation of molybdoeznymes incorporating MGD by an unknown

mechanism (McLuskey et al., 2003).

Figure 1.3.13 Tertiary (a and c) and quaternary (b and d) structures of E. coli MobA

(PDB-id: 1FRW; Lake et al., 2000) and E. coli MobB (PDB-id: 1NP6; McLuskey et al., 2003), respectively. The GTP molecule and metal ions are shown as stick and spheres, respectively.


51

1.3.8.7 Storage of molybdenum cofactor

As Moco is labile and oxygen-sensitive, it needs to be transferred immediately

after biosynthesis to the apo-molybdoenzyme or to a carrier protein that protects and

stores until further use (Rajagopalan and Johnson, 1992; Figure 1.3.14a). The

availability of sufficient amounts of Moco is essential for the cell to meet its changing

demand for synthesizing molybdoenzymes, thus the existence of a Moco carrier protein

(MCP) would provide a way to buffer supply and demand of Moco (Aguilar et al.,

1992). MCP from Chlamydomonas rheinhardtii forms a homo-tetramer in solution

(Witte et al., 1998; Ataya et al., 2003; Figure 1.3.14b).

1.3.8.8 Transfer of Moco to molybdoenzymes

Transfer of Moco into molybdoenzymes is not clearly understood. Using a

defined in vitro system, it was shown that human apo-sulfite oxidase could directly

incorporate Moco (Leimkuhler and Rajagopalan, 2001). However, for transfer of Moco

into the target apo-molybdoenzymes as it occurs in the living cell, either chaperone

proteins (still unknown) would be needed or MCP could become involved at this stage

(Blasco et al., 1998).

Figure 1.3.14 (a) Schematic diagram of the transfer of the matured Moco compounds to the molybdoenzymes. Mature Moco can either be bound to a Moco carrier protein (MCP), to NR and SO or to the Nifs protein, which generates a protein-bound persulfide that is the source of the terminal sulfur ligand of Moco in enzymes of the XDH/AO family. (b) Quaternary structure of MCP protein from C. reinhardtii (PDB-id: 2IZ7; Fischer et al., 2006). Interfaces and Moco-binding sites are indicated by arrows.


52

1.4 PLAN OF THE WORK As mentioned at the beginning, a major macromolecular crystallography effort

in this laboratory is concerned with structural studies on bovine pancreatic

phospholipase A2 with the aim to understand the structural basis of the enzyme action

with particular attention to the role of the active-site residues and the mode of the lipid

binding to the enzyme PLA2. These studies have yielded several interesting features

pertaining to the functional role of the active site and surface-loop residues, the calcium

ion binding and the mode of the substrate binding in the active site. The work in the

laboratory encompasses more than 50% of the structural studies done on BPLA2 so far

(Sekar et al., 1997a,b; Sekar et al., 1999; Sekar and Sundaralingam, 1999; Rajakannan

et al., 2002; Sekar et al., 2003; Sekar et al., 2005; Sekar et al., 2006a,b). In fact, all the

structural studies on BPLA2 inhibitor complexes have been carried out in this

laboratory (Sekar, 2007). When the candidate joined the laboratory, the crystal

structures of D49E and H48Q mutants of BPLA2 were available (Sekar et al., 1999). In

addition, site-directed mutagenesis and NMR studies on three active-site mutants

H48N, D49N and D49K were available in the literature (Li and Tsai, 1993; Li et al.,

1994). Although, the active-site mutant H48Q shows almost no activity, the mutant

H48N shows detectable enzymatic activity. Similarly, though the mutant D49E enzyme

binds the functionally important calcium ion (12-fold weaker compared to the wild type

enzyme), two mutants D49N and D49K do not bind the active-site calcium ion.

The author's work, described in this thesis, has been concerned with further and

deeper studies on the structural basis of three active-site mutants H48N, D49N and

D49K of BPLA2 enzyme. The structural studies were also complemented with

molecular-dynamics (MD) simulations. This was followed up by the work on invariant

water molecules identified in all the BPLA2 structures available in the literature. Again

this work was supplemented with the MD studies. As discussed in the body of the

thesis, these crystallographic studies supported by MD studies, provided valuable

insights into the catalytic activity of the mutant enzymes.

In the meantime, the collaboration with RIKEN structural biology group,

JAPAN was established in this laboratory. The cloning, expression and purification of

proteins studied under this collaboration were carried out by the collaborators. The

candidate carried out crystallization, data collection, structure solution and analyses.


53

The candidate was assigned to carry out the structural studies on two proteins MoaC

and MogA involved in Moco biosynthesis pathway. When the candidate started this

work, the crystal structures of a native and a mutant MoaC from E. coli were available

(Wuebbens et al., 2000). However, no structural study on MoaC complex was available

in the literature. The candidate started with co-crystallization of MoaC from Thermus

thermophilus HB8. In fact, the structural and biophysical studies of MoaC with GTP

led to propose the nature of the substrate molecule for this protein. On the other hand,

the structural and functional studies on Cnx1G (homologous protein of MogA) from

plants were already available in the literature (Kuper et al., 2004). The candidate

carried out the crystallographic and MD studies on these proteins. These

crystallographic studies, supplemented by MD results, revealed several insights into the

functional role of these proteins and their oligomerization process.

CHAPTER 2 Materials and Methods

CHAPTER 2: MATERIALS AND METHODS 55

2.1 INTRODUCTION This chapter describes the experiments conducted and the methods employed in two

major approaches listed below, which were used in the work reported in this thesis.

1. Protein crystallography

2. Molecular-dynamics (MD) simulations

2.2 PROTEIN CRYSTALLOGRAPHY This section deals with the experiments conducted and the methods used by the

author during the course of crystallization, data collection, structure solution,

refinement and analysis and the theories underlying them. Details of the experiments

are discussed in the appropriate chapters.

2.2.1 CRYSTALLIZATION The first step towards determination of a protein structure using X-ray

crystallography involves crystallization of the protein. Over the years several methods

have been developed to grow diffraction quality protein crystals. Of these, vapor-

diffusion, free interface diffusion, batch and dialysis methods are most commonly used

(McPherson, 1985, 1990, 1997, 2001, 2004a,b; McPherson et al., 1995; McPherson et

al., 2007). Crystallization can be considered as controlled precipitation where the solute

to be crystallized achieves supersaturation state and comes out of the solvent as

crystals. In the vapor-diffusion method (such as sitting-drop and hanging-drop),

equilibration of the vapor pressure across the precipitant concentration gradient among

the drop containing protein solution and the reservoir results in the supersaturation of

the protein solute. In the present work, sitting-drop and hanging-drop vapor-diffusion

techniques were used to carry out screening and optimization of the crystallization

conditions, respectively.

In the hanging-drop vapor-diffusion technique, a drop comprised of a mixture of

protein and reagent (crystallization condition) is placed in vapor equilibration with a

reservoir of reagent. Typically the drop contains a lower reagent concentration than the

reservoir. As a result, initially the droplet of protein solution contains an insufficient

concentration of reagent for crystallization, but as water vaporizes from the drop, the


reagent concentration in the drop increases to a level appropriate for crystallization.

Since the system is in equilibrium, these optimum conditions are maintained until the

crystallization is complete. The sitting-drop vapor-diffusion technique also works on

the same principle as the hanging-drop vapor-diffusion technique except that the

protein drop is kept seated over a siliconised glass plate surrounded by the reservoir

buffer instead of being suspended from the ceiling. Consequently, a larger volume of

protein solution can be used for growing the crystals.

2.2.2 INTENSITY DATA COLLECTION AND PROCESSING The next step towards determination of a crystal structure involves obtaining the

diffraction pattern of the crystal on exposure to monochromatic X-ray beam.

Instrumental details of data collection systems are given in appropriate chapters.

2.2.2.1 Data collection strategy

Collecting optimum X-ray diffraction data involve several choices and

compromises like crystal-to-detector distance, exposure time, oscillation angle,

redundancy, resolution, etc (Dauter, 1999). During data collection, the distance

between the crystal and the image plate was adjusted so that it could record and resolve

the diffraction maxima on the image plate to the resolution limit of the crystal. As a

rule of thumb, the detector is kept at a distance in mm corresponding to the longest

expected unit cell axis in Angstroms. Since large oscillation angles tend to decrease the

signal to noise ratio and the accuracy in the estimated reflection profiles due to higher

background, an oscillation angle of 1.0° is found to be a good compromise between

speed and the data quality. Exposure time was set long enough to give reasonable

statistics at the highest resolution, but not so long as to overload the detector with the

strong low-angle spots. The exposure time was selected on the basis of size and

diffraction quality of the crystal and the oscillation range.

2.2.2.2 Data processing

All the datasets were indexed, integrated and scaled using the programs

DENZO and SCALEPACK of the HKL suite (Otwinowski and Minor, 1997). The

analysis and reduction of single crystal diffraction data consists of the following steps:


I. Visualization and preliminary analysis of the original, unprocessed data.

II. Indexing of the diffraction pattern.

III. Refinement of the crystal and detector parameters.

IV. Intensity integration of the diffraction maxima by profile fitting.

V. Estimating the scale factors to convert data from all the frames to a common

scale.

VI. Symmetry determinations and merging of the symmetry related reflections.

VII. Statistical summary and estimation of errors.

The first four steps are carried out by XDISPLAYF and DENZO and the last

three steps are carried out by SCALEPACK.

DENZO allows for an interactive visualization, adjustment and input of various

parameters such as shape, size and profile-fitting radius of the spots. The user also has

the choice of visually selecting reflections for input to the auto-indexing routine. After

each cycle of refinement, DENZO updates the display and prints the new values for the

refined parameters and shift in their values. The output gives the χ2 value for the X and

Y positions of the predicted spots. The χ2 values represent the average ratio, squared, of

the error in the fitting divided by the expected error. A good refinement is expected to

have χ2 close to one. A very large value for the χ2 indicates some serious error with

indexing, refinement or detector parameters. At the end of the refinement (for each

individual image), DENZO outputs a list of reflections (hkl) and their unscaled

intensities.

The program SCALEPACK is used to scale the intensities obtained from

DENZO. The program calculates single isotropic scale and B factors for each of the

processed input frames.

The assessment of the high-resolution limit of the diffraction pattern is done in

two ways: the first is the mean ratio of the intensity to the error [I/σ(I)] and the second,

Rmerge, is the agreement between the symmetry related reflections and is given by

( ) ( )

( )∑∑∑∑ −

=

h ii

h ii

merge hI

hIhIR (2.1)

where I(h)i is the ith measurement of the intensity of a reflection h and <I(h)> is its

average intensity.


The first criterion is an indicator of data quality and a value greater than or

equal to 2.0 is generally accepted. Rmerge, on the other hand, is an unweighted statistic,

which is independent of error model. It can be intentionally or unintentionally

manipulated. Low redundancy, omission of weak or partial reflections, use of sigma

cutoffs in the data set leads to artificially low Rmerge.

2.2.3 CALCULATION OF STRUCTURE FACTOR AMPLITUDES The program TRUNCATE in the CCP4 suite (CCP4, 1994) was used to convert

a file of averaged intensities to a file containing mean amplitude |F| and the original

intensities I. TRUNCATE allows two ways of calculating the amplitudes. In the first,

the amplitude is taken as the square root of the intensities, setting any negative intensity

to zero. Alternatively, a best estimate of |F| is calculated from I, σ(I) and the

distribution of intensities in resolution shells (French and Wilson, 1978). By this

method, negative intensities are made positive and the weakest reflections are inflated

so that these observations are not underestimated.

2.2.4 STRUCTURE SOLUTION In macromolecular crystallography, the initial phases are normally obtained

using one of the following methods: (i) Multiple-wavelength anomalous dispersion

(MAD), (ii) Multiple isomorphous replacement (MIR) and (iii) Molecular replacement

(MR). The crystal structures, described in this thesis, were solved using MR method

(Rossmann and Blow, 1962; Rossmann, 1972). The theoretical details of MR method

and the program Phaser (McCoy et al., 2007) are described in the next section.

2.2.4.1 Molecular Replacement

MR is used when the three-dimensional structure of a similar or homologous

molecule is available. Essentially, it involves generating a preliminary model of the

target crystal structure by orienting and positioning the search molecule within the unit

cell of the target crystal so as to best account for the diffraction pattern (Rossmann,

1990). The problem in MR is to find six (three-rotational and three-translational)

parameters to place the search model in the unit cell of the target protein crystal

(Rossmann and Blow 1962; Rossmann, 1972).


The rotation function involves looking for agreement between Patterson

functions of the model and the target structure as a function of their relative orientation.

To evaluate this agreement index, a function R is defined as

( ) ( )∫=u

21 dVCxPxPR (2.2)

where P1 and P2 are Patterson functions, C is the rotation operator that rotates the

Patterson function P2 with respect to P1 and u is the spherical volume of integration

centered at the origin. A maximum in the rotation function (R) indicates a potential

orientation for the search molecule in the target crystal. After finding a potent rotation

function, the translation [T] of the molecule X with an orientation [R] relative to the

model M involves the maximization of the function

( ) ( )duuPt,uP)t(T 1Cell

2∫= (2.3)

where P1(u) is the observed Patterson of the unknown cell, P2(u,t) is the Patterson

corresponding to a homologous model rotated using the results of cross-rotation

function search to an orientation corresponding to that of the unknown molecule and

positioned at t from origin (Rossmann et al., 1964; Crowther and Blow 1967). The

three-dimensional structure solution can be obtained from the combined result of

rotation and translation function searches using

[ ] [ ]TMRX += (2.4)

where [R] is the appropriate rotation and [T] the required translation to correctly orient

and position the search model in the target unit cell. Several programs such as AMoRe,

Phaser, MOLREP, MrBUMP, and BALBES are available for applying MR. The

program Phaser was used in the present investigation.

2.2.4.2 Phaser

The algorithms in Phaser are based on maximum likelihood probability theory

and multivariate statistics rather than the traditional least squares and Patterson

methods. Maximum likelihood is a branch of statistical inference, which asserts that the

best model on the evidence of the data is the one that explains what has in fact been

observed with the highest probability (Fisher, 1922). The model is a set of parameters,

including the variances describing the error estimates for the parameters. The likelihood

of the model, given the data, is defined as the probability of the data given the model


( ) ( )Model/datapdata/ModelL ii = (2.5)

where i=1 to N and the data have independent probability distributions. Thus, the joint

probability of the data given the model is the product of the individual distributions

( ) ( )∏=

=N

1iii Model/datapdata/ModelL (2.6)

In crystallography, the data are the individual reflection intensities, which are not

strictly independent. However, the assumption of independence is necessary to make

the problem tractable and works well in practice (McCoy et al., 2007). To avoid the

numerical problems of working with the product of potentially hundreds of thousands

of small probabilities (one for each reflection), the log of the likelihood is used in the

following way

( ) ( )[ ]∑=

=N

1iii Model/dataplndata/ModelLL (2.7)

The best solutions in Phaser are identified based on the values of Z-score and

log-likelihood gain (LLG). The Z-score expresses the divergence of the experimental

result from the most probable result mean as a number of standard deviations. The

larger the value of Z-score, the less probable the experimental result is due to chance.

LLG is the difference between the likelihood of the model and the likelihood calculated

from a Wilson distribution. So it measures the improvement in the prediction of the

data with the model than with a random distribution of the same atoms.

2.2.5 STRUCTURE REFINEMENT Structure refinement aims at optimizing the agreement of an atomic model with

both observed diffraction data and chemical restraints. It is an iterative process to

improve the quality of the structure of the model. Three positional (x, y and z) and one

atomic displacement (B-factor) parameters, and sometimes occupancy, for each atom is

adjusted to minimize the difference between the observed structure factor (|Fo|) and

those calculated from the structure model (|Fc|). Crystal structures reported in this thesis

were refined using the program CNS (Crystallography and NMR System, Brunger et al.

1998). Molecular-dynamics methods are exploited by CNS to probe the conformational

space of the molecule while minimizing the difference between the observed and

calculated structure factors (Brunger et al., 1987). There are options for rigid body


refinement, positional refinement, restrained and unrestrained individual B-factor

refinement, group B-factor refinement, occupancy refinement and electron-density map

calculations. In addition, there are options to perform simulated-annealing refinements

both in the cartesian and torsion angle conformational space. Features of the programs

pertinent to the present work are discussed briefly in the following sections.

2.2.5.1 Cross-validation

The reliability of the fit of a model to the diffraction data is given by the R-

factor, which measures the discrepancy between the observed structure factor

amplitudes Fo and calculated structure factor amplitudes |Fc|:

( ) ( )

( )∑∑ −

=

hklo

hklco

hklF

hklFhklFR (2.8)

This value can be made arbitrarily low by increasing the number of adjustable

parameters used to describe the model. The method of statistical cross-validation by

using free R-factor is a more accurate indicator of model quality (Brunger, 1992). For

cross-validation, the diffraction data are divided into two sets: a large working set

(usually comprising of 90-95% of the data) and a small complementary test set

(comprising the remaining of 10-5% of the data). The diffraction data present in the

working set is used for refinement. It provides a more objective guide during model

building and refinement process than the conventional R-factor. If the model is correct

and errors are statistical, Rfree is expected to be close to R-factor.

2.2.5.2 Target functions

Crystallographic refinement can be formulated as a search for the global

minimum of the target function (Jack and Levitt, 1978)

xrayxraychemtotal EwEE += (2.9)

where the term Echem is a function of all atomic positions describing covalent (bond

lengths, bond angles, dihedral angles, chiral centers, planarity) and non-covalent (van

der Waals, hydrogen-bonding and electrostatic) interactions. The term Exray takes into

account the differences between observed and calculated diffraction data. The term

wxray is the weight chosen to balance the contributions from Echem and Exray. The choice


of wxray can be critical. If wxray is too large, the refined structure might have large

deviations from the ideal geometry and if wxray is too small, there will be large

discrepancy between the refined structure and the experimental data. Automated

procedures to calculate initial estimates for optimal weighting are available in CNS, but

cross-validation must be used to obtain the best possible weight for the diffraction data.

Several algorithms have been developed to minimize the target function (Etotal)

based on least squares (Konnert, 1976), conjugate gradient (Jack and Levitt, 1978) and

simulated annealing (Brunger et al., 1987, 1990).

The first term (Echem) of the target function (Etotal) is an empirical potential

energy function and is defined in the program CNS as

( ) ( ) ( )[ ]∑ ∑ ∑ δ+φ+θ−θ+−= φθ ncosKKbbKE 20

20bchem

( ) ∑∑ ⎟⎠⎞

⎜⎝⎛ +++ω−ω+ ω r

drc

raK 612

20 (2.10)

where the symbols b, θ, φ and ω are the ideal bond length, bond angle, torsion angle

and chiral volume, respectively. The symbols bo, θo and ωo are equilibrium values and

Kb, Kθ, Kφ and Kω are energy constants, n is periodicity, δ is the phase shift, r is the

distance between two non-bonded atoms and a, c, d are constants. All the parameters in

the above equation are obtained from a small-molecule database (Engh and Huber,

1991).

The second component (Exray) of the target function (Etotal) in the least square

optimization method is given as

( ) ( )( )∑ −==hkl

2co

LSQxray hklkFhklFEE (2.11)

where the scale factor (k) is usually estimated by minimizing the above equation.

2.2.5.3 Maximum likelihood refinement targets

Energy minimization using least square optimization method can improve the

model, but results in the accumulation of systematic errors in the model by fitting noise

in the diffraction data. An improved target for macromolecular refinement is maximum

likelihood function (Adams et al., 1997; Read, 1997). In this method, the likelihood of

a model is maximized, given the estimates of the errors in the model and the measured

intensities. The effects of model errors on the calculated structure factors are quantified


with σA values, which correspond to the fraction of each structure factor that is

expected to be correct (Read, 1986, 1997). The expected values of <Fo> and the

corresponding variance (σ2) are derived from σA, Fo and Fc structure factor amplitudes

(Pannu and Read, 1996). Thus, the term (Exray) can be modified as

( )2

hklco2

ML

MLxray FF1EE ∑ −

σ== (2.12)

However, the maximum likelihood leads to the least squares if the errors between the

observed and the predicted values follow a Gaussian distribution. In order to achieve an

improvement over the least squares residual, cross-validation was found to be essential

for the computation of σA and its derived quantities.

2.2.5.4 Rigid-body refinement

This procedure minimizes the differences in the observed and the calculated

structure factors by refining three rotational and three translational degrees of freedom

(Head-Gordon and Brooks, 1991) of the user-defined ‘rigid’ groups. Each group is

regarded as a continuous mass distribution located at the center of mass defined by

∑=i

iij

j rmM1R where ∑=

iij mM (2.13)

where mi represents the ith atomic mass and j is the number of rigid groups. The

segments of the molecule not included in any group are kept fixed.

2.2.5.5 Positional refinement

Routines are available in CNS to carry out conventional positional refinement

where energy minimization is carried out by the use of a conjugate-gradient

minimization algorithm (Powell, 1977). The algorithm requires the value of energy and

its first derivative and uses gradient descent minimization for convergence.

2.2.5.6 Simulated annealing

Annealing represents a physical process wherein a solid is heated until all the

particles randomly arrange themselves in a liquid phase and then is slowly cooled so

that all the particles arrange themselves in the lowest energy state. By defining the Etotal

(Equation 2.9) to be equivalent of the potential energy, the annealing process can be


simulated (Brunger et al., 1990, 1997). Compared to conjugate-gradient minimization,

where such directions must follow the gradient, simulated annealing (SA) achieves

more optimal solution by allowing motion against the gradient. The parameter

‘temperature (T)’ used does not have any physical significance and is correlated to the

likelihood of overcoming the energy barriers. The SA algorithm in CNS uses the

molecular-dynamics simulations mechanism to create a Boltzmann distribution at a

given temperature T.

2.2.5.7 Atomic displacement (B-factor) refinement

The B-factor refinement can be performed primarily in two methods, which are

dictated by the resolution of the data. In general, for a data set of resolution better than

or equal to 2.5 Å, individual B-factor refinement is performed. Here, the B factor for

the individual atom is defined by one parameter. For a data set of resolution from 2.5 to

3.5 Å, grouped B-factor refinement is carried out. In this case, a single value for the

whole molecule is defined. For atomic-resolution data, the anisotropic B-factor

refinement can be employed, which involves the refinement of six parameters that

represent the B factor for each atom.

2.2.5.8 Torsion-angle dynamics

Cartesian (flexible bond lengths and angles) molecular dynamics places

restraints on bond lengths and bond angles (Equation 2.10). These restrictions are

implemented as constraints (fixed bond lengths and angles) in the torsion angle

dynamics algorithm (Jain et al., 1993). Molecular dynamics can be carried out at

significantly higher temperatures due to elimination of high frequency bond and angle

vibrations. For crystallographic refinements, this formulation significantly increases

convergence over conventional techniques because of its reduced variable high

temperature sampling strategy. This is particularly significant for the refinement of

macromolecules when the data to parameter ratio is low (Rice and Brunger, 1994).

2.2.5.9 Constraints and restraints

During the course of refinement of macromolecules, some groups of atoms may

have to be constrained or restrained to improve the ratio of observable reflections to


parameters. CNS has options to group atoms so that they move as rigid bodies, or,

restrain or constrain the bond lengths, bond angles, noncrystallographic symmetry

(NCS) and atomic positions to a desired value by use of appropriate force constants.

Restraints are used when limited freedom can be given for a parameter. When a

parameter has to be held to an exact value, then it is constrained. In NCS symmetry

restraints, the molecules in the asymmetric unit are superposed by least squares

superposition and the average coordinates (xav) of individual atoms are computed. If x

represents the coordinates of individual atoms, then each atom can be restrained

according to the mathematical term:

( )2avNCS xxwE −= and

( )2NCS

2av

NCSbb

Bσ−

= (2.14)

where w is a weight function, b and bav are the respective individual and average

temperature factors of NCS related atoms and σNCS is the target deviation for B-factor

restraints.

2.2.5.10 Bulk solvent scattering

The solvent content in protein crystals is typically observed to be in the range of

30-70% of the total crystal volume (Matthews, 1968). Thus, an appropriate model for

the continuous bulk solvent is included in order to avoid an overestimation of the

electron-density contrast at the protein surface (Kostrewa, 1997). In CNS, it is possible

to include all experimental measurements in the refinement by taking into account a

bulk solvent model to compensate for scattering at low resolution. The volume that the

solvent should fill is identified by demarcating a solvent-accessible volume outside the

van der Waals exclusion zone of the protein. The optimal value for the average solvent

density may be obtained by finding the minimum value of

( )( )∑ − 2co totalFF (2.15)

where ( ) ( ) ( ) ⎟⎠⎞

⎜⎝⎛ −+= 2

solcsolcc d4

BexpsolventFKproteinFtotalF (2.16)

the summation is over low resolution reflections, Fc(solvent) is a Fourier transform of a

binary function M (solvent mask) whose value is 1 inside the solvent and 0 outside

(protein region), Ksol is a scale factor, Bsol is an artificially large temperature factor


applied to a flat solvent density, d is the spacing between the planes in the atomic

lattice (Fokine and Urzhumstev, 2002).

2.2.6 ELECTRON DENSITY MAPS AND INTERPRETATION After every cycle of the refinement, model was inspected and manual rebuilding

was done by inspection of 2Fo-Fc and Fo-Fc electron-density maps. CNS has options to

calculate σA-weighted maps where the structure factor amplitudes are weighted in order

to reduce the model bias of an incomplete or partially incorrect structure. The Fourier

coefficients calculated are given by

( ) ( )cco iexpDFmF2 α− and ( ) ( )cco iexpDFmF α− (2.17)

where m is the figure of merit and D is the measure of error in the coordinates of the

model. Both the parameters are 1 for 2Fo-Fc and and Fo-Fc maps.

The maps were usually contoured at 1σ and 3.0σ, respectively, where σ refers

to the root-mean-square deviation (r.m.s.d.) in the mean density (electrons/Å3) in the

maps. Regions of poor electron density were examined with the maps contoured at a

lower level. The molecular-modelling package COOT (Crystallographic Object-

Oriented Toolkit, Emsley and Cowtan, 2004) was used to examine and interpret the

model against the electron-density maps. During the final stages, omit maps were

calculated using the program CNS to check the correctness of the model.

2.2.6.1 Identification of solvent sites

Water molecules were identified using peaks from 2Fo-Fc and Fo-Fc maps

contoured generally at 0.8σ and 2.5σ, respectively. Identification of water molecules

was done both manually and using the automatic water picking routines available in

COOT. The positions of the solvent were verified using omit maps. It was further

verified that no solvent site was located at a distance less than 2.2 Å and more than 3.5

Å from any protein atom and that the R-free values dropped on addition of water

molecules to the model. An initial B factor of 30 Å2 was assigned to the water

molecules, which were subsequently refined. Several rounds of this procedure were

carried out until most of the density in the maps was accounted for and the R-factor of

the model converged.


2.2.6.2 Reducing model bias with omit maps

In order to reduce the effects of model bias, a simulated-annealing omit map is

calculated. In the case of a molecular-replacement solution, it is not certain which parts

of the model contain an error and therefore an omit map that covers the entire molecule

is most useful. During this process, small regions of the model are systematically

excluded and a small map is computed covering the omitted region (Bhat, 1988; Hunt

et al., 1997). These small maps are accumulated and written out as a continuous map

covering the whole molecule (or the defined region). Simulated-annealing refinement

and minimization are used to remove the bias from the omitted region. A composite

omit, cross-validated, σA-weighted map is calculated using the program CNS.

2.2.7 STRUCTURE VALIDATION AND DEPOSITION The refinement programs analyze the geometrical parameters and list the root-

mean-square deviation (r.m.s.d.) in bond lengths, bond angles and dihedral angles,

short contacts between symmetry related atoms etc. of the refined structure. The

program also lists the energies that deviate from weights used for the refinement (Engh

and Huber, 1991). Some of the other validation and deposition tools, which were used

extensively during the course of investigation, are described below.

2.2.7.1 PROCHECK

The program PROCHECK (Laskowski et al., 1993) was used to check the

stereo chemical quality of protein structures. The output consists of a comprehensive

residue-by-residue listing of the parameters and its graphical representation. One of

them is the Ramachandran plot (Ramachandran et al., 1963) to analyze the

stereochemistry of the model and identify the secondary structural features of the

protein molecule. The program compares and assesses the quality of the model vis-a-vis

other structures at comparable resolutions. This program was used to assess the quality

of the model after every refinement cycle.

2.2.7.2 MolProbity

MolProbity is a general-purpose web service offering quality validation for

three-dimensional structures of proteins, nucleic acids and complexes (Davis et al.,


2007). The program adds both polar and nonpolar hydrogen atoms to the structure and

calculates any short contact present in the structure or appeared because of the addition

of hydrogen. In addition, the program also examines the local chemical environment,

and if required, suggests the possible flips for the residues Asn, Gln, and His. It also

enlists Ramachandran-map outliers, torsion-angles outliers and puckered residues,

particularly carbohydrates in nucleic acid structures.

2.2.7.3 ADIT

The Auto Dep Input Tool (ADIT) was developed by the RCSB for depositing

structures to the Protein Data Bank (http://pdbdep.protein.osaka-u.ac.jp/validate/).

ADIT allows users to check the format of coordinates and structure factor files and to

perform a variety of validation tests on a structure prior to deposition. In addition, the

program identifies isolated water molecules and automatically assigns water molecules

to their respective subunits based on their closeness. It also lists missing residues in the

structure model given the amino acid sequence.

2.2.8 ANALYSIS OF SEQUENCES AND STRUCTURES A number of programs used for the analysis of sequences and structures to gain

insights into structures and functions are described below.

2.2.8.1 Sequence analysis

The program ProtParam was used to compute various physical and chemical

parameters such as molecular weight, theoretical pI, molar extinction coefficient etc.

for a given protein sequence (Gasteiger, 2005).

The programs BLAST (Basic Local Alignment Search Tool; Altschul et al.,

1990) and ClustalW (Thompson et al. 1994) were used for pairwise and multiple

sequence alignments, respectively.

The program ESPript (Easy Sequencing in PostScript; Gouet et al. 2003) was

used to decorate the visualization, via PostScript output, of sequences aligned with

programs such as ClustalW. It offers a palette of markers to highlight important regions

in the alignment.


The program MUSTANG (MUltiple STructural AligNment AlGorithm;

Konagurthu et al., 2006) was used to align the sequences based on their three-

dimensional structures.

The web server PSAP (Protein Structure Analysis Package; Balamurugan et al.,

2007) was used during model building of the structures described in this thesis. The

program contains a module to align amino acid sequence and those derived from the

atomic model of the structure. The program essentially helps in locating the missing

residues in the structure model.

2.2.8.2 Phylogenetic tree

A program DrawTree from the Phylip suite (Felsenstein, 2005) was used to

draw the phylogenetic tree obtained from multiple sequence alignment. The program

interactively plots an unrooted tree diagram with many options including orientation of

tree and branches and label sizes.

2.2.8.3 Secondary-structure elements

The program DSSP (Kabsch and Sander, 1983) was used to define secondary

structure, geometrical features and solvent exposure of proteins, given atomic

coordinates in Protein Data Bank format. Two programs Helixang (from CCP4 suite)

and an independent C program Interhlx (Yap et al., 2002) were used to calculate the

angles among secondary structural elements of the protein structures.

2.2.8.4 Structural comparison

Comparing three-dimensional structures may reveal biologically interesting

similarities that are not detectable by comparing sequences. The program ALIGN was

used to perform superposition of two coordinate sets (Cohen, 1997). A web server

3dSS (3-dimensional Structural Superposition) was used to superpose structures to

identify the invariant water molecules (Sumathi et al., 2006). The web server DALI

(Holm and Sander 1995) was used to search the structural homologues in the PDB.


2.2.8.5 Structural rigidity

The program ESCET (Error-inclusive Structure Comparison and Evaluation

Tool; Schneider, 2004) was used to delineate the conformationally rigid and flexible

regions of the protein structures. This tool makes use of error scaled difference distance

matrices and employs a genetic algorithm. The delineation is effected on the basis of a

multiple of a parameter called σ, which can be chosen by the user on the basis of

structural and other relevant considerations.

2.2.8.6 Hydrogen bonds

The program CONTACT (a part of the CCP4 suite) was used for computing

various types of contacts in protein structures. In addition, the program HBPLUS was

used to identify hydrogen bonds (McDonald and Thornton, 1994). In D−H---A−X

systems, where D (Donor) is nitrogen or oxygen and A (Acceptor) is oxygen, contacts

with D---A distances less than 3.5 Å and with angles at H and A greater than 120°,

were treated as hydrogen bonds.

2.2.8.7 Electrostatic potentials and surfaces

The program NACCESS was used to calculate the atomic accessible surface

area using a probe of radius 1.4 Å (Hubbard and Thornton, 1993). In addition, the web-

server PISA (Krissinel and Henrick, 2007) was used to explore the macromolecular

interfaces, prediction of probable quaternary structures (assemblies), etc. The program

SURFNET (Laskowski, 1995) was used to identify the surface cavities and their

volumes. It can compute van der Waals surfaces, gaps between molecules, clefts and

cavities and three-dimensional density distributions.

2.2.8.8 Identification of functional sites

Two programs PatchFinder (Nimrod et al., 2008) and ConSurf (Landau et al.,

2005) were used to identify the functionally important regions in proteins with known

three-dimensional structure. Both the programs are based on the idea that evolutionary

conserved regions are often functionally important.


2.2.8.9 Protein-protein docking

The program ClusPro (Comeau et al., 2007) was employed to perform protein-

protein docking. The program is the first fully automated, web-based program for

docking protein structures. In addition, the web server PPI-Pred (Bradford and

Westhead, 2005) was used to predict protein-protein binding sites using a combination

of surface patch analysis.

2.2.8.10 Others

A freely available web server PDB Goodies (Hussain et al., 2002) was used at

the various stages of the structure refinements and analysis. Another freely available

web server WAP (Water Analysis Package; Praveen et al., 2008) developed locally was

used for water analysis.

2.2.9 STRUCTURE VISUALIZATION The program PyMOL, created by W.L. DeLano and commercialized by DeLano

Scientific LLC (DeLano Scientific LLC http://www.pymol.org), was used to generate

all the figures. A program APBS (Adaptive Poisson-Boltzmann Solver) plugged into

PyMOL was used to calculate the electrostatic potentials of the protein structures

(Baker et al., 2001). It is a cost-effective but uncompromised alternative to GRASP.

2.3 MOLECULAR DYNAMICS SIMULATIONS In this section, a brief introduction to the theory of molecular-dynamics (MD)

simulations, algorithms and parameters pertaining to its theory have been described.

2.3.1 INTRODUCTION For biomolecules specific processes occur over a wide range of time scales. For

example (i) local motions (0.01 to 5 Å, 10-15 to 10-1 s) - atomic fluctuations, side chains

and loops, (ii) rigid body motions (1 to 10 Å, 10-9 to 1 s) - helices, domains and

subunits and (iii) large-scale motions (>5 Å, 10-7 to 104 s)- helix coil transitions,

dissociation/association, folding and unfolding. So, bio-macromolecules do not exist as

static structures in solution; rather they are represented by an ensemble of


conformations in equilibrium. Moreover, macroscopic properties determined by

experimental methods are averaged over ensemble.

Macromolecular MD is a method of choice for understanding properties

associated with such conformation ensembles at the atomic level (Dodson et al., 2008).

It permits the study of complex dynamic processes such as protein stability,

conformational changes, protein folding and provide a means to carry out the studies

like drug design. With the development of fast algorithms and computational powers in

conjunction with our improved understanding of physicochemical properties of

macromolecular systems, a near reality representation of such systems in in silico

environment now seems possible.

2.3.2 GENERAL THEORY OF MOLECULAR DYNAMICS Computation of macroscopic properties as a function of ensemble states,

employing statistical mechanical calculations, is an extremely difficult and expensive

approach. However, in the MD simulations, conformations are achieved as a function

of time and hence if the simulation has been run for substantially longer time it is

reasonable to assume that all the possible states of the particle have been sampled. The

time average property calculated through simulations thus can be correlated with

ensemble average property with the help of Ergodic hypothesis which states that the

time average equals the ensemble average. In other words,

EnsembleTimeAA = (2.18)

For a system of N interacting atoms, the MD formalism to simulate atomic

motion is given by the Newton's equation

⎟⎟⎠

⎞⎜⎜⎝

⎛∂∂

= 2i

2

ii tm

rF where i=1,…,N (2.19)

where mi is the mass, ri is the position vector of the ith atom, t is the time and Fi is the

force on that atom. Force on an atom can be calculated from the derivative of a

potential function V(r1,…,rn) with respect to the atom's position

⎟⎟⎠

⎞⎜⎜⎝

⎛∂∂

−=i

iVr

F (2.20)

The simultaneous solution of above equations in small time steps computes and

updates the force on an atom, its velocity and hence position. New coordinates as a


function of time are stored as a simulation trajectory and analyzed to understand the

desired properties.

2.3.3 PROTOCOLS AND PARAMETERS OF MOLECULAR DYNAMICS

SIMULATIONS

The fundamentals of macromolecular MD simulations, in general, are the same

and are independent of the programs used. However, differences lie in the details of the

algorithms implemented depending on the program used or the problem addressed.

Simulation studies carried out in the present thesis work employed GROMACS

versions 3.3, 3.3.3. and 4.0.4. The general outline of the simulation protocol and

parameters are discussed below and are largely adapted from GROMACS version 4.0

documentation (Lindahl et al., 2001; Hess et al., 2008). The general MD algorithm

involves four steps.

1. System representation, input assignment and parameters decision

2. Computation of forces

3. Configuration update

4. Output

2.3.3.1 System representation, input assignment and parameters

The first step of the MD simulations involves the representation of the

molecular system with required parameters and conditions under which the simulation

has to be carried out. In GROMACS, the macromolecule of interest (protein in the

present case) is given as an input in the form of a set of atomic coordinates. The protein

molecule is first placed in a periodic simulation box. The user can decide the shape and

size of the box. GROMACS supports triclinic box of any shape, some of which can be

specifically described as cubic, rhombic dodecahedron and octahedron box types.

Simulations involving biomolecules often include explicit water molecules. Since the

speed of simulation also depends on the number of atoms present in the simulation box,

the decision for selecting the simulation box and number of water molecules to be used


during simulation has to be made judiciously. As a rule of thumb, the biomolecule is

preferred to be solvated with at least primary hydration layer around it.

After the assignment of initial coordinates and thus positions of atoms,

velocities are assigned for them individually. In the absence of a prior knowledge of

velocities, initial velocities are assigned using Maxwell distribution for a given

temperature in the following way

( ) ⎟⎟⎠

⎞⎜⎜⎝

⎛−

π=

kT2vm

expkT2

mvp

2iii

i (2.21)

where i is the identity of the atom, p(vi) is the probability that the ith atom in the

Maxwell distribution takes velocity vi, m is the mass of the atom, T is the temperature

and k is the Boltzmann's constant.

2.3.3.2 Computation of forces

Once initial positions and velocities are assigned for atoms, forces for them are

computed as the negative gradient of potential functions (Equation 2.20). The

computation also involves the calculation of potential energy associated with various

interaction terms such as Lennard-Jones, Coulomb and bonded terms, which will be

discussed in the section describing force fields.

Temperature coupling

In the MD simulations, it becomes necessary to control the temperature of the

system because of drift during equilibration, drift as a result of force truncation and

integration errors and heating due to external or frictional forces. The instantaneous

temperature during the simulation is calculated by the kinetic energy of the system.

GROMACS can use either the weak coupling scheme of Berendsen (Berendsen et al.,

1984.) or the extended ensemble Nose-Hoover scheme (Nose, 1984; Hoover, 1985).

The works presented in this thesis use the latter one and has been described here.

Nose-Hoover temperature coupling

In this method of temperature coupling, an additional thermal reservoir and a

frictional parameter are embedded. The method is more accurate in probing the correct


canonical ensemble for equilibrium conformations. The modified equation of motion

incorporated with additional terms associated with the coupling algorithm looks like

dtdξ

mdtd i

i

i2i

2 rFr−= , where ( )0TT

Q1

dtd

−=ξ and 2

02T

4T

Qπ

τ= (2.22)

ξ is heat bath parameter, Q is mass parameter of the reservoir, T0 is the reference

temperature, T is the current instantaneous temperature of the system and τT is another

time constant parameter related with time constant of coupling (τ) in the following way

kN/C2 dfTvτ=τ (2.23)

where Cv is the total heat capacity of the system and Ndf is the total number of degrees

of freedom.

Pressure coupling

In a similar way as the temperature coupling, the system can also be coupled to

a ‘pressure bath’. GROMACS supports both the Berendsen algorithm that scales

coordinates and box vectors every step and the extended ensemble Parrinello-Rahman

approach (Parrinello and Rahman, 1981). Both of these can be combined with any of

the temperature coupling methods above. In the present thesis, Parrinello-Rahman

approach has been used.

Parrinello-Rahman pressure coupling

The algorithm is similar to Nose-Hoover temperature coupling algorithm and

uses constant-pressure simulation. The method is particularly useful when fluctuations

in pressure or volumes are important during simulation. The equation of motion after

the incorporation of Parrinello-Rahman pressure coupling looks like

dtd

mdtd i

i

i2i

2 rMFr−= where 11 ''

dtd

dt'd −−

⎥⎦⎤

⎢⎣⎡ += bbbbbbM (2.24)

and b is the matrix representing box vectors and b' is transpose of b.

2.3.3.3 Configuration update

Several algorithms are available to update the configuration of the molecule

during MD simulations. Out of which, Verlet (Verlet, 1967) and Leap-frog (Hockney et

al., 1974) algorithms are widely used. In GROMACS, Leap-frog algorithm is utilized


for the integration of the equations of motion. This algorithm uses positions r at time t

and velocities v at time t − ∆t/2; it updates positions and velocities using the forces F(t)

determined by the positions at time t:

( )Δtm

t2Δttv

2Δttv F

+⎟⎠⎞

⎜⎝⎛ −=⎟

⎠⎞

⎜⎝⎛ + (2.25)

( ) ( ) Δt2ΔttvtΔtt ⎟

⎠⎞

⎜⎝⎛ ++=+ rr (2.26)

It is equivalent to the Verlet algorithm:

( ) ( ) ( ) ( ) ( )42 ΔtOΔtm

tΔttt2Δtt ++−−=+Frrr (2.27)

where r is the position vector, v is the velocity, t is the time, F is the force and m is the

mass.

2.3.3.4 Output

Output of MD simulations in the form of coordinates and optionally velocities

can be saved at regular intervals as a function of time. Saving output at each time step

results in the generation of very large output files hence, for practical purposes,

coordinates can be saved at some discrete but small intervals.

2.3.4 FORCE FIELDS

Force fields in molecular mechanics can be described in terms of the potential

functions and the parameters used in them to generate the potential energy of the

system. Potential functions in GROMACS have been subdivided in three parts, which

include non-bonded and bonded interactions and special terms like position and

distance restraints.

2.3.4.1 Non-bonded interaction terms

Non-bonded interactions in GROMACS are pair-additive and centro-symmetric

( ) ( )∑<

=ji

ijijN1 V,...V rrr and ( )

jij

ij

j ij

ijiji rdr

rdVF

rF −== ∑ (2.28)


It contains repulsion, dispersion and a Coulomb term. The first two terms are

combined in either the Lennard-Jones (6-12 interaction) or the Buckingham (exp-6

potential).

The Lennard-Jones potential (VLJ) and Coulombic potential (Vc) for non-

bonded interaction terms are given as

( )( ) ( )

6ij

6ij

12ij

12ij

ijLJ rC

rC

rV −= and ( )ijr

jiijc r

qqfrVε

= , respectively (2.29)

12ijC and 6

ijC parameters depend on pairs of atom type, εr is the dielectric constant of the

medium and 935485.1384

1f0

=πε

= , where ε0 is the permittivity of free space. In

GROMACS the relative dielectric constant εr may be set in the in the input for the

module grompp.

2.3.4.2 Long-range electrostatics

Computation of long-range non-bonded interactions such as electrostatic

interactions with periodic boundary conditions (see later) requires a huge amount of

computational power and is very slow in convergence. To solve this problem, there are

many algorithms like Ewald summation (Ewald, 1921), Particle-Mesh Ewald (PME;

Darden et al., 1993; Essmann et al., 1995), Particle-Particle Particle-Mesh (PPPM;

Hockney and Eastwood, 1981; Luty et al., 1995) and optimizing Fourier transforms

(FFT). The summation is modified using Ewald summation or PME methods for

practical efficiency.

Ewald Summation

In case of Ewald summation, the interaction potential is divided in three terms.

The first term (Vdir) computes short-range electrostatic interactions and converges fast

in the direct space. Second term (Vrec) computes long-range interactions and converges

fast in the reciprocal space and the third term is a constant (V0). Thus, the electrostatic

potential (Vc) can be written as

0recdirc VVVV ++= (2.30)

where


( )∑∑∑∑=

N

ji, n n n ij

ijjidir

x y z,r

,βrerfcqq

2fV

nn

(2.31)

( )∑∑∑∑

⎥⎥⎦

⎤

⎢⎢⎣

⎡−π+⎟⎟

⎠

⎞⎜⎜⎝

⎛βπ

π=

x y zm m m2

j

2

N

j,ijirec

2-exp

qqV2

fVm

rrmmi.

and (2.32)

∑−=N

i

2i0 q

πfβV (2.33)

β determines the relative weight of the direct and reciprocal sums, m = (mx, my, mz) and

n = (nx, ny, nz) is the box index vector.

Particle-Mesh Ewald

Particle-Mesh Ewald (PME) is an improved and more efficient way for the

summation in the reciprocal space (Darden et al., 1993). Instead of directly summing

wave vectors, the charges are assigned to a grid using cardinal B-spline interpolation.

The PME algorithm scales as Nlog(N) and is substantially faster than ordinary Ewald

summation on medium to large systems. On very small systems, it might still be better

to use Ewald to avoid the overhead in setting up grids and transforms.

2.3.4.3 Bonded interaction terms

Bonded interactions are based on a fixed list of atoms. They are not exclusively

pairwise interactions, but include 3-body and 4-body interactions as well. There are

bond stretching (2-body), bond angle (3-body) and dihedral angle (4-body) interactions.

A special type of dihedral interaction (called improper dihedral) is used to force atoms

to remain in a plane or to prevent transition to a configuration of opposite chirality (a

mirror image). These particular components depend upon the force field type.

Bond Stretching

Bond stretching can be harmonic or anharmonic. The harmonic potential (Vb)

between two bonded atoms i and j, respectively are given by following expressions

( ) ( )2ijijbijijb brk

21rV −= (2.34)


where k is the force constant associated with the bond, r is the instantaneous bond

length and b is the reference value of bond length.

However, for the sake of computational efficiency various modifications of

above formulations are being made. Morse potential is an example of a harmonic bond

stretching potential. The functional form of the potential reads as

( ) ( )[ ]{ } 2ijijijijijmorse brexp1DrV −β−−= (2.35)

where Dij is the depth of the well in kJ mol-1, βij is the steepness of the well in nm-1 and

bij is the equilibrium distance in nm. βij can be expressed in terms of Dij as

ij

ijijij D2

μω=β where ωij is the fundamental vibrational frequency.

Bond angle

The bond angle vibration among a triplet of atoms i-j-k is represented by the

harmonic potential on the angle θijk in the following form

( ) ( ) 20ijkijkijkijka k

21V θ−θ=θ θ (2.36)

θ0 is the reference angle value.

Dihedral angles

The form of the potential for 4-body interactions employing dihedral angles

depends on the type of dihedral angle. Dihedral angle can be proper or improper.

Proper dihedrals

The potential (Vd) associated with periodic type proper dihedral angle (φ) has

functional form:

( ) ( )[ ]0ijkld ncos1kV φ−φ+=φ φ (2.37)

where kφ is the dihedral force constant and φ0 is the reference value of the angle.

Improper dihedral

Improper dihedral angles are important to maintain the planarity of the planar

groups (e.g. aromatic rings). The potential associated with improper dihedral is


harmonic. The functional form of the potential associated with improper dihedral

angles can be represented as

( ) ( )20ijklijklid k21V ξ−ξ=ξ ξ (2.38)

where Vid is the potential associated with the improper dihedral angle made up with

atoms i, j, k and l. kξ is the force constant for the dihedral angle, ξijkl is the angle

between planes (i, j, k) and (j, k, l) and ξ0 is the reference value of the angle.

2.3.4.4 Restraints

Special potentials are used for imposing restraints on the motion of the system

to either avoid disastrous deviations or include knowledge from experimental data.

Several restraints like position, angle, dihedral, distance and orientation can be applied.

Position restraints function discussed below is the most commonly used function for

biomolecular simulations.

Position restraints

Position restraints are potential functions used to restrain atoms at a fixed

reference position. These potentials are often used during solvent equilibration to avoid

too drastic rearrangement of critical regions of the protein molecule because of large

solvent forces from the unequilibrated solvent system. The potential form can be

written as following:

( ) ( ) ( ) ( )[ ]zZzkyYykxXxk21rV 2

iizpr

2ii

ypr

2ii

xpripr −+−+−= (2.39)

where k is force constant, xi and Xi are current and fixed positions of an ith atom. Force

constants in each direction can be used independently of each other and thus the

restraints can be restricted to a line or a plane also.

2.3.5 FORCE FIELDS USED

Various force fields have been developed for different purposes and with some

differences in their method of implementation of above discussed basic functional

forms of different parameters (Wang et al., 2001; Ponder and Case, 2003; Guvench and


MacKerell, 2008). AMBER (Assisted Model Building and Energy Refinement; Cornell

et al., 1995), CHARMM (Chemistry are HARvard Molecular Mechanics; MacKerell et

al., 1998), ENCAD (ENergy Calculation And Dynamics; Levitt et al., 1995),

GROMOS (GROningen Molecular Simulation package; Stocker and van Gunsteren,

2000), MMFF (Merck Molecular Force Field; Halgren, 1996a,b,c), OPLS (Optimized

Potentials for Liquid Simulations; Jorgensen et al., 1996) are some of the force fields

used for biomolecular simulations. In the current work, two force fields OPLS-AA and

AMBER03 have been used for MD simulations.

2.3.5.1 OPLS-AA

The OPLS-AA force field is among the most widely used force field for

biomolecular simulations. The force field parameters have been developed for liquid

simulations. The derivation and the validation of parameters involve statistical

mechanical approach like Monte Carlo simulations on various neutral and charged

molecules (Jorgensen and Tirado-Rives, 1988). The functional form of the OPLS force

field is given as

( ) abdihanglebondN VVVVrV +++= (2.40)

where

( ) 2eq

bondsrbond rrKV −= ∑ , ( ) 2

angleseqangle KV ∑ θ−θ= θ (2.41)

( )[ ] ( )[ ] ( )[ ] ( )[ ]φ−+φ++φ−+φ+= 4cos12

V3cos1

2V

2cos12

Vcos1

2V

V 4321dih (2.42)

∑∑⎪⎭

⎪⎬⎫

⎪⎩

⎪⎨⎧

+⎥⎥

⎦

⎤

⎢⎢

⎣

⎡

⎟⎟⎠

⎞⎜⎜⎝

⎛ σ−⎟

⎟⎠

⎞⎜⎜⎝

⎛ σε=

aon

i

bon

jij

ij

2ji

6

ij

ij

12

ij

ijijab f

reqq

rr4V (2.43)


2.3.5.2 AMBER03

AMBER is a family of force fields for molecular dynamics of biomolecules.

The functional form of the AMBER force field is

( ) ( ) ( ) ( )[ ]∑∑∑ γ−ω++θ−θ+−=torsions

n

angles

20a

bonds

20

bN ncos12

Vkll2

krV

∑ ∑−=

=

=

+= ⎪⎭

⎪⎬⎫

⎪⎩

⎪⎨⎧

πε+

⎥⎥

⎦

⎤

⎢⎢

⎣

⎡

⎟⎟⎠

⎞⎜⎜⎝

⎛ σ−⎟

⎟⎠

⎞⎜⎜⎝

⎛ σε+

1Nj

1j

Ni

1ji ij0

2ji

6

ij

ij

12

ij

ijij r4

eqqr

2r

(2.44)

where the terms represent the energy between covalently bonded atoms, the energy due

to the geometry of electron orbital involved in the covalent bonding, the energy for

twisting a bond due to bond order and neighboring bonds or lone pair of electrons and

the non-bonded energy between all atom pairs composed of two components van der

Waals and electrostatics.

2.3.6 WATER MODEL Proteins are surrounded by a large number of solvent molecules when

performing their functions. In addition, MD simulation characterizes protein internal

motions accurately, including the effects of solvent water (Hayward et al., 1993).

Because water plays essential roles in chemistry and biology, a large number of

classical water models have been proposed for example SPC, TIP3P, SPC/E, TIP4P,

TIP5P, etc. In the present thesis work, SPC (simple point charge) water molecules were

used.

Simple Point Charge water model

This is one of the most used water model in macromolecular MD simulations.

In this model, the water molecule has three centers of concentrated charge: a

predominance of positive charge on the hydrogen-atoms and excess negative charge on

the oxygen-atom (Berendsen et al., 1981). The results of the charge concentration and

the widened V-shaped bond angle are such that the permanent dipole moment of the

SPC water molecule has a value close to that measured in experiment.


2.3.7 LIGAND PARAMETERS A proper description of the molecular system requires various parameters and

constants, which depend on atom or bond types. These parameters have been largely

well characterized and optimized for protein molecules. However, the description of

ligand molecules requires special methods of parameterization. In the present thesis,

several small molecules were used to study the protein-ligand interactions. Most of

them are either nucleotides or pterin based molecules and their derivatives. Two

programs AMBER (Case et al., 2006) and Gaussian03 (Frish et al., 2004) were

extensively used to generate the charges, topology and parameters for ligands. Initially,

the program AMBER was used to generate the topology and parameters, which were

then converted to GROMACS format using a widely distributed Perl script

amb2gmx.pl. Subsequently, ab initio charges were computed using the program

Gaussian03. The electrostatic charge potentials (ESP) have been determined from

Hartree-Fock method employing the basis set of 6-31G*. The energies were calculated

using self-consistent field calculations using unrestricted open-shell Hartree-Fock

(UHF) wave functions (Pople and Nesbet, 1954). The basis set 6-31G* adds

polarization to all atoms and improves the modeling of core electrons and often

considered the best compromise of speed and accuracy. It is the most commonly used

basis set.

2.3.8 ENERGY MINIMIZATION METHODS If the starting conformation is very far from the equilibrium, the force

experienced on atoms will be very large and MD simulation fails to proceed. Moreover,

in most of the bimolecular simulations, the input model is an X-ray structure, which is

often devoid of hydrogen atoms. In MD simulations, hydrogen atoms are added to the

input model and it becomes necessary to relieve short contacts arising due to addition

of hydrogen atoms through energy minimizations. Among various energy minimization

protocols available in the literature, those, which have been used by the author, are

discussed below.


2.3.8.1 Steepest descent

Steepest descent is the simplest of the gradient methods. Although steepest

descent is certainly not the most efficient algorithm for searching, it is robust and easy

to implement.

Let r is the vector for all 3N coordinates. First the forces F and potential energy

are calculated. New positions are calculated by

( ) nn

nn1n h

|F|maxFrr +=+ (2.45)

where hn is the maximum displacement and Fn is the force. The notation max(|Fn|)

means the largest of the absolute values of the force components. The forces and

energy are again computed for the new positions

If (En+1 < En) the new positions are accepted and hn+1 = 1.2hn.

If (En+1 ≥ En) the new positions are rejected and hn = 0.2hn.

The algorithm stops when either a user specified number of force evaluations

have been performed (e.g. 100) or when the maximum of the absolute values of the

force (gradient) components is smaller than a specified value.

2.3.8.2 Conjugate gradient

On the contrary to steepest-descent method, the conjugate-gradient method is an

attempt to mend the problem by ‘learning’ from experience. It is slower than steepest

descent in the early stages of the minimization, but becomes more efficient closer to the

energy minimum. The parameters and stop criterion are the same as for steepest

descent.

2.3.8.3 L-BFGS

The method is a quasi-Newtonian algorithm for energy minimization. It utilizes

low-memory Broyden-Fletcher-Goldfarb-Shanno approach to approximate the inverse

of Hessian matrix by a fixed number of corrections from the previous step. The method

is particularly more useful for cases where the systems involve numerous short

contacts, which cannot be otherwise relieved by steepest-descent and conjugate-

gradient methods.


2.3.9 PERIODIC BOUNDARY CONDITION The classical way to minimize edge effects in a finite system is to apply

periodic boundary conditions. Atoms of the system to be simulated are put into a space-

filling box, which is surrounded by translated copies of it. Thus, there are no

boundaries of the system. The artifact caused by unwanted boundaries in an isolated

cluster is now replaced by the artifact of periodic conditions. GROMACS uses periodic

boundary conditions combined with the minimum image convention: only one, the

nearest image of each particle is considered for short-range non-bonded interaction

terms. For long-range electrostatic interactions this is not always accurate enough and

GROMACS therefore also incorporates lattice sum methods like Ewald Sum, PME and

PPPM.

2.3.10 VISUALIZATION Several programs from GROMACS and other sources were used to visualize

the trajectories generated from MD simulations. NGMX is the GROMACS trajectory

viewer, which reads a trajectory and an index file and plots a 3D structure of the

molecule on the standard X Window screen. The trjconv program of the GROMACS

suite can be used to convert the trajectories in PDB file format, which can be viewed

using PyMOL, RasMOL, etc. Graphs have been prepared using Xmgrace (Paul J.

Turner, Center for Coastal and Land-Margin Research Oregon Graduate Institute of

Science and Technology Beaverton, Oregon).

2.3.11 ANALYSIS There are several programs available in GROMACS to analyze the MD

trajectories. Some modules that were extensively used during the present work are

g_energy (to calculate various types of energies), make_ndx (generating indexes for

atoms, residues, etc.), g_analyze (analyzing data sets, generally output of another

program), g_rms (calculates rmsd's with a reference structure), g_rmsf (calculates

atomic fluctuations), g_confirms (fits two structures and calculates the rmsd), g_cluster

(clusters structures), g_mindist (calculates the minimum distance between two groups),

g_dist (calculates the distances between the centers of mass of two groups) g_bond


(calculates bond length distributions), g_angle (calculates distributions and correlations

for angles and dihedrals), g_hbond was used to analyze the hydrogen bonds, etc.

2.4 OTHER TECHNIQUES USED 2.4.1 ISOTHERMAL TITRATION CALORIMETRY

Although the investigations presented in this thesis primarily used X-ray

crystallographic techniques and MD, isothermal titration calorimetry (ITC) was used to

obtain the thermodynamic parameters of the protein-ligand interactions. ITC is the gold

standard for measuring biomolecular interactions. ITC simultaneously determines all

the binding parameters (n, Kb, ∆H and ΔS) in a single experiment. When substances

bind, heat is either generated or absorbed. ITC is a thermodynamic technique that

directly measures the heat released or absorbed during a biomolecular binding event.

Measurement of this heat allows accurate determination of binding constants (Kb),

reaction stoichiometry (n), enthalpy (∆H) and entropy (ΔS), thereby providing a

complete thermodynamic profile of the molecular interaction in a single experiment.

ITC goes beyond binding affinities and can also elucidate the mechanism of the

molecular interaction, so it has become the method of choice for characterizing

biomolecular interactions.

CHAPTER 3 Structure and Molecular Dynamics Studies of Three Active-

site Mutants of Bovine Pancreatic Phospholipase A2

CHAPTER 3: ACTIVE SITE MUTANTS OF BPLA2 88

3.1 INTRODUCTION As described in the introductory chapter, the enzyme PLA2 specifically

hydrolyzes the sn-2 fatty-acid acyl bond of phospholipids producing a free fatty acid

and a lysophospholipid in a calcium-dependent reaction (van Deenen and de Haas,

1964) and is involved in several cellular processes (van den Berg et al., 1995). The

enzyme PLA2 is widely distributed in snakes, lizards, bees and mammals. Irrespective

of its origin, the primary structures of most PLA2s show a high degree of homology

(Verheij et al., 1981). Bovine pancreatic PLA2 consists of 123 amino-acid residues with

a molecular weight of ~14 kDa. It contains five α-helices, two β-strands and seven

disulfide bonds (Dijkstra et al., 1978). The catalytic network of PLA2s is characterized

by a catalytic dyad (His48-Asp99) and a water molecule, which acts as a nucleophile

during the enzymatic reaction (Figure 3.1). The residue His48 is essential for the

enzymatic activity (Leatherbarrow and Fersht, 1987; Li and Tsai, 1993). Biochemical

studies on the single mutant H48N indicated that the mutant retains 6 × 10-5 of the

original catalytic activity of the wild-type enzyme (Li and Tsai, 1993). Thus, it is of

interest to study the crystal structure of the single mutant H48N. On the other hand, the

residue Asp49 is required for retaining the functionally important calcium ion in the

active site of PLA2 (van den Bergh et al., 1988; Davidson and Dennis, 1990; Li et al.,

1994). The calcium ion is liganded by three backbone oxygen atoms from Tyr28,

Gly30 and Gly32, both the carboxylate oxygen atoms of Asp49 and two water

molecules (Dijkstra et al., 1981a; Sekar, 2007). Li and coworkers used site-directed

mutagenesis and NMR studies to provide insights into the structural and functional

roles of the highly conserved residue Asp49 and observed that the mutants D49N and

D49K do not bind the calcium ion, whereas the mutant D49E binds the calcium with

12-fold weaker affinity (Li et al., 1994; Sekar et al., 1999). Furthermore, structural

analysis using two-dimensional proton NMR indicated no global perturbation in the

single mutants D49N and D49K. Thus, the aim of the proposed work was to solve the

three-dimensional crystal structures of the active-site mutants H48N, D49N and D49K

in order to gain a better understanding of the nature of the structural perturbations

caused by Asp49 mutants and to study the effect of asparagine at position 48. In

addition, molecular-dynamics (MD) simulations of the active-site mutants and three in

silico generated mutants (1MKT_H48N, 1MKT_D49N and 1MKT_D49K) were


carried out to observe the effect of these mutants on calcium binding. The root-mean-

square deviation (r.m.s.d.) plot for Cα atoms of all the simulations is given in Figure

3.2.

Figure 3.1 Active-site hydrogen-bonding network of the trigonal form of bovine pancreatic phospholipase A2. The protein molecule used for the illustration is that of BPLA2 (PDB-id: 1MKT; Sekar et al., 1998a).


3.2 RESULTS AND DISCUSSION 3.2.1 H48N MUTANT

Previous NMR studies on the H48N mutant suggested that the mutation at

His48 affects the overall tertiary structure (Li and Tsai, 1993). In contrast to the NMR

studies, the tertiary structure is intact in the crystal structure of H48N and is highly

similar to that of the trigonal wild-type PLA2 (PDB-id: 1MKT; Sekar et al., 1998a),

with a root-mean-square deviation (r.m.s.d.) of 0.3 Å (for the backbone atoms). The

five protein ligands of the active-site calcium ion superpose with an r.m.s.d. of 0.15 Å

with the wild type enzyme. The calcium-ligand distances vary between 2.26 Å and 2.60

Å, with an average of 2.46 Å, which is slightly higher than the average of 2.39 Å

obtained from the atomic-resolution structure (Steiner et al., 2001). This is probably a

result of the low resolution of the present structure. An MPD molecule is observed in

Figure 3.2 Root-mean-square deviation (r.m.s.d.) plot of Cα atoms from the starting structure of (a) two native simulations (PDB-ids: 1MKT and 1UNE), (b) four simulations related to H48N mutant, (c) two simulations related to D49N mutant and (d) two simulations related to D49K mutant. The two conformations in the case of H48N mutant denote the position of Oδ1 of Asn48 towards the catalytic water and Asp99, respectively.


the vicinity of the active site and is hydrogen bonded to the equatorial water molecule

(W5) as found previously (PDB-id: 1VL9; Sekar et al., 2005). A water molecule (W13)

is hydrogen bonded to the structural water W11 (Figure 3.3). This water molecule is

observed in the orthorhombic wild-type PLA2 (PDB-id: 1UNE; Sekar and

Sundaralingam, 1999) and the single mutants H48Q (PDB-id: 1KVW; Sekar et al.,

1999) and D49N (PDB-id: 2ZP3; present work). A previous study of the single mutant

H48Q (Sekar et al., 1999) revealed that Gln48 Nε2 is hydrogen bonded to Asp99 Oδ1

and Gln48 Oε1 is hydrogen bonded to the catalytic water molecule (W6). However, in

the single mutant H48N both catalytic water molecules (W6 and W7) are hydrogen

bonded to Asn48 Oδ1 (Figure 3.4).

Interestingly, the hydrogen bond between Asn48 Nδ2 and Asp99 Oδ1 is retained

in the structure. However, unlike His48, the mutant Asn48 cannot act as a base to

accept a proton from the catalytic water (W6). Although, solution studies showed a low

(6×10-5) residual catalytic activity in the H48N mutant compared with the wild-type

Figure 3.3 Stereoview of the active-site hydrogen-bonding network of the H48N mutant. The 2Fo-Fc electron-density map for the mutated residue asparagine is contoured at 1.0σ.


enzyme (Li and Tsai, 1993). It is widely accepted that His48 is involved in the

activation of the catalytic water (W6) to initiate the reaction. However, the possibility

exists that a similar role could be fulfilled by Asp49, which is on the other side (Figure

3.1) of the catalytic water (W6...Asp49 Oδ1 = 2.9 Å). The very low catalytic activity in

the present mutant H48N is presumably the consequence of the acceptance of a proton

by the residue Asp49 from the catalytic water W6 (Figure 3.4).

The MD simulations were performed after solvating the crystal structure of the

H48N mutant for a time period of 3 ns. Careful examination of the average structure of

the protein molecule and water molecules, retained in the structure obtained from the

Figure 3.4 Active-site superposition of wild type BPLA2 (PDB-id: 1G4I, green), H48Q mutant (PDB-id: 1KVW, blue) and H48N mutant (PDB-id: 2ZP4, purple) is shown to compare the orientation of the residues His48 and its mutants, Asp49, Asp99 and catalytic water molecule (W6). The distances are given in Angstrom (Å).


last step of the MD trajectories, shows that five water molecules provide coordination

to the calcium ion. Two of these occupy the positions of axial (W12) and equatorial

(W5) water molecules required for the calcium ion coordination generally found in the

crystal structure of PLA2. Interestingly, after the MD simulation, three water molecules

are found at the positions of W7, W11 and W13, as in the crystal structure. In order to

verify the chosen orientation of the amide group of the mutated residue Asn48 (Oδ1

hydrogen bonded to W6, Nδ2 hydrogen bonded to Asp99 Oδ1), a molecular-dynamics

(MD) simulation was also carried out for the alternative orientation. The amide group

was found to be flipped to the chosen orientation after 2.76 ns of simulations.

3.2.2 D49N MUTANT The main-chain atoms superpose well with the trigonal form of the wild-type

PLA2, with a root-mean-square deviation (r.m.s.d.) of 0.3 Å. The functionally important

calcium ion is absent possibly owing to the loss of the negative charge of the mutated

residue asparagine, which reduces the affinity of the enzyme for the calcium ion

(Figure 3.5). However, the catalytic framework along with an extra water molecule

(W13) is intact. A Tris molecule is observed and its oxygen atoms are hydrogen bonded

to the backbone oxygen atoms of Phe106, Ser107 and Val109. Furthermore, an MPD

molecule is also observed near the active-site mouth as in another structure (PDB-id:

1VL9; Sekar et al., 2005). Examination of the MD trajectories of the single mutant

D49N shows water molecules occupying similar positions to W7 (one of the histidine

water molecules), W11 (structural water) and its neighbor W13 in the dynamically

equilibrated system as observed in the crystal structure. Subsequently, Asp49 was

mutated to Asn (in silico) in the wild-type structure (PDB-id: 1MKT; Sekar et al.,

1998a) in order to observe the movement of the calcium ion in the 1MKT_D49N

mutant during the MD simulations. The calcium ion was found to move approximately

5 nm away from the active site of the enzyme (Figure 3.6), well into the solvent region.

Figure 3.6 shows the protein-calcium ion interaction energy and the distance between

the residue Asp49 and calcium ion as a function of time. For comparison, the protein-

calcium ion interaction energy and the distance between the protein and calcium ion for

the H48N mutant are also shown. From the graph, it is clear that the interaction energy


and the distance between the calcium ion and protein are stable throughout the

simulation for the mutant H48N in contrast to the single mutant D49N.

Figure 3.5 Stereoview of the active-site hydrogen-bonding network of the D49N mutant.

The 2Fo-Fc electron-density map of Asn49 is contoured at 1.0σ.


3.2.3 D49K MUTANT Overall tertiary structure of the D49K mutant is similar to that of the wild type

with a root-mean-square deviation (r.m.s.d.) of 0.4 Å (backbone atoms). Superposition

of the backbone atoms of the protein ligands (Tyr28, Gly30, Gly32 and Lys49) with the

wild-type structure shows a greater change with an r.m.s.d. of 0.6 Å. The large

deviation is a consequence of the longer side chain of the mutated residue lysine. The

calcium ion and three water molecules (W5, W7 and W12) are found to be absent. Only

the catalytic water molecule W6 is present. It is hydrogen bonded to His48 Nδ1. In fact,

the atom Nζ of Lys49 occupies the position of the calcium ion. Furthermore, Nζ of

Lys49 is hydrogen bonded to the carbonyl oxygen atoms of Tyr28 and Gly30 and the

catalytic water W6 (Figure 3.7). However, the catalytic framework comprising Ala1,

Tyr52, Pro68, Asp99 and the structural water W11 is preserved. Although the active-

site calcium ion is not present in the crystal structure of D49K, it was retained in the in

Figure 3.6 Graphs show (a) protein-calcium ion interaction energy and (b) protein-calcium ion distance during the molecular-dynamics simulation of 3 ns. Graphs were generated using the program Xmgrace.


silico 1MKT_D49K mutant essentially to observe the effect of the mutated residue

lysine on the calcium during molecular dynamics. The mutated residue lysine was

modelled using the program COOT (Emsley and Cowtan, 2004) in such a way that

there is no short contact with any other atoms including the active-site calcium ion. The

MD calculations of the in silico 1MKT_D49K mutant shows that the interaction energy

of the calcium ion with protein decreases as a function of simulation time. Therefore,

the calcium ion is considered to have moved away from the active site. Interestingly, as

observed in the crystal structure, the MD averaged structure of 1MKT_D49K shows

that Nζ atom of the mutated residue lysine occupies a similar position to that of the

calcium ion.

Figure 3.7 Stereoview of the active-site hydrogen-bonding network of the D49K mutant. The 2Fo-Fc electron density of the mutated residue lysine is contoured at 1.0σ. Only two active-site water molecules (W6 and W11) of the five commonly found water molecules are observed. The distance between Nζ (Lys49) and the catalytic water W6 is 3.34 Å.


In summary, the comparison of the wild type enzyme, D49E and D49N mutants

show that in D49N mutant, only two water molecules (W6 and W11) are conserved.

Also due to the loss of calcium ion in the active site, one water molecule is observed to

be present at the position of one of the carboxylate oxygen atoms of Asp49 (Figure

3.8). Similarly, in D49K mutant, only two water molecules (W6 and W11) are found in

the active site, whereas the ε–amino group of Lys49 occupies the position of the

calcium ion (Figure 3.9).

Figure 3.8 Active-site superposition of the wild type (PDB-id: 1G4I, green), D49E mutant

(PDB-id: 1KVY, blue) and D49N mutant (PDB-id: 2ZP3, purple).


3.2.4 ACTIVE SITE AND SURFACE LOOP RESIDUES The active site of the H48N mutant is not disturbed due to the presence of the

calcium ion. However, in both Asp49 mutants the active site is perturbed owing to the

absence of the calcium ion and water molecules. In the Asp49 mutants, two calcium-

coordinating residues Gly30 and Gly32 have moved away from the active site due to

the absence of the calcium ion (Figure 3.10). Furthermore, the movement of Gly32 in

the case of the D49K mutant may be a consequence of the longer side chain of Lys49.

As expected, the axial and equatorial calcium-coordination water molecules (W12 and

W5) are missing from the Asp49 mutants. It is noteworthy that the equatorial calcium-

coordinated water molecule (W5) is involved in hydrogen bonding (Figure 3.1) to the

backbone nitrogen atom of Gly30 (Sekar and Sundaralingam, 1999). Thus, the

Figure 3.9 Active-site superposition of the wild type (PDB-id: 1G4I, green), D49E mutant (PDB-id: 1KVY, blue) and D49K mutant (PDB-id: 2ZP5, purple).


functionally important calcium ion is also essential for the integrity of the active site.

The electron density (2Fo-Fc) for the 11 surface-loop residues (60-70) in all three

mutants is not clear at the 1.0σ level. However, the surface loop is modeled at a low

contour (0.4σ) level in all three mutants. Figure 3.11 shows a comparison of the surface

loop of all three mutants and two forms (orthorhombic and trigonal) of the wild-type

structure. Most of the crystal structures of BPLA2 determined to date indicate that the

surface loop is ordered either in the presence of inhibitors or in the presence of a second

calcium ion or both, with the exception of two structures (PDB-ids: 1UNE, Sekar and

Sundaralingam, 1999 and PDB-id: 1G4I, Steiner et al., 2001). It is generally observed

that the number of water molecules near the surface-loop residues is greater in the case

of ordered structures.

Figure 3.10 Stereoview of the active-site residues superposition and water molecules of the single mutants (H48N, red; D49N, green; D49K, blue) along with the trigonal form of the wild-type enzyme (PDB-id: 1MKT, yellow).


The average fluctuation of the Cα atom of each residue throughout the MD

simulation reveals that the calcium-binding loop (Gly30 and Gly32) is highly flexible

in addition to the surface-loop region in all three mutants (Figure 3.12). However, in

the Asp49 mutants the large fluctuation of Gly30 and Gly32 is primarily because of the

mutation at residue 49 and the loss of the calcium ion in the active site. Figure 3.12

clearly shows the large movement of residues Gly30 and Gly32 in the D49N (green)

and D49K (blue) mutants. Furthermore, in the MD average structures water molecules

occupied the void created by these two residues.

Figure 3.11 Comparison of the surface-loop region in the three active-site mutants (H48N, red; D49N, green; D49K, blue) with the orthorhombic (PDB-id: 1UNE, magenta) and trigonal (PDB-id: 1MKT, yellow) forms of the wild type.


3.2.5 INVARIANT WATER MOLECULES The numbers of water molecules located in the H48N, D49N and D49K mutant

crystal structures are 132, 139 and 127, respectively. Approximately 85% of the water

molecules are found in the first hydration shell in all three mutants (Shanthi et al.,

2003). Invariant water molecules are identified upon the superposition of the H48N and

D49K structures on the structure of D49N to be those which lie within 0.5 Å of the

equivalent water molecules in the fixed structure. Since the number of water molecules

is greater in the D49N mutant, it was used as a fixed molecule. This analysis revealed

that a total of 41 water molecules (including the structural water W11 and the catalytic

water W6) are invariant, with average B factors of 29.47, 23.82 and 31.18 Å2 for the

H48N, D49N and D49K mutants, respectively. The structural water molecule (W11)

and the catalytic water molecule (W6) are present in all three mutants (Figure 3.13).

Furthermore, all 41 invariant crystallographic water molecules are present in the

corresponding MD average structure computed between 2 and 3 ns. These invariant

water molecules form 127, 136 and 128 hydrogen bonds in the H48N, D49N and D49K

mutants, respectively. Similar analysis with the MD average structures reveals 192, 198

Figure 3.12 Graph depicts the average fluctuation of all 123 Cα atoms using 1500 structures computed every 2 ps. The molecular-dynamics simulations were performed for 3 ns.


and 193 interactions, respectively. Interestingly, the hydrogen-bonding interactions of

the invariant water molecules are also found in the MD average structure. An increased

number of hydrogen bonds in the MD simulated structures are found since they contain

more water molecules (643, 638 and 664 in H48N, D49N and D49K, respectively). A

total of nine water molecules (excluding W6 and W11) are found in the core of the

enzyme. Interestingly, these nine buried water molecules were observed in almost all

crystal structures of bovine and porcine pancreatic PLA2s, indicating possible

involvement in the folding of the enzyme. The remaining 30 invariant water molecules

are on the surface of the enzyme and are likely to be involved in providing stability to

the enzyme by hydrating surface polar residues.

3.3 CONCLUSION The overall tertiary structure of all three mutants is similar to that of the wild-

type enzyme. However, the active site is disturbed in the case of the Asp49 mutants,

whereas it is intact in the H48N mutant. Thus, the crystal structures and molecular-

dynamics (MD) simulations of the three single mutants confirm that the residue Asp49

is important for both calcium binding and the integrity of the active site. On the other

hand, His48 is not crucial for the stability of the active site. However, it is important for

Figure 3.13 Stereoview of invariant water molecules in H48N (red), D49N (green) and D49K (blue) mutants. The protein molecule shown here corresponds to the H48N mutant (PDB-id: 2ZP3).


the catalytic activity of the enzyme. It is clear that the active-site framework remains

intact in all the mutant structures and this is further supported by molecular dynamics.

Furthermore, it is interesting to note that the structural water W11 is retained in its

position during molecular dynamics. This suggests the importance of this water

molecule in maintaining the framework intact. Approximately, 20% of the

crystallographic water molecules are conserved in all the three mutants. In addition,

water molecules occupy similar positions in the average structures obtained from

molecular dynamics.

3.4 MATERIALS AND METHODS 3.4.1 PROTEIN PURIFICATION AND CRYSTALLIZATION

Professor M.-D. Tsai of the Chemistry Department, Ohio State University

supplied the three active-site mutants reported here. The procedures for the purification

of these mutants were similar to those described elsewhere (Noel et al., 1991; Dupureur

et al., 1992a,b; Li and Tsai, 1993). The proteins were concentrated to ~15 mg ml-1 in 50

mM Tris-HCl buffer pH 7.2 and 5.0 mM CaCl2. Crystals of all three mutants were

obtained using the hanging-drop vapor-diffusion method at room temperature (293 K).

In the case of the H48N mutant, the droplet consisted of 5 µl protein solution and 2 µl

50% (v/v) 2-methyl-2,4-pentanediol (MPD) and was equilibrated against 60% (v/v)

MPD in 0.5 ml reservoir solution. In the case of the D49N mutant, the droplet

contained 5 µl protein solution and 1 µl 60% (v/v) MPD. In the case of the D49K

mutant, the droplet contained 5 µl protein solution and 3 µl 60% (v/v) MPD. In both the

cases of D49N and D49K mutants, the droplets were equilibrated against 70% (v/v)

MPD. The X-ray diffraction-quality crystals for all three mutants appeared within a

week (Figure 3.14).


3.4.2 DATA COLLECTION AND PROCESSING The intensity data for all three mutants were collected at 100 K using a MAR

345 imaging-plate detectors mounted on a Rigaku RU-300 generator (operated at 40 kV

and 80 mA) using the home source available at the Molecular Biophysics Unit, Indian

Institute of Science, Bangalore, India. The data were processed and scaled using

DENZO and SCALEPACK from the HKL suite (Otwinowski and Minor, 1997). The

intensities were converted to structure factors using the program TRUNCATE from the

CCP4 suite (Collaborative Computational Project, Number 4, 1994). Crystal data-

collection statistics are given in Table 3.1.

Figure 3.14 Crystal images of (a) H48N mutant, (b) D49N mutant and (c) D49K mutant.


3.4.3 STRUCTURE REFINEMENT, VALIDATION AND ANALYSIS The crystals were isomorphous to the trigonal form of the recombinant enzyme with

space group P3121 and unit-cell parameters a = b = 46.78, c = 102.89 Å (PDB-id:

1MKT; Sekar et al., 1998a). Even though the crystals are isomorphous to the native

structure, the three-dimensional structures of the three mutants were solved using the

molecular-replacement program Phaser (Read, 2001; McCoy et al., 2007). The atomic

coordinates of the trigonal form of the wild type (PDB-id: 1MKT; Sekar et al., 1998a)

were used as the search model for the molecular-replacement calculations. The log

likelihood gain (Z-score) for H48N, D49N and D49K were 1188.35 (41.44), 978.67

(33.21) and 1170.18 (34.89), respectively.

3.4.3.1 Refinement of H48N mutant

The molecular-replacement solution was used as the initial model without the mutated

residue Asn48. A total of 10% (1073) of the reflections were set aside for Rfree

calculations (Brunger, 1992). After a total of 50 cycles of rigid-body refinement

followed by 50 cycles of positional refinement using CNS (Brunger et al., 1998), Rwork

and Rfree were 30% and 31%, respectively, for 8976 reflections in the resolution range

30.0-1.9 Å. Subsequently, the mutated residue Asn48 was modelled and fitted using

difference electron-density (2Fo-Fc and Fo-Fc) maps and the model was subjected to

simulated annealing by heating the system to 3000 K and slowly cooling to 100 K in 10

K steps. Strong electron density was observed for the functionally important active-site

calcium ion and a chloride ion and these were added to the refined model. At this stage,

Rwork and Rfree dropped to 24% and 28%, respectively. A large electron density (up to

7σ in the Fo-Fc map) near the C-terminus was observed which was identified as a

calcium ion as found previously (Sekar et al., 2006b). Water molecules were located

and added using difference electron-density (2Fo-Fc and Fo-Fc) maps with peak heights

greater than 0.8σ and 2.5σ, respectively and at hydrogen-bonding distances of 3.5 Å or

less to protein atoms or other water molecules. The final refined model contains 955

protein atoms, two calcium ions, one chloride ion, 132 water oxygen atoms and one

MPD molecule.


Table 3.1 X-ray data-collection and refinement statistics for all three active-site mutants of bovine pancreatic phospholipase A2. Values in parentheses are for the highest resolution shell.

H48N D49N D49K Data collection and processing

Wavelength (Å) 1.5418 1.5418 1.5418

Temperature (K) 100 100 100

Space group P3121 P3121 P3121

Unit-cell parameters (Å) a=45.82, c=101.50 a=45.79, c=101.94 a=46.19, c=101.92

Crystal dimensions (mm) 0.8×0.3×0.3 0.6×0.5×0.4 0.6×0.3×0.2

Resolution (Å) 30.0-1.9 (1.97-1.90) 30.0-1.9 (1.97-1.90) 30.0-1.9 (1.97-1.90)

Observed reflections 116803 116580 69065

Unique reflections 10227 (970) 9760 (918) 10359 (1011)

Completeness (%) 99.7 (99.5) 94.8 (93.7) 98.8 (99.6)

Matthews coeff. (Å3 Da-1) 2.2 2.2 2.2

Solvent content (%) 44.0 44.2 44.2

Multiplicity 12.7 (11.8) 12.6 (12.1) 7.0 (6.6)

I/σ(I) 39.3 (6.0) 33.5 (14.3) 30.7 (6.4)

Rmerge# (%) 5.8 (41.8) 5.4 (14.6) 5.3 (30.8)

Refinement statistics

Rwork (%) 17.8 19.1 19.7

Rfree (%) 20.1 23.8 23.5

Protein Model

Protein atoms 955 957 958

Water molecules 132 139 127

Metals (Ca2+) 2 1 1

Others 2 3 1

Deviations from ideal geometry

Bonds lengths (Å) 0.004 0.006 0.004

Bond angles (°) 1.2 1.3 1.3

Dihedral angles (°) 21.8 22.3 22.8

Improper angles (°) 0.69 0.79 0.79

Average temperature factors (Å2)

Protein atoms 28.5 21.6 32.2

Water molecules 39.8 36.7 42.6

Metals (Ca2+) 28.5 23.5 36.1

Others 30.0 43.8 22.8

Ramachandran plot (%)

Most favored 92.7 92.7 90.9

Additionally allowed 7.3 7.3 9.1 # Rmerge = ΣhΣi |I(h)i - <I(h)>| / ΣhΣiI(h)i, where I(h) is the intensity of reflection h, Σh is the sum over all reflections and Σi is the sum over i measurements of reflection h.


Although correct assignments of the atoms of the amide group (side-chain

atoms Oδ1 and Nδ2) of asparagine are considered to be difficult, the widely accepted

hydrogen-bonding environment schemes were followed. Furthermore, positional and B-

factor refinements were carried out for both orientations of the amide group and that

with the low average temperature factor (19.84 Å2 compared with 20.13 Å2) was

considered in further analysis. Subsequently, this was verified using the programs

MolProbity (Davis et al., 2004) and HBPLUS (McDonald and Thornton, 1994), which

suggested the preferred orientation (W6 hydrogen bonded to Asn48 Oδ1 and Asp99 Oδ1

hydrogen bonded to Asn48 Nδ2) to be that identified above, in contrast to previous

studies (Li and Tsai, 1993).

3.4.3.2 Refinement of D49N and D49K

A similar approach was followed to refine the other two mutant structures

(D49N and D49K). After initial refinement of the model, it was observed that there was

no electron density for the functionally important calcium ion in the active site. As

observed in the H48N mutant, there was a strong electron density near the C-terminus

of the Asp49 mutant structures (up to 12σ and 7σ for D49N and D49K, respectively in

the Fo-Fc map). In both of the Asp49 mutants, a chloride ion was found near Lys12.

In summary, for all of three mutants, the program CNS (Brunger et al., 1998)

was used for the refinement. The molecular-modelling program COOT (Emsley and

Cowtan, 2004) was used to display the electron-density maps for model fitting and

adjustments. All atoms were refined with unit occupancies. Simulated-annealing omit

maps were calculated using the program CNS and were used to check or correct the

final protein models using the modelling program COOT. The simulated annealing

omit maps calculated at the end of the refinement were also used to check the final

protein models. The program PROCHECK (Laskowski et al., 1993) was used to check

and validate the quality of the final refined models. The final refined models were

checked and validated using the web-server ADIT before depositing to RCSB-PDB.

The final atomic coordinates and structure factors for H48N (PDB-id: 2ZP4), D49N

(PDB-id: 2ZP3) and D49K (PDB-id: 2ZP5) have been deposited in the RCSB Protein

Data Bank (Berman et al., 2000). Figures were generated using the program PyMOL

(DeLano Scientific LLC; http://www.pymol.org). The web-based programs PDB


Goodies (Hussain et al., 2002) and 3dSS (Sumathi et al., 2006) were used for the

analyses and superposition, respectively. The details of the refinement of all three

mutants are given in Table 3.1.

3.4.4 MOLECULAR DYNAMICS SIMULATION Energy minimization and simulations were performed using the program GROMACS

v.3.3 (van der Spoel et al., 2005) with the OPLS-AA force field (Jorgensen et al., 1996;

Kaminski et al., 2001). The recombinant wild-type PLA2 structure (PDB-id: 1MKT)

was mutated in silico using the program COOT at positions 48 (for H48N) and 49 (for

D49N and D49K) with the corresponding residues and these structures are abbreviated

as 1MKT_H48N, 1MKT_D49N and 1MKT_D49K, respectively. However, the

functionally important calcium ion was retained in all three in silico mutants in order to

observe the effect of the mutation on the calcium during molecular dynamics. The in

silico mutants were checked for stereochemistry. The crystallographic water molecules

were removed and protein models were solvated with the SPC (simple point charge)

water model using the genbox module available in the GROMACS suite. The box size

of the system was 6.4×6.4×6.4 nm. Sodium and chloride ions, wherever needed, were

used to neutralize the overall charge of the system. Simulations utilized NPT ensembles

with isotropic pressure coupling (τp = 0.5 ps) to 1 bar and temperature coupling (τt =

0.1 ps) to 300 K. Parrinello-Rahman and Nose-Hoover coupling protocols were used

for pressure and temperature, respectively. Energy minimization was performed using

the conjugate-gradient method for 200 ps with the maximum force-field cutoff being 1

kJ mol-1 nm-1. Long-range electrostatics were computed using the Particle Mesh Ewald

(PME; Darden et al., 1993) method and Lennard-Jones energies were cut off at 1.0 nm.

Bond lengths were constrained with the LINCS (Hess et al., 1997) algorithm.

Simulations were carried out at a dielectric constant of unity (a value used under an

explicit solvent MD simulations). Parameters and topology files for MPD and Tris

molecules were generated using the PRODRG web server (Schuttelkopf and van

Aalten, 2004). Analyses were primarily performed with tools available in the

GROMACS suite. The average structures used for comparison and analyses were

calculated using ensembles generated between 2 and 3 ns. The structures were

computed every 2 ps.

CHAPTER 4 Structural and Functional Role of Water Molecules in Bovine

Pancreatic Phospholipase A2: A Data-mining Approach

CHAPTER 4: ROLE OF INVARIANT WATER MOLECULES 110

4.1 INTRODUCTION Water molecules are quantitatively considered to be an integral part of

biomolecular systems and are crucial in the protein-folding process and in their

function (Cheung et al., 2002; Halle, 2004; Papoian et al., 2004; Eisenmesser et al.,

2005; Smolin et al., 2005). It is also known that protein hydration plays an important

role in biological processes (Otting et al., 1991; Franks, 2002; Chaplin, 2006; Zhang et

al., 2007) and that hydration forces are responsible for the packing and stabilization of

three-dimensional protein structure (Raschke, 2006). In addition, water molecules are

found to be involved in many hydrogen-bonding networks (Meyer, 1992). The common

hydrophilic nature of the interfaces of protein-protein, protein-DNA and protein-ligand

complexes and the abundance of water molecules at the interface suggest that water

molecules are an indispensable component of biomolecular recognition and self-

assembly (Tame et al., 1996; Jayaram and Jain, 2004).

The location of many of these water molecules is conserved in identical or

similar positions in the crystal structures of highly homologous proteins and their

spatial conservation is common in active sites and metal coordination as well as in

polar cavities. Furthermore, water molecules deeply buried in the core of the protein are

considered to be important in the folded structure and make strong hydrogen bonds to

polar groups. They are therefore believed to tighten the protein molecules. It is well

known that such ordered water molecules are best identified using crystallographic

methods (X-ray crystallography or neutron diffraction) or in special cases by NMR

spectroscopy (Otting, 1997). In general, for a water molecule to be described as

ordered, it must make at least one contact (with a maximum distance of 3.5 Å) to the

polar atoms of the protein molecule (Baker and Hubbard, 1984). With the availability

of a large number of three-dimensional protein structures at higher resolutions, it is now

possible to analyze and compare water structures. As indicated in the literature, studies

have been performed on the water structures of T4 lysozyme (Zhang and Matthews,

1994), ribonuclease A (Kishan et al., 1995), hen egg-white lysozyme (Biswal et al.,

2000), aspartic proteinases (Prasad and Suguna, 2002), legume lectins (Loris et al.,

1994) and serine proteases (Sreenivasan and Axelsen, 1992; Krem and Enrico, 1998).

In these studies, homologous protein structures solved under varying experimental

conditions with different solvent contents and with minor mutations were studied and


resulted in the identification of a number of water molecules that are invariant. These

findings suggest that the positions of certain water molecules must be conserved for

structural and/or functional reasons. Therefore, we planned to carry out a similar study

by analyzing the three-dimensional crystal structures of bovine pancreatic

phospholipase A2 (BPLA2; Figure 4.1).

As mentioned in the introductory chapter, the enzyme phospholipase A2 (PLA2,

EC 3.1.1.4) catalyzes the hydrolysis of the sn-2 fatty-acid ester bond of phospholipids

producing a free fatty acid and a lysophospholipid in a calcium-dependent reaction (van

Deenen and de Haas, 1964). The enzyme PLA2 is found in almost all organisms and is

involved in several physiological cellular processes (van den Berg et al., 1995). The

enzyme PLA2 consists of 123 amino-acid residues (molecular weight of ~14 kDa) and

contains seven disulfide bonds. In total, 32 recombinant bovine pancreatic PLA2

structures are available. These include four crystal forms (P2, C2, P212121 and P3121)

and native, inhibitor complexes and mutant structures. The present study aims to better

Figure 4.1 Overall three-dimensional structure of bovine pancreatic phospholipase A2(PDB-id: 1MKT, trigonal form; Sekar et al., 1998a). The disulfide bonds and the functionally important calcium ion are shown in Figure 1.2.12.


understand the role and the involvement of invariant water molecules in the three-

dimensional architecture and function of the enzyme BPLA2. To further strengthen our

findings, molecular dynamics (MD) studies have also been carried out on these

structures and the results are compared with the crystal structures.

4.2 RESULTS AND DISCUSSION 4.2.1 ALL 24 INVARIANT WATER MOLECULES

The relevant details of all 25 crystal structures used in the present study are

given in Table 4.1. A total of 24 invariant water molecules were identified and were

further classified into three groups based on their location in the three-dimensional

structure: Cluster-1, Cluster-2 and Cluster-3 (Figure 4.2). In addition, the details of the

interaction(s) of these invariant water molecules are listed in Table 4.2. The high-

resolution crystal structure of BPLA2 (PBD-id: 1G4I, Steiner et al., 2001) was taken as

the reference structure throughout the discussion and unless otherwise mentioned, the

numbering scheme used for the residues and water molecules corresponds to that of the

reference structure. Although the identification of a few well-conserved water

molecules was straightforward, difficulties were encountered with regard to the others.

Such difficulties were anticipated even when analyzing the same protein in different

crystal forms with different resolutions and space groups. Critical examination of

protein-water interaction(s) and visual inspection methods were used to identify the

invariant water molecules. A close examination of Table 4.2 reveals that the invariant

water molecules are mainly hydrogen bonded to the main-chain polar atoms of the

protein molecule. Furthermore, analysis of these interactions suggests that a significant

number of water molecules are conserved in the vicinity of the active site and at the

interfacial site of the enzyme. As expected, no invariant water molecules are observed

near the surface loop (residues 60-70), except for the structural water molecule

(hereafter referred to as IW3) which is hydrogen bonded to the backbone oxygen atom

of Pro68. Water-numbering scheme (first row), normalized B factor (second row) and

solvent-accessible surface area (third row) of these invariant water molecules are listed

in Tables 4.3, 4.4 and 4.5. In addition, the residence frequency for each water molecule

calculated using the trajectories obtained from the MD simulations are given in Tables


4.3, 4.4 and 4.5. The root-mean-square deviation (r.m.s.d.) plot for Cα atoms of all the

simulations is given in Figure 4.3.

Table 4.1 List of the three-dimensional crystal structures of bovine pancreatic phospholipase A2 used in the present analysis.

A# B C D E F G+ 1 1G4I 0.97 P212121 9.4/NA 247 2 1VL9 0.97 P2 11.4/13.4 243 K53,56,121M 3 2BCH 1.1 P3121 10.4/13.3 228 K53,56,120,121M 4 2BAX 1.1 P3121 11.4/15.6 209 K53,56M 5 1VKQ 1.6 P3121 17.9/21.7 165 K53,56,120M 6 2ZP3 1.9 P3121 19.1/23.8 139 D49N 7 1UNE 1.5 P212121 18.4/22.8 134 8 2ZP4 1.9 P3121 17.8/20.1 132 H48N 9 2ZP5 1.9 P3121 19.7/23.5 127 D49K

10 1GH4 1.9 P2 19.6/25.9 125 K56,120,121M 11 2B96 1.7 P3121 20.2/22.1 125 ANN/K53,56,121M 12 2BD1 1.9 C2 20.7/24.0 218$ K53,56,120,121M 13 1BP2 1.7 P212121 17.1/NA 106 14 1C74 1.9 P3121 18.9/22.4 106 K53,56M 15 1MKT 1.72 P3121 19.5/28.4 106 16 1KVX 1.9 P212121 20.0/31.3 98 D99A 17 1MKV 1.89 P3121 18.0/NA 88 TSA 18 1FDK 1.91 P3121 18.4/28.0 86 MJ33 19 1O3W 1.85 P3121 19.3/23.2 85 K53,56,120M 20 1CEH 1.9 P3121 18.5/NA 81 D99N 21 1IRB 1.9 P3121 19.2/NA 81 K120,121A 22 1MKU 1.8 P212121 19.6/NA 80 Y52,73F, D99N 23 1MKS 1.9 P3121 18.6/NA 77 Y52,73F, D99N 24 1KVY 1.9 P3121 19.8/27.7 70 D49E 25 1KVX 1.95 P3121 20.9/31.4 68 H48Q

A, References; B, PDB-id; C, Resolution (Å); D, Space group; E, Rwork/Rfree (%); F, Number of water molecules in the structure; G, Ligand/mutant; NA, Not applied. #References: 1: Steiner et al., 2001; 2 and 4: Sekar et al., 2005; 3 and 12: Sekar et al., 2006a 5: Sekar et al., 2004; 6, 8 and 9: Chapter 3; 7: Sekar and Sundaralingam, 1999; 10: Rajakannan et al., 2002; 11: Sekar et al., 2006b; 13: Dijkstra et al., 1981b; 14: Yu et al., 2000; 15: Sekar et al., 1998a; 16, 24 and 25: Sekar et al., 1999; 17: Sekar et al., 1998b; 18: Sekar et al., 1997b; 19: Sekar et al., 2003; 20: Kumar et al., 1994; 21: Huang et al., 1996; 22 and 23: Sekar et al., 1997a. +ANN, 4-methoxybenzoic acid; TSA, 1-O-octyl-2-heptylphosphonyl-sn-glycero-3-phosphoethanolamine; MJ33, 1-decyl-3-trifluoroethyl-sn-glycero-2-phosphomethanol. § Number of water molecules given for PDB-id: 2BD1 is for both the molecules found in the asymmetric unit.


Table 4.2 List of hydrogen-bonding interactions between invariant water molecules and protein atoms calculated using the program HBPLUS (McDonald and Thornton, 1994).

A# B C D E F 131 IW1 His48 Nδ1, Asp49 Oδ1 147 132 IW2 Cys45 O His48 Nδ1, Asp49 Oδ1 136 133 IW3 Ala1 N, Pro68 O Tyr52 OH, Asp99 Oδ2 134 IW4 Cys98 O 136 IW5 Tyr28 O, Gly30 N 132 MPD126 O4 140 IW6 Ala93 O 145, 167 141 IW7 Asp40 O 185 143 IW8 Gly32 O 147, 276 144 IW9 Leu41 N, Pro110 O 157 145 IW10 Ser85 O Asn88 Nδ2 140, 207 148 IW11 Glu17 Oε2 358 149 IW12 Leu19 N 316 151 IW13 Arg100 NH1, Asn101 Nδ2 154 IW14 Glu81 O Thr83 Oγ1 197, 230 157 IW15 Asp40 N Asp40 Oδ2 144, 204 159 IW16 Thr36 O, Asn122 O 163, 225 162 IW17 Ser15 O Asp21 Oδ2 137 MPD129 O2

167 IW18 Asn97 Nδ2 140, 173, 248

168 IW19 Phe106 O, Val109 O 298, 328 173 IW20 Cys84 N 167, 281 183 IW21 Ser15 N Ser15 Oγ 208 IW22 Leu31 N Asn23 Oδ1 266, 302

232 IW23 Ser107 Oγ 177, 255, 343

236 IW24 Gln46 Nε2, Thr47 Oγ1 A, Water number; B, Numbering scheme of the invariant water molecules; C, Main-chain protein atoms; D, Side-chain protein atoms; E, Water molecules; F, Others, MPD, 2-methyl-2,4-pentanediol. # Numbering scheme is taken from the reference structure (PDB-id: 1G4I; Steiner et al., 2001).


Figure 4.2 All 24 invariant water molecules observed in all of 25 crystal structures of bovine pancreatic phospholipase A2 are shown in different colors (Cluster-1: wheat, Cluster-2: cyan and Cluster-3: magenta) grouped into three clusters based on their location in the three-dimensional structure. The protein model shown belongs to the reference structure (PDB-id: 1G4I; Steiner et al., 2001).

Figure 4.3 Root-mean-square deviation (r.m.s.d.) of Cα atoms from the starting structure of all of the 25 MD simulations. PDB codes are given in the legend box.


4.2.2 INVARIANT WATER MOLECULES IN CLUSTER-1 Nine invariant water molecules grouped into Cluster-1 (Figure 4.2) are mainly

in the vicinity of the active site. Of these, a total of six water molecules (IW1, IW2,

IW4, IW5, IW8 and IW22) are very close to the active site of the enzyme (Figure 4.4).

Moreover, water molecules IW1, IW2, IW4, IW5 and IW8 are buried and are highly

stable with very low B factor (Table 4.3). They are considered to be internal water

molecules. However, three water molecules (IW1, IW5 and IW8) are exposed with

solvent-accessible surface area of more than 5.0 Å2 in some of the mutant structures

(PDB-ids: 2ZP4, 1KVX, 1O3W, 1CEH, 1MKU and 1MKS) and two native structures

(PDB-ids: 1MKT and 1IRB) owing to disturbances in the active-site hydrogen-bonding

network.

Water molecule IW1, which is hydrogen bonded to His48 Nδ1 and Asp49 Oδ1,

has previously been shown to be involved in the tautomerization of the catalytically

important imidazole of the residue His48 (Sekar and Sundaralingam, 1999). Moreover,

analysis reveals that water molecule (IW1) may be involved in the stabilization of the

functionally important residue Asp49 by donating a proton and is also hydrogen bonded

to another water molecule, which stabilizes the surface-loop residue Tyr69 in the active

site. As indicated in the literature, Tyr69 has been suggested to be involved in the

Figure 4.4 Stereoview of the nine invariant water molecules of Cluster-1 shown as green spheres together with their hydrogen-bonding interactions. Other water molecules that interact with the invariant water molecules are shown as red spheres. The active-site calcium ion is shown as orange sphere.


binding of the pro-S non-bridging oxygen atom of sn-3 phosphate group of the

substrate molecule (Scott and Sigler, 1994a). Furthermore, observations from the

crystal structure of PLA2 bound with a transition-state analogue (PDB-id: 1MKV;

Sekar et al., 1998b) show that it may be involved in holding the substrate molecule

during the enzyme catalysis. In addition, the MD simulations also show the presence of

a water molecule with 100% residence frequency at the position of IW1 (Table 4.3).

Internal water molecule IW2, which is hydrogen bonded to the backbone

oxygen atom of Cys45, His48 Nδ1 and Asp49 Oδ1, is very important for the catalytic

activity of the enzyme and is known to act as a nucleophile during the enzyme

hydrolysis (Steiner et al., 2001; Sekar et al., 2005). In fact, irrespective of the

resolution, space group and biochemical properties of the enzyme, water molecule IW2

is present in all the structures of BPLA2. Recently, the crystal structure of the active-

site single mutant H48N of PLA2 suggested the involvement and a possible role of the

water molecule IW2 in the low enzyme activity of the mutant enzyme (PDB-id: 2ZP4;

Chapter 3). Interestingly, as expected, the catalytic water molecule (IW2) is present in

all of the MD simulations with a residence frequency of 100% (Table 4.3), with the

exception of a structure (PDB-id: 1FDK; Sekar et al., 1997) in which one of the ligand

atoms is hydrogen bonded to His48 Nδ1. Surprisingly, in the case of ligand-bound

structure (PDB-id: 1MKV; Sekar et al., 1998b), the catalytic water molecule (IW2) is

observed to be present during the MD simulations (Table 4.3). According to the

proposed enzyme mechanism for the catalytic activity (Scott et al., 1990b), these

observations correspond to the last step of the catalytic process, in which the substrate

molecule is cleaved and the product is displaced from the active site. During this

process, three water molecules (IW2, IW5 and IW8) occupy the void created by the

substrate molecule in the active site.


Table 4.3 Invariant water molecules corresponding to the reference structure (PDB-id: 1G4I) along with their normalized B factor in Å2 (second row), solvent-accessible surface area in Å2 (third row) and percentage residence frequency (fourth row) calculated from the structures generated using the MD simulations for

Cluster-1. Symbols used: M, mutant structure; D, electron density map (2Fo-Fc) present at 0.8σ; L, a ligand molecule is present at the spatial position of the water molecule; O, other atoms of the protein molecule; NSF,

structure factor not available; AVG, average.

PDB-id IW1 IW2 IW4 IW5 IW7 IW8 IW16 IW22 IW24

1G4I

131 -0.6 0.0 100

132 -0.8 0.0 100

134 -0.6 1.4

98.8

136 -0.7 0.0 100

141 -0.5 11.3 94.4

143 -0.2 0.0 100

159 0.0 0.2

99.6

208 0.6 1.2 100

236 0.1

15.4 100

1VL9

205 -0.6 0.0 100

430 -0.8 0.0 100

202 -0.7 0.0

94.2

206 -0.8 0.0 100

306 -0.3 8.8

38.4

204 -0.7 0.0 100

213 -0.5 15.7 72.9

251 0.4

18.1 100

225 -0.1 21.8 100

2BCH

240 -0.1 0.0 100

217 -0.5 0.0 100

206 -0.8 2.4

99.4

231 -0.1 0.0 100

204 -0.6 9.5

99.8

297 0.4 0.0 100

227 0.1 1.5

99.9

236 1.0 3.0 100

215 -0.5 15.1 100

2BAX

214 -0.4 0.0 100

205 -0.6 0.0 100

203 -0.8 2.2

99.5

257 -0.5 0.0 100

230 -0.6 11.2 98.8

262 -0.2 0.0 100

228 0.2 1.2

99.7

232 0.8 8.7 100

207 -0.7 15.8 100

1VKQ

7 -0.5 0.0 100

2 -0.6 0.0 100

3 -0.8 2.7

93.2

17 -0.6 0.0 100

13 -0.7 6.9

99.8

20 -0.3 0.1 100

28 0.0 3.4

94.5

40 0.8 8.9 100

8 -0.6 14.9 100

2ZP3

M - -

99.9

253 0.7 1.1 100

218 -0.7 2.3

97.9

M - -

100

210 -0.6 0.0

95.2

M - -

100

D - -

63.2

282 0.9

25.3 100

204 -0.6 6.4 100

1UNE

266 -0.5 0.0 100

268 -0.8 0.0 100

201 -1.0 1.8

95.7

263 -1.1 0.0 100

203 0.0

13.0 97.3

264 0.2 0.8 100

213 -0.1 13.7 99.9

225 2.3 7.1 100

216 0.9 8.8 100

2ZP4

203 -0.6 1.2 100

212 -0.9 0.0 100

228 -0.8 2.1

89.7

208 -0.8 0.0 100

214 -0.6 4.8

95.0

225 -0.2 9.2 100

268 -0.2 18.7 93.1

238 0.4

22.7 100

213 -0.5 7.7 100

2ZP5

M - -

99.9

229 0.4 0.0

99.9

254 -0.8 2.7

98.3

M - -

99.9

207 -0.9 4.3

97.0

M - -

100

238 0.3

12.5 99.9

268 1.6 0.0 100

255 -0.5 5.9 100

1GH4

206 -0.4 0.0 100

218 -0.5 0.9 100

209 -0.4 0.0

97.4

201 -1.4 0.2 100

321 1.7

20.0 96.7

202 -0.8 0.0 100

212 0.0

15.5 98.5

230 0.9

15.3 100

233 0.9

20.1 100

2B96

L - -

100

204 -0.1 0.0 100

212 -0.8 2.5

97.0

L - -

34.2

216 -0.1 10.8 96.6

203 -0.3 0.5

99.4

225 0.2 4.5

89.7

259 1.3 6.3

99.9

275 0.8 4.9 100

2BD1

204 -0.4 0.0 100

231 -0.6 0.3 100

210 -0.8 0.0

77.9

202 -1.2 0.0 100

300 0.9

20.2 94.4

208 -0.8 0.2 100

213 -0.8 13.5 98.5

356 0.3

13.0 100

D - -

100


1BP2

7 -0.5 2.2 100

6 -0.6 0.0 100

O - -

92.1

5 -0.6 2.3 100

16 -0.2 6.7

99.6

12 -0.3 1.4 100

13 -0.2 13.1 99.1

47 1.1

10.3 100

O - -

100

1C74

223 -0.3 0.1 100

205 -0.5 1.2 100

212 -0.7 2.5

88.5

202 -0.8 0.0 100

207 -0.2 12.9 97.5

203 0.5 2.1 100

219 0.3

15.4 99.5

228 0.7

20.5 100

293 0.3 4.8 100

1MKT

226 0.6 0.0 100

205 0.3 0.2 100

212 -1.2 2.8

99.5

202 -0.2 0.0 100

207 -1.0 22.0 98.6

203 0.1 5.3 100

222 0.9

13.9 100

234 1.3

17.1 100

220 0.3

10.5 100

1KVX

259 1.5 5.7 100

228 -0.3 2.5 100

201 -0.7 2.2

80.0

227 -1.0 8.9 100

202 -0.9 12.0 95.6

226 2.1 1.8 100

208 0.8

14.3 99.2

214 0.2

14.1 100

NSF - -

100

1MKV

L - -

0.0

L - -

100

208 -1.1 2.8

99.4

L - -

0.0

204 -0.2 9.5

99.6

L - -

0.0

D - -

96.6

226 0.4

24.2 100

214 0.6 6.4 100

1FDK

L - -

100

L - -

0.0

206 -0.9 2.8

99.4

L - -

0.0

202 -0.4 11.4 99.7

D - -

100

209 0.0 9.5 100

L - -

100

222' - -

100

1O3W

224 0.1 0.7 100

205 -0.6 0.0 100

212 -1.0 2.6

98.4

202 -0.4 0.0 100

207 -0.4 11.7 99.3

203 0.3 8.1 100

220 0.6

13.9 100

229 0.9

12.9 100

219 0.9 5.8 100

1CEH

237 0.3 0.6 100

216 -0.3 0.0 100

215 -0.7 2.5

0.73

204 -0.5 1.7 100

210 -0.2 10.0 96.9

238 1.2

11.7 100

214 -0.4 11.8 97.5

259 1.2

27.8 99.9

242 0.7

11.8 100

1IRB

221 0.5 5.5 100

213 -0.5 2.9 100

212 -0.6 2.9

91.8

204 1.1

16.6 100

210 -0.3 7.5

98.1

205 1.3

15.1 100

240 0.1

15.5 99.1

231 0.8

16.6 100

222 0.3

10.5 100

1MKU

211 -0.2 11.5 100

230 -0.2 2.2 100

201 -0.6 2.1

97.0

235 -0.3 7.4 100

204 -0.4 19.7 96.2

236 0.6

20.1 100

221 -0.1 17.0 97.0

249 1.3

13.7 100

228 0.5

12.4 100

1MKS

238 0.1 6.5 100

205 0.0 0.5 100

210 -1.2 2.5

89.7

202 0.4 8.2 100

207 -0.5 9.8

98.2

201 0.8

23.3 100

215 0.6

17.4 92.9

253 1.1

23.2 100

214 0.9 6.3 100

1KVY

250 1.0 0.5 100

251 0.9 0.6 100

208 -0.9 3.0

93.0

M - -

100

204 -0.3 21.9 92.0

M - -

100

252 0.7 8.3

91.7

NSF - -

100

215 0.7

11.5 100

1KVW

M - -

0.0

204 0.1 0.5 100

208 0.4 3.8

91.1

202 0.2 0.4 100

245 1.4

23.9 99.8

243 1.0 0.2 100

NSF - -

100

NSF - -

100

215 0.5

17.2 100

AVG -0.2 0.9 100

-0.5 0.3 100

-0.7 1.9

95.9

-0.4 1.2

98.5

-0.4 11.3 90.1

0.0 2.4 100

0.0 7.9

94.4

0.8 10.6 100

-0.4 13.6 100


Water molecule IW4 seems to be structurally crucial in positioning the

important residue Cys98. It is notable that the water molecule IW4 is present in all of

the structures, irrespective of the crystal form and chemical modifications of protein

molecule, with the exception of a structure (PDB-id: 1BP2; Dijkstra et al., 1981), in

which one MPD molecule replaces this water molecule and is hydrogen bonded to the

backbone oxygen atom of Cys98 with a distance of 2.98 Å. Interestingly, water

molecule IW4 is buried in a cavity that is relatively hydrophobic in nature.

Furthermore, it is interesting to note that the disulfide bond (Cys51-Cys98) plays a key

role in maintaining the two longest helices α3 and α5 of the enzyme (Figure 4.1). This

water molecule is also present during the MD simulations with a residence frequency of

96%, except in one structure (PDB-id: 1CEH; Kumar et al., 1994; residence frequency

of 73%). The low residence frequency of this water molecule in the single-mutant

structure (D99N; Kumar et al., 1994) may possibly be a consequence of the mutation at

position 99. Furthermore, in the MD averaged structure, it is observed that two residues

(Thr47 and Asn101) have moved towards the spatial position of IW4 making it almost

non-accessible to solvent molecules. Thus, water molecule IW4 can be treated as an

integral part of the structure. Water molecules IW5 and IW8 provide coordination to

the functionally important calcium ion. It has been shown that these two water

molecules are replaced by the phosphoryl oxygen atoms of the substrate (Scott et al.,

1990b). Interestingly, the residence frequencies computed from the MD simulations for

these two calcium-coordinated water molecules reveal their absence only in the case of

ligand-bound structures, in which two of the ligand atoms are coordinated to the

functionally important calcium ion (Table 4.3). Water molecule IW7, together with

another water molecule (185), is involved in stabilizing the surface residue Asp40

(Figure 4.4), which forms an ion pair with Arg43. Together with the disulfide bond

between Cys27 and Cys123, water molecule IW16 basically connects the active-site

loop and the C-terminal residues of the enzyme. Another water molecule (IW22),

which is hydrogen bonded to the backbone nitrogen atom of Leu31 and the side-chain

oxygen atom of Asn23, is likely to anchor the active-site calcium-binding loop.

Furthermore, water molecule IW24 seems to be involved in stabilizing the active-site

residue His48 and is hydrogen bonded to Gln46 and Thr47. In addition, all these water

molecules (IW7, IW16, IW22 and IW24) are present during the MD simulations with


more than 90% residence frequency and have very low normalized B factor with the

exception of IW22 (Table 4.3).

4.2.3 INVARIANT WATER MOLECULES IN CLUSTER-2 In total, eight invariant water molecules are grouped into Cluster-2. Water

molecule IW9 is totally buried, has a low normalized B factor (Table 4.4) and is

hydrogen bonded to the backbone nitrogen atom of Leu41, the backbone oxygen atom

of Pro110 and another water molecule IW15 (Figure 4.5). Leu41 has been suggested to

contribute to the back wall of the hydrophobic channel of the enzyme (Scott and Sigler,

1994a). Thus, water molecule IW9 may play a role in holding the residue Leu41 and

two helices α3 and α5 in place by maintaining the overall tertiary structure. The MD

analysis also supports that water molecule IW9 has 100% residence frequency during

the simulations. Interestingly, water molecule IW11 is highly exposed but has a very

low normalized B factor (Table 4.4). It stabilizes the N-terminal capping of Glu17 in

the helix α2 (Figure 4.5). The residence frequency computed from the MD simulations

for water molecule IW11 is 100% (Table 4.4). A semi-accessible (solvent accessibility

of 4.3 Å2) water molecule IW12, with a very low normalized B factor (Table 4.4), is

hydrogen bonded to the backbone nitrogen atom of Leu19 and a water molecule (316).

In most of the structures, water molecule (316) is replaced by one of the side-chain

atoms of Asn6, which connects two helices (α1 and α2) through the water molecule

IW12. However, the residence frequency for this water molecule is only 74%, which

may be a consequence of the hydrophobic nature of residue Leu19 that contributes to

the interfacial adsorption surface (Scott and Sigler, 1994a). Water molecule IW15 is

hydrogen bonded to the backbone nitrogen atom and one of the side-chain carboxylate

oxygen atoms of Asp40. Together with another water molecule (IW7; discussed

earlier), water molecule IW15 may play a role in anchoring Asp40 (Figure 4.5). The

residence frequency computed from the MD simulations for this water molecule is

100% (Table 4.4).

It is interesting to note that water molecule IW17 is hydrogen bonded to the

backbone oxygen atom of Ser15, Asp21 Oδ2, an MPD molecule and a water molecule

(137). The triangle formed by the backbone oxygen atom of Ser15, IW17 and water

molecule 137 is an approximate equilateral triangle (Figure 4.5). In fact, another water


molecule (IW21) is also observed to be hydrogen bonded to the backbone nitrogen and

Oγ atoms of residue Ser15. Thus, it is very much possible that these two invariant water

molecules stabilize residue Ser15, which is on the surface of the molecule and is

located in a loop connecting two helices (α1 and α2). Interestingly, the PLA2 enzymes

from almost all groups consist of five or six short loops. However, the loop consisting

of the residues 14-16 is relatively rigid with a proline residue (Pro14) at the beginning

and a characteristic type I β-turn. Furthermore, the short loop also provides structural

support to the substrate-binding channel and it has been found that the amino-acid

sequences and conformation of this loop are almost identical in all PLA2 enzyme

variants (Jabeen et al., 2005b). In addition, water molecule IW17 is present during the

MD simulations with 100% residence frequency; the corresponding value for IW21 is

83%. However, the reason for the low residence frequency of IW21 is not clear. The

buried water molecule IW19 is involved in a hydrogen-bonding network that connects

the backbone oxygen atoms of Phe106 and Val109. In most of the crystal structures, a

Tris (tris(hydroxymethyl)aminomethane) molecule is observed near IW19 preventing

exposure of this water molecule. However, in three structures (PDB-ids: 1UNE, 1GH4

and 1MKU), water molecule IW19 is more exposed because of the absence of the Tris

molecule. Water molecule IW23 seems to stabilize the C-terminal capping surface

residue Ser107 of helix α5 (Figure 4.1). Both these water molecules (IW19 and IW23)

have a residence frequency of greater than 95% during the MD simulations.


Table 4.4 Invariant water molecules corresponding to the reference structure (PDB-id: 1G4I; Steiner et al., 2001) along with their normalized B factor in Å2 (second row), solvent-accessible

surface area in Å2 (third row) and percentage residence frequency (fourth row) calculated from the structures generated using the MD simulations for Cluster-2. Symmetry-related water molecules

are indicated with a prime; D, electron-density map (2Fo-Fc) present at 0.8σ; O, other atoms of the protein molecule; NSF, structure factor not available; AVG, average.

PDB-id IW9 IW11 IW12 IW15 IW17 IW19 IW21 IW23

1G4I

144 -0.4 0.0 100

148 -0.6 34.1 100

149 0.1 0.0

79.0

157 0.1 2.7 100

162 -0.3 18.9 100

168 0.6 0.0 100

183 0.9

14.0 79.4

232 1.2 1.0 100

1VL9

214 -0.4 0.0 100

302 0.7

43.0 100

244 0.8

10.3 64.7

224 0.2 4.0 100

252 0.7

12.5 100

226 0.0 0.0 100

405 -0.8 21.1 76.9

339 2.1

10.6 99.9

2BCH

211 -0.4 0.3 100

203 -0.7 39.3 100

202 -0.7 6.1

53.8

420 0.0 6.1 100

219 -0.3 17.7 100

212 -0.3 0.0 100

238 0.3

25.3 76.3

220 0.1 0.0

99.9

2BAX

210 -0.5 0.0 100

209 -0.7 47.0 100

202 -0.7 0.1

79.4

225 -0.3 0.5 100

218 -0.3 36.7 100

271 -0.3 0.2

57.5

242 0.3

13.0 76.9

222 0.0 0.0

99.2

1VKQ

10 -0.6 0.0 100

18 -0.7 38.0 100

4 -0.7 6.5

82.1

23 -0.1 5.5 100

37 -0.1 15.7 100

27 -0.3 0.1

98.0

53 0.3 3.4

89.8

22 -0.1 9.0

99.4

2ZP3

234 -0.4 0.0 100

221' - -

100

215 -0.8 7.5

76.1

227 0.0

19.5 100

237 0.0

31.5 100

229 -0.4 0.0

99.9

223 0.0

25.5 82.6

216 -0.1 9.9

97.9

1UNE

208 -0.4 0.0 100

215 -0.7 47.2 100

209 0.8

10.3 89.8

207 1.6 4.1 100

223 0.3

27.7 100

228 0.1 9.3 100

226 1.9

24.3 100

327 1.8

17.5 99.8

2ZP4

226 -0.5 0.0 100

233' - -

100

218 -0.6 5.3

57.6

273 0.2

18.5 100

257 0.0

27.8 100

236 -0.2 0.0 100

256 0.2

26.4 96.8

216 -0.4 7.2

98.8

2ZP5

217 -0.5 0.0 100

242 -0.8 43.5 100

226 -0.5 0.0

95.6

235 0.4

16.5 100

222 -0.6 22.3 100

220 -0.3 3.5 100

210 -0.2 14.5 87.3

204 -0.4 8.8

98.8

1GH4

229 0.2 0.0 100

282 2.7

22.6 100

264 1.3

32.3 99.7

222 0.2

19.2 100

320 1.4

26.4 100

234 0.5

10.0 100

268 0.8

23.9 74.7

232 2.2

22.4 99.9

2B96

228 -0.3 0.0 100

215 -0.5 47.7 100

278 0.3 2.1

83.8

227 -0.3 29.4 100

D - -

100

O - -

100

250 0.0

20.8 81.0

314 1.4 9.7

99.9

2BD1

242 -0.4 0.0 100

O - -

100

414 1.3

13.7 97.1

224 0.5 6.6 100

264 0.3

22.6 100

240 0.2 8.5 100

274 0.5

27.2 93.8

338 32.0 10.1 99.9


1BP2

14 -0.2 0.0 100

O - -

100

31 0.5 5.2

77.1

55 1.3

15.4 100

NSF - -

100

38 0.7 2.8 100

NSF - -

80.6

O - -

10.7

1C74

210 -0.7 0.0 100

214 -0.7 46.0 100

213 -0.6 0.0

77.6

249 1.0

16.8 100

217 -0.1 20.1 100

287 0.3 0.2 100

236 1.1

22.7 73.7

240 0.6 9.5

99.8

1MKT

210 -0.2 0.0 100

215 -0.1 34.4 100

213 -0.3 0.0

79.8

237 0.7

10.0 100

219 0.3

24.0 100

245 0.5 1.1 100

252 1.1

26.5 100

269 1.1 9.6

99.9

1KVX

206 -1.0 0.0 100

250 -0.1 42.0 100

207 -0.7 0.0

95.8

233 1.1 7.9 100

213 -0.2 25.5 100

NSF - -

100

241 0.8

26.6 81.8

O - -

99.5

1MKV

207 -0.2 0.0

99.9

288 -0.3 41.5 100

209 -0.4 0.7

95.4

229 0.4 4.1 100

213 0.3

29.4 100

211 0.0 0.0 100

232 1.3

33.3 80.8

246 0.9 9.3

99.7

1FDK

204 -0.3 0.0 100

207 0.0

47.5 100

263 0.4 0.1 100

219 0.5

13.3 100

257 1.4

19.2 100

249 0.6 1.4 100

268 0.3

40.0 71.7

O - -

99.8

1O3W

210 -0.3 0.0 100

214 -0.7 42.6 100

213 -0.6 0.0

74.7

230 0.9 8.1 100

218 0.1

19.0 100

235 0.4 0.2 100

238 0.7

32.0 80.6

243 0.8

15.8 99.7

1CEH

203 -0.5 0.0 100

209 -0.4 45.9 100

243 0.3 8.4

76.7

218 -0.5 15.3 100

258 0.8

29.3 100

233 -0.2 0.0

99.9

239 0.6

15.2 62.5

241 2.1 1.1

95.6

1IRB

203 -0.2 0.0 100

209 0.0

42.4 100

D - -

92.7

214 0.6

10.6 100

243 0.0

22.0 100

218 0.1 0.0 100

D - -

92.0

242 1.9

15.7 99.2

1MKU

210 -0.5 0.0 100

226 -0.8 44.2 100

212 -0.5 0.1

39.7

208 0.4

14.9 100

246 0.3

26.3 100

257 1.2 8.6 100

252 1.1

17.1 74.1

O - -

100

1MKS

209 -0.2 0.0 100

233 0.0

49.6 100

211 -0.6 0.3

51.2

255 1.1

11.4 100

236 0.6

22.3 100

213 0.3 2.7 100

D - -

77.8

256 0.8

14.1 99.7

1KVY

207 -0.4 0.0 100

210 -0.5 45.0 100

209 -0.6 0.4

30.2

225 0.9 9.8 100

214 -0.4 26.7 100

212 -0.4 0.0 100

NSF - -

78.6

243 1.2

25.4 99.4

1KVW

207 -0.2 0.0 100

210 -0.1 55.9 100

209 -0.1 0.1

96.0

225 1.4

13.0 100

214 0.6

20.6 100

228 0.5 0.2 100

NSF - -

85.4

NSF - -

99.7

AVG -0.4 0.0 100

-0.3 38.0 100

-0.1 4.3

74.4

0.3 8.1 100

0.2 21.7 100

0.1 1.2

94.7

0.4 18.4 80.3

0.9 7.4

97.6


4.2.4 INVARIANT WATER MOLECULES IN CLUSTER-3 In Cluster-3, there are seven water molecules identified to be invariant (Table

4.5). Water molecule IW3 has been suggested to maintain the geometry of the

catalytically important residue Asp99 (Kumar et al., 1994; Sekar et al., 1997a; Sekar

and Sundaralingam, 1999). It is totally buried and stable with low B factor. The water

molecule IW3 forms a hydrogen-bonding network with the surrounding residues (Ala1,

Tyr52, Pro68 and Asp99) and has been proposed to serve as a link between the active

site and the interfacial recognition site (Verheij et al., 1980; Yuan and Gelb, 1988;

Sekar and Sundaralingam, 1999). It is also noteworthy that the N-terminal residue Ala1

is believed to be involved in the interfacial catalysis. Ala1 performs the activation of

the phospholipid while the other end performs the hydrolysis of monomeric

phospholipids (Sekar et al., 1997a). This suggests that the water molecule IW3

stabilizes the N-terminal residue Ala1 as well as Asp99. Analysis of the structures of

Figure 4.5 Stereoview of the eight invariant water molecules shown as green spheres of Cluster-2 together with their hydrogen-bonding interactions with the protein molecule. Other water molecules that interact with the invariant water molecules are shown as red spheres.


single mutants at Asp99 (Kumar et al., 1994; Sekar et al., 1997a; Sekar et al., 1999)

showed the absence of this water molecule at this position and it was observed that

Ala1 occupies the void created by the missing structural water molecule (IW3). MD

analysis of all of the crystal structures investigated reveals the presence of this water

molecule at a similar position with a residence frequency of more than 84%. These

observations suggest a requirement for the structural water at this position to anchor the

N-terminal residue, Ala1.

Furthermore, four water molecules (IW6, IW10, IW18 and IW20) form an

extended water bridge near the N-terminal region (Figure 4.6) and stabilize a stretch of

residues (Cys84, Ser85, Asn88, Ala93 and Asn97). It is interesting to note that these

four water molecules fill a small cavity that is observed on the surface of the enzyme.

However, the importance of this cavity is not known. Moreover, IW20 seems to be

involved in anchoring Cys84 (stabilizing the β-wing of the enzyme), which forms a

disulfide bond between the β-sheet and one of the longest helices α5. In addition, IW10

(a buried water molecule) stabilizes the loop residue Ser85 and the N-terminal capping

residue Asn88 of the helix α5. Furthermore, the second calcium-coordinating residues

Asn71 and Glu92 have been suggested to be responsible for the binding of the enzyme

to the membrane (Sekar et al., 2006a). It has also been observed that water molecules

IW6 and IW20 are mediated through another water molecule IW18, which is further

hydrogen bonded to Asn97. Thus, it may be suggested that the water bridge formed by

these four invariant water molecules is responsible for keeping the residues in place at

the interfacial site (Figure 4.6). Moreover, all four of these invariant water molecules

are present during the MD simulations with more than 97% residence frequency, with

the exception of IW6 (83%). Water molecule IW13 is buried in a positively charged

environment very close to the small cavity filled by an extended water bridge (IW6,

IW10, IW18 and IW20). Water molecule IW14, which is situated between two short β-

strands (known as the neurotoxic region and located in the highly negatively charged

cavity of the enzyme), is semi-accessible (solvent-accessible area = 3.0 Å2) with a low

normalized B factor (Table 4.5). It may play a role in stabilizing the strand β1 of the

enzyme. The residence frequencies of these two water molecules (IW13 and IW14) are

greater than 90% during the MD simulations.


Table 4.5 Invariant water molecules corresponding to the reference structure (PDB-id: 1G4I; Steiner et al., 2001) along with their normalized B factor in Å2 (second row), solvent-accessible

surface area in Å2 (third row) and percentage residence frequency (fourth row) calculated from the structures generated using the MD simulations for Cluster-3. Symbols used: SC, symmetry-related

contact; NSF, structure factors are not available in the PDB; D, electron density map (2Fo-Fc) present at 0.8σ; O, other atoms of the protein molecule; AVG, average.

PDB-id IW3 IW6 IW10 IW13 IW14 IW18 IW20

1G4I

133 -0.3 0.0 100

140 -0.3 4.3

88.7

145 -0.5 0.0

92.4

151 -0.3 0.3 100

154 0.5 0.0

99.6

167 0.3 0.0 100

173 0.3 2.0

97.0

1VL9

203 -0.6 0.0 100

237 0.7

22.0 89.4

208 -0.4 1.7

98.0

228 0.1

12.7 100

286 1.2 3.5

79.8

279 0.7

21.7 100

235 0.1 0.0

98.6

2BCH

208 -0.6 0.0

54.1

226 -0.5 10.0 87.6

209 -0.7 0.0

98.5

419 -0.3 0.0 100

207 -0.7 1.1

99.3

424 0.3 7.4 100

235 0.2 3.4

96.3

2BAX

211 -0.5 0.0

52.3

215 -0.5 0.0

70.0

208 -0.6 0.0

99.5

313 -0.4 1.4 100

206 -0.6 0.5

93.9

231 -0.1 12.7 100

221 -0.5 6.3

99.2

1VKQ

19 -0.3 0.0 100

16 -0.5 14.0 59.6

5 -0.6 2.6

97.9

35 0.3 0.3 100

1 -0.7 0.4

66.4

54 0.2

15.0 100

12 -0.2 11.7 98.7

2ZP3

224 -0.4 0.0 100

231 -0.4 9.6

93.7

203 -0.9 3.0

99.1

212 0.0 0.0 100

208 -0.5 5.3

99.9

273 0.3 0.8 100

205 -0.9 4.9

92.4

1UNE

276 -0.4 0.0 100

224 0.4 7.7

90.0

218 0.0 0.4

99.8

235 1.6 0.2 100

211 0.9 0.7

97.5

289 1.7 9.2 100

288 1.4

12.2 91.7

2ZP4

227 -0.5 0.0 100

253 -0.4 14.1 74.1

201 -0.7 3.0

93.3

219 -0.1 7.9 100

207 -0.5 4.7

60.6

210 0.0

16.9 99.9

205 -0.6 21.1 98.5

2ZP5

219 -0.6 0.0 100

215 -0.4 10.8 59.3

225 -0.7 3.5

96.8

218 -0.1 3.6 100

213 -0.6 4.4

83.2

246 -0.3 7.1

99.9

216 -0.7 6.4

98.4

1GH4

204 -1.3 0.0 100

311 1.3 0.3

93.8

208 0.2 0.5

98.2

245 0.6

14.6 100

SC - -

98.9

265 1.7

23.0 100

316 1.3

19.6 98.3

2B96

202 -0.5 0.0 100

258 0.1

24.7 92.2

236 -0.1 0.0

99.5

255 0.3 5.5 100

234 -0.4 6.6

82.5

322 2.1 3.7 100

235 -0.9 18.7 95.5

2BD1

214 -0.8 0.0 100

296 1.4

25.2 89.4

244 0.2 0.4

99.3

259 0.5 0.0 100

SC - -

99.4

1.3 1.3

12.9 100

0.2 0.2

15.8 96.4


1BP2

11 -0.3 0.0 100

23 0.2 4.9

84.8

18 -0.1 0.0

97.7

36 0.7 2.1 100

32 0.5 2.7

77.6

NSF - -

100

50 1.2

25.3 96.8

1C74

211 -0.5 0.0 100

206 0.0

13.2 75.3

204 -0.4 0.0

96.5

254 0.6 7.4 100

209 0.5

13.2 99.8

215 0.7 2.3 100

220 0.2 2.2

92.1

1MKT

211 -0.3 0.0 100

206 0.2 2.2

94.3

204 0.1 0.1

99.4

D - -

100

209 -0.5 7.4

90.0

216 1.3 2.5 100

223 1.2 4.0

94.1

1KVX

M - -

100

251 0.9 3.2

86.2

210 -0.2 4.1

99.2

235 -0.3 3.2 100

SC - -

98.3

217 0.1

12.3 100

222 -0.6 13.2 98.3

1MKV

286 -0.1 0.0 0.0

203 0.0

14.0 90.2

202 -0.7 0.5

98.2

258 0.5 4.0 100

206 -0.7 6.4

97.2

D - -

100

216 -0.7 18.0 98.4

1FDK

205 -0.1 0.0 100

248 0.2

15.9 93.1

201 -0.3 0.8

99.6

D - -

100

203 -0.4 6.7

98.4

D - -

100

210 -0.7 25.6 98.9

1O3W

211 -0.6 0.3

59.7

206 -0.1 1.0

48.4

204 -0.7 4.4

98.3

257 0.5 8.4 100

209 -0.3 4.7

76.9

215 0.3

13.6 99.8

221 -0.7 7.2

98.8

1CEH

O - -

0.0

220 -0.5 31.3 70.2

206 -0.7 0.2

96.7

226 -0.2 4.5 100

205 -0.7 0.0

64.0

222 0.0

19.3 99.9

234 -0.4 9.6

97.4

1IRB

201 0.3 0.6 100

215 1.7

23.5 86.1

206 -0.8 2.5

97.8

235 -0.1 7.5 100

225 -0.1 11.9 79.3

262 0.9 8.4 100

219 -0.1 5.5

96.0

1MKU

O - -

99.8

248 0.3 4.3

86.2

231 -0.5 0.3 100

277 0.1 6.1 100

217 0.3 1.7

96.5

247 1.3

13.5 100

220 0.2

10.8 96.4

1MKS

O - -

0.0

206 -0.4 18.6 86.1

203 -0.6 3.4

95.2

245 -0.3 8.5 100

208 -0.7 6.4

87.8

234 0.6

16.4 100

216 0.4

20.3 99.4

1KVY

261 0.0 0.0 100

203 -0.1 3.5

87.0

202 -0.9 0.0

98.1

239 0.6 2.5 100

206 -0.2 6.5

72.8

NSF - -

100

216 -0.1 14.9 97.9

1KVW

255 0.8 0.0 100

244 0.6

16.9 89.1

203 -0.2 4.4

99.8

NSF - -

100

206 -0.1 7.2

82.6

211 0.4

13.9 100

218 1.0

14.7 96.6

AVG -0.4 0.0

84.7

0.1 10.2 83.2

-0.5 1.0

97.6

0.1 3.9 100

-0.1 3.0

89.7

0.5 9.8 100

0.0 8.2

97.3


4.3 CONCLUSION There are 24 invariant water molecules that are conserved in all of the structures

of PLA2. Of these, nine water molecules (IW1, IW2, IW3, IW4, IW5, IW8, IW9, IW10

and IW19) are located in the core of the enzyme and are likely to be involved in the

folding of the enzyme. Water molecules IW1 and IW2 are also involved in the catalytic

activity of the enzyme. In contrast, water molecules IW5 and IW8 are structurally

essential and provide coordination to the functionally important active site calcium ion.

These invariant water molecules are also important to maintaining the correct active-

site geometry. In addition, a few invariant water molecules are involved in mediating

ion pairs that play an important role in stabilizing the tertiary structure. A set of water

molecules forms a water bridge that stabilizes the functionally important residues.

Approximately half of the invariant water molecules play a role in stabilizing the

surface residues of the enzyme. Thus, it can be concluded that in addition to the

Figure 4.6 Stereoview of the seven invariant water molecules shown as green spheres of Cluster-3 together with their hydrogen-bonding interactions. Other interacting water molecules are shown as red spheres.


structurally and functionally important water molecules, the present study helps to

rationalize the water molecules that are significant to the folding and stability of the

enzyme PLA2.

4.4 MATERIALS AND METHODS 4.4.1 DATA SET

Seven out of the total 32 structures of BPLA2 were excluded from the study

owing to either the absence of water molecules in the structure (PDB-ids: 1BPQ,

1BVM, 2BP2 and 2BPP) or low resolution (>2.0 Å; PDB-ids: 1O2E and 3BP2) or

because they were structures of pro-PLA2 (PDB-id: 4BP2). The three-dimensional

atomic coordinates of the remaining 25 crystal structures (21 high-resolution and four

atomic-resolution structures) of the enzyme BPLA2 were downloaded from the locally

maintained anonymous PDB-FTP server at the Bioinformatics Centre, Indian Institute

of Science, Bangalore, India. The resolution of the investigated structures varies from

0.97 to 1.95 Å. In this resolution range, the positions of the solvent molecules are

determined with high accuracy, particularly those in the first hydration shell. The

crystal structure (PDB-id: 1G4I, Steiner et al., 2001) was taken as the reference or fixed

molecule because it contained the highest number of water molecules. All the

remaining structures were treated as mobile molecules and were superimposed on the

reference structure (PDB-id: 1G4I) to find invariant water molecules using the 3dSS

web server (Sumathi et al., 2006). The distance cutoff between pairs of superposed

water molecules was taken to be 1.7 Å and a water molecule that had at least one

common hydrogen bond to the protein atoms was considered to be invariant. While the

hydrogen-bonding distance (maximum 3.4 Å) and angle (greater than 90°) criteria were

generally followed, in some cases water molecules were considered to be equivalent if

similar hydrogen bonds were observed even if the pairwise distance cutoff (1.7 Å) was

not satisfied owing to the variations in the side-chain conformation(s). Furthermore, in

order to ascertain the water molecules, an investigation of the electron-density maps for

the structures used in the present study was also carried out. The corresponding

structure factors were downloaded from the RCSB-PDB (Berman et al., 2000) and the

CCP4 suite (Collaborative Computational Project, Number 4, 1994) was used to

generate mtz files. Subsequently, the modelling program COOT (Emsley and Cowtan,


2004) was used to visualize the difference electron-density maps. A water molecule

was considered to be invariant if it was observed in all the structures, except when the

sites were occupied by other molecules such as MPD and Tris (which were ingredients

of the crystallization mixture) in some structures. In the case of alternate

conformations, the conformation with the higher occupancy was considered for the

analysis. However, in the case of equal occupancies of the alternate conformations, the

first conformation was taken for the study. A similar approach was followed for the

water molecules also. Hydrogen-bonding interactions were calculated using the

program HBPLUS (McDonald and Thornton, 1994). The solvent-accessible surface

area of invariant water molecules was computed using the program NACCESS

(Hubbard and Thornton, 1993) with a probe radius of 1.4 Å. Water molecules with a

solvent-accessible surface area less than or equal to 2.5 Å2 were considered to be

internal or buried water molecules. The normalized B factor (Bi') for all invariant water

molecules was calculated using the formula Bi' = (Bi - )/σ(B), where Bi is the B

factor of each atom, is the mean B factor of the protein molecule and σ(B) is the

standard deviation of the B factor (Smith et al., 2003).

4.4.2 MOLECULAR DYNAMICS SIMULATION Molecular-dynamics (MD) simulations were performed using the GROMACS

v.3.3 package (van der Spoel et al., 2005) running on parallel processors with the

OPLS-AA force field (Jorgensen et al., 1996; Kaminski et al., 2001). During the MD

simulations, crystallographic water molecules were removed from the protein models.

However, the ligand, the molecules from the crystallization conditions and the calcium

ions were retained during the MD simulations. A cubic box of dimensions 6.4×6.4×6.4

nm was generated using the module editconf of the GROMACS suite. The necessary

parameters and the topology file for MPD and Tris molecules were generated using the

PRODRG web server (Schuttelkopf and van Alten, 2004). Subsequently, protein

models were solvated with the SPC (simple point charge) water model using the

program genbox available in the GROMACS suite. Energy minimization was

performed using the conjugate-gradient method for 200 ps with the maximum force-

field cutoff being 1 kJ mol-1 nm-1. Sodium and chloride ions, wherever needed, were

used to neutralize the overall charge of the system. Simulations utilized NPT ensembles


with isotropic pressure coupling (τp = 0.5 ps) to 100 kPa and temperature coupling (τt =

0.1 ps) to 300 K. Parrinello-Rahman and Nose-Hoover coupling protocols were used

for pressure and temperature, respectively. Long-range electrostatics were computed

using the Particle Mesh Ewald (PME; Darden et al., 1993) method and Lennard-Jones

energies were cut off at 1.0 nm. Bond lengths were constrained with the LINCS

algorithm (Hess et al., 1997). The dielectric constant was taken as unity required for an

explicit solvent MD simulations. Simulations were performed for a time period of 3 ns

for all the structures considered in the present study. Analyses were performed using

the tools available in the GROMACS suite. The average structures used for comparison

and analyses were calculated using the ensembles generated between 1 and 3 ns. The

atomic coordinates were computed every 2 ps and calculation of the residence

frequency of the invariant water molecules was carried out using these structures. The

interactions between the protein atoms and the solvent molecules were calculated with

a hydrogen-bond distance of 3.4 Å. The residence of a particular invariant water

molecule near the interacting residue(s) was assumed if at least one of the polar atoms

made contact with any solvent molecule within a distance of 3.4 Å.

CHAPTER 5 Crystal Structures of Apo and GTP-bound Molybdenum

Cofactor Biosynthesis Protein MoaC from Thermus

thermophilus HB8

CHAPTER 5: STRUCTURES AND DYNAMICS OF MoaC 134

5.1 INTRODUCTION As mentioned in the introductory chapter, several enzymes (collectively known

as molybdoenzymes) require the metal molybdenum for their functions through a

cofactor called as molybdenum cofactor (Moco). These enzymes play essential roles in

the global cycles of carbon, nitrogen and sulfur (Kisker et al., 1997; Mendel and

Bittner, 2006). Moco, which consists of a mononuclear molybdenum, is synthesized in

an evolutionary conserved pathway in archaea, eubacteria and eukaryotes including

humans (Chan et al., 1995). In Escherichia coli, at least five loci (moa, mob, mod, moe

and mog) are known to be involved in synthesis of Moco (Rajagopalan and Johnson,

1992). Of these, two loci moa and moe are required for the initial steps of Moco

biosynthesis (Nohno et al., 1988; Pitterle and Rajagoplan, 1989, 1993; Rivers et al.,

1993). Two gene products (MoaA and MoaC) of moa loci start the process of Moco

biosynthesis by converting GTP to cPMP or precursor Z (Wuebbens and Rajagopalan,

1993; Hanzelmann et al., 2002; Hanzelmann et al., 2004) in a similar process as

observed in the biosynthesis of folate, riboflavin and biopterin. However, in contrast to

these processes, Moco synthesis is involved in the rearrangement of the guanosine C8

atom as the first carbon of the precursor Z side chain.

Several ligand-free crystal structures of MoaC from E. coli (EcMoaC; PDB-id:

1EKR; Wuebbens et al., 2000), Pyrococcus horikoshii (PhMoaC; PDB-id: 2EKN; N.

K. Lokanth, K. J. Pampa, T. Kamiya and N. Kunishima, unpublished work), Sulfolobus

tokodaii (StMoaC; PDB-id: 2OHD; Yoshida et al., 2008) and Geobacillus kaustophilus

(GkMoaC; PDB-id: 2EEY; N. K. Lokanath, K. J. Pampa, T. Kamiya and Kunishima,

unpublished work) are available. However, no structural study of a ligand-bound form

is available in the literature. Here, we report three crystal structures (two apo forms and

one GTP-bound form) of MoaC from a highly thermophilic eubacterium (Thermus

thermophilius HB8; TtMoaC). The ligand-bound form of the crystal structure provides

the first structural evidence of the binding of a GTP molecule to MoaC. In addition,

ITC experiments have been carried out to support the findings obtained from the

crystallographic results. Furthermore, MD simulations have been carried out on both

known and modelled protein-ligand complexes to corroborate the above results. Thus,

the present study should enhance the existing knowledge of the Moco-biosynthesis

pathway, particularly the first step, which is not clearly understood.


5.2 RESULTS AND DISCUSSION The results obtained from crystal structures of TtMoaC, ITC experiments and

MD simulations have been discussed separately.

5.2.1 CRYSTALLOGRAPHIC RESULTS 5.2.1.1 Overall structure

All three forms of the crystal structure of TtMoaC were solved by molecular-

replacement method using the atomic coordinates of EcMoaC (PDB-id: 1EKR;

Wuebbens et al., 2000) as the search model. The asymmetric units of two apo forms

(P21, Form I and R32, Form II) and a GTP-bound form (C2221, Form III) contain 12,

one, and nine subunits, respectively. The refinement statistics of all three crystal

structures are given in Table 5.1. The final refined model in all three forms lacked the

first ten residues at the N-terminus and three residues at the C-terminus. Each monomer

of MoaC has a α+β structure and is composed of a four-stranded anti-parallel β-sheet

with three helices α1, α2 and α3, located on the same side of the β-sheet (Figure 5.1). In

addition, there is a short 310-helix (residues 90-92). MoaC belongs to ferredoxin-like

(βαββαβ) fold, with an insertion of a helix (βααββαβ). It forms a hexamer made up of

three dimers (Wuebbens et al., 2000). Almost ~42% surface area of each monomer is

buried upon hexamerization. A total of 22% and 18% of the accessible surface area of

each monomer is buried upon dimerization and trimerization, respectively.

Furthermore, each monomer of the hexamer contacts another three subunits (two from

the trimeric subunits and one from the dimeric subunits), as in a similar feature

observed in the case of EcMoaC. Each dimer of TtMoaC is stabilized by 11 inter-

subunit hydrogen bonds, compared with the eight hydrogen bonds in EcMoaC

(Wuebbens et al., 2000). Each trimer and hexamer of TtMoaC has 36 and 115

intersubunit hydrogen bonds, respectively. The solvation-free (SF) energy gain upon

the assembly formation is predicted to be –124.9 kcal mol-1 using the PISA web server

(Krissinel and Henrick, 2007). The SF energies for the phosphate-bound structures in

the apo and the GTP-bound structures in the complex form show an increase of ~50

and ~30 kcal mol-1, respectively, compared with those of the ligand-free hexamer.


Table 5.1 X-ray data and refinement statistics for free and GTP-bound forms of TtMoaC. Values in the parenthesis are for the highest-resolution shell.

Form I (P21) Form II (R32) Form III (C2221) Data collection and processing Wavelength (Å) 1.0 0.97243 1.5418 Temperature (K) 100 100 100

Unit-cell parameters (Å, °) a=64.81, b=109.84, c=115.19, β=104.9

a=b=106.58, c=59.25

a=69.93, b=111.57, c=311.4

Resolution (Å) 50.0-1.9 (1.97-1.9) 50.0-1.75 (1.81-1.75) 30.0-2.5 (2.59-2.5) Observed reflections 667221 288658 309014 Unique reflections 121501 (12072) 13115 (1288) 41061 (3786) Completeness (%) 99.9 (99.7) 99.9 (100) 96.1 (89.6) Matthews coeff. (Å3 Da-1) 1.94 1.9 1.99 Solvent content (%) 36.7 35.5 38.1 Multiplicity 5.5 (5.1) 22.0 (21.3) 7.5 (7.2) I/σ(I) 23.9 (5.3) 57.4 (14.6) 22.8 (3.1) Rmerge

# (%) 4.5 (17.2) 8.0 (17.7) 8.7 (53.0) Refinement Statistics Rwork (%) 18.8 19.9 20.2 Rfree (%) 21.9 21.8 27.0 Protein Model Subunits/ASU 12 1 9 Protein atoms 13230 1078 9883 Water molecules 1181 132 427 Phosphate ions 12 1 - GTP molecules - - 9 Others 22 2 39 Deviations from ideal geometry Bonds lengths (Å) 0.005 0.004 0.007 Bond angles (°) 1.3 1.3 1.4 Dihedral angles (°) 23.4 23.2 23.1 Improper angles (°) 0.75 0.75 0.92 Average temperature factors (Å2) Protein atoms 23.2 24.4 53.3 Water molecules 36.8 38.6 52.1 Phosphate ions 34.8 19.4 - GTP molecules - - 83.0 Others 45.6 54.2 84.9 Ramachandran plot (%) Most favored 93.1 93.6 87.9 Additionally allowed 6.9 6.4 12.1 # Rmerge = ΣhΣi |I(h)i - <I(h)>| / ΣhΣiI(h)i, where I(h) is the intensity of reflection h, Σh is the sum over all reflections and Σi is the sum over i measurements of reflection h.


5.2.1.2 Active-site geometry

The active site of MoaC is located at the dimer interface and is composed of

residues located in six loops L1, L3, L4, L5, L6 and L8 and helix α3, as observed in

EcMoaC (Wuebbens et al., 2000). The residues Lys19, Arg24, Lys49, Gly50, His75,

Thr107, Gly108, Glu110, Met111, Glu112, Asp126, Met127, Lys129, Lys145, Gly147

and Gly148 make up the active site of TtMoaC. In addition, Lys65, which is located in

helix α2 of the third subunit of the hexamer, is involved in active-site formation.

Interestingly, Nζ atom of Lys65 is tri-coordinated to the main-chain carbonyl oxygen

atoms of the conserved residues Ile71, Pro72 and Cys74. Furthermore, the surface-

charge distribution of TtMoaC is uniform, however the active site is positively charged

owing to the presence of basic residues (Figure 5.2).

Figure 5.1 Overall tertiary structure of TtMoaC represented as cartoon. Secondary-structure elements are assigned using the program DSSP (Kabsch and Sander, 1983).


5.2.1.3 Phosphate ion and GTP-binding site

Both the apo crystal structures of TtMoaC contain a phosphate ion (present in

the precipitant solution) in the active site of the protein molecule. It was found that

phosphate ions in the apo forms were located at the position of Pγ of GTP in the

complex structure. The residues Lys49, Cys74, His75, Asp126 and Lys129 are

involved in the hydrogen bonding to the phosphate ions. However, hydrogen bonding

owing to Cys74 is observed in only two subunits. In addition, two water molecules are

observed to be coordinating to the phosphate ion in the structure of form I. In form II,

Lys49 is located at distant and cannot make hydrogen bond to the phosphate ion.

Instead, three water molecules are hydrogen bonded to the phosphate ion. Thus, the

binding of phosphate ions in the active site of the ligand-free forms provides a possible

clue to the binding of a molecule with terminal phosphate groups. In the initial stage of

the refinement of GTP-bound crystal structure, a clear difference electron density (up to

Figure 5.2 Electrostatic potentials of the dimeric subunits of the protein molecule calculated using the module APBS (Baker et al., 2001) plugged into PyMOL. Surface electrostatic potentials that are less than –10 kT, neutral and greater than 10 kT are displayed in red, white and blue, respectively. GTP molecule is shown as ball-and-stick.


8σ in the Fo-Fc map) appropriate for triphosphate groups was observed in the active site

(Figure 5.3).

The GTP-bound crystal structure revealed that the hydrogen-bonding

interactions primarily contributed by the phosphate group stabilize the GTP molecule in

the active site. The residues interacting with GTP are Val47, Lys49, Asp126, Lys129

(from one subunit of the dimer), Cys74, His75, Thr107 (from other subunit of the

dimer) and three water molecules (Figure 5.4). However, interactions owing to Val47

and Cys74 are not observed in all subunits present in the asymmetric unit.

Figure 5.3 Unbiased difference electron-density (2Fo-Fc) map for GTP and citrate ion (FLC) contoured at 0.8σ. The electron densities for both the molecules are shown prior to their addition into the model. Both the molecules are shown as ticks.


5.2.1.4 Other molecules bound in the active site

In addition to the phosphate ions, glycerol (GOL) molecules and acetate (ACT)

ions were observed in the active site of two apo forms I and II, respectively.

Interestingly, in form III, difference electron density (up to 4.7σ in the Fo-Fc map), in

addition to the GTP, was observed in the active site. Based on the ingredients used in

the crystallization, a citrate (FLC) ion was fitted. It is hydrogen bonded to the residues

Arg24, Glu110, Lys145, Lys149, Lys150 and in some subunits, N1 atom of GTP

(Figure 5.4). However, interactions owing to Lys149 and Lys150 are perturbed in most

of the subunits. The average closest distance between GTP and FLC is approximately

3.32 Å. Thus, these observations confirm that a longer molecule such as FPT (an

MoaA-generated intermediate compound) will tightly bind to MoaC.

5.2.1.5 Changes due to substrate binding in the active site

Overall Cα-atom superposition of the apo and GTP-bound crystal structures of

TtMoaC shows a root-mean-square deviation (r.m.s.d.) of 0.4 Å, indicating no

significant change in the overall tertiary structure of the protein molecule. However, the

C-terminal loop regions (residues 148-151) show greater deviation with an average

Figure 5.4 Stereoview of the hydrogen-bonding interactions owing to GTP and FLC at the dimer interface of TtMoaC. The residues involved in hydrogen-bonding interactions are shown as ball-and-stick models in different colors for each subunit. Water molecules are shown as spheres.


r.m.s.d. of 1.7 Å. Furthermore, the electron densities for these residues are not clear in

the GTP-bound crystal structure. This may be due to the binding of the GTP molecule

in the active site, causing a local structural change (Figure 5.5).

5.2.1.6 Invariant water molecules

To study the role of water molecules, a total of 13 crystallographically

independent subunits from two apo forms of TtMoaC were used to identify the

invariant water molecules. The identification of invariant water molecules was carried

out in a similar way as performed previously in our laboratory (see Chapter 4 for

details). Water molecules in a pair of subunits were considered to be equivalent if they

were less than or equal to 1.8 Å apart when the subunits, together with their hydration

shells, were superposed on each other and if they had at least one common interaction

with the protein molecule. Water molecules that are equivalent in all possible pairs

Figure 5.5 Active-site superposition of two apo forms (P21, red; R32, green) and one complex form (C2221, yellow). Phosphate (PO4) ions observed in the apo forms of the crystal structures are also shown. For comparison, glycerol (GOL) molecules and acetate (ACT) and citrate (FLC) ions observed in the crystal structures are also shown.


among the subunits considered are termed as invariant. A total of 16 invariant water

molecules (Table 5.2) were identified.

Table 5.2 Invariant water molecules and hydrogen bonds due to them in the crystal structure of TtMoaC.

1 2 3 4 5 6 IW1 817 Oγ1 Thr25, Oγ1 Thr27, Nδ1 His144 -0.7 1.6 0.97

IW2 818 N Leu37, Oδ2 Asp126, Sδ Met127 -1.0 27.6 0.99

IW3 820 O Leu62, O HOH690 -0.6 13.7 0.68

IW4 822 Sγ Cys74, O His75, Oε2 Glu112, O HOH1185 0.3 1.5 0.73

IW5 823 O Thr25, N Val109 -0.6 0.0 0.83

IW6 824 N Val14, O Leu73, O4 PO4 169, O HOH861 -0.2 14.0 1.00

IW7 827 O Val54, Oε1 Gln57, O HOH849 -0.5 9.1 0.03

IW8 830 O Thr100, O HOH1153 0.0 3.3 1.00

IW9 832 O Asp126, N Ala130, O HOH1178 -0.5 0.2 0.89

IW10 834 N Leu88, O HOH857 0.3 25.0 0.87

IW11 837 N & O Val86, O HOH840 & 1044 -0.1 9.6 0.69

IW12 840 O Leu53, O HOH 837, 846 & 849 1.0 1.6 0.46

IW13 841 Oδ1 Asp69, O HOH845 & 867 0.5 33.0 1.00

IW14 849 O HOH827 & 840 1.1 27.1 0.21

IW15 856 N Val47, N Lys49, O Met127, O HOH 862 -0.2 0.4 0.77

IW16 862 O Gly45, N Gly48, N Gly50, O HOH 1036 0.1 7.5 0.53

1. Invariant water-numbering scheme, 2. Water molecule number in the crystal structure of form I, 3. Hydrogen-bonding interactions observed in the crystal structure, 4. Average normalized B factor (Å2) calculated using the subunits from the apo crystal structures of TtMoaC, 5. Average solvent accessible surface area (Å2), 6. Average occupancy computed from the MD calculations.

5.2.1.7 Plasticity of TtMoaC

The 22 copies of the TtMoaC subunits in two apo forms and one complex form

provide a database for exploring relatively rigid and flexible regions of the protomer.

Analysis was performed using the program ESCET (Schneider, 2004), which

categorizes the molecule into conformationally rigid and flexible regions by automated

analysis of pairs of error-scaled difference distances (Cruickshank, 1999) of an

ensemble of conformers (e.g. crystal structures from different crystal forms or


molecules related by a noncrystallographic symmetry). To use ESCET, a parameter σ is

employed to divide a subunit into rigid and flexible regions and the value of σ is

calculated from the error estimate using the coordinates of the structures being

compared. In the present calculation, the parameter σ was so chosen as to have roughly

69% and 31% [these values were derived using all the structures (approximately 11671)

belonging to the same class, fold, architecture and molecular topology of MoaC

available in the Protein Data Bank] of residues in the invariant and variable regions,

respectively (Schneider, 2002). Thus, out of 143 consistent residues, 55 are predicted to

be conformationally invariant including all/most residues in the three helices.

Interestingly, more than 90% of the β-sheet residues are predicted to be in an

intermediate state. Most of the loops belong to the flexible regions, except for loops L3

and L5 (Figure 5.6).

Figure 5.6 Cartoon representations of structurally rigid and flexible regions of TtMoaC. The rigid and flexible regions are shown in blue and red, respectively. The residues colored green correspond to an intermediate state. GTP molecules are shown as ball-and-stick.


Comparison of ESCET analysis and the B factors of the each

crystallographically independent subunit in all three structures show a significant

difference (Figure 5.7). Although both methods categorize loops as highly flexible,

there is a difference for α-helices and β-strands. According to ESCET analysis, most of

β-strands falls in an intermediate state, whereas B-factor analysis shows them to be

rigid. A similar difference is found in the case of α-helices (Figure 5.7).

5.2.1.8 Comparison with MoaC from other organisms

A structure-based sequence alignment of TtMoaC, EcMoaC, PhMoaC, StMoaC

and GkMoaC is shown in Figure 5.8. Although the overall sequence identity among

them is low (21%), pairwise structure-based sequence alignment of these proteins

shows that TtMoaC has the most similarity to EcMoaC and GkMoaC (~48%) than to

those with PhMoaC and StMoaC (~35%). In addition, Figure 5.8 shows that almost a

Figure 5.7 B-factor comparisons of all 22 crystallographically independent subunits across all three crystal structures. Segments of each subunit are thickened according to their B-factor values. Segments with low B factors are thinner than those with high B factors. The secondary-structure elements (α-helices, β-strands and loops) are shown in red, yellow and green, respectively.


quarter of the residues (36 of 146) are highly conserved among the species. Out of 36, a

total of five residues Lys49, His75, Thr107, Asp126 and Lys129 are important for

substrate binding and another three residues Arg24, Lys133 and Lys145, are involved

in the binding of the citrate ions. The remaining 28 residues may play a role in

stabilizing the overall tertiary structure. The residue Gly50, which is part of the GTP-

binding motif (GKG), is located in the active site of the protein molecule. Four

residues, Ala56, Gly60, Ala63 and Leu65 are mainly involved in the oligomerization of

the protein molecule. In addition, a LIPXCHP motif (residues 70-76) and the residues

Leu122 and Ile137 are involved in the dimerization of the protein molecule.

Figure 5.8 Structure-based sequence alignment of the MoaC proteins using the program MUSTANG (Konagurthu et al., 2006). Secondary-structure elements are shown for TtMoaC. Highly-conserved and semi-conserved residues are shown in red and blue and are marked by symbols (* and +, respectively) at the bottom of the alignment. Residues possibly involved in the catalytic mechanism of the MoaC protein are shown in green.


The tertiary structures of EcMoaC, PhMoaC, StMoaC and GkMoaC to that of

TtMoaC are similar with a root-mean-square deviation (r.m.s.d.) of 1.0, 1.2, 1.0 and 1.0

Å, respectively. Some minor deviations are observed in the region corresponding to the

helix α1 and loop L3 (Figure 5.9). In addition, amino acids at the N- and C-termini,

absent in the other species, are observed in the case of GkMoaC. In the case of

PhMoaC and StMoaC, there is an insertion of seven residues in loop L6. Moreover, the

C-terminal loop L8 could be traced only in the case of TtMoaC. It is interesting to

recall that this loop shows structural change upon substrate binding.

5.2.2 RESULTS FROM ISOTHERMAL TITRATION CALORIMETRY

EXPERIMENTS Isothermal titration calorimetry (ITC) experiments were also carried out to

ensure the binding of GTP to TtMoaC. The ITC results reveal a dissociation constant of

44.4 ± 8 μM and a binding stoichiometry of 0.4 ± 0.1 sites per monomer for GTP

molecules (Figure 5.10 and Table 5.3). In addition, the compounds GDP and GMP,

which were also used for ITC experiments, showed weak binding affinity (fivefold and

Figure 5.9 For comparison, structural superposition of individual subunits of TtMoaC (red), EcMoaC (green), PhMoaC (blue), StMoaC (yellow) and GkMoaC (orange) is shown here.


15-fold weaker) compared with that of GTP (Figure 5.10 and Table 5.3). However, the

binding parameters for GDP and GMP are approximate as the iterations of the curve

fitting were never saturated. Comparison of GTP-binding to MoaC and MoaA suggests

that it is greater with MoaA than with MoaC (~150 fold weaker; Hanzelmann and

Schindelin, 2006). These results, together with the crystal structure of GTP-bound

TtMoaC, suggest that GTP is a true substrate for MoaA rather than for MoaC. Thus, it

can be concluded that the substrate molecule of MoaA and MoaC share a common

motif, which is a terminal triphosphate group.

Figure 5.10 Isothermal Titration Calorimetry (ITC) for the binding of GTP to TtMoaC. (a) The upper panel shows the heat change elicited upon successive injections of GTP into TtMoaC. The lower panel shows the binding isotherm as a function of the molar ratio of GTP to TtMoaC. A theoretical curve was fitted to the integrated data using a single-site model. (b) The relative binding isotherm as a function of the molar ratio of ligands (GTP, blue filled dots; GDP, cyan triangles; GMP, pink inverted triangles; dialyzed buffer used as a control, red squares) to TtMoaC are shown. 1 kcal = 4.186 kJ.


Table 5.3 Isothermal Titration Calorimetry (ITC) data for GTP, GDP and GMP to TtMoaC.

T (K)

n Kb

(M-1) ΔHb

(kcal M-1) TΔS

(kcal M-1) ΔGb

(kcal M-1) GTP 293 0.4±0.08 2.25x104±0.38 -11.05±2.59 -5.21 -5.84

GDP$ 293 0.4±0.79 0.47x104±0.21 -9.37±20.0 -4.69 -4.68

GMP$ 293 1.0±24.29 0.15x104±1.02 -4.74±131.4 -0.50 -4.24 $ The values corresponding to GDP and GMP are not accurate, as the iterations of the nonlinear curve fitting were never saturated. 1kcal = 4.186 kJ.

5.2.3 RESULTS FROM MOLECULAR DYNAMICS SIMULATIONS 5.2.3.1 General features

A total of 16 simulations (15 protein-ligand complexes and one protein-alone)

each of 10 ns were carried out to study the protein-ligand interactions and protein

dynamics. Based on the GTP-bound crystal structure of TtMoaC (present study), 11

different ligands were modelled in the active site to study the interactions containing

triphosphate, diphosphate and monophosphate groups (Figure 5.11). All of the protein-

ligand simulations were carried out in the presence of citrate ion (FLC), which was

observed in the GTP-bound crystal structure. Thus, simulations with GTP and two

probable intermediate compounds (FPT and PBT) without FLC (GTPWF) were also

performed (Figure 5.11). The root-mean-square deviation (r.m.s.d.) plots for all of the

simulations are shown in Figure 5.12. The conformations accessed by ligands

containing triphosphate groups during the MD simulations are shown in Figure 5.13.

The figure shows that phosphate groups are rigid compared to the base and sugar rings

of the ligand molecules.


Figure 5.11 Schematic representations of the basic units of the ligands (a) GTP, (b) ATP, (c) XTP, (d) ITP, (e) FPT and (f) PBT.

Figure 5.12 Root-mean-square deviation (r.m.s.d.) plot of Cα atoms from the starting structure of (a) seven simulations corresponding to GTP, (b) three simulations related to ATP, (c) three simulations related to XTP and (d) three simulations related to ITP molecules.


5.2.3.2 Energetics

The protein-ligand interaction energies given in Table 5.4 are so large as to

render the actual values is somewhat meaningless. However, the fact that the

interaction energies calculated from the MD simulations led to a difference value (∆E)

in favor of the correct ligand type is in itself satisfying. Analysis of these results

revealed that the ligands containing triphosphate groups are more favorable when

compared with those of diphosphates and monophosphates. The results obtained from

the ITC experiments are consistent with the above conclusion. Interestingly, among the

ligands containing triphosphates GTP, XTP and PBT show similar interaction energies

compared with that of ITP, GTPWF and FPT, which shows slightly greater interaction

energy and a greater number of hydrogen bonds (Table 5.4). Furthermore, it is observed

that FPT shows better binding energy among all of the molecules considered in the

present study. Expectedly, it is observed that few of the atoms of the FPT molecule

occupy the positions of FLC during the MD simulations (Figure 5.14). Thus, it may be

suggested that a molecule containing triphosphate group and an open sugar ring

(similar to FPT) would be a better substrate for MoaC. However, given the range of

Figure 5.13 Conformations accessed by the ligands (a) GTP, (b) ATP, (c) XTP, (d) ITP (e) GTPWF (f) FPT and (g) PBT at each 100 ps during MD simulations are shown here.


standard deviations, no clear distinction between the different substrates can be made

except for the NTP, NDP and NMP trend.

Table 5.4 Broad average parameters of TtMoaC–ligand interactions derived from calculations.

1 2 3 4 5 6 GTP -236.1 (12.4) - 6.5 5.5 (0.6) 8.6

GDP -202.3 (7.1) -33.8 2.5 5.2 (0.5) 9.0

GMP -151.2 (5.1) -84.9 1.2 3.3 (0.4) 6.3

ATP -212.9 (5.2) -23.2 4.7 5.0 (0.5) 7.5

ADP -167.4 (5.2) -68.7 2.0 3.7 (0.4) 6.7

AMP -132.6 (3.7) -103.5 1.8 3.1 (0.4) 5.6

XTP -233.9 (7.6) -2.2 4.0 5.6 (0.5) 7.2

XDP -199.2 (4.1) -36.9 2.7 5.1 (0.4) 7.5

XMP -123.5 (4.8) -112.6 1.5 3.1 (0.4) 7.1

ITP -250.4 (5.9) 14.3 5.2 5.7 (0.5) 8.1

IDP -199.3 (5.1) -36.8 2.5 5.0 (0.4) 8.4

IMP -127.8 (4.9) -108.3 1.3 3.2 (0.4) 6.5

GTPWF -253.0 (5.5) 16.9 6.5 6.0 (0.6) 7.2

FPT -263.3 (6.5) 27.2 8.5 7.6 (0.6) 7.8

PBT -235.0 (6.4) -1.1 5.3 6.1 (0.6) 7.9

1. Ligands used for the calculations, 2. Interaction energies (Eprotein-ligand; kcal mol-1). The standard deviations (s.d.) are given in parentheses, 3. Interaction-energy difference (∆E) between GTP and the respective ligands (kcal mol-1), 4. Average number of hydrogen bonds observed in the crystal and modelled structures, 5. Average number of hydrogen bonds calculated from structures generated using the MD simulations. Standard deviations (s.d.) are given in parenthesis, 6. Average life of hydrogen bonds during the MD simulations (in ps). 1 kcal = 4.186 kJ. GTPWF denotes GTP without citrate (FLC).


5.2.3.3 Protein dynamics

The residue-wise root-mean-square fluctuation (r.m.s.f.) in Cα positions

averaged over all the simulations together with the average atomic displacement

derived from B factors obtained from crystallographic studies, is shown in Figure 5.15.

The different indicators presented in Figure 5.15 provide valuable insights into the

plasticity and the dynamics of the protein molecule. The regions 19-23, 40-50, 89-92

and 147-152 are highly flexible. Of these, 40-50 and 147-152 are involved in the

substrate binding. Interestingly, the protein dynamics obtained from the

crystallographic B factors and the MD simulations differ in several regions of the

subunit (Figure 5.16).

Figure 5.14 Structures (in total 45) of FPT generated at every 100 ps from the trajectories obtained from the MD simulations. Each conformation of FPT accessed during the MD simulations is shown at the dimeric interface of TtMoaC. In addition, nine citrate (FLC) ions (one from each subunit) observed in the crystal structure of GTP-bound TtMoaC are shown in green. Each subunit of a dimer is colored differently. Active-site residues involved in the substrate and citrate (FLC) ion binding are also displayed and labelled.


Figure 5.15 Different representations of the plasticity of TtMoaC. Panel (a) Average r.m.s.f.s calculated from B factors obtained from crystal structures of TtMoaC. (b) Average r.m.s.f.s computed using the structures generated from the MD simulations. (c) Relatively rigid (bottom line), intermediate (middle line) and flexible (upper line) regions of the subunit predicted using the program ESCET (Schneider, 2004) and (d) Secondary-structural elements of TtMoaC.

Figure 5.16 Overlay of all of 22 crystallographically independent subunits from all of three crystal structures are shown as ribbon. Secondary-structural elements α-helices, β-strands and loops are shown in cyan, magenta and brown, respectively.


5.2.3.4 Role of invariant water molecules

The location of invariant water molecules identified from crystal structures,

together with their hydrogen-bond interactions, is shown in Figure 5.17. Their

normalized B factors, solvent-accessible surface area and occupancies computed from

the MD calculations are provided in Table 5.2. Of the 16 identified invariant water

molecules, seven (IW2, IW4, IW5, IW6, IW9, IW15 and IW16) are located in the

vicinity of the active site and show a low average normalized B factor and a high

occupancy (≥70%) computed using the structures generated during the MD

simulations, with the exception of IW16. In addition, most of them are buried water

molecules (Table 5.2). Interestingly, three of them (IW2, IW4 and IW9) are hydrogen

bonded to the highly conserved residues His75 and Asp126 that are crucial for the

substrate binding (Figure 5.4). Another water molecule IW6 makes a hydrogen bond to

the phosphate ion observed in the active site of the apo forms. Furthermore, two water

molecules IW15 and IW16 seem to stabilize the active-site loop L3. A set of five water

molecules, IW7, IW10, IW11, IW12 and IW14 (which are located on the protein

surface with an average solvent accessibility of 14 Å2), are involved in a water bridge

near the active site and are flexible with a high average normalized B factor (Table

5.2). As expected, most of these water molecules have low occupancy as computed

from the MD calculations, with the exception of IW10 (Table 5.2). Other four water

molecules IW1, IW3, IW8 and IW13 (which are also located on the protein surface)

show low normalized B factors and high occupancy (Table 5.2). However, the role for

these water molecules is not clear.


5.2.4 A POSSIBLE MECHANISMS FOR THE FIRST STEP OF Moco-

BIOSYNTHESIS PATHWAY Based on the previous studies on MoaA (Hanzelmann and Schindelin, 2006)

and MoaC (Wuebbens et al., 2000) along with the present work, it may be suggested

that an intermediate compound (FPT) generated by MoaA is the most potent substrate

molecule for MoaC. However, the possibility of another compound (PBT) as a

substrate molecule for MoaC cannot be neglected. Thus, two possible sets of

mechanisms are proposed here. Firstly, in the case where FPT is the substrate for

MoaC, precursor Z (a final stable compound in the first step of Moco-biosynthesis

pathway) can be generated in two ways (Figure 5.18a). In the second case, the ring

formation of the FPT molecule is completed first and the resulting compound (PBT)

may play the role of the substrate for MoaC (Figure 5.18b). However, the interaction

Figure 5.17 Hydrogen-bonding interactions of invariant water molecules. The invariant water molecules are shown as spheres in cyan. Only the polar groups of the interacting residues of the protein molecule are shown for clarity. One of the residues (carbon atoms are shown in green) is from the other monomer of the dimer.


energies computed using the MD simulations suggest that the first assumption is more

favorable.

5.3 CONCLUSION The crystallographic study of two apo forms and one GTP-bound crystal

structures of MoaC from T. thermophilus coupled with the ITC experiments and the

MD simulations provide insights into the substrate binding, structure dynamics and

possible mechanism. The GTP-bound crystal structure reveals that residues Lys49,

His75, Asp126 and Lys129 are critical for the activity of the protein molecule.

Together with the interaction energies calculated from the MD simulations, the ITC

results provide insights into the differentiation of the molecules binding to the protein

molecule. They suggest that the molecules with triphosphates are more potent for

Figures 5.18 Schematic diagram of possible mechanisms proposed for the first step of the Moco-biosynthesis pathway involving two probable substrate molecules (a) FPT and (b) PBT for MoaC (see text for details). R and R1 denote the triphosphate and monophosphate groups, respectively.


binding to MoaC. The study of the protein plasticity reveals that all the α-helices,

which are conserved among MoaC from different species, are highly rigid, whereas

most of the β-strands are flexible. In addition, 16 invariant water molecules are

identified, some of which are located in the vicinity of the active site. Interestingly,

water molecules IW2, IW4 and IW9 may play a functional role in the catalytic activity

of the protein molecule. The interaction energies obtained from the MD simulations for

the protein-ligand complexes revealed no clear distinction between the different

substrates, except for the NTP, NDP and NMP trend. In addition, these results support

the crystallographic and ITC results.

5.4 MATERIALS AND METHODS 5.4.1 CLONING, EXPRESSION AND PROTEIN PURIFICATION

The moaC gene (TTHA1789) was amplified by PCR using Thermus

thermophilus HB8 genomic DNA as the template. The amplified fragment was cloned

under the control of the T7 promoter of the E. coli expression vector pET-11a

(Novagen). The expression vector was introduced into the E. coli BL21(DE3) strain

(Novagen) and the recombinant strain was cultured in 6 l LB medium supplemented

with 50 μg ml-1 ampicillin. The cells (20.5 g) were collected by centrifugation, washed

with 20 ml buffer A (20 mM Tris–HCl pH 8.0) containing 50 mM NaCl and

resuspended in 70 ml of the same buffer. The cells were then disrupted by sonication in

a chilled water bath and the cell lysate was incubated at 343 K for 10 min. The sample

was centrifuged at 150000g for 1 h at 277 K and ammonium sulfate was then added to

the supernatant to a final concentration of 1.5 M. The sample was then applied onto a

Resource Phe column (GE Healthcare Biosciences) pre-equilibrated with 50 mM

sodium phosphate buffer pH 7.0 containing 1.35 M ammonium sulfate, which was

eluted with a linear gradient of 1.5 M ammonium sulfate. The eluted fractions

containing the recombinant MoaC were collected, desalted by fractionation on a HiPrep

26/10 desalting column pre-equilibrated with buffer A and applied onto a Resource Q

column (GE Healthcare Biosciences) pre-equilibrated with the same buffer, which was

eluted with a linear gradient of 0–0.5 M NaCl. The eluted fractions containing the

recombinant protein were collected and desalted by fractionation on a HiPrep 26/10

desalting column pre-equilibrated with sodium phosphate buffer pH 7.0 and applied


onto a hydroxyapatite CHT10 column (Bio-Rad Laboratories), which was eluted with a

linear gradient of 10–500 mM potassium phosphate buffer pH 7.0. The eluted fractions

containing the recombinant protein were collected and concentrated with a Vivaspin 20

concentrator (5 kDa molecular-weight cutoff, Sartorius) and loaded onto a HiLoad

16/60 Superdex 75 pg column (GE Healthcare Biosciences) pre-equilibrated with

buffer A containing 150 mM NaCl.

5.4.2 PROTEIN CRYSTALLIZATION The purified protein was concentrated using a Vivaspin 20 concentrator (5 kDa

molecular-weight cutoff, Sartorius). The protein concentration was determined by

measuring the absorbance at 280 nm (Kuramitsu et al., 1990). The concentration of the

purified protein was 11 mg ml-1 in 20 mM Tris–HCl pH 8.0, 150 mM NaCl.

Crystallization experiments were performed using the sitting-drop vapour-diffusion

method. Preliminary screenings were performed using several available kits. The

diffraction-quality crystals were obtained in two different conditions using Emerald

Biostructures Cryo I and II and Hampton Research SaltRX kits. In the first case,

crystals were obtained when 1 μl protein solution was mixed with 1 μl reservoir

solution and allowed to equilibrate against 100 μl reservoir solution at 293 K. The

reservoir solution consisted of 25% (v/v) 1,2-propanediol, 5% (w/v) PEG 3000, 0.1 M

phosphate–citrate buffer pH 4.2 and 10% (v/v) glycerol. The crystals appeared in about

three days (Figure 5.19a).

In the second case, the chosen conditions were further optimized using narrow

intervals of pH (4.6–5.0). Crystals were obtained when 1 μl protein solution was mixed

with 1 μl reservoir solution and allowed to equilibrate against 100 μl reservoir solution

at 293 K. The reservoir solution contained 0.1 M sodium acetate buffer pH 5.0 and 1.0

M ammonium dihydrogen phosphate. Diffraction-quality crystals appeared in about

five days (Figure 5.19b). 100% (v/v) paraffin oil was used as a cryoprotectant.

The crystallization of the GTP-bound form of TtMoaC was carried out using the

conditions that were used for the apo form. The protein solution (0.6 mM) was

incubated overnight with GTP at the final concentration of 10 mM before the

crystallization. A droplet consisting of 2 μl of protein solution and 2 μl of reservoir

solution was equilibrated against 200 μl of reservoir solution consisted of 0.1 M


phosphate-citrate buffer pH 4.2, 25% (v/v) 1,2-propanediol, 5% (w/v) PEG 3350 and

10% (v/v) glycerol. Crystals appeared within a week (Figure 5.19c).

5.4.3 DATA COLLECTION AND PROCESSING Diffraction data were collected from the monoclinic crystal at 100 K at the

RIKEN Structural Genomics Beamline II (BL26B2) at SPring-8 (Hyogo, Japan) using a

Jupiter210 CCD detector (Rigaku MSC Co., Tokyo, Japan). In the case of the primitive

rhombohedral crystal, diffraction data were collected at 100 K using beamline 22-BM

(SER-CAT) at the Advanced Photon Source, Argonne National Laboratory (Argonne,

IL, USA) using a MAR 225 CCD detector (MAR Research USA, Evanton, IL, USA).

In both cases, the distance between the crystal and the detector was maintained at 180

Figure 5.19 Crystal images of two apo forms (a) monoclinic (P21) and (b) rhombohedral (R32) and a GTP-bound form (c) orthorhombic (C2221) of TtMoaC.


mm. The monoclinic and the rhombohedral crystals diffracted to 1.9 and 1.75 Å

resolution, respectively. The intensity data for the GTP-bound crystal were collected at

100 K using the home source MAR 345 imaging-plate detectors mounted on a Rigaku

RU-300 generator (operated at 40 kV and 80 mA). Crystal-to-detector distance was

kept at 220 mm. The GTP-bound crystal diffracted to 2.5 Å resolution. The data were

processed and scaled using the HKL suite (Otwinowski and Minor, 1997). The details

of the data collection statistics are given in Table 5.1.

5.4.4 STRUCTURE SOLUTION, REFINEMENT AND VALIDATION The crystal structures of two apo forms (P21 and R32) were solved by

molecular-replacement calculations using the program Phaser (McCoy et al., 2007).

The atomic coordinates of EcMoaC (PDB-id: 1EKR; Wuebbens et al., 2000) were used

as the search model, which has 53% sequence identity to TtMoaC. A total of 5% of the

reflections were kept aside for the calculations of Rfree (Brunger, 1992). In both apo

forms, difference electron density (up to 12σ in the Fo-Fc map) appropriate for a

phosphate ion was observed in the active site. However, water molecules were first

located and added using the difference electron-density (2Fo-Fc and Fo-Fc) maps with a

criterion of peak heights greater than 0.8σ and 2.8σ, respectively. Subsequently,

phosphate ions were also modelled and refined. The details of the refinement statistics

are given in Table 5.1.

A similar approach to that described above was used to refine the GTP-bound

form. Preliminary calculations (Matthews coefficient of 1.99 Å3 Da-1, solvent content

of 38.1%) suggested the presence of nine subunits in the asymmetric unit (Matthews,

1968). In the initial stages of the refinement, a clear difference electron density (up to

8σ in the Fo-Fc map) appropriate for triphosphate group was observed in the active site.

However, water molecules were first located and fitted into the model to improve the

electron density for the bound ligand molecule. Subsequently, GTP molecules were

added to the model and refined. Topology parameters for GTP were generated using the

HIC-UP web server (Kleywegt, 2007). The refinement statistics are given in Table 5.1.

The molecular modelling program COOT (Emsley and Cowtan, 2004) was used to

display the electron-density maps for model fitting and adjustments. All atoms were

refined with unit occupancies. Refinement was carried out using the program CNS


v.1.2 (Brunger et al., 1998). Simulated annealing omit maps were calculated to correct

or check the final protein models. Two programs PROCHECK (Laskowski et al., 1993)

and MolProbity (Chen et al., 2010) were used to check and validate the quality of the

final refined models. The final refined models and structure factors were checked and

validated using the ADIT server. The atomic coordinates and structure factors of both

the apo forms (PDB-ids: 3JQJ and 3JQK) and the GTP-bound form (PDB-id: 3JQM)

have been deposited in the RCSB Protein Data Bank (Berman et al., 2000).

5.4.5 ISOTHERMAL TITRATION CALORIMETRY All the isothermal titration calorimetry (ITC) experiments were performed in a

VP-ITC MicroCalorimeter (MicroCal Inc., Northampton, Massachusetts, USA) at 293

K. In each experiment, the purified TtMoaC protein was dialyzed against 20 mM Tris-

HCl buffer pH 8.0, 0.15 M NaCl for 12 h with three changes. The ligand solutions were

prepared in the final dialyzed protein buffer. The sample cell (volume 1.4 ml) was filled

with 75 μM of the purified TtMoaC protein solution. The ligand concentrations in the

ITC syringe (volume 298 μl) were 1 mM. Thus, the ITC experiments were performed

under conditions in which the C value (Kb × Mt, where Kb and Mt represent the binding

constant and the enzyme concentration, respectively) was greater than 1. Titrations

were performed by a stepwise addition of small volumes (7 μl) of ligand solutions from

the stirred syringe (307 rev min-1) into the sample cell. A time interval of 180 s was

used between successive injections. The values of the change in binding enthalpy

(∆Hb), binding constant (Kb) and binding stoichiometry (n) for the titration were

determined by a nonlinear least squares fitting of the data using the program Origin 7.0.

The change in entropy (∆S) was obtained using the equation ∆Gb = ∆Hb - T∆S, where

∆Gb = -RTlnKb; the parameters R and T represent the gas constant and the absolute

temperature (K), respectively.

5.4.6 MOLECULAR DYNAMICS SIMULATION Molecular-dynamics (MD) simulations were performed using the package

GROMACS v.3.3.3 running on parallel processors (Berendsen et al., 1995; Lindahl et

al., 2001). The widely distributed AMBER all atom force-field ports for the

GROMACS suite were used (Duan et al., 2003; Sorin and Ponde, 2005). During the


MD simulations, crystallographic water molecules were removed from the protein

models. A cubic box was generated using the editconf module of GROMACS with a

criterion that the minimum distance between the solute and edge of the box was at least

0.75 nm. The protein models were solvated with SPC (simple point charge) water

model using the genbox program available in the GROMACS suite. Hydrogen atoms

were added to the ligand molecules using the PRODRG web server (Schuettelkopf and

van Aalten, 2004). Parameters derived from AMBER03 (Case et al., 2006) were used

to generate ligand topologies, which were further converted to GROMACS format

using a Perl script (amb2gmx.pl). Furthermore, the partial charges of the ligands were

optimized using the ab initio program Gaussian03 (Frisch et al., 2004). Chloride ions

(in the range of 10-37 mM) were used (wherever needed) to neutralize the overall

charge of the system. Energy minimization was performed using the conjugate-gradient

and steepest-descent methods with the frequency of latter of 1 in 1000 with a maximum

force cutoff of 1 kJ mol-1 nm-1 for convergence of minimization. Subsequently, solvent

equilibration by position-restrained dynamics of 10 ps was carried out. Simulations

utilized the NPT ensembles with Parrinello-Rahman isotropic pressure coupling (τp =

0.5 ps) to 1 bar and Nose-Hoover temperature coupling (τt = 0.1 ps) to 300 K. Long-

range electrostatics was computed using the Particle Mesh Ewald (PME; Darden et al.,

1993) method with a cutoff of 1.2 nm. A cutoff of 1.5 nm was used to compute the

long-range van der Waals interactions. Bond lengths were constrained with the LINCS

algorithm (Hess et al., 1997). The value for dielectric constant was used as unity as

needed in the case of an explicit solvent MD simulations. Simulations were performed

for a time period of 10 ns for all the structures discussed in the present study. However,

analyses were performed for a time period of last 9 ns. The protein-ligand interaction

energies were calculated using the equation

Eprotein-ligand = (Eprotein-ligand)elec + (Eprotein-ligand)vdw (5.1)

where Eprotein-ligand denotes the interaction energy between protein and the ligand and

‘elec’ and ‘vdw’ denote the electrostatics and van der Waals components of the energy,

respectively. The relative interaction energies among different ligands were obtained

using the formula

∆E = Eprotein-gtp – Eprotein-ligand (5.2)


where Eprotein-gtp is the interaction energy between the protein and the GTP molecule and

Eprotein-ligand is that of the other ligands considered in the MD simulations.

5.4.7 STRUCTURAL ANALYSIS The three-dimensional atomic coordinates of the crystal structures of the

homologous proteins were downloaded from a locally maintained PDB-FTP

anonymous server at the Bioinformatics Centre, Indian Institute of Science, Bangalore,

India. Invariant water molecules were identified using the 3dSS web server (Sumathi et

al., 2006). Most of the analyses of the MD simulations were performed using the

GROMACS tools and locally developed Perl scripts. A freely available PDB Goodies

web server (Hussain et al., 2002) was used to renumber the residues and to analyze the

temperature factors. Figures were generated using the program PyMOL (DeLano

Scientific LLC http://www.pymol.org). Graphs were prepared using Xmgrace (Paul J.

Turner, Center for Coastal and Land-Margin Research, Oregon Graduate Institute of

Science and Technology Beaverton, Oregon). Structures were superposed using the

program ALIGN (Cohen, 1997). Hydrogen bonds were calculated using the program

HBPLUS (McDonald and Thornton, 1994). A donor-hydrogen-acceptor angle greater

than or equal to 120° and donor-acceptor distance less than or equal to 3.5 Å were used

as a criteria for delineating hydrogen bonds. Solvent-accessible surface area of the

invariant water molecules was computed using the program NACCESS (Hubbard and

Thornton, 1993) with a probe radius of 1.4 Å. Water molecules with an accessible

surface area less than or equal to 2.5 Å2 were considered to be internal/buried. The

normalized temperature factor (Bi') for all the invariant water molecules was calculated

using the formula Bi' = (Bi - )/σ(B), where Bi is the B factor of each atom, is

the mean B factor and σ(B) is the standard deviation of the B factors. Structure-based

sequence alignment has been generated using the program MUSTANG (Konagurthu et

al., 2006). Secondary-structural elements for the protein molecule were assigned using

the program DSSP (Kabsch and Sander, 1983). Electrostatic potentials were calculated

using the module APBS (Baker et al., 2001) plugged into PyMOL.

CHAPTER 6 Structure, Dynamics and Functional Implications of

Molybdenum Cofactor Biosynthesis Protein MogA from Two

Thermophilic Organisms

CHAPTER 6: STRUCTURES AND DYNAMICS OF MogA 165

6.1 INTRODUCTION As discussed in the introductory chapter, the trace element of molybdenum is

required for almost all organisms and forms the catalytic center of a large variety of

enzymes that carry out chemical reactions in carbon, nitrogen and sulfur cycles

(Rajagopalan, 1991; Hille, 2002a,b). Molybdoenzymes are found in almost all

organisms, with Saccharomyces as sole exception amongst the well-known model

organisms (Zhang and Gladyshev, 2008). A genetic deficiency of these enzymes leads

to various autosomal recessive diseases with severe neurological symptoms which may

even lead to early childhood death (Johnson et al., 1989a; Reiss, 2000). Molybdenum is

bioavailable as molybdate. It is incorporated into metal factors such as iron-Moco

(FeMoco) and pterin-based Moco which are synthesized through a similar mechanism

relating to scaffold formation, metal activation and cofactor insertion into

molybdoenzymes (Dos Santos et al., 2004; Schwarz, 2005; Schwarz et al., 2009). As

mentioned in the introduction and the previous chapter, the biosynthesis of Moco is

highly conserved among all organisms including humans and is broadly divided into

five steps (Rajagopalan and Johnson, 1992; Schwarz, 2005). The penultimate step of

Moco synthesis, i.e. the adenylation of molybdopterin (MPT) and insertion of

molybdenum into MPT to make Moco, is carried out by two proteins MogA and MoeA,

respectively (Schwarz et al., 1997; Kuper et al., 2004; Llamas et al., 2004; Llamas et

al., 2006).

Here, in this chapter, three crystal structures of MogA from two thermophilic

Gram-negative bacterium Thermus thermophilus HB8 and Aquifex aeolicus VF5 have

been determined and a comparative study has been discussed. The enzymes of

thermophilic organisms are not only thermostable but are also more resistant to

chemical agents than their mesophilic homologues (Sterner and Liebl, 2001; Vieille and

Zeikus, 2001). Although Moco biosynthesis is quite well understood in bacteria and

eukaryotes, it is still not clear in the case of archaeal systems. Nearly all the archaeal

organisms contain MoaB (homologue of MogA), whereas bacterial systems contain

either MoaB or MogA, with E. coli as an exception that contains both. Since MoaB

from E. coli (EcMoaB) is inactive despite binding MPT, its functional role is still

unclear (Bevers et al., 2008). Both organisms in the present study (T. thermophilus and

A. aeolicus) contain MogA. Interestingly, the gene id TTHA0341 of T. thermophilus


HB8 has been annotated as MoaB in the genomic database (CMR). However, based on

our comparative analysis with known structural and experimental results, TTHA0341 is

considered as MogA in the following (see the first section of results and discussion for

details). Comparative analysis of MogA and its homologues MogA from E. coli

(EcMogA; Liu et al., 2000) and Shewanella oneidensis (SoMogA), MoaB from E. coli

(EcMoaB; Bader et al., 2004; Sanishvili et al., 2004), Bacillus cereus (BcMoaB) and

Sulfolobus tokodaii (StMoaB; Antonyuk et al., 2009), Cnx1G from Arabidopsis

thaliana (AtCnx1G; Kuper et al., 2004) and GephG from Homo sapiens (HsGephG;

Schwarz et al., 2001) and Rattus norvegicus (RnGephG; Sola et al., 2001,) revealed the

functional role of the TtMogA and AaMogA proteins.

6.2 RESULTS AND DISCUSSION 6.2.1 ANNOTATION OF TTHA0341 AS MogA

Gene TTHA0341 of T. thermophilus HB8 was annotated as MoaB in the

genomic database (CMR). All the MoaB and MogA proteins belong to a single family

called MoaB-MogA like family owing to their identical function. However, they differ

in their oligomeric states. Our analyses suggest that the TTHA0341 gene is more like

MogA than MoaB and the following points support this conclusion. (i) The other strain

of T. thermophilus (i.e. HB27) contains the same protein with a single mutation

(K159R) and has been annotated as MogA, (ii) On searching the operon databases

(Okuda et al., 2006), only mog operon could be found in T. thermophilus, (iii) Multiple

sequence alignment of TTHA0341 with other MoaB and MogA proteins clearly shows

higher sequence identity to MogA than MoaB (see sequence comparison for details),

(iv) Phylogenetic analysis obtained from multiple sequence alignment grouped

TTHA0341 into the cluster containing MogA proteins (see sequence comparison for

details) and (v) It is known that MoaB proteins form hexamer in addition to trimer

(Sanishvili, et al., 2004) and the surface analysis of TTHA0341 suggested that it is

stable in the trimeric form (see oligomerization for details). Thus, by considering the

above points, it is concluded that the gene id TTHA0341 corresponds to MogA protein

(hereafter referred to as TtMogA).


6.2.2 PROTEIN ACTIVITY Hereafter, unless otherwise mentioned, the numbering scheme and analysis are

those of the TtMogA structure. A previous study of MoaB from Pyrococcus furiosus

(PfMoaB) and EcMoaB suggested that EcMoaB was inactive. However, it can bind to

MPT (Bevers et al., 2008). Thus, it was important to determine whether TtMogA and

AaMogA proteins are active or inactive. Therefore, we analyzed the sequences of all

known active and inactive proteins. We found that Glu46, Arg77 and Thr80 (Asp57,

Arg87 and Thr90, respectively, in PfMoaB), which were suggested to be the residues

most likely to affect the activity of EcMoaB, are conserved in both the TtMogA and

AaMogA proteins. Studies by Llamas et al. (2004) have shown that the residues Ser10,

Asp20, Asp45 and Arg77 are responsible for MPT binding. In addition, the single

mutants D32A and D56A from site-directed mutagenesis of PfMoaB (Asp20 and

Asp45, respectively, in TtMogA) showed almost no activity. Moreover, a mutation

study of the residue Ser112 (PfMoaB), which is highly conserved among all these

proteins, except for TtMogA (in which it is substituted by Gly103), showed the mutant

to be active (Bevers et al., 2008). Thus, comparing the sequences of and experimental

results for proteins homologous to TtMogA and AaMogA, it can be concluded that the

TtMogA and AaMogA proteins are active and are likely to play a role in MPT-

adenylation. However, it is known that Thr83 and Ser114 in AtCnx1G (Thr72 and

G103, respectively, in TtMogA) are crucial for its catalytic activity. Ser114 is important

as it directly interacts with the N2 atom of MPT. In TtMogA, loop L9 containing this

residue is distant from the active site. Since the interaction takes place through side

chain of the serine residue, it might have some consequence for the activity of TtMogA.

6.2.3 CRYSTALLOGRAPHIC RESULTS 6.2.3.1 Overall structure and active site of TtMogA and AaMogA

The asymmetric unit of TtMogA consists of three crystallographically

independent molecules (Table 6.1) containing residues 1-159, 1-163 and 1-159 (out of

164). The monomeric dimensions of TtMogA are ~42 X 38 X 45 Å. Each monomer

consists of seven α-helices, six β-strands and two 310–helices.


Table 6.1 X-ray data and refinement statistics of TtMogA and AaMogA. Values in the parenthesis are for the highest-resolution shell.

TtMogA (P21) AaMogA (P21) AaMogA (P1) Data collection and processing Wavelength (Å) 1.0 1.0 1.0

Temperature (K) 100 100 100

Space group P21 P21 P1 Unit-cell parameters a, b and c (Å) α, β, γ (°)

33.94, 103.32, 59.59 β=101.3

39.41, 113.16, 55.98 β=93.4

40.02, 64.07, 102.34 95.1, 98.1, 106.9

Resolution (Å) 50.0-1.64 (1.70-1.64) 50.0-1.7 (1.76-1.70) 50-1.9 (1.97-1.90)

Observed reflections 253272 283367 205366

Unique reflections 48481 (4585) 53022 (5250) 71606 (7110)

Completeness (%) 98.7 (93.0) 99.6 (99.9) 96.4 (95.5)

Matthews coeff. (Å3 Da-1) 1.90 2.19 2.28

Solvent content (%) 35.2 43.8 46.1

Multiplicity 5.2 (4.0) 5.3 (5.4) 2.9 (2.9)

I/σ(I) 23.3 (2.5) 33.5 (3.6) 21.7 (4.3)

Rmerge# (%) 6.3 (25.5) 5.4 (26.7) 4.4 (21.9)

Refinement Statistics

Rwork (%) 19.2 19.5 20.8

Rfree (%) 21.7 22.7 23.9

Protein Model Subunits/ASU 3 3 6

Protein atoms 3675 3990 7815

Water molecules 562 625 953

Others 2 8 1

Deviations from ideal geometry Bonds lengths (Å) 0.005 0.006 0.006

Bond angles (°) 1.3 1.3 1.2

Dihedral angles (°) 22.8 23.0 23.1

Improper angles (°) 0.96 0.97 0.93

Average temperature factors (Å2) Protein atoms 23.8 25.1 31.6

Water molecules 35.1 36.6 35.2

Others 36.7 43.5 35.5

Ramachandran plot (%) Most favored 91.7 91.8 90.8

Additionally allowed 8.3 8.2 9.6 # Rmerge = ΣhΣi |I(h)i - <I(h)>| / ΣhΣiI(h)i, where I(h) is the intensity of reflection h, Σh is the sum over all reflections and Σi is the sum over i measurements of reflection h.


The overall tertiary structure of the protein belongs to the Rossmann-like fold

(Figure 6.1). The twisted central β-sheet is sandwiched between the seven α-helices

(five on one side and two on the other). All of the β-strands in the sheet are parallel,

except for β5. The two C-terminal α-helices (α6 and α7) are connected by a 310-helix

(η2). The conserved residue Pro144 in η2 induces a kink of 72° between the α-helices.

Helix α1 is perpendicular to all the helices. The other six helices α2-α7 are parallel to

each other, with the exception of α5.

The crystal structure of AaMogA has been solved in two forms. The asymmetric

unit of the two forms contains three and six subunits, respectively (Table 6.1). In both

forms, most of the residues were clearly observed in the difference electron-density

(2Fo-Fc and Fo-Fc) maps, except for two or three residues at the N- and/or C-terminus in

some subunits. In form II, the electron density for residues 15-22 was not clear in two

subunits. The overall three-dimensional structure of AaMogA is similar to that of

TtMogA with an r.m.s.d. of 1.4 Å, except at the terminal residues.

Figure 6.1 (a) Overall tertiary structure of TtMogA. The AMP and MPT binding sites (AMPBS and MPTBS, respectively) are shown by arrows. (b) Overall tertiary structure of AaMogA. In both the cases, secondary-structure elements and loops are labelled.


Residues from helices α5, α6 and η1, strand β1 and loops L1, L2 and L6

surround the active site depression and can be divided in two parts based on the crystal

structure of AtCnx1G (PDB-id: 1UUY), namely, the MPT-binding site (MPTBS) and

AMP-binding site (AMPBS; Figure 6.1). Gly70 and Gly130 in loops L6 and L12,

respectively, separate the two sites. Thr72, Met99, Ala110, Ser113, Pro129 and Ser138

contribute to forming the floor of the MPTBS depression. Similarly, Val9, Ser10,

Asp20, Thr22, Asp45, Asn69, and Asp78 are involved in formation of the floor of the

AMPBS.

6.2.3.2 Sequence comparison

A search of MoaB and MogA proteins in the Swiss-Prot sequence database

resulted in a total of 31 reviewed and manually curated nonredundant sequences. A

multiple sequence alignment of 15 sequences (six for MogA, one each for Cnx1G,

GephG and Cinnamon and six for MoaB) are shown in Figure 6.2. Protein sequences

were chosen based on the criterion that its structure and/or experimental result were

known. The sequence alignment shows that the GGTG signature motif is highly

conserved in these proteins across species. Of these, Thr72 is involved in

pyrophosphate-bond formation and/or pyrophosphate release (Llamas et al., 2004). The

functionally important residues Ser10, Asp20, Asp45 (except in EcMoaB) and Asp78

are also conserved in these proteins (Kuper et al., 2003; Llamas et al., 2004). Another

sequence motif PGX is also conserved with a mutation in the third position. In MoaB

proteins X is a serine residue. However, MogA proteins show no conservation at this

position. In MogA proteins, X can be asparagine, lysine or glutamine (Figure 6.2). In

the crystal structure of the ligand bound form of AtCnx1G, the Nδ2 atom of Asn142

(Ser131 in TtMogA) of the PGX motif interacts with the O4 atom of MPT. This

suggests that the mutation of Asn142 to a lysine or a glutamine may be acceptable,

whereas that to serine is not. However, a study on PfMoaB, which contains serine at

this position, showed that the PfMoaB protein is active (Bevers et al., 2008).


In addition, the semi-conserved residue Asp11 forms ion pairs with Arg77. The

corresponding residue in EcMoaB is replaced by a glycine, which affects its activity

(Bevers et al., 2008). While most of the homologous proteins maintain the conservation

of this ion pair, StMoaB and EcMoaB show differences (natural mutation to threonine).

Notably, the ion pair is involved in raising the wall near the AMPBS. Furthermore, two

conserved residues Asp45 and Asp78 have been shown to be essential for MPT binding

Figure 6.2 Multiple sequence alignment of MogA, MoaB, Cnx1G and GephG sequences. The protein sequences are taken from Swiss-Prot. The alignment was generated using ClustalW (Larkin et al., 2007). Consensus sequence calculation was performed using a threshold of 80% for the conserved residues. Completely conserved residues are shown as white letters on a red background and the semi-conserved residues are shown in red and boxed. Secondary structural elements are shown for TtMogA (top) and EcMoaB (bottom).


and/or Mg2+ coordination (Sola et al., 2001; Llamas et al., 2004; Sanishvili et al.,

2004). Another feature, which might play a role in inactivating EcMoaB, is the binding

of molybdenum in the active site. It has been observed that a water molecule and

His148 (Tyr154 in AaMogA) or two water molecules in AtCnx1G are responsible for

holding the metal copper (Kuper et al., 2004). In contrast, in EcMoaB this position is

replaced by alanine. Sequence comparison also revealed that the residue Ala83 is only

replaced by threonine or serine in the archaeal proteins StMoaB and PfMoaB, although

its role is not clear (Figure 6.2).

6.2.3.3 Sequence determinants of quaternary structure

The phylogenetic tree obtained from the MSA of MoaB, MogA and its

homologues reveals that proteins that form the hexameric (MoaB) and trimeric (MogA)

quaternary structures were clustered separately (Figure 6.3). It suggests that sequences

of these two types of the proteins determine their oligomeric states. Thus, an analysis of

the sequences and available structures was carried out in order to identify the residues

involved in this feature. Firstly, the residues involved in the trimer-trimer interactions

were identified in the crystal structure of EcMoaB, BcMoaB and StMoaB. The

identified residues were Arg54, Tyr55, Arg58, Ala59, Ser62, Ala63, Ile65, Ala66,

Pro93, Leu94, Asp96, and Asn129 in EcMoaB (see Figure 6.2 for corresponding

residues in BcMoaB and StMoaB). A pairwise sequence alignment of TtMogA and

EcMoaB revealed that Glu46, Asp59, Arg120 and Gly121 (in TtMogA) might be

involved in hexamerization. However, Glu46 and Gly121 are less favorable since these

are chemically similar to the corresponding residues of EcMoaB. Thus, Asp59 and

Arg120 are the residues that strongly contribute to the formation of the oligomer.

Furthermore, the reduced entropy was calculated for all the ungapped sites in

the MSA (Figure 6.2), which resulted in 135 such sites (referred to in the following as

alignment sites). The sequences were grouped in two clusters: (i) the MogA group

containing TtMogA, AaMogA, EcMogA, SoMogA, Helicobacter pylori MogA

(HpMogA), Haemophilus influenzae MogA (HiMogA), HsGephG and AtCn1xG and

(ii) the MoaB group containing EcMoaB, BcMoaB, StMoaB, PfMoaB, Bacillus subtilis

MoaB (BsMoaB) and Staphylococcus aureus MoaB (SaMoaB). The entropy values

were calculated for both of the clusters separately. The amino acids were grouped into


the following physicochemical classes: aromatic (Phe, Tyr and Trp), bulky aliphatic

(Leu, Ile, Val and Met), small nonpolar (Gly and Ala), acidic or amide (Glu, Asp, Gln

and Asn), basic (Lys, Arg and His), those with hydroxyl groups (Ser and Thr) and

others (Pro and Cys) (Ptitsyn, 1998). The entropy values were calculated using the

formula

( ) ( )( )[ ]n2

1miplnipSc

1i

−+= ∑

=σσσ (6.1)

where σ is the given class of amino acids, c is the number of classes considered and

pσ(i) is the frequency of residues belonging to an amino-acid type σ at position i in the

sequence alignment. m is the number of amino-acid types for which pσ(i) ≠ 0 and n is

the number of sequences analyzed. The second term corrects a systematic bias in the

estimation of the entropy (Roulston, 1999). To study the entropic effect between

hexameric and trimeric proteins, we calculated the entropy difference for each

alignment site,

trimerichexameric SSS −=Δ (6.2)

where the first term is the reduced entropy of a site in a hexameric cluster and the

second term is that of a trimeric cluster.

Figure 6.3 Phylogenetic tree obtained from the multiple sequence alignment of MogA, MoaB, Cnx1G and GephG. The proteins that are known to form be hexameric and trimeric oligomers are labelled.


The calculated entropy difference for each alignment site is shown in Figure

6.4. Although, the number of sites above and below the baseline (with zero entropy) is

similar, a total of eight sites (23, 34, 47, 64, 98, 106, 115 and 131) show a significant

entropy difference of less than 1.0. Of these, five sites (47, 64, 106, 115 and 131) are

worth mentioning. At site 47, the hexameric proteins contain positively charged

residues, whereas trimeric proteins show no amino-acid conservation. However,

AaMogA and HpMogA contain positively charged residue (arginine) at this site. Thus,

this site alone is not responsible in determining the oligomeric state. A similar pattern is

found at sites 64 and 106 also. However, two sites 115 and 131, along with other sites

and possibly other properties of hexameric proteins, seem to have a high probability of

being involved in determining the oligomeric state. At the site 115, hexameric proteins

contain a conserved alanine residue, whereas this site is dominated by a glutamine

residue in the trimeric proteins. Site 131 belongs to the PGX motif (see sequence

comparison for details). It is interesting to note that although the GGTG motif is

conserved among homologues, another sequence motif PGS is conserved only in MoaB

proteins, with the exception of HsGephG. The serine residue in this motif is replaced by

a lysine or glutamine, with the exception of AtCnx1G (where it is replaced by an

asparagine).

Figure 6.4 Entropy difference (ΔS) as a function of alignment site. The differences are

taken between the hexameric and the trimeric clusters. The alignment sites are given according to the TtMogA sequence in the multiple sequence alignment.


6.2.3.4 Structure comparison

Pairwise structural superposition of all of the structures shows a high similarity

at the tertiary level. Remarkably, even though the sequence similarities among these

proteins are low (ranging from 16% to 69%), their overall three-dimensional structures

are very similar (Table 6.2). The values for the root-mean-square deviation (r.m.s.d.)

show that TtMogA is very similar to HsGephG, whereas AaMogA to SoMogA (Table

6.2). In general, the N- and C-terminal residues show greater dissimilarity. In addition,

the regions 12-18, 25-35 and 95-115 show high r.m.s.d.s compared with the others

regions (Figure 6.5). The regions 12-18 and 95-115 belong to loops L2 and L9 and are

very close to the AMPBS and the MPTBS, covering the active-site-like wall from both

sides. It is notable that loop L9 also shows movement during opening and closing

process of the active-site channel (see next section for details). The region 25-35

belongs to α2 and L2. However, the reason for its high flexibility is not clear.

Table 6.2 Pairwise root-mean-square deviation (r.m.s.d.) values for all the structures. The pairwise sequence-similarity scores obtained from multiple sequence alignment of these sequences are given

in parenthesis. The diagonal elements have 100% sequence similarity.

TtMogA AaMogA EcMogA SoMogA AtCnx1G HsGephG EcMoaB BcMoaB StMoaB

TtMogA 0.0 1.4 (43) 1.5 (40) 1.5 (42) 2.3 (44) 1.1 (45) 1.3 (30) 1.3 (26) 1.5 (23)

AaMogA 0.0 0.9 (55) 0.6 (69) 1.0 (39) 1.1 (41) 1.1 (17) 1.0 (18) 1.3 (17)

EcMogA 0.0 0.9 (55) 1.1 (32) 2.0 (35) 1.5 (18) 1.2 (18) 1.5 (16)

SoMogA 0.0 1.3 (40) 1.2 (41) 1.3 (48) 1.3 (20) 1.6 (16)

AtCnx1G 0.0 1.0 (49) 1.4 (22) 1.4 (26) 1.4 (22)

HsGephG 0.0 1.2 (25) 1.3 (24) 1.9 (27)

EcMoaB 0.0 1.0 (38) 1.0 (25)

BcMoaB 0.0 0.9 (39)

StMoaB 0.0


6.2.3.5 Protein surface charge distribution

An analysis of the charge distribution of all MogA (TtMogA, AaMogA,

EcMogA, SoMogA) and MoaB (EcMoaB, BcMoaB and StMoaB) proteins and their

eukaryotic homologues (AtCnx1G and HsGephG) shows that the active sites of these

proteins are more or less uniform in nature, with the MPTBS positively charged and the

AMPBS negatively charged. However, the overall charge distribution of these proteins

varies substantially. The protein surface of TtMogA is mostly positively charged,

whereas those of other homologues are negatively charged (Figure 6.6). A close

investigation of the amino-acid compositions of all of these proteins revealed that

TtMogA contains marginally more positively charged residues (14%) than negatively

charged residues (13%), in contrast to other homologous proteins which consist of

fewer positively charged residues compared with their negatively charged residues

(Table 6.3). Interestingly, 91% of the positively charged residues of TtMogA are on the

Figure 6.5 Overall tertiary structural superposition of TtMogA, AaMogA, EcMogA, SoMogA, AtCnx1G, HsGephG, EcMoaB, BcMoaB and StMoaB. For clarity, all structures are shown in the same colors. Two loops (L2 and L9 in TtMogA) are labelled.


protein surface and the remaining residues are buried upon trimerization. In contrast,

only 76% of the negatively charged residues of TtMogA are on the protein surface.

Furthermore, the number of ion pairs found in TtMogA and AaMogA are also high

compared with other proteins (Table 6.3). It is known that ion pairs (in addition to other

factors) play a significant role in stabilizing the structure and function of thermophilic

proteins (Karshikoff & Ladenstein, 2001).

Analysis of protein surfaces results in another interesting feature of these

proteins. Near the MPTBS, a surface channel (hereafter, referred to as the active site

channel; ASC) is observed which has two states (open or closed). It is observed that

TtMogA has an open ASC, whereas AaMogA, AtCnx1G, HsGephG and EcMoaB have

closed ASCs (Figures 6.7 and 6.8). Interestingly, BcMoaB and StMoaB show an

intermediate state (Figure 6.8). The crystal structure of TtMogA shows that the residues

in loop L9 forming the ASC are compelled away. However, the other proteins contain

helices in this region and are observed to be in the closed state (Figure 6.9). Although

the structural and/or functional role of ASC does not seem to be trivial, it is tempting to

speculate that it might play a role in substrate (MPT) entry into the active site. Analysis

of surface cavities of all the available crystal structures of MoaB and MogA and their

eukaryotic homologues revealed that TtMogA and AtCnx1G proteins have similar

active-site cavities (volume ~2000 Å3), whereas active-site volume of other

homologues range from 1000 to 1500 Å3. This suggests that TtMogA can bind a similar

molecule as in the case of AtCnx1G.


Figure 6.6 Electrostatic potential charge distributions on the protein surface of (a) TtMogA and (b) EcMogA. For clarity, both the trimeric interfaces (front - left and back -right) are shown.


Table 6.3 Charged amino-acid compositions, hydrogen bonds and ion pairs in all the MoaB and MogA proteins. The percentage values are given in parenthesis.

Protein 1 2 3 4 5 TtMogA 164 23 (14) 21 (12.8) 71 (43.3) 10

AaMogA 178 23 (12.9) 24 (13.4) 77 (43.3) 14

EcMogA 195 18 (9.2) 26 (13.3) 85 (43.6) 11

SoMogA 177 19 (10.7) 27 (15.2) 68 (38.4) 9

AtCnx1G 161 16 (9.9) 22 (13.6) 58 (36.0) 4

HsGephG 167 18 (10.7) 23 (13.7) 78 (46.7) 7

EcMoaB 170 18 (10.5) 22 (12.9) 72 (42.4) 6

BcMoaB 169 22 (13.0) 24 (14.2) 68 (40.2) 5

StMoaB 178 24 (13.4) 24 (13.4) 72 (40.5) 3

1. Protein length, 2. Number of positively charged residues (percentage), 3. Number of negatively charged residues (percentage), 4. Number of hydrogen bonds owing to charged residues (percentage), 5. Number of ion pairs.

Figure 6.7 Electrostatic potential charge distributions at the active sites of TtMogA (top

left), AaMogA (top right), EcMogA (bottom left) and AtCnx1G (bottom right). The two binding sites (AMPBS and MPTBS) and the active-site channel (ASC) are indicated by arrows.


Figure 6.8 Electrostatic potential charge distributions at the active sites of EcMoaB (top left), BcMoaB (top right), StMoaB (bottom left) and HsGephG (bottom right). The two binding sites (AMPBS and MPTBS) and the active-site channel (ASC) are also indicated.

Figure 6.9 Structural superposition of the crystal structures of MogA (cyan, limegreen and orange), MoaB (red, green, blue and yellow), Cnx1G (wheat) and GephG (white), comparing the active-site channel (ASC). The secondary structural elements of TtMogA (red) and EcMogA (cyan) are labelled.


6.2.3.6 Oligomerization

The asymmetric units of TtMogA and one form of AaMogA contain one trimer,

whereas that of the other form of AaMogA contains two trimers. The trimers are

generated by a noncrystallographic 3-fold axis. The MogA, Cnx1G and GephG proteins

have been shown to be active as trimers in solution (Schwarz et al., 2000, 2001; Llamas

et al., 2004). In contrast, EcMoaB, BcMoaB and StMoaB are predicted to be present in

both trimeric and hexameric states. An investigation of the surface-charge distribution

on the hexameric interface of these proteins shows that they have a combination of

alternating positive and negative charges, which aid in the formation of hexamer

(Figure 6.10). In EcMoaB, residues Arg54, Tyr55, Arg58, Ala59, Ser62, Ala63, Ile65,

Ala66, Leu94, Asp96 and Asn129 are located at the trimer-trimer interface. However, it

is not clear why the MoaB proteins form hexamer whereas the MogA proteins form

trimer. The best possible utilization of a hexameric MoaB would seem to be to form

heterohexamer (MoaB-MogA) that facilitates substrate-product exchange without their

dissociation into the external solvent (Sanishvili et al., 2004).

Figure 6.10 Electrostatic potentials of the trimeric interface of EcMoaB (left) and EcMogA (right).


Thus, we analyzed the oligomerization states of all of the crystal structures of

MoaB, MogA, Cnx1G and GephG proteins using the PISA web server (Krissinel and

Henrick, 2007). The results suggest that EcMoaB, BcMoaB and StMoaB are predicted

to be stable in both the trimeric and hexameric states, whereas TtMogA, AaMogA,

EcMogA, SoMogA, AtCnx1G and HsGephG are stable only in trimeric state. A

detailed analysis of the buried surface area and the solvation-energy gain upon

oligomerization of all these proteins are given in Table 6.4. Interestingly, Cnx1G and

GephG are observed to be more stable as trimer compared with other homologues. The

regions involved in trimer formation are 74-78, 82, 90-98, 100-114, 144-145, 148-149

and 152-153 (80-84, 88, 96-104, 106-120, 150-151, 154-155 and 158-159 in AaMogA,

respectively). As expected, almost 60% of these residues are hydrophobic in nature.

Table 6.4 Protein surface analyses using the web server PISA (Krissinel and Henrick, 2007).

Protein 1 2 3 4

TtMogA 19850 4500 -29.6 Trimer

AaMogA 22090 4370 -33.7 Trimer

EcMogA 21770 4460 -27.8 Trimer

SoMogA 21655 4235 -33.4 Trimer

AtCnx1G 18800 7580 -94.6 Trimer

HsGephG 20340 4730 -51.7 Trimer

EcMoaB 19920 5085 -33.1 Trimer

36770 13240 -84.8 Hexamer

BcMoaB 20350 4880 -33.6 Trimer

37520 12940 -79.8 Hexamer

StMoaB 21470 4700 -27.4 Trimer

39260 13070 -76.4 Hexamer

1. Solvent accessible surface area (Å2), 2. Buried surface area (Å2), 3. Free energy difference (∆Gint) in kcal mol-1, 4. Predicted oligomer. 1 kcal = 4.186 kJ.


6.2.3.7 Role of the N- and C-terminal residues

Pairwise sequence alignment of the EcMoaB and EcMogA proteins revealed

that the two regions 1-13 and 106-118 of EcMoaB match region 103-115 of EcMogA

(Figure 6.11). Thus, the region 103-115 of EcMogA has similar sequence repeats in

EcMoaB. The 103-115 region of EcMogA corresponds to loop L9 of TtMogA.

Superposition of all of the crystal structures available for the MoaB, MogA, Cnx1G and

GephG families revealed that the N-terminus of the MoaB proteins extends to the top of

the MPTBS (Figure 6.12). Superposition of AtCnx1G (bound with MPT-AMP) and

EcMoaB shows that the residues at the N-terminus of EcMoaB can easily interact with

MPT. It is interesting to note that StMoaB has the similar N-terminal conformation. In

contrast, in MogA proteins, the C-terminal residues show a conformation that covers

the MPTBS.

These observations clearly distinguish between MoaB and MogA proteins.

However, the eukaryotic homologues (AtCnx1G and HsGephG) do not show the above

feature. Notably, these two proteins are fused with the E-domain in a single two-

domain polypeptide chain. Multiple sequence alignment of these proteins also shows

insertion at the N- and C-termini of MoaB and MogA proteins, respectively. Thus, it

can be concluded that the N- and C-termini of MoaB and MogA proteins, respectively,

play a similar role, possibly in stabilizing the substrate molecule in the active site.

Figure 6.11 Pairwise sequence alignment of EcMoaB (query) and EcMogA (sbjct). The alignment was generated using the program BLAST (Altschul et al., 1990).


6.2.3.8 MogA-MoeA protein-protein complex

It is known that two proteins Cnx1G and Cnx1E (which are homologues of

MogA/MoaB and MoeA in bacteria) are involved in adenylation and metal insertion

into MPT. Also, Cnx1G and Cnx1E both bind MPT with different affinities (Schwarz

et al., 2000). The protein MoeA contains four domains, of which domain III has the

same fold as MogA. It is a molybdate-binding protein and is involved in the transfer of

metal molybdenum into MPT (Schwarz et al., 2000). Owing to the intrinsic instability

of MPT, Moco has to remain bound to protein during the whole biosynthetic process

until its final delivery to apomolybdoenzymes (Magalon et al., 2002). Also, compared

with MPT synthase (MoaD-MoaE complex), MogA and MoeA proteins bind MPT

more strongly (Magalon et al., 2002). Thus, two proteins MogA and MoeA are

believed to form protein-protein complex to carry out the comparatively fast and

unstable MPT-adenylation reaction (Liu et al., 2000; Schwarz et al., 2000; Magalon et

al., 2002). In addition, the MoaB proteins have been suggested to form protein-protein

Figure 6.12 Stereoview of an active-site structural superposition of all of the proteins. The N- and C-termini are labelled in different colors (TtMogA, firebrick; AaMogA, lightblue; EcMogA, limon; EcMoaB, forest; StMoaB, pink; HsGephG, cyan). The two binding sites (AMPBS and MPTBS) and the active-site channel (ASC) are also labelled.


complex with MobB and MoeA in much the same way as MogA does with MoeA

(Sanishvili et al., 2004). Thus, we carried out protein-protein complex docking using

the ClusPro server (Comeau et al., 2007). The proteins EcMogA (PDB-id: 1DI6; Liu et

al., 2000) and EcMoeA (PDB-id: 1G8L; Xiang et al., 2001) were taken as receptor

(trimer) and ligand (dimer), respectively, during docking. The highest ranked

conformer was used for further analysis. Since only a dimeric molecule of MoeA was

taken as the ligand, the best conformation of MoeA showed a possible site for binding

with respect to the single subunit of MogA. Thus, we generated the same MoeA

conformation with respect to the other subunits of MogA by superposition. Similarly,

MogA was generated with respect to the other end of the MoeA dimer interface (Figure

6.13). Interestingly, the MPTBS of MogA is very close (in the range of 10-15 Å) to the

active-site cavity of domain III of MoeA in MogA-MoeA protein complex, which has

been proposed to be more stable in the presence of MPT/Moco (Magalon et al., 2002;

Figure 6.14).

The residues observed in the protein-protein interactions of MogA (MoeA) were

Arg5B (Asp121L), Glu150B (Glu270M), Asn152B (Glu266M), Val153B (Glu266M),

Glu170B (Val76L, Gly78L), Ala183B (Glu257M), Arg185B (Glu257M), Ser188B

(Ala82L, Gly83L, Gln84L), Ala189B (Gln84L), Arg190B (Glu257M), Arg191B

(Arg97L), Asp13C (Gly88L), Glu50C (Glu89L), Arg81C (Glu89L), Phe110C

(Tyr260M), Gln135C (Asp187M) and Lys147C (His231M). The last uppercase letter

denotes the chain identity. Most of the residues of domain III of MoeA interact with the

active-site residues of MogA, whereas residues from domain II of MoeA interact with

those of the N- and C-termini of MogA. As expected, almost 30% and 70% of the

interacting residues of MogA and MoeA, respectively, are predicted to be involved in

protein-protein interactions using the PPI-Pred server (Bradford & Westhead, 2005).

Sequence comparison of EcMogA with AtCnx1G and HsGephG reveals that

almost 50% of the residues involved in protein-protein interactions are similar in

nature. Of these, four residues Asp13, Glu50, Arg81 and Gln135, of EcMogA are of

particular importance. Asp13 (Asp11 in TtMogA) is essential for maintaining the ion

pair with the residue Arg81 (Arg77 in TtMogA; see sequence comparison for details).

Glu50 (Glu46 in TtMogA) is similar in nature among all the homologous proteins,

except for EcMoaB (see protein activity section for details). Gln135 (Ser131 in


TtMogA) is possibly involved in oligomerization (see sequence determinants of

oligomerization for details). In addition to the dimeric ligand, protein docking was also

carried out considering monomeric MoeA. A comparison of the two best conformers

obtained from dimeric and monomeric MoeA protein docking shows that the

conformations of the two proteins are different. In the case of dimeric MoeA, most of

the interactions are between the residues belonging to domains II and III from two

different subunits of the dimer, whereas in the case of monomeric MoeA, the

interactions are mainly between the residues of domains III and IV. However, almost

30% of the interactions are common to both conformers.

Figure 6.13 Protein-protein interactions (EcMogA-EcMoeA). EcMogA (trimeric) is shown in red and EcMoeA (dimeric) in blue and green. The interacting domains of MoeA are labelled. Two dimers of MoeA with respect to two subunits of MogA were obtained by superposition of the conformations obtained from protein-protein docking program ClusPro (Comeau et al., 2007). In a similar way, three trimers of MogA were also generated with respect to the other dimeric interface of MoeA.


6.2.3.9 Invariant and interfacial water molecules

Water molecules are known to play an important role in the structure and/or

function of many proteins (Halle, 2004; Smolin et al., 2005; Chapter 4). Thus, invariant

water molecules and those located at the subunit interfaces were identified. A total of

12 (nine from AaMogA and three from TtMogA) crystallographically independent

subunits were used separately to identify the invariant water molecules. Identification

of invariant water molecules was carried out using a similar method to that described in

the previous work of this laboratory (see Chapter 4 for details). In total, 12 water

molecules were identified as invariant (Table 6.5, Figures 6.15 and 6.16). Most of them

interacts with the polar backbone atoms of the residues and thus are independent of the

amino-acid types. Out of 12, five water molecules IW1, IW2, IW3, IW10 and IW11 are

located in a cavity generated by the trimeric subunits. A further five water molecules

IW4, IW5, IW6, IW9 and IW12 are close to the active site. The remaining two water

molecules IW7 and IW8 are located on the protein surface far from the active site.

Water molecule IW4 forms a hydrogen bond to the Oγ1 atom of Thr80 (Thr86 in

EcMoaB), which is proposed to be one of the residues that possibly affect the activity

Figure 6.14 Protein-protein interactions (EcMogA-EcMoeA). Active sites of two proteins EcMogA (top) and EcMoeA (bottom) are shown by arrows. Two active sites in protein-protein complex are located at a distance of ~10-15 Å.


of EcMoaB. Similarly, water molecules IW5, IW6, IW9 and IW12 are likely to have

essential roles as they form hydrogen bonds to the highly conserved residues Gly73 and

Asp78, Gly70, Asp45 and Gly72, respectively. Most of the invariant water molecules

are buried, with the exceptions of IW3, IW7, IW8, IW11 and IW12 and have low B

factors (Table 6.5). In addition, most of them (with the exceptions of IW1 and IW11)

have greater than 50% occupancy during the MD simulations.

Table 6.5 Invariant water molecules and their hydrogen bond interactions with the protein and water molecules.

Id 1 2 3 4 5

IW1 209 Leu98 O, Gly102 N, Thr122 Oγ1 0.1 -1.1 0.48

IW2 215 Glu91 Oε2, Thr122 Oγ1, HOH363,441 1.2 -0.9 0.69

IW3 224 Ile117 O, Ser119 O 26.1 -1.0 0.81

IW4 258 Arg83 O, Val85 N, Thr86 Oγ1, HOH319 0.0 -0.5 0.77

IW5 277 Gly79 O, Asp84 O, Arg120 NH1 0.1 -0.8 0.69

IW6 281 Thr74 Oγ1, Gly76 O, Leu134 O 0.0 -0.9 0.77

IW7 295 Val143 O, HOH587, 600 8.4 -0.1 0.82

IW8 308 Ser70 Oγ, Ser128 O & Oγ, HOH355,464 8.3 0.2 0.81

IW9 326 Asp51 O, Arg83 O, HOH375 3.4 -0.1 0.56

IW10 441 HOH215,450,497 4.0 -0.6 0.80

IW11 497 Arg120 O, HOH289,441 6.5 -0.8 0.46

IW12 788 Gly77 N, HOH233,687 14.6 0.2 0.76

1. Water number as in chain A of AaMogA, 2. Hydrogen bond interactions with protein and water molecules, 3. Average solvent accessibility (Å2), 4. Average normalized B factor, 5. Average occupancy calculated during the MD simulations. The average was taken over ligand-free simulations.


Figure 6.15 Overall three-dimensional structure of AaMogA (cartoon) with invariant water molecules (sphere) is shown. Both the binding sites (AMPBS and MPTBS), the N-and C-termini and water molecules are labelled.

Figure 6.16 Schematic representation of invariant water molecules. Hydrogen-bond interactions with various residues of the protein molecules are shown as lines. Water molecules are shown as spheres and the residues are as rectangles. Water molecules and residues in lighter colors are deeper relative to the plane of the paper.


In addition, seven interfacial water molecules were identified in the TtMogA

and AaMogA crystal structures (Table 6.6, Figures 6.17 and 6.18). Water molecule

TGI1 was also identified to be invariant (IW3). Two water molecules TGI2 and TGI3

are located almost on a noncrystallographic 3-fold axis and are hydrogen bonded to

Gly103 from all three subunits of the trimer. Water molecule TGI4 is hydrogen bonded

to Asp78 and Arg90. In a similar fashion, TGI5 is hydrogen bonded to Glu91 and

Arg114. Most of these water molecules show a reasonable occupancy calculated using

the trajectories obtained from MD simulations.

Table 6.6 Water molecules observed at chain interfaces and their hydrogen-bond interactions with the protein molecule.

WID One Chain

Water Other Chain

<SASA> <NBF> <Occupancy>

TGI1 Ala111, Ser113

208 Gly94 0.0 -0.8 0.0

TGI2 Gly103 539 Gly103, Arg105

1.0 1.4 0.57

TGI3 Gly103 422 Gly103 1.3 0.2 0.72

TGI4 Asp78 268 Arg90 0.2 0.4 0.81

TGI5 Arg114 388 Glu91 3.1 0.2 0.99

TGI6 Glu97 398 Glu97 0.0 0.1 1.00

TtM

ogA

TGI7 Arg105 480 Arg105 4.6 1.6 0.82

AGI1 Ile117, Ser119

224 Pro99, Gly100

0.0 -0.5 0.58

AGI2 Arg120 238 Met97 5.4 -0.6 0.72

AGI3 Glu103 310 Glu103 0.4 -0.3 0.78

AGI4 Glu88 213 Glu95 8.7 -0.1 0.67

AGI5 Pro82, Asp84

231 Lys96 3.8 -0.3 0.28

AGI6 Pro114 624 Ala147 0.9 0.3 0.66

AaM

ogA

AGI7 Gln107 794 Gln107 0.0 1.7 0.73

<SASA>: Average solvent accessible surface area (Å2), <NBF>: Average normalized B factor.


Figure 6.17 Water molecules observed at the chain interfaces are shown (as spheres) for TtMogA (red) and AaMogA (blue). The tertiary structure shown here belongs to TtMogA. All the subunits, active sites and both the N- and C-termini are labelled.

Figure 6.18 Schematic representation of the interfacial water molecules. Water molecules belonging to TtMogA (red) and AaMogA (blue) are shown as circles. Water molecules observed at similar positions in both structures are labelled in the same circle. Water molecules observed in only one structure are also labelled in black for TtMogA and blue for AaMogA. The hydrogen-bonding interactions of water molecules with protein molecules are represented by lines. The residues belonging to the three subunits are colored according to Figure 6.17. The lighter colors represent a greater depth relative to the plane of the paper.


6.2.4 MOLECULAR DYNAMICS AND DOCKING RESULTS 6.2.4.1 General features

A total of 42 MD simulations (each of 50 ns) and 47 molecular-docking studies

were carried out to study the protein dynamics and protein-ligand binding energies. A

previous study on the plant protein Cnx1G showed the binding of MPT-AMP (Kuper et

al., 2004). Thus, 14 simulations with purine nucleotides, with MPT and with MPT-

AMS (AMP with one fewer phosphoryl oxygen atom) were carried out with TtMogA at

both binding sites to compare the specificities (Table 6.7). The simulations with MPT-

AMS were carried out in order to mimic the intermediate compound MPT-AMP. In

parallel, molecular-docking studies with these compounds at both binding sites were

also performed. Similarly, to compare the binding specificities of these compounds

with EcMoaB, nine MD simulations were independently carried out only at the MPTBS

as proposed in the previous study (Sanishvili et al., 2004). However, molecular docking

was performed at both binding sites. In a similar way, eight MD and nine molecular-

docking studies were also carried out for AaMogA. It is known from a previous study

(Schwarz et al., 2000) that MogA and MoeA both bind MPT but with different

affinities (MogA > MoeA). Thus, MD simulations and docking studies with AMP,

MPT and MPT-AMP were carried out for both proteins to compare the binding

affinities with those of TtMogA, AaMogA and EcMoaB. The graph showing the root-

mean-square deviation (r.m.s.d.) of Cα atoms for all 42 (37 ligand-bound and 5 ligand-

free) simulations are shown in Figure 6.19.


6.2.4.2 Energetics

The interaction energies calculated using the MD simulations and

intermolecular energies obtained from docking studies are given in Table 6.7.

Expectedly, on account of different set of parameters used in these calculations, there is

a difference in some cases. However, MD results combined with that of molecular-

docking studies reveal several features that relate to different ligand-binding

specificities. A comparison of the binding energies obtained from docking studies

suggests that MPT and MPT-AMP show increased binding to TtMogA and AtCnx1G

compared with EcMoaB, AaMogA and EcMoeA (Table 6.7). However, it has been

shown experimentally that Cnx1E only binds MPT-AMP with a higher affinity than

Cnx1G in the presence of molybdate (Llamas et al., 2006). In the case of ATP and

GTP, the binding energies are greater with TtMogA than with EcMoaB.

Figure 6.19 Root-mean-square deviation (r.m.s.d.) plot of Cα atoms from the starting structure of (a) 15 simulations related to TtMogA (b) nine simulations related to AaMogA (c) ten simulations related to EcMoaB and (d) four simulations each related to AtCnx1G


Table 6.7 Energies calculated from molecular-dynamics and docking methods.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 -116.2 -128.1 -131.7 -196.2 -170.5 -178.1 -117.7 -125.9 -142.7 -166.4 -148.9 -183.1 -74.7 -93.7

4.7 3.3 4.7 5.7 9.4 5.4 5.5 5.0 4.1 5.5 5.0 7.0 8.1 3.8

4.4 5.6 5.0 9.5 5.1 6.0 5.3 4.8 5.9 6.1 5.9 7.0 3.1 1.8

1.7 1.4 1.6 1.7 1.7 1.7 1.9 1.8 1.8 1.7 1.4 1.9 1.2 1.2

-12.4 -12.7 -14.3 -12.8 -15.0 -14.2 -12.2 -13.3 -13.3 -14.0 -14.8 -14.4 -12.5 -11.4/ -12.8

TtM

ogA

29 31 67 103 162 153 35 31 125 79 210 143 4 43/2

-82.8 -117.0 -86.8 -103.0 -82.5 -95.8 -96.6 -111.4

4.9 4.9 3.6 3.0 4.7 6.8 4.1 3.9

2.7 6.7 2.5 6.8 3.4 6.0 5.1 2.5

1.3 1.5 0.9 1.5 1.3 2.6 1.1 1.2

-10.3 -11.8 -12.3 -10.5 -11.9 -10.7 -10.2 -11.8/ -10.2

AaM

ogA

46 113 191 68 153 211 27 50/33

-120.5 -86.9 -86.7 -148.8 -127.3 -118.5 -119.8 -64.1 -142.2

3.1 7.2 7.0 6.5 3.1 4.4 6.4 4.8 4.4

6.0 4.7 3.8 7.0 6.1 6.1 4.9 3.7 3.4

1.4 2.2 1.6 3.0 1.3 1.6 1.7 1.5 1.4

-11.7 -12.7 -14.0 -12.5 -12.5 -12.2 -11.6 -12.0 -12.5 -12.8 -13.6 -14.0 -11.7 -11.6/ -11.8

EcM

oaB

45 70 148 137 198 193 71 79 152 151 220 192 21 48/10

-75.5 -46.5 -135.6

7.9 3.2 2.8

3.3 0.6 4.3

1.5 0.8 1.1

-12.5 -12.4 -10.8/ -12.5

AtC

nx1G

21 2 27/2

-76.0 -39.0 -125.7

3.0 6.9 3.6

5.1 1.2 2.1

1.0 1.4 1.2

-10.1 -10.8 1.0/ -10.8

EcM

oeA

15 23 250/17

1. AMPA, 2. AMPM, 3. ADPA, 4. ADPM, 5. ATPA, 6. ATPM, 7. GMPA, 8. GMPM, 9. GDPA, 10. GDPM, 11. GTPA, 12. GTPM, 13. MPT, 14. MPT-AMP. $AMPA denotes AMP at AMPBS and AMPM denotes AMP at MPTBS. For each protein: Row 1 & 2: protein-ligand interaction energy and its standard deviation (s.d.) in kcal mol-1. Row 3 & 4: average number of hydrogen bonds and its standard deviation. Row 5: protein-ligand binding energies calculated using docking method. Row 6: number of clusters obtained from molecular docking.


A previous study (Bevers et al., 2008) showed that binding of ATP is preferred

over GTP. The interaction energies obtained from MD simulations for these two

compounds reveal that ATP and GTP have similar affinities at the MPTBS, however,

ATP clearly shows better binding at the AMPBS. For AMP, the binding energy is

better with TtMogA than with EcMoaB at the AMPBS, however both proteins show

similar affinities at the MPTBS. A comparison of the binding energies at two sites

suggests that these compounds have a preference for the MPTBS compared with the

AMPBS. In addition, comparison of AtCnx1G and EcMoeA reveals that the binding is

better with AtCnx1G than with EcMoeA, supporting the previous studies. Analysis of

the conformational space accessed during the docking calculations reveals that MPT is

more specific for AtCnx1G and TtMogA, which is reflected by the lower number of

clusters (row 6 in Table 6.7). Each cluster represents a particular conformation of the

ligands, the members of each cluster are more or less similar within an r.m.s.d. of 1.0

Å. Interestingly, ATP and GTP show fewer conformations at the MPTBS than at the

AMPBS. However, diphosphate compounds show fewer conformations at the AMPBS

than at the MPTBS. Analysis of hydrogen-bond dynamics during the MD simulations

shows that compounds form more hydrogen bonds at the MPTBS than at the AMPBS

in most cases.

6.2.4.3 Protein dynamics

All of the secondary-structure elements (except for α1 and β4) of both the

TtMogA and AaMogA proteins show low root-mean-square fluctuations (r.m.s.f.s)

during the MD simulations. Helix α1 is solvent-accessible and forms the active-site

cavity. Interestingly, the residues of helix α1 interact with MoeA in MogA-MoeA

protein complex (see protein-protein complex section). On the other hand, the strand β4

is located in the trimeric interface and is involved in oligomerization. As expected, the

residues of strand β4 show a very low r.m.s.f. in a simulation containing all three

subunits of the trimer (Figure 6.20). Most of the loops show a high fluctuation. The

region 103-115 belongs to loop L9 and is of importance here. Remarkably, r.m.s.f. for

this region in the simulation containing MPT-AMP in the active site is low. Also, the

simulation containing GTP at the MPTBS shows low fluctuation for this region. On the

other hand, r.m.s.f. for loop L9 is high in the simulation containing GTP at the AMPBS


(Figure 6.20). In a similar way, the region of loop L2, which is also part of the active

site, shows low fluctuation in MPT-AMP-bound and GTP-bound simulations. In

addition, the residues in loop L6, which is involved in oligomerization and is part of the

active site, show a low r.m.s.f. in trimeric simulations and in those with MPT-AMP and

GTP at the MPTBS. To some extent, r.m.s.f. values calculated from the B factors

obtained from the crystal structures agree with those of the MD simulations (Figure

6.20). In both proteins, most of the flexible regions identified using the program

ESCET (Schneider, 2004) show a high fluctuation. Similar patterns were observed in

the EcMoaB, AtCnx1G and EcMoeA simulations (Figure 6.21).

Figure 6.20 Root-mean-square fluctuation (r.m.s.f.) of TtMogA (bottom) and AaMogA (top). The secondary-structural elements are shown and labelled. The conformationally rigid (lower) and flexible (upper) regions of the protein molecules obtained using the program ESCET (Schneider, 2004) are shown as brown lines. R.m.s.f. values for the protein only (green), GTP bound at AMPBS (violet), GTP bound at MPTBS (orange), MPT-AMP bound (blue) and the trimer (cyan) are colored differently. The average r.m.s.f. values of all of the simulations are shown in red. The average r.m.s.f. values calculated from the B factors observed in the crystal structures are shown in magenta.


6.3 CONCLUSION The crystal structures of Moco-biosynthesis protein MogA from the

thermophilic organisms T. thermophilus HB8 and A. aeolicus VF5 have been

determined at high resolution. The residues Pro47, Pro48, Lys52, Arg55, Asp59,

Glu86, Gly115, Arg120 and Ser131 (TtMogA) involved in the oligomerization of the

protein molecule have been identified based on a comparative analysis. Furthermore,

five invariant and two interfacial water molecules play a role in oligomerization.

Similarly, a further five invariant water molecules and one interfacial water molecule

are likely to play a role in anchoring the active-site residues. Our comparative analyses

reveal a possible role for the N- and C-terminal residues of MoaB and MogA proteins,

respectively, in stabilizing the substrate and/or product molecule in the active site.

Protein-protein complex prediction leads to the identification of residues (Arg3, Asp11,

Figure 6.21 Root-mean-square fluctuation (r.m.s.f.) of EcMoaB (bottom), AtCnx1G (middle) and EcMoeA (top). R.m.s.f. values for protein only (green), GTP bound at MPTBS (orange), MPT-AMP bound (blue) are shown as lines. The average r.m.s.f. values of all of the simulations are shown in red. The average r.m.s.f. values calculated from the B factors observed in the crystal structures are shown in magenta.


Glu46, Arg77, Lys106, Ser131 and Thr154) that are possibly involved in inter-protein

interactions. Further, MD simulations and molecular-docking studies of several small-

molecule ligands with the proteins support the experimental results reported in the

literature. The results show that MPT and MPT-AMP can bind more strongly to MogA

proteins than to MoaB proteins. In addition, in most of the cases, the MPTBS is

preferred over the AMPBS, except for the ATP molecule. Furthermore, the results of

the MD simulations show that the active-site loops are stabilized upon substrate and/or

product binding.

6.4 MATERIALS AND METHODS 6.4.1 CLONING, EXPRESSION AND PROTEIN PURIFICATION Thermus thermophilus MogA (TTHA0341) protein consists of 164 amino-acid

residues with a predicted molecular weight of 17.9 kDa. The gene mogA was amplified

by PCR using T. thermophilus HB8 genomic DNA as the template. The amplified

fragment was cloned under the control of the T7 promoter of the E. coli expression

vector pET-11a (Novagen, Madison, WI, USA). The expression vector was introduced

into E. coli BL21(DE3) strain (Novagen) and the recombinant strain was cultured in 6 l

LB medium supplemented with 50 μg ml-1 ampicillin in shake flasks. The cells (26 g)

were collected by centrifugation, washed with 20 ml of 20 mM Tris–HCl pH 8.0

containing 50 mM NaCl and resuspended in 70 ml of the same buffer. The cells were

then disrupted by sonication in a chilled water bath and the cell lysate was incubated at

343 K for 10 min. The sample was centrifuged at 150000g for 1 h at 277 K and

ammonium sulfate was then added to the supernatant to a final concentration of 1.5 M.

The sample was then applied onto a Resource PHE column (GE Healthcare

Biosciences) pre-equilibrated with sodium phosphate buffer pH 7.0 containing 1.5 M

ammonium sulfate and was eluted with a linear gradient of 1.5–0 M ammonium sulfate.

The eluted fractions containing the MogA were collected, desalted by fractionation on a

HiPrep 26/10 Desalting column (GE Healthcare Biosciences) pre-equilibrated with 20

mM Tris–HCl pH 8.0 and then applied onto a Resource Q column (GE Healthcare

Biosciences) pre-equilibrated with the same buffer. The flow-through fraction was

collected and applied onto a Resource S column (GE Healthcare Biosciences) pre-

equilibrated with 20 mM MES pH 6.0, which was eluted with a linear gradient of 0–0.5


M NaCl. The eluted fractions containing MogA were pooled, desalted by fractionation

on a HiPrep 26/10 Desalting column pre-equilibrated with 10 mM sodium phosphate

buffer pH 7.0 containing 0.15 M NaCl and then applied onto a hydroxyapatite CHT10

column (Bio-Rad Laboratories) pre-equilibrated with the same buffer, which was eluted

with a linear gradient of 10–250 mM sodium phosphate buffer pH 7.0. The sample

containing MogA was then loaded onto a HiLoad 16/60 Superdex 75pg column (GE

Healthcare Bioscience Corp.) pre-equilibrated with 20 mM Tris–HCl pH 8.0 containing

150 mM NaCl. The purified MogA was concentrated with a VivaSpin 20 concentrator

(10 kDa molecular-weight cutoff; Sartorius). The purified protein was homogeneous as

determined by SDS–PAGE. The protein concentration was determined by measuring

the absorbance at 280 nm (Kuramitsu et al., 1990). The yield of the purified protein

was 5.6 mg per litre of culture.

The cloning, expression and protein purification of AaMogA protein was

carried out in the following way. The mog gene (aq_061) was amplified by PCR using

Aquifex aeolicus VF5 genomic DNA as the template. The amplified fragment was

cloned under the control of the T7 promoter of the E. coli expression vector pET-21a

(Novagen). The expression vector was introduced into the E. coli BL21-CodonPlus

(DE3)-RIL strain (Stratagene) and the recombinant strain was cultured in 4.5 l LB

medium supplemented with 50 μg ml-1 ampicillin. The cells (15.4 g) were collected by

centrifugation, washed with 20 ml of buffer A (20 mM Tris-HCl, pH 8.0) containing

0.5 M NaCl, 5 mM 2-mercaptoethanol and 1 mM phenylmethanesulfonyl fluoride and

resuspended in 15 ml of the same buffer. The cells were then disrupted by sonication in

a chilled water bath and the cell lysate was incubated at 90°C for 11.5 min. The sample

was centrifuged at 15000g for 30 min and the supernatant was desalted by fractionation

on a HiPrep 26/10 desalting column (GE Healthcare Bio-Sciences Corp.) pre-

equilibrated with buffer A. The sample was then applied to a Toyopearl SuperQ-650M

(Tosoh Corp., Japan) column pre-equilibrated with the same buffer, which was eluted

with a linear gradient of 0–0.4 M NaCl. The eluted fractions containing the

recombinant MogA protein were collected, desalted by fractionation on a HiPrep 26/10

desalting column pre-equilibrated with buffer A and applied onto a Resource Q column

(GE Healthcare BioSciences) pre-equilibrated with the same buffer, which was eluted

with a linear gradient of 0–0.3 M NaCl. The eluted fractions containing the MogA


protein were pooled, desalted by fractionation on a HiPrep 26/10 desalting column pre-

equilibrated with 10 mM potassium phosphate buffer pH 7.0 and then applied to a

hydroxyapatite CHT20-I column (Bio-Rad Laboratories), which was eluted by a linear

gradient of 10–500 mM potassium phosphate buffer pH 7.0. The sample containing the

MogA protein was then loaded onto a HiLoad 16/60 Superdex 200 pg column (GE

Healthcare BioSciences) pre-equilibrated with buffer A containing 0.2 M NaCl. The

fractions containing MogA protein were concentrated to 2.7 ml with a Vivaspin 20

concentrator (5000 molecular-weight cutoff; Sartorius). The protein concentration was

24 mg ml-1 as determined by measuring the absorbance at 280 nm (Kuramitsu et al.,

1990).

6.4.2 CRYSTALLIZATION EXPERIMENTS The TtMogA protein was crystallized in the following manner. Freshly purified

protein (11 mg ml-1 in buffer 20 mM Tris–HCl pH 8.0, 150 mM NaCl) was used for

crystallization trials using the PEG/Ion kit (Hampton). Crystals were obtained using the

sitting-drop vapour-diffusion method from a drop of 1 μl protein solution and 1 μl

reservoir solution [20%(w/v) PEG 3350 and 0.2 M tripotassium citrate monohydrate

pH 8.3] at 293 K. Crystals appeared in about two weeks (Figure 6.22a). 20%(v/v) PEG

400 was used as a cryoprotectant.

For the crystallization of AaMogA, the purified protein sample (24 mg ml-1)

was screened for preliminary crystallization conditions using Wizard Cryo II.

Diffraction-quality crystals were obtained as two forms from two different conditions.

The first crystal form (P21) was obtained from 1 μl protein solution and 1 μl reservoir

solution equilibrated against 200 μl reservoir solution using the sitting-drop vapor-

diffusion method. The reservoir solution consisted of 40% (v/v) PEG600, 100 mM

CHES buffer pH 9.5. The second form of the crystal (P1) was also obtained using the

same drop ratio with a reservoir solution consisting 0.2 M ammonium acetate, 0.1 M

bis-tris pH 5.5, 25% (w/v) PEG3350. Diffraction-quality crystals of both forms

appeared within a week (Figure 6.22b). The first crystal form was mounted without any

cryoprotectant. However, the second crystal form was soaked in precipitant solution

consisting 20% (w/v) PEG3350 for a short while prior to flash-freezing and the X-ray

exposure.


6.4.3 DATA COLLECTION AND PROCESSING For both the proteins (TtMogA and AaMogA), the X-ray diffraction intensity

data were collected at 100 K on the RIKEN Structural Genomics Beamline II

(BL26B2) at SPring-8 (Hyogo, Japan) using a Jupiter 210 CCD detector (Rigaku MSC

Co., Tokyo, Japan). The crystal-to-detector distance was maintained at 150 mm. The

data were processed using the HKL suite (Otwinowski & Minor, 1997). Data-collection

and processing for all three crystals are given in Table 6.1.

Figure 6.22 Crystal images of (a) TtMogA (P21) and (b) and (c) AaMogA (P1 and P21) forms, respectively.


6.4.4 STRUCTURE SOLUTION, REFINEMENT AND VALIDATION All three crystal structures were solved by the molecular-replacement (MR)

method using the program Phaser (McCoy et al., 2007). In the case of TtMogA, the

atomic coordinates of gephyrin (PDB-id: 1JLJ; Schwarz et al., 2001) were used as the

search model. The search model has 50% amino-acid sequence identity to TtMogA.

Preliminary calculations (Matthews, 1968) suggested the presence of three monomers

in the asymmetric unit. The crystal structure solution of AaMogA was obtained using

the atomic coordinates of SoMogA (PDB-id: 2FUW; C. Chang, L. J. Bigelow and A.

Joachimiak, unpublished work) as a search model. The search model used in MR has

69% sequence identity to AaMogA. The Matthews coefficient VM (Matthews, 1968)

was calculated to be 2.19 Å3 Da-1, suggesting the presence of three monomers in the

asymmetric unit. The solution of the structure of the other form of AaMogA was

obtained using the refined model of the first form. As suggested by the Matthews

coefficient (2.28 Å3 Da-1), six monomers were searched for in the asymmetric unit

using a monomer as the search model.

In summary, a total of 5% of reflections were kept aside for the calculation of

Rfree (Brunger, 1992). The solution obtained from the MR calculation was subjected to

rigid-body refinement using CNS v.1.2 (Brunger et al., 1998). Subsequently, positional

refinement (50 cycles) was performed. The models were subjected to simulated

annealing by heating the system to 3000 K and slowly cooling to 100 K at the rate of 10

K per step. Furthermore, the models were subjected to 30 cycles of B-factor refinement.

In the next step, the amino acids in the models were replaced by the corresponding

primary structure and refined. In all three cases, R and Rfree fell to below 30% at this

stage. Subsequently, water oxygen atoms were located at 0.8σ and 2.8σ and in 2Fo-Fc

and Fo-Fc difference electron-density maps, respectively, and at a distance of 3.5 Å

from polar groups of the protein molecule or water molecules. The final refinement

statistics of all the crystal structures are given in Table 6.1. In brief, the molecular-

modeling program COOT (Emsley and Cowtan, 2004) was used to display the electron-

density maps for model fitting and adjustments. All atoms were refined with unit

occupancies. Refinement was carried out using the program CNS (Brunger et al.,

1998). Simulated-annealing omit maps were calculated to correct or check the final

protein models. The programs PROCHECK (Laskowski et al., 1993) and MolProbity


(Davis et al., 2007) were used to check and validate the quality of the final refined

models. The final refined models and structure factors were checked and validated

using the ADIT server. The atomic coordinates and structure factors of TtMogA (PDB-

id: 3MCH) and AaMogA (PDB-ids: 3MCI and 3MCJ) have been deposited in the

RCSB Protein Data Bank (Berman et al., 2000).

6.4.5 MOLECULAR DYNAMICS SIMULATION Molecular-dynamics (MD) simulations were performed using the package

GROMACS v.4.0.4 running on parallel processors (van der Spoel et al., 2005; Hess et

al., 2008). The AMBER force-field port for the GROMACS suite was used for all of

the simulations (Duan et al., 2003; Sorin and Pande, 2005). All crystallographic water

molecules were removed from the protein models before MD simulations. A cubic box

was generated using the module editconf of GROMACS with the criterion that the

minimum distance between the solute and the edge of the box was at least 0.75 nm. The

protein models were solvated with the SPC (simple point charge) water model using the

program genbox available in the GROMACS suite. All of the ligand molecules were

modeled (using the program COOT) in the active site of the respective protein

molecules based on the crystal structure of AtCnx1G (PDB-id: 1UUY; Kuper et al.,

2004) bound to adenylated molybdopterin (MPT-AMP). Hydrogen atoms were added

to the ligand molecules using the PRODRG web server (Schuettelkopf and van Aalten,

2004). The parameters derived from AMBER03 (Case et al., 2006) were used to

generate ligand topologies, which were further converted to GROMACS format using a

Perl script (amb2gmx.pl). Furthermore, the partial charges of the ligands were

optimized using the ab initio program Gaussian03 (Frisch et al., 2004). Chloride and

sodium ions were used (wherever needed) to neutralize the overall charge of the

system. Energy minimization was performed using conjugate-gradient and steepest-

descent methods with a frequency of the latter of 1 in 1000 with a maximum force

cutoff of 1 kJ mol-1 nm-1 for convergence of minimization. Subsequently, the solvent

equilibration by position-restrained dynamics for 10 ps was carried out. Simulations

utilized the NPT ensembles with Parrinello-Rahman isotropic pressure coupling (τp =

0.5 ps) to 1 bar and Nose-Hoover temperature coupling (τt = 0.1 ps) to 300 K. Long-

range electrostatics were computed using the Particle Mesh Ewald (PME; Darden et al.,


1993) method with a cutoff of 1.2 nm. A cutoff of 1.5 nm was used to compute the

long-range van der Waals interactions. Bond lengths were constrained with the LINCS

algorithm (Hess et al., 1997). The value for the dielectric constant was taken as unity

required in the case of an explicit solvent MD simulations. MD was performed for a

time period of 50 ns for all of the simulations discussed in the present study. However,

the first 5 ns of the trajectories were excluded from the analysis to allow the system to

equilibrate. The protein-ligand interaction energies were calculated using the equation

( ) ( )vdwligandproteinelecligandproteinligandprotein EEE −−− += (6.3)

where Eprotein-ligand denotes the interaction energy between protein and ligand and ‘elec’

and ‘vdw’ denote the electrostatics and van der Waals components of the energy,

respectively.

6.4.6 MOLECULAR DOCKING Molecular docking of the compounds with the protein molecules was performed

using the program AutoDock v.3.0.5 (Morris et al., 1998). The program AutoDock is

based on a Lamarckian genetic algorithm (LGA). Basically, this program determines

total interaction energies between random pairs of ligands and various selected portions

of protein to determine docking poses. The three-dimensional atomic coordinates of

TtMogA and AaMogA were taken from the final refined model, whereas in the cases of

EcMoaB (PDB-id: 1MKZ; Sanishvili et al., 2004), AtCnx1G (PDB-id: 1UUX; Kuper et

al., 2004) and EcMoeA (PDB-id: 1G8L; Xiang et al., 2001), they were downloaded

from the locally maintained anonymous FTP server at the Bioinformatics Centre,

Indian Institute of Science, Bangalore, India. All crystallographic water molecules were

removed from the protein molecule. For comparison, the partial charge for each atom

of the ligand molecules were kept the same as in the MD simulations. The solvation

parameters were added using the addsol module of AutoDock. A grid box of 60 × 60 ×

60 points in x, y, and z dimensions was used with a grid spacing of 0.375 Å. The grid

was automatically centered at the central point of the ligand molecules modeled in the

active site. The electrostatic and atomic interaction maps for all atom types of the

ligand molecules were calculated using the module autogrid of the AutoDock program.

The docking calculations were allowed to run for 250 runs using LGA for global and a

Solis and Wets algorithm for the local search with an initial population size of 50. The


values for the maximum number of generations (27000) and energy evaluations

(2500000) were taken for the simulations. Additional docking parameters like elitism

(1), mutation rate (0.02), crossover rate (0.8) and local search rate (0.06) were taken as

default values as implemented in the program. The final docked conformations of the

ligand molecules in the active site were clustered using a root-mean-square deviation

(r.m.s.d.) tolerance of 1 Å.

6.4.7 STRUCTURAL ANALYSIS The three-dimensional atomic coordinates of the homologous structures were

downloaded from a locally maintained PDB-FTP anonymous server at the

Bioinformatics Centre, Indian Institute of Science, Bangalore, India. The freely

available web server PDB Goodies (Hussain et al., 2002) was used at various stages of

the refinement and analysis. Multiple sequence alignment (MSA) was performed using

the program ClustalW v.2 (Larkin et al., 2007) and was rendered using the program

ESPript (Gouet et al., 1999). The secondary-structure elements of the protein were

assigned using the program DSSP (Kabsch and Sander, 1983). Invariant water

molecules were identified using the 3dSS web server (Sumathi et al., 2006). Protein

surface cavities were identified and measured using the program SURFNET

(Laskowski, 1995). Figures were generated using the program PyMOL (DeLano

Scientific LLC http://www.pymol.org). Electrostatic potentials were calculated using

the APBS (Baker et al., 2001) module plugged into PyMOL. Structures were

superposed using the program ALIGN (Cohen, 1997). Hydrogen bonds were calculated

using the program HBPLUS (McDonald and Thornton, 1994). A donor-hydrogen-

acceptor angle greater than or equal to 120° and donor-acceptor distance less than or

equal to 3.5 Å were used as a criteria for the identification of hydrogen bonds. The

solvent-accessible surface area of invariant water molecules was computed using the

program NACCESS (Hubbard and Thornton, 1993) with a probe radius of 1.4 Å. Water

molecules with an accessible surface area less than or equal to 2.5 Å2 were considered

to be internal/buried water molecules. The normalized temperature factor (Bi') for all

the invariant water molecules was calculated using the formula Bi' = (Bi - )/σ(B),

where Bi is the B factor of each atom, is the mean B factor and σ(B) is the

standard deviation of the B factors. Most of the MD dynamics analyses were performed


using the GROMACS tools and locally developed Perl scripts. Graphs were prepared

using Xmgrace (Paul J. Turner, Center for Coastal and Land-Margin Research Oregon

Graduate Institute of Science and Technology Beaverton, Oregon).

SUMMARY AND FUTURE PERSPECTIVES

SUMMARY AND FUTURE PERSPECTIVES 208

SUMMARY AND FUTURE PERSPECTIVES The work presented in thesis involves the structural studies on bovine

pancreatic phospholipase A2 (BPLA2) and proteins involved in molybdenum cofactor

biosynthesis. In the first part of this work, three crystal structures of the active-site

mutants of the BPLA2 enzyme have been determined. The results suggest that the

overall structures of all three mutants are similar to that of the wild-type enzyme.

However, the active-site geometry is perturbed in the case of Asp49 mutants, whereas it

is intact in the case of H48N mutant. These observations suggest that the residue Asp49

is responsible for the stabilization of the active-site calcium ion, whereas the residue

His48 is essential for the catalytic activity of the enzyme.

In addition, all the crystal structures of BPLA2 available in the PDB were

investigated to identify the invariant water molecules. This study resulted in 24 water

molecules found to be conserved among all the structures. Out of which, nine water

molecules are proposed to be involved in the folding of the enzyme. Further, several

water molecules stabilize the overall tertiary structure of the enzyme by forming ion

pairs and water bridges.

In the second part of the present work, six crystal structures including a

complex structure have been determined of two proteins MoaC and MogA involved in

the biosynthesis of molybdenum cofactor. In the case of MoaC, the results suggest that

it can bind to the molecules with terminal triphosphate group. In addition, the GTP-

bound crystal structure of MoaC revealed the residues involved in the substrate

binding. The results obtained from the structural studies on MogA reveals the residues

involved in the oligomerization of the protein molecule. Also, the role of the residues at

the N- and C-termini has been suggested.

The previous studies and the present work on three active-site mutants (H48N,

D49N and D49K) of bovine pancreatic phospholipase A2 show the role of calcium ion

in the active site. Furthermore, a possible mechanism for the low activity of H48N

mutant has been proposed in the present study. However, it will be interesting to see the

binding mode of ligand molecules in the case of H48N mutant.

To further prove the binding of the molecules with terminal triphosphate groups

to MoaC, crystallographic studies can be performed with several other analogues. In the

case of MogA, as the substrate molecules are not commercially available, the isolation

SUMMARY AND FUTURE PERSPECTIVES 209

from the species or the synthesis of these molecules would help in the co-crystallization

study of this protein.

REFERENCES

REFERENCES 211

REFERENCES Adamich, M., Roberts, M.F. and Dennis, E.A. (1979). Phospholipid activation of cobra venom phospholipase A2. 2. Characterization of the phospholipid-enzyme interaction. Biochemistry, 18, 3308-3314. Adams, P.D., Pannu, N.S., Read, R.J. and Brunger, A.T. (1997). Cross-validated maximum likelihood enhances crystallographic simulated annealing refinement. Proc. Nat. Acad. Sci. USA, 94, 5018-5023. Adler, P. and Straub, J. (1953). A water-borne caries-protective agent other than fluorine. Acta Med. Acad. Sci. Hung. 4, 221-227. Aguilar, M., Kalakoutskii, K., Cardenas, J. and Fernandez, E. (1992). Direct transfer of molybdopterin cofactor to aponitrate reductase from a carrier protein in Chlamydomonas reinhardtii. FEBS Lett. 307, 162-163. Allen, R.M., Chatterjee, R., Madden, M.S., Ludden, P.W. and Shah, V.K. (1994). Biosynthesis of the iron-molybdenum cofactor of nitrogenase. Crit. Rev. Biotechnol. 14, 225-249. Altschul, S.F., Gish, W., Miller, W., Myers, E.W. and Lipmanl, D.J. (1990). Basic Local Alignment Search Tool. J. Mol. Biol. 215, 403-410. Anderson, L.A., Palmer, T., Price, N.C., Bornemann, S., Boxer, D.H. and Pau, R.N. (1997). Characterisation of the molybdenum-responsive ModE regulatory protein and its binding to the promoter region of the modABCD (molybdenum transport) operon of Escherichia coli. Eur. J. Biochem. 246, 119-126. Andersson, T., Drakenberg, T., Forsen, S., Wieloch, T. and Lindstrom, M. (1981). Calcium binding to porcine pancreatic prophospholipase A2 studied by 43Ca NMR. FEBS Lett. 123, 115-117. Anke, M. and Seifert, M. (2007). The biological and toxicological importance of molybdenum in the environment and in the nutrition of plants, animals and man. Part 1: Molybdenum in plants. Acta Biol. Hung. 58, 311-324. Annand, R.R., Kontoyianni, M., Penzotti, J.E., Dudler, T., Lybrand, T.P. and Gelb, M.H. (1996). Active Site of Bee Venom Phospholipase A2: The Role of Histidine-34, Aspartate-64 and Tyrosine-87. Biochemistry, 35, 4591-4601. Antonyuk, S.V., Strange, R.W., Ellis, M.J., Bessho, Y., Kuramitsu, S., Shinkai, A., Yokoyama, S. and Hasnain, S.S. (2009). Structure of hypothetical Mo-cofactor biosynthesis protein B (ST2315) from Sulfolobus tokodaii. Acta Cryst. F65, 1200-1203.

REFERENCES 212

Appleyard, M.V., Sloan, J., Kanan, G.J., Heck, I.S., Kinghorn, J.R. and Unkles, S.E. (1998). The Aspergillus nidulans cnxF gene and its involvement in molybdopterin biosynthesis. Molecular characterization and analysis of in vivo generated mutants. J. Biol. Chem. 273, 14869-14876. Arni, R.K., Ward, R.J., Gutierrez, J.M. and Tulinsky, A. (1995). Structure of a calcium-independent phospholipase-like myotoxic protein from Bothrops asper venom. Acta Cryst. D51, 311-317. Ataya, F.S., Witte, C.P., Galvan, A., Igeno, M.I. and Fernandez, E. (2003). Mcp1 encodes the molybdenum cofactor carrier protein in Chlamydomonas reinhardtii and participates in protection, binding, and storage functions of the cofactor. J. Biol. Chem. 278, 10885-10890. Bacher, A., Eberhardt, S., Eisenreich, W., Fischer, M., Herz, S., Illarionov, B., Kis, K. and Richter, G. (2001). Biosynthesis of riboflavin. Vitam. Horm. 61, 1-49. Bader, G., Gomez-Ortiz, M., Haussmann, C., Bacher, A., Huber, R. and Fischer, M. (2004). Structure of the molybdenum-cofactor biosynthesis protein MoaB of Escherichia coli. Acta Cryst. D60, 1068-1075. Baker, E.N. and Hubbard, R.E. (1984). Hydrogen bonding in globular proteins. Prog. Biophys. Mol. Biol. 44, 97-179. Baker, N.A., Sept, D., Joseph, S., Holst, M.J. and McCammon, J.A. (2001). Electrostatics of nanosystems: application to microtubules and the ribosome. Proc. Natl. Acad. Sci. USA, 98, 10037-10041. Balamurugan, B., Md. Roshan, M.N.A., Shaahul Hameed, B., Sumathi, K., Senthilkumar, R., Udayakumar, A., Venkatesh Babu, K.H., Kalaivani, M., Sowmiya, G., Sivasankari, P., Saravanan, S., Ranjani, C.V., Gopalakrishnan, K., Selvakumar, K.N., Jaikumar, M., Brindha, T., Michael, D. and Sekar, K. (2007). PSAP: protein structure analysis package. J. Appl. Cryst. 40, 773-777. Barceloux, D.G. and Barceloux, D. (1999). Molybdenum. J. Toxicol. Clin. Toxicol. 37, 231-237. Bazan, N.G., Colangelo, V. and Lukiw, W.J. (2002). Prostaglandins and other lipid mediators in Alzheimer's disease. Prostaglandins Other Lipid Mediat. 68-69, 197-210. Begley, T.P., Downs, D.M., Ealick, S.E., McLafferty, F.W., van Loon, A.P., Taylor. S., Campobasso, N., Chiu. H.J., Kinsland, C., Reddick, J.J. and Xi, J. (1999b). Thiamin biosynthesis in prokaryotes. Arch. Microbiol. 171, 293-300. Begley, T.P., Xi, J., Kinsland, C., Taylor, S. and McLafferty, F. (1999a). The enzymology of sulfur activation during thiamin and biotin biosynthesis. Curr. Opin. Chem. Biol. 3, 623-629.

REFERENCES 213

Bekkers, A.C., Franken, P.A., Toxopeus, E., Verheij, H.M. and de Haas, G.H. (1991). The importance of glycine-30 for enzymatic activity of phospholipase A2. Biochim. Biophys. Acta. 1076, 374-378. Bell, J.D. and Biltonen, R.L. (1989a). Thermodynamic and kinetic studies of the interaction of vesicular dipalmitoylphosphatidylcholine with Agkistrodon piscivorus piscivorus phospholipase A2. J. Biol. Chem. 264, 225-230. Bell, J.D. and Biltonen, R.L. (1989b). The temporal sequence of events in the activation of phospholipase A2 by lipid vesicles. Studies with the monomeric enzyme from Agkistrodon piscivorus piscivorus. J. Biol. Chem. 264, 12194-12200. Berendsen, H.J.C., Postma, J.P.M., van Gunsteren, W.F., DiNola, A. and Haak, J.R. (1984). Molecular dynamics with coupling to an external bath. J. Chem. Phys. 81, 3684-3690. Berendsen, H.J.C., Postma, J.P.M., van Gunsteren, W.F. and Hermans, J. (1981). In Intermolecular Forces, edited by Pullman (Reidel, Dordrecht), 331. Berendsen, H.J.C., van der Spoel, D. and van Drunen, R. (1995). A message-passing parallel molecular dynamics implementation. Comp. Phys. Comm. 91, 43-56. Berg, O.G., Yu, B.Z., Rogers, J. and Jain, M.K. (1991). Interfacial catalysis by phospholipase A2: determination of the interfacial kinetic rate constants. Biochemistry, 30, 7283-7297. Berman, H.M., Westbrook, J., Feng, Z., Gilliland, G., Bhat, T.N., Weissig, H., Shindyalov, I.N. and Bourne, P.E. (2000). The Protein Data Bank. Nucl. Acids Res. 28, 235-242. Bevers, L.E., Hagedoorn, P.L. and Hagen, W.R. (2009). The bioinorganic chemistry of tungsten. Coord. Chem. Rev. 253, 269-290. Bevers, L.E., Hagedoorn, P.L., Krijger, G.C. and Hagen, W.R. (2006). Tungsten transport protein A (WtpA) in Pyrococcus furiosus: the first member of a new class of tungstate and molybdate transporters. J. Bacteriol. 188, 6498-6505. Bevers, L.E., Hagedoorn, P.L., Santamaria-Araujo, J.A., Magalon, A., Hagen, W.R. and Schwarz, G. (2008). Function of MoaB proteins in the biosynthesis of the molybdenum and tungsten cofactors. Biochemistry, 47, 949-956. Bhat, T.N. (1988). Calculation of an OMIT map. J. Appl. Cryst. 21, 279-281. Biswal, B.K., Sukumar, N. and Vijayan, M. (2000). Hydration, mobility and accessibility of lysozyme: structures of a pH 6.5 orthorhombic form and its low-humidity variant and a comparative study involving 20 crystallographically independent molecules). Acta Cryst. D56, 1110-1119.

REFERENCES 214

Blasco, F., Dos Santos, J.P., Magalon, A., Frixon, C., Guigliarelli, B., Santini, C.L. and Giordano, G. (1998). NarJ is a specific chaperone required for molybdenum cofactor assembly in nitrate reductase A of Escherichia coli. Mol. Microbiol. 28, 435-447. Bomalaski, J.S. and Clark, M.A. (1990). Activation of phospholipase A2 in rheumatoid arthritis. Adv. Exp. Med. Biol. 279, 231-238. Boyanovsky, B.B. and Webb, N.R. (2009). Biology of Secretory Phospholipase A2. Cardiovasc. Drugs Ther. 23, 61–72. Bradford, J.R. and Westhead, D.R. (2005). Improved prediction of protein-protein binding sites using a support vector machines approach. Bioinformatics, 21, 1487-1494. Brunger, A.T. (1992). Free R value: a novel statistical quantity for assessing the accuracy of crystal structures. Nature (London), 355, 472-475. Brunger, A.T., Adams, P.D. and Rice, L.M. (1997). New applications of simulated annealing in X-ray crystallography and solution NMR. Structure, 5, 325-336. Brunger, A.T., Adams, P.D., Clore, G.M., DeLano, W.L., Gros, P., Grosse-Kunstleve, R.W., Jiang, J.-S., Kuszewski, J., Nilges, M., Pannu, N.S., Read, R.J., Rice, L.M., Simonson, T. and Warren, G.L. (1998). Crystallography and NMR System: A New Software Suite for Macromolecular Structure Determination. Acta Cryst. D54, 905-921. Brunger, A.T., Krukowski, A. and Erickson, J.W. (1990). Slow-cooling protocols for crystallographic refinement by simulated annealing. Acta Cryst. A46, 585-593. Brunger, A.T., Kuriyan, J. and Karplus, M. (1987). Crystallographic R Factor Refinement by Molecular Dynamics. Science, 235, 458-460. Brunie, S., Bolin, J., Gewirth, D. and Sigler, P.B. (1985). The refined crystal structure of dimeric phospholipase A2 at 2.5 Å. Access to a shielded catalytic center. J. Biol. Chem. 260, 9742-9749. Brzozowski, A.M., Derewenda, Z.S., Dodson, E.J., Dodson, G.G. and Turkenburg, J.P. (1992). Structure and molecular model refinement of Rhizomucor miehei triacyglyceride lipase: a case study of the use of simulated annealing in partial model refinement. Acta Cryst. B48, 307-319. Buchanan, G., Kuper, J., Mendel, R.R., Schwarz, G. and Palmer, T. (2001). Characterisation of the mob locus of Rhodobacter sphaeroides WS8: mobA is the only gene required for molybdopterin guanine dinucleotide synthesis. Arch. Microbiol. 176, 62-68. Burgess, B.K. and Lowe, D.J. (1996). Mechanism of molybdenum nitrogenase. Chem. Rev. 96, 2983-3012.

REFERENCES 215

Burke, J.E. and Dennis, E.A. (2009). Phospholipase A2 Biochemistry. Cardiovasc. Drugs Ther. 23, 49–59. Campbell, W.H. (2001). Structure and function of eukaryotic NAD(P)H:nitrate reductase. Cell Mol. Life Sci. 58, 194-204. Case, D.A., Darden, T.A., Cheatham III, T.E., Simmerling, C.L., Wang, J., Duke, R.E., Luo, R., Merz, K.M., Pearlman, D.A., Crowley, M., Walker, R.C., Zhang, W., Wang, B., Hayik, S., Roitberg, A., Seabra, G., Wong, K.F., Paesani, F., Wu, X., Brozell, S., Tsui, V., Gohlke, H., Yang, L., Tan, C., Mongan, J., Hornak, V., Cui, G., Beroza, P., Mathews, D.H., Schafmeister, C., Ross, W.S. and Kollman, P.A. (2006). AMBER 9, University of California, San Francisco. Chan, M.K., Mukund, S., Kletzin, A., Adams, M.W. and Rees, D.C. (1995). Structure of a hyperthermophilic tungstopterin enzyme, aldehyde ferredoxin oxidoreductase. Science, 267, 1463–1469. Chang, C.C., and Lee, C.Y. (1963). Isolation of neurotoxins from the venom of Bungarus multicinctus and their modes of neuromuscular blocking action. Arch. Int. Pharmacodyn. Ther. 144, 241-257. Chaplin, M.F. (2006). Do we underestimate the importance of water in cell biology? Nature Rev. Mol. Cell Biol. 7, 861-866. Chen, J., Engle, S.J., Seilhamer, J.J. and Tischfield, J.A. (1994a). Cloning and recombinant expression of a novel human low molecular weight Ca2+-dependent phospholipase A2. J. Biol. Chem. 269, 2365-2368. Chen, J., Engle, S.J., Seilhamer, J.J. and Tischfield, J.A. (1994b). Cloning, expression and partial characterization of a novel rat phospholipase A2. Biochim. Biophys. Acta. 1215, 115-120. Chen, V.B., Arendall, W.B., Headd, J.J., Keedy, D.A., Immormino, R.M., Kapral, G.J., Murray, L.W., Richardson, J.S. and Richardson, D.C. (2010). MolProbity: all-atom structure validation for macromolecular crystallography. Acta Cryst. D66, 12-21. Cheung, M.S., Garcia, A.E. and Onuchic, J.N. (2002). Protein folding mediated by solvation: water expulsion and formation of the hydrophobic core occur after the structural collapse. Proc. Natl. Acad. Sci. USA, 99, 685-690. Cho, W., Tomasselli, A.G., Heinrikson, R.L. and Kezdy, F.J. (1988). The chemical basis for interfacial activation of monomeric phospholipases A2. Autocatalytic derivatization of the enzyme by acyl transfer from substrate. J. Biol. Chem. 263, 11237-11241. Cohen, G.E. (1997). ALIGN: a program to superimpose protein coordinates, accounting for insertions and deletions. J. Appl. Cryst. 30, 1160-1161.

REFERENCES 216

Collaborative Computational Project, Number 4 (1994). Acta Cryst. D50, 760-763. Comeau, S.R., Kozakov, D., Brenke, R., Shen, Y., Beglov, D. and Vajda, S. (2007). ClusPro: Performance in CAPRI rounds 6–11 and the new server. Proteins, 69, 781-785. Cornell, W.D., Cieplak, P., Bayly, C.I., Gould, I.R., Merz, Jr. K.M., Ferguson, D.M., Spellmeyer, D.C., Fox, T., Caldwell, J.W. and Kollman, P.A. (1995). A Second Generation Force Field for the Simulation of Proteins, Nucleic Acids, and Organic Molecules. J. Am. Chem. Soc. 117, 5179−5197. Coughlan, M.P. (1983). The role of molybdenum in human biology. J. Inherit. Metab. Dis. 6, 70-77. Cove, D.J. and Pateman, J.A. (1963). Independently segregating genetic loci concerned with nitrate reductase activity in Aspergillus nidulans. Nature, 198, 262-263. Crowther, R.A. and Blow, D.M. (1967). A method of positioning a known molecule in an unknown crystal structure. Acta Cryst. 23, 544-548. Cruickshank, D.W. (1999). Remarks about protein structure precision. Acta Cryst. D55, 583-601. Cummings, B.S. (2007). Phospholipase A2 as targets for anti-cancer drugs. Biochem. Pharmacol. 74, 949-959. Cupillard, L., Koumanov, K., Mattei, M.G., Lazdunski, M. and Lambeau, G. (1997). Cloning, chromosomal mapping, and expression of a novel human secretory phospholipase A2. J. Biol. Chem. 272, 15745-15752. Curzon, M.E.J., Kubota, J. and Bibby, B.G. (1971). Environmental Effects of Molybdenum on Caries. J. Dent. Res. 50, 74-77. Dalm, D., Palm, G.J., Aleksandrov, A., Simonson, T., Hinrichs, W. (2010). Nonantibiotic properties of tetracyclines: structural basis for inhibition of secretory phospholipase A2. J. Mol. Biol. 398, 83-96. Daniels, J.N., Wuebbens, M.M., Rajagopalan, K.V. and Schindelin, H. (2008). Crystal structure of a molybdopterin synthase-precursor Z complex: insight into its sulfur transfer mechanism and its role in molybdenum cofactor deficiency. Biochemistry, 47, 615-626. Darden, T., York, D. and Pedersen, L. (1993). Particle mesh Ewald: An N·log(N) method for Ewald sums in large systems. J. Chem. Phys. 98, 10089-10092. Dauter, Z. (1999). Data-collection strategies. Acta Cryst. D55, 1703-1717.

REFERENCES 217

Davidson, F.F. and Dennis, E.A. (1990). Evolutionary relationships and implications for the regulation of phospholipase A2 from snake venom to human secreted forms. J. Mol. Evol. 31, 228-238. Davis, I.W., Leaver-Fay, A., Chen, V.B., Block, J.N., Kapral, G.J., Wang, X., Murray, L.W., Arendall III, W.B., Snoeyink, J., Richardson, J.S. and Richardson, D.C. (2007). MolProbity: all-atom contacts and structure validation for proteins and nucleic acids. Nucl. Acids Res. 35, W375-W383. Davis, I.W., Murray, L.W., Richardson, J.S. and Richardson, D.C. (2004). MOLPROBITY: structure validation and all-atom contact analysis for nucleic acids and their complexes. Nucl.c Acids Res. 32, W615-W619. Dennis, E.A. (1994). Diversity of group types, regulation, and function of phospholipase A2. J. Biol. Chem. 269, 13057-13060. Dennis, E.A. Rhee, S.G., Billah, M.M. and Hannun, Y.A. (1991). Role of phospholipase in generating lipid second messengers in signal transduction. FASEB J. 5, 2068-2077. Dijkstra, B.W., Drenth, J. and Kalk, K.H. (1981a). Active site and catalytic mechanism of phospholipase A2. Nature, 289, 604-606. Dijkstra, B.W., Drenth, J., Kalk, K.H. and Vandermaelen, P.J. (1978). Three-dimensional structure and disulfide bond connections in bovine pancreatic phospholipase A2. J. Mol. Biol. 124, 53-60. Dijkstra, B.W., Kalk, K.H., Drenth, J., de Haas, G.H., Egmond, M.R. and Slotboom, A.J. (1984). Role of the N-terminus in the interaction of pancreatic phospholipase A2 with aggregated substrates. Properties and crystal structure of transaminated phospholipase A2. Biochemistry, 23, 2759-2766. Dijkstra, B.W., Kalk, K.H., Hol, W.G. and Drenth, J. (1981b). Structure of bovine pancreatic phospholipase A2 at 1.7 Å resolution. J. Mol. Biol. 147, 97-123. Dobbek, H., Gremer, L., Kiefersauer, R., Huber, R. and Meyer, O. (2002). Catalysis at a dinuclear [CuSMo(==O)OH] cluster in a CO dehydrogenase resolved at 1.1 Å resolution. Proc. Natl. Acad. Sci. USA, 99, 15971-15976. Dodson, G.G., Lane, D.P. and Verma, C.S. (2008). Molecular simulations of protein dynamics: new windows on mechanisms in biology. EMBO Rep. 9, 144-150. Dos Santos, P.C., Dean, D.R., Hu, Y. and Ribbe, M.W. (2004). Formation and insertion of the nitrogenase iron-molybdenum cofactor. Chem. Rev. 104, 1159-1173. Dua, R., Wu, S.K. and Cho, W. (1995). A structure-function study of bovine pancreatic phospholipase A2 using polymerized mixed liposomes. J. Biol. Chem. 270, 263-268.

REFERENCES 218

Duan, Y., Wu, C., Chowdhury, S., Lee, M. C., Xiong, G., Zhang, W., Yang, R., Cieplak, P., Luo, R., Lee, T., Caldwell, J., Wang, J. and Kollman, P. (2003). A point-charge force field for molecular mechanics simulations of proteins based on condensed-phase quantum mechanical calculations. J. Comp. Chem. 24, 1999-2012. Dupureur, C.M., Yu, B.Z., Jain, M.K., Noel, J.P., Deng, T., Li, Y., Byeon, I.J. and Tsai, M.-D. (1992a). Phospholipase A2 engineering. Structural and functional roles of highly conserved active site residues tyrosine-52 and tyrosine-73. Biochemistry, 31, 6402-6413. Dupureur, C.M., Yu, B.Z., Mamone, J.A., Jain, M.K. and Tsai, M.-D. (1992b). Phospholipase A2 engineering. The structural and functional roles of aromaticity and hydrophobicity in the conserved phenylalanine-22 and phenylalanine-106 aromatic sandwich. Biochemistry, 31, 10576-10583. Duran, M., Beemer, F.A., van de Heiden, C., Korteland, J., de Bree, P.K., Brink, M., Wadman, S.K. and Lombeck, I. (1978). Combined deficiency of xanthine oxidase and sulphite oxidase: a defect of molybdenum metabolism or transport? J. Inherit. Metab. Dis. 1, 175-178. Eady, R.R. (1995). Vanadium nitrogenases of Azotobacter. Met. Ions Biol. Syst. 31, 363-405. Eaves, D.J., Palmer, T. and Boxer, D.H. (1997). The product of the molybdenum cofactor gene mobB of Escherichia coli is a GTP-binding protein. Eur. J. Biochem. 246, 690-697. Edwards, S.H., Thompson, D., Baker, S.F., Wood, S.P. and Wilton, D.C. (2002). The crystal structure of the H48Q active site mutant of human group IIA secreted phospholipase A2 at 1.5 Å resolution provides an insight into the catalytic mechanism. Biochemistry, 41, 15468-15476. Eisenmesser, E.Z., Millet, O., Labeikovsky, W., Korzhnev, D.M., Wolf-Watz, M., Bosco, D.A., Skalicky, J.J., Kay, L.E. and Kern, D. (2005). Intrinsic dynamics of an enzyme underlies catalysis. Nature (London), 438, 117-121. Emsley, P. and Cowtan, K. (2004). Coot: model-building tools for molecular graphics. Acta Cryst. D60, 2126-2132. Engh, R.A. and Huber, R. (1991). Accurate bond and angle parameters for X-ray protein structure refinement. Acta Cryst. A47, 392-400. Essmann, U., Perera, L., Berkowitz, M.L. Darden, T., Lee, H. and Pedersen, L.G. (1995). A smooth particle mesh Ewald method. J. Chem. Phys. 103, 8577-8593. Ewald, P. (1921). Die Berechnung optischer und elektrostatischer Gitterpotentiale. Ann. Phys. b, 253–287.

REFERENCES 219

Fairbairn, D. (1945). The phospholipase of the venom of the cottonmouth moccasin (Agkistrodon piscivorus L). J. Biol. Chem. 157, 633-645. Felsenstein, J. (2005). PHYLIP (Phylogeny Inference Package). Department of Genome Sciences, University of Washington, Seattle. Ferreira, J.P., Sasisekharan, R., Louie, O. and Langer, R. (1994). A study on the functional subunits of phospholipases A2 by enzyme immobilization. Biochem. J. 303, 527-530. Fischer, K., Llamas, A., Tejada-Jimenez, M., Schrader, N., Kuper, J., Ataya, F.S., Galvan, A., Mendel, R.R., Fernandez, E. and Schwarz, G. (2006). Function and structure of the molybdenum cofactor carrier protein from Chlamydomonas reinhardtii. J. Biol. Chem. 281, 30186-30194. Fisher, R.A. (1922). On the Mathematical Foundations of Theoretical Statistics. Phil. Trans. R. Soc. Lond. A222, 309-368. Fleer, E.A., Verheij, H.M. and de Haas, G.H. (1981). Modification of carboxylate groups in bovine pancreatic phospholipase A2. Identification of aspartate-49 as Ca2+-binding ligand. Eur. J. Biochem. 113, 283-288. Fohlman, J., Eaker, D., Karlsoon, E. and Thesleff, S. (1976). Taipoxin, an extremely potent presynaptic neurotoxin from the venom of the australian snake taipan (Oxyuranus s. scutellatus). Isolation, characterization, quaternary structure and pharmacological properties. Eur. J. Biochem. 68, 457-469. Fokine, A. and Urzhumstev, A. (2002). Flat bulk-solvent model: obtaining optimal parameters. Acta Cryst. D58, 1387-1392. Franks, F. (2002). Protein stability: the value of 'old literature'. Biophys. Chem. 96, 117-127. Fremont, D.H., Anderson, D.H., Wilson, I.A., Dennis, E.A. and Xuong, N.H. (1993). Crystal structure of phospholipase A2 from Indian cobra reveals a trimeric association. Proc. Natl. Acad. Sci. USA, 90, 342-346. French, S. and Wilson, K. (1978). On the treatment of negative intensity observations. Acta Cryst. A34, 517-525. Frey, P.A. and Magnusson, O.T. (2003). S-Adenosylmethionine: a wolf in sheep's clothing, or a rich man's adenosylcobalamin? Chem. Rev. 103, 2129-2148. Friesen, J.A. and Rodwell, V.W. (2004). The 3-hydroxy-3-methylglutaryl coenzyme-A (HMG-CoA) reductases. Genome Biol. 5, 248-254.

REFERENCES 220

Frisch, M.J., Trucks, G.W., Schlegel, H.B., Scuseria, G.E., Robb, M.A., Cheeseman, J. R., Montgomery, Jr. J.A., Vreven, T., Kudin, K.N., Burant, J.C., Millam, J.M., Iyengar, S.S., Tomasi, J., Barone, V., Mennucci, B., Cossi, M., Scalmani, G., Rega, N., Petersson, G.A., Nakatsuji, H., Hada, M., Ehara, M., Toyota, K., Fukuda, R., Hasegawa, J., Ishida, M., Nakajima, T., Honda, Y., Kitao, O., Nakai, H., Klene, M., Li, X., Knox, J.E., Hratchian, H.P., Cross, J.B., Bakken, V., Adamo, C., Jaramillo, J., Gomperts, R., Stratmann, R.E., Yazyev, O., Austin, A.J., Cammi, R., Pomelli, C., Ochterski, J.W., Ayala, P.Y., Morokuma, K., Voth, G.A., Salvador, P., Dannenberg, J.J., Zakrzewski, V.G., Dapprich, S., Daniels, A.D., Strain, M.C., Farkas, O., Malick, D.K., Rabuck, A.D., Raghavachari, K., Foresman, J.B., Ortiz, J.V., Cui, Q., Baboul, A.G., Clifford, S., Cioslowski, J., Stefanov, B.B., Liu, G., Liashenko, A., Piskorz, P., Komaromi, I., Martin, R.L., Fox, D.J., Keith, T., Al-Laham, M.A., Peng, C.Y., Nanayakkara, A., Challacombe, M., Gill, P.M.W., Johnson, B., Chen, W., Wong, M.W., Gonzalez, C. and Pople, J.A. (2004). Gaussian Inc., Wallingford CT. Funk, C.D. (2001). Prostaglandins and leukotrienes: advances in eicosanoid biology. Science, 294, 1871-1875. Gambero, A., Landucci, E.C., Toyama, M.H., Marangoni, S., Giglio, J.R., Nader, H.B., Dietrich, C.P., De Nucci, G. and Antunes, E. (2002). Human neutrophil migration in vitro induced by secretory phospholipases A2: a role for cell surface glycosaminoglycans. Biochem. Pharmacol. 63, 65-72. Gasteiger, E., Hoogland, C., Gattiker, A., Duvaud, S., Wilkins, M.R., Appel, R.D. and Bairoch, A. (2005). Protein Identification and Analysis Tools on the ExPASy Server; (In) John M. Walker (ed): The Proteomics Protocols Handbook. Humana Press, 571-607. Gelb, M.H., Jain, M.K., Hanel, A.M. and Berg, O.G. (1995). Interfacial enzymology of glycerolipid hydrolases: lessons from secreted phospholipases A2. Annu. Rev. Biochem. 64, 653-688. Ghannoum, M.A. (2000). Potential role of phospholipases in virulence and fungal pathogenesis. Clin Microbiol Rev. 13, 122-143. Giri, S., Khan, M., Rattan, R., Singh, I. and Singh, A.K. (2006). Krabbe disease: psychosine-mediated activation of phospholipase A2 in oligodendrocyte cell death. J. Lipid Res. 47, 1478-1492. Glaser, J.H. and DeMoss, J.A. (1971). Phenotypic restoration by molybdate of nitrate reductase activity in chlD mutants of Escherichia coli. J. Bacteriol. 108, 854-860. Gomez, F., Vandermeers, A., Vandermeers-Piret, M.C., Herzog, R., Rathe, J., Stievenart, M., Winand, J. and Christophe, J. (1989). Purification and characterization of five variants of phospholipase A2 and complete primary structure of the main phospholipase A2 variant in Heloderma suspectum (Gila monster) venom. Eur. J. Biochem. 186, 23-33.

REFERENCES 221

Gouet, P., Robert, X. and Courcelle, E. (2003). ESPript/ENDscript: extracting and rendering sequence and 3D information from atomic structures of proteins. Nucl. Acids Res. 31, 3320-3323. Grunden, A.M., Ray, R.M., Rosentel, J.K., Healy, F.G. and Shanmugam, K.T. (1996). Repression of the Escherichia coli modABCD (molybdate transport) operon by ModE. J. Bacteriol. 178, 735-744. Grunden, A.M. and Shanmugam, K.T. (1997). Molybdate transport and regulation in bacteria. Arch. Microbiol. 168, 345-354. Guy, J.E., Stahl, U. and Lindqvist, Y. (2009). Crystal structure of a class XIB phospholipase A2 (PLA2): rice (Oryza sativa) isoform-2 pla2 and an octanoate complex. J. Biol. Chem. 284, 19371-19379. Guvench, O. and MacKerell, A.D. Jr. (2008). Comparison of protein force fields for molecular dynamics simulations. Methods Mol. Biol. 443, 63-88. Halgren, T.A. (1996a). Merck molecular force field. I. Basis, form, scope, parameterization, and performance of MMFF94. J. Comp. Chem. 17, 490-519. Halgren, T.A. (1996b). Merck Molecular Force Field. II. MMFF94 van der Waals and Electrostatic Parameters for Intermolecular Interactions. J. Comp. Chem. 17, 520-552. Halgren, T.A. (1996c). Merck Molecular Force Field. III. Molecular Geometries and Vibrational Frequencies. J. Comp. Chem. 17, 553-586. Hall, D.R., Gourley, D.G., Leonard, G.A., Duke, E.M., Anderson, L.A., Boxer, D.H. and Hunter, W.N. (1999). The high-resolution crystal structure of the molybdate-dependent transcriptional regulator (ModE) from Escherichia coli: a novel combination of domain folds. EMBO J. 18, 1435-1446. Halle, B. (2004). Protein hydration dynamics in solution: a critical survey. Philos. Trans. R. Soc. Lond. B Biol. Sci. 359, 1207–1224. Han, S.K., Yoon, E.T., Scott, D.L., Sigler, P.B. and Cho, W. (1997). Structural aspects of interfacial adsorption. A crystallographic and site-directed mutagenesis study of the phospholipase A2 from the venom of Agkistrodon piscivorus piscivorus. J. Biol. Chem. 272, 3573-3582. Hanzelmann, P., Hernandez, H.L., Menzel, C., Garcia-Serres, R., Huynh, B.H., Johnson, M.K., Mendel, R.R. and Schindelin, H. (2004). Characterization of MOCS1A, an oxygen-sensitive iron-sulfur protein involved in human molybdenum cofactor biosynthesis. J. Biol. Chem. 279, 34721-34732. Hanzelmann, P. and Schindelin, H. (2004). Crystal structure of the S-adenosylmethionine-dependent enzyme MoaA and its implications for molybdenum cofactor deficiency in humans. Proc. Natl. Acad. Sci. USA, 101, 12870-12875.

REFERENCES 222

Hanzelmann, P, and Schindelin, H. (2006). Binding of 5'-GTP to the C-terminal FeS cluster of the radical S-adenosylmethionine enzyme MoaA provides insights into its mechanism. Proc. Natl. Acad. Sci. USA. 103, 6829-6834. Hanzelmann, P., Schwarz, G. and Mendel, R.R. (2002). Functionality of alternative splice forms of the first enzymes involved in human molybdenum cofactor biosynthesis. J. Biol. Chem. 277, 18303-18312. Harrison R. (2002). Structure and function of xanthine oxidoreductase: where are we now? Free Radic. Biol. Med. 33, 774-797. Hasona, A., Ray, R.M. and Shanmugam, K.T. (1998). Physiological and genetic analyses leading to identification of a biochemical role for the moeA (molybdate metabolism) gene product in Escherichia coli. J. Bacteriol. 180, 1466-1472. Hattori, M., Adachi, H., Tsujimoto, M., Arai, H. and Inoue, K. (1994). The catalytic subunit of bovine brain platelet-activating factor acetylhydrolase is a novel type of serine esterase. J. Biol. Chem. 269, 23150-23155. Hayward, S., Kitao, A., Hirata, F. and Go, N. (1993). Effect of solvent on collective motions in globular protein. J. Mol. Biol. 234, 1207-1217. Hazlett, T.L. and Dennis, E.A. (1985). Aggregation studies on fluorescein-coupled cobra venom phospholipase A2. Biochemistry, 24, 6152-6158. Head-Gordon, T. and Brooks, C.L. (1991). Virtual rigid body dynamics. Biopolymers, 235, 1021-1031. Heinrikson, R.L. (1991). Dissection and sequence analysis of phospholipases A2. Methods Enzymol. 197, 201-214. Hershko, A. and Ciechanover, A. (1998). The ubiquitin system. Annu. Rev. Biochem. 67, 425-479. Hess, B., Bekker, H., Berendsen, H.J.C. and Fraaije, J.G.E.M. (1997). LINCS: a linear constraint solver for molecular simulations. J. Comput. Chem. 18, 1463-1472. Hess, B., Kutzner, C., van der Spoel, D. and Lindahl, E. (2008). GROMACS 4: Algorithms for Highly Efficient, Load-Balanced, and Scalable Molecular Simulation. J. Chem. Theo. Comp. 4, 435-447. Hille, R. (1996). The Mononuclear Molybdenum Enzymes. Chem Rev. 96, 2757-2816. Hille, R. (2002a). Molybdenum and tungsten in biology. Trends Biochem. Sci. 27, 360-367. Hille, R. (2002b). Molybdenum enzymes containing the pyranopterin cofactor: an overview. Met. Ions Biol. Syst. 39, 187–226.

REFERENCES 223

Hiraoka, M., Abe, A. and Shayman, J.A. (2002). Cloning and characterization of a lysosomal phospholipase A2, 1-O-acylceramide synthase. J. Biol. Chem. 277, 10090-10099. Ho, C.L. and Lee, C.Y. (1981). Presynaptic actions of Mojave toxin isolated from Mojave rattlesnake (Crotalus scutulatus) venom. Toxicon, 19, 889-892. Hockney, R.W. and Eastwood, J.W. (1981). Computer Simulation Using Particles. New York: McGraw-Hill. Hockney, R.W., Goel, S.P. and Eastwood, J.W. (1974). Quite high resolution computer models of a plasma. J. Chem. Phys. 14, 148-158. Hoff, T., Schnorr, K.M., Meyer, C. and Caboche, M. (1995). Isolation of two Arabidopsis cDNAs involved in early steps of molybdenum cofactor biosynthesis by functional complementation of Escherichia coli mutants. J. Biol. Chem. 270, 6100-6107. Holleman, A.F. and Wiberg, E. (2001). Inorganic chemistry. Academic Press. 1384. Holm, L. and Sander, C. (1995). DALI: a network tool for protein structure comparison. Trends Biochem. Sci. 20, 478-480. Hollenstein, K., Frei, D.C. and Locher, K.P. (2007). Structure of an ABC transporter in complex with its binding protein. Nature, 446, 213-216. Holm, L. and Sander, C. (1995). Dali: a network tool for protein structure comparison. Trends Biochem. Sci. 20, 478-480. Hoover, W.G. (1985). Canonical dynamics: Equilibrium phase-space distributions. Phys. Rev. A31, 1695-1697. Hough, E., Hansen, L.K., Birknes, B., Jynge, K., Hansen, S., Hordvik, A., Little, C., Dodson, E. and Derewenda, Z. (1989). High-resolution (1.5 Å) crystal structure of phospholipase C from Bacillus cereus. Nature, 338, 357-360. Huang, B., Yu, B.Z., Rogers, J., Byeon, I.J., Sekar, K., Chen, X., Sundaralingam, M., Tsai, M.-D. and Jain, M.K. (1996). Phospholipase A2 engineering. Deletion of the C-terminus segment changes substrate specificity and uncouples calcium and substrate binding at the zwitterionic interface. Biochemistry, 35, 12164-12174. Hubbard, S.J. and Thornton, J.M. (1993). NACCESS Computer Program. Department of Biochemistry and Molecular Biology, University College, London. Hunt, J.F., Vellieux, F.M.D. and Deisenhofer, J. (1997). A Connected Set Algorithm for the Identification of Spatially Contiguous Regions in Crystallographic Envelopes. Acta Cryst. D53, 434-437.

REFERENCES 224

Hussain, A.S.Z., Shanthi, V., Sheik, S.S., Jeyakanthan, J., Selvarani, P. and Sekar, K. (2002). PDB Goodies - a web-based GUI to manipulate the Protein Data Bank file. Acta Cryst. D58, 1385-1386. Ichida, K., Aydin, H.I., Hosoyamada, M., Kalkanoglu, H.S., Dursun, A., Ohno, I., Coskun, T., Tokatli, A., Shibasaki, T. and Hosoya, T. (2006). A Turkish case with molybdenum cofactor deficiency. Nucleosides Nucleotides Nucleic Acids, 25, 1087-1091. Ichida, K., Matsumura, T., Sakuma, R., Hosoya, T. and Nishino, T. (2001). Mutation of human molybdenum cofactor sulfurase gene is responsible for classical xanthinuria type II. Biochem. Biophys. Res. Commun. 282, 1194-1200. Iobbi-Nivol, C., Palmer, T., Whitty, P.W., McNairn, E. and Boxer, D.H. (1995). The mob locus of Escherichia coli K12 required for molybdenum cofactor biosynthesis is expressed at very low levels. Microbiology, 141, 1663-1671. Jabeen, T., Sharma, S., Singh, N., Singh, R.K., Kaur, P., Perbandt, M., Betzel, Ch., Srinivasan, A. and Singh T.P. (2005c). Crystal structure of a calcium-induced dimer of two isoforms of cobra phospholipase A2 at 1.6 Å resolution. Proteins, 59, 856-863. Jabeen, T., Sharma, S., Singh, N., Singh, R.K., Verma, A.K., Paramasivam, M., Srinivasan, A. and Singh, T.P. (2005d). Structure of the zinc-induced heterodimer of two calcium-free isoforms of phospholipase A2 from Naja naja sagittifera at 2.7 Å resolution. Acta Cryst. D61, 302-308. Jabeen, T., Singh, N., Singh, R.K., Ethayathulla, A.S., Sharma, S., Srinivasan, A. and Singh, T.P. (2005b). Crystal structure of a novel phospholipase A2 from Naja naja sagittifera with a strong anticoagulant activity. Toxicon, 46, 865-875. Jabeen, T., Singh, N., Singh, R.K., Jasti, J., Sharma, S., Kaur, P., Srinivasan, A. and Singh, T.P. (2006). Crystal structure of a heterodimer of phospholipase A2 from Naja naja sagittifera at 2.3 Å resolution reveals the presence of a new PLA2-like protein with a novel Cys32-Cys49 disulphide bridge with a bound sugar at the substrate-binding site. Proteins, 62, 329-337. Jabeen, T., Singh, N., Singh, R.K., Sharma, S., Somvanshi, R.K., Dey, S. and Singh, T.P. (2005a). Non-steroidal anti-inflammatory drugs as potent inhibitors of phospholipase A2: structure of the complex of phospholipase A2 with niflumic acid at 2.5 Å resolution. Acta Cryst. D61, 1579-1586. Jack, A. and Levitt, M. (1978). Refinement of large structures by simultaneous minimization of energy and R factor. Acta Cryst. A34, 931-935. Jain, M.K., Ranadive, G., Yu, B.Z. and Verheij, H.M. (1991). Interfacial catalysis by phospholipase A2: monomeric enzyme is fully catalytically active at the bilayer interface. Biochemistry, 30, 7330-7340.

REFERENCES 225

Jain, A., Vaidehi, N. and Rodriguez, G. (1993). A fast recursive algorithm for molecular dynamics simulation. J. Comp. Phys. 106, 258-268. Jain, M.K. and Vaz, W.L. (1987). Dehydration of the lipid-protein microinterface on binding of phospholipase A2 to lipid bilayers. Biochim. Biophys. Acta. 905, 1-8. Janssen, M.J.W., Wendy, A.E.C., van de Wiel, Beiboer, S.H.W., van Kampen, M.D., Verheij, H.M., Slotboom, A.J. and Egmond, M.R. (1999). Catalytic role of the active site histidine of porcine pancreatic phospholipase A2 probed by the variants H48Q, H48N and H48K. Protein Engineering, 12, 497–503. Jarrett, J.T. (2003). The generation of 5'-deoxyadenosyl radicals by adenosylmethionine-dependent radical enzymes. Curr. Opin. Chem. Biol. 7, 174-182. Jayaram, B. and Jain, T. (2004). The role of water in protein-DNA recognition. Annu. Rev. Biophys. Biomol. Struct. 33, 343-361. Jenkins, C.M., Mancuso, D.J., Yan, W., Sims, H.F., Gibson, B. and Gross, R.W. (2004). Identification, cloning, expression, and purification of three novel human calcium-independent phospholipase A2 family members possessing triacylglycerol lipase and acylglycerol transacylase activities. J. Biol. Chem. 279, 48968-48975. Johnson, J.L. and Duran, M. (2001) Molybdenum cofactor deficiency and isolated sulfite oxidase deficiency. In: Scriver CR, Beaudet AL, Sly WS, Valle D, eds; Childs B, Kinzler KW, Vogelstein B, assoc. eds: The Metabolic and Molecular Bases of Inherited Disease, 8th edn. New York: McGraw-Hill, 3163-3177. Johnson, J.L., Hainline, B.E. and Rajagopalan, K.V. (1980). Characterization of the molybdenum cofactor of sulfite oxidase, xanthine, oxidase, and nitrate reductase. Identification of a pteridine as a structural component. J. Biol. Chem. 255, 1783-1786. Johnson, J.L., Hainline, B.E., Rajagopalan, K.V. and Arison, B.H. (1984). The pterin component of the molybdenum cofactor. Structural characterization of two fluorescent derivatives. J. Biol. Chem. 259, 5414-5422. Johnson, J.L., Indermaur, L.W. and Rajagopalan, K.V. (1991). Molybdenum cofactor biosynthesis in Escherichia coli. Requirement of the chlB gene product for the formation of molybdopterin guanine dinucleotide. J. Biol. Chem. 266, 12140-12145. Johnson, J.L. and Rajagopalan, K.V. (1987a). In vitro synthesis of molybdopterin. J. Bacteriol. 169, 110-116. Johnson, J.L. and Rajagopalan, K.V. (1987b). Involvement of chlA, chlE, chlM and chlN loci in Escherichia coli molybdopterin biosynthesis. J. Bacteriol. 169, 117-125.

REFERENCES 226

Johnson, J.L., Wuebbens, M.M., Mandell, R. and Shih, V.E. (1989a). Molybdenum cofactor biosynthesis in humans. Identification of two complementation groups of cofactor-deficient patients and preliminary characterization of a diffusible molybdopterin precursor. J. Clin. Invest. 83, 897–903. Johnson, J.L., Wuebbens, M.M. and Rajagopalan, K.V. (1989b). The structure of a molybdopterin precursor. Characterization of a stable, oxidized derivative. J. Biol. Chem. 264, 13440-13447. Jorgensen, W.L., Maxwell, D.S. and Tirado-Rives, J. (1996). Development and Testing of the OPLS All-Atom Force Field on Conformational Energetics and Properties of Organic Liquids. J. Am. Chem. Soc. 118, 11225-11236. Jorgensen, W.L. and Tirado-Rives, J. (1988). The OPLS [optimized potentials for liquid simulations] potential functions for proteins, energy minimizations for crystals of cyclic peptides and crambin. J. Am. Chem. Soc. 110, 1657-1666. Joshi, M.S., Johnson, J.L. and Rajagopalan, K.V. (1996). Molybdenum cofactor biosynthesis in Escherichia coli mod and mog mutants. J. Bacteriol. 178, 4310-4312. Kabsch, W. and Sander, C. (1983). Dictionary of protein secondary structure: pattern recognition of hydrogen-bonded and geometrical features. Biopolymers, 22, 2577-2637. Kaminski, G.A., Friesner, R.A., Tirado-Rives, J. and Jorgensen, W.L. (2001). Evaluation and Reparametrization of the OPLS-AA Force Field for Proteins via Comparison with Accurate Quantum Chemical Calculations on Peptides. J. Phys. Chem. B105, 6474-6487. Karshikoff, A. and Ladenstein, R. (2001). Ion pairs and the thermotolerance of proteins from hyperthermophiles: a "traffic rule" for hot roads. Trends in Biochem. Sci. 26, 550-556. Keith, C., Feldman, D.S., Deganello, S., Glick, J., Ward, K.B., Jones, E.O. and Sigler, P.B. (1981). The 2.5 Å crystal structure of a dimeric phospholipase A2 from the venom of Crotalus atrox. J. Biol. Chem. 256, 8602-8607. Kimura, S., Tanaka, T., Shimada, I., Shiratori, Y., Nakagawa, S., Nakamura, H., Inagaki, F. and Ota, Y. (1990). Deletion of D-helix in bovine pancreatic phospholipase A2. Agric. Biol. Chem. 54, 633-639. Kini, R.M. and Evans, H.J. (1987). Structure-function relationships of phospholipases. The anticoagulant region of phospholipases A2. J. Biol. Chem. 262, 14402-14407. Kishan, R.V.R., Chandra, N.R., Sudarsanakumar, C., Suguna, K. and Vijayan, M. (1995). Water-dependent domain motion and flexibility in ribonuclease A and the invariant features in its hydration shell. An X-ray study of two low-humidity crystal forms of the enzyme. Acta Cryst. D51, 703-710.

REFERENCES 227

Kisker, C., Schindelin, H. and Rees, D.C. (1997). Molybdenum-cofactor-containing enzymes: structure and mechanism. Annu. Rev. Biochem. 66, 233–267. Kisker, C., Schindelin, H., Baas, D., Retey, J., Meckenstock, R.U. and Kroneck, P.M. (1998). A structural comparison of molybdenum cofactor-containing enzymes. FEMS Microbiol. Rev. 22, 503-521. Kleywegt, G.J. (2007). Crystallographic refinement of ligand complexes. Acta Cryst. D63, 94-100. Konagurthu, A.S., Whisstock, J.C., Stuckey, P.J. and Lesk, A.M. (2006). MUSTANG: a multiple structural alignment algorithm. Proteins, 64, 559-574. Konnert, J.H. (1976). A restrained-parameter structure-factor least-squares refinement procedure for large asymmetric units. Acta Cryst. A32, 614-617. Koshiba, T., Saito, E., Ono, N., Yamamoto, N. and Sato, M. (1996). Purification and Properties of Flavin- and Molybdenum-Containing Aldehyde Oxidase from Coleoptiles of Maize. Plant Physiol. 110, 781-789. Kostrewa, D. (1997). CCP4 Newsl. 34, 9-22. Kramer, S., Hageman, R.V. and Rajagopalan, K.V. (1984). In vitro reconstitution of nitrate reductase activity of the Neurospora crassa mutant nit-1: specific incorporation of molybdopterin. Arch. Biochem. Biophys. 233, 821-829. Kramer, S.P., Johnson, J.L., Ribeiro, A.A., Millington, D.S. and Rajagopalan, K.V. (1987). The structure of the molybdenum cofactor. Characterization of di-(carboxamidomethyl)molybdopterin from sulfite oxidase and xanthine oxidase. J. Biol. Chem. 262, 16357-16363. Kraut, J. (1977). Serine proteases: structure and mechanism of catalysis. Ann. Rev. Biochem. 46, 331-358. Krem, M.M. and Enrico, D.C. (1998). Conserved water molecules in the specificity pocket of serine proteases and the molecular mechanism of Na+ binding. Proteins, 30, 34-42. Krissinel, E. and Henrick, K. (2007). Inference of macromolecular assemblies from crystalline state. J. Mol. Biol. 372, 774-797. Kuchler, K., Gmachl, M., Sippl, M.J. and Kreil, G. (1989). Analysis of the cDNA for phospholipase A2 from honeybee venom glands. The deduced amino acid sequence reveals homology to the corresponding vertebrate enzymes. Eur. J. Biochem. 184, 249-254. Kudo, I. and Murakami, M. (2002). Phospholipase A2 enzymes. Prostaglandins and Other Lipid Mediators, 68-69, 3-58.

REFERENCES 228

Kumar, A., Sekharudu, C., Ramakrishnan, B., Dupureur, C.M., Zhu, H., Tsai, M.-D. and Sundaralingam, M. (1994). Structure and function of the catalytic site mutant Asp99Asn of phospholipase A2: absence of the conserved structural water. Protein Sci. 3, 2082-2088. Kuper, J., Llamas, A., Hecht, H.J., Mendel, R.R. and Schwarz, G. (2004). Structure of the molybdopterin-bound Cnx1G domain links molybdenum and copper metabolism. Nature, 430, 803-806. Kuper, J., Palmer, T., Mendel, R.R. and Schwarz, G. (2000). Mutations in the molybdenum cofactor biosynthetic protein Cnx1G from Arabidopsis thaliana define functions for molybdopterin binding, molybdenum insertion, and molybdenum cofactor stabilization. Proc. Natl. Acad. Sci. USA, 97, 6475-6480. Kuper, J., Winking, J., Hecht, H.J., Mendel, R.R. and Schwarz, G. (2003). The active site of the molybdenum cofactor biosynthetic protein domain Cnx1G. Arch. Biochem. Biophys. 411, 36-46. Kuramitsu, S., Hiromi, K., Hayashi, H., Morino, Y. and Kagamiyama, H. (1990). Pre-steady-state kinetics of Escherichia coli aspartate aminotransferase catalyzed reactions and thermodynamic aspects of its substrate specificity. Biochemistry, 29, 5469-5476. Lake, M.W., Temple, C.A., Rajagopalan, K.V. and Schindelin, H. (2000). The crystal structure of the Escherichia coli MobA protein provides insight into molybdopterin guanine dinucleotide biosynthesis. J. Biol. Chem. 275, 40211-40217. Lake, M.W., Wuebbens, M.M., Rajagopalan, K.V. and Schindelin, H. (2001). Mechanism of ubiquitin activation revealed by the structure of a bacterial MoeB-MoaD complex. Nature, 414, 325-329. Landau, M., Mayrose, I., Rosenberg, Y., Glaser, F., Martz, E., Pupko, T. and Ben-Tal, N. (2005). ConSurf: the projection of evolutionary conservation scores of residues on protein structures. Nucl. Acids Res. 33, W299-W302. Larkin, M.A., Blackshields, G., Brown, N.P., Chenna, R., McGettigan, P.A., McWilliam, H., Valentin, F., Wallace, I.M., Wilm, A., Lopez, R., Thompson, J.D., Gibson, T.J. and Higgins, D.G. (2007). Clustal W and Clustal X version 2.0. Bioinformatics Applications Note, 23, 2947–2948. Laskowski, R.A. (1995). SURFNET: a program for visualizing molecular surfaces, cavities, and intermolecular interactions. J. Mol. Graph. 13, 323-330. Laskowski, R.A., MacArthur, M.W., Moss, D.S. and Thornton, J.M. (1993). PROCHECK: a program to check the stereochemical quality of protein structures. J. Appl. Cryst. 26, 283-291.

REFERENCES 229

Leatherbarrow, R.J. and Fersht, A.R. (1987). Investigation of transition-state stabilization by residues histidine-45 and threonine-40 in the tyrosyl-tRNA synthetase. Biochemistry, 26, 8524-8528. Lee, B.I., Yoon, E.T. and Cho, W. (1996). Roles of surface hydrophobic residues in the interfacial catalysis of bovine pancreatic phospholipase A2. Biochemistry, 35, 4231-4240. Leimkuhler, S. and Rajagopalan, K.V. (2001). In vitro incorporation of nascent molybdenum cofactor into human sulfite oxidase. J. Biol. Chem. 276, 1837-1844. Leimkuhler, S., Wuebbens, M.M. and Rajagopalan, K.V. (2001). Characterization of Escherichia coli MoeB and its involvement in the activation of molybdopterin synthase for the biosynthesis of the molybdenum cofactor. J. Biol. Chem. 276, 34695-34701. Leiros, I., Secundo, F., Zambonelli, C., Servi, S. and Hough, E. (2000). The first crystal structure of a phospholipase D. Structure, 8, 655-667. Le Trong, I. and Stenkamp, R.E. (2007). An alternate description of two crystal structures of phospholipase A2 from Bungarus caeruleus. Acta Cryst. D63, 548-549. Levitt, M., Hirshberg, M., Sharon, R. and Daggett, V. (1995). Potential energy function and parameters for simulations of the molecular dynamics of proteins and nucleic acids in solution. Comp. Phys. Comm. 91, 215-231. Lewis, N.J., Hurt, P., Sealy-Lewis, H.M. and Scazzocchio, C. (1978). The genetic control of the molybdoflavoproteins in Aspergillus nidulans. IV. A comparison between purine hydroxylase I and II. Eur. J. Biochem. 91, 311-316. Li, Y. and Tsai, M.-D. (1993). Phospholipase A2 engineering. 10. The aspartate...histidine catalytic diad also plays an important structural role. J. Am. Chem. Soc. 115, 8523–8526. Li, Y., Yu, B.Z., Zhu, H., Jain, M.K. and Tsai, M.-D. (1994). Phospholipase A2 engineering. Structural and functional roles of the highly conserved active site residue aspartate-49. Biochemistry, 33, 14714-14722. Lin, G., Noel, J., Loffredo, W., Stable, H.Z. and Tsai, M.-D. (1988). Use of short-chain cyclopentano-phosphatidylcholines to probe the mode of activation of phospholipase A2 from bovine pancreas and bee venom. J. Biol. Chem. 263, 13208-13214. Lindahl, E., Hess, B. and van der Spoel, D. (2001). GROMACS 3.0: a package for molecular simulation and trajectory analysis. J. Mol. Modl. 7, 306-317. Liu, M.T., Wuebbens, M.M., Rajagopalan, K.V. and Schindelin, H. (2000). Crystal structure of the gephyrin-related molybdenum cofactor biosynthesis protein MogA from Escherichia coli. J. Biol. Chem. 275, 1814-1822.

REFERENCES 230

Liu, X., Zhu, H., Huang, B., Rogers, J., Yu, B.Z., Kumar, A., Jain, M.K., Sundaralingam, M. and Tsai, M.-D. (1995). Phospholipase A2 engineering. Probing the structural and functional roles of N-terminal residues with site-directed mutagenesis, X-ray, and NMR. Biochemistry, 34, 7322-7334. Liu, Y. and Eisenberg, D. (2002). 3D domain swapping: as domains continue to swap. Protein Sci. 11, 1285-1299. Llamas, A., Mendel, R.R. and Schwarz, G. (2004). Synthesis of adenylated molybdopterin: an essential step for molybdenum insertion. J. Biol. Chem. 279, 55241-55246. Llamas, A., Otte, T., Multhaup, G., Mendel, R.R. and Schwarz, G. (2006). The Mechanism of nucleotide-assisted molybdenum insertion into molybdopterin. A novel route toward metal cofactor assembly. J. Biol. Chem. 281, 18343-18350. Lo, Y.C., Lin S.C., Shaw, J.F. and Liaw, Y.C. (2003). Crystal structure of Escherichia coli thioesterase I/protease I/lysophospholipase L1: consensus sequence blocks constitute the catalytic center of SGNH-hydrolases through a conserved hydrogen bond network. J. Mol. Biol. 330, 539-551. Loris, R., Stas, P.P.G. and Wyns, L. (1994). Conserved waters in legume lectin crystal structures. The importance of bound water for the sequence-structure relationship within the legume lectin family. J. Biol. Chem. 269, 26722-26733. Ludescher, R.D., Johnson, I.D., Volwerk, J.J., de Haas, G.H., Jost, P.C. and Hudson, B.S. (1988). Rotational dynamics of the single tryptophan of porcine pancreatic phospholipase A2, its zymogen, and an enzyme/micelle complex. A steady-state and time-resolved anisotropy study. Biochemistry, 27, 6618-6628. Luty, B.A., Tironi, I.G. and van Gunsteren, W.F.J. (1995). Lattice-sum methods for calculating electrostatic interactions in molecular simulations. Chem. Phys. 103, 3014−3021. MacKerell, A.D. Jr., Bashford, D., Bellott, M., Dunbrack, R.L., Evanseck, J.D., Field, M.J., Fischer, S., Gao, J., Guo, H., Ha, S., Joseph-McCarthy, D., Kuchnir, L., Kuczera, K., Lau, F.T.K., Mattos, C., Michnick, S., Ngo, T., Nguyen, D.T., Prodhom, B., Reiher, W.E., Roux, B., Schlenkrich, M., Smith, J.C., Stote, R., Straub, J., Watanabe, M., Wiorkiewicz-Kuczera, J., Yin, D. and Karplus, M. (1998). All-Atom Empirical Potential for Molecular Modeling and Dynamics Studies of Proteins. J. Phys. Chem. B102, 3586-3616. Magalon, A., Frixon, C., Pommier, J., Giordano, G. and Blasco, F. (2002). In Vivo Interactions between Gene Products Involved in the Final Stages of Molybdenum Cofactor Biosynthesis in Escherichia coli. J. Biol. Chem. 277, 48199-48204.

REFERENCES 231

Magro, A.J., Takeda, A.A., Soares, A.M. and Fontes, M.R. (2005). Structure of BthA-I complexed with p-bromophenacyl bromide: possible correlations with lack of pharmacological activity. Acta Cryst. D61, 1670-1677. Mansfeld, J., Gebauer, S., Dathe, K. and Ulbrich-Hofmann, R. (2006). Secretory phospholipase A2 from Arabidopsis thaliana: insights into the three-dimensional structure and the amino acids involved in catalysis. Biochemistry, 45, 5687-5694. Matoba, Y. and Sugiyama, M. (2002). Atomic resolution structure of prokaryotic phospholipase A2: analysis of internal motion and implication for a catalytic mechanism. Proteins, 51, 453-469. Matthews, B.W. (1968). Solvent content of protein crystals. J. Mol. Biol. 33, 491–497. Matthies, A., Rajagopalan, K.V., Mendel, R.R. and Leimkuhler, S. (2004). Evidence for the physiological role of a rhodanese-like protein for the biosynthesis of the molybdenum cofactor in humans. Proc. Natl. Acad. Sci. USA, 101, 5946-5951. Matthies, A., Nimtz, M. and Leimkuhler, S. (2005). Molybdenum cofactor biosynthesis in humans: identification of a persulfide group in the rhodanese-like domain of MOCS3 by mass spectrometry. Biochemistry, 44, 7912-7920. Maupin-Furlow, J.A., Rosentel, J.K., Lee, J.H., Deppenmeier, U., Gunsalus, R.P. and Shanmugam, K.T. (1995). Genetic analysis of the modABCD (molybdate transport) operon of Escherichia coli. J. Bacteriol. 177, 4851-4856. McCoy, A.J., Grosse-Kunstleve, R.W., Adams, P.D., Winn, M.D., Storoni, L.C. and Read, R.J. (2007). Phaser crystallographic software. J. Appl. Cryst. 40, 658-674. McDonald, I.K. and Thornton, J.M. (1994). Satisfying hydrogen bonding potential in proteins. J. Mol. Biol. 238, 777-793. McIntosh, J.M., Ghomashchi, F., Gelb, M.H., Dooley, D.J., Stoehr, S.J., Giordani, A.B., Naisbitt, S.R. and Olivera, B.M. (1995). Conodipine-M, a novel phospholipase A2 isolated from the venom of the marine snail Conus magus. J. Biol. Chem. 270, 3518-3526. McLuskey, K., Harrison, J.A., Schuttelkopf, A.W., Boxer, D.H. and Hunter, W.N. (2003). Insight into the role of Escherichia coli MobB in molybdenum cofactor biosynthesis based on the high resolution crystal structure. J. Biol. Chem. 278, 23706-23713. McPherson, A. (1985). Crystallization of macromolecules: general principles. Methods Enzymol. 114, 112-120. McPherson, A. (1990). Current approaches to macromolecular crystallization. Eur. J. Biochem. 189, 1-23.

REFERENCES 232

McPherson, A. (1997). Recent advances in the microgravity crystallization of biological macromolecules. Trends Biotechnol. 15, 197-200. Mcpherson, A. (2001). A comparison of salts for the crystallization of macromolecules. Protein Sci. 10, 418–422. McPherson, A. (2004a). Introduction to protein crystallization. Methods, 34, 254-265. McPherson, A. (2004b). Protein crystallization in the structural genomics era. J. Struct. Funct. Genomics, 5, 3-12. McPherson, A., Malkin, A.J. and Kuznetsov, Y.G. (1995). The science of macromolecular crystallization. Structure, 3, 759-768. McPherson, A., Nguyen, C., Larson, S.B., Day, J.S. and Cudney, B. (2007). Development of an alternative approach to protein crystallization. J. Struct. Funct. Genomics, 8, 193-198. Mendel, R.R. (1992). The plant molybdenum cofactor (MoCo)-its biochemical and molecular genetics. In: Gresshoff PM, ed. Plant biotechnology and development–current topics in plant molecular biology. Boca Raton: CRC Press, 11–16. Mendel, R.R. and Bittner, F. (2006). Cell biology of molybdenum. Biochim. Biophys. Acta, 1763, 621-635. Mendel, R.R. and Schwarz, G. (2002). Biosynthesis and molecular biology of the molybdenum cofactor (Moco). Met. Ions. Biol. Syst. 39, 317-368. Menendez, C., Siebert, D. and Brandsch, R. (1996). MoaA of Arthrobacter nicotinovorans pAO1 involved in Mo-pterin cofactor synthesis is an Fe-S protein. FEBS Lett. 391, 101-103. Meyer, E. (1992). Internal water molecules and hydrogen bonding in biological macromolecules: A review of structural features with functional implications. Protein Sci. 1, 1543-1562. Millar, L.J., Heck, I.S., Sloan, J., Kanan, G.J., Kinghorn, J.R. and Unkles, S.E. (2001). Deletion of the cnxE gene encoding the gephyrin-like protein involved in the final stages of molybdenum cofactor biosynthesis in Aspergillus nidulans. Mol. Genet. Genomics, 266, 445-453. Mitchell, P.C.H. (2003). Overview of Environment Database. International Molybdenum Association. Mitsuishi, M., Masuda, S., Kudo, I. and Murakami, M. (2006). Group V and X secretory phospholipase A2 prevents adenoviral infection in mammalian cells. Biochem. J. 393, 97-106.

REFERENCES 233

Miyamoto, Y., Kawamata, Y., Yokoigawa, H., Yamamoto, H., Kataoka, K. and Kashima, K. (1993). Usefulness of serum pancreatic phospholipase A2 determination in patients with various pancreatic diseases. Kaku Igaku, 30, 1103-1110. Montalbini, P. (1992a). Inhibition of hypersensitive response by allopurinol applied to the host in the incompatible relationship between Phaseolus vulgaris and Uromyces phaseoli. J. Phytopathology, 134, 218–228. Montalbini, P. (1992b). Ureides and enzymes of ureide synthesis in wheat seeds and leaves and effect of allopurinol on Puccinia recondita f.sp. tritici infection. Plant Sci. 87, 225–231. Montalbini, P. and Della Torre, G. (1996). Evidence of a two-fold mechanism responsible for the inhibition by allopurinol of the hypersensitive response induced in tobacco by tobacco necrosis virus. Physiol. Mol. Plant Pathol. 48, 273–287. Moolenaar, W.H., van Meeteren, L.A. and Giepmans, B.N. (2004). The ins and outs of lysophosphatidic acid signaling. Bioessays, 26, 870-881. Morris, G.M., Goodsell, D.S., Halliday, R.S., Huey, R., Hart, W.E., Belew, R.K. and Olson, J.A. (1998). Automated Docking Using a Lamarckian Genetic Algorithm and an Empirical Binding Free Energy Function. J. Compl. Chem. 19, 1639-1662. Moura, J.J., Brondino, C.D., Trincao, J. and Romao, M.J. (2004). Mo and W bis-MGD enzymes: nitrate reductases and formate dehydrogenases. J. Biol. Inorg. Chem. 9, 791-799. Mukherjee, A.B., Miele, L. and Pattabiraman, N. (1994). Phospholipase A2 enzymes: regulation and physiological role. Biochem. Pharmacol. 48, 1-10. Muller, A.J. and Mendel, R.R. (1989). Biochemical and somatic cell genetics of nitrate reductase in Nicotiana. In: Wray JL, Kinghorn JR, eds. Molecular and genetic aspects of nitrate assimilation. Oxford: Oxford University Press, 166–185. Muller, G. and Petry, S. (2005). Lipases and Phospholipases in Drug Development: From Biochemistry to Molecular Pharmacology, Wiley-VCH Verlag GmbH & Co. KGaA. Murakami, M. and Kudo, I. (2002). Phospholipase A2. J. Biochem. 131, 285-292. Muralikrishna, A.R. and Hatcher, J.F. (2006). Phospholipase A2, reactive oxygen species, and lipid peroxidation in cerebral ischemia. Free Radic. Biol. Med. 40, 376-387. Murzin, A.G., Brenner, S.E., Hubbard, T. and Chothia, C. (1995). SCOP: a structural classification of proteins database for the investigation of sequences and structures. J. Mol. Biol. 247, 536-540.

REFERENCES 234

Nason, A., Lee, K.Y., Pan, S.S., Ketchum, P.A., Lamberti, A. and DeVries, J. (1971). In vitro formation of assimilatory reduced nicotinamide adenine dinucleotide phosphate: nitrate reductase from a Neurospora mutant and a component of molybdenum-enzymes. Proc. Natl. Acad. Sci. USA, 68, 3242-3246. Nevalainen, T.J., Graham, G.G. and Scott, K.F. (2008). Antibacterial actions of secreted phospholipases A2. Biochim. Biophys. Acta, 1781, 1-9. Nevalainen, T.J., Haapamaki, M.M. and Gronroos, J.M. (2000). Roles of secretory phospholipases A2 in inflammatory diseases and trauma. Biochim. Biophys. Acta, 1488, 83-90. Nichols, J.D. and Rajagopalan, K.V. (2002). Escherichia coli MoeA and MogA. Function in metal incorporation step of molybdenum cofactor biosynthesis. J. Biol. Chem. 277, 24995-25000. Nichols, J.D. and Rajagopalan, K.V. (2005). In vitro molybdenum ligation to molybdopterin using purified components. J. Biol. Chem. 280, 7817-7822. Nieder, J., Stallmeyer, B., Brinkmann, H. and Mendel, R.R. (1997). Molybdenum cofactor biosynthesis: identification of A. thaliana cDNAs homologous to the E. coli sulfotransferase MoeB. In: Cram WJ, De Kok LJ, Stulen I, Brunold C, Rennenberg H (eds) Sulfur metabolism in higher plants: molecular, ecophysiological and nutritional aspects. SPB Academic Publishers, Amsterdam, 275-277. Nieter Burgmayer, S.J., Pearsall, D.L., Blaney, S.M., Moore, E.M. and Sauk-Schubert, C. (2004). Redox reactions of the pyranopterin system of the molybdenum cofactor. J. Biol. Inorg. Chem. 9, 59-66. Nimrod, G., Schushan, M., Steinberg, D.M. and Ben-Tal, N. (2008). Detection of Functionally Important Regions in “Hypothetical Proteins” of Known Structure. Structure, 16, 1755-1763. Noel, J.P., Bingman, C.A., Deng, T.L., Dupureur, C.M., Hamilton, K.J., Jiang, R.T., Kwak, J.G., Sekharudu, C., Sundaralingam, M. and Tsai, M.-D. (1991). Phospholipase A2 engineering. X-ray structural and functional evidence for the interaction of lysine-56 with substrates. Biochemistry, 30, 11801-11811. Nohno, T., Kasai, Y. and Saito, T. (1988). Cloning and sequencing of the Escherichia coli chlEN operon involved in molybdopterin biosynthesis. J. Bacteriol. 170, 4097-4102. Nose, S. (1984). A unified formulation of the constant temperature molecular dynamics methods. J. Chem. Phys. 81, 511-519. Ohto, T., Uozumi, N., Hirabayashi, T. and Shimizu, T. (2005). Identification of novel cytosolic phospholipase A2s, murine cPLA2δ, ε, and ζ, which form a gene cluster with cPLA2β. J. Biol. Chem. 280, 24576-24583.

REFERENCES 235

Okuda, S., Katayama, T., Kawashima, S., Goto, S. and Kanehisa, M. (2006). ODB: a database of operons accumulating known operons across multiple genomes. Nucl. Acids Res. 34, D358-D362. Otting, G. (1997). Protein hydration in aqueous solution. Prog. Nucl. Magn. Res. Spectrosc. 31, 259-285. Otting, G., Liepinsh, E. and Wuthrich, K. (1991). Protein hydration in aqueous solution. Science, 254, 974-980. Otwinowski, Z. and Minor, W. (1997). Processing of X-ray diffraction data collected in oscillation mode. Methods Enzymol. 276, 307-326. Palmer, T., Goodfellow, I.P., Sockett, R.E., McEwan, A.G. and Boxer, D.H. (1998). Characterisation of the mob locus from Rhodobacter sphaeroides required for molybdenum cofactor biosynthesis. Biochim. Biophys. Acta. 1395, 135-140. Palmer, T., Santini, C.L., Iobbi-Nivol, C., Eaves, D.J., Boxer, D.H. and Giordano G. (1996). Involvement of the narJ and mob gene products in distinct steps in the biosynthesis of the molybdoenzyme nitrate reductase in Escherichia coli. Mol. Microbiol. 20, 875-884. Palmer, T., Vasishta, A., Whitty, P.W. and Boxer, D.H. (1994). Isolation of protein FA, a product of the mob locus required for molybdenum cofactor biosynthesis in Escherichia coli. Eur. J. Biochem. 222, 687-692. Pan, Y.H., Yu, B.Z., Singer, A.G., Ghomashchi, F., Lambeau, G., Gelb, M.H., Jain, M.K. and Bahnson, B.J. (2002). Crystal structure of human group X secreted phospholipase A2. Electrostatically neutral interfacial surface targets zwitterionic membranes. J. Biol. Chem. 277, 29086-29093. Pannu, N.S. and Read, R.J. (1996). Improved structure refinement through maximum likelihood. Acta Cryst. A52, 659-668. Papoian, G.A., Ulander, J., Eastwood, M.P., Luthey-Schulten, Z. and Wolynes, P.G. (2004). Water in protein structure prediction. Proc. Natl. Acad. Sci. USA, 101, 3352-3357. Parrinello, M. and Rahman, A. (1981). Polymorphic transitions in single crystals: A new molecular dynamics method. J. Appl. Phys. 52, 7182-7190. Pastori, G.M. and Del Rio, L.A. (1997). Natural Senescence of Pea Leaves (An Activated Oxygen-Mediated Function for Peroxisomes). Plant Physiol. 113, 411-418. Pateman, J.A., Cove, D.J., Rever, B.M. and Roberts, D.B. (1964). A common co-factor for nitrate reductase and xanthine dehydrogenase which also regulates the synthesis of nitrate reductase. Nature, 201,58-60.

REFERENCES 236

Pitterle, D.M., Johnson, J.L. and Rajagopalan, K.V. (1993). In vitro synthesis of molybdopterin from precursor Z using purified converting factor. Role of protein-bound sulfur in formation of the dithiolene. J. Biol. Chem. 268, 13506-13509. Pitterle, D.M. and Rajagopalan, K.V. (1989). Two proteins encoded at the chlA locus constitute the converting factor of Escherichia coli chlA1. J. Bacteriol. 171, 3373-3378. Pitterle, D.M. and Rajagopalan, K.V. (1993). The biosynthesis of molybdopterin in Escherichia coli. Purification and characterization of the converting factor. J. Biol. Chem. 268, 13499-13505. Plesniak, L.A., Yu, L. and Dennis, E.A. (1995). Conformation of micellar phospholipid bound to the active site of phospholipase A2. Biochemistry, 34, 4943-4951. Ponder, J.W. and Case, D.A. (2003). Force fields for protein simulations. Adv. Protein Chem. 66, 27-85. Pople, J.A. and Nesbet, R.K. (1954). Self-consistent Orbitals for Radicals. J. Chem. Phys. 22, 571. Powell, M.J.D. (1977). Restart procedures for the conjugate gradient method. Math. Program. 12, 241-254. Prasad, B.V.L.S. and Suguna, K. (2002). Role of water molecules in the structure and function of aspartic proteinases. Acta Cryst. D58, 250-259. Praveen, S., Ramesh, J., Sivasankari, P., Sowmiya, G. and Sekar, K. (2008). WAP(2.0): An updated computing and visualization server for water molecules. J. Appl. Cryst. 41, 952-954. Pruzanski, W., Bogoch, E., Wloch, M. and Vadas, P. (1991). The role of phospholipase A2 in the physiopathology of osteoarthritis. J. Rheumatol. Suppl. 27, 117-119. Ptitsyn, O.B. (1998). Protein folding and protein evolution: common folding nucleus in different subfamilies of c-type cytochromes. J. Mol. Biol. 278, 655-666. Puijk, W.C., Verheij, H.M. and de Haas, G.H. (1977). The primary structure of phospholipase A2 from porcine pancreas. A reinvestigation. Biochim. Biophys. Acta, 492, 254-259. Rajagopalan, K.V. (1991). Novel aspects of the biochemistry of the molybdenum cofactor. Adv. Enzymol. Relat. Areas. Mol. Biol. 64, 215-290. Rajagopalan, K.V. (1996). E. coli and Salmonella typhimurium, ASM press. Rajagopalan, K.V. and Johnson, J.L. (1992). The pterin molybdenum cofactors. J. Biol. Chem. 267, 10199-10202.

REFERENCES 237

Rajakannan, V., Yogavel, M., Poi, M.J., Jeyaprakash, A.A., Jeyakanthan, J., Velmurugan, D., Tsai, M.-D. and Sekar, K. (2002). Observation of additional calcium ion in the crystal structure of the triple mutant K56,120,121M of bovine pancreatic phospholipase A2. J. Mol. Biol. 324, 755-762. Ramachandran, G.N., Ramakrishnan, C. and Sasisekharan, V. (1963). Stereochemistry of polypeptide chain configurations. J. Mol. Biol. 7, 95-99. Raschke, T.M. (2006). Water structure and interactions with protein surfaces. Curr. Opin. Struct. Biol. 16, 152-159. Raykova, D. and Blagoev, B. (1986). Hydrolysis of short-chain phosphatidylcholines by bee venom phospholipase A2. Toxicon, 24, 791-797. Read, R.J. (1986). Improved Fourier coefficients for maps using phases from partial structures with errors. Acta Cryst. A42, 140-149. Read, R.J. (1997). Model phases: probabilities and bias. Methods Enzymol. 278, 110-128. Read, R.J. (2001). Pushing the boundaries of molecular replacement with maximum likelihood. Acta Cryst. D57, 1373-1382. Rebelo, J., Auerbach, G., Bader, G., Bracher, A., Nar, H., Hosl, C., Schramek, N., Kaiser, J., Bacher, A., Huber, R. and Fischer, M. (2003). Biosynthesis of pteridines. Reaction mechanism of GTP cyclohydrolase I. J. Mol. Biol. 326, 503-516. Reiss, J. (2000). Genetics of molybdenum cofactor deficiency. Hum. Genet. 106, 157-163. Reiss, J., Cohen, N., Dorche, C., Mandel, H., Mendel, R.R., Stallmeyer, B., Zabot, M.T. and Dierks, T. (1998). Mutations in a polycistronic nuclear gene associated with molybdenum cofactor deficiency. Nat. Genet. 20, 51-53. Reiss, J. and Johnson, J.L. (2003). Mutations in the molybdenum cofactor biosynthetic genes MOCS1, MOCS2, and GEPH. Hum. Mutat. 21, 569-576. Renetseder, R., Brunie, S., Dijkstra, B.W., Drenth, J. and Sigler, P.B. (1985). A comparison of the crystal structures of phospholipase A2 from bovine pancreas and Crotalus atrox venom. J. Biol. Chem. 260, 11627-11634. Renetseder, R., Dijkstra, B.W., Huizinga, K., Kalk, K.H. and Drenth J. (1988). Crystal structure of bovine pancreatic phospholipase A2 covalently inhibited by p-bromo-phenacyl-bromide. J. Mol. Biol. 200, 181-188. Rice, L.M. and Brunger, A.T. (1994). Torsion angle dynamics: Reduced variable conformational sampling enhances crystallographic structure refinement. Proteins: Struct. Funct. Genet. 19, 277-290.

REFERENCES 238

Richmond, B.L., Boileau, A.C., Zheng, S., Huggins, K.W., Granholm, N.A., Tso, P. and Hui, D.Y. (2001). Compensatory phospholipid digestion is required for cholesterol absorption in pancreatic phospholipase A2-deficient mice. Gastroenterology, 120, 1193-1202. Rieder, C., Eisenreich, W., O'Brien, J., Richter, G., Gotze, E., Boyle, P., Blanchard, S., Bacher, A. and Simon, H. (1998). Rearrangement reactions in the biosynthesis of molybdopterin-an NMR study with multiply 13C/15N labelled precursors. Eur. J. Biochem. 255, 24-36. Rivers, S.L., McNairn, E., Blasco, F., Giordano, G. and Boxer, D.H. (1993). Molecular genetic analysis of the moa operon of Escherichia coli K-12 required for molybdenum cofactor biosynthesis. Mol. Microbiol. 8, 1071-1081. Robinson, D.R., Tashjian, A.H. Jr. and Levine, L. (1975). Prostaglandin-induced bone resorption by rheumatoid synovia. Trans. Assoc. Am. Physicians, 88, 146-160. Robson, R., Woodley, P. and Jones, R. (1986). Second gene (nifH*) coding for a nitrogenase iron protein in Azotobacter chroococcum is adjacent to a gene coding for a ferredoxin-like protein. EMBO J. 5, 1159-1163. Rogers, J., Yu, B.Z., Tsai, M.-D., Berg, O.G. and Jain, M.K. (1998). Cationic residues 53 and 56 control the anion-induced interfacial k*cat activation of pancreatic phospholipase A2. Biochemistry, 37, 9549-9556. Ross, B.M., Moszczynska, A., Erlich, J. and Kish, S.J. (1998). Low activity of key phospholipid catabolic and anabolic enzymes in human substantia nigra: possible implications for Parkinson's disease. Neuroscience, 83, 791-798. Rossmann, M.G. (1972). The locked rotation function. J. Mol. Biol. 64, 246-249. Rossmann, M.G. (1990). The molecular replacement method. Acta Cryst. A46, 73-82. Rossmann, M.G. and Blow, D.M. (1962). The detection of sub-units within the crystallographic asymmetric unit. Acta Cryst. 15, 24-31. Rossmann, M.G., Blow, D.M., Harding, M.M. and Coller, E. (1964). The relative positions of independent molecules within the same asymmetric unit. Acta Cryst. 17, 338-342. Roulston, M.S. (1999). Estimating the errors on measured entropy and mutual information. Physica, D125, 285-294. Rudolph, M.J., Wuebbens, M.M., Rajagopalan, K.V. and Schindelin, H. (2001). Crystal structure of molybdopterin synthase and its evolutionary relationship to ubiquitin activation. Nat. Struct. Biol. 8, 42-46.

REFERENCES 239

Rudolph, M.J., Wuebbens, M.M., Turque, O., Rajagopalan, K.V. and Schindelin, H. (2003). Structural studies of molybdopterin synthase provide insights into its catalytic mechanism. J. Biol. Chem. 278, 14514-14522. Sanishvili, R., Beasley, S., Skarina, T., Glesne, D., Joachimiak, A., Edwards, A. and Savchenko, A. (2004). The crystal structure of Escherichia coli MoaB suggests a probable role in molybdenum cofactor synthesis. J. Biol. Chem. 279, 42139-42146. Santamaria-Araujo, J.A., Fischer, B., Otte, T., Nimtz, M., Mendel, R.R., Wray, V. and Schwarz, G. (2004). The tetrahydropyranopterin structure of the sulfur-free and metal-free molybdenum cofactor precursor. J. Biol. Chem. 279, 15994-15999. Schaloske, R.H. and Dennis, E.A. (2006). The phospholipase A2 superfamily and its group numbering system. Biochim. Biophys. Acta, 1761, 1246-1259. Schevitz, R.W., Bach, N.J., Carlson, D.G., Chirgadze, N.Y., Clawson, D.K., Dillard, R.D., Draheim, S.E., Hartley, L.W., Jones, N.D., Mihelich, E.D., Olkowski, J.L., Snyder, D.W., Sommers, C. and Wery, J.-P. (1995). Structure-based design of the first potent and selective inhibitor of human non-pancreatic secretory phospholipase A2. Nat. Struct. Biol. 2, 458-465. Schneider, T.R. (2002). A genetic algorithm for the identification of conformationally invariant regions in protein molecules. Acta Cryst. D58, 195-208. Schneider, T.R. (2004). Domain identification by iterative analysis of error-scaled difference distance matrices. Acta Cryst. D60, 2269-2275. Schuttelkopf, A.W. and van Aalten, D.M.F. (2004). PRODRG: a tool for high-throughput crystallography of protein-ligand complexes. Acta Cryst. D60, 1355-1363. Schuttelkopf, A.W., Boxer, D.H. and Hunter, W.N. (2003). Crystal structure of activated ModE reveals conformational changes involving both oxyanion and DNA-binding domains. J. Mol. Biol. 326, 761-767. Schwarz, G. (2005). Molybdenum cofactor biosynthesis and deficiency. Cell. Mol. Life Sci. 62, 2792-2810. Schwarz, G., Boxer, D.H. and Mendel, R.R. (1997). Molybdenum cofactor biosynthesis. The plant protein Cnx1 binds molybdopterin with high affinity. J. Biol. Chem. 272, 26811-26814. Schwarz, G. and Mendel, R.R. (2006). Molybdenum cofactor biosynthesis and molybdenum enzymes. Annu. Rev. Plant Biol. 57, 623-647. Schwarz, G., Mendel, R.R. and Ribbe, M.W. (2009). Molybdenum cofactors, enzymes and pathways. Nature, 460, 839-847.

REFERENCES 240

Schwarz, G., Schrader, N., Mendel, R.R., Hecht, H.J. and Schindelin, H. (2001). Crystal structures of human gephyrin and plant Cnx1G domains: comparative analysis and functional implications. J. Mol. Biol. 312, 405-418. Schwarz, G., Schulze, J., Bittner, F., Eilers, T., Kuper, J., Bollmann, G., Nerlich, A., Brinkmann, H. and Mendel, R.R. (2000). The molybdenum cofactor biosynthetic protein Cnx1 complements molybdate-repairable mutants, transfers molybdenum to the metal binding pterin, and is associated with the cytoskeleton. Plant Cell, 12, 2455-2472. Scott, D.L., Otwinowski, Z., Gelb, M.H. and Sigler, P.B. (1990a). Crystal structure of bee-venom phospholipase A2 in a complex with a transition-state analogue. Science, 250, 1563-1566. Scott, D.L. and Sigler, P.B. (1994a). Structure and catalytic mechanism of secretory phospholipases A2. Adv. Protein Chem. 45, 53–88. Scott, D.L. and Sigler, P.B. (1994b). The structural and functional roles of calcium ion in secretory phospholipases A2. Adv. Inorg. Biochem. 10, 139-155. Scott, D.L., White, S.P., Browning, J.L., Rosa, J.J., Gelb, M.H. and Sigler, P.B. (1991). Structures of free and inhibited human secretory phospholipase A2 from inflammatory exudate. Science, 254, 1007-1010. Scott, D.L., White, S.P., Otwinowski, Z., Yuan, W., Gelb, M.H. and Sigler, P.B. (1990b). Interfacial catalysis: the mechanism of phospholipase A2. Science, 250, 1541-1546. Segelke, B.W., Nguyen, D., Chee, R., Xuong, N.H. and Dennis, E.A. (1998). Structures of two novel crystal forms of Naja naja naja phospholipase A2 lacking Ca2+ reveal trimeric packing. J. Mol. Biol. 279, 223-232. Seilhamer, J.J., Plant, S., Pruzanski, W., Schilling, J., Stefanski, E., Vadas, P. and Johnson, L.K. (1989). Multiple forms of phospholipase A2 in arthritic synovial fluid. J. Biochem. 106, 38-42. Seilhamer, J.J., Randall, T.L., Yamanaka, M. and Johnson, L.K. (1986). Pancreatic phospholipase A2: isolation of the human gene and cDNAs from porcine pancreas and human lung. DNA, 5, 519-527. Sekar, K. (2007). Structural biology of recombinant bovine pancreatic phospholipase A2 and its inhibitor complexes. Curr. Topics Med. Chem. 7, 779-785. Sekar, K., Biswas, K., Li, Y., Tsai, M.-D. and Sundaralingam, M. (1999). Structures of the catalytic site mutants D99A and H48Q and the calcium-loop mutant D49E of phospholipase A2. Acta Cryst. D55, 443-447.

REFERENCES 241

Sekar, K., Eswaramoorthy, S., Jain, M.K. and Sundaralingam, M. (1997b). Crystal structure of the complex of bovine pancreatic phospholipase A2 with the inhibitor 1-hexadecyl-3-(trifluoroethyl)-sn-glycero-2-phosphomethanol. Biochemistry, 36, 14186-14191. Sekar, K., Gayathri, D., Velmurugan, D., Jeyakanthan, J., Yamane, T., Poi, M.J. and Tsai, M.-D. (2006b). Third calcium ion found in an inhibitor-bound phospholipase A2. Acta Cryst. D62, 392-397. Sekar, K., Kumar, A., Liu, X., Tsai, M.-D., Gelb, M.H. and Sundaralingam, M. (1998b). Structure of the Complex of Bovine Pancreatic Phospholipase A2 with a Transition-State Analogue. Acta Cryst. D54, 334-341. Sekar, K., Rajakannan, V., Gayathri, D., Velmurugan, D., Poi, M.J., Dauter, M., Dauter, Z. and Tsai, M.-D. (2005). Atomic resolution (0.97 Å) structure of the triple mutant (K53,56,121M) of bovine pancreatic phospholipase A2. Acta Cryst. F61, 3-7. Sekar, K., Rajakannan, V., Velmurugan, D., Yamane, T., Thirumurugan, R., Dauter, M. and Dauter, Z. (2004). A redetermination of the structure of the triple mutant (K53,56,120M) of phospholipase A2 at 1.6 Å resolution using sulfur-SAS at 1.54 Å wavelength. Acta Cryst. D60, 1586-1590. Sekar, K., Sekharudu, C., Tsai, M.-D. and Sundaralingam, M. (1998a). 1.72 Å Resolution Refinement of the Trigonal Form of Bovine Pancreatic Phospholipase A2. Acta Cryst. D54, 342-346. Sekar, K. and Sundaralingam, M. (1999). High-resolution refinement of orthorhombic bovine pancreatic phospholipase A2. Acta Cryst. D55, 46-50. Sekar, K., Vaijayanthi, M.S., Yogavel, M., Velmurugan, D., Poi, M.J., Vishwanath, B.S., Gowda, T.V., Jeyaprakash, A.A. and Tsai M.-D. (2003). Crystal structures of the free and anisic acid bound triple mutant of phospholipase A2. J. Mol. Biol. 333, 367-376. Sekar, K., Yogavel, M., Kanaujia, S.P., Sharma, A., Velmurugan, D., Poi, M.J., Dauter, Z. and Tsai, M.-D. (2006a). Suggestive evidence for the involvement of the second calcium and surface loop in interfacial binding: monoclinic and trigonal crystal structures of a quadruple mutant of phospholipase A2. Acta Cryst. D62, 717-724. Sekar, K., Yu, B.Z., Rogers, J., Lutton, J., Liu, X., Chen, X., Tsai, M.-D., Jain, M.K. and Sundaralingam, M.A. (1997a). Phospholipase A2 engineering. Structural and functional roles of the highly conserved active site residue aspartate-99. Biochemistry, 36, 3104-3114. Sekharudu, C., Ramakrishnan, B., Huang, B., Jiang, R.T., Dupureur, C.M., Tsai, M.-D. and Sundaralingam, M. (1992). Crystal structure of the Y52F/Y73F double mutant of phospholipase A2: increased hydrophobic interactions of the phenyl groups compensate for the disrupted hydrogen bonds of the tyrosines. Protein Sci. 1, 1585-1594.

REFERENCES 242

Shanmugam, K.T., Stewart, V., Gunsalus, R.P., Boxer, D.H., Cole, J.A., Chippaux, M., DeMoss, J.A., Giordano, G., Lin, E.C. and Rajagopalan, K.V. (1992). Proposed nomenclature for the genes involved in molybdenum metabolism in Escherichia coli and Salmonella typhimurium. Mol. Microbiol. 6, 3452-3454. Shanthi, V., Rajesh, C.K., Jayalakshmi, J., Vijay, V.G. and Sekar, K. (2003). WAP: water analysis package - a Web-based package to calculate geometrical parameters between water oxygen and protein atoms. J. Appl. Cryst. 36, 167-168. Sigel, A. and Sigel, H. (2002). Molybdenum and tungsten, Their Roles in Biological Processes, vol. 39, Marcel Dekker, New York. Singh, G., Gourinath, S., Sarvanan, K., Sharma, S., Bhanumathi, S., Betzel, Ch., Yadav, S., Srinivasan, A. and Singh, T.P. (2005b). Crystal structure of a carbohydrate induced homodimer of phospholipase A2 from Bungarus caeruleus at 2.1 Å resolution. J. Struct. Biol. 149, 264-272. Singh, G., Gourinath, S., Sharma, S., Paramasivam, M., Srinivasan, A. and Singh, T.P. (2001). Sequence and crystal structure determination of a basic phospholipase A2 from common krait (Bungarus caeruleus) at 2.4 Å resolution: identification and characterization of its pharmacological sites. J. Mol. Biol. 307, 1049-1059. Singh, G., Jasti, J., Saravanan, K., Sharma, S., Kaur, P., Srinivasan, A. and Singh, T.P. (2005a). Crystal structure of the complex formed between a group I phospholipase A2 and a naturally occurring fatty acid at 2.7 Å resolution. Protein Sci. 14, 395-400. Singh, N., Somvanshi, R.K., Sharma, S., Dey, S., Kaur, P. and Singh, T.P. (2007). Structural elements of ligand recognition site in secretory phospholipase A2 and structure-based design of specific inhibitors. Curr. Top. Med. Chem. 7, 757-764. Singh, R.K., Vikram, P., Makker, J., Jabeen, T., Sharma, S., Dey, S., Kaur, P., Srinivasan, A. and Singh, T.P. (2003). Design of specific peptide inhibitors for group I phospholipase A2: structure of a complex formed between phospholipase A2 from Naja naja sagittifera (group I) and a designed peptide inhibitor Val-Ala-Phe-Arg-Ser (VAFRS) at 1.9 Å resolution reveals unique features. Biochemistry, 42, 11701-11706. Six, D.A and Dennis, E.A (2000). The expanding superfamily of phospholipase A2 enzymes: classification and characterization. Biochimica et Biophysica Acta, 1488, 1-19. Slotboom, A.J., Jansen, E.H., Vlijm, H., Pattus, F., de Araujo, P.S. and de Haas, G.H. (1978). Ca2+ binding to porcine pancreatic phospholipase A2 and its function in enzyme-lipid interaction. Biochemistry, 17, 4593-4600. Slotta, K. H. and Fraenkel-Conrat, H. (1938). Schlangengifte III. Mitteilung: Reinigung und Kristallisation des Klapperschlangengiftes. Ber. Dtsch. Chem. Ges. 71, 1076-1081.

REFERENCES 243

Smith, D.K., Radivojac, P., Obradovic, Z., Dunker, A.K. and Zhu, G. (2003). Improved amino acid flexibility parameters. Protein Sci. 12, 1060-1072. Smolin, N., Oleinikova, A., Brovchenko, I., Geiger, A. and Winter, R. (2005). Properties of Spanning Water Networks at Protein Surfaces. J. Phys. Chem. B109, 10995-11005. Smolinsky, B., Eichler, S.A., Buchmeier, S., Meier, J.C. and Schwarz, G. (2008). Splice-specific functions of gephyrin in molybdenum cofactor biosynthesis. J. Biol. Chem. 283, 17370-17379. Sofia, H.J., Chen, G., Hetzler, B.G., Reyes-Spindola, J.F. and Miller, N.E. (2001). Radical SAM, a novel protein superfamily linking unresolved steps in familiar biosynthetic pathways with radical mechanisms: functional characterization using new analysis and information visualization methods. Nucl. Acids Res. 29, 1097-1106. Sola, M., Bavro, V.N., Timmins, J., Franz, T., Ricard-Blum, S., Schoehn, G., Ruigrok, R.W., Paarmann, I., Saiyed, T., O'Sullivan, G.A., Schmitt, B., Betz, H. and Weissenhorn, W. (2004). Structural basis of dynamic glycine receptor clustering by gephyrin. EMBO J. 23, 2510-2519. Sola, M., Kneussel, M., Heck, I.S., Betz, H. and Weissenhorn, W. (2001). X-ray crystal structure of the trimeric N-terminal domain of gephyrin. J. Biol. Chem. 276, 25294-25301. Sorger, G.J. and Giles, N.H. (1965). Genetic control of nitrate reductase in Neurospora crassa. Genetics, 52, 777-788. Sorin, E.J. and Pande, V.S. (2005). Exploring the Helix-Coil Transition via All-Atom Equilibrium Ensemble Simulations. Biophys. J. 88, 2472-2493. Sosa, B.P., Alagon, A.C., Martin, B.M. and Possani, L.D. (1986). Biochemical characterization of the phospholipase A2 purified from the venom of the Mexican beaded lizard (Heloderma horridum horridum Wiegmann). Biochemistry, 25, 2927-2933. Spiegel, S., Foster, D. and Kolesnick, R. (1996). Signal transduction through lipid second messengers. Curr. Opin. Cell Biol. 8, 159-167. Sreenivasan, U. and Axelsen, P.H. (1992). Buried water in homologous serine proteases. Biochemistry, 31, 12785-12791. Stahl, U., Lee, M., Sjodahl, S., Archer, D., Cellini, F., Ek, B., Iannacone, R., MacKenzie, D., Semeraro, L., Tramontano, E. and Stymme, S. (1999). Plant low-molecular-weight phospholipase A2s (PLA2s) are structurally related to the animal secretory PLA2s and are present as a family of isoforms in rice (Oryza sativa). Plant Mol. Biol. 41, 481-490.

REFERENCES 244

Stallmeyer, B., Drugeon, G., Reiss, J., Haenni, A.L. and Mendel, R.R. (1999a). Human molybdopterin synthase gene: identification of a bicistronic transcript with overlapping reading frames. Am. J. Hum. Genet. 64, 698-705. Stallmeyer, B., Nerlich, A., Schiemann, J., Brinkmann, H. and Mendel, R.R. (1995). Molybdenum co-factor biosynthesis: the Arabidopsis thaliana cDNA cnx1 encodes a multifunctional two-domain protein homologous to a mammalian neuroprotein, the insect protein Cinnamon and three Escherichia coli proteins. Plant J. 8, 751-762. Stallmeyer, B., Schwarz, G., Schulze, J., Nerlich, A., Reiss, J., Kirsch, J. and Mendel, R.R. (1999b). The neurotransmitter receptor-anchoring protein gephyrin reconstitutes molybdenum cofactor biosynthesis in bacteria, plants, and mammalian cells. Proc. Natl. Acad. Sci. USA, 96, 1333-1338. Steiner, R.A., Rozeboom, H.J., de Vries, A., Kalk, K.H., Murshudov, G.N., Wilson, K.S. and Dijkstra, B.W. (2001). X-ray structure of bovine pancreatic phospholipase A2 at atomic resolution. Acta Cryst. D57, 516-526. Stephens, W.W.J. and Myers, W. (1898). The action of cobra poison on the blood: a contribution to the study of passive immunity. J. Pathol. Bacteriol. 5, 279–301. Sterner, R. and Liebl, W. (2001). Thermophilic adaptation of proteins. Crit. Rev. Biochem. Mol. Biol. 36, 39-106. Stevenson, C.E., Sargent, F., Buchanan, G., Palmer, T. and Lawson, D.M. (2000). Crystal structure of the molybdenum cofactor biosynthesis protein MobA from Escherichia coli at near-atomic resolution. Structure, 8, 1115-1125. Stiefel, E.I. (2002). The biogeochemistry of molybdenum and tungsten. Met. Ions Biol. Syst. 39, 1-29. Stocker, U. and van Gunsteren, W.F. (2000). Molecular dynamics simulation of hen egg white lysozyme: A test of the GROMOS96 force field against nuclear magnetic resonance data. Proteins: Struct. Funct. Bioinformatics, 40, 145-153. Strynadka, N.C. and James, M.N. (1989). Crystal structures of the helix-loop-helix calcium-binding proteins. Annu. Rev. Biochem. 58, 951-998. Su, M.J., Coulter, A.R., Sutherland, S.K. and Chang, C.C. (1983). The presynaptic neuromuscular blocking effect and phospholipase A2 activity of textilotoxin, a potent toxin isolated from the venom of the Australian brown snake, Pseudonaja textilis. Toxicon, 21, 143-151. Sumandea, M., Das, S., Sumandea, C. and Cho, W. (1999). Roles of aromatic residues in high interfacial activity of Naja naja atra phospholipase A2. Biochemistry, 38, 16290-16297.

REFERENCES 245

Sumathi, K., Ananthalakshmi, P., Roshan, M.N.A. and Sekar, K. (2006). 3dSS: 3D structural superposition. Nucl. Acids Res. 34, W128-W138. Taketo, M.M. and Sonoshita, M. (2002). Phospolipase A2 and apoptosis. Biochim. Biophys. Acta, 1585, 72-76. Tame, J.R.H., Sleigh, S.H., Wilkinson, A.J. and Ladbury, J.E. (1996). The role of water in sequence-independent ligand binding by an oligopeptide transporter protein. Nat. Struct. Biol. 3, 998-1001. Tao, H., Hasona, A., Do, P.M., Ingram, L.O. and Shanmugam, K.T. (2005). Global gene expression analysis revealed an unsuspected deo operon under the control of molybdate sensor, ModE protein, in Escherichia coli. Arch. Microbiol. 184, 225-233. Thompson, J.D., Higgins, D.G. and Gibson, T.J. (1994). CLUSTALW: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucl. Acids Res. 22, 4673-4680. Thony, B., Auerbach, G. and Blau, N. (2000). Tetrahydrobiopterin biosynthesis, regeneration and functions. Biochem. J. 347, 1-16. Thunnissen, M.M., Ab, E., Kalk, K.H., Drenth, J., Dijkstra, B.W., Kuipers, O.P., Dijkman, R., de Haas, G.H. and Verheij, H.M. (1990). X-ray structure of phospholipase A2 complexed with a substrate-derived inhibitor. Nature, 347, 689-691. Tjoelker, L.W., Wilder, C., Eberhardt, C., Stafforini, D.M., Dietsch, G., Schimpf, B., Hooper, S., Le Trong, H., Cousens, L.S., Zimmerman, G.A., Yamadat, Y., McLntyre, T.M., Prescott, S.M. and Gray, P.W. (1995). Anti-inflammatory properties of a platelet-activating factor acetylhydrolase. Nature, 374, 549-553. Tomasselli, A.G., Hui, J., Fisher, J., Zurcher-Neely, H., Reardon, I.M., Oriaku, E., Kezdy, F.J. and Heinrikson, R.L. (1989). Dimerization and activation of porcine pancreatic phospholipase A2 via substrate level acylation of lysine 56. J. Biol. Chem. 264, 10041-10047. Tomatsu, H., Takano, J., Takahashi, H., Watanabe-Takahashi, A., Shibagaki, N. and Fujiwara, T. (2007). An Arabidopsis thaliana high-affinity molybdate transporter required for efficient uptake of molybdate from soil. Proc. Natl. Acad. Sci. USA, 104, 18807-18812. Tomoo, K., Ohishi, H., Doi, M., Ishida, T., Inoue, M., Ikeda, K., Hata, Y. and Samejima, Y. (1992). Structure of acidic phospholipase A2 from the venom of Agkistrodon halys blomhoffii at 2.8 Å resolution. Biochem. Biophys. Res. Commun. 184, 137-143. Tomsett, A.B. and Garrett, R.H. (1980). The isolation and characterization of mutants defective in nitrate assimilation in Neurospora crassa. Genetics, 95, 649-660.

REFERENCES 246

Tsuboi, K., Sugimoto, Y. and Ichikawa, A. (2002). Prostanoid receptor subtypes. Prostaglandins Other Lipid Mediat. 68-69, 535-556. Turnlund, J.R. (2002). Molybdenum metabolism and requirements in humans. Met. Ions Biol. Syst. 39, 727-739. Unkles, S.E., Heck, I.S., Appleyard, M.V. and Kinghorn, J.R. (1999). Eukaryotic molybdopterin synthase. Biochemical and molecular studies of Aspergillus nidulans cnxG and cnxH mutants. J. Biol. Chem. 274, 19286-19293. Unkles, S.E., Smith, J., Kanan, G.J., Millar, L.J., Heck, I.S., Boxer, D.H. and Kinghorn, J.R. (1997). The Aspergillus nidulans cnxABC locus is a single gene encoding two catalytic domains required for synthesis of precursor Z, an intermediate in molybdenum cofactor biosynthesis. J. Biol. Chem. 272, 28381-28390. Valentin, E., Ghomashchi, F., Gelb, M.H., Lazdunski, M. and Lambeau, G. (2000). Novel human secreted phospholipase A2 with homology to the group III bee venom enzyme. J. Biol. Chem. 275, 7492-7496. van Deenen, L.L. and de Hass, G.H. (1964). The synthesis of phosphoglycerides and some biochemical applications. Adv. Lipid Res. 2, 167-234. van den Berg, B., Tessari, M., Boelens, R., Dijkman, R., de Haas, G.H., Kaptein, R. and Verheij, H.M. (1995). NMR structures of phospholipase A2 reveal conformational changes during interfacial activation. Nat. Struct. Biol. 2, 402-406. van den Bergh, C.J., Slotboom, A.J., Verheij, H.M. and de Haas, G.H. (1988). The role of aspartic acid-49 in the active site of phospholipase A2. A site-specific mutagenesis study of porcine pancreatic phospholipase A2 and the rationale of the enzymatic activity of [lysine49]phospholipase A2 from Agkistrodon piscivorus piscivorus' venom. Eur. J. Biochem. 176, 353-357. van der Spoel, D., Lindahl, E., Hess, B., Groenhof, G., Mark, A.E. and Berendsen, H.J.C. (2005). GROMACS: fast, flexible, and free. J. Comput. Chem. 26, 1701-1718. Verheij, H.M., Slotboom, A.J. and de Haas, G.H. (1981). Structure and function of phospholipase A2. Rev. Physiol. Biochem. Pharmacol. 91, 91-203. Verheij, H.M., Volwerk, J.J., Jansen, E.H., Puyk, W.C., Dijkstra, B.W., Drenth, J. and de Haas, G.H. (1980). Methylation of histidine-48 in pancreatic phospholipase A2. Role of histidine and calcium ion in the catalytic mechanism. Biochemistry, 19, 743-750. Verlet, L. (1967). Computer "Experiments" on Classical Fluids. I. Thermodynamical Properties of Lennard-Jones Molecules. Phys. Rev. 159, 98–103. Vieille, C. and Zeikus, G.J. (2001). Hyperthermophilic enzymes: sources, uses, and molecular mechanisms for thermostability. Microbiol. Mol. Biol. Rev. 65, 1-43.

REFERENCES 247

Wang, W., Donini, O., Reyes, C.M. and Kollman, P.A. (2001). BIOMOLECULAR SIMULATIONS: Recent Developments in Force Fields, Simulations of Enzyme Catalysis, Protein-Ligand, Protein-Protein, and Protein-Nucleic Acid Noncovalent Interactions. Ann. Rev. Biophys. Biomol. Struct. 30, 211-243. Wang, X.Q., Yang, J., Gui, L.L., Lin, Z.J., Chen, Y.C. and Zhou, Y.C. (1996). Crystal structure of an acidic phospholipase A2 from the venom of Agkistrodon halys pallas at 2.0 Å resolution. J. Mol. Biol. 255, 669-676. Warner, C.K. and Finnerty, V. (1981). Molybdenum hydroxylases in Drosophila. II. Molybdenum cofactor in xanthine dehydrogenase, aldehyde oxidase and pyridoxal oxidase. Mol. Gen. Genet. 184, 92-96. Webb, N.R. (2005). Secretory phospholipase A2 enzymes in atherogenesis. Curr. Opin. Lipidol. 16, 341-344. Wei, S., Ong, W.Y., Thwin, M.M., Fong, C.W., Farooqui, A.A., Gopalakrishnakone, P. and Hong, W. (2003). Group IIA secretory phospholipase A2 stimulates exocytosis and neurotransmitter release in pheochromocytoma-12 cells and cultured rat hippocampal neurons. Neuroscience, 121, 891-898. White, S.P., Scott, D.L., Otwinowski, Z., Gelb, M.H. and Sigler, P.B. (1990). Crystal structure of cobra-venom phospholipase A2 in a complex with a transition-state analogue. Science, 250, 1560-1563. Winget, J.M., Pan, Y.H. and Bahnson, B.J. (2006). The interfacial binding surface of phospholipase A2s. Biochim. Biophys. Acta. 1761, 1260-1269. Witte, C.P., Igeno, M.I., Mendel, R., Schwarz, G. and Fernandez, E. (1998). The Chlamydomonas reinhardtii MoCo carrier protein is multimeric and stabilizes molybdopterin cofactor in a molybdate charged form. FEBS Lett. 431, 205-209. Wuebbens, M.M., Liu, M.T., Rajagopalan, K. and Schindelin, H. (2000). Insights into molybdenum cofactor deficiency provided by the crystal structure of the molybdenum cofactor biosynthesis protein MoaC. Structure, 8, 709-718. Wuebbens, M.M. and Rajagopalan, K.V. (1993). Structural characterization of a molybdopterin precursor. J. Biol. Chem. 268, 13493-13498. Wuebbens, M.M. and Rajagopalan, K.V. (1995). Investigation of the early steps of molybdopterin biosynthesis in Escherichia coli through the use of in vivo labeling studies. J. Biol. Chem. 270, 1082-1087. Xiang, S., Nichols, J., Rajagopalan, K.V. and Schindelin, H. (2001). The crystal structure of Escherichia coli MoeA and its relationship to the multifunctional protein gephyrin. Structure, 9, 299-310.

REFERENCES 248

Xiong, L., Ishitani, M., Lee, H. and Zhu, J.K. (2001). The Arabidopsis LOS5/ABA3 locus encodes a molybdenum cofactor sulfurase and modulates cold stress- and osmotic stress-responsive gene expression. Plant Cell, 13, 2063-2083. Xu, S., Gu, L., Wang, Q., Shu, Y., Song, S. and Lin, Z. (2003). Structure of a king cobra phospholipase A2 determined from a hemihedrally twinned crystal. Acta Cryst. D59, 1574-1581. Yap, K.L., Ames, J.B., Swindells, M.B. and Ikura, M. (2002) Vector Geometry Mapping: a method to characterize the conformation of helix-loop-helix calcium binding proteins. Methods Mol. Biol. 173, 317-324. Yoshida, H., Yamada, M., Kuramitsu, S. and Kamitori, S. (2008). Structure of a putative molybdenum-cofactor biosynthesis protein C (MoaC) from Sulfolobus tokodaii (ST0472). Acta Cryst. F64, 589–592. Yu, B.Z., Berg, O.G. and Jain, M.K. (1993). The divalent cation is obligatory for the binding of ligands to the catalytic site of secreted phospholipase A2. Biochemistry, 32, 6485-6492. Yu, B.Z., Poi, M.J., Ramagopal, U.A., Jain, R., Ramakumar, S., Berg, O.G., Tsai, M.-D., Sekar, K. and Jain, M.K. (2000). Structural basis of the anionic interface preference and kcat* activation of pancreatic phospholipase A2. Biochemistry, 39, 12312-12323. Yu, L., Deems, R.A., Hajdu, J. and Dennis, E.A. (1990). The interaction of phospholipase A2 with phospholipid analogues and inhibitors. J. Biol. Chem. 265, 2657-2664. Yuan, C. and Tsai, M. (1999). Pancreatic phospholipase A2: new views on old issues. Biochim. Biophys. Acta. 1441, 215-222. Yuan, W. and Gelb, M.H. (1988). Phosphonate-containing phospholipid analogs as tight-binding inhibitors of phospholipase-A2. J. Am. Chem. Soc. 110, 2665-2666. Zadori, Z., Szelei, J., Lacoste, M.C., Li, Y., Gariepy, S., Raymond, P., Allaire, M., Nabi, I.R. and Tijssen, P. (2001). A viral phospholipase A2 is required for parvovirus infectivity. Dev. Cell, 1, 291-302. Zhang, L., Wang, L., Kao, Y.-T., Qiu, W., Yang, Y., Okobiah, O. and Zhong, D. (2007). Mapping hydration dynamics around a protein surface. Proc. Natl. Acad. Sci. USA, 104, 18461-18466. Zhang, N.Z. and Matthews, B.W. (1994). Conservation of solvent-binding sites in 10 crystal forms of T4 lysozyme. Protein Sci. 3, 1031-1039.

REFERENCES 249

Zhang, W., Urban, A., Mihara, H., Leimkuhler, S., Kurihara, T. and Esaki, N. (2010). IscS functions as a primary sulfur-donating enzyme by interacting specifically with MoeB and MoaD in the biosynthesis of molybdopterin in Escherichia coli. J. Biol. Chem. 285, 2302-2308. Zhang, Y. and Gladyshev, V.N. (2008). Molybdoproteomes and evolution of molybdenum utilization. J. Mol. Biol. 379, 881-899. Zhao, K., Song, S., Lin, Z. and Zhou, Y. (1998). Structure of a basic phospholipase A2 from Agkistrodon halys Pallas at 2.13 Å resolution. Acta Cryst. D54, 510-521. Zhao, K., Zhou, Y. and Lin, Z. (2000). Structure of basic phospholipase A2 from Agkistrodon halys Pallas: implications for its association, hemolytic and anticoagulant activities. Toxicon, 38, 901-916. Zhu, H., Dupureur, C.M., Zhang, X. and Tsai, M.-D. (1995). Phospholipase A2 engineering. The roles of disulfide bonds in structure, conformational stability, and catalytic function. Biochemistry, 34, 15307-15314.

Documents

Structural Studies on Bovine Pancreatic Phospholipase · PDF fileStructural Studies on Bovine Pancreatic Phospholipase A 2 and Proteins involved in Molybdenum Cofactor Biosynthesis