RESEARCH

Unscrambling the Effect of C-Terminal Tail Deletionon the Stability of a Cold-Adapted, Organic Solvent Stable Lipasefrom Staphylococcus epidermidis AT2

Nor Hafizah Ahmad Kamarudin • Raja Noor Zaliha Raja Abd Rahman •

Mohd Shukuri Mohamad Ali • Thean Chor Leow • Mahiran Basri •

Abu Bakar Salleh

� Springer Science+Business Media New York 2014

Abstract Terminal moieties of most proteins are long

known to be disordered and flexible. To unravel the func-

tional role of these regions on the structural stability and

biochemical properties of AT2 lipase, four C-terminal end

residues, (Ile–Thr–Arg–Lys) which formed a flexible, short

tail-like random-coil segment were targeted for mutation.

Swapping of the tail-like region had resulted in an

improved crystallizability and anti-aggregation property

along with a slight shift of the thermostability profile. The

lipolytic activity of mutant (M386) retained by 43 %

compared to its wild-type with 18 % of the remaining

activity at 45 �C. In silico analysis conducted at 25 and

45 �C was found to be in accordance to the experimental

findings in which the RMSD values of M386 were more

stable throughout the total trajectory in comparison to its

wild-type. Terminal moieties were also observed to exhibit

large movement and flexibility as denoted by high RMSF

values at both dynamics. Variation in organic solvent sta-

bility property was detected in M386 where the lipolytic

activity was stimulated in the presence of 25 % (v/v) of

DMSO, isopropanol, and diethyl ether. This may be worth

due to changes in the surface charge residues at the

mutation point which probably involve in protein–solvent

interaction.

Keywords C-terminal region � Deletion �Crystallizability � Thermostability � Organic solvent

stability

Introduction

Proteins in general are built of the main globular structural

domain and regions that lack defined secondary structure

called disordered regions. Often, these non-structured

regions are missing in the electron density map due to their

inherent flexibility and motion which limit the scattering of

X-rays but notably discovered to be important in protein

function and folding [1]. A number of works have been

devoted to understand the role of disordered region in

different aspects including its involvement in promoting

large assembly, signal recognition, and its influence on

structural stability.

Proteins with high number of unstructured and flexible

segments such as long loop and coil are generally postu-

lated to be less stable and often associated with low-tem-

perature tolerance. Increased conformational flexibility has

been known as an essential hallmark in cold-adapted

enzymes to facilitate an easier accommodation of sub-

strates at cold temperature [2]. In contrast, reduced number

of disordered flexible regions which consequently con-

tribute to high conformational rigidity has been regarded as

one of the key structural feature for thermostable enzymes

[3]. Apart from the influence of these segments on struc-

tural stability, the occurrence of flexible regions in proteins

is known to impede crystal formation—a primary pre-

requisite in X-ray crystallography—by creating

N. H. A. Kamarudin � R. N. Z. R. A. Rahman (&) �M. S. M. Ali � T. C. Leow � M. Basri � A. B. Salleh

Enzyme and Microbial Technology Research Centre, Universiti

Putra Malaysia, 43400 Serdang, Selangor, Malaysia

e-mail: rnzaliha@upm.edu.my

N. H. A. Kamarudin � R. N. Z. R. A. Rahman �M. S. M. Ali � T. C. Leow � A. B. Salleh

Faculty of Biotechnology and Biomolecular Sciences, Universiti

Putra Malaysia, 43400 Serdang, Selangor, Malaysia

M. Basri

Faculty of Science, Universiti Putra Malaysia, 43400 Serdang,

Selangor, Malaysia

123

Mol Biotechnol

DOI 10.1007/s12033-014-9753-1

microheterogeneity and interfering with the crystal packing

[4]. Clipping of the flexible regions has been proven to

facilitate the crystal formation of many proteins and now

seemingly to become a routine practice in crystal

engineering.

Of many parts of the proteins, terminal moieties are

commonly regarded to be highly flexible, surface

exposed, and frequently observed to be disordered. As

for some proteins, surface-exposed terminal is important

for the interaction with water and polar solvents, while

others involve in post-translational process [5]. On the

downside, disordered termini have been observed to

induce aggregation, precipitation, and sample instability

[6]. Modification of the terminal moieties has been

considerably investigated which resulted in changes in

biological function, biochemical properties, as well as

crystallization propensity. Fernando et al. [7], for

instance, had investigated on the C-terminal end trun-

cation of a small heat shock protein, hsp30. The trun-

cation hampered its chaperone activity and led to

disruption of the secondary structure. Bovine poly

(A) polymerase was subjected to C-terminal deletion of

226 residues and the resulting outcome showed the for-

mation of high-quality crystals [8]. Also in many cases,

removal of N- and C-terminal has become a standard

protocol in expression studies [4].

Previously, a lipase gene from Staphylococcus epide-

rmidis AT2 (termed as AT2 lipase) was cloned and

expressed in Escherichia coli system [9]. It was later

identified as a cold-adapted alkaline lipase which exhibited

optimal activity at 25 �C and displayed total deactivation at

50 �C. The lipolytic activity was enhanced in the presence

of long-chain triglycerides and natural oils. In addition, it

showed excellent stability in various organic solvents

namely 25 % (v/v) DMSO, methanol, ethanol, acetone,

diethyl ether, toluene, and n-hexane. This hydrolytic

enzyme was also found prone to aggregation and recalci-

trant to crystallization (data not shown). Based on these

findings, it is hypothesized that this enzyme confers high

conformational flexibility, in which the presence of

unstructured segments could be one of the contributing

factors. Such assumption is made through the notion that

flexibility is the dominating signature of heat-labile

enzymes. Preliminary investigation on the amino acid

composition agrees that AT2 lipase is intrinsically flexible

as denoted by high number of glycine residues (11 % out

of total amino acids), lack of disulfide bonds, and low

number of proline and arginine residues. In this study, we

attempted to modify the flexible disordered region, mainly

targeted on the terminal segment of AT2 lipase. The effects

of the C-terminal truncation on the enzyme stability, bio-

chemical properties, and crystallization propensity were

investigated.

Materials and Methods

Construction of Wild-Type AT2 Lipase and Its

Truncated Mutant, M386

A 1.2 kb of DNA fragment encoding for the mature form

of AT2 lipase was generated by PCR amplification using

two primers: forward, 50 CG GAA TTC GCA GCA GCA

ATG GCG CAA GCT CAA TAT AAA AAT 30 and

reverse, 50 CCG CTC GAG TCA CTA ATG ATG ATG

ATG ATG ATG CTT ACG TGT GAT ACC ATC TAG 30.The amplified fragment was cloned in pGEX-6p-1 vector

(GE Healthcare, UK) at EcoR1 and Xho1 sites (bold). This

vector encodes for a prescission protease cleavage site for

the separation of target protein and glutathione-S-transfer-

ase tag (GST) during purification. Mutant M386 was con-

structed by incorporating a stop codon immediately after

Gly386, downstream the reverse primer 50 CCG CTC

GAG TCA CTA ACC ATC TAG CTC TTC GAC TTT

CAA 30 (stop codon—underlined, XhoI site in the primer is

indicated in bold), and subjected to PCR amplification. The

PCR product was cloned in the same vector and trans-

formed into chemically competent E. coli DH5a (Novagen,

Germany). Positive transformants were analyzed through

single, double digestion, and colony PCR. The gene

encoding for the lipolytic enzyme was confirmed correct

through sequencing. Both recombinant pGEX-6p-1/AT2

and pGEX-6p-1/M386 were transformed into chemically

competent E. coli BL21 (DE3) (Novagen, Germany) for

over-expression.

Expression and Purification of Wild-Type AT2 Lipase

and Its Truncated Mutant, M386

Single colony of E. coli BL21 (DE3) harboring pGEX-6p-

1/AT2 and pGEX-6p-1/M386 was grown in 10 mL LB

broth supplemented with 50 lg/mL of ampicillin and

incubated at 37 �C overnight. Five mL of the overnight

culture was further cultivated in 500 mL of LB broth in 5-L

shake flask for 1 L bacterial culture preparation. Cells were

cultivated at 37 �C for 2–3 h, induced with 0.5 mM IPTG

at OD600 = 0.6, and further incubated at 16 �C for 16 h.

Cells were then harvested at 5,0009g for 15 min, at 4 �C.

Bacterial pellet was resuspended in pre-chilled binding

buffer (50 mM phosphate buffer saline pH 7.4) and lysed

by sonication using Branson Sonifier 250 (Branson, USA).

Sample was cleared by centrifugation at 10,0009g for

30 min, 4 �C. Supernatant which contained the soluble

fraction was loaded into first affinity chromatography using

Glutathione Sepharose High Performance column, GST HP

(GE Healthcare, UK) at a flow rate of 1 mL/min. The

column was continuously flowed with binding buffer until

A280 readings reached baseline. The bound protein was

Mol Biotechnol

123

then directly eluted using a buffer composed of 50 mM

Tris–HCl and 10 mM reduced glutathione at pH 8 with a

flow rate of 3 mL/min. Fractions containing target protein

were pooled and cleaved by prescission protease cleavage

at 4 �C for 16 h in 50 mM Tris–HCl pH 8, 100 mM NaCl

cleavage buffer. The cleaved fractions were separated by a

second GST column following the same procedures

described above. In this second GST affinity, target pro-

teins were collected in the flow through fractions and

analyzed for purity and activity.

SDS-PAGE and Native PAGE Analysis

Purified proteins were examined through SDS-PAGE ana-

lysis which was carried out according to Laemmeli’s method

[10] using 12 % of resolving gel. The molecular weight of

the protein samples was determined through comparison

with a standard protein marker (Thermo Scientific, USA).

Native PAGE was conducted using non-denaturing gel, and

protein sample was prepared in non-reducing, non-denatur-

ing sample buffer. Gel was electrophoresed using appropri-

ate voltage and stained with Coomassie Blue.

Lipase Activity Assay

Lipase assay was performed according to Kwon and Rhee

[11] method with slight modification. Reaction mixture

containing 1 mL of enzyme, 2.5 mL of olive oil emulsion,

and 20 lL of 20 mM CaCl2 was incubated at 25 �C for

30 min with 200 rpm agitation. The enzyme reaction was

then terminated using 1 mL of 6 N HCl and 5 mL of iso-

octane, which later resulted in the formation of two layers.

Four milliliters of the upper layer was drawn out and added

into test tubes containing 1 mL of 5 % (w/v) cupric acetate

pyridine pH 6.1. The mixture was mixed vigorously and

allowed to settle for color development. The absorbance

reading was measured at 715 nm, and lipase activity was

determined by oleic acid standard curve. One unit of

activity is defined as the rate of free fatty acids released in

1 min. For all activity measurements, the experiments were

performed in triplicates.

Thermostability Study

Purified rWT-AT2 and M386 were pre-incubated for

30 min at different temperatures ranging from 10 to 50 �C

with 5 �C interval. The treated enzymes were then sub-

jected to lipase assay at 25 �C (optimal temperature) for

30 min. The residual activity was measured against control

reaction (enzyme without pre-incubation) which was

regarded as 100 %.

Organic Solvent Stability Study

Fifteen organic solvents with different log P were

employed for organic solvent stability study on recombi-

nant wild-type AT2 lipase and its mutant, M386. Organic

solvents involved were DMSO (-1.3), methanol (-0.76),

acetonitrile (-0.33), ethanol (-0.24), acetone (-0.24),

1-propanol (0.28), isopropanol (0.28), ethyl acetate (0.68),

diethyl ether (0.85), chloroform (2.0), benzene (2.0), tolu-

ene (2.5), octanol (2.9), n-hexane (3.5), and n-heptane

(4.0). The enzyme was pre-incubated with organic solvent

at 25 % (v/v) for 30 min, with 150 rpm agitation, at 20 �C.

Lipase assay was subsequently conducted at 25 �C, and the

lipolytic activity was determined in relative to the activity

of the non-treated enzyme (enzyme in aqueous solution).

Structure Modeling of AT2 Lipase

Structure prediction of AT2 lipase was conducted using

automated homology modeling by a program called Yet

Another Scientific Artificial Reality Application (YA-

SARA) [12]. Template search was performed by aligning

the target sequence with other bacterial lipases using PSI-

BLAST from the National Center for Biotechnological

Information (NCBI) (http://www.ncbi.nlm.nih.gov/blast)

based on PDB database. Top homologs were ranked based

on alignment score and structure quality. S. hyicus lipase

(PDB ID: 2HIH) with sequence identity of 51 % was

selected and used as the template. Backbone, loop, and side

chains were built, optimized, and fine-tuned. Simple min-

imization was carried out using AMBER03 force field.

Model was refined in explicit solvent molecules, and the

quality of the final model was evaluated using Rama-

chandran Plot (www.ebi.ac.uk/thornton-srv/software/PRO

CHECK), Errat2 http://nihserver.mbi.ucla.edu/ERRATv2

and Verify 3D (http://nihserver.mbi.ucla.edu/Verify_3D).

In silico Analysis of Wild-Type AT2 Lipase (rWT-

AT2) and M386

MD simulation was conducted using YASARA program

[13]. Both rWT-AT2 and M386 models were simulated at

two temperatures, 298 K (25 �C) and 318 K (45 �C).

Simulation was carried out in explicit water environment at

constant pressure using AMBER03 force field in a periodic

cell boundary condition with a total trajectory of 12 and

10 ns for 25 and 45 �C, respectively. A 7.86 A cutoff was

assigned to all atoms for non-bonded interaction. Simula-

tion snapshots were saved every 25 ps. The root mean

square deviation and fluctuation (RMSD and RMSF) val-

ues were analyzed for each simulation conducted.

Mol Biotechnol

123

Results and Discussions

Structure Modeling

In the absence of crystal structure, structural information of

AT2 lipase was obtained from a predicted model built by a

modeling program called YASARA. Template recognition

was initially performed by identifying its homolog struc-

ture from Protein Data Bank (PDB). The search was carried

out by a dynamic well-known open program, PSI-BLAST

(http://www.ncbi.nlm.nih.gov/blast/) which aligned the

query sequence with other sequences from the PDB data-

base. The primary sequence of AT2 lipase which com-

prised 390 residues was subjected to PSI-BLAST analysis,

and the top ten homologs were identified.

Of the ten homologs, crystal structure from S. hyicus

lipase (PDB ID: 2HIH) showed the highest sequence

identity to AT2 lipase (51 %) and query coverage of 98 %.

Others, in contrast, displayed relatively poor sequence

identity to AT2 lipase with percentage identity of below

36 %. Lacking of other staphylococcal lipase crystal

structures, S. hyicus lipase crystal structure serves as the

most reliable template for the prediction of AT2 lipase

structure. The secondary structure was first predicted to

assist in the alignment correction between the selected

template and for further loop construction. Side chains

were added, and the rotamers were optimized by employ-

ing different hundreds conformation in fine tuning the best

conformation. The model was subjected to energy mini-

mization using AMBER03 force field.

The quality of the final AT2 lipase model was assessed

by PROCHECK (Ramachandran Plot), Verify 3D and

Errat2 (Table 1). Based on the output obtained from Ra-

machandran plot analysis, no residues were resided in the

disallowed region. The most favored region comprised 304

residues which corresponded to 91.8 %, while 7.9 % of the

residues were localized in the additionally allowed region.

Typically, catalytic serine (Ser116) was detected in the

generously allowed region where similar observation has

been obtained in other lipases [14, 15]. The active serine is

located in a so-called ‘‘nucleophilic elbow,’’ a tightly

constrained turn between an alpha helix and a beta strand.

The overall quality of the model as determined by Errat2

was 96.809, while Verify 3D profile showed that the scores

fall within the range of correctly folded protein. No major

errors were detected indicating that there was no mis-

alignment that would cause unusual contacts. Taken toge-

ther all the validation assessments, the quality of the final

AT2 lipase model is of satisfactory and deemed to be

highly reliable.

Predicted Structure of AT2 Lipase

AT2 lipase adopted the typical a/b hydrolase fold similar

to that displayed by other lipases. The heart-shape like

structure comprises nine b-strands in which five of the bstrands buried in the central hydrophobic core surrounded

by fourteen a-helices (a1 to a14). As predicted by multiple

sequence alignment, Ser116, Asp307, and His349 are in the

catalytic configuration, buried in the protein core. The

model was predicted in an open conformation as the tem-

plate used for modeling, S. hyicus lipase (2HIH) was elu-

cidated in similar form. Two a-helices (a8 and a9)

corresponding to residue 181–198 and a stretch from res-

idue 218–229 are likely to form the so-called lid region

covering the active site (Fig. 1). The lid region which

involves in interfacial activation represents the unique

structural feature observed in most lipases.

Two metal-binding sites were also identified in AT2

lipase model. Ca2? ion was held by a pentahedral coordi-

nation composed of Asp348, Asp351, Asp356, Asp359,

and Arg283. Among the five Ca2? binding residues,

Asp348 and Asp351 were localized adjacent to the loop

holding the catalytic His349. A second metal-binding site

found on the model was proposed as Zn2? binding site. The

Zn2? ion was bound by the side chain residues in a tetra-

hedral coordination comprising Asp64, Tyr80, His90, and

Asp236. The binding distance was 2–3 A. Based on 2HIH

crystal structure, Zn2? exhibited stabilizing role in the

structural conformation as it is localized in between several

helices and loops keeping two helices (a3 and a4) in their

correct position and orientation [16].

Identification of Mutation Point

The predicted model was superimposed with its template

structure, 2HIH, to identify any structural difference par-

ticularly on the presence of unstructured, flexible loop

region that may be germane to modify. Apparently, major

dissimilarity between the two structures was spotted at the

terminal regions. Both regions were highly surface exposed

and seem to be distantly apart for any plausible interaction

to occur between the two termini. N-terminal of 2HIH

Table 1 Validation results for the structure quality of AT2 lipase

model

Validation tool Predicted structure

Ramachandran plot

a) Most favored region 91.6 %

b) Additional allowed region 7.9 %

c) Generously allowed region 0.3 %

d) Disallowed region 0.0 %

Errat2 (overall quality factor) 96.809

Verify 3D (3D-ID) 0.0–0.79

Mol Biotechnol

123

crystal structure which started at residue Ala7 constituted

3-residues longer than that found in AT2 lipase (Fig. 2).

Meanwhile, the C-terminal region of AT2 lipase formed a

short tail-like extension on the long 22-residues terminal a-

helix. The tail-like structure composed of Ile387, Thr388,

Arg389, and Lys390 in which Lys390, a bulky charged

residue, protrudes outward to the surface, making no con-

tacts and bonds with adjacent residues. Such observation

was lacking in 2HIH, suggesting that modification of this

region could be promising to stabilize the enzyme. High

mobility and movement of this tail-like region are likely to

reduce the conformational rigidity, interfere with the

crystallization process, and promote larger assembly of

proteins which leads to aggregation. Consistent with the

predicted model, analysis from an open access server called

XtalPred (http://ffsd.burham.org/XtalPred-cgi/xtal.pi) sug-

gested that the C-terminal segment of AT2 lipase,

LDGITRK, constituted of a disordered region. The server

is designed to facilitate prediction of protein crystalliz-

ability which includes determination of disordered region.

On this relevance, the C-terminal tail extension was tar-

geted for mutation.

Expression and Purification of rWT-AT2 and M386

For C-terminal deletion mutant construction, stop codons

(TAA TGA) were incorporated immediately after Gly386

generating a truncated lipase gene containing 386 residues

(termed as M386). The newly constructed mutant was

expressed in E. coli BL21 (DE3). Expression study

Fig. 1 Predicted structure of

AT2 lipase in an open

conformation. This enzyme

adopted a typical a/b hydrolase

fold. Catalytic triad is

positioned at the center core of

the protein and shown in circle.

A close view of the active site is

denoted on the bottom right

figure. Potential lid regions are

colored rosy brown. Metal-

binding sites are indicated

through heteroatom colors.

Calcium and zinc ions are

colored green and gray,

respectively. Terminal regions

are labeled accordingly. Figure

is generated using Chimera

1.5.3 (Color figure online)

Fig. 2 Superimposition of AT2

lipase predicted structure

(green) and S. hyicus lipase

crystal structure, 2HIH (pink).

The RMSD value for the

superimposed structures is

0.554 A over 378 matched

atoms. Both N- and C-terminal

moieties display variation in

length and type of residues. AT2

lipase confers a tail-like

extension at the C-terminal

helix (bottom right) and a

slightly shorter N-terminal coil

(up right) compared to 2HIH.

Figure is generated using

Chimera 1.5.3 (Color figure

online)

Mol Biotechnol

123

demonstrated that the lipolytic activity of both mutant and

wild-type was comparable (data not shown). This pre-

liminary observation suggested that the truncation per-

formed did not cause any major effect to neither the

catalytic site nor the structural folding. It also indicated that

no alteration occurred in the active site; hence, the sub-

strate preference profile of M386 remained similar to that

of wild-type. Wild-type AT2 lipase (rWT-AT2) and its

mutant, M386, were purified in the same manner by

undergoing affinity chromatography using GST HP column

followed by cleavage of the fusion tag and separation of the

target protein from the GST tag by a second GST affinity

column. The purity of the proteins was examined by SDS-

PAGE and Native PAGE analysis. SDS-PAGE analysis

was of satisfactory as purification of the newly constructed

mutant and its wild-type resulted in a pure single band at

the final stage of purification (Fig. 3a).

Effect of Truncation on Solubility of AT2 Lipase

On the basis of Native PAGE analysis, a significant result

was obtained where a sharp distinct band was clearly

viewed on the gel as denoted in Fig. 3b. Such observation

was not obtained when rWT-AT2 was subjected to Native

PAGE analysis. In the previous work, rWT-AT2 had

encountered restricted mobility in Native PAGE where in

most of the trials, only smearing appeared on the gel.

Native PAGE separation is based on the surface charge of

the protein and the hydronomic size [17]. Low mobility of

bands in Native PAGE corresponds to large conformational

state of the protein, or simply, the protein exists in the form

of oligomer or aggregates. Presumably, rWT-AT2 is prone

to aggregation though being identified as monomeric pro-

tein (data not shown). Incomplete solubility of protein will

often form smear on native gel [18]. The significant

improvement in mobility and separation on Native PAGE

analysis of M386 suggested that the C-terminal tail region

of AT2 lipase influenced its native conformational state. In

many cases of complex proteins, disordered region is

essential to serve as linker to form large assembly of pro-

tein [19, 20]. It can, therefore, be deduced that in the

absence of the flexible disordered tail region of AT2 lipase,

no intermolecular interaction occurred between the

monomeric species, and hence providing a more stabilized

conformational state and almost a complete protection

from aggregation. Similar phenomenon was also monitored

in porcine ab-crystallin in which truncation of 10 residues

or less at the C-terminal extension segment improved its

anti-aggregation property [21]. A marginal shift in iso-

electric point (pI) value from 6.65 to 6.43 was also denoted

in M386 mutant through removal of the four end residues,

including Lys390, a surface-exposed charged amino acid.

Dissimilarity in surface charge distribution may also be

responsible to the variation in Native PAGE profile

between M386 and rWT-AT2.

Effect of Truncation on AT2 Lipase Crystallizability

Attempts to crystallize AT2 lipase (rWT-AT2) had resulted

in no promising leads (data not shown). It was presumed

that the inherent flexibility of the enzyme was the major

drawback for crystal formation. To probe the effect of

C-terminal tail deletion in promoting the crystallizability of

AT2 lipase, crystal screenings were conducted using sitting

drop vapor diffusion method at protein concentration of

4.5 mg/mL. The resulting outcomes were somewhat

promising. Positive hits were obtained in two formulations

from PEG/Ion Screen and Index Screen (Hampton

Research, USA) after approximately 40 days of incubation

at 10 �C. Clusters of needle-like crystals were observed in

Formulation 46 comprising 0.2 M sodium citrate tribasic

and 20 % (v/v) PEG 3350 (PEG/Ion Screen) and Formu-

lation 87 of Index Screen which composed of 0.2 M

sodium malonate pH 7 and 20 % PEG 3350. In a study

performed by McPherson [22], sodium malonate had sig-

nificantly emerged as a salt of choice for crystallization of

various types of proteins. This salt exhibits high solubility

in water and is deduced to stabilize the protein and local

water structure. Interestingly, both formulations contain

identical precipitant, PEG 3350 suggesting that this pre-

cipitant is important in inducing the growth of M386

crystal. Figure 4 depicts the crystal-grown wells of M386.

Fig. 3 SDS-PAGE and Native PAGE analysis of M386 and its wild-

type. a SDS-PAGE of pure fraction of rWT-AT2 and M386 after the

final step of purification (GST affinity), lane 1 Prestained Protein

Ladder (Thermo, USA), lane 2 purified rWT-AT2, lane 3 purified

M386 and b Native PAGE comparative examination of the purified

lipase of rWT-AT2 (lane 1) and its mutant, M386 (lane 2). Native gel

was prepared using the normal Tris–Glycine [10]. For comparison

study, 10 lL of the protein sample at final concentration of 0.4 mg/

mL was loaded into the well. Arrows indicate the target protein band

Mol Biotechnol

123

It is long known that in most proteins, the N- and C-ter-

minal regions are often disordered and flexible [23]. This

lead to potential increased in entropic impediment for

crystallization and problem associated with multiple

intermolecular contacts which resulted in multiple copies

of target protein in the asymmetric unit [24]. A dramatic

improvement in crystallization can be achieved through

minimization of local disorder [25]. Interestingly, the

removal of unstructured C-terminal region of AT2 lipase

agreed with the hypothesis that in the absence of this

segment, AT2 lipase exhibited greater propensity to pro-

duce crystal. Deletion of these residues probably left a

more stable core of AT2 lipase with improved anti-aggre-

gation prone property. Diffraction study was performed to

further verify the crystal hits as protein crystals. Simple

scan was conducted by in-house BRUKER X8-Proteum

single-crystal diffractometer system (BRUKER, Germany);

however, poor diffraction spots were obtained. Regardless

to the weak intensity spots of M386 crystals, this result

demonstrated a positive indication of that diffracted crys-

tals were protein crystals albeit not suitable for structure

determination.

Effect of Truncation on the Thermostability Profile

of AT2 Lipase

AT2 lipase was earlier identified as a cold-adapted enzyme

which was remarkably active at temperature below 30 �C.

Apart from the effect of C-terminal tail swapping on

crystallizability, removal of disordered, flexible region

could synergistically increase protein stability and rigidity.

Enzyme stabilization is one of the key factors to improve

crystallization while conformational rigidity is often asso-

ciated with enhanced thermostability. On this relevance,

the thermostability profile of M386 as compared to its

wild-type was investigated and shown in Fig. 5.

Both M386 and the rWT-AT2 were pre-incubated at

temperature ranging from 10 to 50 �C with 5 �C interval

for 30 min and subsequently underwent lipase activity

assay. Generally in both cases, the lipolytic activity

decreased as the temperature increased. M386 and rWT-

AT2 displayed optimal stability at 20 �C. At lower tem-

perature, 10 and 15 �C, a slight reduction in lipolytic

activity by 30 % was detected in rWT-AT2, but not in

M386. On the other hand, at 45 �C, M386 exhibited greater

stability than the wild-type. The residual activity remained

by 43 %, whereas in the wild-type, only 18 % of the

activity retained. At 50 �C, a total inhibition of lipolytic

activity was observed in rWT-AT2, while 24 % of the

residual activity was measured in M386. This result indi-

cates a slight improvement in thermal stability profile of

AT2 lipase. The absence of the flexible tail-like random

coil is likely to contribute to increased molecular com-

pactness and stabilization of the C-terminal helix which

consequently drives the shift in its thermostability

Fig. 4 Positive crystal hits

obtained from C-terminal

deletion mutant, M386.

a Crystal-grown well in PEG/

Ion Screen (Formulation 46)

and b crystal-grown well in

Index Screen (Formulation 87).

Both showing clusters of thin

needle-like crystals obtained in

sitting drop method. The size of

the crystals is 83.7 9 1.6 lm

for (a) and 259 9 1.6 lm for

(b). Arrows indicate the

measurement of size taken of

the respective single thin needle

crystal

Fig. 5 Thermal stability profile of rWT-AT2 (-r-) and its mutant,

M386 (-j-). Purified enzyme was pre-incubated for 30 min at

different temperatures prior to lipase assay at 25 �C. Lipolytic

activities were measured in relative to the activity of enzyme without

pre-incubation. Error bars represent standard deviations of means

(n = 3). The absence of error bars is due to larger size of symbols

Mol Biotechnol

123

behavior. Improved conformational compactness by

shortening of loop and reducing number of flexible disor-

dered region are regarded as factors enhancing thermosta-

bility [26]. Alteration on such segment has been employed

as one of the routes to improve thermostability of enzyme.

For example, deletion of a single C-terminal disordered

residue of xylanase from Aspergillus niger showed an

improvement in its thermostability by exhibiting a 6 �C

higher in the optimal temperature and threefold longer in

stability compared to its wild-type [27].

Together with thermal stability studies, thermal dena-

turation of both rWT-AT2 and M386 (data not shown) was

conducted using circular dichroism spectrophotometer,

JASCO J810-spectrapolarimeter (JASCO, Japan). Mea-

surement was conducted at 222 nm for temperature rang-

ing from 10 to 70 �C. A 0.1-cm path-length quartz cuvette

(Hellma, Concord, ON, Canada) was used for the study.

The midpoint of the unfolding transition (Tm) of 25 �C was

measured at 222 nm. The range of denaturation of both

rWT-AT2 and M386 was set to 48–60 �C, and the resulting

Tm were 56.89 and 56.27 �C, respectively. No significant

alteration in Tm behavior was observed between the two

samples, suggesting that removal of C-terminal residues

did not cause a significant change in the global confor-

mational state of AT2 lipase.

In silico Analysis of rWT-AT2 and M386 at Low-

and High-Temperature Simulation

In order to gain both structural insights into the natural

dynamics of rWT-AT2 and M386, and in silico analysis of

the mutagenesis study, molecular dynamic (MD) simula-

tion was carried out. MD simulation provides detailed

access on protein behavior at different time scales at which

not achievable by experimental work [28]. MD simulation

was performed in parallel with the experimental work, in

order to assure that mutagenesis did not cause any unnec-

essary distortion in AT2 lipase structure. Simulation was

conducted for both recombinant wild-type AT2 lipase and

its mutant (M386) at two different temperatures: 298 K

(25 �C) and 318 K (45 �C). Other parameters were made

constant. After 10 ns and 12 ns trajectories, the resulting

outcomes were analyzed by means of Ca root-mean-square

deviation (RMSD) and fluctuation (RMSF).

Analysis of MD Simulation by RMSD

RMSD measures the average scalar distance of atoms

between two structures. The deviation of the Ca backbone

from the initial structure as a function of time was calcu-

lated as a measurement of structural flexibility. As shown

in Fig. 6a, at 25 �C, during the beginning of the trajectory,

M386 conferred higher RMSD value (1.7 A) compared to

rWT-AT2 (1.3 A). This suggested that at this point where

equilibration and thermalization occurred, M386 exhibited

greater movement in Ca backbone compared to its wild-

type. However, the RMSDs of M386 remained in a stable

pattern throughout the whole trajectory (t = 12 ns) with a

deviation value of 2 A. Conversely, rWT-AT2 showed a

gradual increment in RMSD value by time. At 7–8 ns, the

RMSD value increased to 1.9 A, while at 10 ns, the

increment continued and resulted in RMSD value to reach

2.3 A. It is deduced that at a longer dynamic, the RMSDs

of rWT-AT2 will gradually increase. From this result, it

can be concluded that M386 displayed greater stability than

rWT-AT2 lipase as denoted by the small deviation in

RMSD value throughout the simulation at 25 �C.

As determined experimentally, M386 possessed the

ability to conserve its lipolytic activity by 43 % at 45 �C.

To obtain mechanistic details of the enzyme behavior,

simulation was conducted at 45 �C. Similarly, the RMSD

values of both rWT-AT2 and M386 during the overall

trajectory (t = 10 ns) were analyzed and compared

(Fig. 6b). M386 displayed higher RMSD value (2.0 A)

during the early stage of trajectory (t = 2 ns) than the

rWT-AT2. The deviation, however, remained stable until

the end of the simulation. The dynamic of rWT-AT2, on

the other hand, showed a smaller deviation at the beginning

Fig. 6 RMSD of the Ca backbone position of rWT-AT2 and M386.

a The RMSD values at 25 �C dynamic as measured for 12 ns

trajectory and b The RMSD profile at 45 �C simulation as analyzed

over 10 ns trajectory. rWT-AT2 and M836 are indicated by gray line

and black line, respectively

Mol Biotechnol

123

(1.8 A) but the RMSD value increased continuously along

the trajectory reaching to a value of 3.3 A at 10 ns. M386

exhibited lesser Ca backbone movement at 45 �C and most

likely the structural conformation remained stable. In

contrast, rWT-AT2 exhibited greater conformational

changes at 45 �C. Thermal vibration caused the structural

elements to confer high motion and consequently triggered

the unfolding and denaturation of the protein.

In general comparison between the two simulations per-

formed, higher temperature induced a larger deviation in

RMSD value of AT2 lipase structure. This observation is in

line with the property of AT2 lipase having good stability at

low temperature but rapidly undergoes denaturation at high

temperature. Deletion of the C-terminal tail residues (Ile387,

Thr388, Arg389, and Lys390) had slightly improved the

enzyme stability at 25 �C as well as at higher temperature of

45 �C. The tail segment which constituted of unstructured

region was proven to contribute to the structural flexibility of

AT2 lipase for which in their absence, the structural stability

increased. Low number of flexible regions enhance protein

rigidity, and hence its thermostability.

The influence of termini moiety in enzyme thermal

stability was also investigated in L1 lipase by MD simu-

lation. The study demonstrated that the N-terminal region

as well as a small domain was critical in determining the

stability of the enzyme at high temperature. The N-terminal

moiety showed high flexibility and dynamic throughout the

simulation conducted at 300, 400, and 500 K [29].

Analysis of MD Simulation by RMSF

To attain information on the local structural flexibility

particularly of the terminal regions, the average RMSF

values for each residue throughout the whole trajectories

were calculated. High RMSF value represents high flexi-

bility and often, such observation is found on terminal and

loosely structured loop [30]. Figure 7 depicts the evolution

of RMSF values during the dynamics at 25 and 45 �C. For

both simulations, major fluctuation in Ca backbone was

clearly seen on terminal residues. Terminal region is

regarded to be more flexible than other parts of the protein

[31]. This was also observed in ARM lipase and its mutant

where N-terminal moieties were largely mobile as denoted

by high RMSF value at high-temperature simulation [32].

Simulation of rWT-AT2 at 25 �C (Fig. 7a) showed that

four residues at the N-terminal region (Ala1, Gln2, Ala3,

and Gln4) displayed high RMSF values of 5.556, 5.336,

3.681, and 2.611 A, respectively. Meanwhile, at the

C-terminal region, the RMSF values were higher: 4.029,

3.747, 5.051, and 6.332 A for Ile387, Thr388, Arg389, and

Lys90, respectively. Residues that stretched from 181 to

229 are proposed to be the helix-loop-helix motif forming

the lid region. In contrary, the RMSF values for M386

showed lower fluctuations compared to rWT-AT2. The

absence of flexible C-terminal region promoted stabiliza-

tion of the structure in almost a fully relaxed model.

Analysis of RMSF at high-temperature simulation

(45 �C) for both rWT-AT2 and M386 reflected almost

similar fluctuation pattern to that obtained at 25 �C

(Fig. 7b). However, the RMSF values of virtually all res-

idues for both rWT-AT2 and M386 were larger compared

to dynamic at 25 �C reflecting that high temperature caused

greater movement of residues by thermal energy. The N-

and C-terminal regions of rWT-AT2 exhibited higher

RMSF values compared to M386 suggesting the plasticity

property of the native protein. Similar to that of low tem-

perature dynamic, high fluctuating region was also

observed at the predicted lid motif of AT2 lipase (residues

181–229). In the absence of the tail-like extension, M386

displayed a more stable fluctuation profile than rWT-AT2

throughout the total trajectory. Terminal residues have also

been proposed to influence the stability of helix and strands

[27]. This might explain that lacking the short random-coil

fragment, the stability of the C-terminal helix of AT2

lipase that spun from residue 364–385, was improved. The

long 22-residues helix is deemed to be crucial in holding

the Ca2? binding site in position near to the catalytic triad.

Consequently, it increased the stability of the global

structural conformation.

Fig. 7 RMSF of the Ca backbone position for each residue in rWT-

AT2 and M386. The average RMSFs were measured at dynamic a 25

and b 45 �C. Simulations were performed in aqueous environment.

Gray line denotes the RMSF fluctuation of rWT-AT2, while black

line indicates the RMSF fluctuation of M386 calculated for each

residue

Mol Biotechnol

123

Effect of Truncation on the Organic Solvent Stability

of AT2 Lipase

Recent review has underlined that surface charged residue

is one of the major determinants for organic solvent sta-

bility property in lipases [33]. In this study, the C-terminal

end deletion involved a solvent-exposed, charged residue

namely lysine. Hence, to investigate the effect of this

modification on the organic solvent stability profile of AT2

lipase, fifteen organic solvents with different log P were

tested.

Table 2 depicts the effects of organic solvents on the

lipolytic activity of rWT-AT2 and M386. M386 demon-

strated improvement in lipolytic activity in DMSO, iso-

propanol and diethyl ether by 50–150 % as compared to

rWT-AT2. In other polar organic solvents namely metha-

nol, ethanol, and acetone, the lipolytic activity of both

rWT-AT2 and M386 was fairly comparable. Treatment

with acetonitrile and 1-propanol had resulted in almost a

complete inhibition of the lipolytic activity of rWT-AT2

and M386. In non-polar organic solvents, the enzymatic

activity of M386 was retained by 54 % in n-hexane.

Meanwhile, suppression of the activity to over 50 % was

observed when M386 was treated with toluene, octanol,

xylene, and n-heptane. In contrast to M386, rWT-AT2

displayed stability in toluene, xylene, and n-hexane by

retaining more than 50 % of the lipolytic activity. Under-

standing of the activation and inhibition events of a specific

organic solvent toward enzyme, however, remains ambig-

uous and unclear since there is no particular trend or pat-

tern has yet been observed literally. Hence, this result

though may not be able to specifically conclude the specific

effects of the missing terminal end toward the organic

solvent stability of AT2 lipase, but it generally shows that

the mutation has caused a slight change in the stability

pattern and that the deleted residues are likely to involve in

protein–solvent interaction.

Conclusions

Flexible C-terminal region of AT2 lipase was found to be

an important determinant to AT2 lipase stability. A trun-

cated C-terminal mutant, M386, showed improved stability

and changes in enzymatic properties. Swapping of the tail-

like extension also yielded a crystallizable variant which

could be further optimized for structure determination. A

slight variation in organic solvent stability pattern of AT2

lipase was also observed indicating that the surface-

exposed residues were involved in protein–solvent inter-

action. This strategy may serve as a new route in lipase

engineering studies especially for stabilization of intrinsi-

cally disordered protein.

Acknowledgments This project was financially supported by

Ministry of Science and Technology (MOSTI), Malaysia (10-01-04-

SS07).

References

1. Dunker, A. K., Lawson, J. D., Brown, C. J., Williams, R. M.,

Romero, P., et al. (2001). Intrinsically disordered protein. Journal

of Molecular Graphics and Modelling, 19, 26–59.

2. Joseph, B., Ramteke, P. W., & Thomas, G. (2008). Cold active

microbial lipases: Some hot issues and recent developments.

Biotechnology Advances, 26, 457–470.

3. Li, W. F., Zhou, X. X., & Lu, P. (2005). Structural features of

thermozymes. Biotechnology Advances, 23, 271–281.

4. Dale, G. E., Oefner, C., & D’Arcy, A. (2003). The protein as a

variable in protein crystallization. Journal of Structural Biology,

142, 88–97.

5. Alexandrov, N. (1993). Structural argument for N-terminal ini-

tiation of protein folding. Protein Science, 2, 1989–1991.

6. Sharma, S., Zheng, H., Huang, Y. J., Ertekin, A., Hamuro, Y.,

et al. (2009). Construct optimization for protein NMR structure

analysis using amide hydrogen/deuterium exchange mass spec-

trometry. Proteins, 76, 882–894.

7. Fernando, P., Abdulle, R., Mohindra, A., Guillemette, J. G., &

Heikkila, J. J. (2002). Mutation or deletion of the C-terminal tail

affects the function and structure of Xenopus laevis small heat

shock protein, hsp30. Comparative Biochemistry and Physiology

Part B: Biochemistry and Molecular Biology, 133, 95–103.

Table 2 Effects of organic solvents on the lipolytic activities of

rWT-AT2 and M386

Organic solvent (log P) Relative activity, %

rWT-AT2 M386

Control (without organic

solvent)

100.00 100.00

DMSO (-1.3) 137.21 ± 8.26 245.76 ± 16.27

Methanol (-0.76) 238.84 ± 11.36 211.42 ± 5.65

Acetonitrile (-0.33) 22.87 ± 0.64 9.80 ± 3.90

Ethanol (-0.24) 150.36 ± 8.88 142.13 ± 5.29

Acetone (-0.24) 132.21 ± 2.56 111.55 ± 11.79

1-propanol (0.28) 14.25 ± 8.98 6.05 ± 1.08

Isopropanol (0.28) 111.80 ± 17.68 252.62 ± 23.01

Diethyl ether (0.85) 243.83 ± 16.68 379.03 ± 27.81

Chloroform (2.0) 24.68 ± 4.46 5.43 ± 5.61

Benzene (2.0) 49.18 ± 10.74 42.88 ± 10.56

Toluene (2.5) 60.07 ± 7.06 39.14 ± 1.87

Octanol (2.9) 48.73 ± 6.42 23.22 ± 6.62

Xylene (3.1) 64.61 ± 14.85 34.15 ± 9.43

n-Hexane (3.5) 107.26 ± 15.83 54.12 ± 10.59

n-Heptane (4.0) 45.10 ± 2.57 41.01 ± 7.95

log P is the partition coefficient of water/octanol which used to

measure the solvent polarity [34]. Purified enzyme was pre-incubated

with 25 % (v/v) of organic solvents for 30 min at 20 �C with shaking

at 150 rpm and assayed for remaining lipolytic activity. Data are

presented as relative activity (±) standard deviation

Mol Biotechnol

123

8. Martin, G., Keller, W., & Doublie, S. (2000). Crystal structure of

mammalian poly (A) polymerase in complex with an analog of

ATP. EMBO Journal, 19, 4193–4203.

9. Rahman, R. N. Z. R. A., Kamarudin, N. H. A., Yunus, J., Salleh,

A. B., & Basri, M. (2010). Expression of an organic solvent stable

lipase from Staphylococcus epidermidis AT2. International

Journal of Molecular Sciences, 11, 3195–3208.

10. Laemmli, U. K. (1970). Cleavage of structural proteins during the

assembly of the head of bacteriophage T4. Nature, 227, 680–685.

11. Kwon, D. Y., & Rhee, J. S. (1986). A simple and rapid colori-

metric method for determination of free fatty acids for lipase

assay. Journal of the American Oil Chemist’s Society, 63, 89–92.

12. Krieger, E., Koraimann, G., & Vriend, G. (2002). Increasing the

precision of comparative models with YASARA NOVA—A self-

parameterizing force field. Proteins, 47, 393–402.

13. Krieger, E., Darden, T., Nabuurs, S., Finkelstein, A., & Vriend,

G. (2004). Making optimal use of empirical energy functions:

Force-field parameterization in crystal space. Proteins, 57,

678–683.

14. Noble, M. E., Cleasby, A., Johnson, L. N., Egmond, M. R., &

Frenken, L. G. (1993). The crystal structure of triacylglycerol

lipase from Pseudomonas glumae reveals a partially redundant

catalytic aspartate. FEBS Letters, 331, 123–128.

15. Tyndall, J. D. A., Sinchaikul, S., Fothergill-Gilmore, L. A.,

Taylor, P., & Walkinshaw, M. D. (2002). Crystal structure of a

thermostable lipase from Bacillus stearothermophilus P1. Journal

of Molecular Biology, 323, 859–869.

16. Tiesinga, J. J. W., Pauderoyan, G. V., Nardini, M., Ransac, S., &

Dijkstra, B. W. (2007). Structural basis of phospholipase activity

of Staphylococcus hyicus lipase. Journal of Molecular Biology,

371, 447–456.

17. Arakawa, T., Philo, J. S., & Kita, Y. (2001). Kinetic and ther-

modynamic analysis of thermal unfolding of recombinant eryth-

ropoietin. Bioscience, Biotechnology, and Biochemistry, 65,

1321–1327.

18. Nybo, K. (2012). Troubleshooting forum, molecular biology and

techniques: Native PAGE. BioTechniques, 52, 20–21.

19. Gu, J., & Hilser, V. (2009). The significance and impacts of

protein disorder and conformational variants. In J. Gu & P.

E. Bourne (Eds.), Structural bioinformatics (pp. 939–962).

Hoboken: Wiley.

20. Buske, P. J., & Levin, P. A. (2013). A flexible C-terminal linker

is required for proper FtsZ assembly in vitro and cytokinetic ring

formation in vivo. Molecular Microbiology, 89, 249–263.

21. Liao, J. H., Lee, J. S., Wu, S. H., & Chiou, S. H. (2009). COOH-

terminal truncations and site-directed mutations enhance ther-

mostability and chaperone-like activity of porcine aB-crystallin.

Molecular Vision, 15, 1429–1444.

22. McPherson, A. (2001). A comparison of salts for the crystalli-

zation of macromolecules. Protein Science, 10, 418–422.

23. Thornton, J. M., & Sibanda, B. L. (1983). Amino and carboxy-

terminal regions in globular proteins. Journal of Molecular

Biology, 167, 443–460.

24. Derewenda, Z. S. (2010). Application of protein engineering to

enhance protein crystallizability and improve crystal properties.

Acta Crystallographica Section D, 66, 604–615.

25. Evdokimov, A. G., Pokross, M., Walter, R., Mekel, M., Cox, B.,

et al. (2006). Engineering the catalytic domain of human protein

tyrosine phosphatase b for structure-based drug discovery. Acta

Crystallographica Section D, 62, 1435–1445.

26. Russel, R. J. M., Ferguson, J. M. C., Hough, D. W., Danson, M.

J., & Taylor, G. L. (1997). The crystal structure of citrate syn-

thase from hyperthermophilic archeon Pyrococcus furiosus at

1.9 A resolution. Biochemistry, 36, 9983–9994.

27. Liu, L., Zhang, G., Zhang, Z., Wang, S., & Chen, H. (2011).

Terminal amino acids disturb xylanase thermostability and

activity. Journal of Biological Chemistry, 286, 44710–44715.

28. Haran, G., Haas, E., & Rapaport, D. C. (1994). Molecular

dynamics simulations of simple peptide models: Solvent effects

and comparison with experiment. The Journal of Physical

Chemistry, 98, 10294–10302.

29. Karjiban, R. A., Rahman, M. B. A., Salleh, A. B., Basri, M.,

Rahman, R. N. Z. R. A., & Chor, A. L. T. (2009). On the

importance of the small domain in the thermostability of ther-

moalkalophilic lipases from L1 and T1: Insights from molecular

dynamics simulation. Protein and Peptide Letters, 17, 699–707.

30. Kamal, M. Z., Mohammad, T. A. S., Krishnamoorthy, G., & Rao,

N. M. (2012). Role of active site rigidity in activity: MD simu-

lation and fluorescence study on a lipase mutant. PLoS ONE, 7,

1–8.

31. Iwakura, M., & Honda, S. (1996). Stability and reversibility of

thermal denaturation are greatly improved by limiting terminal

flexibility of Escherichia coli dihydrofolate reductase. Journal of

Biochemistry, 119, 414–420.

32. Salleh, A. B., Abd Rahim, A. S. M., Abdul Rahman, R. N. Z. R.,

Leow, T. C., & Basri, M. (2012). The role of Arg157Ser in

improving the compactness and stability of ARM lipase. Journal

of Computational Science and System Biology, 5, 039–046.

33. Chakravorty, D., Parameswaran, S., Dubey, V. K., & Patra, S.

(2012). Unraveling the rationale behind organic solvent stability

of lipases. Applied Biochemistry and Biotechnology, 167,

439–461.

34. Laane, C., Boeren, S., Vos, K., & Veeger, C. (1987). Rules for

optimization of biocatalysis in organic solvents. Biotechnology

and Bioengineering, 30, 81–87.

Mol Biotechnol

123