Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Medicine 1073

_____________________________ _____________________________

Studies on the DifferentialSpecificity of Protein Kinases

and Its Applications

BY

MART LOOG

ACTA UNIVERSITATIS UPSALIENSISUPPSALA 2001

Dissertation for the Degree of Doctor of Philosophy (Faculty of Medicine) in MedicalBiochemistry presented at Uppsala University in 2001

ABSTRACT

Loog, M. 2001. Studies on the differential specificity of protein kinases and itsapplications. Acta Universitatis Upsaliensis. Comprehensive Summaries of UppsalaDissertations from the Faculty of Medicine 1073. 48 pp. Uppsala. ISBN 91-554-5116-0.

Protein kinases are enzymes that catalyse the phosphoryl transfer from the -phosphateof ATP to acceptor amino acids in proteins. The specificity of selected model proteinkinases was studied at three different levels using a) novel bi-substrate-analogueinhibitors, b) synthetic peptide substrates and c) mutated protein substrate analogues.

A new class of protein kinase bi-substrate-analogue inhibitors was designed onthe basis of adenosine-5’-carboxylic acid derivatives, where a short arginine containingpeptide was attached to the 5'-carbon atom of the adenosine sugar moiety via a linkerchain. These compounds showed high inhibitory potential against two basophilicprotein kinases, the protein kinase A (PKA) and protein kinase C (PKC), with IC50values in the nanomolar range, but no inhibitory activity towards the acidophilic kinasesCKI and CK2. The inhibitors were efficiently applied for affinity purification of PKAusing MgATP as well as L-arginine as eluting agents.

Ca2+ -dependent protein kinase (CDPK-1) was purified from maize seedlingsand its substrate specificity was studied using a set of synthetic peptides. These werederived from the phosphorylatable sequence RVLSRLHS(15)VRER of maize sucrosesynthase 2 (SuSy2), and a consensus sequence motif A/LXRXXSXRZR (where Xdenotes a position with no strict amino acid requirements and Z a position strictly nottolerating arginine) was defined from a study using arrays of systematically variedpeptides attached to cellulose membrane (SPOTsTM membranes). The SuSy2 derivedpeptides were also found to be efficient substrates for mammalian PKC, but showed lowreactivity in the case of PKA. On the basis of this peptide motif, a positionally orientedpeptide library approach based on ESI-MS detection of phosphopeptides in initialvelocity conditions was designed for quantitative kinetic characterization of proteinkinase specificity profiles. On the basis of the obtained data an optimal peptide substratefor PKC, FRRRRSFRRR, was designed.

The specificity of protein kinase A was studied using site-directed mutagenesisin the phosphorylation site of L-type pyruvate kinase (L-PK), and comparison of theobtained data with the data from previous studies on structurally altered peptidesubstrates revealed that amino acid alterations in short peptide substrates cause strongereffects on the phosphorylation rate than the corresponding alterations in the proteinsubstrate L-PK.

Key words: Protein kinases, protein kinase inhibitors, substrate specificity, peptidelibraries, affinity chromatography.

Mart Loog, Department of Medical Biochemistry and Microbiology, Biomedical Center,Box 582, Uppsala, SE-751 23, Sweden

© Mart Loog 2001ISSN 0282-7476ISBN 91-554-5116-0Printed in Sweden by Eklundshofs Grafiska AB, Uppsala 2001

3

This thesis is based on the following papers, which are referred to by their Romannumerals

I. Loog, M., Uri, A., Raidaru, G., Järv, J., and Ek, P. (1999) Adenosine 5’-carboxylic acid peptidyl derivatives as inhibitors of protein kinases. Bioorg.Med. Chem. Lett. 9, 1447-52

II. Loog, M., Uri, A., Järv, J., and Ek, P. (2000). Bi-substrate analogue ligands foraffinity chromatography of protein kinases. FEBS Lett. 480, 244-248

III. Loog, M., Toomik, R., Sak, K., Muszynska, G., Järv, J., and Ek, P. (2000).Peptide phosphorylation by calcium-dependent protein kinase from maizeseedlings. Eur. J. Biochem. 267, 337-43

IV. Loog, M., Närvänen, A., Järv, J., Ek, B., and Ek, P. Systematic analysis ofsubstrate specificity profile for protein kinase C using a positionally orientedpeptide library approach coupled with electrospray mass-spectrometry.Manuscript

V. Loog, M., Oskolkov, N., O’Farrell, F., Järv, J., and Ek, P. Substrate specificityof protein kinase A in reaction with protein and peptide substrates. Manuscript

Reprints were made with permissions from the publishers.

4

Table of contents

Abbreviations……………………………………………………………………… 5Introduction……………………………………………………………………….. 61. The protein kinase superfamily…….…………………………………………... 72. The catalytic domain of protein kinase A as a model for the protein kinase superfamily…………………………………………………………………….. 83. Regulation of protein kinase activities………………………………………… 124. Protein kinase C……………………………………………………………….. 125. Calcium-dependent protein kinase from plants……………………………….. 146. Studies on the substrate specificity of protein kinases…..……………………. 15

6. 1. Phosphorylation site specificity of protein kinases…………………. 156. 1. 1. Structurally varied peptides for study of substrate specificity of protein kinases………………………………………….. 166. 1. 2. Use of peptide libraries for study of substrate specificity of protein kinases……………………………………………. 17

6. 2. Overview of the substrate specificity of serine/threonine protein kinases……………………………………………………………… 19

6. 2. 1. Basophilic protein kinases………………………………… 216. 2. 2. Proline directed protein kinases…………………..………. 216. 2. 3. Acidophilic and phosphate-directed protein kinases……… 22

6. 3. Docking sites and anchoring proteins as secondary specificity determining factors of protein kinases……………………………… 22

7. Inhibitors of protein kinases………………………………………………….. 237. 1. Protein and peptide site directed inhibitors………………………… 237. 2. ATP site directed inhibitors………………………………………… 267. 3. Bi-substrate inhibitors………………………………………………. 27

8. Affinity isolation of protein kinases using specific ligands targeted to substrate sites…………………………………………………………………………… 28Present investigation…………………………………………………………….. 29

Paper I……………………………………………………………………. 29Paper II…………………………………………………………………… 30Paper III………………………………………………………………….. 31Paper IV………………………………………………………………….. 32Paper V…………………………………………………………………… 33

Future perspectives……………………………………………………………… 34Acknowledgements……………………………………………………………… 35References………………………………………………………………………. 36

5

Abbreviations

AKAP A-kinase anchoring proteinCaMKII calcium- and calmodulin-dependent protein kinasecAMP cyclic AMPCDK cyclin-dependent protein kinaseCDPK calcium-dependent protein kinaseCK1 casein kinase 1CK2 protein kinase CK2, formerly casein kinase IIDAG diacylglycerolESI-MS electrospray mass-spectrometryGSK glycogen synthase kinaseL-PK L-type pyruvate kinaseMAPK mitogen-activated protein kinaseMLCK myelin light-chain kinasePEP phosphoenol pyruvatePKA protein kinase A, cAMP-dependent protein kinasePKC protein kinase CPKG cGMP-dependent protein kinasePKI heat-stable protein kinase inhibitorPS phosphatidylserineRACK receptor for activated C-kinase

6

Introduction

Protein kinases are enzymes which catalyse the phosphorylation of substrate proteins bytransfer of the -phosphate of ATP to the acceptor amino acid. Depending on theacceptor amino acids, the protein kinases can be classified as either serine/threoninekinases or tyrosine kinases, which are the two most important and numeroussubdivisions. This introduction aims to give an overview of the important aspects of thespecificity of serine/threonine protein kinases, with emphasis on the examples of twothoroughly studied model enzymes: protein kinase A (PKA), and protein kinase C(PKC).

The phosphorylation modulates the activity of many proteins and protein kinasesplay a key role in multiple signalling and regulatory processes in cell. It has beenestimated that as much as 20-50% of all the cellular proteins undergo phosphorylationin vivo [124, 173]. The regulation of protein activities by phosphorylation is a reversibleprocess, with the reverse direction catalysed by protein phosphatases. The extracellularsignals are relayed from the plasma membrane to specific intracellular targets mostlythrough several phosphorylation events catalysed by different protein kinases, each ofwhich, in order to maintain the exactness of the signalling flux, should possessindividual target recognition specificity and localisation specificity. There is a greatdiversity of protein kinase forms involved in eukaryote signalling and according to thepreliminary estimation of Celera’s human genome project, the human genome codes for868 protein kinases [168], corresponding to 2.8% of all genes, which makes this thelargest enzyme family in eukaryotic cells.

Ubiquitous involvement of protein kinases in the regulation of importantprocesses including cell growth, differentiation and apoptosis, associates them withmany diseases. Abnormal protein kinase activities have been identified as key factors inseveral cancers [76, 86]. These findings have validated several protein kinases aspotential drug targets and the design of specific protein kinase inhibitors presents apromising field of therapeutic intervention with diseases [5, 10, 147].

The functioning of protein kinases as “molecular swiches” in checkpoints ofcellular signalling also makes these enzymes attractive as biomarkers. Recently it hasbeen shown that differential protein kinase expression patterns can reflect the molecularphenotype of different disease states [14, 15, 48, 80, 135]. Determination of proteinkinase activity profiles in normal and diseased cells as fingerprints of the signallingstatus of the cell may lead to important medical applications including molecularclassification and diagnosis of disease, design of individual treatment strategy,identification of therapeutic markers and drug targets, and profiling of response totoxins and pharmaceuticals.

Despite the large amount of sequence data revealing the existence of hundreds ofprotein kinase forms, still only a small fraction of these enzymes have beencharacterised at the protein level and the substrate specificity has been investigated forfewer still. Understanding the subtle differences in specificity requirements for substratebinding regions of different representatives of this conserved family of enzymes wouldfacilitate the design of selective inhibitors, which can be considered as potential drugs.

7

There is also a great demand for selective inhibitors in the field of signal transductionresearch, where these compounds can be used for dissecting the signalling pathwaysmediated by phosphorylation of particular kinases and thereby unravelling the signallingnetworks. Secondly, the understanding of the structural requirements for substrates ofprotein kinases leads to the design of selective and highly efficient peptide substrates fordetection of individual protein kinase activities in cells. The latter direction isparticularily interesting in the light of the promising new trends in developing proteinkinase chips for high throughput simultaneous kinase activity screening [87, 173, 186].

New specific ligands are also needed for development of group specific affinityisolation methods for protein kinases. Protein kinases as regulatory catalyst moleculeshave low abundance in cells and the methods for their enrichment using specific affinityisolation methods are therefore in great demand in field of proteomics research. Suchaffinity adsorbents shorten the gap between genomics and proteomics by providing thepossibilities to study kinase function at protein level.

The current thesis focuses on the questions of phosphorylation site-specificity ofpeptide and protein substrates, and also on the possibilities and strategies for developinghighly efficient inhibitors and affinity ligands for protein kinases. Since the research forachieving these goals can be substantially accelerated by development of novelcombinatorial library methods and rational ligand design strategies, proportionally moreemphasis is paid to these questions.

1. The protein kinase superfamily

The classification of protein kinases and the definition of the eukaryotic protein kinasesuperfamily were done by Hanks, Quinn and Hunter [40]. The common motif which allprotein kinases share and are related by is the homologous “kinase domain”, also knownas the “catalytic domain”, which consists of 250-300 amino acid residues. The kinasedomain contains 12 conserved subdomains and it folds to closely similar three-dimensional core structures in different protein kinases (reviewed in [38, 39, 153]). Theclassification given by Hanks et al. [40] was based one the phylogenetic trees derivedfrom an alignment of kinase domain amino acid sequences. It was found that thesequence similarity of the kinase domains was a good indicator of other characteristicscommon to different kinases, revealing four main families with related substratespecificities and modes of regulation. The rest of the kinases falling outside these majorgroups were difficult to classify into defined subsections. The four groups are shownbelow together with the list of the few representatives which are more frequentlyreferred to in subsequent chapters of this thesis:

1) The AGC group units protein kinases with mainly basic amino acidspecificity determinants including the cyclic nucleotide-regulated proteinkinase family with the well known representatives PKA and PKG, the

8

diacylglycerol-activated/phospholipid-dependent protein kinase C familyand the RAC family of kinases

2) The CaMK group covers the family of kinases regulated byCa2+/calmodulin. The plant Ca2+-dependent protein kinases, studied in oneof the papers of this thesis also belong to this group

3) The CMGC group includes the family of cyclin-dependent kinases (Cdks),the MAP kinase family and casein kinase II

4) The PTK group: the protein-tyrosine kinase group

The first 3-dimentional structure of a protein kinase was the crystal structure ofthe catalytic subunit of protein kinase A (cAMP-dependent protein kinase, PKA) incomplex with its peptide inhibitor PKI and MgATP, solved in 1991[79]. Lateradditional structures of PKA in different ligand complexes were solved [24, 106-108,126, 184, 185] and PKA has become the most thoroughly studied representative of theprotein kinase superfamily serving as a model system for research on newly identifiedand isolated protein kinases.

Protein kinases have been identified and characterised from other animal phyla,besides the mammals, including other vertebrates, as well as plants, fungi andmicroorganisms, and the comparative studies have demonstrated that the basic aspectsof protein phosphorylation pathways regulating the cell functions have been maintainedthroughout the course of eukaryotic evolution [38].

2. The catalytic domain of protein kinase A as a model for the protein kinasesuperfamily

PKA is the most investigated and understood of all the protein kinases of thesuperfamily and many general rules derived from these extensive structural and kineticstudies will certainly apply to the entire protein kinase superfamily. This protein kinaseis unique due to its dissociative mechanism of activation in response to the secondmessenger cyclic AMP (cAMP). The inactive holoenzyme of PKA is a tetramerconsisting of two regulatory and two catalytic subunits. Binding of cAMP to theregulatory subunits in the holoenzyme leads to its dissociation into two active catalyticsubunits and a dimer of regulatory subunits [151]:

R2C2 + 4 cAMP R2(cAMP)4 + 2 C

There are three known genes that code for the catalytic subunit. The C is thepredominant form and is expressed in most tissues [162]. The C is more tissuespecific, being highly expressed in neuronal tissues [9, 140, 163] and the C has beenfound in testis and its expression is most limited [3].

The schematic representation of the catalytic and regulatory subunit structure isgiven in Fig. 1. The general three-dimensional architecture of the catalytic subunit is

9

Figure 1. Schematic view of the catalytic subunit and regulatory subunit of PKA.

shown in form of the ribbon diagram in Fig. 2, as a ternary complex, where therecombinant catalytic subunit was co-crystallized with MgATP and a 20-residue peptideinhibitor PKI(5-24) [79]. The catalytic subunit consists of a smaller upper lobe and alarger lower lobe. The small lobe is dominated by antiparallel -sheets and isresponsible for ATP binding. The loop between the beta sheets 1 and 2 of the smalllobe, encompassing the signature motif Gly50-X-Gly52-X-X-Gly55, is called thephosphate anchor since it is binding the phosphate groups of ATP via several hydrogenbonds [6, 184, 185]. The third beta strand contains a lysine residue (Lys72), which isconserved throughout the protein kinase superfamily. This lysine coordinates the - and

-phosphates of the ATP and also forms a salt bridge with Glu91 located in the middleof helix C. The latter interaction is important for maintaining the active state in case ofother studied protein kinases [56, 64]. The adenine ring of ATP binds in a confined,mostly hydrophobic pocket, located in the cleft under the strands 1 and 2. Most of theinteractions of ATP, except the contacts with the -phosphate and the hydrophobicinteractions with the adenine ring, are associated with the small lobe.

The large lobe is predominantly -helical and is associated with catalysis andpeptide binding. An important structural element of the large lobe is the catalytic loopconnecting the strands 6 and 7. The loop is preceded by the conserved Arg165, theside chain of which forms a tight electrostatic contact with phosphorylated Thr197. This

Peptide binding/catalysis

Regulatory subunit

ATP bindingA-helix

Conserved kinase core

Tail

Catalytic subunit

cAMP binding sitesDimerizationdomain Pseudo-

substrate

Site A Site B

10

Figure 2. Ribbon diagram of PKA catalytic subunit cocrystallized with MgATP andpeptide inhibitor PKI[5-24].

11

interaction is essential for maintaining the active conformation of the kinase. Theneighbouring amino acid, the conserved Asp166, is proposed to act as catalytic base inthe reaction of phosphoryl transfer [95].

-strand 8 is followed by a conserved Asp184-Phe185-Gly186 (DFG) motif orthe Mg2+ binding loop. The Asp184, correctly positioned by the anchoring hydrophobicinteractions of the neighbouring phenylalanine, is one of the ligands of the magnesiumion coordination sphere, which in turn coordinates the - and -phosphates of ATP. TheDFG motif is followed by -strand 9 and the activation loop. The activation loopcontains the previously mentioned Thr197, which phosphorylation byautophosphorylation or the PDK1 kinase [13] is essential for the activity of PKA. Mostprotein kinases possess threonine or tyrosine residues in the activation loop that must bephosphorylated, either autocatalytically or by an activating kinase in order to be active(reviewed in [66]). The three additional residues in the large lobe of the kinase core,which are conserved throughout the superfamily act as general stabilizators of thestructure. The Asp 220 forms hydrogen bonds to the backbone nitrogens of Arg165 andTyr164, and thereby stabilizes the catalytic loop, while Glu208 and Arg280 form aburied ion pair [77].

The large lobe also contains the binding region of the peptide and proteinsubstrates, whose structural details determine the peptide specificity of protein kinasesand thus are particularly interesting in relation to the practical aims of the current thesis.The specific binding of peptide pseudosubstrate inhibitorTTYADFIASGRTGRRNA*IHD, or PKI(5-24), is facilitated by several specificcontacts between consensus motif amino acid side chains, highlighted in bold in thepreceding sequence, and the amino acids in the binding site at the catalytic subunit. Inthe current thesis we have adopted a nomenclature that designates the phosphorylationsite as P-site. In the case of PKI the P-site is occupied by a non-phosphoacceptor aminoacid alanine (*). The amino acids carboxy-terminal to this site are denoted as plus andthe amino terminal residues as minus. The residues that are important for recognition ofthe main basic specificity determinants in PKI and peptide substrates are Glu170 andGlu230 for P-2 arginine, Asp328 and Glu127 for P-3 arginine, Glu203 and Tyr204 forP–6 arginine. The hydrophobic pocket accommodating the side chain of P+1 Ile isformed by Leu198, Pro202 and Leu205 of the P+1 loop, which follows the activationloop. The phenylalanine of P-11 located in the N-terminal amphipathic helix, a criticalresidue responsible for high affinity binding of PKI, lies in the hydrophobic pocket thatconsists of Tyr235, Pro236 and Phe239.

The kinetic mechanism of PKA-catalysed phosphorylation of peptide substratesis formally random sequential, but under standard assay conditions and presumably alsoin physiological conditions, the mechanism is predominantly an ordered sequentialsteady-state, with ATP binding first [1]. In the same study it was found that kcat wasvery sensitive to viscosity, indicating that ADP release and not the chemical step is ratelimiting.

12

3. Regulation of protein kinase activities

There are several control mechanisms for protein kinase catalytic activity. The first andthe most essential is the control by additional subunits or domains that may function inresponse to second messengers. PKA is a specific example of regulation of a proteinkinase by regulatory subunits. The regulatory subunit contains a pseudosubstrate typesequence with the substrate consensus sequence RRGAI, which at low levels of cAMPoccupies the substrate binding site and thereby hinders the binding of the substrate. Sofar, both the catalytic subunit and the regulatory subunit of PKA have been crystallizedbut the absence of a crystal structure for the regulatory and catalytic subunit complexhampers the precise determination of the structural basis of the regulatory mechanism.Downregulation by the pseudosubstrate regions is also the case for the regulation byregulatory domains, with the simple difference that the pseudosubstrate motif in thesecases is found in the same polypeptide chain as the catalytic domain. The twocharacteristic examples of this type of regulation, are the protein kinase C (PKC) andcalcium dependent protein kinase (CDPK) from plants.

Another type of regulation is the control mechanism by additional subunitswhose level of expression varies depending on the functional state of the cell. Examplesof such regulation are the cyclin-dependent kinases, activated through binding ofdifferent cyclins and thereby controlling the timing of the cell cycle progression. Thecrystal structures of inactive cyclin-dependent protein kinase 2 (cdk2) [20] and the sameenzyme activated by cyclin A have been solved explaining the control mechanism issufficient detail [64].

The third most important type of regulation is the regulation by phosphorylationof the activation loop. In PKA and also several other studied kinases, a significantcomponent of activation comes from the interaction of basic residues of catalytic loopwith a phosphate covalently attached to the activation loop. In case of PKA thisabsolutely essential phosphorylation occurs at the Thr 197 residue. In other kinasesseveral individual varieties of this regulatory mechanisms have been discovered(reviewed in [66]).

4. Protein kinase C

Protein kinase C (PKC) was initially discovered as a histone kinase which could beactivated by limited proteolysis by a Ca2+-activated protease [61]. The protease cleavesoff the regulatory domain of the inactive enzyme releasing the active catalytic kinasecore. Later it was found that the discovered kinase could also be activated byphosphatidylserine (PS) and diacylglycerol (DAG) in a Ca2+-dependent manner andadditionally by tumor promoting phorbol esters such as PMA (reviewed in [111]). Untilnow, a number of PKC isoforms have been discovered, defining a PKC subfamily,consisting of at least twelve distinct genes [100]. There are three main groups in thissubfamily, the conventional PKCs (cPKC), novel PKCs (nPKC) and atypical PKCs

13

Figure 3. Schematic view of the conventional PKC.

(aPKC). The most studied and understood of them are the conventional PKCs, whichcomprise of the , I, II and isoforms and are activated by Ca2+/PS and also bybinding of DAG. The PKC gene is alternatively spliced yielding the two isoforms,which differ only in their C-terminal ends. The nPKCs, consist of the , , and isotypes and are Ca2+ insensitive but activated by DAG or phorbol esters in the presenceof PS [118]. The aPKC (isoforms and ) are both insensitive to Ca2+ and PMA/DAG[119]. The schematic view of the functional domain building blocks for cPKC group isgiven in Fig. 3. For cPKCs, the N-terminal regulatory domain starts with thepseudosubstrate sequence followed by a cystein-rich C1 domain, which contains tworepeated zink-finger motifs and has been shown to accommodate the binding of phorbolester [117] and DAG [137]. The C1 domain is present also in nPKCs , but the aPKCscontain only one zink-finger motif. The C2 domain that follows the C1 domain confersCa2+/PS binding to the cPKCs. Initially it was thought that there was no C2 domain innPKC and aPKC, but later analysis showed that the N-terminal regions preceding theC1 domain, or the single zink-finger motif in the case of aPKC, reveal homology withthe C2 domains, but lack certain conserved aspartate residues, which explains why thesePKCs are incapable to bind Ca2+ but are still stimulated by PS [100].

PKCs have a multitude of substrates in cells and the general physiologicalpathway for cPKCs and nPKCs starts with the stimulation of phospholipase C activityby particular cell surface receptors, which is followed by the hydrolysis ofphosphatidylinositol 4,5 bisphosphate (PtdIns4,5P2) yielding DAG and as aconsequence activated PKC. Upon activation, DAG binds to the PKC/Ca2+/phospholipidunproductive complex, formed by association with the cell membrane (or vesicle),forming the tight active ternary complex, where the pseudosubstrate region is removedfrom the active site groove [122].

The kinase core of mammalian PKCs shares identity in the functionallyimportant conserved residues listed for PKA above, except for several differences inamino acids located in the substrate peptide binding pockets, these being the structural

Peptide bindingATPbinding

Conserved kinase core

V1 V2 V5V3C1 C2 C4C3

Ca bindingPMA/DAG

Regulatory domain

Pseudo-substrate

14

origin of the differential substrate specificity of these two kinase groups. Within thePKC family, the substrate specificity requirements are broadly overlapping due to highhomology of the kinase core [47], and therefore the complete understanding of thespecificity factors governing the isoenzyme specific signalling is not yet completelyattained.

Figure 4. Schematic view of the CDPK structure. The calmodulin-like domain containsfour calcium-binding EF-hand motifs.

5. Calcium-dependent protein kinase from plants

Calcium-dependent protein kinases (CDPKs) form a family of protein kinases in plants[42, 130]. The molecule of CDPK consists of a single polypeptide with a typical proteinkinase catalytic core, a linker region containing the pseudosubstrate autoinhibitorysequence and an adjoining calmodulin-like domain with four calcium-binding EF-handmotifs (Fig. 4). CDPK activity is regulated by cytosolic Ca2+, which act, similarly tomammalian cells, as second messengers in plants.

The CDPKs belong to the CaMK group of the eukaryotic protein kinasesuperfamily, where they form a large family together with Ca2+/calmodulin-activatedprotein kinases including CaMKII, CaMKIV, PhK- , SmMLCK and SkMLCK [38].The physiological roles and protein targets for CDPKs have yet been poorlydocumented, however a few reports prove their involvement in such central mechanismsas stress signal transduction, regulation of cell growth, metabolism and gene expression[22, 57, 139]. As one of the putative physiological roles for CDPK, also studied by ourgroup, is the regulatory phosphorylation of one of the key enzymes of sucrosemetabolism in maize, the sucrose synthase 2, which is phosphorylated in vivo on itsSer15 by a CDPK from maize seedlings [58, 91].

A number of recent discoveries have revealed a complex isoenzyme pattern ofCDPKs expressed in plant cells and many of these closely related kinase forms havebeen isolated or cloned from a number of plant sources [7, 23, 25, 62]. The high degree

Conserved kinase coreCaM-likedomain

Autoinhibitory domain

15

of the isoform versatility indicates to a sophisticated network where one of the keyfactors maintaining the exactness of signal transduction must be the substrate specificityof particular isoenzymes.

An interesting aspect making the CDPKs attractive for research is its functionalsimilarity to mammalian PKCs. There are no structural PKC analogues found ingenomic data on plant model organisms and therefore it has been proposed that CDPKsmay be the functional analogues of PKCs in plants. CDPKs, similarily to PKCs, areregulated by intracellular Ca2+ signals. But strikingly, there is also experimentalevidence that some of the CDPK forms are additionally activated by phospholipids,making them candidate mediators of the inositol phospholipid-based signalling in plants[149].

6. Studies on the substrate specificity of protein kinases

Substrate specificity of protein kinases is controlled on three main structural levels.Firstly, most of the protein kinases exhibit a certain degree of site-specificity, whichmeans that they are able to recognize primary structure motifs in the vicinity of thephophoacceptor amino acid. Secondly, the recognition of a substrate protein may bemediated by a high-affinity interaction between a site of the kinase other than the P-siteaccomodating region, and a short sequence of the substrate protein called “dockingsite”. Thirdly, the targeting of a kinase to its substrate may be achieved by specificanchoring protein(s) containing “anchoring motif”, which binds the kinase, and the“targeting domain”, which directs the subcellular localisation of the kinase to thevicinity of its substrate. A protein kinase may use all of these three levels of recognitionfor different phosphorylation events, but a certain degree of phosphorylation sitespecificity is the minimum necessary factor needed for correct docking and positioningof the phosphoacceptor residue into the active site. In many cases the other two levelsare not absolutely required to be present, but their presence may decrease theimportance for the substrate to possess the full P-site recognition consensus motif. In thepresent thesis, emphasis is put on the study of phosphorylation site primary structurespecificity factors of different model protein kinases.

6.1. Phosphorylation site specificity of protein kinases

Study of the phosphorylation site specificity of protein kinases started about 40 yearsago with phosphorylase kinase, which was shown to phosphorylate both phosphorylase[28] at Ser14 and the chymotryptic peptide containing the same site [112]. Due togrowing evidence that, in a number of cases, synthetic peptide substrates based on thephosphorylation site sequence in proteins, showed Km values in the same range or evenlower than their protein progenitors, a series of studies was conducted in order to

16

characterise the primary structure preferences for different kinases (reviewed in [72,124, 183]). The overview of the specificity studies given below is mainly focused onPKA and PKC, since these two kinases are the model systems used in the current thesiswork. The current knowledge on the phosphorylation site-specificity consensussequences of other protein kinases are taken together in Table 1.

6.1.1. Structurally varied peptides for study of substrate specificity of proteinkinases

The first systematic studies using structurally varied synthetic peptide substrates, basedon the sequence of identified phosphorylation sites in protein substrates were performedfor PKA [27, 70, 98, 181, 182]. It was found that the short peptide substrates based onthe phosphorylation site of L-type pyruvate kinase were just as good substrates or evenbetter than the pyruvate kinase itself and it was found that the minimal short sequenceRRASV was still efficiently phosphorylated by PKA with Km in the low micromolarrange [182]. Further shortening of this structure led to dramatic decreases in reactivity[182]. Thus the structure RRASV, was denoted “minimal substrate” for PKA, i.e. theshortest substrate structure containing the whole set of important specificitydeterminants for efficient phosphorylation by the protein kinase. The two arginines werefound to be exclusive substrate specificity determinants for recognition by PKA sincethey could not be replaced by even lysine without serious detriment to substrateefficiency [70]. These studies resulted in the definition of a minimal consensussequence RRXS/TB, for PKA, where X denoted any amino acid and B is a largehydrophobic amino acid [74]. More recently, the determination of three-dimensionalstructures of PKA-pseudosubstrate inhibitor peptide complexes [152, 161] have shedlight on the peptide recognition mechanism of PKA in atomic detail and the system hasbecome a model of substrate recognition for the whole protein kinase superfamily.

Extensive studies using peptide substrates have also been performed on PKC(reviewed in [47]), however, the site-specificity of this kinase was not as defined as forPKA. The studies resulted in definition of several possible consensus sequences (Table1), the most frequent of which was RXXS/TXR. In addition, a favourable role ofhydrophobic residue in position +1 was found [176], indicating the similarity to PKA,but later systematic studies showed that also basic amino acids are specificitydeterminants in the position P+1 [68].

However, despite of these extensive efforts, no systematic positional analysis ofthe structural impact of each possible amino acid on protein kinase substrate reactivitywas done until the early nineties. The attempt to rationalise and systematize thecollected peptide data was performed for PKA by Järv and Ragnarsson [67]. In thisstudy the amino acid positions from –3 to +2 around the serine in synthetic peptideanalogues were analysed by linear free energy relationships. Different structuralparameters for each amino acid side chain were correlated with logarithmic values ofkinetic constants obtained for the positionally varied peptide analogues and the

17

structural specificity of each position was characterized by the intensity factors obtainedfrom these correlations. A similar approach was used, and developed further in twostudies where the structural specificity of PKC and PKA was studied using the peptideanalogues derived from a phosphorylation site of myelin basic protein, varied inposition P+1 [68, 92]. It was found that the specificity requirements of these kinases inposition P+1 could be quantitatively described by the obtained correlations.Hydrophobicity, bulkiness and charge were all important factors for recognition by PKCwhile only the hydrophobicity and bulkiness were important for PKA. From thesestudies it was understood that it is not possible to express the primary structure sitespecificity of protein kinases by a defined and rigid consensus sequence, since it mayhave a far more complex character, involving different structural factors, which mayhave positive or negative influences on substrate reactivity.

6.1.2. Use of peptide libraries for study of substrate specificity of protein kinases

Despite the extensive knowledge obtained from the studies described above, theapproach using the separately synthesized peptide substrate models for determination ofspecificity requirements for protein kinases is time consuming and costly. As analternative, in the last decade, different combinatorial peptide libraries techniques wereapplied for determination of the consensus sequences and optimal substrates for severalprotein kinases [11, 12, 59, 82, 84, 85, 94, 109, 110, 114, 116, 141-145, 154, 156, 177,178, 180]. These innovations were introduced to the protein kinase studies following thedevelopment of automated combinatorial peptide synthesis methods [32] and successfulapplications of peptide libraries for development of new bioactive peptides [51, 52]. Thepioneering attempt to apply these techniques for study of kinase recognition motifs wasdone by Songyang and co-workers, where the new strategy was used to determineoptimal sequences of phosphopeptides recognised by SH2 domains [146]. One of thefirst studies on the protein kinase substrate specificity was done by the same group forPKA, cyclin B-Cdc2, cyclin A CDK2 and SLK1 kinases, using a random peptide librarywith sequence MAXXXXSXXXXAKKK, where X indicates all amino acids excepttryptophan, cystein, tyrosine, serine and threonine [143]. The methodology developed inthis work has become a standard protocol for several later kinase specificity studies[109, 110, 114, 116, 141, 142, 144, 145]. According to this protocol, after partialphosphorylation of the library, the phosphopeptides are isolated using Fe-chelatingchromatography and submitted to peptide sequencing. The specificity profiles areconstructed using the relative abundances of each amino acid in each subsequentpositions. The optimal peptide consensus motifs obtained in these studies weregenerally in agreement with the known phosphorylation sites in proteins, showing thatsuch library methods can be used for qualitative substrate specificity studies.

In one of these studies the substrate specificity profiles for PKC isoenzymeswere investigated [110] using the positionally oriented (fixed phosphorylation site)library with general structure MAXXXXRXXSXXXXXAKKK. The arginine was

18

“locked in” since previous studies had shown the importance of Arg at P-3. Thisdetailed study allowed determination of the optimal substrate sequences for nine PKCisoenzymes. It was also possible to design optimal peptide substrates, which showed adegree of selectivity between the isoenzymes, rebutting the standard opinion that theisoenzymes of PKC are too homologous to allow discrimination between them usingpeptide substrates. However, certain disadvantages of these random libraries are evidentwhen the specificity diagrams are examined. The specificity profiles in several positionsdo not show strict structural specificity, i.e. in several positions the amino acids ofdifferent chemical nature show only minor differences in relative abundance, likelyindicating that the specificity is not strictly determined by defined binding pockets onthe enzyme molecule. The reason for this may be that the assay of such a large librarywas not made under the initial velocity conditions, leading to library“overphosphorylation”. Application of the initial velocity conditions is impossible dueto low sensitivity and separation efficiency of the detection methods used. It could beproposed that the correct sensitivity of the positional structure-activity relationships canbe achieved by using the step-by-step library design where every generation of librarieshas only a subset of amino acid positions varied. This can be combined with the rationaldesign of libraries, where the structural variations are diminished to a minimum, but stillcover a sufficiently wide range of structural details.

Another drawback, affecting the random library study in the opposite way, waspointed out by Pinna [124]. It was shown that the results from the library study for PKA[143] revealed 13 fold preference for arginine over alanine in position P–1, while thesame substitution performed on individual peptide analogues showed almost equalreactivity for the two peptides. This was explained by the situation where the arginine inP–1 position, in course of library phosphorylation starts to compensate the absence ofarginines in positions P-2 and P-3 and therefore becomes more abundant than the P-1noncharged alanine in the collected phosphopeptide pool.

As an alternative to the soluble peptide library strategies, peptide librariesattached to solid matrix have been developed. In the first attempt at this approach,random synthetic combinatorial peptide libraries were generated by the split peptidesynthesis on polystyrene beads of 90 m, where each bead expressed only one peptideentry [177]. This ”one bead one peptide” approach, also called the Selectide Process,had been previously described in [83] as a general method for peptide ligand search.Using this method, several optimal substrates for PKA were “fished out” from the largepool of 500000 beads with random hexapeptide sequences. These substrates containedthe PKA consensus sequence RRXS. However, the shortcoming of this approach is thatonly few optimal sequences are selected and no real knowledge about the detailedpositional structure-activity relationships and of the specificity profiles can be obtained.

Another powerful method was developed, where peptide libraries weresimultaneously and separately synthesised on cellulose membranes as SPOTsTM

peptides [29]. The successful utilisation of the SPOTs peptides for protein kinase assaywas demonstrated in our laboratory [157] and it was shown that this method wasefficient and less time consuming for selection of optimal peptide substrate for PKCfrom a number of candidate sequences based on the substrate sites of proteins [158]. A

19

more systematic approach for determination of kinase specificity profiles using theSPOTs peptide arrays, was developed by Tegge et al. [154], synthesizing severalgenerations of libraries. The first generation had the structure AcXXX12XXX, where Xrepresents mixtures of all twenty natural amino acids and the numbers 1 and 2 representindividual amino acids, each fixed on a separate spot. After phosphorylation by [ -32P]ATP the sublibraries AcXXXRRXX and AcXXXRKXXX were identified as the bestsubstrates for PKA and PKG. In the next library generations, these two positions werefixed and the rest of the positions of the octapeptides were analysed in a similar wayresulting in the optimal peptide structures KRAERKASIY and TQKARKKSNA forPKA and PKG, respectively, with Km values in low micromolar range when measuredin solution. However, this method suffers from the relatively high background bindingof [ -32P] ATP (unpublished results from our laboratory) and also due to the possibilityof formation of inhibitor structures causing the affinity docking of the enzyme to thespot area. Furthermore, the kinetic meaning of the radioactive spot intensities remainsunclear, as the values did not correspond to the Km values estimated for the peptides insolution [157].

In general, the following different aims in the peptide library studies on proteinkinase specificity can be outlined:

First, selection of some of the most optimal substrate structures from the largepool of possible substrate sequences.Secondly, determination of the substrate specificity profiles, where structure-activity relationships for each critical position around the phosphorylatableamino acid are obtained. This approach provides detailed knowledge about thespecificity requirements of closely related kinase forms and thus can be used fordesign of selective substrates discriminating between different kinase forms.Thirdly, the obtained consensus motifs may be valuable in predicting thepotential phosphorylation sites using protein sequence databases [69, 114, 142,143] However, this perspective must be analysed separately, because of theknown flexibility of the consensus motifs [81] and also in respect to the questionhow the sites predicted and tested in vitro, reflect the real physiological sites invivo.

6.2. Overview of the substrate specificity of serine/threonine protein kinases

With few exceptions, most of the studied serine/threonine kinases are markedly“sequence specific”. The serine/threonine kinases can be categorized into three maingroups with respect to their specificity determinants [124],

basophilic kinases, using basic and often also hydrophobic residues as specificitydeterminantsproline-directed kinases, some of which also require basic residues besidesproline and

20

Table 1.Phosphorylation specificity consensus motifs of serine/threonine protein kinases. Thetable is based on reviews [123, 124] and peptide library studies [114, 145]. B denotesthe hydrophobic amino acid, Z any amino acid except arginine and X any possibleamino acid. The phosphoacceptor residues are underlined.

Protein kinase Consensus sequence

Basophilic protein kinasesPKA R-(R/K)-X-S/T-BPKG R-(R/K)-X-S/T-BPKC X-R-X-X-S/T-X-R-X

S/T-X-K/R(R/K)-(R/K)-X-S/T-B-(R/K)-(R/K)(R/K)1-3-(X)2-0-S/T-(X)2-0-(R/K)1-3

AKT/PKB kinase R-X-R-X-X-S/TCaM kinase I B-X-R-X-X-S/T-X-X-X-BCaM kinase II B-X-(R/K)-X-X-S/T-BPhosphorylase kinase K/R-X-X-S-BCDPK maize L/A-X-R-X-X-S-X-R-Z-Rsm MLCK K-K-R-X-X-S-X-B-BAMP-stimulated kinase B-(X,R/K/H)-X-X-S/T-X-X-X-BTGF receptor kinase K-K-K-K-K-K-S/T-X-X-XHSV protein kinase R-R-R-R-X-S/T-XP70S6K (K/R)-X-R-X-X-S/T-BMAPKAP kinase 1 X-X-(R/K)-X-R-X-X-S-X-X

R-R-R-X-S-X-XMAPKAP kinase 2 X-X-B-X-R-X-X-S-X-XNIMA F-R-X-S-R/BHRI (Haem-regulated eIF-2 ) kinase (E)-X-S-R-X-X-R

Proline-directed protein kinasesCdc2 X-S/T-P-X-K/RCDK2 X-S/T-P-X-K/RP35/CDK5 K-S/T-P-X-H/R/K-H/R/KErk1/2 P-X(1-2) -S/T-P-P

Acidophilic protein kinasesCK1 (Sp/Tp)-X-X-(X)-S/T-B

(D/E)>4-X-X-S/T-B

CK2 X-X-X-S/T-X-X-(E/D/Sp/Yp)-XX is preferably acidic

GSK3 S/T-X-X-X-SpARK (D/E)n-S/T-X

21

acidophilic/phosphate directed kinases, which require carboxylic andphosphorylated residues.

A short overview of the consensus sequences of these three specificity groups is givenbelow, and more precisely taken together in Table 1.

6.2.1. Basophilic protein kinases

This group is comprised mainly of members of ACG and CaMK groups of proteinkinases and also of several kinases not falling into major groups according to the Hanksand Hunter nomenclature [38]. All three model enzymes used in the current work: PKA,PKC, and plant CDPK, are representatives of this specificity group. By examination ofthe determined consensus sequences collected in Table 1, it can be concluded that themost conserved basophilic determinant is the requirement of a basic amino acid, andmostly arginine, in position –3. Secondly, hydrophobic residues are outlined asimportant specificity determinants for basophilic kinases, and most importantlyrequirement for the large hydrophobic residue in position +1. Alternatively, thehydrophobic residues can be determinants, which provide selectivity between closelyrelated basophilic kinase forms, as has been shown in the case of PKA and PKG [16]. Inthis study it was found that the phenylalanine in position P+4 was a detrimental factorfor PKG, but not for PKA, being thus an illuminating example of the promising and stillhidden opportunities for selective substrate design for closely related kinases.

As a negative determinant for the basophilic group of kinases, the proline inposition P+1 should be mentioned. This crucially required determinant for the secondkinase specificity group overviewed below, is shown to be universally detrimental forbasophilic kinases [124].

6.2.2. Proline-directed protein kinases

The two main proline-directed kinase groups are the cyclin-dependent protein kinasesand the MAP kinases, both belonging to the CMGC group of protein kinasesuperfamily. As the general substrate motif for both subgroups is Ser/Thr-Pro, the CDKkinases additionally require the basic residues around the phosphorylation site. ForMAP kinases a second proline is often present in position P-2 [35] or even the third atposition P+2 as shown for Erk1 [145]. The peptide substrates derived from thephosphorylation site containing the motif PXT/SP, were poor substrates for MAP kinasep38, compared to its corresponding protein substrate [43]. This finding was expectedsince docking site interactions between the MAP kinases and their substrates has beenshown to play an important role in substrate phosphorylation [138]. Similarily, thesubstrate recognition of cyclin-dependent protein kinases is in large part determined bythe nature of cyclin associated with the kinase subunit [36].

22

6.2.3. Acidophilic and phosphate directed protein kinases

The acidophilic kinases are a considerably smaller group compared to the basophilickinases. The CK2 kinase (formerly casein kinase II) and GSK3 belong the CMGCgroup, while two other examples, the -adrenergic receptor kinase and the rhodopsinkinase belong to ACG group. The casein kinase 1 (CK1) is a family of enzymes notfalling into any one of the major groups according to the Hanks and Hunterclassification. An interesting example of substrate specificity of an acidophilic kinase isthe glycogen synthase kinase GSK3, whose crystal structure in complex withpseudosubstrate region was solved by two separate groups recently [18, 155]. GSK3requires the prephosphorylation of the substrate at position P+4 in a consensus motifS/TXXXSpX by CK2. Prior to the phophorylation by GSK3, the glycogen synthase isphosphorylated at Ser 656 by CK2. After this event the GSK3 will sequentiallyphosphorylate Ser652, Ser648, Ser644 and Ser640, each of which locates four positionsupstream from the previous one. Both solved crystal structures provide the structuralexplanation for this peculiar priming mechanism, showing that the phosphate added byCK2 kinase localises at the same site as the covalently attached phosphate in theactivation loop of many other protein kinases. This is possible because the GSK3 doesnot require phosphorylation of any residues in the activation loop for its activity, and asfor alternative, displays the described substrate-coupled activation mechanism. Thephosphate directed phosphorylation is also observed in case of CK1, but only in rarecases for CK2, which generally preferres carboxylic acids in recognition of itssubstrates [124].

6.3. Docking sites and anchoring proteins as secondary specificity determiningfactors of protein kinases

The specific recognition of substrates by a protein kinase may be partly achieved byhigh affinity docking sites located away from the phosphorylation site. Such amechanism is used by the MAP kinases [49, 138]. These docking motifs are rich inbasic amino acids and often contain a characteristic motif LXL. Depending on substrate,these motifs may be located 50-150 amino acids upstream or downstream from thephosphoacceptor motifs. An additional docking motif FXF is also present in manyMAP-kinase substrates. The phosphoacceptor motifs are usually sandwiched between adocking domain and FXF domain, forming the MAP-kinase recognition module.However many substrates contain only either the docking domain or FxF motif, andthere are substrates which do not contain these domains.

The “targeting hypothesis” proposes that the specificity of phosphorylationevents is controlled in part by the specific localisation of the kinases and phosphatasesin the cell [55]. This is determined by association of these enzymes with “targetingsubunits” or “anchoring proteins” [103]. Until now many of such anchoring proteinshave been isolated. For PKA they are called AKAPs (A kinase anchoring proteins) [26].

23

AKAPs contain two conserved structural modules: (i) a targeting domain that serves asscaffold and membrane anchor, and (ii) PKA specific domain interacting with theregulatory subunits. There are several variants of AKAPs, with different targetingspecificity and it has been shown that they may also bind other molecules, mainlyphosphatases, forming multiprotein complexes called “transduceosomes”, capable ofassembling and integrating signals from multiple pathways [26]. Similarily, in the caseof PKC, the anchoring proteins help to translocate the isoenzymes to specificsubcellular sites after stimulation with hormones or phorbol esters. Several proteins thatbind PKC have been cloned including several annexins, cytosceletal proteins andnuclear proteins [103]. A group of 30-36 kDa proteins collectively termed as RACKs(receptos for activated C-kinase) bind PKC through sites in the region of the C2 and V5domains [104, 148]. Differential localisation of PKC isoenzymes is determined byisoenzyme specific RACKs [75, 131]. Several PKC binding proteins are also substrateslike the myristoylated alanine rich C-kinase substrate (MARCKS) [60].

7. Inhibitors of protein kinases

In recent years much of attention has been paid to the design of the selective and potentinhibitors of protein kinases. The rise in interest has been caused mainly by thediscoveries revealing the involvement of protein kinases in several diseases.

There are five main classes of reversible protein kinase inhibitors:

a) ATP-site inhibitors are compounds competing with ATP for its bindingpocket in the enzyme active site cleft

b) Peptide-site inhibitors are compounds competing with the peptide or proteinsubstrate for their binding site on the enzyme molecule

c) Bi-substrate inhibitors are inhibitors directed simultaneously into the ATPand peptide binding sites

d) Inhibitors targeting regulatory domains by acting on the sites of allostericeffectors

e) Inhibitors blocking the docking sites for anchoring proteins, inhibiting thespecific localisation of a kinase (localisation inhibitors).

7.1. Protein and peptide site directed inhibitors

This class of inhibitors competes with the peptide or protein substrates for their bindingsites on the protein kinase molecule and is mostly comprised of peptides containing theelements of substrate binding motifs. The peptide inhibitors containing the proteinkinase P-site consensus motif are generally denoted as pseudosubstrate inhibitors.According to the definition, any peptide chain that folds back into the active site to turn

24

a protein kinase off using substrate like recognition groups can be termed aspseudosubstrate [71].

The first discovered and the most thoroughly studied members of thepseudosubstrate class of inhibitors are the heat stable protein kinase inhibitors (PKIs) ofPKA [165]. The PKIs contain the substrate-like sequence RRNAI, where thephosphoacceptor serine is replaced by alanine, and they possess nanomolar Ki values[165, 171]. Low nanomolar affinity is retained in a shorter synthetic variant PKI(5-24),but a variant containing only the short substrate consensus motif, the PKI(16-24), hasseveral hundred fold higher Ki than the PKI protein [33]. According to the crystalstructure of the PKA-ATP-PKI(5-24) complex [79] and studies on structurally alteredvariants of PKI, it was found that arginine in position P–6 also plays important role intight binding of the inhibitor [136]. The same studies prove the importance ofphenylalanine in position P–11, which is the determinant not generally present in knownpeptide substrates, and therefore, proposed to determine the tight binding of theinhibitor [34].

As described for the PKA, PKC and CDPK, in many protein kinases there areautoinhibitory regions outside the catalytic core, which act by mimicking the features ofthe protein substrate phosphorylation site. These regions have been identified bysearching for the consensus motifs in the sequences of regulatory domains, bytruncation mutagenesis or point mutations and proteolysis. Several synthetic peptidescorresponding to these sequences have been found to act as potent competitiveinhibitors with Ki values in the nanomolar or low micromolar range, including inhibitorsfor PKC [53], CaM-II [17] and soybean CDPK [41] and smooth muscle MLCK [78].

However, these inhibitors do not work in all cases. For example, the peptidesfrom the regulatory regions of PKA and PKG show very weak inhibitory effects with Ki

values in the millimolar range [73]. The mechanism of a kinase autoinhibition can beillustrated with the crystal structure of twichin kinase, a member of the MLCKsubfamily [54]. On the carboxy-terminal tail of the catalytic core of twichin kinase thereis a 60-residue autoinhibitory sequence which extends over the surface of the cleftbetween the two lobes of the catalytic core. There are three different functions in thisinhibitory interaction: (i) substrate-like binding contacts, (ii) steric superpositioningwith the ATP and (iii) contacts with residues crucial for catalysis. This shows thatdownregulation by a pseudosubstrate peptide chain is more complex than a simplebinding via a few interactions of the pseudosubstrate consensus motif, but involvesfunctional interactions all over the active site. Such a multi-contact binding mode mayexplain why in some cases shorter pseudosubstrate peptide fragments are not highaffinity inhibitors.

Sequences of the autoinhibitory regions have been used as templates forcombinatorial libraries, designed to obtaine selective and potent inhibitors for MLCK[93]. Another recent library study describes the design of highly specific peptideinhibitors for the cGMP-dependent protein kinase (PKG) by a rational stepwiseapproach using arrays of octameric peptide libraries on cellulose paper, screened for thebinding of 32P-autophosphorylated PKG [21]. The obtained peptideFLLKKKKKKHHK, did not contain the defined pseudosubstrate motif, but did contain

25

the basic specificity determinants and showed selectivity relative to PKA. This resultsuggests that the active site of protein kinases hides many possibilities for design ofhighly efficient ligands/inhibitors, and that the pseudosubstrate consensus motif is notan absolutely necessary element in these structures.

Due to their peptidic nature, the pseudosubstrate peptide inhibitors of proteinkinases are not classified as drug-like molecules and the main efforts in the field ofprotein kinase inhibitor design, largely promoted by pharmaceutical companies, havebeen applied to development of small-molecule ATP site directed inhibitors. However,quite a breakthrough for the practical application of peptide based inhibitors is theabove quoted recent work by Dostmann et al. [21]. The fusion peptides designed bycombining the peptide inhibitors for PKG and membrane translocation signals fromHIV-1 tat protein and Drosophila antennapedia homeo-domain, proved to be efficientlydelivered through the cell membrane. As an in vivo effect of the inhibitors, it was shownthat the NO-induced dilation was substantially decreased in pressurized cerebralarteries. The described library-based inhibitor design combined with novel methods forintracellular delivery is a promising way for modulation of protein kinase activities in invivo systems.

In several cases, the peptide substrates designed on the basis of thepseudosubstrate sequences have been shown to be very potent substrates. The PKI (14-22)S21 analogue is the best known peptide substrate for PKA with Km of 0.11 M[102]. On the other hand, the alanine substituted peptide based on the optimal peptidesubstrate Kemptide, LRRASLG, with Km value of 16 M [70], has very low inhibitoryeffect, with Ki of 320 M [101, 170]. This was not surprising since the Kemptide itselfhad relatively low affinity to the PKA catalytic subunit (Kd = 210 M) [170], which isan order of magnitude lower than would have been expected from the Km value. Thediscrepancy between the Km and Kd was found to be at least in part due to the fastchemical step of the phosphoryl transfer reaction [1]. The finding may seemdiscouraging with respect to the use of substrate specificity data for inhibitor design.However, this Km vs Kd difference is not general and was shown only for a very specificexample, the minimal short substrate Kemptide, not containing the specific arginine inP–6. On the other hand, Walsh and co-workers [33, 102] have shown that the Ki forPKI(14-22)amide, a peptide containing the P-6 arginine, is 4-fold lower than the Km forthe corresponding serine containing substrate. In order to reduce the possibility of beingmisled in the substrate specificity based inhibitor design, the use of aromatic andsecondary alcohols as phosphate acceptor groups in substrates was proposed [169].These substrates were less efficiently phosphorylated but displayed Km valuescomparable to their dissociation constants since the altered phosphoacceptor site causedthe phosphoryl transfer to be the rate limiting step.

Proceeding from the optimal structures for the peptide-competitors, severalattempts to obtain irreversible (covalent) inhibitors have been reported. The irreversibleinactivation of PKC via S-thiolation of an active site cystein in peptide RKRCLRRLwas shown [166]. The bound peptide hindered sterically catalysis of peptide and proteinsubstrates and even the PKC catalysed ATP hydrolysis [167]. A selective affinity labelthat inactivates PKG but not PKA was constructed using D-alanine instead of the

26

phosphoacceptor-site serine in a pseudosubstrate peptide with an aminoterminalcystein[3-nitro-2-pyridinesulfenyl-(Npys)]-amide [179].

For protein kinases from the MAP kinase cascades, the peptide inhibitors basedon the docking domain motifs and FxF motifs in substrates have been designed [50, 63,138]. Finally, nonpeptidic peptide/protein site sirected inhibitors have been described,comprising mainly of different polyanionic or polycationic compounds, inhibitingbasophilic or acidophilic protein kinases, respectively [4, 45, 120, 164]. An interestingexample is the inhibition of CK2 kinase by heparin. The kinetic mechanism of thisinhibition system has been previously studied by our group [113]. Despite the negativecharges in heparin, which most likely mimic the P-site substrate recognitiondeterminants for CK2, it was found that the inhibitor binding region does not overlapwith the binding site for the peptide substrate. On the other hand, it competitivelyhindered the binding of the protein substrate casein. This example shows that in thevicinity of the minimal peptide substrate binding site of a protein kinase, there may beother possible sites, which may be considered to be the docking sites of inhibitors.

7.2. ATP site directed inhibitors

ATP binding site in protein kinases is in the cleft between the two lobes of the enzymecatalytic domain. Due to the high conservation of the ATP binding pocket region, theinitial assumption was that it is not possible to design ATP-site directed inhibitors thatdifferentiate between different protein kinases. However, when several protein kinasecrystal structures became available, this opinion changed. Using crystallographic studiescoupled to computer modelling, regions inside the nucleotide binding pocket wereidentified, which allowed the design of more highly selective compounds [31]. Untilnow, both computer modelling and pharmacophore-based approach have helped todesign several classes of inhibitors which have an already noteworthy degree ofselectivity between closely related protein kinases [24, 30, 105, 127, 134, 160].

The first discovered potent ATP-site directed inhibitor was staurosporine, amicrobial alkaloid, isolated from Streptomyces staurosporeus [150]. This inhibitor hasKi values in the low nanomolar range for several protein kinases including PKA andPKC [99], and it has remained in wide use in research as a potent universal inhibitor ofprotein kinase activity. Various related compounds, developed on the basis ofstaurosporine have relatively good selectivity for several kinases. These includebisindolyl maleimides, indolcarbazoles, and bisindolylmaleimide macrocycles [65, 96,159, 172]. Another historically important class of protein kinase inhibitors is the H-series inhibitors, include different derivatives of isoquinolinesulphonamide, firstreported by Hidaka et al. [46]. Today, the crystal structures of the kinase-inhibitorcomplexes are available for the staurosporine in complex with PKA [126] and Cdk2[88] and three variants of complexes of the H-series of inhibitors with PKA [24]. Thesecrystallographic studies together with several other structures confirm the existence ofthree specificity regions in the adenosine binding pocket: the “hydrophobic pocket”, the

27

“ribose pocket”, and the “linker region”. The existence of the hydrophobic pocket isconfirmed by crystal structures of the p38 and FGFR protein kinases in complex withpyridinylimidasole and PD173074 inhibitors [105, 174]. The ”ribose pocket” is used bythe isoquinoline and dianalinophtalamide inhibitors [24]. The diversity of the structuralenvironment of the “linker region” is utilised by the purine based class of olomoucineinhibitors [134]. In general, there are two key components of the protein kinasestructure that contribute to the specificity of ATP-site directed inhibitors. (1) amino acidsequence diversity within the ATP binding cleft, and (2) conformational diversity,which allows inhibitors to “adopt” and “lock” the protein kinase into a specificconformation.

7.3. Bi-substrate inhibitors

Combination of the structural elements of the peptide-site inhibitors and ATP-siteinhibitors into one molecule by connecting them via a linker is the basic principle of thebi-substrate inhibitor design. The general principle of the bi-substrate inhibitor approachfor enzymes catalysing bimolecular reactions has been described in [8, 37]. A classicalsuccessful example of a bisubstrate inhibitor is the P1, P5-Di(adenosine-5’)pentaphosphate, a molecule combining two adenosines connected via apentaphosphate linker, which inhibits the adenylate kinase with Ki values in lownanomolar range [90]. The first attempt to apply the bi-substrate approach for design ofinhibitors for serine/threonine protein kinases was performed by Ricouart et al. [128] bycombining one of the derivatives of a H-series inhibitor to a tetra- and hexaargininepeptide anchors via different linkers. The best inhibitor for both PKA and PKC, withIC50 values of 3 and 300 nM, respectively, was found to be a isoquinoline-5-sulfonamide coupled to a Ser-Arg6 peptidic moiety via a –NH(CH2)2NH(CH2)2CO-linker. Further, an attempt was made to design transition-state analogue inhibitors,resembling the bi-substrate reaction complex of PKA, by coupling the serine residue ofthe specific PKA substrate Kemptide to adenosine nucleotides via phosphate linkers ofdifferent length [97]. The most potent of these compounds had an inhibitory effect inmicromolar range. Similarly, a moderately successful attempt, with inhibitory effects inthe low micromolar range, was performed to generate selective inhibitors for cdc2kinase using the bi-substrate approach by coupling potent ATP-site inhibitor 3,4-bis(indol-3-yl)maleimide analogues to a pseudosubstrate peptide Ac-C(S)PKK-NHMevia ethyloxy linkers of different length [133]. Thus, the exact structural requirements forsuccessful bi-substrate inhibitor design are still undefined and according to theknowledge available, the most crucial variable seems to be the length and chemicalnature of the linker region. A recent crystallographic study has shed some light onseveral of the unclear aspects of the bi-substrate approach [121]. In this study a bi-substrate inhibitor for the insulin receptor tyrosine kinase was synthesised by linkingATP S to a peptide substrate analogue via a two-carbon spacer. The design of thismolecule was inspired by the knowledge about the dissociative transition state of the

28

protein kinase catalysed phosphoryl transfer. The crystal structure confirmed thepredicted bi-substrate mode of binding and showed also that the linker takes part in theoctahedral coordination of an active site Mg2+.

Knowledge obtained from the studies quoted above, reveals that the idea toconnect two highly specific structural elements into one molecule via a linker region is apowerful method for potentiation of inhibitory effects. While the studies carried out sofar have used the ATP-site binding blocks originated from inhibitor structures ofmedium potency and the peptide parts with lower affinities compared to the mostoptimal pseudosubstrate structures, it is most probable that when combining the recentlydesigned ultrapotent ATP-site inhibitors via a suitable linker to the best inhibitorpeptides at an optimal connection point, the unprecedented inhibitory affinities could beachieved.

8. Affinity isolation of protein kinases using specific ligands targeted to substratesites

Specific ligands that bind to the substrate sites in the protein kinase active site groovehave been used for affinity purification of protein kinases. These attempts have shownsome success, but there has been no general breakthrough. A method using peptidesubstrates for isolation of glycogen synthase kinase and protein kinase CK2 has beenreported [175], but the low affinity of the kinases to the column did not allow stringentwashing conditions. Peptide inhibitor PKI has been used with relative success forpurification of catalytic subunit of PKA [115]. This method, required a long incubationtime for protein kinase binding and gave low yields of purification. The ATP bindingpocket directed affinity adsorbent -ATP-Sepharose has been tested for isolation ofseveral kinases [19, 44]. This method, however, cannot be strictly specific for proteinkinases, due to the high number of different enzymes utilizing adenosine nucleotides assubstrates and cofactors.

29

Present investigation

Aims of the thesisThis thesis focuses on study of the specificity of different model protein kinases at threedifferent levels using

a) novel bi-substrate analogue inhibitorsb) synthetic peptide substratesc) mutated protein substrates.

The specific aims for the projects were:to investigate the possibilities to obtain high affinity inhibitors and affinityligands for protein kinases by combination of structural elements of bothsubstrates into one bi-substrate analogue inhibitor molecule.to isolate the 59 kDa form of CDPK from maize seedlings and determine thesubstrate specificity consensus motif for this kinaseto develop the positionally oriented peptide library methodology forquantitative substrate specificity study of protein kinases using ESI-MS fordetection of phosphopeptides. To investigate the substrate specificity of PKCusing the new methodto study the phosphorylation site specificity determinants for PKA using thesite-dierected mutagenesis into the consensus motif positions of the L-typepyruvate kinase (L-PK) phosphorylation site and compare the kinetic dataobtained with the data on similar alterations performed on peptide substrates inprevious studies.

Paper I

Adenosine 5’-carboxylic acid peptidyl derivatives as inhibitors of protein kinases In collaboration with a research group in Tartu University, we have designed anddeveloped a novel class of inhibitors for protein kinases. These compounds contain thestructural elements of both substrates of protein kinases: the peptide part and theadenosine part. These two parts are coupled to each other via a linker. A set of eleveninhibitors was synthesised at the organic synthesis facility of Tartu University, by Dr.Asko Uri. The novel compounds were potent inhibitors of PKA and PKC with IC50values in nanomolar range. It was also shown that the two main building blocks of thesecompounds, the adenosine and hexaarginine peptide had inhibitory effects inmicromolar range indicating that the coupling of these two structural elements into onebi-substrate analogue has a positive synergistic effect on the inhibitory efficiency. Theselective effect was also gained as these compound were relatively poor inhibitors forprotein kinases with acidic substrate specificity determinants CK1 and CK2(IC50>30 M). A third basophilic protein kinase, CDPK from maize, was also inhibitedby the new compounds but the inhibitory effects were in micromolar range. It was

30

found that the hexaarginine peptide moiety was necessary for the high affinity bindingof the inhibitors, since the compounds with four and two arginines showed ten and onehundred-fold weaker inhibitory effects, respectively. The main focus of this study wasto optimise the requirement of the linker region of the novel bi-substrate inhibitors ofgeneral structure AdoC-Linker-Arg6. The linkers of different charge and length weretested. For introducing negative charges, the aspartic acids were used, and for positivecharges the amino acids containing secondary amines were included. Thirdly, thealiphatic chains of different length were tested as linkers. It was found that the optimallinker structures for PKA and PKC were the aliphatic chains comprised of six and eightcarbons, respectively. Interestingly, in spite of much weaker inhibitory effects forCDPK compared to PKA and PKC, the structural preferences for the linker region wereclosely similar for all three enzymes, pointing to the common structural features of theregion locating the linker chain of the bound inhibitor. Since CDPK is quite a distantkinase relatively to PKA and PKC, belonging to the CaMKII group of the protein kinasesuperfamily, the results obtained for linker optimisation may apply universally to thekinase superfamily. In this case, the AdoC-linker part could stay unchanged in thedesign process and the peptide anchor could be varied according to the peptide substratespecificity of the kinase of interest resulting in selective inhibitors. For such variationsthe peptide substrate libraries studies are a potential source of valuable information.

Paper II

Bi-substrate analogue ligands for affinity chromatography of protein kinasesNovel purification methods are needed to characterize the growing number ofdiscovered protein kinases. Relatively slow progress in this area can partly be related tocomplex purification protocols, which use conventional chromatographic steps forprotein isolation (ion-exchange chromatography, hydrophobic chromatography, sizeexclusion chromatography, etc.). Attempts to use affinity chromatography for thispurpose have still not lead to a general breakthrough. In this paper we proposed a novelaffinity medium for purification of PKA. This medium was based on the application ofa new type of protein kinase bi-substrate inhibitor, described in paper I of the currentthesis. Two ligands, AdoC-Aoc-Arg4-Lys (ligand A), with Ki value of 330 nM for PKAand AdoC-Aoc-Arg4- NH(CH2)6NH2 (ligand B), with Ki value of 83 nM, were designedfor purification of PKA. The ligands A and B were coupled to the epoxy-activatedSepharose and NHS-activated Sepharose, respectively. The efficiency of the media wastested for purification of recombinant catalytic subunit of PKA from bacterial extract.Both MgATP and L-arginine were successfully used as eluting agents yieldinghomogeneous PKA preparation. The binding of PKA was not distorted by 1 M NaClconfirming the specific mode of binding. PKA was also isolated from pig heart usingthe column with ligand B. Since PKA is a very sparse enzyme in tissues, a preliminarystep of DE-52 was included for elimination of a part of the proteins present in thesupernatant of the heart homogenate and thereby the PKA was enriched. After the

31

subsequent affinity chromatography step a preparation of PKA was obtained with noessential contaminants as visualised in SDS-page. Additionally it was found that theaffinity column did not bind the purified protein kinase CK2, which belongs to proteinkinase class with acidic specificity determinants. No binding of proteins possessingnucleotide binding site (the L-type pyruvate kinase), or sites for wide variety ofdifferent ligands (the bovine serum albumin) was observed. These findings point to highselectivity of the bi-functional binding mode of the affinity ligand.

Paper III

Peptide phosphorylation by Ca2+-dependent protein kinase (CDPK-1) from maizeseedlingsA 59 kDa form of Ca2+-dependent protein kinase (CDPK-1) was partially purified frommaize seedlings and the substrate specificity of this enzyme was studied using a set ofsynthetic peptides, derived from the phosphorylatable sequence RVLSRLHS(15)VRERof maize sucrose synthase 2. The isolated enzyme was activated by Ca2+ -ions at lowmicromolar range, but its activity was not stimulated by the phospholipid-diacylglycerolmixture. The preparation showed both a single band of autophosphorylation and a singleimmunoreactive band with the monoclonal antibody rised against the calmodulin-likedomain of the soybean CDPK, confirming that the CDPK preparation contained onlyone protein kinase. The molecular size estimated from both autophosphorylation andblotting experiments was close to the predicted molecular mass value of 59.4 kDa for acloned CDPK from maize seedlings [132]. The predicted molecular mass of 60 kDa hasbeen reported also for a cloned maize pollen-specific CDPK [25]. The purified CDPKwas different from the 54 kDa CDPK form stimulated by phospholipids [149]. Besidesthe molecular size and the activation conditions, the characteristic difference betweenthese two major forms of CDPK in maize seedlings was the different ionic strengthrequired for elutions of the two enzyme forms from the DEAE ion exchanger.

For study of the substrate specificity of the newly isolated kinase we applied twosequential approaches. First, a series of synthetic peptides with different lengths,shortened by one amino acid at a time, was used for determination of a minimal peptidepeptide substrate which would be efficiently phosphorylatable by the kinase. In thisway, the deca-peptide LARLHSVRER, was identified as a minimal unit containing thecritical determinants for recognition by kinase and showing Vmax/Km value still in thesame order of magnitude as the parent peptide. The same set of peptides was tested assubstrates for mammalian PKC , and it was found that the phosphorylation motif fromsucrose synthase 2 also contains a specific substrate motif for this kinase. Furthermore,PKC had similar optimal peptide length requirement to CDPK-1. Conversely, thesepeptides revealed low reactivity in the case of PKA.

In the second step of the study a systematic analysis of the positional specificityof CDPK-1 within the minimal substrate structure LARLHSVRER was performed usingpositionally varied substrate sets synthesised as spots on cellulose membranes

32

(SPOTsTM membranes). This method has been developed and used previously by Frankand co-workers [29] and also partly in our laboratory by Toomik et al. [157]. Thesystematic variation of the hydrophobic (Ala, Leu) and ionic (Arg, Glu) amino acidssubsequently in each position of the minimal peptide lead to the definition of theconsensus sequence motif A/LXRXXSXRZR, for CDPK-1, where X denotes a positionwith no strict amino acid requirements and Z a position strictly not tolerating argininecompared to the other three varied amino acids. The consensus motif for maize CDPK-1was in agreement with the results of specificity studies made with CDPK forms fromwheat germ [125], soybean [89] and from a lower plant Mougeotia [129], where thelocation of basic amino acids in position –3 from the substrate phosphorylation siteseemed to be critical for substrate recognition. The leucine in position –5 has beenshown to be an important substrate recognition element for two forms of CDPK fromspinach leaf [2]. However, the CDPK forms from these plants did not exhibit any strictrequirements for basic or acidic amino acids at the C-terminal side of the serine. Thusthe fragment RZR, in positions from +2 to +4 seems to be specifically characteristic formaize CDPK-1. The recognition of this motif was probably based on electrostaticinteractions, as the enzyme did not tolerate amino acids of opposite charge in thisregion. However, this fragment alone was not sufficient for efficient phosphorylation ofpeptides. For example, the peptides that lack the Leu in position –5 but possess thefragment RER, were poor substrates for CDPK-1. From these observations it can beconcluded that CDPK-1 has complex and rather original substrate specificity.Interestingly, the obtained consensus motif was close to the analogous sequenceK/RXXS/TXK/R proposed for mammalian PKC, but different from the consensus motifRRXS/TX for PKA.

Paper IV

Systematic analysis of substrate specificity profile of protein kinase C using apositionally oriented peptide library approach coupled with electrospray mass-spectrometryIn the this study a positionally oriented peptide library approach based on ESI-MSdetection of phosphopeptides in initial velocity conditions was designed for quantitativekinetic characterization of protein kinase specificity profiles. The leading structure ofthe peptide library, LLRLHSLRER, is a slightly modified sequence of aphosphorylation site of maize sucrose synthase. Peptides derived from this sequencewere previously shown to be potent substrates for maize calcium-dependent proteinkinase (CDPK) and also for mammalian PKC.

Each subsequent position in the peptide library was successively substituted witha mixture of amino acids resulting in nine different libraries. PKC was used as a modelenzyme for optimization of the method. For quantitative detection of thephosphopeptides in true initial velocity conditions, the ESI-MS was used. It was shownthat five amino acid positions upstream and four downstream from the serine had

33

defined and sensitive structural specificity for PKC, and that the positively chargedarginines showed strong preference over negatively charged glutamic acid in all variedpositions. On the basis of the data obtained, an optimal peptide substrate for PKC,FRRRRSFRRR, was designed, showing the Km value in nanomolar range.

Additionally, the possibilities to use the obtained quantitative specificity profilefor prediction of phosphorylation sites for PKC by screening the protein sequencedatabases were analysed. It was found that the optimal substrate structures for PKC arevery rare in known protein substrates and the majority of phosphorylation site sequencesin proteins contain only a minimal number of substrate specificity determinants.

Paper V

Phosphorylation of peptide and protein substrates by protein kinase AThe results from papers III and IV showed that the substrate specificity profiles ofprotein kinases obtained from the studies with synthetic peptide substratescorresponding to the sequences of phosphorylation sites in protein substrates may reveala sensitive profile of structural requirements for amino acid positions surrounding thephosphorylatable residue. However, the discussion in the paper IV revealed that thesespecificity profiles are not strictly present in many phosphorylation sites in proteins.Therefore, for validation of the possibilities for prediction of phosphorylation sites inproteins it is important to estimate the impact of both the primary structure sitespecificity and that of secondary determinants, such as protein-protein surface contactsoutside of the phosphporylation site. In the paper V, we addressed this question bystudying a kinase P-site specificity using site-directed mutagenesis in a phosphorylationsite of a physiological substrate of PKA, L-type pyruvate kinase (L-PK). Amino acidswith basic, acidic, small and large hydrophobic side chains, and also with hydrophilicside chains were systematically introduced in the positions P-3, P-2 and P+1 ofphosphorylation motif RRASV of L-PK and the purified mutated proteins were testedas substrates for PKA. The obtained kinetic data confirmed the basic substratespecificity requirements for PKA, i.e. the preference of arginine in P-2 and P-3 and andhydrophobic residue in P+1 position. Comparision of the relative logarithmic values ofthe obtained kinetic data with the available data on similar structural alterations inseparately tested peptide substrates collected from literature revealed that amino acidalterations in peptides cause stronger effects on the phosphorylation rate than thecorresponding alterations in protein substrate L-PK.

This finding indicates that the importance of the primary structure specificityprofile around the phosphorylation site in protein substrates may be overestimated andsimilarily to the conclusion from paper IV, other structural factors play an importantrole in the efficiency of protein phosphorylation reactions. These factors may be thedocking sites or additional docking proteins for protein kinases. But, importantly, in thecase of L-PK, these factors are most plausibly the nonspecific interactions between thesurfaces of two encountering proteins, since there are no specific docking sites known,

34

which would mediate the recognition of L-PK and PKA. Since the phosphorylationrates of the wild type L-PK and wild type peptide RRASVA did not differ very much, itmay conceived that these additional interactions between the surfaces of the twoencountering proteins compensate the negative effect of unfavorable residues inconsensus motif positions, which otherwise when introduced in peptide substrateswould be detrimental for the reactivity. In this context, the number of potentialphysiological phosphorylation sites should be very high, possibly making the kinasestoo promiscuous mediators for signaling systems requiring exactness. However, thisdrawback is overcome and the specificity of kinase action is shown to be determined bycompartmentation and specific subcellular targeting of kinase, as shown for PKAsystem exemplified by the A-kinase anchoring proteins (AKAPs).

Future perspectives

Further development of the bi-substrate analogue inhibitors by varying the structure ofthe peptide anchor may lead to design of selective inhibitors for protein kinases. Thestructure of the peptide anchor could be optimised according to the results from the ESI-MS coupled peptide library studies for different kinases. By this way even minorsubstrate specificity differences can be used for modulation of selectivity of the bi-substrate molecules. Such substrate specificity based inhibitor design would combinethe two main themes of the current thesis. The obtained inhibitors may be used forfurther optimisation of the group specific affinity isolation methods for protein kinases.

On the other hand, the highly detailed specificity data possible to obtain from thepeptide library methodology, designed for kinetic measurements in initial velocityconditions, opens new perspectives for isoenzyme specific substrate design. Thedesigned substrates can be used for high-throughput protein kinase activity assays fordrug screening or for validation of protein kinase activity profiles as potential biomarkerpanels for different diseases.

35

Acknowledgements

This study was carried out at the Department of Medical Biochemistry andMicrobiology, Uppsala University, in collaboration with the Laboratory of BioorganicChemistry at Tartu University, Estonia.

I would like to express my sincere gratitude to all the people who have supported methroughout this study.

In particular I would like to thank Docent Pia Ek for being an excellent supervisor, forsupport and stimulating discussions.I would like to thank my supervisor Professor Jaak Järv, for introducing me to thepractical field of biochemistry, for establishing my connection with Uppsala University,and for many good ideas and useful discussions.Special thanks to:Doctor Asko Uri for being a good collaboration partner.Professor Lars Rask, my examinator, for guidance and good suggestions.Professor Örjan Zetterqvist and Professor Lorentz Engström, the classics in the field ofprotein phosphorylation studies, for the possibility to learn from the collective practicalexperience settled in the lab during the past decads.Doctor Bo Ek, for helping me with ESI-MS experiments.Present and past members of our group: Reet, Gunilla, Susanne, Inger, Stefan, Nikita,Fergal.Special thanks to Åsa for good company and help.Special thanks to Carina for keeping mood up and for helping me with mutagenesis.

I would like to thank all my fellow-Estonians in Uppsala, and specially: Annika, Anu,Anneli, Sunne, Tanel Käämbre, Tanel Punga, Teet, Taavo and Priit.My fellow-Estonians in Stockholm: Anton and Ursel.The people at the laboratory of Bioorganic Chemistry, Tartu University: Gerda, Katrin,Marika, Ago, Andres and Martin.

36

References

1. Adams, J. A. and Taylor, S. S. (1992) Energetic limits of phosphotransfer in the catalytic subunit ofcAMP-dependent protein kinase as measured by viscosity experiments. Biochemistry 31, 8516-22.

2. Bachmann, M., Shiraishi, N., Campbell, W. H., Yoo, B.-C., Harmon, A. C. and Huber, S. C.(1996) Identification of Ser-543 as the major regulatory phosphorylation site in spinach leaf nitratereductase. Plant Cell 8, 505-517.

3. Beebe, S. J., Oyen, O., Sandberg, M., Froysa, A., Hansson, V. and Jahnsen, T. (1990) Molecularcloning of a tissue-specific protein kinase (C gamma) from human testis--representing a third isoformfor the catalytic subunit of cAMP-dependent protein kinase. Mol. Endocrinol. 4, 465-75.

4. Benovic, J. L., Stone, W. C., Caron, M. G. and Lefkowitz, R. J. (1989) Inhibition of the beta-adrenergic receptor kinase by polyanions. J. Biol. Chem. 264, 6707-10.

5. Blume-Jensen, P. and Hunter, T. (2001) Oncogenic kinase signalling. Nature 411, 355-65.6. Bossemeyer, D., Engh, R. A., Kinzel, V., Ponstingl, H. and Huber, R. (1993) Phosphotransferase

and substrate binding mechanism of the cAMP-dependent protein kinase catalytic subunit fromporcine heart as deduced from the 2.0 A structure of the complex with Mn2+ adenylylimidodiphosphate and inhibitor peptide PKI(5-24). EMBO J. 12, 849-59.

7. Breviario, D., Morello, L. and Giani, S. (1995) Molecular cloning of two novel rice cDNAsequences encoding putative calcium-dependent protein kinases. Plant Mol. Biol. 27, 953-67.

8. Broom, A. D. (1989) Rational design of enzyme inhibitors: multisubstrate analogue inhibitors. J.Med. Chem. 32, 2-7.

9. Cadd, G. and McKnight, G. S. (1989) Distinct patterns of cAMP-dependent protein kinase geneexpression in mouse brain. Neuron 3, 71-9.

10. Carter, C. A. (2000) Protein kinase C as a drug target: implications for drug or diet prevention andtreatment of cancer. Curr. Drug. Targets 1, 163-83.

11. Chan, P. M. and Miller, W. T. (1998) Use of peptide libraries to determine optimal substrates oftyrosine kinases. Methods Mol. Biol. 84, 75-86.

12. Chan, P. M., Nestler, H. P. and Miller, W. T. (2000) Investigating the substrate specificity of theHER2/Neu tyrosine kinase using peptide libraries. Cancer Lett. 160, 159-69.

13. Cheng, X., Ma, Y., Moore, M., Hemmings, B. A. and Taylor, S. S. (1998) Phosphorylation andactivation of cAMP-dependent protein kinase by phosphoinositide-dependent protein kinase. Proc.Natl. Acad. Sci. U S A 95, 9849-54.

14. Cho, Y. S., Lee, Y. N. and Cho-Chung, Y. S. (2000) Biochemical characterization of extracellularcAMP-dependent protein kinase as a tumor marker. Biochem. Biophys. Res. Commun. 278, 679-84.

15. Clegg, R. A., Gordge, P. C. and Miller, W. R. (1999) Expression of enzymes of covalent proteinmodification during regulated and dysregulated proliferation of mammary epithelial cells: PKA, PKCand NMT. Adv. Enzyme. Regul. 39, 175-203.

16. Colbran, J. L., Francis, S. H., Leach, A. B., Thomas, M. K., Jiang, H., McAllister, L. M. andCorbin, J. D. (1992) A phenylalanine in peptide substrates provides for selectivity between cGMP-and cAMP-dependent protein kinases. J. Biol. Chem. 267, 9589-94.

17. Colbran, R. J., Fong, Y. L., Schworer, C. M. and Soderling, T. R. (1988) Regulatory interactionsof the calmodulin-binding, inhibitory, and autophosphorylation domains of Ca2+/calmodulin-dependent protein kinase II. J. Biol. Chem. 263, 18145-51.

37

18. Dajani, R., Fraser, E., Roe, S. M., Young, N., Good, V., Dale, T. C. and Pearl, L. H. (2001)Crystal structure of glycogen synthase kinase 3 beta: structural basis for phosphate-primed substratespecificity and autoinhibition. Cell 105, 721-32.

19. Davies, S. P., Hawley, S. A., Woods, A., Carling, D., Haystead, T. A. and Hardie, D. G. (1994)Purification of the AMP-activated protein kinase on ATP-gamma-sepharose and analysis of itssubunit structure. Eur. J. Biochem. 223, 351-7.

20. De Bondt, H. L., Rosenblatt, J., Jancarik, J., Jones, H. D., Morgan, D. O. and Kim, S. H. (1993)Crystal structure of cyclin-dependent kinase 2. Nature 363, 595-602.

21. Dostmann, W. R., Taylor, M. S., Nickl, C. K., Brayden, J. E., Frank, R. and Tegge, W. J. (2000)Highly specific, membrane-permeant peptide blockers of cGMP-dependent protein kinase Ialphainhibit NO-induced cerebral dilation. Proc. Natl. Acad. Sci. U S A 97, 14772-7.

22. Douglas, P., Pigaglio, E., Ferrer, A., Halfords, N. G. and MacKintosh, C. (1997) Three spinachleaf nitrate reductase-3-hydroxy-3-methylglutaryl-CoA reductase kinases that are required byreversible phosphorylation and/or Ca2+ ions. Biochem. J. 325, 101-9.

23. Ellard-Ivey, M., Hopkins, R. B., White, T. J. and Lomax, T. L. (1999) Cloning, expression and N-terminal myristoylation of CpCPK1, a calcium-dependent protein kinase from zucchini (Cucurbitapepo L.). Plant Mol. Biol. 39, 199-208.

24. Engh, R. A., Girod, A., Kinzel, V., Huber, R. and Bossemeyer, D. (1996) Crystal structures ofcatalytic subunit of cAMP-dependent protein kinase in complex with isoquinolinesulfonyl proteinkinase inhibitors H7, H8, and H89. Structural implications for selectivity. J. Biol. Chem. 271, 26157-64.

25. Estruch, J. J., Kadwell, S., Merlin, E. and Crossland, L. (1994) Cloning and characterization of amaize pollen-specific calcium-dependent calmodulin-independent protein kinase. Proc. Natl. Acad.Sci. U S A 91, 8837-41.

26. Feliciello, A., Gottesman, M. E. and Avvedimento, E. V. (2001) The biological functions of A-kinase anchor proteins. J. Mol. Biol. 308, 99-114.

27. Feramisco, J. R., Glass, D. B. and Krebs, E. G. (1980) Optimal spatial requirements for thelocation of basic residues in peptide substrates for the cyclic AMP-dependent protein kinase. J. Biol.Chem. 255, 4240-5.

28. Fischer, E. H., Graves, D. J., Snyder Crittender, E. R. and Krebs, E. G. (1959) Structure of thesite phosphorylated in the phosphorylase b to a reaction. J. Biol. Chem. 234, 1698-1704.

29. Frank, R. (1992) Spot-synthesis: an easy technique for the positionally addressable, parallelchemical synthesis on a membrine support. Tetrahedron 48, 9217-32.

30. Fry, D. W., Kraker, A. J., McMichael, A., Ambroso, L. A., Nelson, J. M., Leopold, W. R.,Connors, R. W. and Bridges, A. J. (1994) A specific inhibitor of the epidermal growth factorreceptor tyrosine kinase. Science 265, 1093-5.

31. Furet, P., Caravatti, G., Lydon, N., Priestle, J. P., Sowadski, J. M., Trinks, U. and Traxler, P.(1995) Modelling study of protein kinase inhibitors: binding mode of staurosporine and origin of theselectivity of CGP 52411. J. Comput. Aided Mol. Des. 9, 465-72.

32. Gausepohl, H., Boulin, C., Kraft, M. and Frank, R. W. (1992) Automated multiple peptidesynthesis. Pept. Res. 5, 315-20.

33. Glass, D. B., Cheng, H. C., Mende-Mueller, L., Reed, J. and Walsh, D. A. (1989) Primarystructural determinants essential for potent inhibition of cAMP-dependent protein kinase by

38

inhibitory peptides corresponding to the active portion of the heat-stable inhibitor protein. J. Biol.Chem. 264, 8802-10.

34. Glass, D. B., Lundquist, L. J., Katz, B. M. and Walsh, D. A. (1989) Protein kinase inhibitor-(6-22)-amide peptide analogs with standard and nonstandard amino acid substitutions for phenylalanine10. Inhibition of cAMP-dependent protein kinase. J. Biol. Chem. 264, 14579-84.

35. Gonzalez, F. A., Raden, D. L. and Davis, R. J. (1991) Identification of substrate recognitiondeterminants for human ERK1 and ERK2 protein kinases. J. Biol. Chem. 266, 22159-63.

36. Gould, K. L. (1994) in Protein Kinases (Woodgett, J. R., ed.), pp. 149-176, Oxford University Press37. Hampton, A., Hai, T. T., Kappler, F. and Chawla, R. R. (1982). Species- or isozyme-specific

enzyme inhibitors. 6. Synthesis and evaluation of two-substrate condensation products as inhibitorsof hexokinases and thymidine kinases. J. Med. Chem. 25, 801-5.

38. Hanks, S. K. and Hunter, T. (1995) in The protein kinase factsbook (Hardie, D. G. and Hanks, S.K., eds.), pp. 7-47, Academic Press, London

39. Hanks, S. K. and Hunter, T. (1995) Protein kinases 6. The eukaryotic protein kinase superfamily:kinase (catalytic) domain structure and classification. Faseb J. 9, 576-96.

40. Hanks, S. K., Quinn, A. M. and Hunter, T. (1988) The protein kinase family: conserved featuresand deduced phylogeny of the catalytic domains. Science 241, 42-52.

41. Harmon, A. C., Yoo, B. C. and McCaffery, C. (1994) Pseudosubstrate inhibition of CDPK, aprotein kinase with a calmodulin-like domain. Biochemistry 33, 7278-87.

42. Harper, J. F., Sussman, M. R., Schaller, G. E., Putnam-Evans, C., Charbonneau, H. andHarmon, A. C. (1991) A calcium-dependent protein kinase with a regulatory domain similar tocalmodulin. Science 252, 951-4.

43. Hawkins, J., Zheng, S., Frantz, B. and LoGrasso, P. (2000) p38 map kinase substrate specificitydiffers greatly for protein and peptide substrates. Arch. Biochem. Biophys. 382, 310-3.

44. Haystead, C. M., Gregory, P., Sturgill, T. W. and Haystead, T. A. (1993) Gamma-phosphate-linked ATP-sepharose for the affinity purification of protein kinases. Rapid purification tohomogeneity of skeletal muscle mitogen-activated protein kinase kinase. Eur. J. Biochem. 214, 459-67.

45. Hegazy, M. G., Schlender, K. K., Wilson, S. E. and Reimann, E. M. (1989) Inhibitory effect ofpolycations on phosphorylation of glycogen synthase by glycogen synthase kinase 3. Biochim.Biophys. Acta 1011, 198-204.

46. Hidaka, H., Inagaki, M., Kawamoto, S. and Sasaki, Y. (1984) Isoquinolinesulfonamides, noveland potent inhibitors of cyclic nucleotide dependent protein kinase and protein kinase C.Biochemistry 23, 5036-41.

47. Hoffmann, J. (1997) The potential for isoenzyme-selective modulation of protein kinase C. FASEBJ. 11, 649-669.

48. Hojo, H., Morimura, Y., Abe, M. and Wakasa, H. (1996) Expression of classical protein kinase Csubspecies in non-neoplastic lymphocytes and non-Hodgkin's lymphomas: an immunohistochemicalstudy. Pathol. Int. 46, 148-54.

49. Holland, P. M. and Cooper, J. A. (1999) Protein modification: docking sites for kinases. Curr.Biol. 9, R329-31.

50. Horiuchi, K. Y., Scherle, P. A., Trzaskos, J. M. and Copeland, R. A. (1998) Competitiveinhibition of MAP kinase activation by a peptide representing the alpha C helix of ERK.Biochemistry 37, 8879-85.

39

51. Houghten, R. A., Appel, J. R., Blondelle, S. E., Cuervo, J. H., Dooley, C. T. and Pinilla, C.(1992) The use of synthetic peptide combinatorial libraries for the identification of bioactivepeptides. Biotechniques 13, 412-21.

52. Houghten, R. A., Pinilla, C., Blondelle, S. E., Appel, J. R., Dooley, C. T. and Cuervo, J. H.(1991) Generation and use of synthetic peptide combinatorial libraries for basic research and drugdiscovery. Nature 354, 84-6.

53. House, C. and Kemp, B. E. (1990) Protein kinase C pseudosubstrate prototope: structure-functionrelationships. Cell Signal. 2, 187-90.

54. Hu, S. H., Parker, M. W., Lei, J. Y., Wilce, M. C., Benian, G. M. and Kemp, B. E. (1994)Insights into autoregulation from the crystal structure of twitchin kinase. Nature 369, 581-4.

55. Hubbard, M. J. and Cohen, P. (1993) On target with a new mechanism for the regulation of proteinphosphorylation. Trends Biochem. Sci. 18, 172-7.

56. Hubbard, S. R., Wei, L., Ellis, L. and Hendrickson, W. A. (1994) Crystal structure of the tyrosinekinase domain of the human insulin receptor. Nature 372, 746-54.

57. Huber, S. C., Bachmann, M. and Huber, J. L. (1996) Post-translational regulation of nitratereductase activity: a role for Ca2+ and 14-3-3 proteins. Trends Plant Sci. 1, 432-438.

58. Huber, S. C., Huber, J. L., Liao, P.-C., Gage, D. A., McMichael, R. W. J., Chourey, P. S.,Hannah, L. C. and Koch, K. (1996) Phosphorylation of Serine-15 of maize leaf sucrose synthase.Plant Physiol. 112, 793-802.

59. Huyer, G., Kelly, J., Moffat, J., Zamboni, R., Jia, Z., Gresser, M. J. and Ramachandran, C.(1998) Affinity selection from peptide libraries to determine substrate specificity of protein tyrosinephosphatases. Anal. Biochem. 258, 19-30.

60. Hyatt, S. L., Klauck, T. and Jaken, S. (1990) Protein kinase C is localized in focal contacts ofnormal but not transformed fibroblasts. Mol. Carcinog. 3, 45-53.

61. Inoue, M., Kishimoto, A., Takai, Y. and Nishizuka, Y. (1977) Studies on a cyclic nucleotide-independent protein kinase and its proenzyme in mammalian tissues. II. Proenzyme and its activationby calcium-dependent protease from rat brain. J. Biol. Chem. 252, 7610-6.

62. Iwata, Y., Kuriyama, M., Nakakita, M., Kojima, H., Ohto, M. and Nakamura, K. (1998)Characterization of a calcium-dependent protein kinase of tobacco leaves that is associated with theplasma membrane and is inducible by sucrose. Plant Cell Physiol. 39, 1176-83.

63. Jacobs, D., Glossip, D., Xing, H., Muslin, A. J. and Kornfeld, K. (1999) Multiple docking sites onsubstrate proteins form a modular system that mediates recognition by ERK MAP kinase. GenesDev. 13, 163-75.

64. Jeffrey, P. D., Russo, A. A., Polyak, K., Gibbs, E., Hurwitz, J., Massague, J. and Pavletich, N. P.(1995) Mechanism of CDK activation revealed by the structure of a cyclinA-CDK2 complex. Nature376, 313-20.

65. Jirousek, M. R., Gillig, J. R., Gonzalez, C. M., Heath, W. F., McDonald, J. H., 3rd, Neel, D. A.,Rito, C. J., Singh, U., Stramm, L. E., Melikian-Badalian, A., Baevsky, M., Ballas, L. M., Hall, S.E., Winneroski, L. L. and Faul, M. M. (1996) (S)-13-[(dimethylamino)methyl]-10,11,14,15-tetrahydro-4,9:16,21-dimetheno-1H,13H-dibenzo[e,k]pyrrolo[3,4-][1,4,13]oxadiazacyclohexadecene-1,3(2H)-d ione (LY333531) and related analogues: isozyme selective inhibitors of protein kinase Cbeta. J. Med.Chem. 39, 2664-71.

66. Johnson, L. N., Noble, M. E. and Owen, D. J. (1996) Active and inactive protein kinases: structuralbasis for regulation. Cell 85, 149-58.

40

67. Järv, J. and Ragnarsson, U. (1991) Linear free energy relationships in cAMP-dependent proteinkinase reactions with synthetic substrates. Bioorg. Chem. 19, 77-87.

68. Järv, J., Sak, K., Eller, M., Ek, P., Engström, Å. and Engström, L. (1996) Quantitative structure-activity relationships in the protein kinase C reaction with synthetic peptides derived from myelinbasic protein. Bioorg. Chem. 24, 159-168.

69. Kameshita, I. and Fujisawa, H. (1995) Preparation and characterization of calmodulin-dependentprotein kinase IV (CaM-kinase IV) free of CaM-kinase IV kinase from rat cerebral cortex. J.Biochem. (Tokyo) 117, 85-90.

70. Kemp, B. E., Graves, D. J., Benjamini, E. and Krebs, E. G. (1977) Role of multiple basic residuesin determining the substrate specificity of cyclic AMP-dependent protein kinase. J. Biol. Chem. 252,4888-94.

71. Kemp, B. E., Parker, M. W., Hu, S., Tiganis, T. and House, C. (1994) Substrate andpseudosubstrate interactions with protein kinases: determinants of specificity. Trends Biochem. Sci.19, 440-4.

72. Kemp, B. E. and Pearson, R. B. (1991) Design and use of peptide substrates for protein kinases.Methods Enzymol. 200, 121-34.

73. Kemp, B. E., Pearson, R. B. and House, C. M. (1991) Pseudosubstrate-based peptide inhibitors.Methods Enzymol. 201, 287-304.

74. Kennelly, P. J. and Krebs, E. G. (1991) Consensus sequences as substrate specificity determinantsfor protein kinases and protein phosphatases. J. Biol. Chem. 266, 15555-8.

75. Kiley, S. C., Jaken, S., Whelan, R. and Parker, P. J. (1995) Intracellular targeting of proteinkinase C isoenzymes: functional implications. Biochem. Soc. Trans. 23, 601-5.

76. Kiley, S. C., Welsh, J., Narvaez, C. J. and Jaken, S. (1996). Protein kinase C isozymes andsubstrates in mammary carcinogenesis. J Mammary Gland Biol. Neoplasia 1, 177-87.

77. Knighton, D. R., Bell, S. M., Zheng, J., Ten Eyck, L. F., Xuong, N.-h., Taylor, S. S. andSowadski, J. M. (1993) 2.0 Arefined crystal structure of the catalytic subunit of cAMP-dependentprotein kinase complexed with a peptide inhibitor and detergent. Acta Cryst. D49, 357.

78. Knighton, D. R., Pearson, R. B., Sowadski, J. M., Means, A. R., Ten Eyck, L. F., Taylor, S. S.and Kemp, B. E. (1992) Structural basis of the intrasteric regulation of myosin light chain kinases.Science 258, 130-5.

79. Knighton, D. R., Zheng, J. H., Ten Eyck, L. F., Ashford, V. A., Xuong, N. H., Taylor, S. S. andSowadski, J. M. (1991) Crystal structure of the catalytic subunit of cyclic adenosinemonophosphate-dependent protein kinase [see comments]. Science 253, 407-14.

80. Koren, R., Langzam, L., Paz, A., Livne, P. M., Gal, R. and Sampson, S. R. (2000) Protein kinaseC (PKC) isoenzymes immunohistochemistry in lymph node revealing solution-fixed, paraffin-embedded bladder tumors. Appl. Immunohistochem. Molecul. Morphol. 8, 166-71.

81. Kreegipuu, A., Blom, N., Brunak, S. and Järv, J. (1998) Statistical analysis of protein kinasespecificity determinants. FEBS Lett. 430, 45-50.

82. Lam, K. S. (1998) Determination of peptide substrate motifs for protein kinases using a "one-beadone-compound" combinatorial library approach. Methods Mol. Biol. 87, 83-6.

83. Lam, K. S., Salmon, S. E., Hersh, E. M., Hruby, V. J., Kazmierski, W. M. and Knapp, R. J.(1991) A new type of synthetic peptide library for identifying ligand-binding activity. Nature 354,82-4.

41

84. Lam, K. S., Sroka, T., Chen, M. L., Zhao, Y., Lou, Q., Wu, J. and Zhao, Z. G. (1998)Application of "one-bead one-compound" combinatorial library methods in signal transductionresearch. Life Sci. 62, 1577-83.

85. Lam, K. S., Wu, J. and Lou, Q. (1995) Identification and characterization of a novel syntheticpeptide substrate specific for Src-family protein tyrosine kinases. Int J Pept Protein Res 45, 587-92.

86. Lamorte, L. and Park, M. (2001) The receptor tyrosine kinases: role in cancer progression. Surg.Oncol. Clin. N. Am. 10, 271-88.

87. Lawrence, D. S. (2001) Functional proteomics: large-scale analysis of protein kinase activity.Genome Biol. 2, Reviews 1007.

88. Lawrie, A. M., Noble, M. E., Tunnah, P., Brown, N. R., Johnson, L. N. and Endicott, J. A.(1997) Protein kinase inhibition by staurosporine revealed in details of the molecular interaction withCDK2. Nat. Struct. Biol. 4, 796-801.

89. Lee, J. Y., Yoo, B. C. and Harmon, A. C. (1998) Kinetic and calcium-binding properties of threecalcium-dependent protein kinase isoenzymes from soybean. Biochemistry 37, 6801-9.

90. Lienhard, G. E. and Secemski, II. (1973) P 1 ,P 5 -Di(adenosine-5')pentaphosphate, a potentmultisubstrate inhibitor of adenylate kinase. J. Biol. Chem. 248, 1121-3.

91. Lindblom, S., Ek, P., Muszynska, G., Ek, B., Szczegielniak, J. and Engstrom, L. (1997)Phosphorylation of sucrose synthase from maize seedlings. Acta Biochim. Pol. 44, 809-17.

92. Loog, M., Sak, K., Ek, P. and Järv, J. (1999) Differential specificity of protein kinases A and C inreaction with synthetic peptides. Bioorg. Chem. 27, 189-196.

93. Lukas, T. J., Mirzoeva, S., Slomczynska, U. and Watterson, D. M. (1999) Identification of novelclasses of protein kinase inhibitors using combinatorial peptide chemistry based on functionalgenomics knowledge. J. Med. Chem. 42, 910-9.

94. Luo, K., Zhou, P. and Lodish, H. F. (1995) The specificity of the transforming growth factor betareceptor kinases determined by a spatially addressable peptide library. Proc. Natl. Acad. Sci. U S A92, 11761-5.

95. Madhusudan, Trafny, E. A., Xuong, N. H., Adams, J. A., Ten Eyck, L. F., Taylor, S. S. andSowadski, J. M. (1994) cAmp-dependent protein kinase: crystallographic insights into substraterecognition and phosphotransfer. Protein Sci. 3, 176-87.

96. Martiny-Baron, G., Kazanietz, M. G., Mischak, H., Blumberg, P. M., Kochs, G., Hug, H.,Marme, D. and Schachtele, C. (1993) Selective inhibition of protein kinase C isozymes by theindolocarbazole Go 6976. J. Biol. Chem. 268, 9194-7.

97. Medzihradszky, D., Chen, S. L., Kenyon, G. L. and Gibson, B. W. (1994) Solid-phase synthesisof adenosine phosphopeptides as potential bisubstrate inhibitors of protein kinases. J. Amer. Chem.Soc. 116, 9413.

98. Meggio, F., Chessa, G., Borin, G., Pinna, L. A. and Marchiori, F. (1981) Synthetic fragments ofprotamines as model substrates for rat liver cyclic AMP-dependent protein kinase. Biochim.Biophys. Acta 662, 94-101.

99. Meggio, F., Donella Deana, A., Ruzzene, M., Brunati, A. M., Cesaro, L., Guerra, B., Meyer, T.,Mett, H., Fabbro, D., Furet, P. and et al. (1995) Different susceptibility of protein kinases tostaurosporine inhibition. Kinetic studies and molecular bases for the resistance of protein kinaseCK2. Eur. J. Biochem. 234, 317-22.

100. Mellor, H. and Parker, P. J. (1998) The extended protein kinase C superfamily. Biochem. J. 332,281-92.

42

101. Mendelow, M., Prorok, M., Salerno, A. and Lawrence, D. S. (1993) ATPase-promoting dead endinhibitors of the cAMP-dependent protein kinase. J. Biol. Chem. 268, 12289-96.

102. Mitchell, R. D., Glass, D. B., Wong, C.-W., Angelos, K. L. and Walsh, D. A. (1995) Heat-stableinhibitor protein derived peptide substrate analogs: phosphorylation by cAMP-dependent and cGMP-dependent protein kinase. Biochemistry 34, 528-534.

103. Mochly-Rosen, D. (1995) Localization of protein kinases by anchoring proteins: a theme in signaltransduction. Science 268, 247-51.

104. Mochly-Rosen, D., Smith, B. L., Chen, C. H., Disatnik, M. H. and Ron, D. (1995) Interaction ofprotein kinase C with RACK1, a receptor for activated C-kinase: a role in beta protein kinase Cmediated signal transduction. Biochem. Soc. Trans. 23, 596-600.

105. Mohammadi, M., Froum, S., Hamby, J. M., Schroeder, M. C., Panek, R. L., Lu, G. H.,Eliseenkova, A. V., Green, D., Schlessinger, J. and Hubbard, S. R. (1998) Crystal structure of anangiogenesis inhibitor bound to the FGF receptor tyrosine kinase domain. EMBO J. 17, 5896-904.

106. Narayana, N., Akamine, P., Xuong, N. H. and Taylor, S. S. (1998) Crystallization and preliminaryX-ray analysis of the unliganded recombinant catalytic subunit of cAMP-dependent protein kinase.Acta Crystallogr. D Biol. Crystallogr. 54, 1401-4.

107. Narayana, N., Cox, S., Shaltiel, S., Taylor, S. S. and Xuong, N. (1997) Crystal structure of apolyhistidine-tagged recombinant catalytic subunit of cAMP-dependent protein kinase complexedwith the peptide inhibitor PKI(5-24) and adenosine. Biochemistry 36, 4438-48.

108. Narayana, N., Diller, T. C., Koide, K., Bunnage, M. E., Nicolaou, K. C., Brunton, L. L., Xuong,N. H., Ten Eyck, L. F. and Taylor, S. S. (1999) Crystal structure of the potent natural productinhibitor balanol in complex with the catalytic subunit of cAMP-dependent protein kinase.Biochemistry 38, 2367-76.

109. Nishikawa, K., Sawasdikosol, S., Fruman, D. A., Lai, J., Songyang, Z., Burakoff, S. J., Yaffe, M.B. and Cantley, L. C. (2000) A peptide library approach identifies a specific inhibitor for the ZAP-70 protein tyrosine kinase. Mol. Cell 6, 969-74.

110. Nishikawa, K., Toker, A., Johannes, F. J., Songyang, Z. and Cantley, L. C. (1997) Determinationof the specific substrate sequence motifs of protein kinase C isozymes. J. Biol. Chem. 272, 952-60.

111. Nishizuka, Y. (1984) The role of protein kinase C in cell surface signal transduction and tumourpromotion. Nature 308, 693-8.

112. Nolan, C., Novoa, W. B., Krebs, E. G. and Fischer, E. H. (1964) Further studies on the sitephosphorylated in the phosphorylase b ta a reaction. Biochemistry 3, 542-551.

113. O' Farrell, F., Loog, M., Janson, I. M. and Ek, P. (1999) Kinetic study of the inhibition of CK2 byheparin fragments of different length. Biochim. Biophys. Acta 1433, 68-75.

114. Obata, T., Yaffe, M. B., Leparc, G. G., Piro, E. T., Maegawa, H., Kashiwagi, A., Kikkawa, R.and Cantley, L. C. (2000) Peptide and protein library screening defines optimal substrate motifs forAKT/PKB. J. Biol. Chem. 275, 36108-15.

115. Olsen, S. R. and Uhler, M. D. (1989) Affinity purification of the C alpha and C beta isoforms of thecatalytic subunit of cAMP-dependent protein kinase. J. Biol. Chem. 264, 18662-6.

116. O'Neill, T., Dwyer, A. J., Ziv, Y., Chan, D. W., Lees-Miller, S. P., Abraham, R. H., Lai, J. H.,Hill, D., Shiloh, Y., Cantley, L. C. and Rathbun, G. A. (2000) Utilization of oriented peptidelibraries to identify substrate motifs selected by ATM. J. Biol. Chem. 275, 22719-27.

43

117. Ono, Y., Fujii, T., Igarashi, K., Kuno, T., Tanaka, C., Kikkawa, U. and Nishizuka, Y. (1989)Phorbol ester binding to protein kinase C requires a cysteine-rich zinc-finger-like sequence. Proc.Natl. Acad. Sci. U S A 86, 4868-71.

118. Ono, Y., Fujii, T., Ogita, K., Kikkawa, U., Igarashi, K. and Nishizuka, Y. (1988) The structure,expression, and properties of additional members of the protein kinase C family. J. Biol. Chem. 263,6927-32.

119. Ono, Y., Fujii, T., Ogita, K., Kikkawa, U., Igarashi, K. and Nishizuka, Y. (1989) Protein kinaseC zeta subspecies from rat brain: its structure, expression, and properties. Proc. Natl. Acad. Sci. USA86, 3099-103.

120. Palczewski, K., Arendt, A., McDowell, J. H. and Hargrave, P. A. (1989) Substrate recognitiondeterminants for rhodopsin kinase: studies with synthetic peptides, polyanions, and polycations.Biochemistry 28, 8764-70.

121. Parang, K., Till, J. H., Ablooglu, A. J., Kohanski, R. A., Hubbard, S. R. and Cole, P. A. (2001)Mechanism-based design of a protein kinase inhibitor. Nat. Struct. Biol. 8, 37-41.

122. Parker, P. J. (1994) in Protein Kinases (Woodgett, J. R., ed.), pp. 68-84, Oxford University Press.123. Pearson, R. B. and Kemp, B. E. (1991) Protein kinase phosphorylation site sequences and

consensus specificity motifs: tabulations. Methods Enzymol. 200, 62-81.124. Pinna, L. A. and Ruzzene, M. (1996) How do protein kinases recognize their substrates? Biochim.

Biophys. Acta. 1314, 191-225.125. Polya, G. M., Morrice, N. and Wettenhall, R. E. H. (1989) Substrate specificity of wheat embryo

calcium-dependent protein kinase. FEBS Lett. 253, 137-140.126. Prade, L., Engh, R. A., Girod, A., Kinzel, V., Huber, R. and Bossemeyer, D. (1997)

Staurosporine-induced conformational changes of cAMP-dependent protein kinase catalytic subunitexplain inhibitory potential. Structure 5, 1627-37.

127. Rewcastle, G. W., Murray, D. K., Elliott, W. L., Fry, D. W., Howard, C. T., Nelson, J. M.,Roberts, B. J., Vincent, P. W., Showalter, H. D., Winters, R. T. and Denny, W. A. (1998)Tyrosine kinase inhibitors. 14. Structure-activity relationships for methylamino-substitutedderivatives of 4-[(3-bromophenyl)amino]-6-(methylamino)-pyrido[3,4-d]pyrimidine (PD 158780), apotent and specific inhibitor of the tyrosine kinase activity of receptors for the EGF family of growthfactors. J. Med. Chem. 41, 742-51.

128. Ricouart, A., Gesquiere, J. C., Tartar, A. and Sergheraert, C. (1991) Design of potent proteinkinase inhibitors using the bisubstrate approach. J. Med. Chem. 34, 73-8.

129. Roberts, D. M. (1989) Detection of a calcium-activated protein kinase in Mougeotia by usingsynthetic peptide substrates. Plant Physiol. 91, 1613-1619.

130. Roberts, D. M. and Harmon, A. C. (1992) Calcium-modulated proteins: targets of intracellularcalcium signals in higher plants. Annu. Rev. Plant Physiol. Plant Mol. Biol. 43, 375-414.

131. Ron, D., Luo, J. and Mochly-Rosen, D. (1995) C2 region-derived peptides inhibit translocation andfunction of beta protein kinase C in vivo. J. Biol. Chem. 270, 24180-7.

132. Saijo, Y., Hata, S., Sheen, J. and Izui, K. (1997) cDNA cloning and prokaryotic expression ofmaize calcium-dependent protein kinases. Biochim. Biophys. Acta 1350, 109-14.

133. Sasaki, S., Hashimoto, T., Obana, N., Yasuda, H., Uehara, Y. and Maeda, M. (1998) Design ofnew inhibitors for cdc2 kinase based on a multiple pseudosubstrate structure. Bioorg. Med. Chem.Lett. 8, 1019-22.

44

134. Schulze-Gahmen, U., Brandsen, J., Jones, H. D., Morgan, D. O., Meijer, L., Vesely, J. and Kim,S. H. (1995) Multiple modes of ligand recognition: crystal structures of cyclin-dependent proteinkinase 2 in complex with ATP and two inhibitors, olomoucine and isopentenyladenine. Proteins 22,378-91.

135. Sclafani, R. A. (2000) Cdc7p-Dbf4p becomes famous in the cell cycle. J. Cell Sci. 113, 2111-7.136. Scott, J. D., Glaccum, M. B., Fischer, E. H. and Krebs, E. G. (1986). Primary-structure

requirements for inhibition by the heat-stable inhibitor of the cAMP-dependent protein kinase. Proc.Natl. Acad. Sci. U S A 83, 1613-6.

137. Sharkey, N. A. and Blumberg, P. M. (1985) Kinetic evidence that 1,2-diolein inhibits phorbol esterbinding to protein kinase C via a competitive mechanism. Biochem. Biophys. Res. Commun. 133,1051-6.

138. Sharrocks, A. D., Yang, S. H. and Galanis, A. (2000) Docking domains and substrate-specificitydetermination for MAP kinases. Trends. Biochem. Sci. 25, 448-53.

139. Sheen, J. (1996) Ca2+-dependent protein kinases and stress signal transduction in plants. Science274, 1900-2.

140. Showers, M. O. and Maurer, R. A. (1986) A cloned bovine cDNA encodes an alternate form of thecatalytic subunit of cAMP-dependent protein kinase. J. Biol. Chem. 261, 16288-91.

141. Sondhi, D., Xu, W., Songyang, Z., Eck, M. J. and Cole, P. A. (1998) Peptide and proteinphosphorylation by protein tyrosine kinase Csk: insights into specificity and mechanism.Biochemistry 37, 165-72.

142. Songyang, Z. (2001) Analysis of protein kinase specificity by peptide libraries and prediction of invivo substrates. Methods Enzymol. 332, 171-83.

143. Songyang, Z., Blechner, S., Hoagland, N., Hoekstra, M. F., Piwnica-Worms, H. and Cantley, L.C. (1994) Use of an oriented peptide library to determine the optimal substrates of protein kinases.Curr. Biol. 4, 973-82.

144. Songyang, Z. and Cantley, L. C. (1998) The use of peptide library for the determination of kinasepeptide substrates. Methods Mol. Biol. 87, 87-98.

145. Songyang, Z., Lu, K. P., Kwon, Y. T., Tsai, L. H., Filhol, O., Cochet, C., Brickey, D. A.,Soderling, T. R., Bartleson, C., Graves, D. J., DeMaggio, A. J., Hoekstra, M. F., Blenis, J.,Hunter, T. and Cantley, L. C. (1996) A structural basis for substrate specificities of protein Ser/Thrkinases: primary sequence preference of casein kinases I and II, NIMA, phosphorylase kinase,calmodulin-dependent kinase II, CDK5, and Erk1. Mol. Cell Biol. 16, 6486-93.

146. Songyang, Z., Shoelson, S. E., Chaudhuri, M., Gish, G., Pawson, T., Haser, W. G., King, F.,Roberts, T., Ratnofsky, S., Lechleider, R. J. and et al. (1993) SH2 domains recognize specificphosphopeptide sequences. Cell 72, 767-78.

147. Sridhar, R., Hanson-Painton, O. and Cooper, D. R. (2000) Protein kinases as therapeutic targets.Pharm. Res. 17, 1345-53.

148. Stebbins, E. G. and Mochly-Rosen, D. (2001) Binding Specificity for RACK1 Resides in the V5Region of beta II Protein Kinase C. J. Biol. Chem. 276, 29644-50.

149. Szczegielniak, J., Liwosz, A., Jurkowski, I., Loog, M., Dobrowolska, G., Ek, P., Harmon, A. C.and Muszynska, G. (2000) Calcium-dependent protein kinase from maize seedlings activated byphospholipids. Eur. J. Biochem. 267, 3818-27.

45

150. Tamaoki, T., Nomoto, H., Takahashi, I., Kato, Y., Morimoto, M. and Tomita, F. (1986)Staurosporine, a potent inhibitor of phospholipid/Ca++dependent protein kinase. Biochem. Biophys.Res. Commun. 135, 397-402.

151. Tao, M., Salas, M. L. and Lipmann, F. (1970) Mechanism of activation by adenosine 3':5'-cyclicmonophosphate of a protein phosphokinase from rabbit reticulocytes. Proc. Natl. Acad. Sci. U S A67, 408-14.

152. Taylor, S. S., Knighton, D. R., Zheng, J., Sowadski, J. M., Gibbs, C. S. and Zoller, M. J. (1993)A template for the protein kinase family. Trends Biochem. Sci. 18, 84-9.

153. Taylor, S. S., Knighton, D. R., Zheng, J., Ten Eyck, L. F. and Sowadski, J. M. (1992) Structuralframework for the protein kinase family. Annu. Rev. Cell. Biol. 8, 429-62.

154. Tegge, W., Frank, R., Hofmann, F. and Dostmann, W. R. (1995) Determination of cyclicnucleotide-dependent protein kinase substrate specificity by the use of peptide libraries on cellulosepaper. Biochemistry 34, 10569-77.

155. ter Haar, E., Coll, J. T., Austen, D. A., Hsiao, H. M., Swenson, L. and Jain, J. (2001) Structure ofGSK3beta reveals a primed phosphorylation mechanism. Nat. Struct. Biol. 8, 593-6.

156. Till, J. H., Annan, R. S., Carr, S. A. and Miller, W. T. (1994) Use of synthetic peptide librariesand phosphopeptide-selective mass spectrometry to probe protein kinase substrate specificity. J.Biol. Chem. 269, 7423-8.

157. Toomik, R., Edlund, M., Ek, P., Obrink, B. and Engström, L. (1996) Simultaneously synthesizedpeptides on continuous cellulose membranes as substrates for protein kinases. Pept. Res. 9, 6-11.

158. Toomik, R. and Ek, P. (1997) A potent and highly selective peptide substrate for protein kinase Cassay. Biochem. J. 322, 455-60.

159. Toullec, D., Pianetti, P., Coste, H., Bellevergue, P., Grand-Perret, T., Ajakane, M., Baudet, V.,Boissin, P., Boursier, E., Loriolle, F. and et al. (1991) The bisindolylmaleimide GF 109203X is apotent and selective inhibitor of protein kinase C. J. Biol. Chem. 266, 15771-81.

160. Traxler, P. M., Furet, P., Mett, H., Buchdunger, E., Meyer, T. and Lydon, N. (1996) 4-(Phenylamino)pyrrolopyrimidines: potent and selective, ATP site directed inhibitors of the EGF-receptor protein tyrosine kinase. J. Med. Chem. 39, 2285-92.

161. Tsigelny, I., Grant, B. D., Taylor, S. S. and Ten Eyck, L. F. (1996) Catalytic subunit of cAMP-dependent protein kinase: electrostatic features and peptide recognition. Biopolymers 39, 353-65.

162. Uhler, M. D., Carmichael, D. F., Lee, D. C., Chrivia, J. C., Krebs, E. G. and McKnight, G. S.(1986) Isolation of cDNA clones coding for the catalytic subunit of mouse cAMP-dependent proteinkinase. Proc. Natl Acad. Sci. U S A 83, 1300-4.

163. Uhler, M. D., Chrivia, J. C. and McKnight, G. S. (1986) Evidence for a second isoform of thecatalytic subunit of cAMP-dependent protein kinase. J. Biol. Chem. 261, 15360-3.

164. Vaglio, P., Sarno, S., Marin, O., Meggio, F., Issinger, O. G. and Pinna, L. A. (1996) Mapping theresidues of protein kinase CK2 alpha subunit responsible for responsiveness to polyanionicinhibitors. FEBS Lett. 380, 25-8.

165. Walsh, D. A. and Glass, D. B. (1991) Utilization of the inhibitor protein of adenosine cyclicmonophosphate-dependent protein kinase, and peptides derived from it, as tools to study adenosinecyclic monophosphate-mediated cellular processes. Methods Enzymol. 201, 304-16.

166. Ward, N. E., Gravitt, K. R. and O'Brian, C. A. (1995) Irreversible inactivation of protein kinase Cby a peptide-substrate analog. J. Biol. Chem. 270, 8056-60.

46

167. Ward, N. E., Pierce, D. S., Stewart, J. R. and CA, O. b. (1999) A peptide substrate-based affinitylabel blocks protein kinase C-catalyzed ATP hydrolysis and peptide-substrate phosphorylation.Arch. Biochem. Biophys. 365, 248-53.

168. Venter, J. C. et al. (2001) The sequence of the human genome. Science 291, 1304.169. Werner, D. S., Lee, T. R. and Lawrence, D. S. (1996) Is protein kinase substrate efficacy a reliable

barometer for successful inhibitor design? J. Biol. Chem. 271, 180-5.170. Whitehouse, S., Feramisco, J. R., Casnellie, J. E., Krebs, E. G. and Walsh, D. A. (1983) Studies

on the kinetic mechanism of the catalytic subunit of the cAMP-dependent protein kinase. J. Biol.Chem. 258, 3693-701.

171. Whitehouse, S. and Walsh, D. A. (1983) Mg X ATP2-dependent interaction of the inhibitor proteinof the cAMP-dependent protein kinase with the catalytic subunit. J. Biol. Chem. 258, 3682-92.

172. Wilkinson, S. E., Parker, P. J. and Nixon, J. S. (1993) Isoenzyme specificity ofbisindolylmaleimides, selective inhibitors of protein kinase C. Biochem. J. 294, 335-7.

173. Williams, D. M. and Cole, P. A. (2001) Kinase chips hit the proteomics era. Trends Biochem. Sci.26, 271-3.

174. Wilson, K. P., McCaffrey, P. G., Hsiao, K., Pazhanisamy, S., Galullo, V., Bemis, G. W.,Fitzgibbon, M. J., Caron, P. R., Murcko, M. A. and Su, M. S. (1997) The structural basis for thespecificity of pyridinylimidazole inhibitors of p38 MAP kinase. Chem. Biol. 4, 423-31.

175. Woodgett, J. R. (1989) Use of peptide substrates for affinity purification of protein-serine kinases.Anal. Biochem. 180, 237-41.

176. Woodgett, J. R., Gould, K. L. and Hunter, T. (1986) Substrate specificity of protein kinase C. Useof synthetic peptides corresponding to physiological sites as probes for substrate recognitionrequirements. Eur. J. Biochem. 161, 177-84.

177. Wu, J., Ma, Q. N. and Lam, K. S. (1994) Identifying substrate motifs of protein kinases by arandom library approach. Biochemistry 33, 14825-33.

178. Wu, J. J., Phan, H. and Lam, K. S. (1998) Comparison of the intrinsic kinase activity and substratespecificity of c-Abl and Bcr-Abl. Bioorg. Med. Chem. Lett. 8, 2279-84.

179. Yan, X., Corbin, J. D., Francis, S. H. and Lawrence, D. S. (1996) Precision targeting of proteinkinases. An affinity label that inactivates the cGMP- but not the cAMP-dependent protein kinase. J.Biol. Chem. 271, 1845-8.

180. Yonemoto, W., Garrod, S. M., Bell, S. M. and Taylor, S. S. (1993) Identification ofphosphorylation sites in the recombinant catalytic subunit of cAMP-dependent protein kinase. J.Biol. Chem. 268, 18626-32.

181. Zetterqvist, Ö., Ragnarsson, U., Humble, E., Berglund, L. and Engström, L. (1976) Theminimum substrate of cyclic AMP-stimulated protein kinase, as studied by synthetic peptidesrepresenting the phosphorylatable site of pyruvate kinase (type L) of rat liver. Biochem. Biophys.Res. Commun. 70, 696-703.

182. Zetterqvist, Ö. and Ragnarsson, U. (1982) The structural requirements of substrates of cyclicAMP-dependent protein kinase. FEBS Lett. 139, 287-90.

183. Zetterqvist, Ö., Ragnarsson, U. and Engström, L. (1990) in Peptides and protein phosphorylation(Kemp, B. E., ed.), pp. 171-184, CRC Press, Boca Raton.

184. Zheng, J., Knighton, D. R., ten Eyck, L. F., Karlsson, R., Xuong, N., Taylor, S. S. andSowadski, J. M. (1993) Crystal structure of the catalytic subunit of cAMP-dependent protein kinasecomplexed with MgATP and peptide inhibitor. Biochemistry 32, 2154-61.

47

185. Zheng, J., Knighton, D. R., Xuong, N. H., Taylor, S. S., Sowadski, J. M. and Ten Eyck, L. F.(1993) Crystal structures of the myristylated catalytic subunit of cAMP-dependent protein kinasereveal open and closed conformations. Protein Sci. 2, 1559-73.

186. Zhu, H., Klemic, J. F., Chang, S., Bertone, P., Casamayor, A., Klemic, K. G., Smith, D.,Gerstein, M., Reed, M. A. and Snyder, M. (2000) Analysis of yeast protein kinases using proteinchips. Nat. Genet. 26, 283-9.