Specificity determinants of substrate recognition by the protein kinase DYRK1A

DYRK1A is a dual-specificity protein kinase that is thought to be involved in brain development. We identified a single phosphorylated amino acid residue in the DYRK substrate histone H3 (threonine 45) by mass spectrometry, phosphoamino acid analysis, and protein sequencing. Exchange of threonine 45 for alanine abolished phosphorylation of histone H3 by DYRK1A and by the related kinases DYRK1B, DYRK2, and DYRK3 but not by CLK3. In order to define the consensus sequence for the substrate specificity of DYRK1A, a library of 300 peptides was designed in variation of the H3 phosphorylation site. Evaluation of the phosphate incorporation into these peptides identified DYRK1A as a proline-directed kinase with a phosphorylation consensus sequence (RPX(S/T)P) similar to that of ERK2 (PX(S/T)P). A peptide designed after the optimal substrate sequence (DYRKtide) was efficiently phosphorylated by DYRK1A (K(m) = 35 microM) but not by ERK2. Both ERK2 and DYRK1A phosphorylated myelin basic protein, whereas only ERK2, but not DYRK1A, phosphorylated the mitogen-activated protein kinase substrate ELK-1. This marked difference in substrate specificity between DYRK1A and ERK2 can be explained by the requirement for an arginine at the P -3 site of DYRK substrates and its presumed interaction with aspartate 247 conserved in all DYRKs.     

Assessment of protein-tyrosine phosphatase 1B substrate specificity using "inverse alanine scanning"

An "inverse alanine scanning" peptide library approach has been developed to assess the substrate specificity of protein-tyrosine phosphatases (PTPases). In this method each Ala moiety in the parent peptide, Ac-AAAApYAAAA-NH(2), is separately and sequentially replaced by the 19 non-Ala amino acids to generate a library of 153 well defined peptides. The relatively small number of peptides allows the acquisition of explicit kinetic data for all library members, thereby furnishing information about the contribution of individual amino acids with respect to substrate properties. The approach was applied to protein-tyrosine phosphatase 1B (PTP1B) as a first example, and the highly potent peptide substrate Ac-ELEFpYMDYE-NH(2) (k(cat)/K(m) 2.2 +/- 0.05 x 10(7) M(-1) s(-1)) has been identified. More importantly, several heretofore unknown features of the substrate specificity of PTP1B were revealed. This includes the ability of PTP1B to accommodate acidic, aromatic, and hydrophobic residues at the -1 position, a strong nonpreference for Lys and Arg residues in any position, and the first evidence that residues well beyond the +1 position contribute to substrate efficacy.     

Utilization of oriented peptide libraries to identify substrate motifs selected by ATM

The ataxia telangiectasia mutated (ATM) gene encodes a serine/threonine protein kinase that plays a critical role in genomic surveillance and development. Here, we use a peptide library approach to define the in vitro substrate specificity of ATM kinase activity. The peptide library analysis identified an optimal sequence with a central core motif of LSQE that is preferentially phosphorylated by ATM. The contributions of the amino acids surrounding serine in the LSQE motif were assessed by utilizing specific peptide libraries or individual peptide substrates. All amino acids comprising the LSQE sequence were critical for maximum peptide substrate suitability for ATM. The DNA-dependent protein kinase (DNA-PK), a Ser/Thr kinase related to ATM and important in DNA repair, was compared with ATM in terms of peptide substrate selectivity. DNA-PK was found to be unique in its preference of neighboring amino acids to the phosphorylated serine. Peptide library analyses defined a preferred amino acid motif for ATM that permits clear distinctions between ATM and DNA-PK kinase activity. Data base searches using the library-derived ATM sequence identified previously characterized substrates of ATM, as well as novel candidate substrate targets that may function downstream in ATM-directed signaling pathways.     

Effect of tumor necrosis factor-alpha on insulin signal transduction in rat adipocytes: relation to PKCbeta and zeta translocation

Although much evidence has been accumulated suggesting that tumor necrosis factor-alpha (TNF-alpha) is an important mediator of insulin resistance, the precise mechanism involved is still unclear. Recently, it has been reported that insulin-induced glucose uptake is mediated by activation of second messengers such as insulin receptor substrate 1 (IRS-1), phosphatidylinositol 3-kinase (PI3K), and diacylglycerol (DG)-protein kinase C (PKC). We have examined the effect of TNF-alpha on insulin-induced glucose uptake and activations of tyrosine kinase, IRS-1, PI3K and PKC in rat adipocytes. Pretreatment with 0.1-100 nM TNF-alpha for 60 min resulted in a significant decrease in 10 nM insulin- or 1 microM 12-O-tetradecanoyl phorbol-13-acetate (TPA)-induced [3H]2-deoxyglucose uptake without affecting basal glucose uptake. 10 nM insulin-stimulated activation of tyrosine kinase, IRS-1 and PI3K was suppressed by preincubation with 0.1-10 nM TNF-alpha for 60 min. 10 nM TNF-alpha pretreatment also suppressed 10 nM insulin- and 1 microM TPA-induced increases in membrane-associated PKCbeta and PKCzeta. Furthermore, 10 nM TNF-alpha, by itself, altered PKCbeta translocation from the membrane to cytosol. These results suggest that TNF-alpha inhibits insulin-stimulated activation of both the tyrosine kinase-IRS-1-PI3K-PKCzeta pathway and DG-PKC pathway. Finally, TNF-alpha contributes to insulin resistance in rat adipocytes.     

Aminopeptidase from Sphingomonas capsulata

A novel aminopeptidase with unique substrate specificity was purified from a culture broth of Sphingomonas capsulata. This is the first reported aminopeptidase to demonstrate broad substrate specificity and yet release glycine and alanine with the highest efficacy. On a series of pentapeptide amides with different N-terminal amino acids, this enzyme efficiently releases glycine, alanine, leucine, proline, and glutamate with the lowest turnover value of 370 min(-1) for glutamate. At pH 7.5 (pH optimum) and 25 degrees C, the kinetic parameters for alanine para-nitroanilide were found to be k(cat) = 7600 min(-1) and K(m) = 14 mm. For alanine beta-naphthylamide, they were k(cat) = 860 min(-1) and K(m) = 6.7 mm. Polymerase chain reaction primers were designed based upon obtained internal sequences of the wild type enzyme. The subsequent product was then used to acquire the full-length gene from an S. capsulata genomic library. An open reading frame encoding a protein of 670 amino acids was obtained. The translated protein has a putative signal peptide that directs the enzyme into the supernatant. A search of the amino acid sequence revealed no significant homology to any known aminopeptidases in the available data bases.     

Kinetic study of the inhibition of CK2 by heparin fragments of different length.

The structure-activity relationships for the inhibition of protein kinase CK2 by heparin were investigated using purified heparin fragments of different length, varying from 4 to 24 oligosaccharide sugar units. The inhibitory potency was shown to decrease concomitant with the shortening of the heparin fragment length. The fragment of 24 oligosaccharide sugar units was the most potent inhibitor with a K(i) value of 22 nM which is close to the K(i) value for the commercial heparin mixture available. Shortening of the heparin from 24 to 12 sugar units had a moderate influence on the inhibitory potency causing an increase in K(i) values up to 151 nM while fragments shorter than 12 sugar units showed a more drastic increase in K(i) values reaching up to micromolar range. The mode of inhibition was studied in respect to the protein substrate beta-casein and it was shown to be competitive for the long as well as for the short heparin fragments. In contrast, the inhibition mode in respect to a synthetic peptide substrate RRRADDSDDDDD was found to be hyperbolic partial non-competitive mixed-type. Such a kinetic model suggests that heparin binds to a site on CK2 which does not overlap with the peptide substrate binding site and that a productive enzyme complex exists where both heparin and peptide substrate are simultaneously bound. This is in contrast to the competitive inhibition model of the phosphorylation of protein substrate beta-casein where the binding of the protein substrate and inhibitor was mutually exclusive.     

Target specificity analysis of the Abl kinase using peptide microarray data

Protein kinases play an important role in cellular signalling. The reliable prediction of their substrates is of high importance for the deciphering of signalling pathways. A recently developed peptide microarray technology for the charcterisation of protein kinases delivers data on the individual phosphorylation status of each single member of a large peptide library. This data can be used to approximate the substrate specificity of the investigated kinase. We present an approach to process the collected information using a combination of a weight matrix approach and a nearest neighbor approach. Experiments with the protein-tyrosine kinase Abl are conducted to validate the results. Randomly selected peptides (1433) are used to estimate the substrate preferences of the kinase. The obtained prediction results are compared with standard methods. The new approach is tested further on bona fide Abl phosphorylation sites.     

Rapid determination of substrate specificity of Clostridium histolyticum beta-collagenase using an immobilized peptide library.

The molecular basis of the substrate specificity of Clostridium histolyticum beta-collagenase was investigated using a combinatorial method. An immobilized positional peptide library, which contains 24,000 sequences, was constructed with a 7-hydroxycoumarin-4-propanoyl (Cop) fluorescent group attached at the N terminus of each sequence. This immobilized peptide library was incubated with C. histolyticum beta-collagenase, releasing fluorogenic fragments in the solution phase. The relative substrate specificity (k(cat)/K(m)) for each member of the library was determined by measuring fluorescence intensity in the solution phase. Edman sequencing was used to assign structure to subsites of active substrate mixtures. Collectively, the substrate preference for subsites (P(3)-P(4)') of C. histolyticum beta-collagenase was determined. The last position on the C-terminal side in which the identity of the amino acids affects the activity of the enzyme is P(4)', and an aromatic side chain is preferred in this position. The optimal P(1)'-P(3)' extended substrate sequence is P(1)'-Gly/Ala, P(2)'-Pro/Xaa, and P(3)'-Lys/Arg/Pro/Thr/Ser. The Cop group in either the P(2) or P(3) position is required for a high substrate activity with C. histolyticum beta-collagenase. S(2) and S(3) sites of the protease play a dominant role in fixing the substrate specificity. The immobilized peptide library proved to be a powerful approach for assessing the substrate specificity of C. histolyticum beta-collagenase, so it may be applied to the study of other proteases of interest.     

Characterization of yeast pyruvate kinase 1 as a protein kinase A substrate, and specificity of the phosphorylation site sequence in the whole protein

Pyk1 (pyruvate kinase 1) from Saccharomyces cerevisiae was characterized as a substrate for PKA (protein kinase A) from bovine heart and yeast. By designing Pyk1 synthetic peptides containing potential PKA sequence targets (Ser22, Thr94 and Thr478) we determined that the peptide S22 was a substrate for PKA in vitro, with a K(sp)* (specificity constant) 10-fold and 3-fold higher than Kemptide for bovine heart and yeast PKA respectively. In vitro phosphorylation of the Pyk1 S22A mutant protein was decreased by as much as 90% when compared with wild-type Pyk1 and the Pyk1 T94A mutant. The K(sp)* values for Pyk1 and Pyk1 T94A were the same, indicating that both proteins are phosphorylated at the same site by PKA. Two-dimensional PAGE of Pyk1 and Pyk1 S22A indicates that in vivo the S22A mutation prevented the formation of one of the Pyk1 isoforms. We conclude that in yeast the major PKA phosphorylation site of Pyk1 is Ser22. Phosphorylation of Ser22 leads to a Pyk1 enzyme that is more active in the absence of FBP (fructose 1,6-bisphosphate). The specificity of yeast and mammalian PKA towards the S22 peptide and towards whole Pyk1 protein was measured and compared. The K(sp)* for the S22 peptide is higher than that for Pyk1, indicating that the peptide modelled on Pyk1 is a much better substrate than Pyk1, regardless of which tissue was used as the source of PKA. However, the K(m) of Pyk1 protein is lower than that of the better substrate, the S22 peptide, indicating that ground-state substrate binding is not the major determinant of substrate specificity for PKA.     

Phosphorylation by cyclic GMP-dependent protein kinase of a synthetic peptide corresponding to the autophosphorylation site in the enzyme

The substrate specificity of cGMP-dependent protein kinase has been investigated by examining the ability of the enzyme to phosphorylate a series of synthetic peptides that correspond to the amino acid sequence at its site of autophosphorylation. The undecapeptide Ile53-Gly-Pro-Arg-Thr-Thr58-Arg-Ala-Gln-Gly-Ile63 which corresponds to the sequence around threonine-58 in cGMP-dependent protein kinase (Takio, K., Smith, S.B., Walsh, K.A., Krebs, E.G., and Titani, K. (1983) J. Biol. Chem. 258, 5531-5536) was synthesized and tested as a substrate for that enzyme. It was phosphorylated to the extent of 1.0 mol of phosphate/mol of peptide. Analysis of the products of Edman degradation of the phosphopeptide indicated that only threonine-58 was phosphorylated, as is the case for the autophosphorylation reaction in the native enzyme. The peptide was phosphorylated by cGMP-dependent protein kinase with a Km value of 578 +/- 25 microM and a Vmax of 0.069 +/- 0.003 mumol/min/mg of enzyme. This low Vmax value is consistent with the relatively slow rate of the autophosphorylation reaction. An analog peptide that contained serine in place of threonine-58 was also phosphorylated to 1.0 mol of phosphate/mol of peptide. That phosphopeptide contained only phosphoserine. The serine-containing analog peptide had a Km value similar to that of the parent peptide but was phosphorylated with a 70-fold higher Vmax value. Substitution of arginine-56 in the parent peptide by an alanine residue resulted in a peptide that was essentially not a substrate. Substitution of arginine-59, COOH-terminal to the phosphorylatable threonine, yielded a peptide with a Vmax similar to that of the parent peptide but a Km value of almost 22,000 microM. These results indicate that serine is a better phosphate-accepting residue than is threonine and that both arginine residues around the site of autophosphorylation are important specificity determinants for the cGMP-dependent protein kinase.     

Substrate profiling of cysteine proteases using a combinatorial peptide library identifies functionally unique specificities

The substrate specificities of papain-like cysteine proteases (clan CA, family C1) papain, bromelain, and human cathepsins L, V, K, S, F, B, and five proteases of parasitic origin were studied using a completely diversified positional scanning synthetic combinatorial library. A bifunctional coumarin fluorophore was used that facilitated synthesis of the library and individual peptide substrates. The library has a total of 160,000 tetrapeptide substrate sequences completely randomizing each of the P1, P2, P3, and P4 positions with 20 amino acids. A microtiter plate assay format permitted a rapid determination of the specificity profile of each enzyme. Individual peptide substrates were then synthesized and tested for a quantitative determination of the specificity of the human cathepsins. Despite the conserved three-dimensional structure and similar substrate specificity of the enzymes studied, distinct amino acid preferences that differentiate each enzyme were identified. The specificities of cathepsins K and S partially match the cleavage site sequences in their physiological substrates. Capitalizing on its unique preference for proline and glycine at the P2 and P3 positions, respectively, selective substrates and a substrate-based inhibitor were developed for cathepsin K. A cluster analysis of the proteases based on the complete specificity profile provided a functional characterization distinct from standard sequence analysis. This approach provides useful information for developing selective chemical probes to study protease-related pathologies and physiologies.     

A calmodulin-regulated protein kinase linked to neuron survival is a substrate for the calmodulin-regulated death-associated protein kinase

Death-associated protein kinase (DAPK) is a calmodulin (CaM)-regulated protein kinase and a drug-discovery target for neurodegenerative diseases. However, a protein substrate relevant to neuronal death had not been described. We identified human brain CaM-regulated protein kinase kinase (CaMKK), an enzyme key to neuronal survival, as the first relevant substrate protein by using a focused proteomics- and informatics-based approach that can be generalized to protein kinase open reading frames identified in genome projects without prior knowledge of biochemical context. First, DAPK-interacting proteins were detected in yeast two-hybrid screens and in immunoprecipitates of brain extracts. Second, potential phosphorylation site sequences in yeast two-hybrid hits were identified on the basis of our previous results from positional-scanning synthetic-peptide substrate libraries and molecular modeling. Third, reconstitution assays using purified components demonstrated that DAPK phosphorylates CaMKK with a stoichiometry of nearly 1 mol of phosphate per mole of CaMKK and a K(m) value of 3 microM. Fourth, S511 was identified as the phosphorylation site by peptide mapping using mass spectrometry, site-directed mutagenesis, and Western blot analysis with a site-directed antisera targeting the phosphorylated sequence. Fifth, a potential mechanism of action was identified on the basis of the location of S511 near the CaM recognition domain of CaMKK and demonstrated by attenuation of CaM-stimulated CaMKK autophosphorylation after DAPK phosphorylation. The results raise the possibility of a CaM-regulated protein kinase cascade as a key mechanism in acute neurodegeneration amenable to therapeutic targeting.     

ADAM33 enzyme properties and substrate specificity

ADAM33 is an asthma susceptibility gene recently identified through a genetic study of asthmatic families [van Eerdewegh, et al. (2002) Nature 418, 426-430]. To understand the function of the gene product, the recombinant metalloproteinase domain of human ADAM33 was purified and tested for its substrate cleavage specificity using peptides derived from beta-amyloid precursor protein (APP). A single Ala substitution at the P2 position of a 10-residue APP peptide, YEVHHQKLVF, yielded a 20-fold more efficient substrate. Terminal truncation studies identified a minimal nine-residue core (P5-P4') important for ADAM33 recognition and cleavage. Full positional scanning of the 10-mer peptide using the 19 naturally occurring l-amino acids (excluding Cys) revealed a substrate specificity profile. A strong preference for Val or Ile at P3, Ala at P2, and Gln at P1' was observed. The substrate binding model based on the X-ray structure of the ADAM33-inhibitor complex supported the observed substrate specificity profile. On the basis of this, an improved substrate was designed and a fluorescence resonance energy transfer (FRET) assay was developed using a fluorogenic derivative of this substrate. Kinetic studies confirmed that the best substrate, FRET-P2 [K(Dabcyl)YRVAFQKLAE(Edans)K], was approximately 100-fold more efficient than the wild-type APP peptide substrate, with a k(cat)/K(m) value of (3.6 +/- 0.1) x 10(4) s(-)(1) M(-)(1). Using this substrate and the FRET assay, ADAM33 enzyme activity and thermal stability were characterized. ADAM33 dependence on buffer conditions, detergents, and temperature was examined, and optimal conditions were defined. Accurate K(i) values for tissue inhibitors of metalloproteinase and small molecule compounds were obtained.     

The assay of the activity of protein kinase C with the synthetic oligopeptide substrate designed for histone kinase II

The activity of histone kinase II was determined on the basis of its ability to phosphorylate the nonapeptide Ala-Ala-Ala-Ser-Phe-Lys-Ala-Lys-Lys-amide designed previously as a specific substrate for this enzyme. Histone kinase II was purified from calf thymus extract by DEAE-cellulose chromatography followed by hydroxylapatite chromatography and high-performance liquid chromatography on a Protein Analysis column (I-125). The Mr value of histone kinase II estimated by the latter method was 50,000-55,000, but several observations indicated that histone kinase II was a product of a proteolytic process. Since the substrate specificity determinants for histone kinase II known from our previous investigations are very similar to those for protein kinase C, it was presumable that histone kinase II was the proteolytic fragment of protein kinase C. Therefore, the nonapeptide was tested as a substrate for protein kinase C prepared from rabbit brain extract by DEAE-cellulose chromatography. The activity of histone kinase II was also detected in brain extract. Histone kinase II was eluted from the DEAE-cellulose in the known position of the proteolytic fragment of protein kinase C. The nonapeptide Ala-Ala-Ala-Ser-Phe-Lys-Ala-Lys-Lys-amide proved to be a better substrate than H1 histone for the detection of the activity of protein kinase C because it was not phosphorylated by the cAMP-dependent protein kinase and the Vmax of protein kinase C was about one order of magnitude higher with the peptide than with H1 histone. The apparent Km of protein kinase C for the peptide was identical with that of histone kinase II (0.2 mM).     

A short peptide is a protein kinase C (PKC) alpha-specific substrate

The purpose of this study was to find protein kinase C (PKC) isozyme-specific peptides. A peptide library containing 1772 sequences was designed using Scansite and screened by MALDI-TOF MS and kinase activity assays for PKC isozyme-specificity. A peptide (Alphatomega; H-FKKQGSFAKKK-NH(2)) with high specificity for PKC alpha relative to other isozymes was identified. The peptide was phosphorylated to a greater extent by tissue lysates from B16 melanoma, HepG2, and human breast cancer, which had higher levels of activated PKC alpha, when compared to normal skin, liver, and human breast tissue lysates, respectively. Moreover, addition of Ro-31-7549, an inhibitor with great specificity for PKC alpha, to the phosphorylation reaction caused a dose-dependent reduction in phosphorylation, but no inhibition was identified with the addition of rottlerin and H-89. These results show that this peptide has great potential as a PKC alpha-specific substrate.     

Selection of substrate recognition sequence of protein kinase with ferric chelation affinity chromatography

Protein kinase substrate phage (PKS phage) was constructed by fusing the substrate recognition consensus sequence of cAMP-dependent protein kinase (cAPK) with bacteriophage minor coat protein g3p and by dis-playing it on the surface of filamentous bacteriophage fd. Phosphorylation in vitro by cAPK showed a unique labelled band of approximately 60 ku, which was consistent with the molecular weight of the PKS-g3p fusion protein. Some weakly phosphorylated bands for both PKS phage and wild-type phage were also observed. Phage display random 15-mer peptide library phosphorylated by cAPK was selected with ferric (Fe3 ) chelalion affinity resin. After 4 rounds of screening, phage clones were picked out to determine the displayed peptide sequences by DNA sequencing. The results showed that 5 of 14 sequenced phages displayed the cAPK recognition sequence motif (R)RXS/T. Their in vitro phosphorylation analyses revealed the specific labelled bands corresponding to the positive PKS phages with and without the typ     

Analysis of Jak2 catalytic function by peptide microarrays: the role of the JH2 domain and V617F mutation

Janus kinase 2 (JAK2) initiates signaling from several cytokine receptors and is required for biological responses such as erythropoiesis. JAK2 activity is controlled by regulatory proteins such as Suppressor of Cytokine Signaling (SOCS) proteins and protein tyrosine phosphatases. JAK2 activity is also intrinsically controlled by regulatory domains, where the pseudokinase (JAK homology 2, JH2) domain has been shown to play an essential role. The physiological role of the JH2 domain in the regulation of JAK2 activity was highlighted by the discovery of the acquired missense point mutation V617F in myeloproliferative neoplasms (MPN). Hence, determining the precise role of this domain is critical for understanding disease pathogenesis and design of new treatment modalities. Here, we have evaluated the effect of inter-domain interactions in kinase activity and substrate specificity. By using for the first time purified recombinant JAK2 proteins and a novel peptide micro-array platform, we have determined initial phosphorylation rates and peptide substrate preference for the recombinant kinase domain (JH1) of JAK2, and two constructs comprising both the kinase and pseudokinase domains (JH1-JH2) of JAK2. The data demonstrate that (i) JH2 drastically decreases the activity of the JAK2 JH1 domain, (ii) JH2 increased the K(m) for ATP (iii) JH2 modulates the peptide preference of JAK2 (iv) the V617F mutation partially releases this inhibitory mechanism but does not significantly affect substrate preference or K(m) for ATP. These results provide the biochemical basis for understanding the interaction between the kinase and the pseudokinase domain of JAK2 and identify a novel regulatory role for the JAK2 pseudokinase domain. Additionally, this method can be used to identify new regulatory mechanisms for protein kinases that provide a better platform for designing specific strategies for therapeutic approaches.     

Vinculin, a cytoskeletal substrate of protein kinase C

Vinculin, a cytoskeletal protein localized at adhesion plaques, is a phosphoprotein containing phosphoserine, phosphothreonine, and phosphotyrosine. Vinculin has been previously shown to be a substrate for pp60src, a phosphotyrosine protein kinase, but the kinase(s) responsible for phosphorylation of the other amino acid residues is unknown. The present report examines the phosphorylation of vinculin by various serine- and threonine-specific protein kinases. Only protein kinase C, the calcium-activated phospholipid-dependent protein kinase, phosphorylates vinculin at a significant rate (24 nmol/min/mg) and displays marked specificity for vinculin. Both calcium and phosphatidylserine were required for vinculin phosphorylation by protein kinase C. In addition, both phorbol 12,13-dibutyrate (10 nM) and phorbol 12-myristate 13-acetate (10 nM) stimulated vinculin phosphorylation by protein kinase C at a limiting calcium concentration (10(-6) M). Tryptic peptide analysis revealed two major sites of phosphorylation. One site contained phosphoserine and the other contained phosphothreonine. When compared with tryptic maps of vinculin phosphorylated by src kinase, no overlapping phosphorylated peptides were found. The present findings coupled with the plasma membrane location of both these proteins suggest that vinculin may be a physiologic substrate for protein kinase C.     

Substrate specificity of protein kinases and computational prediction of substrates

To ensure signalling fidelity, kinases must act only on a defined subset of cellular targets. Appreciating the basis for this substrate specificity is essential for understanding the role of an individual protein kinase in a particular cellular process. The specificity in the cell is determined by a combination of "peptide specificity" of the kinase (the molecular recognition of the sequence surrounding the phosphorylation site), substrate recruitment and phosphatase activity. Peptide specificity plays a crucial role and depends on the complementarity between the kinase and the substrate and therefore on their three-dimensional structures. Methods for experimental identification of kinase substrates and characterization of specificity are expensive and laborious, therefore, computational approaches are being developed to reduce the amount of experimental work required in substrate identification. We discuss the structural basis of substrate specificity of protein kinases and review the experimental and computational methods used to obtain specificity information.