Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
doi:10.1016/j.jmb.2004.04.043 J. Mol. Biol. (2004) 339, 1025–1039 REVIEW Structural Modes of Stabilization of Permissive Phosphorylation Sites in Protein Kinases: Distinct Strategies in Ser/Thr and Tyr Kinases A. Krupa1, G. Preethi2 and N. Srinivasan1* 1 Molecular Biophysics Unit Indian Institute of Science Bangalore 560012, India 2 Department of Biotechnology Bharathiar University Coimbatore 641043, India Protein kinases phosphorylate several cellular proteins providing control mechanisms for various signalling processes. Their activity is impeded in a number of ways and restored by alteration in their structural properties leading to a catalytically active state. Most protein kinases are subjected to positive and negative regulation by phosphorylation of Ser/Thr/Tyr residues at specific sites within and outside the catalytic core. The current review describes the analysis on 3D structures of protein kinases that revealed features distinct to active states of Ser/Thr and Tyr kinases. The nature and extent of interactions among well-conserved residues surrounding the permissive phosphorylation sites differ among the two classes of enzymes. The network of interactions of highly conserved Arg preceding the catalytic base that mediates stabilization of the activation segment exemplifies such diverse interactions in the two groups of kinases. The N-terminal and the C-terminal lobes of various groups of protein kinases further show variations in their extent of coupling as suggested from the extent of interactions between key functional residues in activation segment and the N-terminal aC-helix. We observe higher similarity in the conformations of ATP bound to active forms of protein kinases compared to ATP conformations in the inactive forms of kinases. The extent of structural variations accompanying phosphorylation of protein kinases is widely varied. The comparison of their crystal structures and the distinct features observed are hoped to aid in the understanding of mechanisms underlying the control of the catalytic activity of distinct subgroups of protein kinases. q 2004 Elsevier Ltd. All rights reserved. *Corresponding author Keywords: protein kinase; phosphorylation; regulation; stabilization of active site; active and inactive state Introduction Phosphorylation of intra-cellular proteins by protein kinases serve as molecular switches of various cellular processes including metabolism, gene expression, cell division, motility and differentiation.1 – 3 Toggling between distinct actiAbbreviations used: PKA, protein kinase A; PKB, protein kinase B; GS, glycine-serine; GSK-3, glycogen synthase kinase-3; GSK, glycogen synthase kinase; CK2, casein kinase-2; CSK, C-terminal Src kinase; HCK, hemapoietic cell kinase. E-mail address of the corresponding author: [email protected] vation states endowed with unique conformational features brings about the tight regulation of these enzymes.4 Multiple modes of regulation of protein kinases have enabled perception of specific signals leading to selective activation of downstream signalling cascades.5,6 Critical structural features of catalytically active states of all protein kinases carrying out phospho transfer are very similar.4,7 Phosphorylation on the activation segment, a key regulatory element, acts as a switch for the catalytic activity of protein kinases belonging to distinct subgroups of Ser/ Thr and Tyr kinase families.8 Protein kinases regulated by such phosphorylation form a network of residue contacts at the active site that are well 0022-2836/$ - see front matter q 2004 Elsevier Ltd. All rights reserved. 1026 Review: Phosphorylation Sites in Kinases Figure 1. The 3D fold of catalytic core of Ser/Thr or Tyr kinase with their key functional elements is shown. This Figure as well as Figures 3, 4, 6, 7, 9, and 11, are generated using SETOR.37 conserved in their “on” or maximally active state. The key residues involved in stabilization of the catalytic site as described above are invariant across different subgroups of kinases with similar mechanism of regulation.5 The focus of this review is to provide our current understanding on the extent of conservation of the nature of residue contacts in kinases at the activation loop as revealed from the survey of available 3D structures in active and inactive states. We describe the variations in the residue contacts observed across Ser/Thr and Tyr kinase families. Further, the extension of these contacts to residues conserved specifically in Ser/ Thr and Tyr kinases and their implications in alternate modes of stabilization of activation segment are discussed. Apart from the positive regulatory phosphorylation sites within the catalytic core, the inhibitory phosphorylation sites and their residue environment have been surveyed to explore the conservation of their interaction network. The varying residue environment of these phosphosites as exemplified by cyclin dependent kinases provides insight into their mechanism of inhibition. Analysis of other critical functional features including the Lys-Glu salt-bridge dyad and nucleotide binding of the protein kinases in active and inactive states, reveal differences that provide a platform for further investigations on their role in catalysis and ligand binding. The 3D structure of kinase catalytic core The first 3D structure of a protein kinase, protein kinase A9 (PKA) as determined by X-ray crystallography revealed the basic bilobal scaffold that has now been observed in all the protein kinase structures (Figure 1) solved to date. The N-terminal lobe of the kinase fold comprises of an anti-parallel b-sheet made of five b-strands and a single a-helix referred to as aC-helix. The C-terminal lobe is larger and is mainly composed of a-helices. The nucleotide binding and the substrate-binding pocket are located in the cleft between the two lobes. The phosphate groups of ATP are positioned for phospho transfer by their interactions with conserved residues in the N and C-terminal lobes. These include a glycine-rich loop characterized by “GXGXXG” (where X represents any amino acid) motif between the b1 and b2 strands, a Lys residue localized by a salt bridge formed by a Lys-Glu pair (K72† and E91) and Mg2þ ions. Conserved Asn (N171) and Asp (D184) further coordinate the metal ions. The catalytic loop situated in the C-terminal lobe contains the aspartate (D166) referred to as catalytic base that facilitates extraction of proton from the hydroxyl side-chains of phospho-sites of the substrates. The activation segment of 20–30 residues in length caps the C-terminal lobe. This segment forms a part of the substratebinding pocket and shows high structural variation in the active and inactive kinase structures. Phosphorylation of the activation segment of protein kinases The common catalytic scaffold shared by the eukaryotic protein kinases switches between two extreme conformations of on and off states of high and least activity, respectively, in a number † Unless otherwise specified residue numbering as in the crystal structure of PKA9 is followed. Review: Phosphorylation Sites in Kinases of ways. The transitions between two distinct functional states are mediated by a variety of strategies in different kinases such as phosphorylation, and interactions of additional domains within the protein kinases and/or by binding to other regulatory proteins.4 Phosphorylation of the activation segment is a conserved feature in the regulation of most protein kinases belonging to the group of “RD” kinases.5 The RD kinases contain an Arg preceding the catalytic aspartate in the catalytic loop. Phosphorylation at the conserved Ser/Thr in the activation segment serves as a nucleation centre, neutralizing a cluster of positively charged residues including Arg of the RD motif.5 Such electrostatic interactions are crucial for the relative orientation of the “DFG” motif of the activation segment and the other catalytic and substrate binding residues to facilitate phosphotransfer. The conformation of the activation segment is controlled by the phosphorylation, which promotes secondary changes including re-orientation between the lobes, relief of autoinhibition and the correct disposition of catalytic residues. The basic residues that interact with the phospho-amino acid on the activation segment are highly conserved across various kinases and are referred to as contact residues (Figures 2(a) and (b) and 3) in the subsequent discussion. Protein kinases in the off state adopt distinct conformation characteristic to their inherent mode of regulation. The nature of interactions between phospho-amino acid and the contact residues observed for various protein kinases in their active and inactive states has been described earlier.5 These interactions are well conserved in the on state of most protein kinases consistent with the notion of a unique conformation and functional features of the protein kinases in their active states. Crystal structures of protein kinases in on state also reveal distinct pattern of interaction confined to either Ser/Thr kinases or tyrosine kinases. These differences at the active site are discussed below. Distinct strategies of stabilization of the activation segment The crystal structures of protein kinases of the RD group requiring phosphorylation at the activation segment and available in both active and inactive states are listed in Table 1. Comparison of the 3D structures of Ser/Thr and tyrosine kinases in on state, reveals a striking difference in the nature of contacts of the phospho-amino acid in the activation segment with the highly conserved Arg of the RD motif. Ion pair between the Arg residue (in RD motif) and the phospho-amino acid that forms a network of electrostatic interactions with other contact residues (Figures 2(a) and 3), are observed in the 3D structures of PKA, PKB, CDK and MAP kinase. The Arg in RD doublet (RD-Arg) is absolutely conserved in all the tyrosine kinases known to date. The phospho-amino acid 1027 in the activation loop of the tyrosine kinases, however, lacks a direct interaction with the RD-Arg. It should be noted that among the four residues equivalent to the contact residues in the Ser/Thr kinases mentioned above, only two are conserved as positively charged residues (RD-Arg and the Arg from the activation segment) in the tyrosine kinases. The contact residues from the N-terminal lobe are either apolar or uncharged and lack interactions with the phospho-amino acid (Figures 2(b) and 4(a) and (b)). The mode of stabilization of activation segment in the tyrosine kinases to reach out for the substrate phospho-sites therefore seems to differ from that of Ser/Thr kinases. Water mediated interaction of the RD-Arg with the phospho-amino acid, has been observed in the crystal structure of LCK.10 A novel cation –p (aromatic ring) interaction11,12 between the RD-R and the aromatic side-chain of well-conserved Phe (Phe1186 in IRK) has been identified as a unique feature (Figure 4(a) and (b)) among the tyrosine kinases. The critical parameters that define such a contact include the inter-atomic distance between the CG, CD, NE, CZ, NH1, or NH2 atoms of arginine and any of the six carbon atoms of the aromatic rings to be less than 4.6 Å.13 Such aromaticcation interactions are characterized by the parallel alignment of the aromatic ring to the plane of the guanidium group. The Phe residue located in the loop joining the activation segment to the aF-helix further interacts with various hydrophobic residues including Met1176 and Leu1171 (p þ 1 and p þ 3 substrate-binding pocket) located in the activation segment of IRK. This Phe is hence likely to be an important mediator in channeling the conformational changes occurring at the activation segment to the catalytic loop via the RD-Arg. Similar interactions are observed between the equivalent phenylalanine and the hydrophobic residues in the activation segment of the other tyrosine kinases (Figure 4(b)). In the receptor kinase, Ret a M918T mutation leads to multiple endocrine neoplasia.14 The mutated residue is equivalent to M1176 of IRK and is involved in the formation of p þ 3 substrate-binding pocket of IRK.15 The apolar contacts of M1176 with the Phe are likely to be directed by the cation – p interactions between the planar ring of Phe and the RD-Arg. The influence of the aromatic-cation interaction between the RD-Arg and the conserved Phe residue in activation loop of tyrosine kinases on substrate binding needs to be further explored. The extended network of interactions involving the RD-Arg of the catalytic loop, the highly conserved phenylalanine (F1186) and the hydrophobic residues (M1176, L1171) can therefore be regarded as novel “local signal integration motifs”8 unique to tyrosine kinases. Occurrence of such interactions in both active and inactive tyrosine kinases as evident from their crystal structures suggests a pre-formed substrate-binding site, which is, however, restricted to substrate access by the activation segment that wraps around the catalytic site in 1028 Review: Phosphorylation Sites in Kinases Figure 2 (legend opposite) inactive tyrosine kinases. The cation– p interaction is also observed between two other highly conserved residues of tyrosine kinases, non-RD Arg in the catalytic loop (R1136 in IRK) and a Trp residue (W1175 in IRK) in the activation segment. An analogous interaction between the Lys (K168 in PKA) and Thr (T201) is found in Ser/Thr kinases only in their active state (Figure 3). The positively charged amino acid residues in the catalytic loop of the tyrosine kinases therefore seek an alternate mode of stabilization mediated through cation –p interactions in contrast to the electrostatic and polar interactions observed for equivalent residues in the active state of Ser/Thr kinases. The crystal structures of receptor tyrosine kinases (IRK and IGFR) in their active state15,16 also reveal contacts of the phospho-amino acid with two basic residues from the activation segment. One of these residues is equivalent to the contact residue (R1155 in IRK) that forms a part of the phospho-amino acid net- work conserved in all the protein kinases described in the previous sections. The other basic residue (R1164 in IRK) is well conserved in tyrosine kinases and has very distinct conformations in the active and inactive structures. Summary of the differences between Ser/Thr and Tyr kinases in the permissive phospho-amino acid network is schematically represented in Figure 5. From the above discussion it is evident that Ser/Thr kinase and Tyr kinases exhibit a few unique features in their phospho-amino acid network. The amino acid residues (T201 of PKA, W1175 of IRK) involved in the interactions with the positively charged residues (K168 of PKA and R1136 of IRK) in the catalytic loop of Ser/Thr and Tyr kinases are specifically conserved within the group of Ser/Thr and Tyr kinases. These distinct modes of interactions provide a structural basis for the conservation of residues specific to Ser/ Thr and tyrosine kinase family.17 Review: Phosphorylation Sites in Kinases 1029 Figure 2. Multiple sequence alignment of Ser/Thr and Tyr kinases showing key functional residues. Sequences of (a) Ser/Thr kinases and (b) tyrosine kinases, with 3D structures and discussed in the text, are shown. Numbers shown in parentheses in the alignment indicate the number of residues inserted in the corresponding region of the alignment. Color coding used in the alignment is: conserved residues, shaded gray; catalytic residues, black and bold; contact residues, green; other residues interacting with contact residues and specifically conserved in Ser/Thr or tyrosine kinases, red. Figure 3. The phospho-amino acid network at the activation segment of protein kinase B (1o6l), a typical RD kinase, is shown. The phospho-amino acid is colored red, and the contact residues are shown in green. Similar color-coding is used to represent them in subsequent Figures. The conserved residues, lysine of the catalytic loop and the threonine (discussed in the text) that are in contact in the phospho-amino acid in on state are shown in blue. 1030 Review: Phosphorylation Sites in Kinases Table 1. List of crystal structures of on and off states of RD kinases requiring phosphorylation for activation, investigated in the current review Kinase CAPK CAPK CDK CDK MAPK MAPK Protein kinase B Protein kinase B Insulin receptor kinase Insulin receptor kinase Insulin-like growth factor 1 receptor kinase Insulin-like growth factor 1 receptor kinase Lymphocyte kinase Src PDB code Conformational state Reference 1atp 1j3h 1hcl 1jst 1erk 2erk 1o6l 1gzn 1ir3 1irk 1k3a 1jqh 3lck 2src Active Apoenzyme (inactive) Inactive Active Inactive Active Active Inactive Active Inactive Active Bis-phosphorylated (partially active) Active Inactive 9 27 38 39 40 30 41 42 15 43 16 44 10 22 Further, preferences exhibited by specific contact residues are observed within Ser/Thr kinases for their interaction with the phospho-amino acids. For instance, the interaction of the N-terminal lobe with the phospho-amino acid is mediated by the N-terminal contact residues (His87 and His196 in PKA and protein kinase B (PKB) of AGC subfamily, respectively). The equivalent contact residues in CDK and MAPK of CMGC subfamily, (namely Thr47 and Arg65) do not interact with the phospho-amino acid residues. Alternately another contact residue from the N-terminal lobe of CDK and MAPK (Arg50 and Arg68) interacts with the phospho-amino acid (Figure 6). These contact residues provide a link between the N-terminal and the C-terminal lobes of the kinase.5 Preferences for such contact residues indicate minor structural differences conserved within the subfamily. Various factors including residue environment, oligomeric state of the catalytically active forms and intrinsic flexibility of the catalytic core are likely to contribute to the conformational features conserved within subgroups of protein kinases. Non-phospho-amino acid interactions with the contact residues The RD-kinases, phosphorylated on the activation segment have well conserved basic residues (contact residues) forming a part of the phosphoamino acid network. The role of these residues in the charge neutralization by phospho-amino acid residues is therefore well established. The available crystal structures of protein kinases in different functional states suggest other conserved interactions restricted to specific functional states of distinct families of protein kinases. For instance, the side-chain of RD-Arg contacts the hydroxyl of tyrosine (Y215 in PKA, Y328 in PKB) located in a loop that continues as aF-helix (Figure 3), in active forms of some of the Ser/Thr kinases. The equivalent residue is replaced by hydrophobic residues incapable of making polar contacts in other RD types of Ser/Thr kinases like checkpoint kinase and TGF-b receptor kinase that are regulated by mechanisms differing from the charge neutralization by phospho-amino acid residues or acidic residues. In the TGF-bR-1, the phosphorylation at the conserved glycine-serine (GS) region switches the binding affinity of the receptor towards the Smad2 and the binding site is no longer recognized by the inhibitor protein FKBP12, thus relieving the inhibition.18 The crystal structures of the above mentioned enzymes are currently available in the inhibited or inactive states.19,20 The 3D structures of these kinases in the on state would help in elucidating the roles of various conserved residues in catalysis and regulation. It has to be noted that the conserved RD-Arg-Tyr pair of Ser/Thr kinases is formed by residues equivalent to the cation – p pair forming residues of tyrosine kinases. The current observations indicate that the location of residues linking the catalytic loop and the activation loop of Ser/Thr and Tyr kinases are conserved in their primary and tertiary structure while the nature of interactions between the residues vary across the two groups of kinases. The contact residues also contribute to the stabilization of nucleotide binding residues in the off state. Such interactions are exemplified by the salt bridge between Glu of Lys-Glu dyad and the contact residue from the activation loop of Src family and Tec families of kinases. Crystal structures of inactive hemapoietic cell kinase,21 Src kinase22 and BTK23 reveal the interactions of contact residues (R385 of HCK, Src and R544 of BTK) with Glu of Lys-Glu dyad. The above interaction (Figure 7) is mediated by the conserved lysine (K72 in PKA) in the active state. The disruption of the salt-bridge contacts between the Glu of aC-helix and the contact residue in the activation segment is likely to mediate further conformational changes that leads to appropriate disposition of residues at the catalytic centre upon phosphorylation. Hence, the contact residues also serve as links between the N-terminal lobe and C-terminal lobe of protein kinases in Src and BTK family in their inactive state. Review: Phosphorylation Sites in Kinases 1031 Figure 4. The phospho-amino acid in the activation segment of (a) IRK and (b) LCK and its interactions with the contact residues in the on state are shown. Aromatic residues involved in cation – p interactions with RD-Arg and other basic residues discussed in the text are shown in blue. Distinct modes of ATP binding in the on and off states of Ser/Thr protein kinases The crystal structures of substrate and ATP bound ternary complexes of various Ser/Thr protein kinases including PKA, CDK24,25 and phosphorylase kinase26 have indicated a common binding mode of the substrate peptides in their active conformations. The backbone of the substrate peptides adopts an extended conformation to align the hydroxyl group suitably at the catalytic site. An analysis on the conformation of the co-substrate ATP, bound to active and inactive states of protein kinases, by pair-wise superposition of ATP atoms reveals distinct features. Average deviations of atoms in base, ribose and phosphate groups 1032 Review: Phosphorylation Sites in Kinases Figure 5. Schematic representation of the interactions of phospho-amino acid in the activation segment of (a) Ser/ Thr and (b) Tyr kinases. PKA (a) and IRK (b) have been chosen as examples to illustrate the phospho-amino acid network in the two groups of protein kinases. Distinct nature of interactions with RD-Arg, characterized by electrostatic interactions in Ser/Thr kinases and cation– p (aromatic) interactions in tyrosine kinases, stabilize the activation segment. Unlike in Ser/Thr kinases, N-terminal helix (aC-helix) of tyrosine kinases lacks interactions with the phosphoamino acid. The arrows (blue) indicate the residues involved in cation – p (aromatic) interactions in tyrosine kinases. arrived at from all pairwise superpositions of ATP molecules in kinases are represented in Figure 8. ATP displays similar conformation in the on state of Ser/Thr protein kinases as indicated by the lower deviation of their adenine ring, ribose and phosphate groups compared to ATP molecules in the inactive states of kinases (Figure 8). This feature is consistent with appropriate disposition Figure 6. Phospho threonine (pT 183) at the activation segment of active MAPK and its interactions with specific contact residues are shown. Arg65 (blue) equivalent to His196 of PKB (Figure 3) is drifted away from the catalytic site and lacks interactions with pT183. Review: Phosphorylation Sites in Kinases 1033 Figure 7. 3D structure of Src kinase (2Src) in an inhibited state. The interaction between residues involved in Lys-Glu dyad (red) in the active state and residues from the activation segment (blue) is shown. Figure 8. Extent of deviation of conformation of ATP bound to active and inactive states of protein kinases. The average deviation (Å) of equivalent atoms of ATP observed for adenine, ribose and phosphate groups among active forms (red) and between active and inactive forms (dark green) is shown. 1034 of the gamma phosphate groups for inline phospho-transfer. ATP molecules bound to inactive forms of protein kinases show wide variations in the conformation of their chemical groups. Although the adenine ring and the ribose groups of ATP deviate least across active and inactive states, deviations greater than 1 Å are observed in the equivalent atoms of the phosphate groups of the ATP. The beta and gamma phosphate groups are highly mobile with equivalent atoms deviating by greater than 3 Å. These variations in the conformation of ATP are likely to arise form the large conformational changes accompanying the key residues involved in ATP binding and coordination with the metal ions (Mg2þ, Mn2þ) during transition into different states. ATP binding residues of the kinase are localized in distinct regions of the fold that includes the Gly-loop, hinge region, catalytic loop and the activation segment. In CDK the key residues comprising the ATP-binding pocket include Gly13, Thr14, Val18 of Gly loop; Lys33 of Lys-Glu dyad; Phe80, Glu81, Phe82 and Leu83 of the hinge region; Asn132 of the catalytic loop and Asp145 of the activation segment. These residues vary in their extent of interactions with ATP in their active and inactive states. While some of these residues (G13, T14, F80, D145) are involved in ATP binding only in the active state, other residues (F82, L83) interact with ATP in the inactive state. The remaining residues participate in ATP binding in both states of activity. High mobility of phosphate groups enables their co-ordination with the metal ions that are displaced due to the variation in the positions of residue ligands (N132, D145) located in the catalytic and activation loop between the two conformational states of kinases. The atoms of ATP bound to active form of kinases in general can be well superimposed. Large variation is, however, observed in the conformation of ribose and phosphate groups of ATP bound to phospho-CDK/cyclin complex in comparison with ATP bound to active states of other Ser/Thr kinases. Accommodation of ATP in altered conformations within the structurally similar catalytically active scaffold without hindering phospho-transfer therefore seems to be influenced by the binding of cyclin to CDK. The conformations of ATP bound to Ser/Thr kinases and tyrosine kinases in the on state also varies to a large extent. The RTKs lack the formation of a correct Lys-Glu ion-pair in their active states as described in a later section. The nature of stabilization of the activation segment of tyrosine kinases in the on state is also distinct from that of Ser/Thr kinases as described earlier. High sequence divergence is observed in the nucleotide binding pockets of Ser/Thr and Tyr kinases. These distinct functional and structural features of Tyr kinases are hence likely to influence the mode of ATP binding in the tyrosine kinases. The different Review: Phosphorylation Sites in Kinases features appear to play an important role in the distinct nature of co-ordination of phosphate groups of ATP with the conserved lysine and metal ions. Influence of phosphorylation and ATP binding on closed state of protein kinases Appropriate spatial orientation of the ATP binding residues and the relative disposition of catalytic residues in the activation and catalytic loop are characteristic to the closed state of protein kinases. The aC-helix is a key structural element in the N-terminal lobe mediating the conformational changes occurring at the catalytic center through an ion pair (Lys72-Glu91 in PKA) that bridges phosphate groups of ATP. The charge neutralization of the basic cluster of residues in the active site5 is another critical feature in RD kinases regulated by phosphorylation as described earlier. The crystal structure of the apo-enzyme of phosphorylated PKA27 reveals the electrostatic interactions of the phospho-amino acid similar to its ternary complex. However, the Lys and Glu residues are spatially distant, incapable of salt bridge formation and coordination with a- and b-phosphate groups of ATP. Phosphorylation of the activation segment in PKA therefore appears to alter the active site locally, with least influence on the formation of ion pair in the N-terminal lobe. In contrast to the Ser/Thr kinases, the receptor tryosine kinases lack the Lys-Glu ion-pair interaction in their active states as observed in the 3D structures of insulin receptor15 and IGF receptor kinases.16 However, the phospho-amino acid in the activation segment is capable of interacting with the basic residues in its vicinity in a unique way as described in the previous sections. In the conformationally restrained state of c-Abl28,29 the Lys-Glu ion pair interaction is retained. However, the residues at the catalytic site are not aligned for phospho transfer, as the contact residues are not drawn towards Tyr412, which is phosphorylated for complete activation of the kinase. The formation of Lys-Glu ion pair and conformational changes at the active site accompanying the activation process are hence suggested to vary across different protein kinases. The mechanisms of acquisition of a closed, functionally competent state in protein kinases is therefore suggested to vary based on the extent of reorganization of the active site necessary to initiate phospho transfer. Interactions of “secondary” phospho-amino acid residues in protein kinases The phosphorylation occurs at multiple sites on protein kinases providing additional controls on their activity. In MAP kinases of CMGC subfamily, the successive dual phosphorylation of the Thr and Tyr residues in TXY motif of the activation Review: Phosphorylation Sites in Kinases segment results in local and global conformational changes30 realigning the positively charged contact residues close to the phospho threonine (pT183). The second dianionic site unique to MAPK (Figure 9(a)) created by the phosphorylation of Tyr residue also interacts with two basic residues (Arg189, Arg192). Glycogen synthase kinase-3 (GSK-3), another member of CMGC subfamily is usually kept active in cells and is known to phosphorylate a wide range of cellular proteins.31 PKB mediated phosphorylation on the N-terminal serine residue of GSK-3b leads to its inactivation. GSK-3b is unique in its choice of substrates with higher preferences for phosphorylated (primed) substrates. The phosphorylation on the activation segment for regulation of this enzyme is nonobligatory. Analogous interactions are observed in the 3D structure32 of GSK between the phospho- 1035 tyrosine (pY) 416 (only phospho-amino acid) in the activation segment and Arg220 and Arg223 (Figure 9(b)). The basic residues hence are likely to play a major role in accommodating the phospho-amino acid on the primed substrates similar to their contacts with pY416 in the absence of the substrate. The four main groups of CMGC kinases (CDK, MAPK, glycogen synthase kinase (GSK) and casein kinase-2 (CK2)) show high conservation of basic residues in equivalent positions with the exception of CDK. An apolar residue (Leu166) replaces the residue equivalent to the Arg189 (MAPK) in CDK while the other basic residue is well conserved. However, the role of these residues in CDK that possess mechanisms of activation distinct from MAPK and GSK is unclear. CK2 is a constitutively active kinase and no phosphorylation at the Figure 9. The interactions of secondary phospho-amino acid residues in the active and inactive states of protein kinases. (a) Active state of MAPK, (b) phospho-glycogen synthase kinase (GSK), (c) inactive state of c-Src kinase (inactive) and (d) inactive state of hemapoietic cell kinase (HCK). 1036 catalytic subunit is so far documented for its regulation. It has to be noted that the Sky1p33 of Saccharomyces cerevisiae (a non-RD kinase) is the only other protein kinase in addition to the RD kinases of the CMGC group, with basic residues conserved at equivalent positions. The inhibitory phosphorylation site of Src family protein kinases is located C-terminal to the kinase domain. Upon phosphorylation the phospho-tyrosine (Tyr527) docks into an intramolecular site located in the SH2 domain.22,34 This interaction is mediated by an ion pair between the wellconserved pair of Arg residues (R155, R175) in the SH2 domain and the phosphate (Figure 9(c)). The crystal structure of inactive HCK,21 another member of Src family of kinases reveals analogous interactions with a pair of basic residues in its SH2 domain. The basic residues include an Arg equivalent to the R175 of Src while the other basic residue (K203) is located in a different strand unlike R155 in the Src (Figure 9(d)). Thus, different basic residues interact with the inhibitory pTyr in HCK and Src as summarized in a schematic Figure (Figure 10). These observations suggest that the spatial location of the phospho-amino acid interacting residues is conserved, while their positions in the primary structure vary. The C-terminal Src kinase (CSK) is similar to Src kinases in their modular organization with SH3, SH2 and kinase domains. The in-built mechanism of regulation is quite distinct from that of Src kinases lacking phospho sites in the activation segment and the C-terminal tail. While basic residues interacting with inhibitory phospho-sites in Src are well conserved in CSK, a salt bridge exists between one of the basic residues and Asp residue located in the helix of SH2 domain in their inactive state35 analogous to the contact of phospho-amino acid. These observations indicate that the phosphoamino acid network at the inhibitory phospho sites of Src show preferences for the basic residues as exemplified by HCK and Src kinases. Among the cytoplasmic tyrosine kinases, with similar domain organization cAbl and CSK lack inhibitory phospho sites (cAbl and CSK). Conserved nega- Review: Phosphorylation Sites in Kinases tively charged residues of cAbl and CSK (E123 in cAbl) form ion pairs with the basic residues equivalent to the inhibitory phospho site interacting residues in the off state. Coupling of inhibitory phospho-sites of CDK with the Lys-Glu dyad Phosphorylation on conserved Thr and Tyr (Thr14 and Tyr15) residues in the Gly rich loop of CDK by wee1 kinase36 leads to its inactivation. The crystal structures of CDKs available in inhibited and inactive states reveal these phosphorylation sites to be distant from the aC-helix. However, the partially active cyclin bound forms of CDK and the fully active phospho-CDK/cyclin complex show contacts between the hydroxyl group of Tyr15 and the Glu51 of the Lys-Glu dyad that anchors ATP. The proximity of the Tyr15 (in glycine-loop) and Glu51 (Figure 11) induced by cyclin binding, can cause a repulsion between like charges upon phosphorylation of Tyr15 drifting the aC-helix away from Gly-rich loop. The electrostatic repulsion is therefore likely to mediate the transition of the partially active and fully active CDKs to the inactive states. Crystal structure of CDK phosphorylated at the inhibitory sites should therefore clarify the role of interactions of the Lys-Glu dyad with inhibitory phospho sites in the Gly-rich loop. Conclusions The active state of Ser/Thr and Tyr kinases adopt grossly similar structures that enable efficient catalysis on diverse substrates. Phosphorylation at the activation segment is the major means of regulation of most of the RD class of kinases. Interactions subsequent to phosphorylation at the active site of these kinases facilitate substrate access and phospho-transfer. While the chemical environment of the active sites in the vicinity of phospho-amino acid is well conserved, the spatial location, extent and mode of interactions vary Figure 10. Schematic representation of interactions of inhibitory phospho-tyrosine in the Src family of protein kinases. (left) c-Src and (right) hemapoietic cell kinase (HCK). The network of phosphoamino acid is comprised of distinct positively charged residues well conserved across the two protein kinases. Review: Phosphorylation Sites in Kinases 1037 Figure 11. 3D structure of cyclin dependent kinase showing interactions between Tyr15 (red), the inhibitory phosphorylation site and conserved Glu51 (red) of Lys (green)-Glu dyad in cyclin boundphosphorylated complex (1JST). across Ser/Thr and Tyr kinases. The Arg of the RD motif stabilizes the activation segment by electrostatic interaction with the phospho-amino acid in most Ser/Thr kinases in their on state. In contrast in the Tyr kinase the RD-Arg lacks the conserved interaction with the phospho-amino acid and alternately makes cation – p interactions in on and off states with phenylalanine well conserved in known Tyr kinases. The phenylalanine further extends its interaction to the substratebinding pocket and is hence likely to be important for the formation of substrate binding sites. Crystal structures of Ser/Thr kinases in the active state also reveal conservation of polar interactions of RD-R with conserved tyrosine that is uniquely conserved among RD kinases regulated by negative charge or phospho-amino acid in the activation segments. While the role of such interactions is currently unclear, it is likely to be another feature of the active site of protein kinases specifically conserved in the on state. The significance of the stabilizing interactions of the contact residues with the phospho-amino acid in the active state of kinases is well established. Ion pair between the contact residue and the LysGlu dyad in the off-state of Src kinases and BTK suggest coupling of the activation segment and aC-helix, which communicates the changes to the catalytic center subsequent to phosphorylation. Preferences for interactions of the phosphoamino acid with specific contact residues in aC-helix, are observed within the group of Ser/ Thr kinases as exemplified by AGC and CMGC family of kinases. The variation in orientation of aC-helix in distinct functional states of these subgroups of kinases and the intrinsic flexibility associated with the catalytic core for such movements, appear to be critical for the choice of residues in the phospho-amino acid network. Inhibitory phosphorylation sites of Src family of kinases also create a cluster of positively charged residues, which are usually exposed in the on states. Distinct set of basic residues, however, participate in such interactions in the c-Src and HCK family of kinases. The tyrosine in the Gly-rich loop of CDKs is proximal to Glu of Lys-Glu dyad in the active and partially active structures. Phosphorylation of the Tyr in the cyclin bound forms of CDKs leading to inhibition of CDK is hence suggested to result from the repulsion between Glu and the dianionic charge. The available 3D structures of protein kinases have enabled the identification of a subset of unique features associated with the active state of Ser/Thr and tyrosine kinases despite many common structural features and mechanisms of phospho transfer. The participation of residues conserved specifically in the two groups of kinases in unique interactions provides a structural basis for their conservation. While interactions of the phospho-amino acid residues located at various sites in the 3D-scaffold of the protein kinases with positively charged residues is observed, the extent of neutralization by the dianionic charge and the spatial location of the positively charged residues may vary. ATP bound to Ser/Thr kinases in the on state, 1038 in general, adopt similar conformation of adenine, ribose and phosphate groups. The conformation of ATP varies between active and inactive forms of ATP, with the phosphate groups showing the large deviation in their relative orientation. Variations are, however, observed in active states of other protein kinases (CDK) which might be partially influenced by additional factors like the interactions of the kinases with regulatory proteins. The vast number of crystal structures of RDkinases in distinct conformational states have aided in the identification of conserved and varied functional features at their active sites subsequent to phosphorylation. The number of 3D structures currently available for protein kinases regulated by other modes of regulation in on and off states is limited. The emergence of crystal structures of non-RD kinases in their on state will enhance the current understanding of the factors stabilizing their active sites. Review: Phosphorylation Sites in Kinases 11. 12. 13. 14. 15. Acknowledgements This research is supported by the Wellcome Trust, London, in the form of Senior Fellowship in Biomedical Sciences to N.S. A.K. is supported by Council of Scientific and Industrial Research, New Delhi. G.P. is a visiting student trainee to the Indian Institute of Science. 16. 17. 18. References 1. Schenk, P. W. & Snaar-Jagalska, B. E. (1999). Signal perception and transduction: the role of protein kinases. Biochim. Biophys. Acta, 1449, 1 – 24. 2. Cohen, P. (2000). The regulation of protein function by multisite phosphorylation—a 25-year update. Trends Biochem. Sci. 25, 596– 601. 3. Hunter, T. (2000). Signaling—2000 and beyond. Cell, 100, 113 – 127. 4. Huse, M. & Kuriyan, J. (2002). The conformational plasticity of protein kinases. Cell, 109, 275–282. 5. Johnson, L. N., Noble, M. E. & Owen, D. J. (1996). Active and inactive protein kinases: structural basis for regulation. Cell, 85, 149– 158. 6. Kobe, B. & Kemp, B. E. (1999). Active site-directed protein regulation. Nature, 402, 373– 376. 7. Johnson, L. N., Lowe, E. D., Noble, M. E. & Owen, D. J. (1998). The eleventh datta lecture. The structural basis for substrate recognition and control by protein kinases. FEBS Letters, 430, 1 – 11. 8. Johnson, L. N. & Lewis, R. J. (2001). Structural basis for control by phosphorylation. Chem. Rev. 101, 2209– 2242. 9. Knighton, D. R., Zheng, J. H., Ten Eyck, L. F., Ashford, V. A., Xuong, N. H., Taylor, S. S. & Sowadski, J. M. (1991). Crystal structure of the catalytic subunit of cyclic adenosine monophosphate-dependent protein kinase. Science, 253, 407– 414. 10. Yamaguchi, H. & Hendrickson, W. A. (1996). Structural basis for activation of human lymphocyte 19. 20. 21. 22. 23. 24. 25. kinase LCK upon tyrosine phosphorylation. Nature, 384, 484– 489. Singh, J. & Thornton, J. M. (1990). SIRIUS. An automated method for the analysis of the preferred packing arrangements between protein groups. J. Mol. Biol. 211, 595– 615. Flocco, M. M. & Mowbray, S. L. (1994). Planar stacking interactions of arginine and aromatic side-chains in proteins. J. Mol. Biol. 235, 709– 717. Momany, F. A., Carruthers, L. M., McGuire, R. F. & Scheraga, H. A. (1974). Intermolecular potentials and applications to the packing configurations and lattice energies in crystals of hydrocarbons, carboxylic acids, amines and amides. J. Phys. Chem. 78, 1595– 1620. Hofstra, R. M., Landsvater, R. M., Ceccherini, I., Stulp, R. P., Stelwagen, T., Luo, Y., Pasini, B., Hoppener, J. W., van Amstel, H. K. & Romeo, G. (1994). A mutation in the RET protooncogene associated with multiple endocrine neoplasia type 2B and sporadic medullary thyroid carcinoma. Nature, 367, 375– 376. Hubbard, S. R. (1997). Crystal structure of the activated insulin receptor tyrosine kinase in complex with peptide substrate and ATP analog. EMBO J. 16, 5572– 5581. Favelyukis, S., Till, J. H., Hubbard, S. R. & Miller, W. T. (2001). Structure and autoregulation of the insulin-like growth factor 1 receptor kinase. Nature Struct. Biol. 8, 1058– 1063. Singh, J. (1994). Comparison of conservation within and between the Ser/Thr and Tyr protein kinase family: proposed model for the catalytic domain of the epidermal growth factor receptor. Protein Eng. 7, 849– 858. Huse, M., Muir, T. W., Xu, L., Chen, Y. G., Kuriyan, J. & Massague, J. (2001). The TGF beta receptor activation process: an inhibitor- to substrate-binding switch. Mol. Cell, 8, 671– 682. Huse, M., Chen, Y. G., Massague, J. & Kuriyan, J. (1999). Crystal structure of the cytoplasmic domain of the type I TGF beta receptor in complex with FKBP12. Cell, 96, 425–436. Chen, P., Luo, C., Deng, Y., Ryan, K., Register, J., Margosiak, S., Tempczyk-Russell, A., Nguyen, B., Myers, P., Lundgren, K., Kan, C. C. & O’Connor, P. M. (2000). The 1.7 Å crystal structure of human cell cycle checkpoint kinase Chk1: implications for Chk1 regulation. Cell, 100, 681– 692. Sicheri, F., Moarefi, I. & Kuriyan, J. (1997). Crystal structure of the Src family tyrosine kinase Hck. Nature, 385, 602 –609. Xu, W., Doshi, A., Lei, M., Eck, M. J. & Harrison, S. C. (1999). Crystal structures of c-Src reveal features of its autoinhibitory mechanism. Mol. Cell, 3, 629– 638. Mao, C., Zhou, M. & Uckun, F. M. (2001). Crystal structure of Bruton’s tyrosine kinase domain suggests a novel pathway for activation and provides insights into the molecular basis of X-linked agammaglobulinemia. J. Biol. Chem. 276, 41435– 41443. Brown, N. R., Noble, M. E., Endicott, J. A. & Johnson, L. N. (1999). The structural basis for specificity of substrate and recruitment peptides for cyclindependent kinases. Nature Cell Biol. 1, 438– 443. Brown, N. R., Noble, M. E., Lawrie, A. M., Morris, M. C., Tunnah, P., Divita, G., Johnson, L. N. & Endicott, J. A. (1999). Effects of phosphorylation of Review: Phosphorylation Sites in Kinases 26. 27. 28. 29. 30. 31. 32. 33. 34. threonine 160 on cyclin-dependent kinase 2 structure and activity. J. Biol. Chem. 274, 8746– 8756. Lowe, E. D., Noble, M. E., Skamnaki, V. T., Oikonomakos, N. G., Owen, D. J. & Johnson, L. N. (1997). The crystal structure of a phosphorylase kinase peptide substrate complex: kinase substrate recognition. EMBO J. 16, 6646 –6658. Akamine, P., Madhusudan, Wu, J., Xuong, N. H., Ten Eyck, L. F. & Taylor, S. S. (2003). Dynamic features of cAMP-dependent protein kinase revealed by apoenzyme crystal structure. J. Mol. Biol. 327, 159– 171. Hantschel, O., Nagar, B., Guettler, S., Kretzschmar, J., Dorey, K., Kuriyan, J. & Superti-Furga, G. (2003). A myristoyl/phosphotyrosine switch regulates c-Abl. Cell, 112, 845– 857. Nagar, B., Hantschel, O., Young, M. A., Scheffzek, K., Veach, D., Bornmann, W., Clarkson, B., SupertiFurga, G. & Kuriyan, J. (2003). Structural basis for the autoinhibition of c-Abl tyrosine kinase. Cell, 112, 859–871. Canagarajah, B. J., Khokhlatchev, A., Cobb, M. H. & Goldsmith, E. J. (1997). Activation mechanism of the MAP kinase ERK2 by dual phosphorylation. Cell, 90, 859– 869. Welsh, G. I., Miyamoto, S., Price, N. T., Safer, B. & Proud, C. G. (1996). T-cell activation leads to rapid stimulation of translation initiation factor eIF2B and inactivation of glycogen synthase kinase-3. J. Biol. Chem. 271, 11410 –11413. Bax, B., Carter, P. S., Lewis, C., Guy, A. R., Bridges, A., Tanner, R., Pettman, G., Mannix, C., Culbert, A. A., Brown, M. J., Smith, D. G. & Reith, A. D. (2001). The structure of phosphorylated GSK-3beta complexed with a peptide, FRATtide, that inhibits beta-catenin phosphorylation. Structure (Camb), 9, 1143– 1152. Nolen, B., Ngo, J., Chakrabarti, S., Vu, D., Adams, J. A. & Ghosh, G. (2003). Nucleotide-induced conformational changes in the Saccharomyces cerevisiae SR protein kinase, Sky1p, revealed by X-ray crystallography. Biochemistry, 42, 9575– 9585. Young, M. A., Gonfloni, S., Superti-Furga, G., Roux, B. & Kuriyan, J. (2001). Dynamic coupling between the SH2 and SH3 domains of c-Src and Hck underlies their inactivation by C-terminal tyrosine phosphorylation. Cell, 105, 115 – 126. 1039 35. Ogawa, A., Takayama, Y., Sakai, H., Chong, K. T., Takeuchi, S., Nakagawa, A., Nada, S., Okada, M. & Tsukihara, T. (2002). Structure of the carboxylterminal Src kinase, Csk. J. Biol. Chem. 277, 14351 – 14354. 36. Russell, P. & Nurse, P. (1987). Negative regulation of mitosis by wee1 þ , a gene encoding a protein kinase homolog. Cell, 49, 559– 567. 37. Evans, S. V. (1993). SETOR: hardware-lighted threedimensional solid model representations of macromolecules. J. Mol. Graph. 11, 134– 138. see also pp. 127 –128. 38. Schulze-Gahmen, U., Brandsen, J., Jones, H. D., Morgan, D. O., Meijer, L., Vesely, J. & Kim, S. H. (1995). Multiple modes of ligand recognition: crystal structures of cyclin-dependent protein kinase 2 in complex with ATP and two inhibitors, olomoucine and isopentenyladenine. Proteins: Struct. Funct. Genet. 22, 378– 391. 39. Russo, A. A., Jeffrey, P. D. & Pavletich, N. P. (1996). Structural basis of cyclindependent kinase activation by phosphorylation. Nature Struct. Biol. 3, 696– 700. 40. Zhang, F., Strand, A., Robbins, D., Cobb, M. H. & Goldsmith, E. J. (1994). Atomic structure of the MAP kinase ERK2 at 2.3 Å resolution. Nature, 367, 704 –711. 41. Yang, J., Cron, P., Good, V. M., Thompson, V., Hemmings, B. A. & Barford, D. (2002). Crystal structure of an activated Akt/protein kinase B ternary complex with GSK3-peptide and AMP-PNP. Nature Struct. Biol. 9, 940– 944. 42. Yang, J., Cron, P., Thompson, V., Good, V. M., Hess, D., Hemmings, B. A. & Barford, D. (2002). Molecular mechanism for the regulation of protein kinase B/ Akt by hydrophobic motif phosphorylation. Mol. Cell, 9, 1227– 1240. 43. Hubbard, S. R., Wei, L., Ellis, L. & Hendrickson, W. A. (1994). Crystal structure of the tyrosine kinase domain of the human insulin receptor. Nature, 372, 746 –754. 44. Pautsch, A., Zoephel, A., Ahorn, H., Spevak, W., Hauptmann, R. & Nar, H. (2001). Crystal structure of bisphosphorylated IGF-1 receptor kinase: insight into domain movements upon kinase activation. Structure (Camb), 9, 955– 965. Edited by J. Thornton (Received 11 February 2004; received in revised form 21 April 2004; accepted 22 April 2004)