Protein modification: Docking sites for kinases

Dispatch

R329

Protein modification: Docking sites for kinases Pamela M. Holland and Jonathan A. Cooper

The substrate specificities of protein kinases have been found, in many cases, to be determined at least in part by short regions within the substrate known as docking sites. Docking sites are specific and modular, and can dramatically increase the efficiency of phosphorylation. Address: Fred Hutchinson Cancer Research Center, 1100 Fairview Avenue North, Seattle, Washington 98109, USA. E-mail: [email protected] Current Biology 1999, 9:R329–R331 http://biomednet.com/elecref/09609822009R0329 © Elsevier Science Ltd ISSN 0960-9822

In eukaryotes, phosphorylation is the most common reversible protein modification, used in the regulation of multiple cellular processes. The vast number of phosphoproteins and protein kinases inside a cell raises the question of how each kinase selectively recognizes its own substrates and not those of closely-related kinases. Without specificity, different control processes would become hopelessly entangled. The problem has clearly not been solved at the level of the phosphoacceptor site sequences, as these are not particularly distinctive. Phosphorylation sites for serine and threonine kinases show considerable degeneracy, and lack the stringent specificity needed to ensure fidelity of phosphorylation. So what does determine the substrate specificity of protein kinases? It is increasingly being found that protein kinases form tight complexes with their substrates. These tight complexes are either bridged by a third protein, which provides a scaffold, or involve a direct, high-affinity interaction between a part of the kinase and a short sequence of the substrate, known as a docking site. Recent studies [1–3] have shown that docking sites increase the efficiency with which a substrate is phosphorylated by a kinase both in vitro and in cells. The sites are modular and self-contained, that is, they can be attached to different proteins to direct their phosphorylation by a specific kinase. Several docking sites can be present in a single substrate, increasing affinity for a kinase. Furthermore, docking site sequences can be predictive, allowing potential substrates for kinases to be detected by sequence. The family of mitogen activated protein (MAP) kinases, which includes the extracellular signal-regulated kinase (ERK), Jun N-terminal kinase (JNK) and p38 subfamilies, is a prime example where phosphorylation site specificity is not adequate to explain substrate selection in the cell. ERK, JNK and p38 phosphorylate different substrates and have different biological functions, despite all similarly

phosphorylating sites containing serine or threonine followed by a proline. The specificity problem is compounded because some MAP kinase substrates also fall into functional families with extensive sequence homology. For example, one family of MAP kinase substrates includes the MAPKAP kinases, which are closely related yet are phosphorylated differentially by ERK or p38. Another family of MAP kinase substrates is the bZip transcription factors, which are selectively phosphorylated by JNK or p38. The identification of distinct docking sites for ERK, JNK and p38 in MAPKAP kinases and bZip transcription factors goes a long way to explain this specificity. Specific docking sites for ERK, but not p38 or JNK, are found at the carboxyl termini of the MAPKAP kinases Rsk and Mnk2 [4,5]. Both of these proteins are found complexed to ERK in cells [5,6], and are phosphorylated by ERK in their kinase domains. The specific docking site in Rsk has now been whittled down to the pentapeptide LAQRR, followed by a leucine four or five residues downstream that may also be important and is conserved in Mnk2 [2,3] (see Table 1). Non-conservative substitution of residues in the LAQRR motif abolishes binding of ERK [2]. This region is necessary for Rsk phosphorylation by ERK in vitro and in Xenopus oocytes, and when transplanted to another protein is sufficient to confer ERK binding [2]. Other MAPKAP kinases related to Rsk and Mnk2 contain a sequence similar to LAQRR at their carboxyl termini, but are activated by other MAP kinase family members (see Table 1). These other candidate docking sites have yet to be tested by mutagenesis, but the subtle differences in the docking site sequence may be enough to provide selective activation by p38, ERK or both. The bZip transcription factor c-Jun contains a distinct docking sequence that is essential for efficient and specific phosphorylation by JNK. The docking site, known as the delta domain, was identified by binding studies with JNK [7,8]. The delta domain is a 14 residue region that includes the motif KX2+X4LXL — where + is R or K, and X is any amino acid — and lies approximately 20 residues from the nearer of two JNK phosphorylation sites [9]. Interestingly, the delta domain is poorly conserved in the Jun relative, JunD, which contains JNK phosphorylation sites, yet is highly conserved in JunB, which lacks JNK phosphorylation sites. It turns out that JNK bound to JunB can phosphorylate JunD contained in JunD–JunB heterodimers [10]. This illustrates how kinase–substrate docking and activation can be accomplished by complex formation with a dockingcompetent partner.

R330

Current Biology, Vol 9 No 9

Table 1 Protein kinase docking sites. Substrate

Docking site

Protein kinase

References

Mnk2, Rsk

LAQRRX4L

ERK

Mnk1, Msk1

LA+RR

ERK, p38

[3]

Msk2/Rskb, PRAK, Lφ++RK MAPKAPK2, MAPKAPK3

p38

[3]

LIN-1, Elk-1, SAP-1a, SAP-2

FXFP

ERK

[1]

Elk-1

D-domain KX2+X3LXL

ERK2, JNK

[2,3]

[1,11,16]

Spi-B

1–108

ERK, JNK

[17]

cJun

Delta domain KX2+X4LXL

JNK2

[7–9]

NF-AT4

A-2 domain SDASSCES

Casein kinase 1α

[18]

p107, p130, p27, p57, p21, E2F1

RRLFG

Cyclin D–Cdk4 (MRAIL in cyclin D)

MKK4

1–87

MEKK1

[23]

MEK1, MEK2

270–307 proline-rich insert

Raf-1

[24]

ATF2

47–66

JNK

[25]

Crk

SH3 domain

Abl (proline-rich region)

[26]

AFAP-110

Proline-rich regions

SH3-containing tyrosine kinases

[27]

p130Cas

pYXXP

Abl

[28]

[19–22]

+, R or K; φ, L or A; pY, phosphotyrosine.

Studies on the ERK docking sites in the Ets family transcription factor LIN-1 have afforded insights into the effects of single, duplicated and compound docking sites [1]. Gain-of-function mutations in LIN-1 cluster in the sequence FQFP, near the protein’s carboxyl terminus. These mutations impair the phosphorylation, and thus negative regulation, of LIN-1 by ERK. Examination of the sequences of mammalian and Drosophila members of the Ets family shows that the FXFP sequence is conserved through evolution. Mutating the FXFP motif reduces the affinity — increases the Km — for phosphorylation in vitro 9-fold. Adding the sequence to a short form of LIN-1 increases the affinity 6–8-fold (depending on position), and adding two copies increases the affinity 29-fold. This cooperativity implies that ERK can bind to two copies of FXFP at one time. Importantly, adding a single docking site to a short peptide substrate increases the affinity for phosphorylation 85-fold, indicating a profound effect on kinase– substrate encounters.

In a protein substrate, the effect of the docking site is less pronounced because the protein has other features that favour interaction with the kinase. Indeed, in the case of Ets proteins, another docking site for ERK has already been identified [11]. This so-called D-domain, KX2+X3LXL, is related in sequence to the delta domain on Jun and to amino-terminal sequences of some protein kinases that phosphorylate ERK relatives, for which the consensus sequence +++X1–5LXL has been defined [12]. Jacobs et al. [1] have shown that the D-domain and FXFP sequence have combinatorial effects: having both in a substrate has a multiplicative effect on the affinity for phosphorylation, consistent with their acting as independent binding sites. Quantitatively, the D-domain is more important, and together they contribute a 170-fold increase in affinity for phosphorylation of the Ets protein Elk-1. They also account for kinase specificity, because the D-domain binds both JNK and ERK, but the FXFP docking site is ERK-specific. The results reported by Jacobs et al. [1] make two other important points about docking sites: they can be predictive and they can be used to generate competitive inhibitors of protein phosphorylation. To test the predictive ability of docking sites, the authors searched protein databases for 50-residue segments containing two potential phosphorylation sites and the FXFP docking sequence. Such segments were found in about 50% of GATA-family transcription factors, MAP kinase phosphatases and the protein kinases A-Raf and KSR. In each case, binding to, or phosphorylation by, ERK is already known or suspected. To test the potential of docking sites as phosphorylation inhibitors, Jacobs et al. [1] used a synthetic peptide containing the FXFP sequence. This peptide inhibited an ERK-dependent response in Xenopus oocytes, while a control peptide did not. The mechanism of inhibition was not determined precisely, but these results encourage hope that it will be possible to develop specific peptide and non-peptide inhibitors of phosphorylation events that depend on docking sites. The preassembly of a kinase with its substrate not only increases specificity but also shortens the time between kinase activation and phosphorylation of the first substrate molecule. Catalytic efficiency could in principle be compromised, however, if the binding is so tight that the phosphorylated substrate does not dissociate from the kinase; in this case, substrate turnover would be limited to one. In actuality, some kinase–docking site complexes are destabilized after substrate phosphorylation [5,6], and others may be sufficiently unstable to allow turnover. This would be consistent with the observation that the phosphorylation reactions for substrates containing docking sites show an increase in Vmax as well as in affinity [1]. The preassembly of kinase–substrate complexes could also aid their localization in the cell. For example, the

Dispatch

MAP kinase p38 is restrained in the nucleus when inactive, in part by binding to its nuclear-localized substrate MAPKAP kinase 2 [13]. Activation induces export of the complex from the nucleus. Despite the multiplicity of docking sites (Table 1), structural details of how they bind to protein kinases are sadly lacking. Indeed, the portion of the kinase that binds the docking site has been identified in very few cases. The binding site for the delta domain of Jun was mapped using chimeras of different JNK family members that differ in their affinity for the delta domain [9]. Binding to the delta domain was found to require a short, surfaceexposed strand located next to the catalytic pocket of the kinase, which varies in sequence between JNK family members. The atomic interactions that engage the delta domain, however, are not known. Indeed, the only example where a docking interaction has been analyzed with atomic resolution may not be typical, as it involves a non-catalytic ‘scaffold’ protein that bridges the kinase to its substrate [14]. The scaffold protein is a cyclin, which connects a cyclin-dependent kinase (Cdk) to its substrate p27. A short docking site on p27, RXLFG, forms a rigid coil that interacts with a conserved, hydrophobic groove in the cyclin [14]. Mutation of the hydrophobic patch in cyclin A was found to block binding to proteins with RXLFG motifs and to disrupt cyclin A function in cells [15]. With the precedents of JNK and cyclin A–Cdk2, it seems that the kinase surfaces that bind to docking sites are likely to be as varied and numerous as the sites with which they interact. References 1. Jacobs D, Glossip D, Xing H, Muslin AJ, Kornfeld K: Multiple docking sites on substrate proteins form a modular system that mediates recognition by ERK MAP kinase. Genes Dev 1999, 13:163-175. 2. Gavin A-C, Nebreda AR: A MAP kinase docking site is required for phosphorylation and activation of p90rsk/MAPKAP kinase-1. Curr Biol 1999, 9:281-284. 3. Smith JA, Poteet-Smith CE, Malarkey K, Sturgill TW: Identification of an extracellular signal-regulated kinase (ERK) docking site in ribosomal S6 kinase, a sequence critical for activation by ERK in vivo. J Biol Chem 1999, 274:2893-2898. 4. Zhao Y, Bjorbaek C, Moller DE: Regulation and interaction of pp90(rsk) isoforms with mitogen-activated protein kinases. J Biol Chem 1996, 271:29773-29779. 5. Waskiewicz AJ, Flynn A, Proud CG, Cooper JA: Mitogen-activated protein kinases activate the serine/threonine kinases Mnk1 and Mnk2. EMBO J 1997, 16:1909-1920. 6. Hsiao KM, Chou SY, Shih SJ, Ferrell JE: Evidence that inactive p42 mitogen-acivated protein kinase and inactive Rsk exist as a heterodimer in vivo. Proc Natl Acad Sci USA 1994, 91:5480-5484. 7. Adler V, Polotskaya A, Wagner F, Kraft AS: Affinity-purified c-Jun amino-terminal protein kinase requires serine/threonine phosphorylation for activity. J Biol Chem 1992, 267:17001-17005. 8. Hibi A, Lin A, Smeal T, Minden A, Karin M: Identification of an oncoprotein- and UV-responsive protein kinase that binds and potentiates the c-Jun activation domain. Genes Dev 1993, 7:2135-2148. 9. Kallunki T, Su B, Tsigelny I, Sluss HK, Dérijard B, Moore G, Davis R, Karin M: JNK2 contains a specificity-determining region responsible for efficient c-Jun binding and phosphorylation. Genes Dev 1994, 8:2997-3007.

R331

10. Kallunki T, Deng T, Hibi M, Karin M: c-Jun can recruit JNK to phosphorylate dimerization partners via specific docking interactions. Cell 1996, 87:929-939. 11. Yang SH, Whitmarsh AJ, Davis RJ, Sharrocks AD: Differential targeting of MAP kinases to the ETS-domain transcription factor Elk-1. EMBO J 1998, 17:1740-1749. 12. Bardwell L, Thorner J: A conserved motif at the amino termini of MEKs might mediate high-affinity interaction with the cognate MAPKs. Trends Biochem Sci 1996, 21:373-374. 13. Ben-Levy R, Hooper S, Wilson R, Paterson HF, Marshall CJ: Nuclear export of the stress-activated protein kinase p38 mediated by its substrate MAPKAP kinase-2. Curr Biol 1998, 8:1047-1057. 14. Russo AA, Jeffrey PD, Patten AK, Massague J, Pavletich NP: Crystal structure of the p27Kip1 cyclin-dependent-kinase inhibitor bound to the cyclin A-Cdk2 complex. Nature 1996, 382:325-331. 15. Schulman BA, Lindstrom DL, Harlow E: Substrate recruitment to cyclin-dependent kinase 2 by a multipurpose docking site on cyclin A. Proc Natl Acad Sci USA 1998, 95:10453-10458. 16. Yang SH, Whitmarsh AJ, Davis RJ, Sharrocks AD: Differential targeting of MAP kinases to the ETS-domain transcription factor Elk-1. EMBO J 1998, 17:1740-1749. 17. Mao C, Ray-Gallet D, Tavitian A, Moreau-Gachelin F: Differential phosphorylations of Spi-B and Spi-1 transcription factors. Oncogene 1996, 12:863-873. 18. Zhu J, Shibasaki F, Price R, Guillemot J-C, Yano T, Dotsch V, Wagner G, Ferrara P, Mckeon F: Intramolecular masking of nuclear import signal on NF-AT4 by casein kinase 1 and MEKK1. Cell 1998, 93:851-861. 19. Chen J, Saha P, Kornbluth S, Dynlacht BD, Dutta A: Cyclin-binding motifs are essential for the function of p21CIP1. Mol Cell Biol 1996, 16:4673-4682. 20. Zhu L, Harlow E, Dynlacht BD: p107 uses a p21CIP1-related domain to bind cyclin/cdk2 and regulate interactions with E2F. Genes Dev 1995, 9:1740-1752. 21. Lacy S, Whyte P: Identification of a p130 domain mediating interactions with cyclin A/Cdk2 and cyclinE/Cdk2 complexes. Oncogene 1997, 14:2395-2406. 22. Adams PD, Sellers WR, Sharma SK, Wu AD, Nalin CM, Kaelin WG Jr: Identification of a cyclin-cdk2 recognition motif present in substrates and p21-like cyclin-dependent kinase inhibitors. Mol Cell Biol 1996, 16:6623-6633. 23. Xia Y, Wu Z, Xu B, Murray B, Karin M: JNKK1 organizes a MAP kinase module through specific and sequential interactions with upstream and downstream components mediated by its aminoterminal extension. Genes Dev 1998, 12:3369-3381. 24. Catling AD, Schaeffer H-J, Reuter CWM, Reddy GR, Weber MJ: A proline-rich sequence unique to MEK1 and MEK2 is required for Raf binding and regulates MEK function. Mol Cell Biol 1995, 15:5214-5225. 25. Livingstone C, Patel G, Jones N: ATF-2 contains a phosphorylationdependent transcriptional activation domain. EMBO J 1995, 14:1785-1797. 26. Ren R, Ye Z-S, Baltimore D: Abl protein-tyrosine kinase selects the Crk adapter as a substrate using SH3-binding sites. Genes Dev 1994, 8:783-795. 27. Guappone AC, Flynn DC: The integrity of the SH3 binding motif of AFAP-110 is required to facilitate tyrosine phosphorylation by, and stable complex formation with, Src. Mol Cell Biochem 1997, 175:243-252. 28. Mayer BJ, Hirai H, Sakai R: Evidence that SH2 domains promote processive phosphorylation by protein-tyrosine kinases. Curr Biol 1995, 5:296-305.