Understanding and exploiting substrate recognition by protein kinases

Protein kinases play a virtually universal role in cellular regulation and are emerging as an important class of new drug targets, yet the cellular functions of most human kinases largely remain obscure. Aspects of substrate recognition common to all kinases in the ATP nucleotide binding site have been exploited in the generation of analog-specific mutants for exploring kinase function and discovering novel protein substrates. Likewise, understanding interactions with the protein substrate, which differ substantially between kinases, can also help to identify substrates and to produce tools for studying kinase pathways, including fluorescent biosensors. Principles of kinase substrate recognition are particularly valuable in guiding bioinformatics and phosphoproteomics approaches that impact our understanding of signaling pathways and networks on a global scale.     

Signaling from the human melanocortin 1 receptor to ERK1 and ERK2 mitogen-activated protein kinases involves transactivation of cKIT

Melanocortin 1 receptor (MC1R), a Gs protein-coupled receptor expressed in melanocytes, is a major determinant of skin pigmentation, phototype and cancer risk. Upon stimulation by αMSH, MC1R triggers the cAMP and ERK1/ERK2 MAPK pathways. In mouse melanocytes, ERK activation by αMSH binding to Mc1r depends on cAMP, and melanocytes are considered a paradigm for cAMP-dependent ERK activation. However, human MC1R variants associated with red hair, fair skin [red hair color (RHC) phenotype], and increased skin cancer risk display reduced cAMP signaling but activate ERKs as efficiently as wild type in heterologous cells, suggesting independent signaling to ERKs and cAMP in human melanocytes. We show that MC1R signaling activated the ERK pathway in normal human melanocytes and melanoma cells expressing physiological levels of endogenous RHC variants. ERK activation was comparable for wild-type and mutant MC1R and was independent on cAMP because it was neither triggered by stimulation of cAMP synthesis with forskolin nor blocked by the adenylyl cyclase inhibitor 2',5'-dideoxyadenosine. Stimulation of MC1R with αMSH did not lead to protein kinase C activation and ERK activation was unaffected by protein kinase C inhibitors. Conversely, pharmacological interference, small interfering RNA studies, expression profiles, and functional reconstitution experiments showed that αMSH-induced ERK activation resulted from Src tyrosine kinase-mediated transactivation of the stem cell factor receptor, a receptor tyrosine kinase essential for proliferation, differentiation, and survival of melanocyte precursors, thus demonstrating a functional link between the stem cell factor receptor and MC1R. Moreover, this transactivation phenomenon is unique because it is unaffected by natural mutations impairing canonical MC1R signaling through the cAMP pathway.     

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.     

Extracellular signal-regulated kinases in T cells: characterization of human ERK1 and ERK2 cDNAs.

Extracellular signal-regulated kinases 1 and 2 are growth factor-sensitive serine/threonine kinases. cDNAs for both human kinases were isolated and sequenced. The nucleic acid and deduced protein sequences of human extracellular signal-regulated kinase 1 were 88% and 96% identical, respectively, to the homologous rat sequences. The nucleic acid and deduced protein sequences of human extracellular signal-regulated kinase 2 were 90% and 98% identical, respectively, to the corresponding rat sequences. A human extracellular signal-regulated kinase 2 specific probe was used to demonstrate that the mRNA for this kinase was present in T cells and did not change with activation. The deduced protein sequences of both human kinases were greater than 95% identical to two Xenopus kinase sequences, indicating that these enzymes are highly conserved across species.     

A bipartite substrate recognition motif for cyclin-dependent kinases

Cy or RXL motifs have been previously shown to be cyclin binding motifs found in a wide range of cyclin-Cdk interacting proteins. We report the first kinetic analysis of the contribution of a Cy motif on a substrate to phosphorylation by cyclin-dependent kinases. For both cyclin A-Cdk2 and cyclin E-Cdk2 enzymes, the presence of a Cy motif decreased the K(m(peptide)) 75-120-fold while the k(cat) remained unchanged. The large effect of the Cy motif on the K(m(peptide)) suggests that the Cy motif and (S/T)PX(K/R) together constitute a bipartite substrate recognition sequence for cyclin-dependent kinases. Systematic changes in the length of the linker between the Cy motif and the phosphoacceptor serine suggest that both sites are engaged simultaneously to the cyclin and the Cdk, respectively, and eliminate a "bind and release" mechanism to increase the local concentration of the substrate. PS100, a peptide containing a Cy motif, acts as a competitive inhibitor of cyclin-Cdk complexes with a 15-fold lower K(i) for cyclin E-Cdk2 than for cyclin A-Cdk2. These results provide kinetic proof that a Cy motif located a minimal distance from the SPXK is essential for optimal phosphorylation by Cdks and suggest that small chemicals that mimic the Cy motif would be specific inhibitors of substrate recognition by cyclin-dependent kinases.     

Cyclin-dependent kinases: inhibition and substrate recognition

Four unresolved issues of cyclin-dependent kinase (CDK) regulation have been addressed by structural studies this year - the mechanism of CDK inhibition by members of the INK4 family of CDK inhibitors, consensus substrate sequence recognition by CDKs, the role of the cyclin subunit in substrate recognition and the structural mechanism underlying CDK inhibition by phosphorylation.     

Identification of ERK2-binding domain of EBITEIN1, a novel ERK2-binding protein

We recently cloned a cDNA encoding a novel extracellular signal-regulated kinase 2 (ERK2) binding protein, EBITEIN1, by yeast two-hybrid screening. In this study, we further characterized EBITEIN1. Binding experiments using various deletion mutants identified a 40-amino acid minimal sequence for binding ERK2. Binding experiments using substitution mutants indicated the crucial role of arginine residues in this sequence. Based on empirical and bioinformatic analyses, we propose two domains in EBITEIN1. One is the minimal sequence for binding ERK2 (EB domain) and the other is the EBITEIN1 C-terminal domain (ECT domain). These results might pave the way for further empirical and bioinformatic analyses of EBITEIN1- and ERK2-mediated events.     

Centaurin-alpha1 is a phosphatidylinositol 3-kinase-dependent activator of ERK1/2 mitogen-activated protein kinases

Centaurin-alpha1 is known to be a phosphatidylinositol 3,4,5-triphosphate (PIP3)-binding protein that has two pleckstrin homology domains and a putative ADP ribosylation factor GTPase-activating protein domain. However, the physiological function of centaurin-alpha1 is still not understood. Here we have shown that transient expression of centaurin-alpha1 in COS-7 cells results in specific activation of ERK, and the activation is inhibited by co-expression of a dominant negative form of Ras. We have also found that a mutant form of centaurin-alpha1 that is unable to bind PIP3 fails to induce ERK activation and that a phosphatidylinositol 3-kinase inhibitor LY294002 inhibits centaurin-alpha1-dependent ERK activation. Furthermore, transient knockdown of centaurin-alpha1 by small interfering RNAs results in reduced ERK activation after epidermal growth factor stimulation in T-REx 293 cells. These results suggest that centaurin-alpha1 contributes to ERK activation in growth factor signaling, linking the PI3K pathway to the ERK mitogen-activated protein kinase pathway through its ability to interact with PIP3.     

SRPK1 and LBR protein kinases show identical substrate specificities

Arginine/serine protein kinases constitute a novel class of enzymes that can modify arginine/serine (RS) dipeptide motifs. SR splicing factors that are essential for pre-mRNA splicing and the lamin B receptor (LBR), an integral protein of the inner nuclear membrane, are among the best characterized proteins that contain RS domains. Two SR Protein-specific Kinases, SRPK1 and SRPK2, have been shown to phosphorylate specifically the RS motifs of the SR family of splicing factors and play an important role in regulating both the spliceosome assembly and their intranuclear distribution, whereas an LBR-associated kinase, that specifically phosphorylates a stretch of RS repeats located at the NH2-terminal region of LBR, has been recently purified and characterized from turkey erythrocyte nuclear envelopes. Using synthetic peptides representing different regions of LBR and recombinant proteins produced in bacteria we now demonstrate that SRPK1 modifies LBR with similar kinetics and on the same sites as the LBR kinase, that are also phosphorylated in vivo. These data provide significant evidence for a new role of SRPK1 in addition to that of pre-mRNA splicing.     

Diverging substrate specificity of pure human thymidine kinases 1 and 2 against antiviral dideoxynucleosides

The two thymidine (dThd) kinases in human cells, the cytosolic, S-phase-specific TK1 and the mitochondrial, constitutively expressed TK2 were purified to homogeneity as judged from sodium dodecyl sulfate-gel electrophoresis. The substrate specificity of TK1 and TK2 toward natural substrates and important nucleoside analogues was compared. With TK1, the Km values for 5-fluorodeoxyuridine (FdUrd), 3'-azido-2',3'-dideoxythymidine (AZT), and 3'-fluoro-2',3'-dideoxythymidine (FLT) were 2.2, 0.6, and 2.1 microM as compared to 0.5 microM for dThd and 9 microM for deoxyuridine (dUrd). With TK2, dUrd, deoxycytidine (dCyd), and 5-fluorodeoxyuridine (FdUrd) were efficiently phosphorylated, but with distinctly different kinetics: Michaelis-Menten kinetics with dCyd, dUrd, and FdUrd; negative cooperativity with dThd. Negative cooperativity was also observed with AZT, although this drug was a very poor substrate for TK2 with a Vmax of 5-6% of that with dThd. FLT, 2',3'-dideoxycytidine (ddCyd), and arabinofuranosylcytosine (araC) were not substrates for TK2, and 2',3'-didehydrodideoxy-thymidine (D4T) was not a substrate for TK1 or TK2. On the other hand, AZT, FLT, and D4T were competitive inhibitors with Ki values of 0.6, 6, and 2073 microM for TK1, and 2, 10, and 78 microM for TK2, respectively. The much lower tolerance for modifications of the deoxyribose moiety of TK2 as compared to TK1 is important for the design of new antiviral nucleoside analogues intended for use in cells with different expression of TK1 and TK2.     

Repression of mitogen-activated protein kinases ERK1/ERK2 activity by a protein tyrosine phosphatase in rat fibroblasts transformed by upstream oncoproteins

The observation that mitogen-activated protein (MAP) kinases ERK1 and ERK2 are constitutively activated in a number of oncogene-transformed cell lines has led to the hypothesis that prolonged activation of these enzymes is required for the transformation process. To investigate this question, we have examined the regulation of the ERK pathway in Rat1 fibroblasts transformed with activated c-Raf-1 (Raf22W), v-Ha-Ras, and v-Src. Expression of these oncoproteins had no effect on the enzymatic activity of ERK1 and ERK2 in either serum-starved or exponentially growing cells. Moreover, the stimulatory effect of serum on ERK1/ERK2 activity was substantially reduced or abrogated in these cells; this impairment was associated with a strong attenuation of c-fos gene induction. In contrast, expression of Raf22w, v-Ha-Ras, or v-Src resulted in the constitutive activation of the upstream kinases MEK1 and MEK2. Treatment of the cells with vanadate completely restored the activation of ERK1/ERK2 in oncogene-transformed cells, suggesting the involvement of a vanadate-sensitive tyrosine phosphatase. Northern blot analysis of VH1-like dual-specificity MAP kinase phosphatases did not reveal any significant difference in the mRNA expression pattern of these genes between parental and transformed Rat1 cells. Phosphoamino acid analysis indicated that ERK1 is phosphorylated on threonine, but not on tyrosine, in oncogene-transformed cells and that vanadate treatment restores tyrosine phosphorylation. We conclude from these results that ERK1/ERK2 activity is repressed by a single-specificity tyrosine phosphatase in oncogene-transformed rat fibroblasts. J. Cell. Physiol. 174:35–47, 1998. © 1998 Wiley-Liss, Inc.     

Potential role of phospholipase D2 in increasing interleukin-2 production by T-lymphocytes through activation of mitogen-activated protein kinases ERK1/ERK2

Hydrolysis of phosphatidylcholine by phospholipase D (PLD) leads to the generation of phosphatidic acid (PA), which is itself a source of diacylglycerol (DAG). These two versatile lipid second messengers are at the centre of a phospholipid signalling network and as such are involved in several cellular functions. However, their role in T-cell activation and functions are still enigmatic. In order to elucidate this role, we generated a human and a murine T-cell line that stably overexpressed the PLD2 isoform. Analysis of the Ras-MAPK pathway upon phorbol myristate acetate (PMA) and ionomycin stimulation revealed that PLD2 promoted an early and sustained increase in ERK1/2 phosphorylation in both cell lines. This response was inhibited by 1-butanol, a well known distracter of PLD activity, or upon overexpression of a dominant negative PLD2, and it was concomitant with a boost of PA/DAG production. As a functional consequence of this PLD2-dependent MAPK activation, interleukin-2 production evoked by PMA/ionomycin stimulation or CD3/CD28 engagement was enhanced in the two T-cell lines overexpressing PLD2. Thus, PLD2 emerged as an early player upstream of the Ras-MAPK-IL-2 pathway in T-cells via PA and DAG production, raising new possibilities of pharmacological manipulation in immune disorders.     

ERK1/2 MAP kinases in cell survival and apoptosis

ERK1/2 is an important subfamily of mitogen-activated protein kinases that control a broad range of cellular activities and physiological processes. ERK1/2 can be activated transiently or persistently by MEK1/2 and upstream MAP3Ks in conjunction with regulation and involvement of scaffolding proteins and phosphatases. Activation of ERK1/2 generally promotes cell survival; but under certain conditions, ERK1/2 can have pro-apoptotic functions.     

Structural basis for substrate recognition and dissociation by human transportin 1

Transportin 1 (Trn1) is a transport receptor that transports substrates from the cytoplasm to the nucleus through nuclear pore complexes by recognizing nuclear localization signals (NLSs). Here we describe four crystal structures of human Trn1 in a substrate-free form as well as in the complex with three NLSs (hnRNP D, JKTBP, and TAP, respectively). Our data have revealed that (1) Trn1 has two sites for binding NLSs, one with high affinity (site A) and one with low affinity (site B), and NLS interaction at site B controls overall binding affinity for Trn1; (2) Trn1 recognizes the NLSs at site A followed by conformational change at site B to interact with the NLSs; and (3) a long flexible loop, characteristic of Trn1, interacts with site B, thereby displacing transport substrate in the nucleus. These studies provide deep understanding of substrate recognition and dissociation by Trn1 in import pathways.     

Molecular basis for substrate specificity of protein kinases and phosphatases

Regulation of various metabolic processes occurs by the phosphorylation/dephosphorylation of enzymes. Both the protein kinases that catalyze the phosphorylations and the protein phosphatases that catalyze the dephosphorylations display relatively broad specificity, reacting with a number of distinct sites in target enzymes. In this way changes in the activity of a particular kinase or phosphatase can cause coordinated and pleiotropic responses. However, the kinases and phosphatases do not exhibit a one-to-one correspondence in their reactions. Residues at different positions may be phosphorylated by a single kinase, yet dephosphorylated by different individual phosphatases. Conversely, sites which are substrates for different individual kinases may be dephosphorylated by a single phosphatase. In exploring the molecular basis for these differences this article shows that whereas kinases react with specific primary structures that often times appear as beta bends, the phosphatases recognize higher order structure, less strictly ruled by amino acid sequence surrounding the phosphorylated site. The differences, seen in the ability of these enzymes to utilize synthetic peptide substrates, might be rationalized in terms of function. Kinases need protruding segments of structure that can be enwrapped to exclude water, thereby minimizing ATP hydrolysis and enhancing phosphotransferase activity. On the other hand phosphatases are hydrolytic enzymes that may operate especially well on protein interfaces. Hydrolytic action often measured with p-nitrophenylphosphate is not necessarily indicative of a protein phosphatase and consideration of the mechanism reveals why this substrate can be misleading.(ABSTRACT TRUNCATED AT 250 WORDS)     

Role of receptor and nonreceptor protein tyrosine kinases in H2O2-induced PKB and ERK1/2 signaling

Excessive generation of reactive oxygen species (ROS) has been implicated in the pathogenesis of many diseases, including atherosclerosis, hypertension, and vascular complications of diabetes. However, the precise mechanisms by which ROS contribute to the development of these diseases are not fully characterized. Hydrogen peroxide (H₂O₂), a ROS, has been shown to activate several signaling protein kinases, such as extracellular signal-regulated kinase (ERK)1/2 and protein kinase B (PKB) in different cell types, notably in vascular smooth muscle cells. Because these pathways regulate cellular mitogenesis, migration, proliferation, survival, and death responses, their aberrant activtion has been suggested to be a potential mechanism of ROS-induced pathologies. The upstream elements responsible for H₂O₂-induced ERK1/2 and PKB activation remain poorly characterized, but a potential role of receptor and nonreceptor protein tyrosine kinases (PTKs) as triggers that initiate such events has been postulated. Therefore, the aim of this review is to highlight the involvement of receptor and nonreceptor PTKs in modulating H₂O₂-induced ERK1/2 and PKB signaling.