Characterization of ERK1 activation site mutants and the effect on recognition by MEK1 and MEK2.

To discern MEK1 and MEK2 specificity for their substrate, extracellular signal-regulated kinase (ERK), site-directed mutagenesis was performed on the amino acid residues flanking the regulatory phosphorylation sites of ERK1. These ERK1 mutants were analyzed for the ability to act as a substrate for MEK1 and MEK2. Based on both phosphorylation and activation analyses, the mutants could be divided into four classes: 1) dramatically decreased phosphorylation and activation, 2) enhanced basal kinase activity, 3) preferentially enhanced phosphorylation of tyrosine and decreased phosphorylation of threonine, and 4) increased threonine phosphorylation with an increase in activation. In general, the residues proximal to the regulatory phosphorylation sites of ERK1 had greater influence on both phosphorylation and activation. This is consistent with the highly specific recognition of the ERK1 regulatory sites by MEK. Mutation of Arg-208 or Thr-207 to an alanine residue significantly altered the relative phosphorylation on Thr-202 and Tyr-204. The Arg-208 to alanine mutant increased the phosphorylation of Tyr-204 approximately 4-fold yet almost completely eliminated the phosphorylation on Thr-202. In contrast, mutation of Gly-199 to alanine resulted in an increased phosphorylation of Thr-202 relative to Tyr-204. This suggests that both Gly-199 and Arg-208 play important roles in determining the relative phosphorylation of Thr-202 and Tyr-204. Our results demonstrate that residues in the phosphorylation lip of ERK play an important role in the recognition and phosphorylation by MEK.

To discern MEK1 and MEK2 specificity for their substrate, extracellular signal-regulated kinase (ERK), sitedirected mutagenesis was performed on the amino acid residues flanking the regulatory phosphorylation sites of ERK1. These ERK1 mutants were analyzed for the ability to act as a substrate for MEK1 and MEK2. Based on both phosphorylation and activation analyses, the mutants could be divided into four classes: 1) dramatically decreased phosphorylation and activation, 2) enhanced basal kinase activity, 3) preferentially enhanced phosphorylation of tyrosine and decreased phosphorylation of threonine, and 4) increased threonine phosphorylation with an increase in activation. In general, the residues proximal to the regulatory phosphorylation sites of ERK1 had greater influence on both phosphorylation and activation. This is consistent with the highly specific recognition of the ERK1 regulatory sites by MEK. Mutation of Arg-208 or Thr-207 to an alanine residue significantly altered the relative phosphorylation on Thr-202 and Tyr-204. The Arg-208 to alanine mutant increased the phosphorylation of Tyr-204 approximately 4-fold yet almost completely eliminated the phosphorylation on Thr-202. In contrast, mutation of Gly-199 to alanine resulted in an increased phosphorylation of Thr-202 relative to Tyr-204. This suggests that both Gly-199 and Arg-208 play important roles in determining the relative phosphorylation of Thr-202 and Tyr-204. Our results demonstrate that residues in the phosphorylation lip of ERK play an important role in the recognition and phosphorylation by MEK.
Mitogen-activated protein kinases (MAPKs) 1 are important components of various signaling pathways, phosphorylate numerous substrates, and may be key components in linking growth factor receptor activation to serine/threonine protein phosphorylation (1)(2)(3). MAPK activation can be induced by numerous mitogenic stimuli including phorbol esters, cytokines, T-cell antigens, and growth factors with tyrosine kinase receptors such as insulin, epidermal growth factor, fibroblast growth factor, and platelet-derived growth factor (3)(4)(5)(6). Stimulation of MAPK in various non-lymphoid cell types leads to a multitude of cellular responses including phosphorylation of microtubule-associated proteins involved in microtubule rearrangement (7,8) and phosphorylation resulting in subsequent activation of transcription factors TCF and STAT (9 -11). In lymphoid cells, activation of MAPK results in stimulation of T-cells to produce cytokines (12)(13)(14). The involvement of MAPK in such a variety of cells and cellular processes emphasizes a potentially vital role of MAPK in mediating cellular signal transduction. One of the most unique features of MAPK is that it must be phosphorylated on both threonine and tyrosine to exhibit full enzymatic activity (15). Dephosphorylation of either phosphothreonine or phosphotyrosine completely inactivates MAPK, emphasizing the importance of both residues for activation (16 -18).
MAPKs are also referred to as extracellular signal-regulated kinases (ERKs). The upstream activator of ERK is a dual specific kinase called MAPK or ERK kinase (MEK), which phosphorylates ERK2 on both threonine 183 and tyrosine 185 (15). Mammalian isoforms of MEK, two of which (MEK1 and MEK2) can phosphorylate and activate ERK, were identified by molecular cloning techniques (18 -24). ERK is the only known substrate for MEK, with all other proteins tested thus far being phosphorylated at least 1000-fold less efficiently by MEK (25).
MEK is activated by a number of different serine/threonine kinases including the proteins encoded by the proto-oncogene c-raf (26 -29). In fact, the regulatory phosphorylation sites in MEK were mapped using immunoprecipitated Raf from epidermal growth factor-stimulated Swiss 3T3 cells (30). Activation was shown to occur through phosphorylation of 2 serine residues at positions 218 and 222 of MEK1 (30). Constitutively active MEK1 can be obtained by mutation of the regulatory phosphorylation sites to glutamic acid (31,32). Alternatively, if these sites are mutated to alanine (S218A/S222A) the resulting MEK mutant is incapable of being activated by Raf (30). Phosphorylation and activation of MEK occur in response to the same mitogenic stimuli leading to ERK activation (2,6,33), again providing indirect evidence of MEK's substrate specificity for ERK.
The crystal structure of the unphosphorylated form of ERK2 reveals that Thr-183 and Tyr-185 are contained within a loop structure (34). In this inactivated form, Tyr-185 is buried in a hydrophobic pocket and Thr-183 is exposed to the surface of the molecule. Since the active form of ERK2 is phosphorylated on both Tyr-185 and Thr-183, ERK2 must undergo a conformational change in this loop structure upon association with MEK. Although the exact mechanism of the MEK/ERK interaction is unknown it is highly likely that some feature of the looped structure is important for determining MEK specificity. Thus, it is quite probable that the amino acid residues con-tained within this phosphorylation lip contribute to both the conformational mobility of ERK and the specificity of ERK recognition by MEK.
In this study, we have performed numerous site-directed mutations in both regions flanking the regulatory phosphorylation sites of ERK1 to further define MEK's specificity for its substrate, ERK. The ERK mutants were tested for differences in phosphorylation and activation by GST-MEK1 or GST-MEK2. Results in this paper indicate that four classes of mutants have been obtained: the first exhibits dramatically decreased phosphorylation and activation, the second shows enhanced basal kinase activity of ERK, the third exhibits preferentially enhanced phosphorylation of tyrosine and substantially decreased phosphorylation of threonine, and the fourth revealed an increase in threonine phosphorylation and an increase in activation. Results in this paper further indicate that the specificity of MEK1 and MEK2 for their substrate ERK1 is rendered, in part, by the amino acid residues flanking the regulatory phosphorylation sites of ERK1. These flanking residues not only influence the overall effectiveness of ERK1 as a substrate of MEK1 and MEK2 but also affect the relative phosphorylation levels of Thr-202 versus Tyr-204.

Methods
Bacterial Purification of GST-MEKs and ERKs-Cells from a 1-liter culture were pelleted by centrifugation at 3500 rpm for 20 min. The supernatant was discarded and the pellets were resuspended in 20 ml of Buffer A. Cells were lysed by two passes through a French press at 1200 p.s.i. The lysate was centrifuged at 12,000 rpm for 10 min to pellet the cell debris. The supernatant was then added to 3 ml of glutathioneagarose beads and mixed gently on ice for 1 h. The samples were briefly centrifuged at 3000 rpm for 5 min, and the supernatant was discarded. The matrix was transferred to a 10-ml polypropylene column and washed with 40 ml of Buffer A. The matrix was then equilibrated with 10 ml of Buffer B. GST-ERKs were cleaved using the thrombin cleavage method described by Guan and Dixon (36). Samples were concentrated by a Centricon 30 (Amicon, Inc.) according to the manufacturer's instructions. Purified proteins were frozen at Ϫ80°C in a final concentration of 10% glycerol, 1 mM phenylmethylsulfonyl fluoride, and 5 mM EDTA (pH 8.0). Proteins were separated on 10% SDS-PAGE and quantitated by a densitometric scanner using bovine serum albumin as a standard.
Purification of MEKs was performed as stated above for the ERKs with the exception of the elimination of the thrombin cleavage procedure. Instead, GST-MEKs were eluted from the glutathione-agarose with 6 ml of Buffer C containing 5 mM glutathione. Samples were dialyzed with two exchanges of Buffer C. GST-MEKs were concentrated by a Centricon 30 according to the manufacturer's instructions. Purified proteins were frozen at Ϫ80°C in 10% glycerol, 1 mM phenylmethylsulfonyl fluoride, and 5 mM EDTA (pH 8.0).
Site-directed Mutagenesis of ERK1-Human ERK1 plasmid was a generous gift from Dr. S. Pelech, University of British Columbia. The entire 5Ј-coding region of human ERK1 to the BglII site was cloned in the pAlter-1 vector (Promega). The mutagenesis was performed according to the manufacturer's instructions using synthetic oligonucleotides that contain the mutated bases. Mutations were confirmed by Sanger dideoxy DNA sequencing (51). The mutated human ERK1s were then subcloned into pGEX-2T vector in frame with GST (36).
MEK Expression Vector-BamHI sites were introduced into the 5Јand 3Ј-ends of the coding sequence of MEK by polymerase chain reaction (24). These polymerase chain reaction products were then sub-cloned, in frame with GST, into the BamHI site of pGEX-KG (24).
In Vitro Phosphorylation Assays-Human ERK1 mutants used in the phosphorylation assays contained the conserved lysine (Lys-71) mutated to an arginine residue. Human ERK1 (0.3 g) mutants were incubated with 0.3 g of either GST-MEK1 or GST-MEK2 (0.3 g) in the presence of Buffer D and [␥-32 P]ATP (5000 cpm/pmol). Samples were incubated at 30°C for 30 min. Reactions were terminated by addition of SDS sample buffer and separated by 10% SDS-PAGE. The gel was then transferred to Immobilon-P and visualized by autoradiography. Bands were quantitated by a Molecular Dynamics, Inc. PhosphorImager.
In Vitro Activation Assays-The activation of ERK1 was performed by incubation with recombinant GST-MEK for 20 min at 30°C in a 30-l reaction volume containing 18 mM HEPES (pH 7.5), 20 M ATP, and 10 mM magnesium acetate. To initiate the kinase reaction, myelin basic protein (20 g) and [␥-32 P]ATP (5000 cpm/pmol) were added in the same reaction buffer and incubated for an additional 30 min at 30°C. Half of the reaction volume (20 l) was applied to P81 phosphocellulose filters. The filters were washed with 180 mM phosphoric acid (5 times) and rinsed with 95% ethanol. Filters were quantitated by liquid scintillation counting.
Phosphoamino Acid Analysis-Phosphorylated ERK1 mutants were resolved on 10% SDS-PAGE and transferred to an Immobilon-P membrane. 32 P incorporated bands were excised and hydrolyzed in 6 N HCl for 2 h at 110°C. Samples were dried and washed in distilled water (2 ϫ 1 ml) and dissolved in 10 l of distilled water. Phosphoamino acid analysis was performed by separation in Buffer E using one-dimensional electrophoresis on cellulose plates (Eastman Kodak Co.) and visualized by autoradiography. Amino acids were quantitated by a Molecular Dynamics, Inc. PhosphorImager.

RESULTS AND DISCUSSION
ERK is activated by a dual specific kinase (MEK) by phosphorylation on specific threonine and tyrosine residues. The native form of ERK is the only substrate identified thus far for MEK (21). This is based on the inability of MEK to phosphorylate various proteins or peptides known to act as substrates for either protein serine/threonine kinases or protein tyrosine kinases. Furthermore, two-dimensional phosphopeptide mapping in combination with mass spectrometry techniques has shown that, for ERK2, only the regulatory residues Thr-183 and Tyr-185 are phosphorylated upon activation by MEK (15). The portion of the ERK2 sequence containing the phosphorylation sites is conserved in ERK1 and thus the same specificity is presumed to hold for residues Thr-202 and Tyr-204 of ERK1. The crystal structure of the unphosphorylated form of ERK2 reveals that Thr-183 and Tyr-185 are contained within a loop structure (34). The Tyr-185 is buried in a hydrophobic pocket, and Thr-183 is exposed to the surface of the molecule. Since ERK2 is more highly phosphorylated on Tyr-185 it has been suggested that the protein must undergo both local and global conformational changes in order for phosphorylation of this residue to occur upon association with MEK (34). Thus, it is conceivable that the amino acid residues flanking the regulatory phosphorylation sites in both ERK1 and ERK2 may contribute to interactions that are vital for their structural mobility. Furthermore, it is plausible to speculate that these flanking residues of the regulatory phosphorylation sites may contribute to the specificity of ERK activation by MEK. We have undertaken mutation assays of ERK1 to investigate this hypothesis.
Expression and Purification of Recombinant Human ERK1 Mutants-Recombinant ERK1 proteins were overexpressed as GST fusion proteins and purified by glutathione-agarose affinity chromatography (36). Purified ERK1 mutants were then cleaved with thrombin as stated in the experimental procedures. Upon analysis by SDS-PAGE, all 32 ERK1 mutants were approximately 85% pure and migrated as a doublet at 44 kDa (data not shown). The amount of purified protein obtained from a 1-liter culture was approximately 1-5 mg for all ERK1 mutants.
Recombinant MEK proteins were overexpressed as GST fusion proteins (GST-MEK1 or GST-MEK2) and purified by glutathione-agarose affinity chromatography for use in ERK activation studies. Since both MEKs have molecular weights similar to that of ERK1, MEKs were not cleaved with thrombin but were purified as GST fusion proteins. This aided in separation on SDS-PAGE, as GST-MEK1 or GST-MEK2 migrated at molecular masses of 69 and 70 kDa, respectively (37).
Mutational Analysis of Human ERK1-In an effort to discern the substrate specificity of MEK for ERK1, in vitro sitedirected mutagenesis studies were performed on the conserved "TEY" sequence and both amino acid flanking regions of ERK1, residues 197-209 (Fig. 1). Mutational analyses initially concentrated on these regions surrounding the regulatory phosphorylation sites of ERK1 due to their probable role in imparting MEK substrate specificity (21,37) and the inferred ability of this part of the molecule to form a conformationally mobile "looped structure" (34). To investigate this specificity, mutations were performed on both wild type kinase-active ERK1 and the catalytically impaired enzyme (ERK1*) made by mutation of the conserved lysine (Lys-71) to an arginine. Residue Lys-71 is located in subdomain II and is known to be essential for kinase activity (38,39). ERK1* mutants were assayed for in vitro phosphorylation by GST-MEK1 or GST-MEK2. In vitro kinase assays were performed on catalytically active ERK1 mutants to determine whether this activation by GST-MEK1 or GST-MEK2 correlated with ERK1* phosphorylation. Mutants that revealed interesting phosphorylation and activation results were further analyzed by phosphoamino analysis.
Phosphorylation of ERK1* in the Presence of GST-MEK2-Phosphorylation sites Thr-202 and Tyr-204 were initially substituted with residues incapable of being phosphorylated. As expected, in vitro phosphorylation by GST-MEK2 followed by phosphoamino acid analysis of mutants ERK1*T202A (data not shown) and ERK1*Y204F (Fig. 2) revealed that MEK2 only phosphorylated the ERK1 mutants on tyrosine or threonine, respectively, results that are consistent with previous observations (25). This result demonstrates that phosphorylation of Thr-202 and Tyr-204 can occur independently. These same two residues were also mutated to serine residues that could potentially be phosphorylated. However, serine phosphorylation was not observed for either ERK1*T202S or ERK1*Y204S (Fig.  2). This inability to phosphorylate serine residues at the regulatory sites further illustrates the extremely high specificity for substrate recognition of MEK. These data demonstrate that in order for phosphorylation by MEK to occur, residues 202 and 204 of ERK1 must be Thr and Tyr, respectively.
To determine whether the high substrate specificity of MEK2 is influenced by residues flanking the phosphorylation sites of ERK1, each amino acid in both flanking regions was initially substituted with an alanine residue. In general, residues in closer proximity to the phosphorylation sites have a more prominent effect than distal residues (Fig. 3). The only exception to this was mutant W209A, which had less than 30% phosphorylation. Mutants L201A and V205A showed less than 50% phosphorylation; residues F200A, E203Q, and Y204F had less than 25% phosphorylation. Interestingly, mutation of residue Arg-208 to alanine dramatically increased the phosphorylation of ERK1* by MEK2 approximately 4-fold. These results indicated that residues flanking Thr-202 and Tyr-204 have a significant influence on the ability of ERK1 to be phosphorylated by MEK2. However, to determine the specific features of this flanking region important for substrate recognition, further mutation analyses were performed on selected residues.
Mutation of the glutamic acid at position 203 to a glutamine (E203Q) was performed to determine what effect the negative charge at this position has on the ability of MEK to phosphorylate human ERK1. In addition, residue Glu-203 was mutated to an aspartic acid (E203D) to determine the effects of side chain length. ERK1*E203Q exhibited a 4-fold decrease in the level of phosphorylation, and even the more conservative mutation, E203D, showed a similar 4-fold decrease in phosphorylation. The dramatic effect of this relatively minor change would seem to indicate a pivotal role for Glu-203 in ERK1 phosphorylation beyond simply providing a negatively charged center. This central amino acid of the dual phosphorylation site motif is not fully conserved among the MAPK family members. In the ERKs and yeast Saccharomyces cerevisiae proteins, encoded by the genes KSS1, FUS3, and MPK1, the motif is TEY (40, 41). However, this site differs in the stress-activated MAPKs, HOG1 and p38 (TGY) and JNK (TPY) (42)(43)(44). With the regu- Site-directed mutagenesis was performed on the conserved TEY sequence and both flanking sequences as illustrated. Kinase-impaired mutants contain the conserved lysine (Lys-71) substituted by an arginine (ERK1*). The phosphorylation target threonine and tyrosine residues are denoted by asterisks.

FIG. 2. Phosphoamino acid analysis of ERK1* mutants.
Phosphorylated ERK1* mutants were resolved on SDS-PAGE and transferred to an Immobilon-P membrane. 32 P incorporated bands were excised and hydrolyzed in 6 N HCl for 2 h at 110°C. Phosphoamino acid analysis was performed by separation using one-dimensional electrophoresis on cellulose plates (Kodak) and visualized by autoradiography. Phosphothreonine, phosphotyrosine, phosphoserine, origin, and unin- latory phosphorylation sites being invariant among the MAPK family members, it is of interest that the residue in between these two sites is not also highly conserved. It may be that this residue is at least partly responsible for conferring specificity for the various MEK isoforms.
Due to the dramatic 4-fold decrease in the phosphorylation level of ERK1*F200A it was decided to make an additional conservative mutation of Phe-200 to a tyrosine in order to assess any role of the aromatic side chain in this position. The ERK1*F200Y mutant also exhibited a 3-fold decrease in the level of phosphorylation. The role of Phe-200 is also unclear, but it may be involved in the conformational integrity of ERK1 enabling Thr-202 and Tyr-204 to be accessible for phosphorylation by MEK.
Activation Assays of Wild Type Kinase-active ERK1 Mutants by MEK2-All but three mutants had basal kinase activities for phosphorylation of myelin basic protein comparable with wild type kinase-active ERK1 (data not shown). Mutants T198A and F200Y exhibited increased basal kinase activities of 8-and 2-fold, respectively. Thus, the activation of T198A mutant may be underestimated because of the increase in basal kinase activity. This observation was of interest since a constitutively active form of ERK has not been reported but would be extremely useful to study the roles of the different ERK isoforms in signal transduction. Unfortunately, neither of these mutants had sufficiently high basal activity to be considered constitutively active. In addition, a double mutation of T198A and F200Y did not exhibit an increase in basal activity in vitro compared with either Thr-198 or Phe-200 mutation alone (data not shown). However, it would be of interest to determine whether any of the mutants that showed an increase in basal activity in vitro could function as a constitutively active mutant in the transformation of cultured cells. An additional mutant, W209A, had no detectable basal activity. Therefore, this residue appears to be crucial to either the structural or mechanistic integrity of ERK1, regardless of MEK activation.
As expected, mutation of either Thr-202 or Tyr-204 eliminated the activation of ERK1 by MEK2. The phosphorylation level of T202S was nearly equal to that of ERK1*; however, it was not an effective substrate for activation by MEK2. These results were consistent with the 6-fold higher level of Tyr-204 versus Thr-202 phosphorylation in wild type ERK1. That is, threonine phosphorylation represents only about 14% of the total ERK1 phosphorylation and will therefore have little effect on the overall phosphorylation level of the protein. However, as phosphorylation of both Thr-202 and Tyr-204 is necessary for activation, this lack of threonine phosphorylation results in a marked decrease in activation for the T202S mutant protein.
The comparable substitutions of Ser-218 and Ser-222 of MEK1 by threonine did not eliminate phosphorylation of MEK1 by Raf, indicating that MEK has a higher substrate specificity than Raf (30).
In general, mutants revealed significantly decreased activation by MEK2 in correlation with decreased phosphorylation in vitro (Fig. 4). This attenuated activation was also generally more severe than the phosphorylation decrease. This may be explained by the fact that ERK must adopt an altered conformation for MEK to bind, leading to subsequent phosphorylation of both tyrosine and threonine. Thus, it may be possible that even though phosphorylation was detected, indicating interaction with MEK, the ERK1 mutants may not be able to adopt some optimal conformation necessary for its full activity. However, some notable exceptions were observed as described below.
Mutation of Glycine 199 Enhances the Phosphorylation of Threonine-One mutant, G199A, actually showed an increased activation of nearly 2-fold. This is in contrast to the slight decrease in phosphorylation observed for the corresponding ERK1*G199A mutant. Phosphoamino acid analysis of ERK1*G199A revealed an increase in the level of threonine phosphorylation relative to tyrosine (1:3 versus 1:6 in wild type, Fig. 5A). This may explain the overall increase in activation since phosphorylation of both threonine and tyrosine on a single molecule is necessary for full enzymatic activity of ERK (21, The filters were washed with 180 mM phosphoric acid (5 times) and rinsed with 95% ethanol. Filters were quantitated by liquid scintillation counting. Each sample was performed in duplicate; error bars denote the standard deviation of two to four experiments. 45,46). In wild type ERK, the limiting factor for activation is threonine phosphorylation of the same molecule containing a phosphorylated tyrosine residue. Therefore, by increasing the level of threonine phosphorylation, a larger number of molecules will be phosphorylated on both threonine and tyrosine residues. That is, although the overall level of phosphorylation decreased, the larger proportion of diphosphorylated molecules results in an overall increase in ERK kinase activity.
Mutation of Threonine 207 or Arginine 208 Decreases the Phosphorylation of Threonine 202 but Increases Phosphorylation on Tyrosine 204 -One of the most perplexing results was obtained with the R208A mutant, which showed a 400% increase in the level of phosphorylation, yet the kinase activation was almost completely abolished (Figs. 3 and 4). Similar results were observed with the T207A mutant, which had only a slight decrease in total phosphorylation while less than 2% activation was observed. Phosphoamino acid analysis of these mutants was performed to determine the molecular basis for this lack of correlation between phosphorylation and activation. Surprisingly, mutant R208A was exclusively phosphorylated on tyrosine, while threonine phosphorylation was not observed (Fig.  5B). Similarly, mutant T207A retained only residual threonine phosphorylation, while the ratio of phosphothreonine to phosphotyrosine was increased to 1:50 versus 1:6 for ERK1* (Fig.  5B). These observations demonstrate that Thr-207 and Arg-208 play a critical role in the recognition of Thr-202 of ERK1 by MEK. The drastic effects on phosphorylation or basal kinase activity of distal residues Thr-207, Arg-208, and Trp-209 suggest an important function for this region in the mechanism of ERK activation.
Effect of ERK1 Mutations on Phosphorylation and Activation by MEK1-MEK1 and MEK2 are two closely related kinases that have been identified as ERK activators (21,45,46). The functional difference between MEK1 and MEK2 is unclear, although MEK2 shows higher basal activity toward both ERK1 and ERK2 (37). These differences were also detected in interactions with Raf. MEK1, but not MEK2, associated with c-Raf (47). Disruption of this MEK1-Raf complex reduced activation of MEK1 by Raf family members in both in vitro and in vivo studies (48). A unique proline-rich sequence has been identified and shown to be important for this MEK1 and Raf interaction (48). Differences in MEK activation by A-Raf were also detected (49). Immunoprecipitated A-Raf preferentially activated MEK1 versus MEK2 (49). Thus, a significant component of specificity of the ERK isoforms for various signaling pathways may be contributed by differences in their upstream activators, MEKs.
To determine whether mutations at the phosphorylation sites and flanking residues of ERK1 have the same effect on both MEK1 and MEK2, ERK1 mutants were assayed for phosphorylation and activation by MEK1 as described previously for MEK2. The phosphorylation results obtained were similar to those for MEK2 (Fig. 6); mutation of the flanking residues generally decreased the phosphorylation of ERK1* by MEK1. These data suggest that a common sequence motif in ERK is recognized by both MEK1 and MEK2; however, noticeable differences were observed. For example, the ERK1*R208A mutant showed an increase in phosphorylation by MEK2 of 4-fold as compared with 2-fold for MEK1. Activation of ERK1 mutants by MEK1 was also comparable with that by MEK2 (Fig. 7). Mutation of flanking residues dramatically reduced the activation potential of ERK1 by both MEK1 and MEK2. Among all the mutants, G199A was the only one to exhibit less than a 70% decrease in activation by MEK1.
Activation of MAPKs is a common event in many signal transduction pathways. At least six identified pathways in the budding yeast S. cerevisiae utilize different MAPK family members for their signal transduction (50). These pathways have very unique functions, which include the mating pheromone response, osmolarity regulation, cell wall construction, pseudohyphal development and spore formation in diploid strains, and invasiveness in haploid strains (50). The number of MAPKs and their upstream activators involved in such a variety of vital functions of the organism raises important questions on how the different isoforms of MAPK are regulated. The results of the mutation experiments presented in this paper provide additional information toward understanding the mechanisms of MAPK activation and the selectivity of MEK. Furthermore, these data will provide valuable insight toward the complete interpretation of the crystal structures of the tyrosine-phosphorylated and dual phosphorylated MAPK in terms of the mechanisms of MAPK dual phosphorylation. Understanding the mechanisms of activation for the different isoforms of MAPK and the specificity of the various activators will indeed lead to an improved general understanding of the mechanisms of regulation involved in the cell signaling pathways in mammalian systems. FIG. 7. Activation of ERK1 mutants by GST-MEK1. To activate ERK1, recombinant GST-MEK1 was incubated for 20 min at 30°C in a 30-l reaction volume containing 18 mM Hepes (pH 7.5), 20 M ATP, and 10 mM magnesium acetate. To initiate the kinase reaction myelin basic protein and [␥-32 P]ATP (5000 cpm/pmol) were added in the same reaction buffer and incubated for an additional 30 min at 30°C. Half of the reaction volume (20 l) was applied to P81 phosphocellulose filters. The filters were washed with 180 mM phosphoric acid (5 times) and rinsed with 95% ethanol. Filters were quantitated by liquid scintillation counting. Each sample was performed in duplicate; error bars denote the standard deviation of two to three experiments.