Mechanistic Studies of the Autoactivation of PAK2

Protein kinase activation, via autophosphorylation of the activation loop, is a common regulatory mechanism in phosphorylation-dependent signaling cascades. Despite the prevalence of this reaction and its importance in biological regulation, the molecular mechanisms of autophosphorylation are poorly understood. In this study, we developed a kinetic approach to distinguish quantitatively between cis- and trans-pathways in an autocatalytic reaction. Using this method, we have undertaken a detailed kinetic analysis for the autoactivation mechanism of p21-activated protein kinase 2 (PAK2). PAK2 is regulated in vivo and in vitro by small GTP-binding proteins, Cdc42 and Rac. Full activation of PAK2 requires autophosphorylation of the conserved threonine, Thr402, in the activation loop of its catalytic kinase domain. Analyses of the time courses of substrate reaction during PAK2 autoactivation suggest that autophosphorylation of Thr402 in PAK2 obeys a two-step mechanism of cis initiation, followed by trans amplification. The unphosphorylated PAK2 undergoes an intramolecular (cis) autophosphorylation on Thr402 to produce phosphorylated PAK2, and this newly formed active PAK2 then phosphorylates other PAK2 molecules at Thr402 in an intermolecular (trans) manner. Based on the kinetic equation derived, all microscopic kinetic constants for the cis and trans autophosphorylation have been estimated quantitatively. The advantage of the new method is not only its usefulness in the study of fast activation reactions, but its convenience in the study of substrate effects on modification reaction. It would be particularly useful when the regulatory mechanism of the autophosphorylation reaction toward certain enzymes is being assessed.

Protein kinase activation, via autophosphorylation of the activation loop, is a common regulatory mechanism in phosphorylation-dependent signaling cascades. Despite the prevalence of this reaction and its importance in biological regulation, the molecular mechanisms of autophosphorylation are poorly understood. In this study, we developed a kinetic approach to distinguish quantitatively between cis-and transpathways in an autocatalytic reaction. Using this method, we have undertaken a detailed kinetic analysis for the autoactivation mechanism of p21-activated protein kinase 2 (PAK2). PAK2 is regulated in vivo and in vitro by small GTP-binding proteins, Cdc42 and Rac. Full activation of PAK2 requires autophosphorylation of the conserved threonine, Thr 402 , in the activation loop of its catalytic kinase domain. Analyses of the time courses of substrate reaction during PAK2 autoactivation suggest that autophosphorylation of Thr 402 in PAK2 obeys a two-step mechanism of cis initiation, followed by trans amplification. The unphosphorylated PAK2 undergoes an intramolecular (cis) autophosphorylation on Thr 402 to produce phosphorylated PAK2, and this newly formed active PAK2 then phosphorylates other PAK2 molecules at Thr 402 in an intermolecular (trans) manner. Based on the kinetic equation derived, all microscopic kinetic constants for the cis and trans autophosphorylation have been estimated quantitatively. The advantage of the new method is not only its usefulness in the study of fast activation reactions, but its convenience in the study of substrate effects on modification reaction. It would be particularly useful when the regulatory mechanism of the autophosphorylation reaction toward certain enzymes is being assessed.
Protein phosphorylation is one of the most important processes for cellular regulation and signal transduction in eukaryotic cells. The enzymes responsible for catalyzing this reaction are protein kinases, which catalyze phosphate transfer reactions from ATP to serine, threonine, or tyrosine residues in target substrate and provide key mechanisms for control of cellular signaling processes (1). The activities of many protein kinases are themselves regulated by phosphorylation (2). Most protein kinases are activated through phosphorylation of amino acid residue(s) within the activation loop. This can be carried out by an upstream kinase as part of a signaling cascade, or a kinase can autophosphorylate its own activation loop by either an intramolecular (cis) or intermolecular (trans) mechanism (3). Intramolecular autophosphorylation is a first-order reaction and can be described by the single exponential kinetic equation. In contrast, intermolecular autophosphorylation involves in a bimolecular autocatalytic event, and hence there are more variations in overall mechanisms. One important question for trans-autophosphorylation to address is what molecular mechanism might be involved in the initiation of autocatalytic reaction. Uncovering the answers would provide a better understanding of the regulation of a protein kinase and some important clues for development of specific kinase inhibitors that selectively block autoactivation of protein kinases (4,5).
There are three possible mechanisms through which an unphosphorylated kinase can be activated by trans-autophosphorylation of a critical residue in the activation loop. The first possibility is that the unphosphorylated kinase is completely inactive, and its activation is catalyzed by a small quantity of contamination, the phosphorylated active kinase in Reaction 1, where E and E* represent unphosphorylated and phosphorylated enzyme, respectively. In this case, phosphorylation by an upstream kinase is required to initiate kinase autoactivation. A second possible mechanism is one in which unphosphorylated kinase itself has inherent catalytic activity that slowly generates the phosphorylated kinase molecule by an intramolecular reaction, and once the activated kinase is formed, the activation is accelerated by the intermolecular reaction shown in Reaction Scheme 1. The third limiting mechanism is an inter-initiation model in which unphosphorylated kinase molecule, which has residual activity catalyzes the initial phosphorylation in an intermolecular manner shown in Reaction Scheme 2.
A considerable ingenious procedure is therefore required to distinguish between these mechanisms on the basis of various lines of experimental evidence. The p21-activated protein kinase 2 (PAK2) 2 is a serine/threonine kinase regulated by the small GTPases, Cdc42 and Rac, and is involved in the regulation of various cytoskeletal functions, including cell motility and membrane blebbing during apoptosis (6 -9). PAK2 consists of a C-terminal kinase domain and an Nterminal regulatory domain containing a GTPase binding domain (GBD) and an autoinhibitory domain (AID). PAK2 is activated by the binding of active Cdc42 and Rac to the GBD, resulting in a conformational change and a subsequent autophosphorylation on several serine and/or threonine residues. Upon activation, PAK2 from human is autophosphorylated at seven sites, six of which have been identified as serine residues in the N-terminal regulatory domain (10,11). The only phosphothreonine, pThr 402 , is in the C-terminal kinase domain, and the PAK2 activation was showed to coincide with the phosphorylation of this threonine residue (12). Although autophosphorylation of some N-terminal regulatory sites occurs in cis, phosphorylation of the activation loop threonine (Thr 402 ) is an inter-molecular process (13). The phosphorylation of the activation loop is clearly a critical step in the activation of the PAK kinases, although how the first molecule of active PAK2 is formed remains obscure.
In this study, we developed a new kinetic approach to distinguish effectively between cis-and trans-pathways in an autocatalytic reaction. The mechanism of the autophosphorylation reaction of PAK2 has been further examined using this approach. We demonstrated that the unphosphorylated PAK2 has an activity inherent in itself by which it can selfactivate, and that autophosphorylation of Thr 402 obeys a twostep mechanism of cis initiation, followed by trans amplification. That is, the unphosphorylated PAK2 undergoes an intramolecular autophosphorylation on Thr 402 to produce the first fully activated PAK2, and this newly formed active PAK2 then phosphorylates other PAK2 molecules at Thr 402 in an intermolecular manner. On the basis of the kinetic equation of the substrate reaction, all microscopic kinetic constants for the cis and trans autophosphorylation at Thr 402 of PAK2 have been estimated quantitatively.

EXPERIMENTAL PROCEDURES
Materials-Adenosine 5Ј-triphosphate (ATP), phospho(enol)pyruvate (PEP), nicotinamide adenine dinucleotide, reduced (NADH), lactate dehydrogenase (LDH), and pyruvate kinase (PK) were purchased from Sigma. Fetal bovine serum, Grace's insect cell culture medium and yeastolate solution were purchased from Invitrogen, MOPS and Tris were purchased from Sigma. Phosphospecific antibodies that react with PAK2 when phosphorylation at Thr 402 , Ser 141 , and Ser 192/197 , respectively, were obtained from Cell Signaling Technology. MLCtide (KKRPQRATSNVFA) was synthesized using standard protocol, purified by reverse-phase preparative HPLC chromatography, and characterized by MALDI-TOF mass spectrometry by Scilight Biotechnology (Beijing) Inc. Other reagents were local products of analytical grade used without further purification. Double-deionized water was used throughout.
Expression and Purification of Proteins-GST-fusion human PAK2 was expressed in Sf9 cells and purified as described previously (13). A constitutively active Cdc42, Histagged Cdc42L61 was expressed in Escherichia coli and purified by Ni-NTA column (Qiagen), followed by an anion exchange Source 15Q HR 10/10 column (GE Healthcare). The protein purity was over 95% as judged by SDS-PAGE. Protein concentrations were determined spectrophotometrically using theoretical molar extinction coefficients at 280 nm, following the method of Gill and von Hippel (14). The purified proteins were made to 20% glycerol and stored at Ϫ80°C.
Enzyme Assays for PAK2-Enzyme activity of PAK2 was determined with the MLCtide as a substrate using a coupled spectrophotometric assay (15,16). The standard assay for PAK2 was carried out at 25°C in a 1.6-ml reaction mixture containing 100 mM MOPS buffer, pH 7.4, 100 mM KCl, 20 mM MgCl 2 , 1 mM ATP, 200 M NADH, 1 mM phospho(enol)pyruvate, 20 units/ml lactate dehydrogenase, and 15 units/ml pyruvate kinase, and different concentrations of substrates . Reactions were initiated by the addition of PAK2 to the reaction mixture. Progress of the reaction was monitored continuously by following the formation of NAD ϩ at 340 nm, on a PerkinElmer Lambda 45 spectrophotometer. The concentrations of ADP formed in PAK2-catalyzed reaction were determined using an extinction coefficient for NADH of 6220 cm Ϫ1 M Ϫ1 at 340 nm. The concentrations of peptide were determined by turnover with the activated-PAK2 under conditions of limiting peptides.
Western Blotting-The phosphorylation states of PAK2 were assessed by Western blot analysis using specific antibodies. Samples were resolved by SDS-PAGE and electrotransferred onto polyvinylidene fluoride membranes using a Bio-Rad TransBlot system. After incubation of the membranes with anti-phospho-Ser 141 , Ser 192/197 , and Thr 402 antibodies (Cell Signaling Technology, Inc., 2606, 2605, 2601), specific immunocomplexes were detected by chemiluminescence using ECL reagents(Vigorous Biotechnology Beijing Co., Ltd.). The membrane was finally exposed to x-ray film (KodakInc.).

RESULTS
Effect of Cdc42 on the PAK2-catalyzed Reaction-Recombinant human PAK2 expressed in insect cell was purified as an inactive enzyme. To examine the phosphorylation state of the enzyme preparation used in this study, samples were subjected to Western blot analysis using the phosphospecific antibodies against phospho-Thr 402 , Ser 192/197 , and Ser 141 of PAK2, respectively (Fig. 1). In the absence of Cdc42L61, no detectable phosphorylation of Thr 402 and Ser 192/197 were detected, whereas Ser 141 exhibited partial phosphorylation. Thus, the first mechanism can be readily excluded since the purified PAK2 contained no phospho-Thr 402 contamination. The full-length PAK2 (final concentration of 3 M) was incubated with 15 M Cdc42L61 in a reaction mixture containing 100 mM MOPS (pH 7.4), 20 mM MgCl 2 , 1 mM DTT, and 2 mM ATP at 25°C, maximal phosphorylation at Thr 402 , Ser 141 , and Ser 192/197 was achieved within 30 min of Cdc42 stimulation, at which time the PAK2 activation was near 100% complete.
To elucidate the mechanism of the autoactivation of fulllength PAK2, we first performed kinase assay to measure the rate of PAK2-catalyzed reaction. Enzyme activity of PAK2 was determined using a continuous spectrophotometric assay that couples production of ADP to oxidation of NADH, measured as a decrease in absorbance at 340 nm. A synthetic peptide derived from myosin light chain, MLCtide (residues 11-23), was used as an exogenous substrate (17). As shown in Fig. 2A, the activated PAK2 is highly efficient toward MLCtide substrate, even in the absence of Cdc42 (curve 1). In contrast, when the unphosphorylated PAK2 was incubated with MgATP and MLCtide, no significant kinase activity (less than 1%) was detected (curve 2 in Fig. 2A). The addition of the activated PAK2 to the reaction mixture resulted in a linear increase of the absorbance at 340 nm (curve 3 in Fig. 2A) with a slope identical to that of the control experiment in the absence of unphosphorylated PAK2 (curve 1 in Fig. 2A). If the unphosphorylated, inactive PAK2 molecule alone can be trans-phosphorylated by other activated PAK2, then the total active PAK2 concentration would be increased during the assay reaction, and the rates of PAK2-catalyzed MLCtide phosphorylation should be significantly greater than that ob-served in the control experiment. Therefore, our results demonstrated that the unphosphorylated PAK2 alone is not a substrate of the activated PAK2 in the absence of Cdc42. We next investigated the effect of Cdc42 on the activated PAK2 activity. The initial rates of the activated PAK2-catalyzed MLCtide phosphorylation in the presence of various concentrations of Cdc42 were measured (Fig. 2B). It can be seen from this figure, the progressive addition of Cdc42 had no effect on the initial velocities of PAK2-catalyzed reactions, suggesting that once PAK2 is activated, the continued binding of Cdc42 is not necessary for PAK2 kinase activity, and even following dissociation of the GTPase, the kinase would remain in an "open" state allowing the catalyzed-reaction to occur.
Theoretical Analysis-The general mechanism for PAK2 autoactivation can be written as Scheme 3 (see supplemental information), where E and E* represent unphosphorylated and phosphorylated enzyme, L represent activator, and S and P represent exogenous substrate and its corresponding product, respectively. According to the results given above, it is reasonable to assume that the binding of Cdc42 to the activated PAK2 does not affect its kinase activity for both autophosphorylation of PAK2 and phosphorylation of exogenous substrate MLCtide (that is, K S ϭ KЈ S , k 2 ϭ kЈ 2 , and k ϩ0 ϭ kЉ ϩ0 , and kЈ ϩ0 ϭ kٞ ϩ0 ). Therefore, the kinetic equation for   where v s ϭ k 2 [T] 0 [S]/(K S ϩ [S]) is the steady state velocity of the substrate reaction catalyzed by PAK2, and the expressions of the observed rate constant k obs and ␤ are given by Equation 2 for the intra-initiation model, or Equation 3 for the inter-initiation model.
[T] 0 and [L] refer to the total concentrations of PAK2 and Cdc42, and A is the apparent second-order rate constant in Equation 4, where k ϩ0 and kЈ ϩ0 are microscopic rate constants for the free enzyme and enzyme-substrate complex, respectively. A comparison of the k obs expressions for the second and third mechanisms suggests that a plot of k obs as a function of [T] 0 should yield a straight line with a slope and a positive y intercept in the intra-initiation case, but a straight line with a y intercept of zero in the inter-initiation case. Thus, these two mechanisms can be distinguished by determining the effects of varying enzyme concentration on the value of k obs at fixed concentrations of substrate and Cdc42. Similar to irreversible inhibitor (18), covalent modification can also be classified as competitive, noncompetitive, and uncompetitive on the basis of the interaction mode of the exogenous (S) and isogenous (EL) substrates with the enzyme. Experimentally, the type of modification can be ascertained by studying the effect of the substrate concentration [S] on the apparent second-order rate constant A. Competitive: Therefore, the three types of substrate competition in covalent modification can be distinguished by suitable plots of A against [S]. For the competitive binding, a plot of 1/A against [S] will give a straight line. For an uncompetitive binding, the plot of 1/A against 1/[S] will be a straight line, whereas A will be independent of [S] for a noncompetitive binding.
Kinetics of PAK2 Autoactivation-We applied the kinetic theory of the substrate reaction during modification of enzyme activity (18,19) to the study of the autoactivation process of PAK2. Fig. 3 shows the time courses of product formation at different PAK2 concentrations. The progress curves start off the lag phases, and go up with increasing time due to the activation of PAK2. To distinguish between the second and third mechanisms, the progress curves were analyzed with Equation 1, and the values of k obs , v s , and ␤ were determined by nonlinear curve-fitting analysis. As shown in the inset of Fig. 3, a plot of k obs against enzyme concentration gives a straight line with a positive intercept at the ordinate. The nonzero intercept indicates that the autoactivation proceeds through the second mechanism: an initial intramolecular step followed by an intermolecular activation event.
To determine the kinetic parameters for the Cdc42-mediated PAK2 autoactivation, two sets of experiments were performed. First, the time courses of product formation at different Cdc42 concentrations, but fixed MLCtide and PAK2 concentrations, were monitored (Fig. 4). From the progress curves of product formation, k obs , v s , and ␤ were determined by fitting the experimental data to Equation 1 as described before. The inset of Fig. 4 shows that the observed rate constant k obs gives rise to the hyperbolic dependence on the Cdc42 concentration, and the data fit to the expression of k obs with a K L of 6.3 Ϯ 0.9 M and an upper limit of 0.066 Ϯ 0.004 SCHEME 3 s Ϫ1 . The dissociation constant for the binding of Cdc42 to the full-length PAK2 (K L ϭ 6.3 M) is 2 orders of magnitude higher than those between Cdc42 and the isolated GBD fragments (20 -22). A possible explanation for this discrepancy might be that there are contacts between the kinase domain and its regulatory domain in full-length PAK2, which modify the interaction of Cdc42 with the regulatory domain (23,24).
To further assess the effect of exogenous substrates on the autophosphorylation of PAK2, we next carried out the autoactivation assays in the presence of varied concentrations of MLCtide, but at a fixed concentrations of Cdc42 and PAK2 (Fig. 5). For a two-step mechanism of cis initiation followed by trans amplification, Equation 1 can be rewritten as Equation    A (Equation 4). It can be seen that the parameter kЈ ϩ0 , as inferred from these data, had a very large standard deviation. The large magnitude of the standard deviation in the estimate of kЈ ϩ0 could signify that the Equation 4 does not contain this parameter. When kЈ ϩ0 ϭ 0, Equation 4 can be simplified to Equation 5. In this case, a plot of 1/A against [S] will give a straight line, as indeed the case for the autoactivation of PAK2 (Fig. 6C). Re-fitting the experimental data with Equation 5 gives a remarkable correspondence, and the continuous lines in Fig. 6B represent the best fit with k ϩ0 ϭ 14.4 Ϯ 0.3 M Ϫ1 s Ϫ1 and a fixed K S ϭ 292 M. This result indicates that transautophosphorylation of PAK2 is competitively inhibited by binding of the exogenous substrate, MLCtide, at the active site. A plot of k e against MLCtide concentration shows the value of k e is not affected by the exogenous substrate (Fig.  6D), consistent with our model in which MLCtide cannot bind to EL to form ELS ternary complex. An average of these experiments yielded k e ϭ 0.0254 Ϯ 0.0007 s Ϫ1 . All kinetic parameters for the autoactivation of the intact PAK2 are summarized in Table 1. According to our definition, the parameter k ϩ0 determined from the autoactivation kinetics is equal to k cat /K m for the isogenous substrate (unphosphorylated PAK2 in complex with Cdc42). The second order rate constant k cat /K m is a physiologically relevant parameter for the reaction of free enzyme with free substrate and reflects both binding affinity and catalytic efficiency. As shown in Table 1, the k ϩ0 value for trans-autophosphorylation of PAK2 is more than 100-fold higher than k 2 /K S for the exogenous substrate, MLCtide. Therefore, the PAK2-Cdc42 complex is a more efficient substrate for the active enzyme, suggesting that there are additional contacts outside the immediate vicinity of the phos-phorylation site that mediate the recognition of one PAK2 molecule by another.

DISCUSSION
The autophosphorylation of the activation loop is a critical step in the activation of PAK2 kinase, but the molecular details of this process remain poorly understood. To further this understanding, we have undertaken a detailed kinetic study of PAK2 activation mechanism. Analyses of the time courses of substrate reaction during PAK2 autoactivation suggest that autophosphorylation of Thr 402 in PAK2 obeys a two-step mechanism of cis initiation, followed by trans amplification, and binding of the exogenous substrate MLCtide at the active site of PAK2 completely abolish the PAK2 autophosphorylation at site Thr 402 . Compared with the conventional method of taking aliquots at time intervals during the activation process and assaying for the enzyme activity or modification extend, the advantage of the present method is not only its continuous nature and usefulness in the study of fast modification reactions, but its convenience in the study of substrate effects on modification reaction. This new method would be particularly useful when the regulatory mechanism of the reversible phosphorylation reaction toward certain enzymes is being assessed.
Ligand-induced autophosphorylation of protein kinase is considered to be a trigger of intracellular signal transduction (2,3). Autophosphorylation can be achieved via two basic routes: intramolecular (in cis) and intermolecular (in trans) reactions. There is an abundance of structural data for protein kinases, which have provided insight into how the kinase transforms from an inactive state into an active state by transautophosphorylation. However, how a protein kinase autophosphorylates its activation loop in cis remains obscure. In the inactive state, the nonphosphorylated activation loop often folds back into the active site and sterically blocks substrate binding. This raises the question of how protein kinases that become self-activated by autophosphorylation of their activation loops in cis overcoming these steric constraints (3). Most of the kinases that autophosphorylate their activation loops in cis require protein binding partners, which temporally regulates the conformation of protein kinases and promote their activation (25)(26)(27)(28). For example, autophosphorylation of the critical activation loop tyrosine of DYRK is intramolecular and mediated by the nascent kinase passing through a transitory intermediate form (27). In contrast, GSK-3␤ is not activated during protein translation but re-

TABLE 1 Kinetic parameters for autoactivation of PAK2
Kinetic parameters for autoactivation of PAK2 were determined by the progress curve method and conventional method at 25°C and pH 7.4 in 100 mM MOPS buffer. The standard deviation for individual parameters was obtained from the nonlinear regression fit of the data to the corresponding equations using the software Sigmaplot 9. quires the binding and activity of the molecular chaperone, heat shock protein 90 kDa (Hsp90), which mediates a cis autophosphorylation event in the activation loop of GSK-3␤ (at Tyr 216 ) in response to growth factors (25). Therefore, we propose that binding of Cdc42 to the N-terminal regulatory domain of PAK2 does more than just relieve inhibitory constraints imposed on the catalytic domain by the inhibitory switch, and may also provide an essential chaperone-like function that transiently coverts the intermediate form of PAK2 into an intramolecular kinase capable of autophosphorylating the activation loop threonine in the protein kinase domain.