If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
* This work was supported in part by National Institutes of Health Grant GM 67969. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. § Supported by National Institutes of Health Training Grant GM 07752. ∥ Supported by the Summer Undergraduate Research Fellowship Program.
Protein kinases phosphorylate the appropriate protein substrate by recognizing residues both proximal and distal to the site of phosphorylation. Although these distal contacts may provide excellent binding affinities (low Km values) through stabilization of the enzyme-substrate complex, these contacts could reduce catalytic turnover (decrease kcat) through slow phosphoprotein release. To investigate how protein kinases can overcome this problem and maintain both high substrate affinities and high turnover rates, the phosphorylation of the yeast RNA transport protein Npl3 by its natural protein kinase, Sky1p, was evaluated. Sky1p bound and phosphorylated Npl3 with a Km that was 2 orders of magnitude lower than a short peptide mimic representing the phosphorylation site and only proximal determinants. Surprisingly, this extraordinary difference is not the result of high affinity Npl3 binding. Rather, Npl3 achieves a low Km through a rapid and favorable phosphoryl transfer step. This step serves as a chemical clamp that locks the protein substrate in the active site without unduly stabilizing the product phosphoprotein and slowing its release. The chemical clamping mechanism offers an efficient means whereby a protein kinase can simultaneously achieve both high turnover and good substrate binding properties.
The phosphorylation of hydroxyl-containing amino acids (i.e. serine, threonine, and tyrosine) in eukaryotic proteins controls numerous physiological responses essential for cell function. The enzymes that catalyze this modification, the protein kinases, have been studied for their critical role in complex signaling cascades and for their potential as valuable chemotherapeutic targets in proliferative diseases (
), less is known about the interactions of these enzymes with their physiological protein substrates. Protein kinases recognize and phosphorylate short peptide segments (consensus sequences) of 4–10 residues based on the known phosphorylation sequences of their targets (
). Although these short segments (consensus sequences) are minimal substrates, they do not provide all the binding energy offered by the full-length protein substrate. For instance, the catalytic efficiency of phosphorylating the protein substrate MAPKAP2 using p38 MAPK
). Studies of this type support a model in which regions beyond the limited consensus sequence interact with the protein kinase scaffold and govern high efficiency protein phosphorylation.
Although an x-ray structure for a protein kinase with a full-length protein substrate bound is currently unavailable, structural and functional studies are now beginning to reveal the nature of the interacting surfaces between the enzyme and substrate. Residues important for recognition on the substrate may, in fact, be very far from the site of phosphorylation. Several transcription factors (e.g. c-Jun, activating transcription factor-2, and myocyte enhancer factor-2A) regulated by MAPK phosphorylation interact with the enzyme using contacts as far as 100 residues away from the site of modification (
). Advances toward defining the regions of the kinase that interact with the protein substrates have also been made. Crystal structures of MAPK with docking peptides based on the distal interacting residues in myocyte enhancer factor-2A show that the large lobe of the kinase domain offers residues not in the active site that recognize the substrate (
). It has also been shown that the high affinity interaction between Csk and its substrate Src relies on a specific docking region in helix αD of the kinase domain, a secondary structural element outside the catalytic cleft (
). Although these examples highlight discrete surfaces within the kinase domain, other studies reveal that substrate-binding sites can lie on non-catalytic regulatory proteins. Cdk2 recognizes substrates containing an RXL motif that binds not to the kinase domain, but rather to a conserved region in the regulatory cyclin A3, ∼40 Å away from the active site (
). Although such studies have begun to unravel some of the mysteries regarding protein kinase substrate recognition, it is not clear how these distal surface patches impact catalysis in the active site.
Although the protein kinases have been well studied for the last half-century, only recently have powerful, fast mixing kinetic methods been applied to understand the mechanism of protein phosphorylation (
). However, these studies have been performed using short consensus sequence peptides that may not possess the conformational properties and full range of ground state contacts available in the full-length protein substrates. Although pre-steady-state kinetic investigations of protein phosphorylation have been reported in two cases, thus far, key features regarding the interaction of the protein target and enzyme have not been explored (
). Given the observation that protein substrates tend to have lower Km values than peptide mimics, it is possible that the current mechanism based on short peptides may differ with real physiological substrates. To investigate this possibility, we explored the kinetic mechanism of Sky1p, a yeast SR protein kinase using a physiological substrate, Npl3 (nuclear protein localization-3) (
). Although the Km for Npl3 is ∼2 orders of magnitude lower than that for the corresponding consensus sequence peptide, the rate-limiting step for protein phosphorylation is still ADP release. Surprisingly, Sky1p does not exploit added contacts in the enzyme-substrate complex for good affinity binding (Km < Kd), but rather recruits these interactions to support a fast and favorable phosphoryl transfer step. Using this strategy, the phosphoryl transfer step functions as a chemical clamp that provides exquisite selectivity, but avoids the dilemma of over-stabilizing the phosphoproduct and limiting the catalytic cycle through slow phosphoprotein release.
Materials—ATP, ADP, Mops, MgCl2, KCl, acetic acid, sucrose, DE52 resin, and liquid scintillant were obtained from Fisher. MANT-ADP was purchased from Molecular Probes, Inc. Poly-prep chromatography columns were obtained from Bio-Rad. The synthetic peptide YRTRDAPRERSPTR was obtained from the Peptide and Oligonucleotide Facility at the University of Southern California. [γ-32P]ATP was obtained from PerkinElmer Life Sciences.
Expression and Purification of Proteins—Sky1p lacking the N terminus and spacer insert (Sky1pΔNS) and Npl3 were expressed and purified using previously published protocols (
). Plasmid pET15b containing Npl3 with a Ser411 → Thr mutation (Thr-Npl3) was transformed into Escherichia coli BL21(DE3) RP cells (Invitrogen). Cells were grown in 2 liters of LB broth containing 200 μg/ml ampicillin to A600 = 0.8. Protein expression was induced with 0.2 mm isopropyl-β-d-thiogalactopyranoside for 12 h at 22 °C. Cells were resuspended in 100 ml of buffer containing 20 mm Tris-HCl (pH 7.5), 5 mm imidazole, 500 mm NaCl, 10% glycerol, 0.5 mm phenylmethylsulfonyl fluoride, and 0.5 ml of protease inhibitor mixture (Sigma) and lysed by sonication. The crude lysate was spun at 15,000 × g for 30 min. The soluble fraction was loaded onto a nickel-nitrilotriacetic acid column, washed with lysis buffer containing 40 mm imidazole and eluted with the same buffer containing 250 mm imidazole. The eluent was diluted to 100 mm NaCl and loaded onto a Q-Sepharose fast flow column. The column was washed with lysis buffer containing 20 mm Tris-HCl (pH 7.5), 100 mm NaCl, 1 mm dithiothreitol, 0.5 mm EDTA, and 10% glycerol. The protein was eluted with a 100–500 mm NaCl gradient. The fractions containing Thr-Npl3 were pooled, and the NaCl concentration was increased to 1 m. After concentration in a Centriprep 30 concentrator (Millipore Corp.), the protein was filtered through a 0.22-μm low bind syringe filter and loaded onto a Superdex 200 size exclusion column (Amersham Biosciences) equilibrated with 20 mm Tris-HCl (pH 7.5), 1 m NaCl, 1 mm dithiothreitol, and 10% glycerol. Pooled fractions were concentrated and flash-frozen. Protein concentrations were determined by the method of Gill and von Hippel (
Steady-state Kinetic Assays—Steady-state kinetic parameters for Sky1pΔNS were determined using Npl3 and Thr-Npl3 proteins as substrates in the presence of 50 mm Mops (pH 7.0), 4 mm free Mg2+, and 600–1000 cpm/pmol [γ-32P]ATP unless stated otherwise. Assays were typically executed by pre-equilibrating the enzyme, MgCl2, and ATP for 2 min and then initiating the reaction by the addition of protein substrate at 23 °C. Generation of the phosphoproduct was measured at various quench times using fixed concentrations of substrate. The reaction mixtures (20 μl) were quenched with 180 μl of 30% acetic acid. A portion of each reaction (180 μl) was then applied to DE52 columns (3 ml of resin) and washed with 5 ml of 30% acetic acid. The collected flow-through fraction containing phosphorylated substrate was then counted on the 32P channel in liquid scintillant. Control experiments were performed to determine the background phosphorylation (i.e. phosphorylation of Npl3 or the peptide substrate in the presence of quench). The specific activity of [γ-32P]ATP was determined by measuring the total counts of the reaction mixture. The time-dependent concentration of phosphorylated Npl3 and peptide substrate was then determined by considering the total counts/min of the flow-through fraction, the specific activity of the reaction mixture, and the background phosphorylation. For steady-state kinetic measurements, product formation is linear with time, and <5% of the substrate was converted to product.
Viscosity Studies—The steady-state phosphorylation of the peptide substrate and Npl3 was monitored using the column assay as described above in the absence and presence of 30% sucrose. The relative solvent viscosity (ηrel) of the buffer (50 mm Mops (pH 7)) containing sucrose was measured using an Ostwald viscometer and a previously published protocol (
). A ηrel of 2.4 was measured for the buffer containing 30% sucrose at 23 °C.
Inhibition Assays—Inhibition studies using Thr-Npl3 were carried out by pre-equilibrating the enzyme, MgCl2, and ATP for 2 min and then initiating the reaction by the addition of Npl3 (4 μm) in the presence of Thr-Npl3 (0–52.5 μm). Inhibition studies using phospho-Npl3 were carried out by generating various amounts of phospho-Npl3 by completely turning over known amounts of Npl3 in the presence of enzyme, MgCl2, and ATP in 50 mm Mops (pH 7.0). The degree of inhibition at various phospho-Npl3 concentrations (0–50 μm) was then measured by the addition of exogenous Npl3. Reactions were processed and analyzed as described above.
Rapid Quench Flow Experiments—The phosphorylation of Npl3 and the peptide substrate was monitored using a Model RGF-3 quench flow apparatus (KinTek Corp.) following a previously published procedure (
). Rapid quench flow experiments were typically executed by loading equal volumes of enzyme, buffer, and MgCl2 into one sample loop and substrate, [γ-32P]ATP (600–1000 cpm/pmol), and MgCl2 in 50 mm Mops (pH 7) into the other. The reactions were quenched using 30% acetic acid, and the phosphoproduct was separated from unreacted [32P]ATP using the column separation assay. Control experiments were performed to determine the background phosphorylation (i.e. phosphorylation of substrate in the presence of quench) using previously published protocols (
). The time-dependent concentration of phosphoproduct was then determined by considering the total counts/min of the flow-through fraction, the specific activity of the reaction mixture, and the background phosphorylation.
Stopped-flow Fluorescence Experiments—All transient state fluorescence studies were performed using an Applied Photophysics stopped-flow spectrometer. In the trapping experiments, Sky1pΔNS, MANT-ADP, and MgCl2 in 50 mm Mops (pH 7) were pre-equilibrated in one 2.5-ml syringe and then mixed at a 1:1 ratio with ADP and MgCl2 in 50 mm Mops (pH 7) in a second 2.5-ml syringe. The excitation wavelength was 290 nm, and emission was measured using a 410-nm cut-on filter fitted between the cell and the photomultiplier tube. For data analysis, an average of five to eight individual traces was recorded, and the data were fit to double exponential functions.
Data Analysis—The initial velocity versus substrate concentration data were fit to the Michaelis-Menten equation to obtain kcat and Km values for Npl3 and peptide substrate phosphorylation. The production of phospho-Npl3 in the rapid quench flow experiments was fit to Equation 1,
where [P] is the concentration of phospho-Npl3, α is the amplitude of the “burst” phase, kb is the burst phase rate constant, L is the linear rate, and t is time. The data generated from the catalytic trapping experiments were fit using the kinetic simulation program KINSIM (
Steady-state Kinetics—To investigate the kinetic properties of Sky1p, a truncated form of the enzyme lacking the N terminus and an insert sequence (Sky1pΔNS) was expressed. Sky1pΔNS can be expressed in large pure amounts possessing the same activity and the same Km for ATP as the full-length enzyme (
). The steady-state kinetic parameters for Sky1pΔNS were measured using either Npl3 or a peptide substrate based on the single phosphorylation site in the C terminus of Npl3 (YRTRDAPRERSPTR). Initial velocities were measured using fixed concentrations of [32P]ATP (2 mm) and varying concentrations of Npl3 (1–40 μm) or peptide (0.1–3mm) in the presence of 4 mm free Mg2+. The kcat, Km, and kcat/Km values for Npl3 phosphorylation are 0.6 ± 0.1 s—1, 3.5 ± 0.5 μm, and 0.17 ± 0.04 μm—1 s—1, respectively. The kcat and Km values for peptide phosphorylation are 1.0 ± 0.2 s—1 and 600 ± 100 μm, respectively. Thus, the kcat/Km for Npl3 is ∼100-fold larger than that for the short peptide.
), we showed that the phosphorylation of Npl3 by Sky1pΔNS leads to classic burst kinetics, indicating that the phosphoryl transfer step does not limit substrate turnover. To investigate the role of distal protein segments in controlling the phosphoryl transfer step, we employed rapid quench flow methods to measure the pre-steady-state kinetics of peptide phosphorylation. In this experiment, Sky1pΔNS (15 μm) was pre-equilibrated with [32P]ATP (2 mm) and then rapidly mixed with peptide substrate (3 mm) in the rapid quench flow instrument. As shown in Fig. 1, the production of the phosphopeptide proceeded in a linear manner, with no evidence of a burst phase in the early portion of the reaction. In these experiments, very high enzyme concentrations (15 μm) were used to maximize the chances of detecting a burst phase. No clear evidence for an appreciable lag in the generation of the phosphopeptide was detected in the early portion of the reaction. Accordingly, the data were fit to a linear function to obtain an initial velocity of 15 μm/s (v/[E] = 1 s—1), a value consistent with the steady-state kinetic parameters for this substrate. The concentrations of ATP and peptide were ∼3- and 5-fold above their Km values, so the absence of a burst amplitude was not due to low substrate saturation. In comparison, the phosphorylation of Npl3 (10 μm) by Sky1pΔNS (1 μm) was accompanied by a rapid large burst phase (kb = 33 ± 6 s—1 and α/[E] = 0.75 ± 0.10 μm), followed by a linear steadystate phase (L/[E] = 0.42 ± 0.04 s—1). These findings imply that the rate of phosphoryl transfer is rate-limiting for peptide phosphorylation, whereas this step is fast for Npl3.
Viscosity Effects on Substrate Phosphorylation
The pre-steady-state kinetic studies suggested that a slow phosphoryl transfer step limits peptide phosphorylation (Fig. 1). To provide further support for this conclusion, the effects of solvent viscosity on the steady-state kinetic parameters for peptide phosphorylation were measured. We have shown previously that kcat for a good substrate displays a large viscosity effect due to rate-limiting bimolecular product release (
). The initial velocity for the Sky1pΔNS reaction was measured as a function of peptide concentration in the presence and absence of 30% sucrose (Fig. 2). The added viscosogen had no effect on kcat, implying that the rate-limiting step in peptide turnover is a unimolecular step, viz. the phosphoryl transfer step. Furthermore, the absence of a viscosity effect implies that the viscous agent does not alter protein structure. As an additional control, the phosphorylation of Npl3 was also monitored at 0 and 30% sucrose (Fig. 2, inset). The observed decrease in kcat, measured as a ratio of the parameter in the absence and presence of sucrose (kcat(o)/kcat = 2.6), is similar in value to the change in solvent viscosity (ηrel = 2.4). These data imply that Npl3 turnover is limited by a diffusive step. Thus, the viscosity studies indicate that the absence of a burst phase for the peptide is due to rate-limiting phosphoryl transfer.
Rate-limiting Step for Npl3 Phosphorylation: Catalytic Trapping Studies
Although a large burst phase and a large viscosity effect on kcat for Npl3 (Figs. 1 and 2) indicate that the phosphoryl transfer step is fast, these results do not provide information as to which product release step (ADP or phospho-Npl3) limits turnover. Although ADP dissociation limits peptide phosphorylation in several protein kinases (
), it is unclear whether the presence of a real physiological protein substrate could change this rate-limiting step. To address this question, a catalytic trapping protocol originally developed in our laboratory for protein kinase A was employed (
). In this method, shown in Scheme 1, Sky1pΔNS is pre-equilibrated with ADP prior to reaction initiation with excess amounts of ATP in the rapid quench flow instrument. In the absence of ADP preequilibration, Npl3 was phosphorylated in a typical biphasic manner consistent with classic burst kinetics (Fig. 3). However, pre-equilibration of Sky1pΔNS with ADP abolished the burst phase without altering the magnitude of the steady-state linear phase. These findings indicate that ATP is an effective catalytic trapping agent. The kinetic transients in Fig. 3 were modeled using a numerical integration program (
) and the simple kinetic mechanism in Scheme 1. The kinetic trace in the absence of ADP is simulated to obtain a set of rate constants that satisfy the burst kinetics (see legend to Fig. 3). Using this fixed set of rate constants, the trace in the presence of ADP is then modeled to provide a visual fit to the data with only one variable, koff. The simulated data best approximate the experimental data when koff (0.5 s—1) is close in value to kcat (0.6 s—1). Thus, the catalytic trapping studies demonstrated that phospho-Npl3 is released quickly from the active site in the steadystate time frame and that the rate-limiting step in kcat is cleanly ADP release.
Dissociation Rate Constant for MANT-ADP: Stopped-flow Fluorescence Methods
To provide further evidence that ADP release limits the k4 step in Scheme 1, the dissociation rate constant for a fluorescent derivative of ADP, MANT-ADP, was measured using stopped-flow fluorescence spectroscopy. We showed previously that MANT-ADP binding can be monitored by direct fluorescence enhancements of the methylanthraniloyl group at 440 nm (
). Given these observations, we measured the dissociation rate constant for MANT-ADP as a surrogate for ADP release using a trapping experiment. Sky1pΔNS was saturated with MANT-ADP in one syringe (400 μm, 30 × Kd) and then mixed with a large excess of ADP in the second syringe (4 mm). The fluorescence data were well fit by a double exponential transient with rate constants of 14 and 0.7 s—1 (Fig. 4). Increasing the concentration of trapping ligand or MANT-ADP by 50% did not alter these rate fits, indicating that the ADP concentration was sufficiently high to ensure that the observed kinetics reflect true dissociation kinetics for MANT-ADP (data not shown). The presence of biphasic kinetics for MANT-ADP dissociation in trapping experiments may be interpreted by a two-step binding mechanism for this ligand as described in Scheme 2. Under conditions in which the two observed kinetic phases are separated by rate as in Fig. 4, the fast phase approximates k—1 (
). Furthermore, the slow phase approximates the net rate for product dissociation from E*·MANT-ADP (kreverse = 0.7 s—1). Although the MANT group could influence the kinetic mechanism of nucleotide binding, the slow phase rate for trapping is close to kcat for Npl3 phosphorylation (0.6 s—1), further supporting the notion that net ADP release limits the turnover of Npl3. Interestingly, if MANT-ADP is an adequate mimic for ADP, the data suggest that a slow conformational change could be associated with ADP release.
Effects of Npl3 Concentration on the Burst Phase Rate
Previously, we showed that the Kd for a peptide substrate to protein kinase A can be measured from the hyperbolic substrate dependence on the burst rate constant (
). To determine whether a Kd for Npl3 could be measured using this approach, we monitored the burst rate constant (kb) as a function of several Npl3 concentrations (1–19 μm). At all concentrations, a single exponential burst phase was observed, with no signs of sigmoidal behavior (i.e. a lag phase) at the lowest concentrations (data not shown). If a lag phase is present in the kinetic traces, then it is expected to be no greater than 5 ms in duration, setting a lower limit of 200 s—1 on the subsequent phosphoryl transfer step (
). However, since no strong evidence for this phase was present, we treated the initial rise in phospho-Npl3 as a simple first-order process and fit the data to Equation 1 to obtain kb. As shown in Fig. 5, over the specified Npl3 concentration range, kb varied in a linear manner (kb(app)), with a slope (Sburst) of 3 ± 0.2 μm—1 s—1. Overall, the data indicate that the phosphoryl transfer rate constant for Sky1p is a minimum of 60 s—1 and that the apparent association rate constant for Npl3 is 3 μm—1 s—1.
Enzyme-dependent Single Turnover Experiments
Due to the substrate expression levels, we were not able to measure burst rate constants above 20 μm Npl3 in Fig. 5. To circumvent this limitation, we took advantage of the higher expression of Sky1pΔNS to measure the Kd for Npl3 in single turnover experiments. These experiments were performed under conditions in which the enzyme concentration exceeds the substrate concentration so that no steady-state reaction component is present in the kinetic transients (
). Under these conditions, the dependence of the rate constant on total enzyme concentration provides the Kd for the substrate provided the reaction obeys simple first-order kinetics over a wide range of enzyme concentrations (
). Fig. 6 shows a typical single turnover experiment using 25 μm Sky1pΔNS and 2 μm Npl3. The complete turnover of Npl3 followed an exponential transient, with a rate constant of 67 ± 5 s—1 and no signs of sigmoidal behavior in the early part of the reaction. To obtain the Kd for Npl3 and the phosphoryl transfer rate constant, the single turnover rate constant was measured as a function of total enzyme concentration. As shown in Fig. 6 (inset), the exponential rate constant increased as a function of Sky1pΔNS concentration. Fitting the data to a hyperbolic function provided a Kd of 52 ± 12 μm for Npl3 and a phosphoryl transfer rate constant of 220 ± 30 s—1. Over the entire enzyme concentration range (12–80 μm), the data fit cleanly to a single exponential equation, implying that the Npl3 binding step is likely to be fast relative to the phosphoryl transfer step. Finally, the initial slope of the rate versus enzyme concentration plot in Fig. 6 provides an apparent association rate constant of 4 μm—1 s—1, a value similar to Sburst from the Npl3-dependent burst phase (Fig. 5).
Generating an Inactive Form of Npl3 with a Ser → Thr Substitution
To provide further support for the Npl3 Kd determination in Fig. 6, we designed an inactive form of the substrate as a surrogate for the protein substrate and measured its affinity for Sky1pΔNS. SR protein kinases have extraordinary substrate specificity requirements, preferring to phosphorylate serine, but not threonine, when flanked by an arginine (
). We took advantage of this requirement by generating a mutant form of Npl3 with the phosphorylatable serine at position 411 replaced with a threonine (Thr-Npl3). As shown in Fig. 7, Thr-Npl3 was not phosphorylated to any appreciable extent; so over the typical time frame for the Npl3 steady-state reaction (0.5 min), the mutant served as an inhibitor. Indeed, the initial velocity for Npl3 phosphorylation decreased as a function of Thr-Npl3 (Fig. 7, inset). Since we showed in separate experiments that Thr-Npl3 was competitive with the substrate (data not shown), the data in Fig. 7 (inset) could be used to obtain a KI of 50 ± 10 μm for Thr-Npl3 using a constant total amount of Npl3 (4 μm) in the assay. This KI value is similar to the Kd for Npl3 measured by different means (Fig. 6), and both are 15-fold larger than the Km for Npl3.
Binding Affinity of Phospho-Npl3
To determine whether phosphorylation increases the affinity of Npl3, the KI for phospho-Npl3 was measured. To perform these experiments, discrete amounts of phospho-Npl3 were generated using Sky1pΔNS (1 μm), excess [32P]ATP (2 mm), and limiting amounts of Npl3 (0–50 μm). In every case, all the initial Npl3 was converted to phospho-Npl3 based on radiolabel incorporation (data not shown). The initial velocity of the enzyme reaction was then measured by adding 3 μm Npl3 at a series of fixed phospho-Npl3 concentrations and constant ATP and enzyme concentrations. Since the amount of phospho-Npl3 generated at the highest substrate concentrations reflected only 2% consumption of the total ATP, the presence of ADP was not expected to affect the initial velocity determination. Background counts from the initial labeled phospho-Npl3 were subtracted from the assays. The initial velocity of the reaction declined as a function of the initial phospho-Npl3 concentration. The data were fit using Dixon plot analysis to obtain a KI of 45 ± 10 μm (data not shown). This value is similar to the KI for Thr-Npl3 and the Kd for Npl3, suggesting that phosphorylation does not enhance the binding affinity of the protein substrate.
Although available structural and functional studies suggest that protein kinases exploit extended binding surfaces for substrate recognition (
), little is known about the thermodynamics of the kinase-protein substrate interaction and how these extended surfaces in the enzyme and substrate facilitate the phosphoryl transfer reaction. Although the physiological substrates tend to have much lower Km values compared with the short peptide substrates (
), it is not clear whether these differences are due to higher affinities of the protein substrates. It is also not clear whether the phosphoryl transfer step plays a role in determining these differences in specificity. To begin to characterize the fundamental parameters associated with this essential protein-protein interaction, we investigated the phosphorylation of Npl3, the physiological protein target for the yeast kinase Sky1p. For these studies, we exploited the unique advantages of pre-steady-state kinetic methods for establishing catalytic mechanisms. In the past, we (
) have demonstrated that these methods are well suited for investigating important catalytic features of both simple and multidomain protein kinase systems.
Distal Residues Regulate the Phosphoryl Transfer Rate— Although Sky1p is capable of phosphorylating a short peptide based on a single phosphorylation site in Npl3, the modification of the full-length substrate is ∼100-fold more efficient compared with the peptide based on kcat/Km measurements. To begin to understand the source of this enhanced specificity, we measured the rate of phosphoryl transfer for both substrates using rapid quench flow techniques. For Npl3, multi-turnover pre-steady-state experiments were inconclusive since only a linear dependence between the burst rate constant and substrate concentration was measured (Fig. 5). However, a phosphoryl transfer rate constant of 200 s—1 was measured for Npl3 using single turnover experiments. This reflects a highly efficient catalytic step since this rate is close to that for protein kinase A, one of the most efficient kinases studied (
). In comparison with Npl3, the phosphoryl transfer rate for the peptide substrate is considerably lower at 1 s—1 (Fig. 1). Thus, residues outside the limited phosphorylation sequence in the C terminus have a profound positive impact on the delivery of the γ-phosphate of ATP to the serine hydroxyl, enhancing the rate for this transfer by ∼200-fold.
Npl3 Does Not Influence the Rate-limiting Step in Catalysis—Our laboratory first demonstrated that ADP release is rate-limiting for protein kinase A (
). These results, although informative, have been derived from short peptide substrates designed from consensus sequence data rather than from full-length physiological protein substrates. Whether large protein substrates modify the release rate of ADP remained a prominent question until now. Using catalytic trapping methods, we showed that ADP release is the rate-limiting step for Npl3 phosphorylation (Fig. 3), a conclusion further supported by stopped-flow trapping and viscosity experiments (Figs. 2 and 4). Thus, although Npl3 is rapidly phosphorylated and released from the active site, overall turnover awaits slow release of the product (ADP), the rate-limiting step in the catalytic cycle. The slightly higher turnover number for the peptide compared with Npl3 (1 versus 0.6 s—1) is consistent with negative binding synergism between the nucleotide and the peptide previously observed in steady-state kinetic investigations (
). In accord with this observation, no changes in Npl3 binding affinity upon phosphorylation were observed in the presence of the nucleotide.
Source of the Low Km for Npl3—The studies presented here indicate that the thermodynamic Kd for Npl3 is ∼15-fold higher than the steady-state Km. To understand the nature of the high Npl3 affinity, we considered the phosphorylation mechanism in Scheme 3 and related the substrate Km to the individual steps using Equation 2.
The complexity of this steady-state kinetic equation can be reduced substantially due to several constraints imposed by the kinetic data. A rapid and large burst phase (Figs. 1 and 3) implies that the phosphoryl transfer step is highly favorable and exceeds the rate of the product release step (i.e. k4 < k3 > k—3). Furthermore, since the association rate constant for Npl3 (Sburst) is ∼10-fold larger than the kcat/Km, the phosphoryl transfer step must be reversible in the time frame of substrate turnover (i.e. k—3 > k4).
Several conclusions about the relative values of Sburst and kcat/Km can be made depending on the rates of the steps in Scheme 3. For example, Sburst will be equivalent to the expression for kcat/Km (kcat/Km = Sburst = k3/Kd) if k4 > k—3 and k>2 — k3 (
), an outcome not observed in our studies (Sburst μ = 3 μm—1 s—1 and kcat/Km = 0.17 μm—1 s—1). If Npl3 is a sticky substrate and k—3 is small (k3 > k2 and k4 > k—3), kinetic simulations show that Sburst approximates k2 provided that k3 is very fast, and appreciable no “lag” phase is observed in the early portion of the pre-steady-state phase (simulations not shown). In this scenario, kcat/Km ≈ Sburst, another outcome that is not consistent with experimental observations. In contrast to these findings, when k—3 > k4, kcat/Km = k2/(1 + k—2/Kintk4) and thus can be lower than k2 when k—2 > Kintk4. In simulation studies, this mechanism permits a large value for Sburst relative to kcat/Km (simulations not shown).
With these conditions, the Km expression now reduces to that shown in Equation 3,
where Kint is the internal equilibrium constant for protein phosphorylation (Kint = k3/k—3). If we assume that Sburst sets a lower limit on k2 and that k4 is limited by the release of ADP, k4/k2 is expected to be >20-fold lower than Km (0.6 s—1/3 μm—1 s—1 ≥ k4/k2), and Equation 3 then reduces to Kd/Kint. Since both Km and Kd are known, we can calculate a value of 15 for the internal equilibrium constant. Thus, we conclude that the high apparent affinity of Npl3 (Km) is due to a fast and thermodynamically favorable phosphoryl transfer step.
Contribution of Distal Sequences to Npl3 Binding Affinity— The measurement of the Npl3 Kd now permits an evaluation of the binding energy offered by distal sequences in the protein substrate. Since kcat for peptide phosphorylation is limited by a slow phosphoryl transfer step (Figs. 1 and 2), its Km is equivalent to its Kd (see Equation 2). Accordingly, the peptide Kd is higher than the Npl3 Kd by 12-fold (600 versus 50 μm). These data suggest that residues outside the consensus sequence contribute ∼1.5 kcal/mol stabilization energy to the enzymesubstrate complex. At this time, it is not known which residues in Npl3 provide this added stability and which residues in Sky1p serve as the docking site for the protein substrate. Nonetheless, although the outlying residues clearly contribute, they do not account for the apparent binding energy of 3 kcal/mol determined from the Km values for Npl3 and the peptide (600 versus 3.5 μm).
Chemical Clamping and Efficient Substrate Phosphorylation—Through detailed kinetic studies, we have shown that the high catalytic efficiency of the protein compared with the peptide substrate does not rest in a tightly bound Npl3 molecule, but rather is largely due to a highly favorable, fast phosphoryl transfer step (see Equation 3). We call this process the chemical clamping step since it effectively locks Npl3 in the active site through a chemical transfer event and generates a low substrate Km without recruiting high affinity interactions. The mechanism depicted in Fig. 8 illustrates how the total apparent binding energy for Npl3 under catalytic conditions (ΔGm) arises from two principal factors: the stability of the enzyme-substrate complex (ΔGd) and the internal equilibrium constant for the phosphoryl transfer step (ΔGint). Using protein-protein interactions in this dual manner provides some important catalytic advantages. Although ΔGcat is controlled by the rate of ADP release, extensive stabilization of the protein-enzyme complex could lead to large increases in ΔGcat due to slow release of phospho-Npl3. We have shown that Npl3 and phospho-Npl3 bind with similar affinities, so any additional stabilization of the protein substrate would be translated to the product. However, Sky1p avoids this problem and maintains high turnover by minimizing residue contacts that would lower the ground state energy of the enzyme-product complex. In this manner, the enzyme can effectively recognize Npl3 and, upon phosphorylation, quickly release the phosphoproduct. It will be interesting to see whether other protein kinases utilize the chemical clamping mechanism to modify substrate affinity through different values for Kint.