Mechanism-based inhibition of enzyme I of the Escherichia coli phosphotransferase system. Cysteine 502 is an essential residue.

Four phosphoenolpyruvate (PEP) derivatives, carrying reactive or activable chemical functions in each of the three chemical regions of PEP, were assayed as alternative substrates of enzyme I (EI) of the Escherichia coli PEP:glucose phosphotransferase system. The Z- and E-isomers of 3-chlorophosphoenolpyruvate (3-Cl-PEP) were substrates, presenting K(m) values of 0.08 and 0.12 mm, respectively, very similar to the K(m) of 0.14 mm measured for PEP, and k(cat) of 40 and 4 min(-1), compared with 2,200 min(-1), for PEP. The low catalytic efficiency of these substrates permits the study of activity at in vivo EI concentrations. Z-Cl-PEP was a competitive inhibitor of PEP with a K(I) of 0.4 mm. E-Cl-PEP was not an inhibitor. Compounds 3 and 4, obtained by modification of the carboxylic and phosphate groups of PEP, were neither substrates nor inhibitors of EI, highlighting the importance of these functionalities for recognition by EI. Z-Cl-PEP is a suicide inhibitor. About 10-50 turnovers sufficed to inactivate EI completely. Such a property can be exploited to reveal and quantitate phosphoryl transfer from EI to other proteins at in vivo concentrations. Inactivation was saturatable in Z-Cl-PEP, with an apparent K(m)(inact) of 0.2-0.4 mm. The rate of inactivation increased with the concentration of EI, indicating a preferential or exclusive reaction with the dimeric form of EI. E-Cl-PEP inactivates EI much more slowly, and unlike PEP, it did not protect against inactivation by Z-Cl-PEP. This and the ineffectiveness of E-Cl-PEP as a competitive inhibitor have been related to the presence of two EI active species. Cys-502 of EI was identified by mass spectrometry as the reacting residue. The C502A EI mutant showed less than 0.06% wild-type activity. Sequence alignments and comparisons of x-ray structures of different PEP-utilizing enzymes indicate that Cys-502 might serve as a proton donor during catalysis.

The bacterial phosphoenolpyruvate:glycose phosphotransferase system (PTS) 1 catalyzes the uptake and phosphorylation of carbohydrates (1). It is further involved in signal transduction (2), e.g. catabolite repression (3), chemotaxis (4), and allosteric regulation of metabolic enzymes and transporters in response to the availability of carbohydrates (5,6). The PTS is widely distributed in eubacteria and absent from eukaryotes. It generally consists of two central phosphoryl carriers, EI and HPr, and depending on the species, between 1 and 20 peripheral phosphotransferases, the integral membrane components that recognize and transport hexoses, hexitols, and disaccharides. EI and HPr are the proteins at the top of this divergent phosphorylation cascade (7).
EI is a 64-kDa two-domain protein (8,9). In the aminoterminal domain (EIN, residues 1-250) is located His-189, the amino acid transiently phosphorylated by PEP. The threedimensional structure of EIN has been elucidated (10,11) and its mode of interaction with HPr characterized by NMR spectroscopy (12,13). It is composed of a HPr binding ␣-helical subdomain and an ␣/␤ subdomain, structurally similar to the phosphohistidine swivel domain of pyruvate phosphate dikinase (PPDK). Phospho-EIN transfers the phosphoryl group to HPr, but it requires the presence of the carboxyl-terminal domain (EIC) to become phosphorylated by PEP (8,14). EIC contains the PEP binding site (15) and plays a crucial role in dimerization (16,17). It also confers species and subunit specificity during the phosphoryl transfer to the next phosphocarrier protein (14). EIC is proteolytically unstable and flexible. Its structure is not yet known. However, the amino acid sequence similarity between EIC and the PEP binding domain of PPDK of Clostridium symbiosum points toward similar fold and structure. There is also sequence similarity with PEP synthase (18).
The mode of action of EI and the way its activity is controlled are not yet fully understood. The protein dimerizes in a temperature-dependent manner (19). The dissociation constant has been reported to shift from 20 M at 6°C to 0.9 M at 30°C. The presence of divalent cations (Mg 2ϩ or Mn 2ϩ ) and PEP also affects the equilibrium and the association and dissociation rate constants (20,21). The association rate constant is 2-3 orders of magnitude slower than in other dimeric proteins, indicating that dimerization is accompanied by major conformational rearrangements of the interacting EIC domains (20,22). Several studies suggest that the dimer is the catalytically active form: (i) in vitro PTS activity increases during a long lag time, which was explained by the slow dimerization of EI to form the active species (23); (ii) PTS activity at rate-limiting concentrations of wild-type EI was enhanced upon addition of the inactive H189G and G338E EI mutants (24); (iii) activity and fluorescence of pyrene-labeled EI changed in parallel as a function of temperature, which was interpreted as a correlation between function and dimerization (20,21).
Controversial issues concerning EI catalysis are: (i) the number of sites per dimer which are phosphorylated at a given time, one (25,26) or two (27); (ii) the kinetic mechanism of EI, simple Bi Bi ping-pong mechanism (28,29) or a flip-flop mechanism (26,30); and (iii) the dynamics of monomer-dimer formation. Lag times suggest that dimer formation is slow but that the dimer, once formed, remains stable (23). Other data indicate that the EI dimer dissociates before or during the transfer of the phosphate to HPr (19,31). In relation to this, it is not clear whether phosphorylation of EI favors dimerization (27,32) or to the contrary promotes monomer formation (19,21).

EXPERIMENTAL PROCEDURES
Materials-NADP (sodium salt) and bromopyruvic acid were purchased from Fluka. PEP (cyclohexylammonium salt) and NADH (disodium salt) were from Sigma. Chloropyruvic acid was synthesized as described previously (44). The synthesis and purification of compounds 1-4 will be reported elsewhere. They were characterized by 1 H, 13 C, and 31 P NMR spectroscopy and mass spectrometry. The purity of compounds 1 and 2 was higher than 99%, as determined by high performance liquid chromatography (not shown). Rabbit muscle lactate dehydrogenase and yeast glucose-6-phosphate dehydrogenase were from Roche Molecular Biochemicals.
Site-directed Mutagenesis-The coding regions for EI-and IIA Glc were cloned in the expression vector pJF119EH (47), and the H189A, C502A mutants of EI and the H90Q mutant of IIA Glc were prepared by recombinant circle PCR to afford plasmids pMSEI189, pMSEI189H6, pMSEI502, pMSEI502H6, and pMSIIA90 (48).
Assay for PEP:Glucose Phosphotransferase Activity-In vitro PTS activity was coupled to the oxidation of glucose 6-phosphate to 6-phosphoglucono-␦-lactone, catalyzed by glucose-6-phosphate dehydrogenase. Reactions were performed in 96-well microtiter plates. The formation of NADPH at 30°C was monitored continuously at ϭ 340 nm in a Spectramax-250 plate reader. Initial rates were determined from the maximal slope of the absorption curves. Assay mixtures contained (150 l/well): 1 M HPr, 20 M IIA Glc , 1 l of membrane extract, 0.1 unit of glucose-6-phosphate dehydrogenase, 1 mM NADP ϩ , 50 mM HEPES pH 7.5, 2.5 mM DTT, 2.5 mM NaF, and 5 mM MgCl 2 . The concentrations of EI and PEP are indicated in the figure legends. EI was always ratelimiting. The rates did not exceed 25% of the activity attained at saturating EI. The reactions were started by the addition of 10 l of 15 mM glucose, PEP, or the PEP analog, to 140 l/well of the assay mixture containing the rest of components preincubated for 10 min at 30°C.
Inactivation Assays-EI (0.051-5.1 M) was shaken for 10 min at 30°C in the presence of 0.5 M HPr, 5 M IIA Glc (or 0.5 M IIAB Man ), 0.01 l/l membranes containing IICB Glc (or IICD Man ), 50 mM HEPES pH 7.5, 2.5 mM DTT, 2.5 mM NaF, 5 mM MgCl 2 , 1 mM glucose, 1 mg/ml bovine serum albumin, 6 units/ml lactate dehydrogenase, and 1 mM NADH. Compounds 1-4, chloro-or bromopyruvic acid were added to a final concentration of 0 -5 mM, as indicated in the figure legends or in the text. Aliquots were withdrawn at intervals, and the reaction was stopped by dilution into at least 10 volumes of cold (4 -10°C) buffer containing 1 mg/ml bovine serum albumin, 50 mM HEPES, 2.5 mM DTT, 2.5 mM NaF, 5 mM MgCl 2 , and 0.3-1 mM PEP. PTS activity of inactivated EI was assayed in the PEP:glucose phosphotransferase assay at 0.017 M final EI concentration and 1 mM PEP. In the case of incubation of 0.051 M EI, the aliquots (50 l) were assayed directly by addition to 100 l of a mixture of the rest of components incubated at 30°C and containing 3 mM PEP. Background activity was determined in the absence of EI, and it was subtracted before subsequent calculations. To measure protection of EI by substrate, the inhibitor and the protecting substrate were mixed in the desired concentration ratio and added simultaneously to the incubation mixture.
Calculation of Kinetic Parameters-The K m and k cat values were calculated by least squares fit of the initial rates measured in the presence of variable concentrations of PEP or PEP analogs 1-4 to a Michaelis-Menten hyperbola. Identification of the Labeled Amino Acid Residue by Mass Spectrometry-EI (30 M, 1 ml) was treated under turnover conditions with 2 mM Z-Cl-PEP (see above). After 60 min NaBH 4 was added to a final concentration of 20 mM and the reaction allowed to proceed at room temperature for 60 min. A second sample of EI was subjected to the same treatment without Z-Cl-PEP. Both samples were cooled on ice, dialyzed against 100 volumes of buffer (10 mM HEPES pH 7.5, 1 mM EDTA, 2 mM DTT, 0.4 M NaCl), and EI was purified by gel filtration chromatography (Superdex-200, Amersham Biosciences, Inc., same buffer). Peak fractions were pooled, concentrated (Centricon-50, Amicon) to a volume of 1 ml, and dialyzed overnight against 100 mM Tris buffer. 50 l of the solution was treated with 20 l of 200 g/ml modified trypsin (sequencing grade, Promega) at 25°C for 5 h. Proteolysis was stopped by the addition of formic acid to a final concentration of 1%. The volume was reduced to 20 l by evaporation in a Speed-Vac centrifuge. Peptides were transferred to tips containing POROS R2 (C-18 reverse phase), SCHEME I
ESI-MS and tandem mass spectrometry (ESI-MS/MS) measurements were done on an Applied Biosystems/Sciex QSTAR Pulsar hybrid quadrupole time-of-flight (QqTOF) mass spectrometer (Sciex Concord, Canada), equipped with a nanoelectrospray ion source (Protana, Odense, Denmark), under standard conditions. The spray capillary supplies a flow rate of 10 -20 nl/min. A typical voltage of 900 V was applied, and nitrogen was used as the curtain gas. Precursor ions were selected in the quadrupole Q 1 within an m/z window of Ϯ0.7 and subjected to collision-induced dissociation using nitrogen as the collision gas. Collision energies in the range of 20 -40 eV were applied. The molecular masses of the peptides were calculated from the series of multiply charged ions obtained by ESI-MS. The product ions of selected peptides were subsequently analyzed by the second stage of mass spectrometry. The accuracy of time of flight (reflectron) lies within 10 ppm. The mass spectrometer was calibrated externally using a mixture of cesium iodide and reserpine. The Applied Biosystems Analyst QS software package was used for data acquisition and processing.
Sequence and Structural Comparisons-Protein sequences (from the Swiss Protein Database) were aligned using the PILEUP program from the Wisconsin package, using standard values for the gap creation and extension penalties (49). The resulting multiple alignments were shaded using the BoxShade 3.21 program (www.ch.embnet.org/ software/BOX_form.html).

RESULTS
PEP Analogs as Substrates-Compounds 1-4 are phosphate and phosphonate enol esters, and as such they are potential substrates of EI. The Z-and E-isomers of 3-Cl-PEP (compounds 1 and 2, respectively) function as alternative substrates in the phosphotransferase assay, whereas the PEP derivatives with modifications in the carboxylic (3) or phosphate group (4) do not ( Fig. 1 and Table I). Because Z-Cl-PEP also acts as a suicide inhibitor (see below), EI becomes irreversibly inhibited during the time of the assay. More than 50% inactivation was detected at the highest concentration of 1 after 2 min. For that reason, the initial rates measured for Z-Cl-PEP have been also corrected for the progressive inactivation.
The K m values of EI for 1 (0.04 -0.08 mM) and for 2 (0.12 mM) are comparable with the K m determined for PEP (0.14 mM), which is also close to the published one (0.18 mM) (28). However, the turnover numbers (k cat ) of EI for Z-Cl-PEP and E-Cl-PEP are 2 and 3 orders of magnitude smaller, respectively, than for PEP (Table I). It is important to point out that the EI concentration had to be increased 100 -1,000-fold to detect turnover with compounds 1 and 2. Consequently, the EI dimer: monomer ratio is expected to increase. Because the dimer probably is the most active species, the difference between the catalytic constants for PEP and the 3-Cl-PEP analogs might be underestimated.
Similar results have been reported for other PEP-consuming enzymes. For example, Z-Cl-PEP and Z-F-PEP are 3 orders of magnitude worse substrates than PEP for pyruvate kinase and enolase (33,34,50). Enzymatic preference for the Z-isomer is also a common feature of PEP-utilizing enzymes, e.g. PEPCK, PPDK, and enolase (33), and it had been already observed with EI (43).
Z-Cl-PEP as Competitive Inhibitor-Because the catalytic efficiencies of 1-4 are very small, compared with PEP, these compounds can also be studied as inhibitors. Z-Cl-PEP inhibits PEP phosphoryl transfer to EI (Fig. 2A). The inhibition is competitive with a K I of 0.4 mM. E-Cl-PEP (Fig. 2B), 3 and 4 do not inhibit EI. The strikingly different inhibitory strength of the Z-and E-isomers is surprising in view of the very similar K m values measured for both compounds. The reason for this discrepancy is not known. Notice that the inhibition assays were started by the simultaneous addition of PEP and inhibitor to minimize irreversible inhibition of EI by the Z-Cl-PEP analog. With this precaution less than 10% activity was lost, even at the highest inhibitor and minimal PEP concentrations used.
Z-Cl-PEP as a Suicide Inhibitor-As has been mentioned, phosphoryl transfer from the 3-Cl-PEP analogs to EI is accompanied by the release of a chloropyruvate molecule, which might then react with nucleophilic residues in the active site. Bearing this idea in mind, EI was incubated at 30°C with compound 1 or 2, and aliquots were withdrawn at different time points to assay for residual activity. As shown in Fig. 3A, EI is inactivated by Z-Cl-PEP under turnover conditions, that is, in the presence of glucose and catalytic amounts of the other PTS components. Incubation of EI with Z-Cl-PEP alone is not inhibitory. Omission of Mg 2ϩ prevents inactivation at all EI concentrations tested (0.051-5.1 M), highlighting the importance of this cation for catalysis (and consequently inactivation). The results are independent of the transporter employed (glucose or mannose PTS components). Irreversible inhibition is the result of modification of EI because only the external addition of fresh EI could restore the activity. Activity was not recovered after gel filtration or extensive dialysis of the modified protein.
E-Cl-PEP is a 100-fold slower suicide inhibitor than the Z-isomer. 0.5 mM E-Cl-PEP half inactivates EI (5.1 M) after ϳ60 min of incubation, and similar to Z-Cl-PEP, turnover conditions are required (not shown). On the other hand, compounds 3 and 4, which are neither substrates nor competitive inhibitors, do not induce any inactivation. Both the phosphate and carboxylic group of PEP appear to be critical for EI binding (34,51,52). In fact, there exist very few examples of PEP analogs carrying modifications in these chemical functions which are still recognized by PEP-utilizing enzymes. The employment of thiophosphates and the remarkable case of sulfoenolpyruvate might be cited (53)(54)(55).
PEP efficiently protects against inactivation of EI by Z-Cl-PEP (Fig. 3B). Notice that at the EI concentration assayed the rate of disappearance of PEP is very fast (limited by other PTS component). Because there are no signs of incipient inactivation after 5 min of incubation, protection has to be also effective at considerably lower PEP concentrations. This is also suggested by the low inactivation observed at the lowest PEP concentrations utilized in the experiments conducted to deter-  Table I. mine Z-Cl-PEP competitive inhibition. Conversely to PEP, E-Cl-PEP does not, or very weakly, protects against inactivation (Fig. 3B).
Two experiments prove that the chloropyruvate produced from Z-Cl-PEP does not equilibrate with the bulk solution before inactivating the protein and that such a reaction involves a nonexposed residue. (i) The addition of lactate dehydrogenase/NADH, a system known to quench halopyruvate reactivity, does not reduce the rate of inactivation by Z-Cl-PEP (35). (ii) Bromo-or chloropyruvate (0.5 mM) does not inactivate EI, provided that an excess of DTT (2.5 mM) is present in the solution. The latter observation indicates that DTT is able to quench these chemicals efficiently, as Hoving et al. pointed out from EI labeling studies carried out with bromopyruvate (30). Because DTT has been used systematically in the Z-Cl-PEP inactivation experiments shown in this work, it can be concluded that chloropyruvate reacts with the protein before it leaves the protein environment, and probably at the active site.
Inactivation of EI in Presence of Phosphoryl Acceptor Proteins-Multiple turnovers, and consequently, time-dependent inactivation of EI by Z-Cl-PEP, are also elicited with a stoichiometric excess of HPr, the phosphoryl acceptor substrate of EI (Fig. 3C). Surprisingly, the addition of IIA Glc also induces inactivation, although this protein is believed to be phosphoryl-ated by HPr and not by EI. The nonphosphorylatable H90Q mutant of IIA Glc and the functionally analogous but structurally different IIAB Man subunit have no effect. It should be noticed that EI, HPr, and IIA Glc were isolated from ptsH deletion strains. It is also important to indicate that phosphoryl transfer from Z-Cl-PEP to EI is the rate-limiting step during the inactivation experiment. Therefore, the difference in the rate of phosphotransfer from EI to HPr or to IIA Glc is not being addressed in this experiment. EI phosphoryl transfer to HPr occurs orders of magnitude faster than to IIA Glc . The "crosstalk" reaction between EI and IIA Glc is unlikely to be relevant in the living cell, being probably a manifestation of the promiscuity of IIA Glc for binding a large number of different proteins. Control experiments with [ 33 P]PEP confirmed that (i) IIA Glc is phosphorylated by EI in the absence of HPr (Fig. 4B, fourth lane) and (ii) phosphotransfer from PEP to HPr or IIA Glc is also mediated by the nonphosphorylatable H189A mutant of EI (second and fifth lanes). No phosphorylation of IIA Glc could be detected in the absence of EI (not shown).
Different degrees of EI inactivation are achieved after some minutes of incubation depending on the amount of acceptor protein used, as shown in Fig. 3C, inset. Accurate estimations of the extent of maximum inactivation could be done with IIA Glc because a clear plateau is reached after a few minutes of incubation, indicative of the stability of phospho-IIA Glc . From the amount of IIA Glc consumed it can be calculated that about 10 turnovers are necessary to render EI inactive (considering EI monomer). In the case of HPr, inactivation is biphasic, initially fast, comparable with what happens with IIA Glc , then slow and linear (not shown). This second phase is probably the result of the continuous regeneration of HPr by slow hydrolysis of its phosphorylated form (56).

Effect of Z-Cl-PEP and EI Concentrations on Inactivation-
The time course of inactivation induced by 0 -1 mM concentrations of Z-Cl-PEP has been measured at three EI concentrations (0.051, 0.51, and 5.1 M).

-4 as phosphoryl donors to EI and as inhibitors of PEP phosphorylation
Activity was measured at 30°C, using 0 -2 mM concentrations of the compounds, at the EI concentrations indicated. The initial rate dependence on the concentration was fitted to a hyperbola to derive the value of the kinetic constants. Inhibition constants were obtained by measuring the initial rates at 30°C, using 17 nM EI and 0 -2 mM PEP in the presence of 0, 0.12, 0.36, and 1.08 mM concentrations of the inhibitor (Fig. 2) c Initial rates were corrected for irreversible inactivation occurring during the assay (Fig. 1) before fitting. d No inhibition detected. ND, it could not be determined. e Under the detection limit of the assay.

FIG. 2. Z-Cl-PEP and E-Cl-PEP as competitive inhibitors. Shown is a Lineweaver-Burk plot of EI inhibition kinetics in presence of Z-Cl-PEP (panel A) and E-Cl-PEP (panel B).
Concentrations were 0.017 M EI, 0 -2 mM PEP, 0 mM inhibitor (squares), 0.12 mM inhibitor, (circles), 0.36 mM inhibitor (triangles), 1.08 mM inhibitor (stars). also measured as a function of the EI concentration at a single Z-Cl-PEP concentration (1 mM) in a second experiment (Fig.  5B). The inactivation rate (k inact ) increased with the EI concentration, gradually approaching a maximum at 1.4 min Ϫ1 when EI Ͼ 1 M (inset). Moreover, the smaller the EI concentration, the higher the residual activity remaining after long incubation times (inset). Taken together these results are compatible with the notion of a dimeric form of EI being the most, if not the only, active species (see "Discussion").
Identification of Cys-502 as the Reacting Amino Acid-EI treated with Z-Cl-PEP and an untreated control were reduced with NaBH 4 and digested with trypsin. The mass spectra of the peptide mixtures are shown in Fig. 6, A and B. The peptide mass fingerprints were almost identical for both samples. However, a doubly charged peak at m/z 756.85 (1,511.7 Da) emerged in the Z-Cl-PEP-treated sample, whereas the ion with m/z 712.83 (1,423.6 Da) present in the untreated control disap-peared (Fig. 6, A and B). The two ions were selected as precursors for collision-induced dissociation. The product ion spectra included a complete series of the Yn ions of the peptide W 498 TGM(X)GELAGDER 510 , as shown in Fig. 6, C and D. Cys is the mass difference corresponding to Cys-502 (103.0 Da) in the control, whereas X matches the residue mass of an Scarboxyhydroxyethyl-labeled Cys-502 (103 Da ϩ 88 Da). Reduction with NaBH 4 was necessary to stabilize the modified cysteine. Without such treatment the corresponding tryptic peptide appeared at m/z 741.80 (1,481.60 Da). The mass difference of 58 Da, instead of the expected 86 Da, points toward the decomposition of the pyruvoyl moiety, as has been reported earlier (57,58).
These results indicate that Cys-502 is modified by Z-Cl-PEP and suggest that this residue is catalytically important. Initially the rate of inactivation of EI is compatible with the involvement of a cysteine (59). To confirm the catalytic role of this residue, the C502A mutant of EI was prepared, purified, and its activity compared with wild type and the H189A mutant of EI (Fig. 4A). C502A is virtually inactive, having less than 0.06% of wild-type activity. It is less active than the H189A mutant, which retains 0.5% activity. The C502A is phosphorylated by [ 33 P]PEP, indicating that its folding is not affected by the mutation (Fig. 4B, third lane). Phosphorylation takes place, in any case, several orders of magnitude slower than with the wild type, for which it is instantaneous (not shown). As expected, the H189A mutant is not phosphorylated (second and fifth lanes).

Cys-502 Counterparts in Homologous EI and in Other PEP-
consuming Enzymes-Cys-502 of EI is the only invariant cysteine residue in all other EI and EI-like proteins such as EI Ntr (Fig. 7). Sequence alignments reveal the presence of an invariant cysteine in equivalent positions of the homologous PPDK and PEPS (Cys-831 of C. symbiosum PPDK, Cys-750 of E. coli PEPS, Fig. 7). Moreover, Cys-831 has been found to lie close to the methylene group of a PEP molecule modeled into the x-ray structure of PPDK, suggesting the involvement of its sulfhydryl group in the protonation of the released enolate (60). PPDK is inactivated irreversibly by Z-Br-PEP (33) and bromopyruvic acid (61). Inactivation was found to be the result of the alkylation of Cys-831, and the C831A PPDK mutant was also inactive (61).
The mechanistic equivalence between Cys-502 of EI and Cys-831 of PPDK, and the x-ray structure of the latter prompted us to look for analogous amino acid residues in other PEP-utilizing enzymes of known three-dimensional structure. These structures were superimposed by matching the orientation of co-crystallized substrates or substrate analogs. The correct spatial orientation of the active sites, relative to each other, was confirmed by the presence of a similar three-dimensional arrangement of functional groups (guanidinium and carboxylate) surrounding the substrate (further information is given in the legend to Fig. 8).
Once all active sites were oriented, we then looked for residues that were overlapping with or close to the thiol group of Cys-831 of PPDK. Two types of side chains could be discerned: (i) potential catalytic residues, Glu-211 of enolase; and (ii) noncatalytic amino acids, Ile-117 of MurA, the hydrocarbon side chain of Lys-340 of EPSPS, Met-290 of PK, Leu-504 of PEPC, and Pro-98 of DAHPS (Fig. 8B). Interestingly, from all of these proteins, enolase is the only one known to catalyze a reaction at the C-3 carbon atom of PEP from the 2-re-face (62). Glu-211 has been proposed to serve as the acid/base catalyst in the activation of water for attacking PEP and produce 2-phospho-D-glycerate. The same kind of stereoreactivity is known for EI (in this case protonation) (43). In view of this observation and the sequence similarity among EI, PPDK, and PEPS, it is likely that protonation by PPDK and PEPS also takes place at the 2-re-face of PEP (as represented in Fig. 8A) and that their active site cysteines act as the acid/base catalyst. The occurrence of neutral residues for the second group of enzymes agrees with their documented 2-si-stereoreactivity: PK (63, 64), PEPC (65), DAHPS (66), EPSPS (67), MurA (68), and PEPCK (69). The potential catalytic residues of these enzymes were localized in these superimposed structures by comparison with the Cys-115 of MurA (Fig. 8B), an amino acid known to protonate PEP from the 2-si-side (68,70): Glu-341 of EPSPS (71) and Thr-327/Ser-361 of PK (72). No clear amino acid matches were found in PEPC and DAHPS, probably because some space has to be available at this side of the protein for E4P (73) and HCO 3 Ϫ binding (74). Neutral residues are present in this region of the PPDK structure (and by analogy, of EI and PEPS) and enolase (Fig. 8B), further supporting the notion that these enzymes catalyze reactions on the 2-re-side. density of the C-C double bond of PEP, compromising catalysis if protonation at C-3 is rate-limiting; and (ii) the bulkier chlorine atom raises detrimental steric interactions during approach to the transition state. Additional experiments are now being undertaken to clarify this issue. In any case, it is interesting to highlight that the small k cat values displayed by the Cl-PEP analogs allow experimentation with EI at protein concentrations that are relevant in the intact cell and under conditions where the slow monomer-dimer equilibration no longer would complicate EI kinetics.
Z-Cl-PEP is a suicide inhibitor of EI. The rate of inactivation of EI by Z-Cl-PEP depends on EI concentration, as expected if monomeric and dimeric EI have different catalytic properties. Both the maximum inactivation rate (k I max ) and the completeness of inactivation increase with the EI concentration, indicating that the dimer is the most, if not the only, active form. The incompleteness of inactivation at low EI concentrations would be a reflection of the higher fraction of monomer which might not, or only slowly react with the inhibitor. Conversely, K m inact does not vary significantly, indicating that the same active form is probably inactivated at all EI concentrations.
Inactivation of EI by Z-Cl-PEP occurs with a frequency of 1 in 50 turnovers, as calculated from the rates of chloropyruvic acid formation ((k cat /K m ϭ 400 -500 mM Ϫ1 min Ϫ1 ) and EI inac- ). An independent estimate, based on inactivation in presence of substrate amounts of IIA Glc , indicates a rate of 1 in 10 turnovers. For comparison, this inactivation efficiency is 20 -100 times higher than that of Z-Br-PEP with PEPC (35).
Cys-502 of EI is the residue labeled by Z-Cl-PEP. The C502A mutant is practically inactive, emphasizing the relevance of this amino acid for catalysis. Sequence alignments established the equivalence among Cys-502 of EI, Cys-831 of PPDK, and Cys-750 of PEPS, confirming the suggestion of Herzberg et al. (60) that Cys-831 of PPDK is involved in the protonation of the pyruvate enolate.  (79). The active site of PEPM was oriented in such a manner that the phosphate group of a bound PEP would face Asp-58, a residue proposed to be transiently phosphorylated (80). EI and PEPS fragments come from sequence comparisons with PPDK. PEPCK (E. coli, Swiss Protein P22259) residues were found by sequence alignments using small sketches of PEPC (regions comprising residues 500 -510 and 530 -545). More intriguing is the comparison of the kinetic properties displayed by Z-and E-Cl-PEP. Z-Cl-PEP is not only a 10-fold better substrate than the E-isomer, but it is also a 100-fold better suicide inhibitor of EI. Initially, this might be the result of a different orientation in the active site of the electrophilic center of the chloropyruvic molecule released from each isomer, relative to the thiol group of Cys-502. Two EI active species (presumably dimers), with different disposition of this cysteine, might be an alternative explanation. Under this hypothesis, Z-Cl-PEP, but not the E-isomer, would bind to and react with one of these species. Such argument might help the understanding of (i) why E-Cl-PEP does not protect EI from inactivation by Z-Cl-PEP, and (ii) the discrepancy observed between Z-Cl-PEP and E-Cl-PEP as competitive inhibitors. Z-Cl-PEP was found to be a much better inhibitor than the E-isomer (K I E /K I Z Ͼ 25), despite their comparable K m values (K m E /K m Z ϭ 1.5-3.0). It should be noted that, theoretically, the K m values of EI for both isomers are close to their real dissociation constants, given their low k cat .
A notion of two active forms, interconverting in presence of the substrate, would also fit in with previous published observations. Hoving et al. (30) showed that inactivation of EI, under conditions where EI is primarily in the dimeric form, by reaction with bromopyruvic acid began to happen after alkylation of one irrelevant cysteine. When the treatment was done in presence of PEP, pyruvate, or oxalate, two cysteines could be labeled without causing inactivation (30). These results were interpreted as an evidence of the presence of two nonidentical subunits composing the dimer. The effect of PEP on the reactivity of EI cysteines has been also documented by Han et al. (59) and related to conformational changes of the EI dimer. In a recent work, Brokx et al. (75) also suggested substrate-induced conformational changes of the PEP binding site of EI.