Study of the 5-oxoprolinase reaction by 13C NMR.

5-Oxoprolinase catalyzes the ATP-dependent decyclization of 5-oxo-L-proline to L-glutamate. Previous studies provided evidence for the intermediate formation of a phosphorylated form of 5-oxoproline (Seddon, A. P., and Meister, A. (1986) J. Biol. Chem. 261, 11538-11541) and of a tetrahedral intermediate (Li, L., Seddon, A. P., and Meister, A. (1987) J. Biol. Chem. 262, 11020-11025). A new approach to the study of the reaction mechanism using the 18O isotope effect on the 13C NMR signals for 5-oxoproline and glutamate is reported here. The 13C chemical shifts induced by 18O substitution for the carbonyl group of 5-oxoproline and the gamma-carboxyl group of glutamate are about 0.03 ppm with respect to the corresponding 16O-compounds. Using 5-[18O]oxo[5-13C]proline (97 and 79.5 atom % excess, 13C and 18O, respectively), the disappearance of the 18O-labeled and unlabeled 5-oxoproline and formation of the corresponding glutamate species were followed in the reactions catalyzed by purified preparations of 5-oxoprolinase isolated from Pseudomonas putida and from rat kidney. This procedure permits simultaneous determinations of the rates of 18O exchange and of the overall decyclization reaction. The ratios of 18O exchange rates to the overall reaction rates for the bacterial and kidney enzyme catalyzed-reactions were 0.28 and 0.14, respectively. The findings support the view that the coupling of ATP hydrolysis to 5-oxoproline decyclization involves formation of a phosphorylated tetrahedal intermediate. Although the exchange phenomena are consistent with the mechanistic interpretations, they seem not to be required for catalysis.

5-Oxoprolinase catalyzes the ATP-dependent decyclization of 5-oxo-~-proline to L-glutamate. Previous studies provided evidence for the intermediate formation of a phosphorylated form of 5-oxoproline (Seddon, A. P., and Meister, A. (1986) J. Biol. Chem. 261,[11538][11539][11540][11541] and of a tetrahedral intermediate (Li, L., Seddon, A. P., and Meister, A. (1987) J. Biol. Chem. 262,[11020][11021][11022][11023][11024][11025]. A new approach to the study of the reaction mechanism using the l80 isotope effect on the 13C NMR signals for 5-oxoproline and glutamate is reported here. The 13C chemical shifts induced by "0 substitution for the carbonyl group of 5-oxoproline and the y-carboxyl group of glutamate are about 0.03 ppm with respect to the corresponding "0-compounds. Using 5-[*80]oxo[5-'SC]proline (97 and 79.5 atom % excess, 13C and "0, respectively), the disappearance of the "0-labeled and unlabeled 5-oxoproline and formation of the corresponding glutamate species were followed in the reactions catalyzed by purified preparations of 5-oxoprolinase isolated from Pseudomonas putida and from rat kidney. This procedure permits simultaneous determinations of the rates of "0 exchange and of the overall decyclization reaction. The ratios of "0 exchange rates to the overall reaction rates for the bacterial and kidney enzyme catalyzedreactions were 0.28 and 0.14, respectively. The findings support the view that the coupling of ATP hydrolysis to 5-oxoproline decyclization involves formation of a phosphorylated tetrahedal intermediate. Although the exchange phenomena are consistent with the mechanistic interpretations, they seem not to be required for catalysis. and was presented at the Meeting of the American Society for Biochemistry and Molecular Biology, Las Vegas, NV, May 2-6, 1988.
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.
$ To whom correspondence should be addressed Dept. of Biochemistry, Cornell University Medical College, 1300 York Ave., New York, NY 10021. coupled cleavage of ATP and that of the amide carbonyl bond of 5-oxoproline, ATP is required thermodynamically to drive the intrinsically endergonic decyclization of 5-oxoproline and also catalytically for activation of the amide carbonyl bond (1-4). 5-Oxoprolinase from Pseudomonas putida (l), unlike that from rat kidney (5) Previous studies (4) showed that, in the reaction catalyzed by the bacterial enzyme complex, "0 from Hk80 is incorporated into the amide carbonyl oxygen of the residual 5-oxoproline. In addition to the predicted incorporation of "0 into Pi and into the y-carboxyl group of glutamate, significant dilabeling of Pi and of the y-carboxyl group was observed. Results obtained in studies in which 5-[180]oxoproline was used were complementary to those conducted in H2"0. The findings show that there is no transfer of l80 from the amide carbonyl oxygen of 5-oxoproline to inorganic phosphate.' Thus, the amide carbonyl oxygen of the residual 5-oxoproline and the y-carboxyl oxygen atoms of glutamate are replaced to a significant extent by oxygen atoms derived from water. These findings led to the proposal of a mechanism that involves phosphorylation of the amide carbonyl oxygen of 5oxoproline (3) and formation of a tetrahedral intermediate (4). Ring opening of the latter gives y-glutamyl phosphate, which undergoes hydrolysis. Randomization of the phosphoryl group and interconversion of the intermediates were proposed to account for these observations (4).
In this study, the "0 isotope effect on the 13C NMR (9-12) ]glutamate and the corresponding 13C,"j0 analogs was utilized to investigate kinetic aspects of the oxygen exchange between water and reaction pathway intermediates. The reaction course and oxygen exchange were simultaneously monitored, and the respective rates were determined.
Reversible transfer of oxygen from the y-carboxyl group of glutamate to inorganic phosphate occurs in the reaction catalyzed by glutamine synthetase (6, 7), which involves intermediate formation of y-glutamyl phosphate (8). Although y-glutamyl phosphate is postulated as an intermediate in the reaction catalyzed by 5-oxoprolinase, this reaction also involves formation of a phosphorylated tetrahedral intermediate (see Fig. 4

EXPERIMENTAL PROCEDURES
Materials-~-[5-'~C]Glutamic acid (97-99 atom % excess) and H2180 (97-99 atom % excess) were purchased from MSD Isotopes. Sources of all other chemicals and materials were as described (2). P. putida (strain ALA) was that previously obtained by soil enrichment culture (1). 5-Oxoprolinase (320 units/mg) was isolated and purified as described (1,2). Rat kidney 5-oxoprolinase (22 units/mg) was purified essentially as described (5), except that chromatography on phenyl-Sepharose was omitted. One unit of enzyme activity is defined as the amount of enzyme required for the conversion of 1 pmol of substrate to product in 1 h at 37 "C under the assay conditions.
described (1). 13C NMR spectra were obtained at 24 "C with a 500-Methods-Assays for 5-oxoprolinase activity were carried out as MHz Bruker spectrometer (AM series) using a 10-mm broad-band probe tuned at 125.78 MHz. Samples for analysis were dissolved in 0.6 ml of deuterium oxide (99 atom % excess 'H). A 500-Hz spectral width and a 90" pulse angle (20 p s ) were applied. Exponential multiplication was used for Fourier transform spectra using a linebroadening factor of 0.2. Relative intensities of the corresponding I3C NMR signals were calculated from integrated areas of the peaks. To ensure the accuracy of the measurements, [5-'3C,-1aO]glutamate was isolated by AG 50W-X8 H' chromatography, and '*O enrichment was determined by mass spectrometry as described (4). The results obtained by NMR and by mass spectrometry were in close agreement (?6%). RESULTS   . The reaction was initiated by addition of 60 units of bacterial 5-oxoprolinase. At the indicated intervals, 0.6-ml portions were removed, adjusted to pH 6.0 by addition of HC1, and rapidly frozen. After sample collection, the solutions were thawed, centrifuged, neutralized by addition of NaOH, and lyophilized. Each sample was then dissolved in 0.6 ml of D20 and analyzed by NMR spectrometry (48 acquisitions/ spectrum). thesized [5-'3C,'BOz]glutamate gave a signal that was 0.06 ppm upfield relative to that for [5-13C]glutamate; thus, the isotope effect is additive, as previously noted (11). The 13C resonance Study of the 5-Oxoprolinase Reaction by 13C NMR signal for the y-carboxyl carbon of glutamate is 0.475 ppm upfield of that for the amide carbonyl carbon of 5-oxoproline (Fig. 3). Thus, the signals for 13C-enriched glutamate and 5oxoproline and the corresponding "0-labeled compounds can be simultaneously observed directly without isolation and purification of the samples prior to analysis. Since the inten- peak areas of the isotopic species (lo), the concentrations of the reactant, 5-oxoproline, and of the product, glutamate, can be quantitated.

5-['s0]0xo-~-[5-'3C]
proline was used as substrate for the reaction catalyzed by bacterial 5-oxoprolinase (Fig. 3). Fig. 3 shows a stacked plot composed of a series of NMR spectra obtained on samples of the reaction mixture removed at the time intervals indicated. The disappearance of 13C,'80-and '3C,'60-labeled 5-oxoproline and the formation of corresponding glutamate species are inversely related. It is apparent that the ratio of "0 to "j0 in the substrate is greater than that in the product glutamate. Similar results were obtained in reactions catalyzed by rat kidney 5-oxoprolinase (Fig. 4).
Conversion of the integrated area for each peak in Fig. 3 to the amount of each 13C-labeled species, plotted as a function of time, is shown in Fig. 5. At saturating levels of substrates and in the presence of an ATP-regenerating system (phosphoenolpyruvate and pyruvate kinase), the decrease of 5oxoproline and the formation of glutamate were linear with time for about 90% of the reaction time course. Initial "0 enrichment of 5-oxoproline was 79.5 atom % excess in the amide carbonyl group. "0-Labeled and unlabeled 5-oxoproline decreased at rates of 0.093 and 0.024 pmol/min, respectively. The rate of formation of "0-labeled glutamate was 0.077 pmol/min, whereas that of unlabeled glutamate was 0.049 pmol/min, giving rise to an enrichment in the y-carboxyl group of glutamate of 59.4 atom % excess. If no exchange between medium water and the amide carbonyl oxygen had occurred, the rate of disappearance of the "0-labeled substrate would be equal to the rate of appearance of the "0labeled glutamate species. Similarly, the rate of disappearance of unlabeled 5-oxoproline would be equal to that of the appearance of unlabeled glutamate. It is apparent that this is not the case. In addition to labeled glutamate, "0-labeled 5oxoproline was also converted to unlabeled glutamate during the course of the reaction. Analyses similar to those described above were performed under various conditions (Table I). Addition of phosphoenolpyruvate and pyruvate kinase to the reaction mixture markedly accelerated the rate of glutamate formation for the reaction catalyzed by bacterial 5-oxoprolinase. The loss of "0 label from glutamate, however, was not significantly influenced by an increase in reaction rate (Table I, Systems 1 and 2). The effect of decreasing pH on isotope exchange was also investigated. At pH 7.6, the rate of glutamate formation was approximately half that at pH 8.2; however, "0 enrichment in glutamate was not significantly different at these two pH values. Thus, a decrease in the pH of the reaction or removal  time. Experimental conditions were identical to those described for Fig. 3,

-0~
except that the total reaction time was 210 min. Aliquots of the reaction mixture (* I were removed at the indicated intervals GLU and the ratios of "0 to l60 in 5-oxopro-( -1 line (5-OP) and glutamate were determined by 13C NMR spectroscopy. The progress of the reaction (as percent of completion) was derived from the sum of the integrated areas for both glutamate species. of ADP from the reaction system had no effect on the isotope enrichment in glutamate. Loss of "0 label was also observed in reactions catalyzed by rat kidney 5-oxoprolinase, but was less than that observed with the bacterial enzyme system. The presence of an ATP-regenerating system increased slightly the rate of glutamate formation; it did not significantly affect "0 enrichment of glutamate (Table I, Systems   3 and 4).
The irreversibility of the 5-oxoprolinase reaction, previously demonstrated by the finding that the enzyme does not catalyze the formation of 5-oxoproline from glutamate in the presence of ATP or ADP and Pi, or of ATP from Pi and ADP in the presence of glutamate (4), was also examined by 13C NMR. The time course of exchange for an experiment in which bacterial 5-oxoprolinase was incubated with 5-["0] ox0 [5-'~C]proline in the presence of ATP and an ATP-regenerating system is shown in Fig. 6. The ratio of "0 to l 6 0 in 5-oxoproline and in glutamate was followed as a function of the reaction time. The reaction was essentially complete in 30 min. Incubation was extended to 210 min to detect possible reversibility of the reaction which would be reflected by a further decrease in the ratio of to "0 in glutamate after the reaction had gone to completion. It is apparent from Fig.  6 that the reaction is irreversible since the decrease in the ratio of "0 to l 6 0 in glutamate was parallel to the reaction course and was not significantly affected by the extended incubation period. A decrease in the ratio of "0 to l60 in the amide carbonyl oxygen of 5-oxoproline was also observed. This is consistent with exchange of oxygen between the amide carbonyl group and medium water during the enzymatic reaction, an observation made earlier in the experiments using mass spectroscopy (4). The results demonstrate that the exchange between water and reaction intermediates cannot be accounted for by nonenzymatic oxygen exchange or by partial reversibility of the reaction.
Oxygen exchange between enzyme-bound intermediates and medium water may be described by the scheme presented in Fig. 7. S1, T1, and P1 are 5-oxoproline, transitional intermediates, and glutamate, respectively, that are labeled with "0. Sz, Tz, and Pz are the corresponding unlabeled species; k,, and kclu are the apparent rate constants for isotope ex-  Table I. The ratio of the rate of oxygen exchange between water and 'EO-labeled intermediates to the rate of glutamate formation for the bacterial enzyme-catalyzed reaction (Systems 1 and 2) was 0.25-0.3 and was greater than found for the rat kidney enzyme-catalyzed reaction.

DISCUSSION
Application of the l80 isotope effect on the 13C NMR signals for 5-oxoproline and glutamate permitted simultaneous determination of the rate of conversion of 5-oxoproline to glutamate and of the rate of oxygen exchange between medium water and intermediates on the reaction pathway. Consistent with previous observations (4), the amide carbonyl oxygen of 5-oxoproline and the y-carboxyl oxygen atoms of glutamate were shown to undergo replacement by oxygen atoms derived from water. These results provide insight into the kinetics of oxygen partitioning in the 5-oxoprolinase reaction. The rate of exchange was about one-fourth of that for glutamate formation in reactions catalyzed by bacterial 5-oxoprolinase, whereas the rate of exchange for the kidney enzyme-catalyzed reaction was about one-eighth of that for glutamate formation. Since the medium water was not enriched with "0, such a partitioning of labeled and unlabeled intermediates can account for the substantial decease in the 180/'60 ratio in both the amide carbonyl group of 5-oxoproline and the y-carboxyl group of glutamate. It is of interest, for the bacterial enzyme system, that despite a markedly accelerated reaction rate in the presence of an ATPregenerating system, the ratio of the exchange rate to the reaction rate was not affected by removal of ADP. This suggests that randomization of the phosphoryl group occurs prior to release of ADP or, alternatively, that ADP is not involved in the phosphoryl transfer.
The reactions catalyzed by kidney and bacterial 5-oxoprolinase show essentially the same labeling patterns, as analyzed here by 13C NMR spectrometry and previously by mass spectrometry (4). The difference between the two enzyme systems is in the extent of "0 loss or incorporation. This may probably be ascribed to the structural differences between the two enzymes. Unlike the two tightly associated subunits of the kidney enzyme ( 5 ) , bacterial 5-oxoprolinase is a reconstituted enzyme. The substrate-dependent interactions between Components A and B in the enzyme complex and the relative strength of these interactions compared to the kidney enzyme may be expected to lead to differences in the microscopic rate constants when ATP hydrolysis is coupled to the decyclization of 5-oxoproline. Some variation in the extent of "0 exchange has been observed with different batches of bacterial enzyme preparations. For example, in previous studies using 5-[180] oxo-L-proline with an "0 enrichment of 45 atom % excess, mass spectral analysis of the glutamate formed showed a final "0 enrichment of only 8% in the y-carboxyl group (4). In this study with 5-['80]oxo[5-'3C]proline, in which "0 enrichment was about 80 atom % excess, 13C NMR analysis of the glutamate formed showed that "0 enrichment of the y-carboxyl group was about 60%. No significant 13C isotope effect to account for these observations was detected when 13C-enriched 5-oxoproline was incubated with the bacterial enzyme in H2180 and compared to an experiment carried out using un-enriched 5-oxoproline. Moreover, "0 enrichment of samples determined by 13C NMR was in close agreement with that determined by mass spectrometry (f6%). Previous studies (13) on the reaction catalyzed by rat kidney 5-oxoprolinase using "0-labeled 5-oxoproline showed complete retention of the isotope in glutamate. Considering that the coupling of ATP hydrolysis to 5-oxoproline decyclization is realized through the formation of the tetrahedal intermediate and that the observed oxygen exchange patterns can be ascribed to the randomization of the phosphoryl group in this intermediate, it is possible that "0 isotope partitioning can be influenced by the extent of randomization or the rate of collapse of the tetrahedral intermediate. If a change in the microenvironment at the active center occurs and imposes a restriction on the phosphoryl transfer reaction, then the extent of the isotope loss would be reduced. Changes in the microenvironment at the active center are most likely to occur in the bacterial enzyme since it is a reconstituted enzyme complex composed of two protein components. It is as yet uncertain as to exactly how Component B participates in the overall reaction. If Component B serves as a catalyst that utilizes phosphorylated 5-oxoproline bound to Component A, it is plausible that the active center of the enzyme complex shares catalytic surfaces on both Components A and B; and thus, these surfaces may be more sensitive to modification or damage during purification.
Exchange between the amide carbonyl oxygen of 5-oxoproline and that of water proceeds without additional hydrolysis of ATP since the ratio of Pi to glutamate is within experimental error (+5%) of unity. An overall or partial reversibility of the reaction was previously excluded by the absence of 5oxoproline formation from glutamate or of ATP from ADP and Pi; this is confirmed by the 13C NMR experiments presented here.
This study provides further support for formation of a tetrahedral intermediate in the 5-oxoprolinase-catalyzed reaction. Mechanistically, oxygen exchange is apparently not required. Thus, it occurs at different rates, relative to the overall catalytic rate. Nevertheless, exchange is a consequence of the coupled reaction and is thus a property that is useful in defining the structure of the tetrahedral intermediate, which is a central complex on the reaction pathway.