Regulation and Mechanism of Phosphoribosylpyrophosphate Synthetase EXCHAXGE REACTIONS CATALYZED BY THE ENZYME*

from an exchange between W-AMP and in the absence of ribose The and of phosphate in to the over-all synthesis reaction) the the of the mixture,


Phosphoribosylpyrophosphate
(PRPP) synthetase from Salmonella fyphimurium LT-2 catalyzes an exchange between W-AMP and ATP in the absence of added ribose 5-phosphate.
The exchange reaction requires Mg++ ions and low concentrations of inorganic phosphate in addition to the substrates.
High concentrations of inorganic phosphate (which stimulate over-all PRPP synthesis and inhibit the over-all reverse reaction) inhibit the exchange reaction. Low concentrations (0.2 m&r) of ribose 5-phosphate stimulate the exchange reaction from lo-to 55-fold depending on the pH of the reaction mixture, but higher concentrations are strongly inhibitory.
The K,,, values for the substrates are not substantially altered by ribose 5-phosphate stimulation. PRPP synthetase also catalyzes an exchange between W-ribose 5-phosphate and PRPP in the absence of added adenine nucleotides. The Mg+f and phosphate requirements for this exchange reaction are similar to those of the AMP-ATP exchange reaction.
Ribose 5-phosphate-PRPP exchange is stimulated more than 200-fold by low concentrations (0.1 mM) of AMP, but is strongly inhibited by higher concentrations of AMP. AMP stimulation is not a consequence of changes in K,,, values for the substrates.
The maximal rates of the unstimulated AMP-ATP exchange reaction and of the unstimulated ribose 5-phosphate-PRPP exchange reaction are slower than the maximal rate of the over-all reverse reaction of PRPP synthetase, but the rates of the stimulated exchanges are at least as fast as the over-all reactions of PRPP synthetase. Evidence is presented that the unstimulated exchange reactions can occur in the complete absence of the stimulating partner substrates. and by a grant from A., AND BOYER, P. D., Biochemisfry, 7, 3608 (1968)).
The regulation of the biosynthesis of 5-phosphoribosyl oc-lpyrophosphate is of interest because of the important roles that this compound plays in metabolism.
A previous paper in this series described the isolation and properties of PRPPI synthetase from Salmonella typhimurium (I). End product inhibition of the enzyme has been reported (2). Studies of the detailed mechanism of action of PRPP synthetase were undertaken both because they would contribute to an understanding of the mechanism of feedback inhibition of the enzyme and because pyrophosphoryl group transfer reactions have not been previously characterized in detail.
The enzymic synthesis of PRPP from ATP and ribose 5-phosphate can be envisioned as proceeding by two general types of mechanisms, each with several possible detailed kinetic pathways. One mechanistic possibility is that an enzyme-pyrophosphate intermediate is formed, which subsequently transfers its pyrophosphoryl group to ribose 5-phosphate: ATP + enzyme + enzyme-PP + AMP (1) Enzyme-PP + ribose 5-P e PRPP + enzyme (2) A second possibility is that the pyrophosphoryl group is transferred directly from ATP to ribose 5-phosphate while the two substrates are bound to the enzyme in a ternary complex: Enzyme.ATP.ribose-5-P =: enzyme.AMP.PRPP (4) En.zyme.AMP.PRPP = enzyme + AMP + PRPP (5) The experiments in this communication constitute an attempt to distinguish between these possibilities and to obtain any possible additional information about the enzyme mechanism by a study of the properties of AMP-ATP exchange and ribose 5-phosphate-PRPP exchange reactions catalyzed by highly purified preparations of PRPP synthetase. MATERIALS AND METHODS Materials-PRPP synthetase was purified from 8. typhimurium cells and assayed as described previously (1). Disodium ATP was purchased from Sigma. Ribose 5-phosphate, sodium salt, and dimagnesium PRPP were products of Calbiochem. PRPP solutions were prepared and assayed at the time of use as previously described (1). AMP was obtained from P-L Biochemicals.
Both radiochemicals were analyzed by cellulose thin layer chromatography with t-amyl alcohol-formic acid-water (3 :2: 1) as the solvent and by high voltage electrophoresis at pH 3.5 and found to be better than 95% radiochemitally pure. The 14C-AMP contained about 2% 14C-adenosine. The 14C-ribose 5-phosphate was contaminated with traces (less than 1%) of an unidentified radioactive impurity, possibly ribose. In each case the traces of radioactive impurities were separated from compounds of interest by the techniques used for study of the exchange reactions.
The reactions were initiated by addition of PRPP synthetase and incubated at 37". The incubation tubes were covered with Parafilm. The reactions were stopped by heating in a 100" bath for 60 set, and the tubes were immediately cooled by plunging into an ice bath. Ten microliters of a "carrier" solution containing 50 mM AMP and 50 mM ATP were added to each reaction mixture and 5-~1 aliquots of each tube were spotted on plastic thin layer cellulose sheets, 20 x 20 cm (MN-Polygram Cel 300, Macherey-Nagel and Company).
The sheet.s were developed by ascending chromatography in t-amyl alcohol-formic acid-water (3 : 2 : 1, by volume) for 3 to 4 hours. After drying the plates, ultravioletabsorbing regions were located with an ultraviolet lamp. AMP and ATP were completely separated; the respective regions were cut out of the plastic sheet and counted directly in 15 ml of Bray's scintillant solution (3). Orientation of the plastic sheets in the vial was found to have negligible effect on the radioactivity observed in a liquid scintillation counter. The total radioactivity found in the AMP and ATP regions of the thin layer sheets was found to be constant regardless of the extent of exchange and accounted for essentially all of the radioactivity added. Exchange rates were calculated from the percentage of exchange equilibrium obtained and the formula (4) : Micromoles exchanged (rmoles of AMP) (pmoles of ATP) = -2'3o3 (pmoles of AMP) + &moles of ATP) Rates of exchange determined in this way were constant for up to 30 min of reaction when suitable amounts of enzyme were added. The rate of 14C-AMP-ATP exchange was proportional to the amount of PRPP synthetase added, except at very low concentrations of enzyme, at which the enzyme appeared to be less active than expected.
The unexpectedly low activity of very small quantities of enzyme probably reflects a requirement for low concentrations of phosphate (transferred with the enzyme solution) either for catalytic activity or for stabilization to denaturation at high dilution (see Reference 1 and data given below). Except where noted, all assays were performed at a final phosphate concentration of 5 my, a concentration at which the exchange rate was linearly dependent on the amount of enzyme added.
14C-Ribose 6-Phosphate-PRPP Exchange Assay-Reaction mixtures contained 50 mM triethanolamine-HCl (pH 7.5), 10 mM MgC12, and variable amounts of 14C-ribose 5-phosphate, PRPP, and PRPP synthetase in a final volume of 50 ~1. The reactions were initiated by addition of the enzyme and incubated at 37". The tubes were covered with Parafilm.
After stopping by heating at 100" for 60 set and immediately cooling on ice, 10 ~1 of a "carrier" containing 4 rnbf PRPP and 20 mM ribose 5-phosphate were added to each reaction mixture.
The entire reaction mixture was spotted on Whatman No. 3 paper with appropriate standards.
After electrophoresis in acetic acid-pyridine-water (100: 10:890, by volume), pH 3.5, for 2 hours at 4 kv (Gilson Electrophorator, model D), the paper strips were dried in air or in a warm air oven. The ribose B-phosphate and PRPP regions were located by staining control strips (5); these regions were cut out and counted in 20 ml of Bray's solution (3) or in 0.5 $JO 2,5diphenyloxazole (PPO) in toluene. In some cases the reaction mixtures were analyzed by electrophoresis of lo-p.1 aliquots in 0.02 M sodium citrate, pH 5.0, at 2' with a Wieland-Pfleiderer apparatus (Brinkmann) with a field strength of 25 volts per cm and a running time of 1.5 hours.
As with the 14C-AMP-ATP exchange reaction, the total radioactivity in the substrates was constant regardless of the extent of exchange and accounted for essentially all of the radioactivity added.
The rate of 14C-ribose 5-phosphate-PRPP exchange was proportional to the amount of PRPP synthetase added, except that, as with the AMP-ATP exchange reaction, the activity was lower than expected when the amount of enzyme per assay was less than 2 pg. This effect was shown to be a consequence of a requirement of the exchange reaction for a low concentration of inorganic phosphate (added with the enzyme solution), since the rate of exchange was shown to be directly proportional to enzyme concentration at all concentrations tested if the potassium phosphate concentration was maintained at 5 to 10 mM. Hence all assays were performed at 5 mM phosphate unless otherwise noted.
Assay of Forward and Reverse Over-all Reactions of PRPP Synthetase-The over-all forward (PRPP synthesis) reaction was assayed with the 32P transfer assay previously described (1). The reaction conditions were as follows: 0.05 M triethanolamine-HCI-0.1 M potassium phosphate buffer (pH 8.0), 10 mM MgC12, 2 mM ATP, and 5 mM ribose S-phosphate.
The reverse of the PRPP synthetase reaction was also assayed as previously described (1). The reaction mixtures contained 2 mM AMP and 0.25 mM PRPP.

W-AMP-ATP Exchange Reaction
Requirements-Exchange between AMP-8-W and nonradioactive ATP requires PRPP synthetase, MgC12, and ATP, in addition to 14C-AMP (Table I). High concentrations of inorganic phosphate (which stimulate PRPP synthesis) inhibit the exchange. This is a true inhibition rather than an activation by the triethanolamine buffer, as is shown by an experiment in which both phosphate and triethanolamine were included. It is especially noteworthy that the addition of ribose 5-phosphate is not required.
Details of the effects of ribose 5-phosphate and inorganic phosphate on AMP-ATP exchange are given in later sections.
Substrate Dependence of AMP-ATP Exchange-Dependence of the rate of AMP-ATP exchange at pH 7.5 on the concentration of AMP at two phosphate levels is shown in double reciprocal form in Fig. 1. Phosphate inhibition appears to be of a mixed competitive and noncompetitive type with respect to AMP. That is to say, the data could not be satisfactorily fit to either the competitive or noncompetitive rate equations with the computer analysis described by Cleland (6). The values obtained for the apparent K, for AMP were 4.4 f 1.1 mM (&S.E.) at 5 mM phosphate and 9.3 i 1.5 mu at 15 mM phosphate.
Both values were much higher than the K, for AMP in the over-all reverse reaction (0.15 to 0.4 mM, Reference 1). The maximal velocity of the exchange is approximately twice as great at 5 InM phosphate (1.60 f 0.24 pmoles per min per mg) as at 15 mM phosphate (0.87 f 0.10 pmole per min per mg). Both the maximal velocity and apparent K, for AMP were significantly (p < 0.01) different at the two phosphate concentrations according to a t test (6).
The effects of varying ATP and Mg++ ion concentration on the  Fig. 2) show that Mg-ATP inhibits at concentrations above 1 mM. Excess Mg++ increases the apparent maximal velocity about IO-fold, but does not significantly alter the K, for Mg-ATP, which was 1.4 mM under these conditions.
Effects of Ribose G-Phosphate on AMP-ATP Exchange-Although PRPP synthetase readily catalyzes an AMP-ATP exchange without addition of ribose 5-phosphate, the addition of low concentrations of ribose 5-phosphate stimulates the exchange appreciably.
The rates of AMP-ATP exchange at various pH values with and without 0.4 mM ribose 5-phosphate are shown in Fig. 3, which shows that the unassisted exchange reaction has a pH optimum at about 7.0. At the pH optimum for PRPP synthesis (8.0 to 8.5, Reference 1) the exchange reaction is almost absolutely dependent on the presence of ribose B-phosphate.
As the pH is lowered, the exchange reaction becomes much less dependent on ribose 5-phosphate.
The concentration dependence of ribose S-phosphate stimulation of AMP-ATP exchange is shown in Fig. 4. This dependence was examined at pH 8.3, where the exchange in the absence of ribose 5-phosphate is very slow (about 0.027 pmole per min per mg), and at pH 7.0, where it is about 15 times as fast. In both cases the exchange reaction was stimulated by concentrations of ribose 5-phosphate up to 0.2 mM and was then sharply inhibited by higher levels. The maximal stimulation appears to be higher at pH 8.3 than at pH 7.0 because the unassisted rate at pH 8.3 is very low; the actual maximal rate of exchange is about twice as high at pH 7.0 as at pH 8.3.
The possibility that ribose 5-phosphate exerts its stimulatory  in the presence and absence of 20 pM ribose 5-phosphate at, pH 8.0, where the effect of ribose 5-phosphate is very pronounced. The substrate inhibition by Mg-ATP appears to be even greater at this pH than at pH 7.5. Ribose 5-phosphate has very little effect, on the K, for ATP; from the linear portions of double reciprocal plots of the data of Fig. 5, the K, is changed only slightly (from 0.1 mM to 0.2 JIIM) by addition of 20 pM ribose 5-phosphate. The maximal velocity, on the other hand, was increased IBfold.
The concentration dependence of AMP for the exchange in the presence and absence of ribose 5-phosphate ( Fig. 6) leads to a similar conclusion.
The AMP saturation curve is sigmoid at this pH. This appears to be true whether ribose S-phosphate is present or not, although the phenomenon is less pronounced when ribose 5-phosphate was added (n = 2.1 in a Hill plot of the data obtained in the absence of ribosk 5-phosphate; when ribose 5-phosphate was present, n = 1.6). Saturation occurs at about the same concentration of AMP in both cases, but the maximal rate of exchange is clearly increased several-fold by ribose 5-phosphate.
The cause of the sigmoid concentration dependence for AMP under these conditions (pH 8.0) is not known.
of heat inactivation of PRPP synthetase activity and AMP-ATP exchange activity (Fig. 7). At 50" PRPP synthesis and the unassisted AMP-ATP exchange reactions are rapidly inactivated until a plateau is reached at about 17% of the original level. This residual activity is inactivated only very slowly.
The rates of inactivation of PRPP synthetase activity and unassisted AMP-ATP exchange activity are identical throughout the experiment. This result strongly suggests that these two activities reside in the same protein molecule.
The shape of the inactivation curve suggests that heating at 50" yields a modified enyme which is intrinsically less active, but not completely inactivated. However, an alternative explanation, the existence of two PRPP synthetases with differing sensitivities to heat cannot be excluded.

Inhibition
of AMP-ATP exchange by high levels of ribose 5-phosphate ( Fig. 4) may be a consequence of depletion of ATP to suboptimal levels by way of PRPP formation, since the equilibrium of the reaction favors PRPP synthesis (I), and the amount of enzyme and substrates present would have permitted extensive reaction.
This suggestion receives support from the experiments shown in Table II, in which inhibition of AMP-ATP exchange by increasing ribose 5-phosphate concentrations was examined at three ATP concentrations.
The inhibition was abolished by increasing the concentration of ATP from 0.4 mM to 2.0 mM, even though 0.4 mM ATP was optimal at 20 PM ribose 5-phosphate (Fig. 5). The inhibition by excess ribose 5-phosphate may also be due in part to a direct effect on the exchange process, as is predicted by the exchange rate expression for a simple ternary mechanism (see "'Appendix").
In the absence of unlabeled PRPP, no radioactivity was found in the PRPP region. It was not necessary to add AMP to observe this exchange, but it was markedly stimulated by AMP.
E$ect of AMP-The relatively slow exchange between I%ribose 5-phosphate and PRPP in the absence of added AMP (Table III) is greatly accelerated by AMP (Fig. 8); at as low as 0.1 mM AMP a 200-fold enhancement of the rate was observed. At higher concentrations AMP becomes very inhibitory. The general effects of AMP on this exchange reaction are similar to the effects of ribose 5-phosphate on the AMP-ATP exchange reaction.

Heat Inactivation of PRPP Synthesis and AMP-A TP Exchange
The dependence of the rate of the ribose 5-phosphate-PRPP Reactions-Since the PRPP synthetase preparations used in this exchange reaction on pH in the presence and absence of a highly study were highly purified, but not homogeneous, it is possible stimulatory level of AMP was examined (Fig. 9). The pH that the AMP-ATP exchange reaction-in particular, the un-optima in both cases were at 7.5, which is somewhat lower than assisted exchange-was catalyzed by an impurity in the prepara-the optimum for the over-all synthesis reaction. The pH rate tion rather than by PRPP synthetase itself. This possibility profile in the presence of AMP is very similar to, although not was rendered extremely unlikely by a study of the relative rates identical with, that observed in the absence of AMP. A solution of PRPP synthetase (30 units per mg, 0.7 mg per ml) in 0.05 M potassium phosphate, pH 7.5, was heated at 50" for the times shown. Aliquots were withdrawn and immediately cooled to 0". PRPP synthesis (0) was assayed as described under "Materials and Methods." The AMP-ATP exchange reaction (m) was assayed at 0.05 mM: imidazole-HCl (pH 7.0), 10 rnM MgCL, 2 mM '%-AMP, 1 mM ATP, without ribose 5-phosphate (BP).
Reactions were for 10 min. A double reciprocal plot of the rate of ribose 5-phosphate-PRPP exchange as a function of PRPP concentration in the presence and absence of 20 pM AMP is shown in Fig. 10. AMP stimulation is shown to be the consequence of a large increase in Conditions were as in the legend to Table III, except that the indicated concentrations of AMP were added. Times of incubation ranged from 1 to 10 min. Rates (V) are compared to VO, the rate observed when no AMP was added. R5P, n-ribose 5-phosphate. the maximal velocity of the exchange rather than an effect on the K, for PRPP, which was 0.1 mM in this experiment.
This value is within the range found for the over-all reverse reaction, which was 0.05 to 0.13 InM (1).
Similarly, the K, for ribose 5-phosphate in the ribose 5-phosphate-PRPP exchange reaction was not substantially altered by the addition of 20 PM AMP (Fig. 11). The K, for ribose 5-phosphate obtained from the data of Fig. 11 was 0.2 mM. The K, for ribose 5-phosphate in the over-all synthetic reaction (2 mu ATP, 5 MM MgClz, 0.1 M potassium phosphate (pH 7.5), Reference 1) was the same within experimental error (0.28 MM). The results indicate that AMP probably does not exert its stimulatory action through changes in the affinity of the system for its substrates.  Table III, except that AMP and PRPP were added as shown.
Times of reaction varied from 1 to 10 min. R6P, n-ribose 5-phosphate. Conditions were as in Fig. 9, except that ribose 5-phosphate (8.76 X lo6 cpm per rmole) and AMP were added as shown and the buffer was 0.05 M triethanolamine-HCI, pH 7.5.
The inhibition of ribose 5-phosphate-PRPP exchange by high levels of AMP deserves further attention.
In the case of the AMP-ATP exchange, inhibition by high levels of ribose 5-phosphate could be explained as resulting from depletion of ATP by way of PRPP formation (see Table II and the accompanying discussion), although other causes were not excluded.
It is possible that high levels of AMP might inhibit the ribose 5-phosphate-PRPP exchange through depletion of the PRPP pool via ATP formation, rather than through some direct effect of AMP on the enzyme.
No experiments were performed to test this possibility, but it appears to be ruled out by the unfavorable equilibrium of the reaction in the direction of ATP formation (1). In order for the highest level of AMP tested (4 mM) to bring about the observed inhibition through PRPP depletion alone, the con- This requires a greater than 90% conversion of PRPP to ATP and ribose 5-phosphate, which is not permitted by the initial concentrations of the substrates and the measured equilibrium constant for the reaction.
Thus, the inhibition of ribose 5-phosphate-PRPP exchange by high levels of AMP is not simply a consequence of PRPP depletion, and results from an effect of AMP itself (or possibly ATP) on the exchange reaction (see "Appendix").
Heat Inactivation of Ribose 6-Phosphate-PRPP Exchange Reaction-The possibility that the ribose 5-phosphate-PRPP exchange studied here was catalyzed by an impurity in the enzyme preparation, rather than by PRPP synthetase itself, was tested by heat inactivation experiments.
In Fig. 12 it may be seen that heating the enzyme solution at 50" very rapidly inactivates slightly over 80% of the PRPP synthetase activity and the ribose 5-phosphate-PRPP exchange activity. 'r'he remaining activity is then only very slowly inactivated by further heating.
These results reproduce very well the results of an earlier experiment on the heat lability of the AMP-ATP exchange reaction (Fig. 7). The lability to heat of the PRPP synthesis and unassisted ribose 5-phosphate-PRPP exchange reactions are clearly identical, but the AMP-stimulated exchange appears to be inactivated to a slightly greater extent.
This difference may not be significant; after 30 min of heating, 19% of the PRPP synthetase activity remained, while 15% of the original assisted exchange activity was found. These rather small differences are exaggerated by the semilogarithmic plot. In any case, the striking similarity between the heat inactivation curves provides strong evidence that the ribose 5-phosphate-PRPP exchange reaction-like the AMP-ATP exchange reaction-is catalyzed by PRPP synthetase rather than an impurity in the preparation.

EJects of Phosphate on Exchan.ge and Over-all
Reactions of PRPP Synthetase The effect's of inorganic phosphate on the reactions catalyzed by PRPP synthetase are complex.
In a previous paper (1) it was shown that the enzyme has a specific and apparently absolute requirement for high levels of phosphate when the synthesis of PRPP was measured.
Double reciprocal plots of the effects of phosphate on PRPP synthesis are bimodal and are described by two linear portions corresponding to apparent K, values of 2.3 mM and 40 mM at pH 7.5. Removal of phosphate from the enzyme by dialysis or by simply diluting the enzyme into colddistilled water inactivates PRPP synthetase; under some conditions activity can be restored by adding phosphate to enzyme solutions that contain very low concentrations of phosphate.
Effects of Phosphate on AMP-A TP Exchange-The AMP-ATP exchange reaction is strongly inhibited by inorganic phosphate (Table I and Fig. 13). Even though this inhibition was observed at all concentrations above 5 mM, removal of phosphate from the enzyme solution by overnight dialysis against Tris-Cl, pH 7.5, or triethanolamine-HCl abolished the ability of the enzyme to catalyze the exchange reaction.
Simultaneously the enzyme lost over 90% of its ability to catalyze PRPP synthesis (32P transfer assay). Phosphate is an effective inhibitor at quite low concentrations (Fig. 13) Table  III or Fig. 15, respectively; 0, ribose 5.phosphate-PRPP exchange reaction in the presence of 4 X 10e6 M AMP (data of Fig. 15) ; 0, AMP-ATP exchange reaction, measured as in Table   I. R5P, n-ribose 5-phosphate; CONC, concentration.
The exchange between AMP and ATP which is catalyzed by PRPP synthetase is sensitive to inhibition by phosphate whether the exchange is stimulated by ribose 5-phosphate or not (Fig. 14); however, the concentration dependence of phosphate inhibition is altered by ribose 5-phosphate. In Fig. 14 it is seen that the presence of 0.1 mM ribose 5-phosphate lowers the sensitivity of the exchange to phosphate inhibition.
It is significant that the relative sensitivity of AMP-ATP exchange to phosphate is shifted by ribose 5-phosphate to a curve which is identical with the phosphate sensitivity curves of the reverse over-all reaction and the ribose 5-phosphate-PRPP exchange reaction (see below). This finding suggests that ribose 5-phosphate may act by altering the rate-determining step of the exchange so that it becomes the same step that determines the rate of these other reactions.
Effects  (Table III and Fig. 13). As a function of phosphate concentra-Con, inhibition of this exchange reaction is more gradual than phosphate inhibition of AMP-ATP exchange. Enzyme solutions which have had phosphate removed by dialysis have less than 3 $& of the exchange activity of an undialyzed control; such a dialyzed preparation has also lost its ability to catalyze PRPP synthesis and the AMP-ATP exchange reaction. A possible interrelation between phosphate inhibition and AMP stimulation of the ribose 5-phosphate-PRPP exchange reaction was probed by examining the concentration dependence of phosphate inhibition of the exchange in the presence and absence of 40 PM AMP (Fig. 15). The results show that the sensitivity of the exchange to phosphate inhibition was nearly the same, if not identical, whether AMP was present or not, even though the rate of exchange was more than 100 times as fast when AMP was added.
Effects of Phosphate on Reverse Over-all Reaction-It has been shown that the removal of inorganic phosphate from PRPP synthetase by dialysis inactivates the enzyme, both with respect to its ability to catalyze PRPP synthesis and the exchange reactions. Since the requirement for phosphate is a very unusual one, this observation may be taken as additional evidence that the exchange reactions and the over-all PRPP synthesis reaction are catalyzed by the same protein, and, further, that the exchange reactions are kinetically significant components of the over-all reaction.
On the other hand, the fact that concentrations of phosphate above 5 mM inhibit the exchange reactions while stimulating the synthetic reaction seems to contradict this conclusion.
The probable resolution of this apparent contradiction has been obtained from studies on the effects of phosphate on the reverse reaction of PRPP synthetase, i.e. on the formation of ATP from PRPP and AMP. Fig. 13 shows that the reverse reaction requires low levels of phosphate but that, when the concentration of phosphate exceeds 5 mM, the reaction is inhibited. For comparative purposes the effects of phosphate on the exchange reactions are superimposed in the same figure. It is clear that the effect of increasing phosphate concentration on the ribose 5-phosphate-PRPP exchange reaction is essentially the same as on the reverse reaction.
The AMP-ATP exchange reaction is also inhibited by phosphate, but it is somewhat more sensitive to increasing phosphate concentration than the reverse reaction.
However, it will be recalled that, when the AMP-ATP exchange reaction is stimulated by ribose 5-phosphate, the phosphate inhibition curve also shifts to one which is the same as that of the reverse reaction (Fig. 14). Data about the effects of phosphate levels below 5 mM on the exchange reactions have not been determined, but it is probable that the exchange reactions also require low concentrations of phosphate since removal of phosphate by dialysis inactivates them and since 5 mM phosphate must be included to yield a linear response of the exchange rates to enzyme level (see "Materials and Methods").

DISCUSSION &faximaZ
Rates of Exchange Reaction-The maximal rates of the unassisted exchange reactions appear to be appreciably slower than the maximal rate of either the reverse or the forward over-all reactions of PRPP synthetase.
Comparison of these rates is somewhat uncertain because of the complex effects of phosphate on the system. Since 5 mM phosphate is optimal for the reverse reaction and for both of the exchange reactions, I have chosen the assay condition containing 5 mM potassium phosphate, pH 7.5, as a "standard state" for comparing these rates. There is evidence that PRPP synthetase is not stable at this phosphate concentration, however.
Dilution of the enzyme into 2.5 mM phosphate brings about a very rapid decline in synthetic activity to a value that is about 30% of the initial value and does not decline further (1). It is possible that a similar partial inactivation occurs during dilution of the enzyme into the 5 mM phosphate which was present in the exchange assays.2 Furthermore, the forward reaction requires much higher concentrations of phosphate and proceeds at only about 20% of the maximal rate at 5 mM phosphate.
These complexities should be kept in mind in assessing the following comparisons.
In a previous paper (I) it was shown that the reverse of the PRPP synthetase reaction is only about 13 70 as fast as t.he synthesis of PRPP.
This comparison was made from reaction rates obtained with 50 mM phosphate, a concentration in which the enzyme is known to be stable. The reverse reaction is about 1.85 times as fast at 5 mM phosphate (Fig. 13); this yields an estimate for the maximal velocity of the reverse reaction of 10.9 pmoles per min per mg of enzyme. The forward reaction can be estimated in a similar fashion to be about 20% of its maximal rate at 0.1 M phosphate, or approximately 13 pmoles per min per mg.
The maximal velocity of AMP-ATP exchange at saturating AMP concentrations, 1 mM ATP (the optimum concentration), and 5 mM phosphate was 1.6 pmoles per min per mg (Fig. 1). Thus, even if appreciable inactivation results from diluting the enzyme into 5 mM phosphate, the maximal rate of AMP-ATP exchange is probably less than half of that of the reverse reaction. The presence of low levels of ribose 5-phosphate stimulates the AMP-ATP exchange from lo-fold at pH 7.0 to 55-fold at pH 8.3, bringing about rates of exchange that are clearly faster than the reverse reaction and are probably as fast as the forward reac-Con. The phosphate inhibition curve and the pH activity profile of the ribose 5-phosphate-stimulated AMP-ATP exchangerather than the unassisted exchange-most closely resemble those of the reverse reaction.
These observations suggest that the stimulated exchange proceeds by steps that are involved in the over-all reaction, while the unstimulated exchange may liot. The maximal rate of the ribose 5-phosphate-PRPP exchange is much slower than the AMP-ATP exchange unless AMP is added. From the data of Fig. 10 a maximal velocit,y of 0.09 pmole per min per mg was obtained at saturating PRPP concentrations and 1 mM ribose 5-phosphate.
Therefore, even after extrapolation to saturating ribose 5-phosphate concentrations and allowing for a a-fold inactivation on dilution into 5 mM phosphate, the maximal rate of the unstimulated ribose 5-phosphate-PRPP exchange is not above 0.4 pmole per min per mg. On the other hand, low concentrations of AMP have been shown to stimulate the exchange by more than 200-fold-yielding rates which are easily as rapid as the reverse reaction.
It might be expected that the maximal rates of the exchange reactions would be equal to or greater than the maximal rate of the reverse or forward reaction, whichever is the slower, if the exchange reactions proceed by the same reaction path as the 2 If such inactivation does take place, it must be very rapid and must cease at a finite value, because the rate of both exchange reactions can be shown to be constant for at least 30 min at 37" in 5 mM potassium phosphate. over-all reaction. Actually, it can be shown for some of the mechanisms applicable to this system that this condition need not be met (see "Appendix").
Nonetheless, the stimulation of the exchange reactions by partner substrates provides valuable information about the mechanism of the PRPP synthetase reaction even if no assumptions about the maximal rate of exchange and over-all reactions are made.
It should be understood that the exchange rates determined in this work were not obtained at reaction equilibrium.
In the cases of the unassisted exchange reactions, this presents no problem since there is no net react'ion during the exchange assay. However, when the stimulating substrates were added, net reaction as well as exchange takes place. This leads to errors in the determination of the exchange rate, especially in those cases in which the level of stimulating substrate is high. The data are adequate to support the qualitative conclusions drawn in this paper, but the use of exchange rates in conjunction with a detailed steady state kinetic analysis will require determination of true equilibrium exchange rates. Do Exchange Reactions Atsolutely Require AMP and Ribose &Phosphate?-It is useful to consider the results of this study in relation to the two general mechanisms described in the introductory section, namely, a mechanism involving formation of an enzyme-pyrophosphate intermediate and a mechanism in which direct pyrophosphate transfer occurs on a ternary complex.
The latter mechanism requires that neither AMP-ATP exchange nor ribose 5-phosphate-PRPP exchange can occur in the absence of the other substrate, i.e. ribose 5-phosphate or AMP, respectively (see "Appendix").
Both of the exchange reactions were readily observed without adding these compounds.
However, since both AMP and ribose 5-phosphate bring about marked stimulation, is it possible that these compounds were present as impurities? In the case of AMP-ATP exchange, the fact that the assisted and unassisted exchange reactions have different pH dependencies (Fig. 3) makes it unlikely that ribose 5-phosphate-independent exchange is an artifact caused by ribose 5-phosphate contamination of the enzyme or substrates.
If the exchange reaction had an absolute requirement for ribose 5-phosphate and was observed in the absence of added ribose 5-phosphate only because of contamination, the assisted and unassisted exchange reactions should have the same pH rate profile, since they would be determined by the same enzymic events. Furthermore, no ribose 5-phosphate has been detected in the reactants with the very sensitive 32P transfer assay (1). With respect to the ribose 5-phosphate-PRPP exchange, the pH dependencies of the exchange with and without AMP (Fig. 9) are very similar, so no argument can be made t.hat an AMP impurity is absent as was the case with AMP-ATP exchange.
However, if either substrate contained AMP as an impurity, one would not expect the exchange rate to reach a maximum as substrate was added until the system was saturated with both substrate and AMP.
Yet the rate enhancement by added AMP is observed when the system is saturated with either ribose 5-phosphate or PRPP (Figs. 10 and 11). Therefore, it is unlikely that either substrate contained AMP as an impurity. Similarly, if the enzyme contained AMP as an impurity, a curve of rate plotted against enzyme concentration might be expected to curve upward; it does not. The enzyme which was used in these studies has been purified some 450-fold by a procedure which includes repeated ammonium sulfate and acid precipitations. Thus, while the presence of very small quantities of very tightly bound AMP cannot be excluded, it is unlikely that the enzyme contains significant amounts of AMP. The results justify the tentative conclusion that at least one and possibly both exchange reactions can occur without the involvement of the stimulating partner substrate.

Mechanism of PRPP Synthetase Reaction-If
one accepts the conclusion that the stimulation of the exchange reactions catalyzed by PRPP synthetase is the consequence of an acceleration of previously existing exchange processes, what can be deduced from these findings about the reaction mechanism?
It is likely that the unassisted exchange reactions proceed by way of an enzyme-pyrophosphate intermediate: and ATP + enzyme= enzyme-PP + AMP PRPP + enzyme = enzyme-PP + ribose-5-P The occurrence of these reactions tends to implicate the two-step mechanism of Equations 1 and 2 (a ping-pong bi bi mechanism in the terminology of Cleland, Reference 7). This simple mechanism must be excluded, however, because it predicts that the only effect of adding the second substrate will be to inhibit the exchange reaction (see "Appendix").
A mechanism involving an enzyme-pyrophosphate intermediate can be accommodated only if binding of the second substrate in some way increases the rate of the exchange process.
A ternary mechanism involving direct transfer of pyrophosphate from ATP to ribose 5-phosphate predicts that the nonexchanging substrate will first stimulate and then inhibit the exchange reaction as the concentration is increased (see the "Appendix" for a discussion of a simple case). Such a model accounts qualitatively for the observed effects of AMP and ribose 5-phosphate on the exchange reactions, but also requires that the exchange cannot occur in the absence of the stimulating substrates.
From the above considerations it appears that only a mechanism involving ternary complexes can account for the observed properties of the enzyme reactions.
There are two general possibilities.
The first, which I favor, envisions a scheme in which an enzyme-pyrophosphate is formed as in the unassisted exchange reactions, but in which the partner substrates remain attached to the active site and alter the rates of individual catalytic steps. Thus, the normal reaction sequence might be as follows (without necessarily requiring the order of binding given).
The postulate must also include reactions involving the free enzyme-pyrophosphate intermediate, which participates in the unassisted exchange reactions (Equations 6 and 7), but which is not a participant in the over-all reaction or the assisted exchange reactions.
The hypothesis is shown in schematic form in Fig. 16 A similar shift in mechanism would presumably apply to AMP stimulation of the ribose 5-phosphate-PRPP exchange reaction.
The over-all reaction mechanism of PRPP synthesis in this formulation would proceed by direct pyrophosphate transfer in ternary complexes as shown in Equations 3,4, and 5. The data in this paper do not permit a conclusive choice between these : d mechanism on the addition of a second substrate seems unwarranted.
If the enzyme is capable of forming a covalent pyro- Enz, enzyme; R-S-P, D-ribose

5-phos-
The suggested mechanism predicts that it should be possible to phate.
isolate an enzyme-pyrophosphate intermediate with substrate amounts of purified PRPP synthetase and 32P-labeled ATP.
In in the laboratory of Dr. Earl R. Stadtman, whose generous support and helpful criticisms are gratefully acknowledged. It is a fact, it has been possible to isolate a phosphorylated derivative of PRPP synthetase (9). A detailed study of this derivative pleasure to acknowledge Professor Paul Eoyer for valuable asmay provide a definitive examination of the mechanistic possi-sistance with the ideas developed in the "Appendix" and Drs.
bilities suggest.ed in this paper. L. N. Ornston and B. V. Plapp for critical readings of the manu-If the PRPP synthetase reaction does indeed proceed by way script.
of an enzyme-pyrophosphate intermediate, it will be of great APPENDIX interest to compare this enzyme with the phosphoenolpyruvate synthases from Escherichiu coZi (10, 11) and propionibacteria (12).

Derivation and Properties of Equilibrium
Exchan.ge Rate Enzyme-pyrophosphate intermediates have been proposed for Expressions for Two Simple Mechanisms for both of these enzymes on the basis of exchange studies. In these PRPP Xynthetase cases, however, the pyrophosphoryl group does not appear to be transferred directly, but rather it is cleaved to yield an enzyme-Mechanism Involving Free Enzyme-Pyrophosphate Intermediatephosphate derivative.
It, seems likely that these enzymes and PRPP synthetase will be found to share chemical and mechanistic properties.
It will also be of interest to learn whether the PRPP synthetase reaction is mechanistically related to other pyrophosphate-transferring enzymes such as thiamine pyrophosphokinase (13) and 7,8-dihydro-2-amino-4-hydroxy&hydroxyethylpteridine pyrophosphokinase (14). The general approach is that of Boyer (15,16). For AMPtional evidence against the two-step mechanism of Equations ATP exchange, as an example, let RI equal the rate of appear-1 and 2.
ante of label from AMP into ATP at equilibrium. R1 = kVl (E.ATP), where (E.ATP), is the concentration of that fraction of E.ATP which is labeled by AMP.  The rate of E -X formation from AMP y k-3 (AMP) (E. The rate expression has the following properties: The relation between R1(,,,) and the maximal velocities clearly depends on (R-5-P).