The Effects of Mevinolin on the Thiol/Disulfide Exchange between 3-Hydroxy-3-methyglutaryl-coenzyme A Reductase and Glutathione*

The feeding of mevinolin plus cholestyramine to rats results in the production of a form of 3-hydroxy-3-methylglutaryl-CoA reductase (HMGR-CM) having thiol/disulfide redox properties different from those of 3-hydroxy-3-methylglutaryl-CoA reductase isolated from animals which had been given only cholestyramine (HMGR-C). The second-order rate constant for the inactivation of HMGR-CM by GSSG is 7-fold slower than for HMGR-C, while the second-order rate constant for the reactivation of oxidized enzyme by GSH is 100-fold slower. However, in the presence of saturating concentrations of both substrates, the rate constants for thiol/disulfide exchange are similar for both forms of the enzyme. HMGR-CM behaves as if a protein-glutathione mixed disulfide having a KO, of 27 f 4 is formed at equilibrium. In contrast, HMGR-C has previously been shown to form a protein-protein disulfide (Cappel, R. E., and Gilbert, H. F. (1988) J. Biol. Chem. 263,12204-12212). Both forms of the enzyme are more difficult to oxidize thermodynamically in the presence of saturating levels of both substrates. For HMGR-CM, NADPH alone has no effect on the equilib- rium constant for oxidation, but hydroxymethylglu-taryl-CoA alone makes the enzyme approximately twice as

including cholesterol, dolichol, and ubiquinone (1-3). The enzyme is a glycosylated integral membrane protein located in the endoplasmic reticulum (4-11). Control of this metabolic pathway is mediated both by alterations in the amount of HMGR present and by changes in the catalytic activity of the enzyme (12). This topic has been extensively reviewed (13).
HMGR is present in low concentration in the cell, less than 0.01% of total cell protein (14). Addition of sterol-sequestering resins such as cholestyramine (Questran) or Amberlite XAD-2 to the diet of mice or rats increases the concentration of HMGR by 3-&fold (15-17). The resin remains in the intestine of the animal and acts by absorbing dietary cholesterol and bile acids, thus depleting the animal of exogenous sterols and inducing the synthesis of HMGR. The discovery of two chemically related fungal metabolites, compactin (18) and mevinolin (19,20), which potently inhibit the activity of HMGR, provided a means by which the level of the enzyme could be further increased. Treatment of rats with both cholestyramine and mevinolin results in the production of significantly higher levels of HMGR than does either agent alone (15,21).
Recent reports in the literature have presented evidence that HMGR isolated from the livers of rats fed both cholestyramine and mevinolin (HMGR-CM) exhibits different properties than the enzyme isolated from the livers of rats fed only cholestyramine (HMGR-C) (4, 15, 21, 22). Both Ness et al. (21) and Roitelman and Schechter (15) have observed that HMGR-C exhibits sigmoidal kinetics with respect to NADPH when assayed in the presence of 5 mM GSH, while HMGR-CM demonstrates normal hyperbolic kinetics under the same conditions. Ness et al. (4) have demonstrated the existence of a disulfide-linked dimeric form of HMGR-C in the absence of thiols. This species could be converted to the monomeric form by the addition of DTT. For HMGR-CM, however, only the monomeric species was observed under both reducing and oxidizing conditions (4). Rogers and Rudney (23) have reported that microsomal HMGR-C incubated in vitro with mevinolin is less reactive with antibody produced against soluble HMGR than is the untreated enzyme. This effect occurs in the absence of a significant loss of enzymatic activity and is not reversed by dialysis.
HMGR is absolutely dependent on the presence of a thiol for activity (21,22,24) and is readily inactivated by disulfides (4,25,26). Because oxidized HMGR is inactive, the possibility that thiol/disulfide exchange, mediated either by GSSG or another cellular disulfide, could serve as a means of metabolic regulation should be considered. The intracellular environment is known to be highly reducing (27-33). Unless the enzyme sulfhydryl group in question can undergo oxidation in this environment and remain oxidized for a significant period of time, regulation of enzymatic activity by thiol/ disulfide exchange would be improbable. Thus, both thermodynamic and kinetic factors must be examined in determining 9180 the susceptibility of a particular enzyme to regulation by thiol/ disulfide exchange. This topic has been recently reviewed (34-36).
The thiol/disulfide redox properties of HMGR-C have been previously reported (26). This enzyme is thermodynamically easier to oxidize than any other mammalian enzyme which has been studied to date. The activity of the enzyme could vary between 35 and 92% in response to normal variations in the thiol/disulfide redox status of the major intracellular redox buffer, glutathione (26). The oxidized species of HMGR-C behaves as a protein-S-S-protein disulfide, rather than a protein-S-S-glutathione mixed disulfide (26). Because significant differences in the redox properties of HMGR-C and HMGR-CM might serve to explain some or all of the observed differences in the behavior of the two forms of the enzyme, experiments were designed to measure the redox properties of the HMGR-CM form of the enzyme. EXPERIMENTAL PROCEDURES*

RESULTS
Kinetics of Thiol/Disulfide Exchange between Glutathione and HMGR-CM in the Absence of Substrate-In the absence of either substrate, rat liver HMGR-CM is totally inactivated by low concentrations of GSSG. As is shown in Fig. 1 (Miniprint), the reaction of the enzyme with GSSG is rapid, monophasic, and first-order in GSSG. The oxidative inactivation is reversible. Incubation of the GSSG-oxidized enzyme with a thiol (DTT or GSH) results in restoration of enzymatic activity. With DTT, the extent of reactivation is >90% of the initial activity. When GSH is used to reduce the enzyme, the extent of reactivation is, within the limits of experimental error, the equilibrium position predicted by the [GSH]/ [GSSG] ratio present in the reaction and the KO, measured for this form of the enzyme (see below). The rate of reactivation by GSH is slow enough to be measured over a GSH concentration range of 2.5-40 mM (Fig. 2, Miniprint). Rate constants are summarized in Table I. Kinetics of ThiollDisulfide Exchange between Glutathione and HMGR-CM under Turnover Conditions-The thiolldisulfide exchange reaction between glutathione and rat liver HMGR-CM occurs even in the presence of saturating levels of both substrates. The rate at which the enzyme is inactivated during the assay could be measured by adding the GSHreduced enzyme to an assay solution containing NADPH, HMG-CoA, and various concentrations of GSSG. When no additional GSSG was added at the time the assay was initiated, the rate of lactone production was linear with time over the 40-min assay period; however, when GSSG was present, the rate of lactone production decreased with time, indicating that the enzyme was being inactivated during the assay (Fig.  3, Miniprint). The apparent first-order rate constant for the loss of enzymatic activity under these conditions was determined by fitting the time course for the formation of lactone to an integrated rate equation describing the effect of enzyme inactivation on the appearance of product (26).
Reactivation of oxidized HMGR-CM by GSH can also occur under turnover conditions. Addition of oxidatively inactivated enzyme to an assay mixture containing various concentrations of GSH resulted in a significant increase in the rate of lactone formation with time ( Fig. 4, Miniprint). A * Portions of this paper (including "Experimental Procedures," Figs. 1-4, and Equation 1) are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are included in the microfilm edition of t,he Journal that is available from Waverly Press. control in which 22 mM DTT was substituted for GSH, was used to define the Vmax. For the DTT control, the rate of lactone production was linear with time, indicating that the time required to fully reduce the enzyme was insignificant compared to the intervals at which time points were taken. When GSH was used to reduce the enzyme, the rate of lactone production increased with time, resulting in a curved assay. Thus, the enzyme was being activated by GSH during the course of the assay. The apparent first-order rate constants for GSH-dependent activation of the enzyme were obtained in a manner analogous to that used for the inactivation in the presence of substrates.
Redox Equilibrium in the Absence of Substrates-The equilibrium constant (K0J for the reaction HMGR-CM,,d + GSSG e HMGR-CMo, + GSH was measured by incubating rat liver HMGR-CM in standard assay buffer containing various [GSH]/[GSSG] ratios and GSH concentrations. Each sample was assayed for enzymatic activity, [GSH], and [GSSG] after equilibrium had been established (>60 min). The rate at which the protein thiols reequilibrate with the redox buffer during the assay was minimized by using a large dilution of the preincubation solution in the assay and a short assay time. In order to determine that the preincubation time was sufficient to allow the system to come to equilibrium, redox curves were constructed by beginning with undialyzed enzyme, which is mostly active, and also with enzyme which has been dialyzed overnight against standard assay buffer and had 6 % activity. The redox curves generated by these two methods were indistinguishable from each other (Fig. 5 ) .
Redox Equilibrium in the Presence of One or Both Substrates-The KO, for rat liver HMGR-CM under turnover conditions was determined in a manner similar to that described above. The enzyme was preincubated with a series of glutathione redox buffers containing various [GSH]/[GSSG] ratios and concentrations of GSH. In order to maximize the rate at which the enzyme could re-equilibrate with the glutathione redox system during the assay, the substrates were added to the preincubation mixture in a minimal volume (1.1fold dilution of the preincubate), the assay time was increased to 60 min, and the concentration of enzyme in the preincubation had to be lowered to ensure that HMG-CoA was not depleted during the assay incubation. Identical results were obtained either starting the assay with fully active HMGR-CM or with HMGR-CM which was brought to redox equilibrium in the preincubation. The KO, for HMGR-CM is significantly lower (enzyme more difficult to oxidize) in the presence of saturating concentrations of both substrates than in 5. Redox equilibrium between HMGR-CM and glutathione in the absence of substrates. HMGR-CM (0.29 mg of protein/ml) was preincubated at 37 "C in 0.1 M phosphate buffer, pH 7.1, containing various concentrations of GSH and GSSG until equilibrium was reached (>75 rnin). The preincubate was diluted 20-fold into standard assay mixtures, and the activity of the enzyme was measured using a 5-min assay incubation time. V, , was determined from a control containing 60 mM DTT in addition to the indicated concentration of GSH. Experiments in which the initial activity of the enzyme was either >70% or less than 5% yielded identical redox curves. A, the data is plotted as a function of R. The curue is drawn according to Equation 7, with a KO. of 27. B, the data is plotted as a function of R. [ Table I).
The redox behavior of HMGR-CM in the presence of a single substrate was also determined. HMGR-CM, briefly preincubated with either 60 W M HMG-CoA or with 3 mM NADPH and the NADPH-regenerating system, was equilibrated with a series of glutathione redox buffers. After equilibration with the glutathione system, an aliquot of the enzyme solution was diluted 20-fold into the usual assay mixture to minimize redox state changes during the 5-min assay. NADPH had no significant effect on the KO, for enzyme inactivation; however, HMG-CoA reduced the KO, by about 50% compared to the enzyme in the absence of substrates (Fig. 7).
Treatment of NADPH-C with Meuinolin in Vitro-The possibility of converting HMGR-C to a form having the redox properties of HMGR-CM was tested by incubating reduced HMGR-C with the sodium salt of mevinolinic acid and measuring the redox properties of this form of the enzyme in the absence of substrates. Fig. 8 shows that the redox behavior of HMGR-C treated with mevinolin in vitro was identical to that of HMGR-CM, having lost the dependence on the absolute concentration of GSH which is characteristic of HMGR-C (26). The specific activity of fully reduced enzyme was significantly lower than that of the untreated HMGR-C form of the enzyme from which it was generated. The V,,, of fully reduced enzyme was decreased to 65 and 27% of its original 3000 O R FIG. 6. Redox equilibrium between HMGR-CM and glutathione in the presence of substrates. HMGR-CM was preincubated at 37 "C in 0.1 M phosphate buffer, pH 7.1, containing various concentrations of GSH and GSSG until equilibrium was reached (275 min) or enzyme was prereduced with 5 mM DTT for at least 60 min. The preincubate was diluted 1.1-fold into standard assay mixtures, and the activity of the enzyme was measured using a 60-min assay incubation time. Both assays contained 18 pg of protein/ml. Ifmax was determined from a control containing 20 mM DTT in addition to the indicated concentration of GSH. Only those samples whose half-life for the approach to equilibrium (calculated according to Equation 1) is less than 6 min were included in the data set. The data is plotted as a function of R. The solid curue is drawn according to Equation 7, with a KO. of   several changes of standard assay buffer. The treated enzyme was then preincubated in 0.1 M phosphate buffer containing various concentrations of GSH and GSSG until equilibrium was reached (>75 min at 37 "C). The preincubate was diluted 20-fold into standard assay mixtures, and the activity of the enzyme was measured using a 5-min assay incubation time. V,,, was determined from a control containing 60 mM DTT in addition to the indicated concentration of GSH. HMGR-CM (A) (1.1 mg of protein/ml) was treated with 100 nM mevinolinic acid in a similar manner as a control. The data are plotted as a function of R. The curve is drawn according to Equation 7, with a KO, of 25.

DO
value by treatment with 10 and 100 nM mevinolin, respectively. HMGR-CM treated in an identical manner showed no changes in either redox behavior or maximum specific activity. Fully reduced HMGR-CM retained >87% of its initial activity.

DISCUSSION
Mechanism of ThiollDisulfide Exchange between HMGR-CM and Glutathione-As with the enzyme isolated from the livers of rats fed only cholestyramine (26), the HMGR-CM form of the enzyme is reversibly inactivated by GSSG oxidation even in the presence of saturating concentrations of both substrates. In the absence of substrates, the HMGR-CM enzyme is inactivated about ?"fold more slowly than the HMGR-C; however, in the presence of saturating levels of both substrates, the inactivation of both the HMGR-CM and HMGR-C enzymes occurs at approximately the same rate. While the presence of the substrate, HMG-CoA, inhibits inactivation of HMGR-C by T-fold, it has only a small effect on the rate of inactivation of HMGR-CM (Table I). With regard to the kinetics of inactivation by GSSG, the HMGR-CM form of the enzyme behaves like the HMGR-C enzyme in the presence of HMG-CoA.
The oxidized enzyme product of the reaction described in Equation 2 might be either a protein-SS-G mixed disulfide (P-SS-G) or a protein-SS-protein disulfide (P-SS-P). The reactions by which the P-SS-G and the P-SS-P species are formed are described in Equations 3 and 4, respectively.
The fraction of the enzyme in the reduced (active) form in the case of the formation of a P-SS-G disulfide and in the case of the formation of a P-SS-P disulfide are described by Equations 7 and 8, respectively (42) where Pt is total protein. In the case of the formation of P-SS-G, the activity is a function of R alone and is independent of [GSH]. Thus, a plot of activity against R will take the form of a single rectangular hyperbola. If a P-SS-P is formed, the activity will be a function of R . [GSH], and a plot of activity as a function of R will take the form of a series of rectangular hyperbolas, one for each fixed concentration of GSH (if the data is replotted as activity uersus R. [GSH], a single rectangular hyperbola will result). For HMGR-CM in the absence of substrates or in the presence of either or both substrates (Table I), the data clearly indicate that the formation of a P-SS-G mixed disulfide can fully account for the behavior of HMGR-CM in glutathione redox buffers (Figs. 5-7).
The presence of NADPH during redox equilibration has no significant effect on the KO,, and HMG-CoA alone only makes the enzyme about 2-fold more difficult to oxidize (lower KO%).
However, the presence of both substrates under turnover conditions makes the enzyme approximately 10-fold less sensitive to oxidation (Table I) Which form of the enzyme is more easily oxidized a t equilibrium, HMGR-C or HMGR-CM? In order to compare two enzymes, one of which is oxidized to a P-SS-G and the other of which is oxidized to a P-SS-P disulfide, a standard GSH concentration must be adopted. The term "R0.5" will be used to refer to the apparent KO, at a given GSH concentration. At equilibrium in 1 mM GSH, HMGR-C is 20-fold easier to oxidize than HMGR-CM. At 100 mM GSH, however, the situation is reversed; HMGR-CM is 5-fold more easily oxidized than HMGR-C. Near 20 mM GSH, the Ro.h values for the two forms of the enzyme are nearly equal.
In Vitro Conversion of HMGR-C to a Form Having the Redox Properties of HMGR-CM-The structural reason for the difference in redox behavior between HMGR-C and HMGR-CM is not yet known. Mevinolin, in the lactone or the diacid form, may either bind directly to HMGR or interact with one or more cellular systems to alter the structure or environment of the enzyme in a manner which does not significantly change its molecular weight. Alterations in the glycosylation state, proteolytic cleavage of a small fragment, or changes in the phospholipid composition of the membrane in the immediate area of the enzyme are possible examples of the latter. However, the observation that in vitro incubation of HMGR-C with mevinolinic acid altered the redox behavior to that shown by the HMGR-CM form of the enzyme favors the argument that the direct binding of mevinolin to a site on the enzyme results in altered enzyme specific activity and changes redox behavior. Similarly, Rogers and Rudney (23) observed that in vitro incubation of HMGR-C with the carboxylic acid form of mevinolin caused a 3-fold decrease in the ability of the enzyme to react with antibody raised against the soluble proteolyzed form of HMGR.
The effect of mevinolin on the redox behavior of HMGR can be produced in a cell-free system and is likely due to a direct interaction of mevinolin with the protein or to some covalent modification which can occur in isolated microsomes. Maximal changes in the reactivity toward antibody (23) and the thiol/disulfide redox properties of the enzyme occur when microsomes are treated with 510 nM mevinolin. At this concentration of mevinolin, the enzyme is <40% inhibited; however, changes in antibody recognition (23) and thiol/disulfide redox behavior are complete. Thus, the effects of mevinolin on the antibody recognition and the thiol/disulfide redox behavior of the enzyme are at least partially separable from the effect on activity, suggesting that mevinolin is capable of binding to more than one site on the enzyme.
With the yeast HMG-CoA reductase, comprehensive kinetic studies of the binding of compactin to the enzyme (45) suggested an interaction which was competitive with HMG-CoA. The rate constant for dissociation of the bound inhibitor from the enzyme was measured to be 6.5 x s-' (half-life of 1.8 min). If the rat liver enzyme behaves similarly, extensive dialysis would be expected to remove the inhibitor. However, even after prolonged dialysis or repeated washing of the microsomal enzyme, the effect of mevinolin on the redox behavior still persists. Thus, there appears to be a site which binds mevinolin very tightly and alters the redox and immunological properties of the protein. The occupancy of this site, however, cannot completely inhibit the enzyme or no activity would be observed.
The molecular basis for the dissimilarity in the properties of the two forms of HMGR remains obscure. The inability of HMGR-CM to form a P-SS-P disulfide could be explained by either a direct blocking of a single sulfhydryl group on the enzyme by bound mevinolin or even a covalent chemical reaction between the lactone form of mevinolin and a protein thiol. The differences in reactivity toward antibody, however, would seem to suggest that the interaction of mevinolin with HMGR results in a more generalized alteration in the tertiary or quaternary structure of the protein. Conformational changes involving the surface of the protein or perturbation of a monomer-dimer equilibrium could provide a plausible explanation for both phenomena.
I n Vitro Consequences of the Redox Properties of HMGR-CM on the Assay of HMGR Activity-Even freshly prepared stock solutions of GSH at neutral pH contain from 0.5-2% GSSG which is formed by autoxidation of GSH. Thus, every glutathione solution actually constitutes a redox buffer. Because the reduction of microsomal HMGR-CM is independent of the total concentration of GSH, at equilibrium HMGR-CM is reduced just as effectively by low GSH as by high GSH. At a [GSH]/[GSSG] ratio of 501, HMGR-CM will exhibit 76% of the DTT activity. Under the same conditions, HMGR-C will be 55% active at 5 mM GSH but 83% active at 20 mM GSH. Also, dilution of HMGR-C into a large volume of assay solution which does not contain GSH or DTT will cause the activity of the enzyme to decrease during the course of the assay (26). For example, if microsomal HMGR-C is preincubated at a total GSH concentration of 50 mM and a [GSH]/ [GSSG] ratio of 50:1, and diluted 10-fold into the assay, the activity would decrease from 71 to 20% of V,,, during the course of the assay. The same treatment would not affect the activity of microsomal HMGR-CM because dilution does not affect the [GSH]/[GSSG] ratio. Thus, smaller redox and activity changes can be expected to occur in the assay of microsomal HMGR-CM than with HMGR-C.
In Vivo Consequences of the Redox Properties of HMGR-&"The fact that the thiol/disulfide exchange reaction between glutathione and both HMGR-CM and HMGR-C (26) can occur at a significant rate in the presence of saturating levels of both substrates suggests that this process might be a viable method for regulation of HMGR activity in vivo. The active site of HMGR faces the cytosolic side of the endoplasmic reticulum. Neither the absolute concentration of GSH nor the [GSH]/[GSSG] ratio of the cytosolic compartment is known with certainty (32, 35, 36); however, the average concentrations of GSH and GSSG in a liver cell under a variety of metabolic conditions have been measured by several methods (26, 27, 32, 35, 46-48). At physiological levels of GSH (2-10 mM), HMGR-CM will be thermodynamically easier to reduce than HMGR-C. For example, under conditions of GSH = 3 mM and GSSG = 30 p M (typical of a fasted rat liver), HMGR-CM would be 79% reduced at equilibrium while HMGR-C would be only 35% reduced. Based on the in vitro data, HMGR activity in liver cells of cholestyramine-plus mevinolin-fed rats could be expected to be more resistant to inhibition by sulfhydryl oxidation than the enzyme from animals which had been fed only cholestyramine. Concenlratlon 01 GSH was measured by HPLC at the beginning and end 01 the sxpenmenl. as described rn the precedlng s8c1ton Redox Equilibrium In the Absencs 01 SubsVstes. HMGR-CM was we-Incubated in standard assay buffer wnh a 58nes of glu?mhione redox bvners wmalning GSH and GSSG at ratlos varying between 0 and 1 w and at GSH CDnCLKltratlonS ranging bePNeen 5 and 110 mM Controls lo determine Vma ymained MI mM D T l in addtlcn 10 the nndhsted cansntratiDn of GSH. The concentration of protean on the pre-incubatlan soiuth was 0 29 mo/ml. ARer an equihbratton of at least 60 mn in the redox buffer a 37'C. an aliqum of the pre-Incuballon mixture was diluted 2(Floid tnto the assay solullon. and the an~vty of the enzyme was determlned mng a fiveminute incubation penad Another alfquot was aaddled by the addmon of HCI and analyzed l o r GSH and GSSG. The large dtlubon factor and Shon assay tlme Sewed to i w e r tne cancentrations d GSH and GSSG I" the assay, thereby decreasing the rate at which the enzyme wwld re-equlbrate with the redox buffer during the assay The halllife for the approach to equ#i,bwm m the assay was calculated for each redox polnt based on the measured m the presence of saturatmg iev~ls of M t h substrates. assumlng.
wncentrat~ons 01 GSH and GSSG and the rate wnsfants for m e reducl~on and Oxldatlon of the enzyme 'obs = ~GSSG IGsGIassay + ~G S H * IGSHlas=.ay 11) detsrmme that the pre-lncubatlon tlme was sut(icrent to allow the system to wme to equdlbnum, redox Only those samples having a han-lle in the assay of 2.25 mln were included m the data set. In order to CUNBS were wnstructed by begnnnlng wlth undtalyzed enzyme. whid IS mostly activa. and also with enzyme which had been dlalyzed overnight aga8nst standard assay buffer and had < 5 % activtry The same BqUlllbrium pOSLlOn was achieved regardless of the lnibal oxidation state of the enzyme.
Redox Equiilbrlum In tho Presence 01 e Single Subsbats. These exprlments were carried out m the pre-lnarbatm minure. In me experlments In Whld the enzyme was pre-Incubated Wlth NADPH, as descrlbed ~n the previous section, except that enher 60 pM HMG-GOA or 3 mM NADPH was included 30 mM gluc0~~-6phosphate and 0 4 ulml of gluwseb-phosphate dehydrogenase were also included In the prsmwbatton solutim to prevent the destruction of NADPH and the accumulation of NADP. The assay mixiures weis prepared vsng the usual concentration 01 both substrates, giUwsB-S~QhoSphate.
the assay, the final wncentratlon of the Substrate Whtch was present In the pre-8ncubation was only 5 % and glucose.6phosphate dehydrogenase Because the premwbation Solut~on was diluted 2Dfold Into greater than m the usual assay system.
Redox Equllibrlum in the Presenc. 01 Both Substratas. Giutath~one redox buffers were prepared as descrlbed above. The enzyme was premcubaled al37' either wlth 5 mM DTT lor s60 mln aCtlVe enzyme). The DlTcontajning enzyme was dlluted 3w-fold mto the assay minure to minimize the (fully actlve enzyme) or the glutathione redox M e r in me absa-of Substrates for >90 mm (panlally carry over of D l T To maXlmlze the rate at whuh the enzyme would re-equlllbrate Wnh the redox buffer during the assay, me dilution d GSH and GSSG lmo the assay was minimized (1.1-fold) and the assay tlme was lncreared to MI minutes The final wncwntration of enzyme m the assay was I 8 pdml. To permit the immdwvon of the substrates In a mbnmal volume. the gluwse-Sphosphate requlted for the NADPH regensralng system was added to the pre-mCubaUon buffer.