Enzymatic synthesis of cytidine diphosphate 3,6-dideoxyhexoses. 8. Studies of the properties of E3 and its role in the formation of cytidine diphosphate-4-keto-3,6-dideoxyglucose.

Abstract Enzyme E3, one of the two enzymes required to promote the formation of CDP-4-keto-3,6-dideoxyglucose from CDP-4-keto-6-deoxyglucose, is thought to catalyze the NADPH requiring reduction in which the 3-deoxy group is actually formed from the intermediate discussed in the preceding paper. One mole of NAD(P)H was oxidized for every mole of sugar reduced when the reaction was carried out in an argon atmosphere. Experiments with both A- and B-[4-3H]NADPH have shown that there is no direct hydride transfer from NADPH to either the sugar being reduced or to pyridoxamine 5'-phosphate, a participant in the reaction. On the other hand, using tritiated water 1 atom of hydrogen is incorporated into the CDP-4-keto-3,6-dideoxyglucose formed. E3 alone also catalyzes the oxidation of NAD(P)H although there is no chromophoric group such as a flavin associated with the enzyme. However, added FMN or 2,6-dichlorophenolindophenol greatly enhances the rate of NAD(P)H oxidation by E3.


SUMMARY
Enzyme E3, one of the two enzymes required to promote the formation of CDP-4-keto-3,6-dideoxyglucose from CDP-4-keto-6-deoxyglucose, is thought to catalyze the NADPH requiring reduction in which the 3-deoxy group is actually formed from the intermediate discussed in the preceding paper.
One mole of NAD(P)H was oxidized for every mole of sugar reduced when the reaction was carried out in an argon atmosphere.

Experiments with both A-and B-[4-3H]NADPH
have shown that there is no direct hydride transfer from NADPH to either the sugar being reduced or to pyridoxamine 5'-phosphate, a participant in the reaction. On the other hand, using tritiated water 1 atom of hydrogen is incorporated into the CDP-4-keto-3,6-dideoxyglucose formed.
E3 alone also catalyzes the oxidation of NAD(P)H although there is no chromophoric group such as a flavin associated with the enzyme. However, added FMN or 2,6-dichlorophenolindophenol greatly enhances the rate of NAD(P)H oxidation by E3.
The formation of CDP-4-keto-3,6-dideoxyglucose from CDP-4-keto- 6-deoxyglucose in Pasteurella pseudotuberculosis type V is catalyzed by two enzymes, El and E3 (l), both of which have been purified to homogeneity (2). NAD(P)H is the reducing agent, and pyridoxamine 5'-phosphate is a required cofactor (3). In the preceding paper, it was shown that the probable role of El in the reaction is to catalyze the formation of an enzyme bound Schiff base-A-3,4-glucoseen adduct between pyridoxamine 5'-phosphate and CDP-4-keto-6-deoxyglucose. The intermediate was termed El .pyridoxamine 5'-phosphate .X (see Fig. 10).
Previous work (2) had revealed that E3, by itself, had the Francisco, California 94122. ability to oxidize NAD(P)H, and it was therefore suggested that the physiological role of the enzyme is to carry out the actual reductive part of the reaction producing CDPI-keto-3,6-dideoxyglucose.
In the present paper, experiments are presented which attempt to define more completely, the mechanism of action of E3. Studies designed to localize the actual site and mechanism of reduction are described and the properties of E3 as a NAD(P)H oxidase are examined. NADP was produced in the reaction when A-[4-3H]NADPH was used as would be expected.
This lack of stereospecificity toward NADPH exhibited by E3 is shared by only an extremely small number of enzymes, one of which is lipoyl dehydrogenase ("diaphorase"), a flavoprotein (6).

Incorporation
of Tritium from 3Hz0 into CDP-4-keto-3,6dideoxyglucose during Its Formation-In the presence of tritiated water, 18 Ci per mmole, there was significant incorporation of tritium from the solvent into CDP-4-keto-3,6-dideoxyglucose (Fig. 4). A total of 41,000 cpm of tritiated product was formed or 205,000 dpm, 0.093 $Zi of 3H (counting efficiency was 20%). In the reaction, 43 nmoles of product were formed; the specific activity of the product was thus 2.2 mCi of 3H per mmole of sugar. Since the specific activity of the solvent hydrogens was 9 Ci per mole atom, 0.24 mole of tritium was incorporated per mole of sugar. Assuming a solvent isotope effect of 3 to 4, this value indicates that 1 mole of solvent hydrogen was incorporated per mole of sugar reduced. No such incorporation was evident when E3 was eliminated from the reaction mixture.
This control would seem to rule out the possibility that nonspecific tritium exchange was occurring with the Cs proton of the sugar or with the original C& proton.
It thus seems that the hydrogen introduced at C& by the reductive part of the reaction is derived from the solvent.
It was previously reported (7) that tritium was not incorporated from solvent into the sugar. This discrepancy might result from the difference in experimental procedures used in purifying the GDP-4-keto-3,6-dideoxyglucose for tritium analysis.
Attempt to Demonstrate Direct Hydride Transfer jrom NADPH to Cd-Methylene Carbon Atom of Pyridoxamine 5'-Phosphate-An alternative reduction mechanism is the reduction by NADPH at the met.hylene carbon at Cq of pyridoxamine 5'-phosphate while it is part of the pyridoxamine 5'-phosphate-Schiff base-A-3,4-glucoseen intermediate (see Fig. 10, Part II).
In an experiment designed to test this possibility (see "Materials and Methods"), El and CDP-4-keto-6-deoxyglucose were present in stoichiometric amounts so that each reaction center would undergo no more than one turnover; pyridoxamine 5'-phosphate was present in a loo-fold excess. When the pyridoxamine 5'phosphate was isolated following the reaction, the total radioactivity isolated with pyridosamine 5'.phosphate was 3,527 cpm of 3H (this value is found after subtracting a blank of 6,084 cpm obtained in the absence of El from a value of 9,954 cpm obtained with the complete reaction mixture).
The value of 3,527 cpm is only 2.5y0 of the amount expected, 142,000 cpm (counting efficiency = 25a/,), if direct hydride transfer had occurred at pyridoxamine 5'.phosphate during the reaction. The expected value takes into consideration a hypothetical isotope effect of 100 representing a factor of 10 from the transfer of tritium from [l-3H]glucose 6-phosphate to NADP and another factor of 10 for the transfer of tritium from NADPH to the pyridoxamine 5'-phosphate.
Since the ratio of NADP to CDP-4-keto-6deoxyglucose was 1:5, tritium would be expected to be transferred from glucose 6-phosphate to both sides of the dihydropyridine ring of NADPH after the first 20% of the over-all reaction had occurred.
Attempt to Trap Covalent ES-Pyridoxamine 5'-Phosphate-Substrate Complex-If the active sulfhydryl group of E3 (2) was added covalently to the Ca position of the sugar by a Michaeladdition while it was involved in the Schiff base-A-3,4-glucoseen intermediate (Fig. 10, Part I; B is the sulfhydryl), addition of NaBH4 and omission of NADPH from the reaction mixture might result in the reduction of the Schiff base by the borohydride, thus causing an irreversible bond formation between the sulfhydryl group and Cp. When such an experiment was attempted (see "Materials and Methods"), using r4C-labeled substrate, the radioactivity found in the total reaction mixture was no greater than that found when either E3 or NaBH4 was omitted from the reaction.

Properties of Enzyme ES
Coincidence of Sugar Reductase and NA DH Oxidase Activities-E3 (Step 6) was subjected to DEAE-cellulose chromatography at pH 5.8 as previously described (2). Aliquots of each fraction were tested for NADH oxidase and sugar reductase activities, and the results are presented in Fig. 5. The two activities were found to be coincidental on the column.
Effect of NEM on NADH Oxidase Activity of ES-Preincubation of E3 with 0.02 M NEM resulted in the loss of 85% of the NADH oxidase activity of the enzyme (Fig. 6). Previous work (7) has shown that under the same conditions used here, the sugar reductase activity of the enzyme is inhibited to the same 3785 Effect of Di$erent Electron Acceptors on NADH Oxidase Activity of Enzyme ES-Using a crude cell extract of a species of Salmonella, Nikaido and Nikaido (8) reported that the addition of FAD to the incubation mixture produced a stimulation of CDP-3,6-dideoxyhexose formation when GDP-glucose was used as the substrate.
Similarly, using a crude enzyme system from P. pseudotuberculosis stimulation of the formation of CDP-4-keto-3,6-dideoxyglucose by the addition of FAD to the reaction mixture; they further reported that there was no FAD-induced stimulation of the conversion of CDP-4-keto-3,6-dideoxyglucose to CDP-3,6-dideoxyhexose. In the present work, using highly purified enzymes, both FAD and FMN markedly inhibited the formation of CDP-4-keto-3,6-dideoxyglucose from CDP-4-keto-6-deoxyglucose when added in concentrations of low4 M and higher (Table I). Perhaps in the crude system, the flavin acts by inhibiting some enzyme activity that would normally degrade the nucleotide sugar substrate.
In any case, using purified enzymes, there is no flavin-catalyzed stimulation of the El, E3 system. In an effort to explain the inhibition of sugar reduction by flavin coenzymes, it was hypothesized that they may be acting by binding to E3 and uncoupling NADH oxidation from sugar reduction.
To test this possibility, the effect of various electron acceptors on the E3-catalyzed rate of NADH oxidation was examined.
Both FMN and DCPIP could be utilized by the enzyme as electron acceptors (Table II); on the other hand, cytochrome c and triphenyltetrazolium chloride were ineffective in this role. When the mode of action of FMN catalysis was examined in greater detail (Fig. 7), it was demonstrated that the Gonzalez-Porque and Strominger (2) had reported previously that addition of a cofactor solution isolated from a cell estract and containing pyridosamine 5'-phosphate resulted in a 2-fold stimulat'ion of the rate of NADH oxidation by E3. In the present study, however, when a pure commercial preparation of the coenzyme was used, no stimulation of NADH oxidation was observed. Based on the results found here, the stimulation by cofactor solution previously rcport,ed was probably due to the presence of a small amount of some electron acceptor in the solution.
Since FMN was shown to bind to the enzyme, a spectrum was taken of a mixture of FMN and E3 in an argon atmosphere.
No decrease in the flavin peak at 450 nm was noted, however.
Since E3 has a single active sulfhydryl group (2), it was thought that this residue might form a covalent bond with the flavin according to the mechanism suggested by Brown and Hamilton (9). The formation of a covalent adduct at the Cla or Ns position of the flavin ring has often resulted in the disappearance of the 450.nm peak (10).

Inhibition of Sugar
Reduction by o-Phenanthroline-As seer1 in Table III, inclusion of 2 mM o-phenantflhroline in the reaction mixture caused an 85% inhibition of sugar reduction. 8-Hydroxyquinoline-5-sulfonic acid at the same concentration i!ihibited the reaction as well but to a much smaller extent.
When three small compounds known for their ability to complex with the IIB metals (II), cyanide, azide, and sulfide all at 10 mM concentrations, were tried as inhibitors, no inhibition was detected. A concentration curve for o-phenanthroline inhibition of sugar reduction is shown in Fig. 8 3787 the experiment presented, 50y0 inhibition occurred at about 7 x lo-' hr. Attempts were made to locate the site of inhibition by o-phenanthroline.
As shown in Fig. 9, this compound had libtle if any effect on the rate of NADH oxidation by E3. Furthermore, attempts to inactivate either El or E3 by dialysis against o-phenanthroline were unsuccessful. Finally, of five different El plus E3 preparations used in this study, one was found which, when used in the production of CDP-4-keto-3,6- The rhr!ltor was added to the reaction mixture as indicated.
Rest:;. ,rc expressed as per cent activity remaining relative to that :~c~riicved in the absence of any chelator. All inhibitors were present at a concentration of 2 rnM, The mechanism of reduction of the sugar remains only partially solved. As shown in this paper, only 1 mole of NADH is required to reduce a mole of CDP-4-keto-6-deoxyglucose.
Using tritium transfer techniques, direct hydride transfer from [3H]-NADPH to the sugar could not be demonstrated. If such a transfer did occur, there are two reasons why it might remain undetected.
First, the presence of an unusually large kinetic isotope effect might cause the total incorporation to be so small that it would remain unnoticed under the conditions of the experiment.
Second, as seen in Fig. 10, after the hypothetical addition of a hydride ion to Cs of the sugar had occurred, a structure involving a Schiff base adjacent to the protons in question would remain; this Schiff base, because of its property as an electron sink, might then facilitate exchange of the Ca protons with the solvent before hydrolysis of the Schiff base occurred. This would imply that if the reaction were carried out in tritiated water, tritium should be incorporated from the solvent into the sugar. In the preceding paper, it was demonstrated that the C) proton originally on the substrate remained attached to the product after the reduction had occurred.
This finding does not entirely eliminate the exchange mechanism just proposed since the proton introduced into the sugar by the reduction might be in a more favorable steric configuration to undergo exchange than the proton originally on the substrate. 8. Effect of o-phenanthroline concentration on the formation of CDP-4-keto-3,6-dideoxyglucose. tion of CDP-4-keto-3,6-dideoxyglucose.
The formation of CDP-The formation of CDP-4-keto-3,6-dideoxyglucose was assayed as outlined under "Ma-4-keto-3,6-dideoxyglucose was assayed as outlined under "Materials and Methods." o-Phenanthroline was added to the terials and Methods." o-Phenanthroline was added to the reaction mixture as indicated. reaction mixture as indicated.
Results are expressed as counts Results are expressed as counts per min of product formed. per min of product formed. 9. Effect of o-phenanthroline on the NAD(P)H oxidase activity of E3. The NAD(P)H oxidase activity of E3 in the presence of 2 mM o-phenanthroline was assayed as described under "Materials and Methods." FMN was omitted from the mixture. An identical experiment was performed in which o-phenanthroline was omitted. dideoxyglucose, was not subject to inhibition by o-phenanthroline. Although o-phenanthroline inhibition is usually thought to indicate the presence of a metal in the system, the failure of other metal chelators to inhibit the El, E3 system, in addition to the failure of dialysis against o-phenanthroline to inhibit suggests that o-phenanthroline may be working in the sugar reduction system by some chaotropic or competitive binding mechanism rather than by metal complexat.ion.
The result shown in Fig. 4 clearly indicates that tritium is incorporated from the solvent into the sugar during reduction. If