Biosynthesis of Uridine Diphosphate D-Xylose

laurentii wheat germ enzymes, respectively, was observed with the C-4-labeled substrate but not with the C-3- or C-5-labeled substrate. Labeled UDP-D- xylose obtained from UDP-D-glucuronic acid-5T was con- verted with phosphodiesterase to labeled oc-D-xylosyl phos- phate. The latter was oxidized with periodic acid to yield D-phosphodiglycohc aldehyde from C-l and C-2 and from C-4 and C-5; hypobromite oxidation of the dialdehyde yielded

UDP-glucuronate carboxy-lyase (EC 4.1.1.35) catalyzes the irreversible decarboxylation of uridine 5'-(cr-n-glucopyranosyl-* This work was supported by Research Grant. GM 08820 from the National Institute of General Medical Sciences, National Institutes of Health. A preliminary report has been presented (1). For the preceding paper of this series see Reference 2.
1 Present address, Department of Biochemistry, Marquette School of Medicine. Milwaukee. Wisconsin 53233.
The enzyme has been purified from Cryptococcus laurentii (3) and wheat germ (4) and has been found to require catalytic amounts of NAD for activity; either enzymebound in the case of the wheat germ enzyme or as an added cofactor for the C. laurentii enzyme. Ankel and Feingold proposed that uridine 5'-(oc-n-xylo-hexopyranosyluronic acid-4ulose-pyrophosphate) would be a likely intermediate in the reaction (3). This would make the first step in the decarboxylation reaction analogous to the 4-epimerization of UDP-glucose catalyzed by UDP-glucose 4-epimerase (EC 5.1.3.2) (5) except that the intermediate, UDP-4-keto-GlcUA,' would be a P-keto acid susceptible to decarboxylation.
The enzymatic epimerization of UDP-Glc-4T has been shown to occur with complete retention of label at C-4 even though a marked inverse isotope effect occurs during the reaction (6). Ankel and Feingold demonstrated that label is retained also during the decarboxylation of UDP-GlcUA-4T; however, they did not measure the effect of isotope on rate or determine the position of label in the product (3, 4). Loewus (7) showed that myo-inositol-2T is converted with inversion of configuration by ripening strawberries to n-xylose-5T. He suggested that UDP-GlcUA and UDP-Xyl might play a role in this conversion, although other possibilities were considered also.
In this paper we show that there is a kinetic isotope effect when UDP-GlcUA-4T is converted to UDP-Xyl by UDPglucuronate carboxy-lyase from either C. Zuurentii or wheat germ. Furthermore, UDP-GlcUA-5T, the absolute configuration of which is S at C-5, is converted by the decarboxylation to UDP-Xyl-5T with the absolute configuration R at C-5.
UDP-Xyl-5T,,b and glycolic acid-T,,b refer to tritiated Materials UDP-glucuronate carboxy-lyase was purified from wheat germ and from C. laurentii according to the procedures of Ankel and Feingold (3,4). The specific activities of the wheat germ and the C. Zaurentii enzymes were 0.34 and 0.18 unit per mg, respectively, where 1 unit is defined as the amount of enzyme which catalyzes the formation of 1 pmole of UDP-Xyl per min at 37". Glucose-32' and glucose-57' were gifts from Dr. J. Katz, Institute for Medical Research, Cedars-Sinai Hospital, Los Angeles, California.
Electrophoresis was carried out on Whatman No. 1 or 3MM filter paper in the solvent-cooled Gilson Medical Electronics model D high voltage electrophorator.
The buffer systems employed were 0.1 M ammonium formate, pH 3.6, or 0.1 M ammonium acetate, pH 5.8 (10). Compounds were detected on paper with the use of the following spray reagents: p-anisidine phosphate (10) or ammoniacal silver nitrate (11) for free sugars; molybdic acid (12) for organic phosphates; and a saturated aqueous solution of ammonium vanadate or the potassium ferrocyanide reagent of Stern et al. (13) for organic acids. Nucleotides and nucleotide sugars were detected by visual inspection under short wave (254 rnp) ultraviolet light.
Radioactive compounds were located with the use of a Tracerlab 4-7r paper strip scanner or by autoradiography with Eastman Kodak No-screen or Blue Brand x-ray film.
Analytical Procedures-Xylose was determined by a modification of the method of Roe and Rice (14) as previously described (15), glycolic acid with 2,7-dihydroxynaphthalene in concentrated sulfuric acid (16), and glyoxalic acid with phenylhydrazine hydrochloride by the procedure of McFadden and Howes (16), except that the total volume was reduced to 2.4 ml. Acidlabile organic phosphate (released by hydrolysis in 0.01 N HCl at 100" for 15 min) was determined by the method of Ames and Dubin (17). Radioactivity measurements were carried out in a Packard Tri-Carb model 3375 scintillation spectrometer in either a toluene scintillation mixture containing 5 g of 2,5diphenyloxazole (PPO) and 100 mg of 1,4-bis[2-(4-methyl-5phenyloxazolyl)]benzene (dimethylPOPOP) per liter or in the dioxane base scintillation mixture described by Bray (18). Quenching caused by color in some of the samples was corrected for with the use of internal standards and with the use of the automatic external standard.
Preparation of Speci$cally Tritiated UDP-GlcUA-Specifically tritiated n-glucose was converted to UDP-Glc with an enzyme preparation from Bakers' yeast according to the procedure of Wright and Robbins (19). UDP-GlcUA-3T, UDP-GlcUA-4T, and UDP-GlcUA-5T were prepared from the correspondingly tritiated UDP-Glc with UDP-Glc dehydrogenase in the presence of a large excess of NAD in 0.1 M Tris-KC1 buffer, pH 8.0. Tritiated UDP-GlcUA was purified by chromatography in Solvent A, followed by high voltage electrophoresis at pH 5.8.
Enzyme Assays-UDP-GlcUA carboxy-lyase was routinely assayed during purification as previously described (4); however, in order to measure the effect of isotope on the rate of decarboxylation, a different type of assay was used. Ninety microliters of 0.1 M sodium-potassium phosphate buffer, pH 7.0, containing 0.5 g of EDTA per liter (Buffer A) and tritiated UDP-GlcUA and UDP-GlcUA uniformly labeled with 14C at a final concentration of 1 to 5 X 10m4 M was incubated at 30". Reaction mixtures for the C. Zaurentii enzyme contained, in addition, NAD at a final concentration of 1 mM. Either before or immediately after the reaction was started by addition of 20 ~1 of suitably diluted enzyme, 10: ~1 of the reaction mixture were sampled. At appropriate time intervals, lo-p1 portions of the reaction mixture were removed in capillary tubes which then were sealed and placed into boiling water for 1 min. The contents of the capillaries were spotted on filter paper and the reaction product was separated from remaining substrate by chromatography in Solvent A. Those areas which cochromatographed with authentic unlabeled UDP-Xyl or UDP-GlcUA were excised and counted in the toluene base scintillation mixture.
The aH:% ratio was determined both in the product and the remaining substrate during the course of the reaction.
This was done to determine the rate of decarboxylation of 14C-labeled and 3H substrate independently and simultaneously in the same reaction mixture. Reaction rates were expressed as percentage of conversion of UDP-GlcUA to UDP-Xyl (UDP-Xyl uniformly labeled with 14C counts were corrected for loss of 14C02).
Reaction mixtures containing UDP-GlcUA (5 to 10 X lo4 cpm) (and 1 pmole of NAD when the C. Zaurentii enzyme was used) and enzyme (0.1 unit) in 1 ml of Buffer A were incubated for 4 hours. The reaction was stopped by heating at 100" for 3 min. Denatured protein was removed by centrifugation and tritiated UDP-Xyl was isolated by chromatography in Solvent A. UDP-Xyl-5T,,d was prepared by carrying out the decarboxylation of unlabeled UDP-GlcUA in tritiated water (100 mCi per g) with the wheat germ enzyme. A reaction mixture containing UDP-GlcUA (4 pmoles), in 0.5 ml of Buffer A was lyophilized; the water was replaced with tritiated water and wheat germ enzyme (0.4 unit in 25 ~1 of Buffer A) was added to start the reaction. After 2 hours at 25", the reaction mixture was lyophilized and UDP-Xyl-5T,,d was isolated by chromatography in Solvent A. Identical reaction mixtures were set up in which UDP-GlcUA was replaced by UDP-Xyl.
Degradation of UDP-Xyl and Isolation of Carbon-bound Hydrogen at C-1 and C-5-UDP-Xyl ( Fig. 1)  UDP-Xyl was degraded as described under "Methods" to allow the isolation of those hydrogen atoms bound at C-l and C-5 of the D-xylosyl moiety. The hydrogen at C-2 is lost during formation of osazone and hydrogen atoms at C-l and C-3 are isolated after periodate oxidation of the osazone to yield crystalline mesoxalaldehyde 1,2-bis (phenylhydraeone). was isolated by electrophoresis at pH 3.6. The xylose-1-P was then degraded on a microscale by a modification of the procedures of Boothroyd et al. (20) and Bevill et al. (6), to yield glyoxalic acid (V) from C-l and C-2 and glycolic acid (VZ) from C-4 and C-5. The degradation involves periodate oxidation to yield formic acid from C-3 and n-phosphodiglycolic aldehyde (III) from C-l and C-2 and from C-4 and C-5, followed by hypobromite oxidation of the dialdehyde to yield n-phosphodiglycolic acid (IV).

I II III
Acid hydrolysis of IV liberates glyoxalic acid (V) and glycolic acid (VI).
The following describes a typical degradation of xylose-1-P uniformly labeled with 1%. Sixty microliters of 1 M periodic acid (60 pmoles) were added to carrier-diluted xylose-1-P uniformly labeled with r4C (22.7 pmoles, 10,700 cpm per pmole per 2-carbon fragment) and the reaction mixture was incubated overnight in a covered vessel at 25". The solution was neutralized with 10% ammonium hydroxide and diluted with water to a total volume of 5 ml. Strontium carbonate (400 mg) and Brz (30 ~1) were added and the mixture was stirred for 30 min, then aerated for 30 min, and filtered through a fine sintered glass funnel.
The filtrate was taken to drvness at 37", dissolved in 2 ml of 1 N HCl and held at 100" for 2 hours, neutralized with silver carbonate, and filtered as before. The filtrate was al)plied to a column (0.7 X 10 cm) of Dowex l-Xl0 (ZOO to 600 mesh, acetate form, previously washed with 4 volumes of 6 s acetic acid and then with water), washed with 2 bed volumes of water, and eluted with 4 M acetic acid (15). Glycolic acid (4.07 pmoles, 10,800 cpm per pmole) was eluted in Fractions 10 to 13 (4.4 ml) and glyoxalic acid (4.26 pmoles, 12,250 cpm per Mmole) was eluted in Fractions 16 to 21 (10.8 ml). Acetic acid was removed by vacuum evaporation and the free acids were dissolved in water.
Glycolic acid isolated by this procedure cochromatographed with authentic nonlabeled glycolic acid in Solvents C and D and coelectrophoresed at pH 3.6. Identification of Carbon-bound Hydrogen at C-2 and C-3 (see Fig. 2)-Labeled UDP-Xyl was hydrolyzed in 0.01 N HCl at 100" for 15 min. The labeled n-xylose (I) was isolated by chromatography in Solvent 13, diluted with unlabeled n-xylose and converted to the osazone (II) as follows.
n-Xylose-42' (144 pmoles), redistilled phenylhydrazine (477 pmoles), and glacial acetic acid (764 pmoles) were dissolved in water to a total volume of 1 ml and heated at 90" for 30 min. The reaction mixture was cooled in ice, the osazone was recovered by filtration, and recrystallized twice from 60% ethanol.
The concentration of the osazone was determined from the optical density at 395 rnp based on the molar absorptivity of the authentic compound. Aliquots of the osazone solution were counted in Bray's solution; the counts were corrected for the quenching caused by the yellow color of the osazone.
The orange precipitate which immediately formed was recovered by filtration, washed with 66% ethanol, and dissolved in absolute ethanol.
The spectrum of the isolated product was compared with authentic mesoxalaldehyde 1,2-bis(phenylhydrazone) (22) and the concentration was determined from the optical density at 420 mp. Aliquots of the orange solution were counted in Bray's solution and the counts were corrected for quenching.

RESULTS
Effect oj Isotope on Rate of Decarboxylation-When the rate of decarboxylation of each of the trit.iated substrates was measured, only UDP-GlcUA-4T was found to show an isotope effect. In Tables I and II  occurs with complete retention of label (Table IV). UDP-GlcUA-42' reacts at 327, of the rate (VT/V, = 0.32) of nonlabeled UDP-GlcUA (measured as the rate of decarboxylation of UDP-GlcUA uniformly labeled with r4C) when decarboxylation is carried out w-it11 the C. laurentii enzyme and at 42y0 of the rate (VT/V, = 0.42) when the reaction is carried out with the wheat germ enzyme.
In contrast, the decarboxylation of UDP-GlcUA-3T and UDP-GlcUA-52' occurs without an isotope effect (VT/V, = 1.0) (Table III).  Location of Label in UDP-Xyl-In order to determine the location of the label in the reaction product, UDP-Xyl was degraded by two different procedures.
The first procedure involved periodate oxidation followed by bromine oxidation of the tritiated xylose-1-P derived from UDP-Xyl (Fig. 1). This procedure permitted the isolation and identification of 3H bound at either C-l or C-5 of the D-xylosyl moiety.
Glyoxalic acid, con-VTIVH UDP-GlcUA-3T taining that hydrogen originally bound at C-l, is derived from C-l and C-2, and glycolic acid, containing those hydrogen atoms originally present at C-5 is derived from C-4 and C-5 of the Dxylosyl moiety.
The second degradation procedure involved acid hydrolysis of UDP-Xyl to yield the free sugar, formation of the osazone of xylose, and periodate oxidation of the osaaone (Fig. 2). That hydrogen atom originally present at C-2 is lost during formation of the osazone, and after periodate oxidation, those hydrogen atoms originally bound at C-l and C-3 are isolated as the phenylhydrazone of mesoxalaldehyde.
The specific activity of xylose-1-P-T derived from the decarboxylation product of UDP-GlcUA-42' (wheat germ enzyme) was 1400 cpm per Mmole; however, glyoxalic acid and glycolic acid obtained from it contained less than 1 y0 of the label originally present.
On the other hand, the specific activity of xylose-1-P-T derived from the decarboxylation product of UDP-GlcUA-52' (wheat germ enzyme) was 2300 cpm per pmole, that of the glycolic acid derived from it was 2600 cpm per pmole, while the glyoxalic acid contained less than 2y0 of the original activity.  These data show that there is no migration of label from C-4 to either C-l or C-5 during the decarboxylation. n-three-Pentose phenylosazone-T (ZZ) (Fig. 2) prepared from the decarboxylation product of UDP-GlcUA-3T (wheat germ enzyme) had a specific activity of 2600 cpm per pmole, the mesoxaldehyde 1,2-bis(phenylhydra.zone) (III) produced by periodate oxidation had a specific activity of 2800 cpm per pmole. This shows the efficacy of the periodate oxidation in distinguishing between C-l through C-3 and C-4 and C-5, and suggests that there is no migration of label from C-3 during the decarboxylation.
The specific activity of n-xylose-2' (Fig. 2  (Z)) obtained from UDP-GlcUA-4T (355 cpm per pmole) was essentially identical with that of the n-three-pentose phenylhydrazone (ZZ) (322 cpm per pmole), showing that no migration of label to C-2 had occurred during the decarboxylation.
However, the mesoxaldehyde 1,2-bis(phenylhydrazone) (Ill) was unlabeled, showing that there was no label at C-l or C-3 of the D-xylosyl moiety of the UDP-Xyl. Hence, by difference, all the label initially present at C-4 of the substrate (UDP-GlcUA4T) is present at C-4 of the decarboxylation product (UDP-Xyl-4T). Absolute Conjguration at C-5 of UDP-Xyl-The absolute configuration at C-5 of the D-xylosyl moiety of UDP-Xyl-ST was determined as follows.
UDP-Xyl-5T,,b (obtained by the decarboxylation of UDP-GlcUA-5T with enzyme from C. Zaurentii as well as from wheat germ) was degraded as before and glycolic acid-T,,b was isolated (Table V). In addition, UDP-Xyl-5T,,d (prepared by the decarboxylation of UDP-GlcUA in tritiated water with wheat germ enzyme) was degraded to yield glycolic acid-T,,d which contained that 3H atom originally incorporated at C-5 of the D-xylosyl moiety (Table V).
This means that during decarboxylation, only one tritium atom is incorporated into UDP-Xyl at C-5 of the D-xylosyl moiety.
UDP-Xyl does not become labeled when it is incubated in tritiated water in the presence of wheat germ enzyme.
The degradation procedure employed to isolate glycolic acid from C-4 and C-5 ( Fig. 1) does not cause inversion of configuration at C-5, so that the a-carbon of glycolic acid-T must have the same absolute configuration as the C-5 of the D-xylosyl moiety from which it was derived.
Glycolic acid-T isolated from UDP-Xyl-5T was oxidized with glycolic oxidase, which is specific for that hydrogen atom in glycolic acid which is sterically equivalent to the a-hydrogen of L-lactic acid (22). Glycolic acid-T,,b, which is derived from the degradation of UDP-Xyl-5T,,b, loses all its label upon oxidation with glycolic oxidase (Table VI). This shows that the absolute configuration at the a-carbon of glycolic acid-T,,b is R, using the nomenclature of Cahn,Ingold,and Prelog (23)) and thus the absolute configuration at C-5 of the xylosyl moiety of UDP-Xyl-5T,,b is R. The absolute configuration at C-5 of the n-glucuronosyl moiety of UDP-GlcUA-5T is X, which means that inversion of configuration occurs during the decarboxylation of UDP-GlcUA-5T by either enzyme. In order to confirm the above results, glycolic acid-T,,d was isolated from UDP-XylL5T,,d.
When glycolic acid-T,,d was oxidized with glycolic oxidase, all of the label was recovered in the product, glyoxalic acid-T (Table IV).
This means that the absolute configuration at C-5 of the xylosyl moiety of UDP-Xyl-5Z',,d is S. These results show that a proton is incorporated at C-5 during the decarboxylation of UDP-GlcUA and that it is incorporated into the side opposite the leaving carboxyl group.

DISCUSSION
Ankel and Feingold proposed that UDP-4-keto-GlcUA would be a likely intermediate in the decarboxylation of UDP-GlcUA by the carboxy-lyases.
The role of NAD in the reaction would be to accept the hydride ion removed from C-4 of the n-glucuronosyl moiety.
This reduced NAD would remain enzyme-bound and the hydride ion would then be stereospecifically introduced into a second intermediate, UDP-4-keto-Xyl, to yield UDP-Xyl (4). The proposed mechanism is similar to that invoked to explain the 4-epimerization of UDP-Glc (5) except that the /3keto acid, UDPI-keto-GlcUA, would be a reaction intermediate. Ankel  Evidence to support the proposed reaction mechanism has now been provided with specifically tritiated UDP-GlcUA as the substrate. A significant isotope effect is noted during the decarboxylation of UDP-GlcUA-47' by UDP-glucuronate carboxy-lyase from either C. laurentii or wheat germ, whereas the decarboxylation of UDP-GlcUA-3T and UDP-GlcUA-57' occurs without an isotope effect. The size of the isotope effect is consistent with a mechanism in which oxidation at C-4 is the rate-limiting step. If a hydride ion is removed upon oxidation at C-4, it must be stereospecifically reintroduced at the same position from which it was removed.
All the label originally present at C-4 of the n-glucuronosyl moiety of UDP-GlcUA-4T is found at C-4 of the D-xylosyl moiety with no inversion of configuration or migration of label.
Additional evidence for the presence of a 4-keto intermediate is provided by the results of the study of the stereochemistry at C-5. The decarboxylation of UDP-GlcUA-5T occurs with complete retention of label at C-5 but with inversion of configuration. The absolute configuration at C-5 of the n-glucuronosyl moiety of UDP-GlcUA-52' is S. After decarboxylation, the configuration at C-5 of the D-xylosyl moiety in UDP-Xyl-5T,,s is R, indicating that inversion of configuration occurs during the decarboxylation.
In addition, the decarboxylation of UDP-GlcUA in tritiated water with the wheat germ enzyme results in the incorporation into UDP-Xyl of only 1 proton with an absolute configuration of S at C-5. These results show that the proton is incorporated at C-5 on the side opposite the leaving carboxyl group.
The decarboxylation of UDP-GlcUA is the second demonstration of a reaction in which inversion of configuration occurs during decarboxylation.
Lienhard and Rose (24) have pointed out that in the case of the decarboxylation catalyzed by 6-phosphogluconate dehydrogenase (EYZ 1.1.1.44), the probable intermediate is the enzyme-bound enol form of D-ribulose 5-phosphate, although they could not rule out the possibility of direct replacement of the carboxyl group with a proton.
Because of the presence of the pyranose ring, it is unlikely that the decarboxylation of UDP-GlcUA occurs by direct replacement.
Such a mechanism would necessitate opening of the pyranose ring, displacement of the carboxyl group by the proton with inversion of configuration, and re-formation of the pyranose ring structure.
In addition, since in a direct displacement reaction there would be no need to facilitate the decarboxylation by formation of a 4-keto intermediate, there should be no isotope effect at C-4 were such a mechanism operative.
The results presented here are consistent with the reaction mechanism presented in Fig. 3. The enzyme would combine with UDP-GlcUA in a reversible step (the C. Zaurentii enzyme would require an additional step to combine with NAD) represented by lci and k-i.
The rate-limiting step, kz, would require the participation of NAD in the extraction of a hydride ion from the substrate at C-4. The 4-keto intermediate (IT) would decarboxylate in an irreversible step leaving a carbanion at C-5 (111). During enolization of 111, C-5 would assume a planar configuration.
Inversion of configuration would occur with the incorporation of the proton at C-5, kh, to form UDP-4-keto-Xyl (IV). The incorporation of the proton has to occur while the substrate is still enzyme-bound because of the complete stereospecificity of the reaction.
If the 4-keto intermediate (111) were to exist free in the medium, the incorporation of the proton (nonenzyme-mediated) would have been randomized. 3. Proposed reaction mechanism for the decarboxylation of UDP-GlcUA.
at C-4, involving the hydride ion originally removed from C-4, is probably the terminal step in the reaction and appears to be irreversible.
UDP-Xyl does not incorporate label when incubated in tritiated water in the presence of UDP-glucuronate carboxy-lyase, which indicates that UDP-Xyl is not activated to the 4-keto intermediate III by the enzyme.
There is no evidence at this time, however, to establish whether kz or I&, is the ratelimiting step in the reaction, but this difference would not affect the over-all mechanism.
This reaction mechanism appears to be consistent with the data which have been presented.
The small difference between the size of the isotope effects for the decarboxylation of UDP-GlcUA-42' by the wheat germ and the C. laurentii enzymes probably does not represent a major difference in mechanism between the two enzymes, and may reflect only the type of binding of NAD to the enzyme.
The mechanism for the decarboxylation of UDP-GlcUA has been compared with the mechanism for the enzymatic epimerization of UDP-Glc to UDP-Gal.
Both mechanisms probably are similar in that an NAD-dependent oxidation at C-4 is involved. The large difference in the type of isotope effect noted with the C-4-tritiated substrates, however, seems to indicate that the enzymatic steps involved after oxidation are entirely different. Whereas reduction of the 4-keto intermediate of the decarboxylation reaction is stereospecific, reduction of the 4-keto intermediate in the epimerase-catalyzed reaction is dictated only by the equilibrium constant.
With the carboxy-lyase, the hydride ion must be held in a fixed orientation in respect to the intermediate, whereas the same may not be true for the epimerase.