THE JOURNAL OF B I O ~ I C A L CHEMISTRY 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc

VOl. 269, No. 15, Issue of April 15, PP. 11509-11513, 1994 Printed in U.S.A.

Xylosyl Transfer to an Endogenous Renal Acceptor PURIFICATION OF THE TRANSFERASE AND THE ACCEPTOR AND THEIR IDENTIFICATION AS GLYCOGENIN*

(Received for publication, June 24, 1993, and in revised form, January 24, 1994)

Lennart RodenS, Sandya AnanthS, Patrick Campbell*, Stephen Manzellas, and Elias MeezanSll From the $Department of Pharmacology, Schools of $Medicine and Dentistry, University of Alabama at Birmingham, Birmingham, Alabama 35294

A xylosyltransferase in rat kidney, tentatively identi- fied as glycogenin (Meezan, E., Ananth, S., Manzella, S., Campbell, P., Siegal, S., Pillion, D. J., and Roden, L. (1994) J. Biol. Chern. 269,11503-11508), was purified by a procedure in which affinity chromatography on UDP- glucuronic acid-agarose was a particularly useful step. The purified material was nearly homogeneous, as shown by SDS-polyacrylamide gel electrophoresis and silver staining, and had an electrophoretic mobility cor- responding to a M, of 32,000. The purified enzyme pos- sessed both glucosyl- and xylosyltransferase activity, and incubation with UDP-[SH]xylose or UDP-[SHlglucose yielded a single macromolecular product, which had the same electrophoretic mobility as the major silver- stained component. These results indicate that the kid- ney transferase was indeed glycogenin and that it was functionally analogous to the larger glycogenin species previously isolated from rabbit muscle. Further exami- nation of the properties of the rat kidney enzyme showed, i.a., that it was inhibited strongly by cytidine 5'-diphosphate. This effect was used to advantage in an alternative purification procedure, which was applied to beef kidney and involved adsorption of the enzyme to LJDP-glucuronic acid-agarose and subsequent elution with cytidine 5"diphosphate. In contrast to glycogenin, glycogen synthase did not catalyze transfer from UDP- xylose, and it is suggested that the incorporation of xy- lose into glycogen observed by other investigators was due to glycogenin-catalyzed xylosyl transfer and subse- quent chain elongation by glycogen synthase.

Initiation of the polysaccharide chains during the biosynthe- sis of certain connective tissue proteoglycans occurs by transfer of xylose residues from UDP-xylose to the hydroxyl groups of specific serine residues in the core proteins of the proteoglycans (1). In some of the first studies of this reaction, it was observed that endogenous xylose acceptors were present in the crude tissue extracts used as enzyme sources and that a portion of the reaction products had properties different from those expected of core proteins with xylosylated serine residues (2, 3). Specifi- cally, the xylose residues were not labile on alkali treatment

DK 39877 and by research grants from the American Diabetes Associa- * This research was supported by National Institutes of Health Grant

tion and the American Heart Association, Alabama Affiliate (to E. M.), and by National Institutes of Health Grants DE 08252 and NS 27353, a grant from DePuy, and a grant-in-aid from the American Heart As-

this article were defrayed in part by the payment of page charges. This sociation, Alabama Affiliate, Inc. (to L. R.). The costs of publication of

with 18 U.S.C. Section 1734 solely to indicate this fact. article must therefore be hereby marked "advertisement" in accordance

ll To whom correspondence should be addressed: Dept. of Pharmacol- ogy, University of Alabama at Birmingham, Birmingham, AL 35294. Tel.: 205-934-4577; Fax: 205-934-1796.

under conditions that cleave the xylose + serine bond in the known proteoglycans. We have recently found that virtually all of the endogenous xylose acceptors in rat kidney yield an alkali- stable product and that a similar product is formed upon incu- bation of the tissue extract with UDP-glucose (4). Since it had been shown previously that xylose may be incorporated into glycogen from UDP-xylose (51, presumably because of a less than absolute specificity for the nucleotide sugar on the par t of glycogen synthase (5), it seemed possible that the endogenous xylose acceptor might be identical to glycogen. This notion was supported by the finding that the xylose-labeled product was susceptible to digestion with a-amylase (4). Other findings, however, showed that the product was not classical glycogen and led to the tentative conclusion (4) that the xylosyl- and glucosyltransferases and their endogenous acceptors were all embodied in a single substance, i e . glycogenin, which is a self- glycosylating protein catalyzing early steps in glycogen biosyn- thesis (6-10). In the present communication, we report the purification of the transferases and their acceptors by a proce- dure in which affinity chromatography on UDP-glucuronic acid-agarose was a particularly effective step. Characterization of the purified material established its identity with glycoge- nin. A comparison with glycogen synthase further demon- strated that the latter has no xylosyltransferase activity.

EXPERIMENTAL PROCEDURES Materials

Frozen rat and rabbit kidneys were obtained from Pel-Freez, Rogers, AR. UDP-[l-3Hlglucose (specific activity, 7.8 CUmmol), and UDP-[l- 3H]xylose (specific activity, 8.9 CUmmol), were purchased from DuPont NEN. The following materials were obtained from Sigma: British anti- Lewisite (2,3-dimercaptopropanol), CHAPS,' UDP-glucose, UDP-xylose, UDP-glucuronic acid-agarose, glycogen synthase (lots 188 F9590 and 29 F9575), glucose, xylose, and maltooligosaccharides (maltose to malto- heptaose). Glucose 6-phosphate was from Boehringer Mannheim. Sepharose CL-GB and PD-10 columns were purchased from Pharmacia LKB Biotechnology Inc. DE53 anion exchange cellulose and Whatman No. 1 and 3MM paper were obtained from Whatman LabSales (Hillsboro, OR). Bio-Rad protein reagent was purchased from Bio-Rad. Quantigold protein assay reagent was obtained from Diversified Biotech (Newton Centre, MA). LiSCN was purchased from Alfa Products (Danvers, MA), phosphotungstic acid from Mallinckrodt (Paris, KY), and ammonium sulfate (enzyme grade) from Serva Biochemicals (New York, NY). Scin- tiverse E was obtained from Fisher Scientific, as were trichloroacetic acid, LiBr, and other analytical reagent grade chemicals. EcoLume liquid scintillation fluid was purchased from ICN (Costa Mesa, CA).

Methods Assay of Glucosyl lXylosyltransferase-Glucosyl- and xylosyltrans-

ferase activities were determined essentially as described in the accom- panying paper (Ref. 4; Method I) or by a modified procedure (Method II),

The abbreviations used are: CHAPS, 3-[(3-cholamidopropyl)dimeth- ylammoniol-1-propanesulfonic acid; MES, 2-(N-morpholino)ethanesul- fonic acid; PAGE, polyacrylamide gel electrophoresis.

11509

11510 Xylosyl Dansfer to an Endogenous Renal Acceptor

TABLE I Purification of glycogenin from rat kidney

volume T o t a l

protein Fraction Total Total activity

Specific activity Purification Recovery

r n l mg cpm cprnlmglh -fold %

1. 105,000 x g supernatant 340 8200 23 x lo6 2875 100 2. Ammonium sulfate precipitate 87 3900 24 x lo6 6340

85 1500 2 104

3. pH 5.0 supernatant 130

25 x lo6 767

17,990 6 109 4. DEAE-cellulose eluate 27 x lo6 35,424 12 117 5. UDP-glucuronic acid-agarose eluate 11 0.055 18.5 x lo6 3.4 x 108 117,000 80

-

used mainly in the purification of glycogenin. For Method I, reaction mixtures (75 pl) contained enzyme in 50 m~ MES, 50 m~ KCl, 12 m~ MgCl,, 6 m~ MnCl,, pH 6.5 (assay buffer I), and 0.15 pCi of UDP- [3Hlglucose or UDP-[3Hlxylose. For Method 11, reaction mixtures (60 pl) contained enzyme in 50 m~ Tris-HC1,0.5 m~ MnCl,, 2 m~ CHAPS, pH 7.4 (assay buffer 11), and 0.047 pCi of UDP-[3H]glucose.

Glycogen Synthase Assay-To determine whether glycogen synthase could use UDP-xylose as a donor substrate, a commercial preparation of glycogen synthase from rabbit muscle (Sigma G-2259; lot 29F 9575; 2.4 unitdmg protein; 0.7 mg/ml) was incubated with UDP-[3H]glucose and UDP-[3Hlxylose essentially as described by Blumenfeld and Krisman (11). Complete reaction mixtures contained the following components in a final volume of 100 pl; glycogen synthase (10 pl; 16.8 milliunits; 0.7 pg of protein), 10% glycogen (10 pl), 0.1 M glucose 6-phosphate (10 pl), UDP-[3Hlglucose or UDP-[3Hlxylose (0.1 pCi; 10 p1) dissolved in 50 m~ UDP-glucose, and 200 m~ glycine-NaOH, pH 8.6,50 m~ EDTA, 10 m~ dithiothreitol(50 pl). In one set of reaction mixtures, the nonradioactive UDP-glucose was omitted, and the radioactive nucleotide sugars (0.1 pCi; 10 pl) were dissolved in 50 m~ MES, pH 6.5, 50 m~ KC1, 12 m~ MgCl,, 6 m~ MnCl,, 75 m~ British anti-Lewisite, 1.5 m~ ATP. After incubation at 37 "C for 20 min, the reaction was terminated by addition of 0.9 ml of 33% KOH, and the tubes were heated in a boiling water bath for 15 min; 1.5 ml of 95% ethanol was added after the tubes had been cooled on ice, and after another 30 min the tubes were centrifuged at 15,000 rpm for 20 min in a microcentrifuge. The supernatant was as- pirated, and the precipitate was dissolved in 1 ml of water. A 0.1-ml sample was taken for radioactivity measurements, and the rest was reprecipitated with 1.5 ml of 95% ethanol and processed as described above. The final precipitate was dissolved in 0.4 ml of water, and ita radioactivity was measured after addition of 4.5 ml of EcoLume.

Enzyme Purification-All operations were carried out at 4 "C with assay buffer I1 unless otherwise specified. Frozen rat kidney (100 g) was homogenized with a blender in 350 ml of buffer, and the suspension was centrifuged at 10,000 x g for 20 min. The resulting pellet was discarded, and the supernatant was centrifuged at 100,000 x g for 60 min. Ammo- nium sulfate was added to the supernatant fraction to 45% saturation, and the mixture was stirred gently for 2 h at 4 "C. After centrifugation at 17,000 x g for 30 min, the precipitated protein was dissolved in 50 ml of buffer, and the solution was dialyzed overnight against 500 ml of the same buffer. The retentate was clarified by centrifugation and adjusted to pH 5.0 by addition of 0.2 M HCl; the resulting suspension was stirred gently for 15 min and was then centrifuged at 17,000 x g for 10 min. The supernatant fraction, readjusted to pH 7.4 with 0.2 M NaOH, was ap- plied directly to a -200-ml column (5 x 10 cm) of DE53 anion exchange cellulose equilibrated in buffer. After rinsing with 300 ml of buffer, the column was eluted with 260 ml of buffer containing 0.1 M KCl. Fractions containing enzyme activity (elution volume, 110-220 ml) were pooled and applied to a 2-ml column of UDP-glucuronic acid-agarose, equili- brated in buffer. The column was rinsed with 10 ml of buffer containing 0.25 M KC1 and was then eluted with eight 0.5-ml aliquots of 2 M LiSCN in buffer. Each fraction was applied to a PD-10 column (Sephadex G-25), which had been equilibrated with buffer I1 and was eluted with 0.5-ml portions of the same buffer. Fractions (0.5 ml) were collected and as- sayed for glucosyltransferase activity (Method II), and enzyme-contain- ing fractions were pooled and stored at 4 "C. In subsequent experi- ments, the procedure was simplified by elution of the PD-10 column with only two portions of buffer (3.5 + 2.5 ml); all enzyme activity emerged in the second fraction. Since the affinity chromatography step was not always equally effective, it was usually repeated at least once. In an experiment with rabbit kidney, performed essentially according to the protocol described above, the enzyme was reapplied to the affinity matrix three times, and additional purification was obtained each time.

Analytical Methods-Reducing power was determined by the Park- Johnson method (12). Paper chromatography was carried out in the descending mode on Whatman No. 1 paper in butanoYethanoYwater (10:3:5) or on Whatman No. 3" paper in butanovpyridinelwater

(4:3:4). Papers were stained with aniline phthalate (13). Radioactive compounds were located by cutting the papers into 0.5-cm strips and soaking in 0.5 ml of water for 1 h, followed by addition of 4.5 ml of EcoLume and counting in a Beckman liquid scintillation counter.

Gel chromatography of oligosaccharides was carried out on a column (1.5 x 200 c m ) of Sephadex G-10, which was eluted with 10% ethanol at a flow rate of 15 ml/h. Fractions of 3 ml were collected and analyzed for reducing power and radioactivity. SDS-PAGE was carried out as de- scribed previously (4). Gels were stained with silver (14) or Coomassie Blue and subjected to autoradiography as described (4).

RESULTS Purification of the Glucosyl IXylosyltransferase from Rat

Kidney-The protocol for the purification of the glucosyV xylosyltransferase from rat kidney encompassed the following steps: ammonium sulfate fractionation of the high speed super- natant from a homogenate of the tissue, precipitation of protein by adjustment to pH 5 , chromatography of the soluble fraction on DE53 anion exchange cellulose, and affinity chromatogra- phy on UDP-glucuronic acid-agarose. The quantitative aspects of the purification are summarized in Table I, from which i t i s seen that the first three steps yielded a 12-fold purification of the enzyme and that the affinity chromatography on UDP- glucuronic acid-agarose was exceptionally effective and re- sulted in close to 10,000-fold purification over the preceding step. It should be pointed out, however, that it was only in the experiment shown in Table I that the affinity chromatography yielded this degree of purification in a single step and that , for four other preparations, the purification ranged from 52-fold to 886-fold. Repetition of the affinity chromatography step, how- ever, consistently yielded a high degree of purification, and, e.g., a 20,000-fold purification of the eluate from the DE53 cellulose column was obtained when a preparation of rabbit kidney enzyme was subjected to four cycles of adsorption fol- lowed by elution with lithium bromide (used as an alternative to LiSCN). The specific activity of this preparation, purified 83,000-fold over the ammonium sulfate-precipitated material in 25% yield, was 402 x lo6 c p d m g p r o t e i f i a n d w a s t h u s comparable to that of the preparation shown in Table I.

Analysis of the purified enzyme by SDS-PAGE gave the re- sults shown in Fig. 1. Upon silver staining (Fig. M), a major band was observed that was located in a position corresponding to a M, of about 32,000. Only a few faint bands were seen in other locations, indicating that the preparation was nearly ho- mogeneous. After incubation of the enzyme with UDP-[3Hlglu- cose (Method I) (Fig. 1B), a single radioactive product was observed by autoradiography, which was located in the same position in the gel as the major silver-stained component.

Similar results were obtained when UDP-[3Hlxylose was used as the glycosyl donor (Method I) (Fig. 1B). It should be emphasized that the approximately 100,000-fold purification of the glucosyltransferase did not result in separation from the xylosyltransferase activity, nor were the acceptors removed in the process. The finding that the major protein component and the reaction products migrated to the same position on SDS- PAGE strongly supports our previous tentative conclusion (4) that the predominant endogenous xylose and glucose acceptors in crude extracts of rat kidney are one and the same compound

Xylosyl Dansfer to an Endogenous Renal Acceptor 11511

66-

42-

31-

2 1-

S E G XG X FIG. 1. SDS-PAGE of purified rat kidney glucosyV

xylosyltransferase.A, a sample of glucosyVxylosyltransferase (325 ng of protein), purified by chromatography on UDP-glucuronic acid-agar- ose, was subjected to SDS-PAGE on a 10% gel together with molecular weight standards. The results of silver staining of enzyme and stan- dards are shown in lanes E and S, respectively. B, samples of purified enzyme, each containing 325 ng of protein, were incubated according to assay method I, with UDP-[3Hlglucose (0.85 pCi) or UDP-[3Hlxylose (0.85 pCi), or both. After SDS-PAGE, the gel was subjected to autora- diography for 4 days. Lanes G , X, and XG show the results for the products labeled with [3Hlglucose, [3Hlxylose, and both sugars, respec- tively.

and that they are the renal form of the self-glycosylating en- zyme glycogenin.

Product formation by the purified enzyme was proportional to enzyme concentration over the range tested (83417 ng/ml). The reaction catalyzed by a partially purified enzyme prepara- tion (specific activity, 2.6 x lo6 cpm/mg/h) was linear with time for 45 min and continued at a decreasing rate for at least 5 h.

Inhibition of Glucosyl Dansfer by Cytidine 5'-Diphosphate- Lomako et al. (15) have observed that ATP inhibits the activity of glycogenin almost completely at a concentration of 4-5 mM. In the present study, we found that CDP is a much more potent inhibitor, causing 93% inhibition of glucosyl transfer (Method I) in a crude rat kidney extract at a concentration of only 25 PM (Fig. 2). Inhibition was also effected by CMP, CTP, CDP-choline, CDP-glucose, and some other nucleotides and nucleotide sug- ars. Detailed results of the testing of these compounds will be reported elsewhere.

In view of the strong inhibition of the glucosyltransferase activity by CDP, it seemed possible that the nucleotide might be useful in the purification of the enzyme as a specific eluant in the affinity chromatography step. This possibility was explored in the course of purification of the glucosyltransferase from beef kidney, which will be reported elsewhere. In these experiments, it was observed that the beef kidney enzyme could be eluted from UDP-glucuronic acid-agarose in good yield with 1 mM CDP.

Glycosyl Donor Specificities of Glycogenin and Glycogen Synthase-Incorporation of xylose from UDP-xylose into glyco- gen has been demonstrated previously by Kimura and Caplan (51, and it was suggested that this reaction was catalyzed by glycogen synthase. In the present study, the ability of glycogen synthase to utilize UDP-xylose as the glycosyl donor was ex- amined directly, with a commercial preparation of rabbit skel- etal muscle glycogen synthase as the enzyme source. As shown in Table 11, the incorporation of radioactivity into glycogen from UDP-["H]glucose, in a reaction mixture without added nonra- dioactive UDP-glucose, was close to 56,000 cpm. In contrast, incubation with UDP-[3H]xylose under comparable conditions yielded only 60 cpm of product (uncorrected for any blank val- ues). I t was thus apparent that glycogen synthase cannot use UDP-xylose as a donor substrate to any significant extent.

I t should also be mentioned that the commercial glycogen synthase preparation was not pure (and was not claimed to be)

CDP concentration (wM)

FIG. 2. Effect of CDP on [SHlglucose incorporation into trichlo- roacetic acid-precipitated endogenous reaction products from a crude rat renal extract. Reaction mixtures (75 p1) contained high speed supernatant from a homogenate of rat kidney (100 pg of protein; see Ref. 4),50 mM MES, 50 m~ KCl, 6 mM MnCl,, 12 m~ MgCl,, pH 6.5, 0.15 pCi (0.26 p ~ ) UDP-["Hlglucose, and CDP as indicated in the figure.

TABLE I1 Glycosyl donor specificity of glycogen synthase

Radioactivitv

Incubation mixture Product formed Product formed from from

UDP-[3HI~lucose UDP-[.'Hlxylose

cpm cpm Complete system 13,709 16 Minus glucose 6-phosphate 1,543 Minus glycogen" 23 13 Zero time control 20 14 Control without enzyme 31 Minus nonradioactive UDP-glucose 55,660 60

Glycogen added after completed incubation.

and that some 30 protein bands were observed on SDS-PAGE (data not shown). One of the contaminants was glycogenin, and assay of two different batches of glycogen synthase for glyco- genin activity (Method I), using 8 pg (25 milliunits) of protein from batch 188 F9590 and 7 pg (42 milliunits) from batch 29F 9575, yielded 1490 and 67 cpm, respectively, when UDP-[3H]xy- lose was the donor substrate. Incubation of the same two batches with UDP-[3H]glucose as donor yielded 28,779 and 1164 cpm, respectively.

Size of Oligosaccharide Chains in the Products of PHIGlu- cosyl- and PHlXylosyl Dansfer-The products of glycosyl transfer from UDP-[3H]glucose and UDP-[3H]xylose generated by incubation of the nucleotide sugars with the most highly purified rat kidney enzyme were characterized further to as- certain that the two sugars had been incorporated unaltered and to determine the size of the labeled oligosaccharides. After acid hydrolysis (1 M HCI; 100 "C, 3 h) of the trichloroacetic acid-precipitated product of transfer from UDP-[3H]glucose, chromatography on Whatman No. 1 paper in butanoWethanoW water (10:3:5) showed the presence of a single radioactive spot in the same location as glucose. Similarly, only xylose was detected in the products from reaction mixtures containing UDP-[3Hlxylose.

The size of the oligosaccharides in the [3H]xylose-labeled products was examined by the experiments shown in Fig. 3. After incubation of the same enzyme preparation with UDP- ['H]xylose, the entire reaction mixture was subjected to mild acid hydrolysis (0.1 M HCI; 100 "C; 15 mid, and the hydrolysate was chromatographed on Sephadex G-10 together with a mix- ture of maltooligosaccharide markers (triose through hep- taose). The profile given by the [3Hlxylose-labeled products (Fig. 3A) exhibited a major, early peak of radioactivity and two smaller, incompletely resolved peaks, followed by the expected

11512 Xylosyl Dansfer to an Endogenous Renal Acceptor

5 0.4 - A T rn

I I I Ill

E 0 C

aa 5 0 100 1 5 0 2 0 0 2 5 0 3 0 0

Effluent volume (ml)

B H.pc.0.. H

0 5 1 0 1 5 2 0 2 5 3 0 3 5 4 0

Distance from origin (cm) FIG. 3. Determination of the size of oligosaccharides in [9I]xyy-

lose-labeled rat kidney glycogenin. A reaction mixture (150 p1) con- taining 700 ng of purified rat kidney glycogenin and 3 pCi of UDP- [3Hlxylose (specific activity, 8.9 Ci/mmol) was prepared as otherwise described for Method I and was incubated for 1 h at 37 "C. The total radioactivity in trichloroacetic acidphosphotungstic acid-precipitable product, determined by analysis of a 10-pl sample, was 13,700 cpm. To the remainder of the reaction mixture (140 pl) was added 140 pl of 0.2 M HCl and 220 pl of 0.1 M HCl to yield a final volume of 0.5 ml. m e r hydrolysis at 100 "C for 15 min, 0.5 ml of a mixture of maltooligosac- charides (5 mg each of tri- through heptasaccharide) was added, and the mixture was applied to a column (1.5 x 200 cm) of Sephadex G-10, which was eluted with 10% ethanol at a flow rate of 15 mVh. Fractions of 3 ml were collected and analyzed for radioactivity (0) and carbohydrate con- tent (0) (ParkJohnson method) and were pooled as indicated in the figure. B, paper chromatography of pooled fractions from the experi- ment shown in A. m e r evaporation to dryness, pools 1-111 were each dissolved in 100 pl of water. A 2 0 4 sample from each pool was sub- jected to descending paper chromatography on Whatman No. 3MM paper in butanollpyridindwater (4:3:4) for 28 h together with standards (0.3 mg) of glucose, maltose, and the five maltooligosaccharides. Radio- activity was measured as described under "Methods." The standards were stained with aniline hydrogen phthalate. 0, pool I; 0, pool 11; X, pool 111.

large peak of xylose formed by hydrolysis of the UDP-[3Hlxylose in the reaction mixture. The first radioactive peak was eluted with or close to V,, as was the leading edge of the major Park- Johnson positive peak, indicating that the labelled oligosaccha- ride was at least the size of a heptaose.

The radioactive fractions were pooled as indicated in Fig. 3A and were analyzed by paper chromatography, which gave the results illustrated in Fig. 3B. In accordance with the result of gel chromatography, most ofthe radioactivity in the major earli-

est peak (pool I) migrated with maltoheptaose. Since xylose migrates faster than glucose and this property must be re- flected to some extent in the oligosaccharides, it is possible that the [3Hlxylose-labeled oligosaccharide in the position of malto- heptaose was actually an octasaccharide. The more retarded peaks contained faster moving components that migrated to the positions of the smaller oligosaccharides.

Analysis of the [3H]glucose-labeled products gave essentially similar results except that the major, early peak observed on gel chromatography after mild acid hydrolysis was eluted a few milliliters earlier than the corresponding [3H]xylose-labeled peak (results not shown). Paper chromatography of this mate- rial showed the presence of a major component, which migrated somewhat more slowly than maltoheptaose (data not shown). These analyses are in accord with findings of Lomako et al. (16), who have reported that the largest oligosaccharide produced by skeletal muscle glycogenin is maltooctaose.

The interpretation of the above results was not entirely straightforward, since the oligosaccharides had been generated by acid hydrolysis, which could have cleaved some glycosidic linkages in addition to the labile linkage between glucose and tyrosine in the glycogenin molecule. In a control experiment, maltoheptaose was hydrolyzed under the same conditions as the labeled material, and the formation of reducing end groups was determined by the Park-Johnson method. It was observed that the reducing power increased 57 and 95% after hydrolysis for 15 and 30 min, respectively, indicating that, after the longer hydrolysis time, on average one glycosidic bond in each oligo- saccharide molecule had been cleaved. Although it could there- fore be assumed that a portion of the oligosaccharides in the hydrolysates of the radiolabeled product had arisen by hydroly- sis, it was nevertheless also apparent that a large proportion of the original oligosaccharides should have escaped cleavage dur- ing the 15-min hydrolysis period used in these experiments. It was thus concluded that the [3Hlxylose-labeled oligosaccha- rides in pool I (Fig. 3) and the corresponding [3Hlglucose-la- beled compounds were indeed the largest oligosaccharides formed in the xylosyl and glucosyl transfer reactions.

DISCUSSION

The objective of this investigation was to isolate the major endogenous xylose acceptor in rat kidney and, likewise, to iso- late and characterize the xylosyltransferase acting on this sub- strate. Through examination of the properties of the enzyme and the reaction product (41, it became clear that we might be dealing with glycogenin (6, 101, and it was then apparent that separation of enzyme and acceptor was not to be expected un- der any circumstances. Purification of the enzyme was there- fore undertaken on this assumption, with the use of UDP- glucose as the glycosyl donor, since this nucleotide sugar is the physiological substrate for glycogenin. Indeed, after purifica- tion to near homogeneity, no separation of the glucosyltrans- ferase from its acceptor had occurred, nor had the xylosyltrans- ferase and the glucosyltransferase activities been separated from each other. These findings strongly indicated that the isolated protein was the renal form of glycogenin and that the xylosyl- and glucosyltransferase activities resided in the same molecule. This conclusion was further supported by the results of SDS-PAGE of the [3Hlglucose- and [3Hlxylose-labeled prod- ucts, which were found to migrate to the same position and were located in the same place as the purified protein. Together with the results reported previously (41, these findings also support the conclusion that the major xylose acceptor in rat kidney is identical to glycogenin.

The rat kidney glycogenin was considerably smaller than the enzyme isolated from skeletal muscle and had a M, of about 32,000 uersus 37,000 for the muscle enzyme (€49). It was there-

Xylosyl Dansfer to an Endogenous Renal Acceptor 11513

fore possible that the kidney enzyme had been subjected to proteolytic cleavage, either during the isolation procedure or already in the intact kidney. We do not yet have sufficient information to allow a definitive conclusion in this regard, but it may be noted that the presence of protease inhibitors during the initial homogenization of the tissue did not affect the size of the glucose-labeled reaction product. In contrast to the rat kid- ney enzyme, purified glycogenin from beef kidney, which had been similarly processed, displayed two bands on SDS-PAGE, which had M, values of 32,000 and 38,000, respectively (data not shown). Again, it was not clear whether the smaller species had been formed before or during the preparative procedure.

Glycogenin has previously been purified from skeletal muscle and liver by several rather laborious procedures (8, 9, 17), which do not include any specific affinity chromatography steps. We reasoned that chromatography on UDP-glucuronic acid-agarose might be a useful addition to existing procedures, since this affinity matrix has been used successfully in other investigations, e.g. in the purification of the glucosyltrans- ferase involved in collagen biosynthesis (18, 19). Indeed, chro- matography on UDP-glucuronic acid-agarose proved to be an extremely efficient step in the isolation of the kidney enzyme, and repeated application of the eluate to the affinity matrix yielded as much as 20,000-fold purification over the preceding step. The same procedure has also been used successfully in the purification of glycogenin from bovine and porcine muscle (data not shown), and it may be noted that a well aged beef tender- loin, purchased in a supermarket, was an excellent source of glycogenin.

Currently available evidence indicates that the self-glycosy- lation of glycogenin stops after a total of 8 glucose residues have been added (16). In agreement with this conclusion, our results indicated that an octasaccharide was the largest prod- uct formed under the conditions chosen.

The ability of glycogenin to utilize UDP-xylose as the glycosyl donor in addition to UDP-glucose raised the question whether the same was true for glycogen synthase, particularly since this possibility had been suggested by Kimura and Caplan (51, who observed incorporation of xylose from UDP-xylose into glycogen in polysome preparations from cultured embryonic chick limb bud cells. The experiments reported here clearly showed that glycogen synthase from rabbit skeletal muscle did not use UDP-xylose as the glycosyl donor instead of UDP-glucose. De- spite the difference in the experimental systems used by Kimura and Caplan (5) and by us, it therefore seems likely that the results obtained by the former investigators were due to the activity of glycogenin rather than glycogen synthase in their preparations.

The question may be raised whether the presence of a non- reducing terminal xylose residue in the oligosaccharide chain of glycogenin represents a stop signal for further chain elongation or whether a glucose residue can be transferred, by glycogenin or glycogen synthase, to such a residue. We have found (data not shown) that, upon incubation in the presence of an excess of nonradioactive UDP-glucose, a small but significant proportion of [3Hlxylose-labeled glycogenin can be elongated by glycogen synthase, indicating that this enzyme recognizes a nonreducing terminal xylose residue as an acceptor and is capable of trans- ferring glucose to such a residue. The rather complex nature of these experiments, however, make it desirable to confirm this conclusion by independent methodology, and it is possible that the use of exogenous substrates may be helpful in this regard. It has been shown by Lomako et al. (20) that p-nitrophenyl glycosides containing maltose and higher maltooligosaccha-

rides may serve as exogenous acceptors in the glycogenin reac- tion, and recent studies in our laboratory have shown that dodecyl maltoside is an even better substrate than any of the p-nitrophenyl compounds. Since dodecyl maltoside serves as an acceptor in the transfer of either glucose or xylose, we antici- pate that testing of the enzymatically xylosylated maltoside will resolve this question unambiguously, and experiments of this nature are presently under way.

The primary structure of glycogenin from rabbit muscle has recently been determined by amino acid sequencing (211, and cDNA cloning and expression of the enzyme in Escherichia coli have been reported by Viskupic et al. (22). No such information is yet available for any of the kidney glycogenins; Viskupic et al. (22), however, have observed that the rabbit kidney glycogenin mRNA transcript is a single entity and has the same size as the skeletal muscle transcript. In as yet incomplete experiments, we have found that, upon Western blotting, the rat kidney enzyme does not react with an antibody toward rabbit muscle glycogenin (23) (kindly provided by Dr. William J. Whelan), nor does it react with an antibody to a peptide from the C-terminal end of rabbit muscle glycogenin (24) (kindly provided by Dr. Peter Roach). The latter antibody, however, reacted strongly with a M, 38,000 component of a crude preparation of beef kidney glycogenin; no reactivity toward any substance in the M , 32,000 region of the gel was observed. These findings sug- gest that the kidney enzyme is synthesized in vivo as a molecu- lar species with an M , of 38,000 and that subsequent proteoly- sis partially or completely removes a peptide with an M, of about 6,000 from the C-terminal end of the molecule. Until detailed studies of the structures of the kidney enzymes have been undertaken, however, the relationship between the differ- ent forms remains hypothetical.

Acknowledgments-We thank Drs. William J. Whelan and Peter J. Roach for generous gifts of antibodies to glycogenin. We are also grate- ful to Dr. Whelan for providing information about the procedure for purification of muscle glycogenin prior to publication.

REFERENCES 1. Roden, L. (1980) in The Biochemistry ofGlycoproteins and Proteoglycans (Len-

2. Grebner, E. E., Hall, C. W., and Neufeld, E. F. (1966) Biochem. Biophys. Res.

3. Grebner, E. E., Hall, C. W., and Neufeld, E. F. (1966)Arch. Biochem. Biophys.

4. Meezan, E., Ananth, S., Manzella, S., Campbell, P., Siegal, S., Pillion, D. J.,

5. Kimura, J. H., and Caplan, A. I. (1978)Arch. Biochem. Biophys. 191,687497 6. Whelan, W. J. (1986) B~WSSQYS 5, 136-140 7. Pitcher, J., Smythe, C., Campbell, D. G., and Cohen, P. (1987) Eur: J. Biochem.

8. Pitcher, J., Smythe, C., and Cohen, P. (1988) Eur J. Biochem. 176,391-395

10. Smythe, C., and Cohen, P. (1991) Eur J. Biochem. 200,625431 9. Lomako, J., Lomako, W. M., and Whelan, W. J. (1988) FASEB J. 2,3097-3103

11. Blumenfeld, M. L., and Krisman, C. R. (1985) J. Biol. Chem. 260,1156&11566 12. Park, J. T., and Johnson, M. J. (1949) J. Biol. Chem. 181, 149-151 13. Partridge, S. M. (1949) Nature 164,443 14. Meml, C. R., Goldman, D., Sedman, S.A., and Ebert, M. H. (1981) Science 211,

15. Lomako, J., Lomako, W. M., and Whelan, W. J. (1990) Biofactors 2, 251-254 16. Lomako, J., Lomako, W. M., and Whelan, W. J. (1990) Biochem. Znt. 21,251-

17. Smythe, C., Villar-Palasi, C., and Cohen, P. (1989) Eur J. Biochem. 183,

18. Myllyla, R., Anttinen, H., Risteli, L., and Kivirikko, K. I. (1977) Biochim.

narz, W. J., ed) pp. 267-371, Plenum Press, New York

Commun. 22, 672477

116, 391-398

and Roden, L. (1994) J. Biol. Chem. 269, 11503-11508

169,497-502

1437-1438

260

205-209

~Btophys. Acta 480, 113-121

20. Lomako. J.. Lomako. W. M.. and Whelan. W. J. (1990) FEES Lett. 264. 13-16 19. Anttinen, H., Myllyla, R., and Kivirikko, K. I . (1978) Biochem. J. 175,737-742

21. Campbell, D. G., and Cohen, P. (1989) Eur: J. Biochem. 185, 119-125 ' 22. Viskupic, E., Cao, Y., Zhang, W., Cheng, C., DePaoli-Roach, A. A,, and Roach,

23. Rodriguez, I . R., and Whelan, W. J. (1985) Biochem. Biophys. Res. Commun.

24. Cao, Y., Mahrenholz, A. M., DePaoli-Roach, A. A,, and Roach, P. J. (1993) J.

P. J. (1992) J. Biol. Chem. 267,25759-25763

132,829436

Biol. Chem. 268, 14687-14693