0031-3998/85/19 12-1240$02.00/0 PEDIATRIC RESEARCH Vol. 19, No. 12, 1985 Copyright 63 1985 International Pediatric Research Foundation, Inc. Printed in U.S. A.

Fetal Cartilage xylosyltransferase Activity and Skeletal Growth in Sheep

FRANK H. MORRISS, JR., BRIAN FITZGERALD, AND LAVON M. RIDDLE Houston Perinatal Nutrition Laboratory, Department ofpediatrics, Medical School, University of Texas Health Science Center at Houston, Houston, Texas 77030

ABSTRAa. The activity of UDP-D-xy1ose:proteoglycan predominantly a Dorset-Rambouillet cross. The ewes were of core protein B-D-xylosyltransferase(EC 2.4.2.26), the en- known gestational age (65 to 138 days; term 147 days) that was zyme that catalyzes the initiation of the polysaccharide established by breeding history and confirmed by fetal radi- chain linkage to the core protein of proteoglycans, was ographs. The ewes were sacrificed by exsanguination or lethal measured in costal cartilage from 20 fetal sheep of 65-138 injection. Each fetus was dried and weighed, the vertebral column days gestation. Activity of the enzyme was estimated from length was measured, and the anterior chest wall of the fetus was the transfer of [14Cjxylosefrom UDP-[14Cjxyloseto silk as removed and stored at -70" C for no longer than 12 months the acceptor protein. The specific activity decreased ap- before further enzyme preparation was done. The anterior chest proximately 10-fold and was found to be highly correlated wall was thawed, and the cartilaginous portions (not including with the decremental rate of growth in length of the fetal the growth plate) of the sixth through the ninth ribs were cleaned vertebral column. These observations, together with the free of tissue while wetted with a buffer which was 0.05 M KC1, known gestational decrease in the in vitro rate of uptake 0.05 M 2(N-morpholino)etanesulfonic acid at pH 6.5, containing of radiolabeled sulfate by ovine fetal cartilage, a subsequent 0.006 M MnC12 and 0.012 M MgC12 at 4" C. The cartilage was step in proteoglycan synthesis, support the hypothesis that chopped into small pieces and shredded with a Tekmar Tissue- normal fetal skeletal growth is dependent during the last mizer (Tekmar Co., Cincinnati, OH) for 9 min in 10 ml of the one-half of gestation on the activity of xylosyltransferase above buffer during which process the cartilage was cooled to 4" in cartilage. (Pediatr Res 19: 1240-1243,1985) C at 3-min intervals. Next, the tissue suspension was homoge- nized by hand with 20 strokes of a Tenbroeck homogenizer while kept ice cold. The homogenate was centrifuged at 10,000 x g for 10 min at 4" C; the supernate was used as the enzyme solution In the synthesis of proteoglycans by chondrocytes, sulfation is in the xylosyltransferase assay. The 10,000 x g supernate may a late step (1). In the sheep fetus the diminishing fractional contain endogenous acceptor proteins in low concentration. growth rate of the skeleton during gestation is paralleled by Preparation of acceptor protein. The xylosyltransferase activity decreasing uptake of radiolabeled sulfate by costal cartilage in was assayed by the rate of transfer of xylose from uniformly vitro (2). Although fetal serum sulfate concentration also de- labeled UDP-['4C]xyloseto an acceptor protein (5). The protein creases during the course of gestation, fetal cartilage sulfate acceptor for the transfer of xylose to connective tissue protein uptake for incorporation into proteoglycans is not limited by catalyzed by the enzyme appears to require a glycine residue on sulfate availability (2). Hence, a rate-limiting step in the synthesis the C-terminal side of a xylose-accepting serine residue; silk of proteoglycans by fetal cartilage may occur prior to the final contains a large number of Ser-Gly-sequences, and thus serves sulfation step. A possibly critical step in the synthesis of proteo- as an acceptor of high activity (6). We used silk as the acceptor glycans is that which initiates polysaccharide chain formation by protein; use of silk as the acceptor protein for the activity of transfer of D-xylose from UDP-D-xylose to the protein core of xylosyltransferase obtained from chick embryo cartilage recently the proteoglycan, a reaction that is catalyzed by UDP-D-xy- has been described (6). Silk worm cocoons were obtained from 1ose:proteoglycan core protein P-~xylosyltransferase (EC the International Silk Association, USA (New York, NY). These 2.4.2.26) (1, 3, 4). were sliced open, and the worms were removed. The cocoons We conducted this survey of fetal sheep costal cartilage xylo- were cut into small pieces and boiled in distilled water until the syltransferase specific activity to determine if enzyme activity cocoon pieces had unraveled into strands; this usually required decreases during gestation. We also sought to determine if a boiling for 2 days or less. The strands were lyophilized to dryness, decrease in enzyme specific activity correlated with the decre- and the dry silk was dissolved in aqueous 60% (wtlvol) LiSCN mental fractional growth rate of the fetal skeleton during the last (Alfa Chemical Co., Danvers, MA) at a ratio of 1 g silk: 10 ml half of gestation, a correlation that would support the hypothesis LiSCN. This solution was centrifuged at 1500 x g for 15 min to that the rate of fetal skeletal growth is dependent on the activity obtain a supernate containing the dissolved silk; to this were of xylosyltransferase in cartilage. added 90 ml of 0.25 M EDTA at pH 6.2. The silk-EDTA solution was dialyzed in Spectrapor 3 dialysis tubing (Spectrum Medical MATERIALS AND METHODS Industries, Inc., Los Angeles, CA) against flowing deionized water (250 ml/h) for 48 h or until the effluent gave a negative test for Extraction of cartilage enzyme activity. Twenty fetal sheep SCN- with FeCl3. The dialized silk solution was centrifuged at were removed from the uteruses of range-fed western ewes, 1500 x g for 15 min, and the supernate was kept at 4" C for no longer than 2 days before being used as the acceptor protein in Received May 15, 1985; accepted July 16, 1985. the xylosyltransferase assay. Address for reprints and correspondence Frank H. Morriss, Jr., M.D., Depart- ment of Pediatrics, University of Texas Medical School at Houston, 6431 Fannin Preparation of substrate. A solution that was 0.015 M KCl, Street, Houston, TX 77030. 0.0165 M 2(N-morpholino)ethanesulfonic acid at pH 6.5, con- Supported by HHS Grant HD-11337. taining 10 pCi/ml of UDP-['4C]xylose (267 mCi/mmol, New 40 XY LOSY LTRANSFERASE ACTIVITY AND GROWTH England Nuclear, Boston, MA) and 0.24 pmol/ml of unlabeled UDP-xylose (Sigma Chemicals, St. Louis, MO) was used as the substrate solution. Enzyme assay. The assay for xylosyltransferase activity was identical to that described by Roden et al. (5), except that the 1% silk solution described above served as the protein acceptor, and that 10 ml Hydrofluor (National Diagnostics, Somerville, NJ) was used instead of the described scintillation mixture. The 50 p1 of the enzyme solution, 20 p1 of the acceptor solution, and 5 p1 of the UDP-['4C]xylose substrate solution were incubated for 40 min then stopped by placement on ice and the addition of 10 p1 of 0.25 M EDTA at pH 8.0. Control incubation mixtures included the enzyme, acceptor protein, and substrate solutions, but were prepared with 10 p1 0.25 M EDTA at pH 8.0 prior to incubation for 40 min on ice. To assay xylose incorporation, 0.05 ml of 1% bovine serum albumin was added to the incuba- tion mixture, proteins were precipitated by the addition of 0.15 ml of 10% trichloroacetic acid-4% phosphotungstic acid, col- lected by centrifugation, washed twice with 0.3 ml of 5% trichlo- roacetic acid, and dissolved in 0.1 ml of 1 M NaOH. Each sample was transferred to a scintillation vial, diluted to 0.4 ml with water, and 0.1 ml of 1 M acetic acid was added to neutralize the NaOH before the radioactivity was measured. Dpm in the final reaction product were converted to total ng xylose transferred/ mg cartilage protein by considering both the labeled and unla- beled xylose in the incubation mixture. Cartilage protein was assayed by a modification of the Bradford method (7, 8). Each fetal cartilage assay was performed in triplicate; the coefficient of variation of three replicates was 4%. Statistics. Linear regression equations were calculated by the least squares method, and the significance of the slope of the equation was tested with the F ratio test.

GESTATIONAL AGE, days Fig. 2. Top, fetal sheep vertebral column length as a function of gestational duration. The increase in length from mid-gestation to term may be expressed by the equation: Y = -13 + 0.39X (r = 0.98; p c 0.0001). Bottom, fractional increase in length of fetal sheep vertebral column as a function of gestational duration. The diminishing growth rate from midgestation to term may be expressed by the equation: Y = 0.39/(- 13 + 0.39X).

RESULTS Fetal sheep costal cartilage xylosyltransferase specific activity decreased approximately 10-fold as a function of gestational duration (Fig. 1). This relationship may be expressed by the linear regression equation: Y = 204 - 1.45 X, where Y is ng xylose transferred to the silk acceptor protein per mg cartilage protein after 40 min incubation at 37" C, and X is gestational duration in days. In Figure 2 (top) is shown the vertebral column length as a function of gestational age. Vertebral column length increased 4-fold during the interval of gestation surveyed. The fractional rate of longitudinal growth of the vertebral column [Fig. 2 0 I,;', 8 I 1 ! I I I I (bottom)] was calculated for each fetus by dividing the 24-h 80 70 80 80 100 110 120 130 140 150 growth increment calculated using the equation from Figure 2 GESTATIONAL AGE, days (top) by the attained vertebral column length. The fractional rate Fig. 1. Fetal sheep costal cartilage specific activity of core protein P- of growth of the fetal vertebral column decreased as gestation ~xylosyltransferaseas a function of gestational duration. The decrease proceeded [Fig. 2 (bottom)]. The relationship during the last half in enzyme activity from mid-gestation to term may be expressed by the of gestation between the fractional rate of growth in length of equation: Y = 204 - 1.45X (r = 0.94; p < 0.0001). the vertebral column and cartilage xylosyltransferase activity is 1242 MORRISS; ET AL. 4). Xylosyltransferase activity has been localized in the cisternae of the rough endoplasmic reticulum of embryonic cartilage cells (9). The activity of the enzyme in rat costal cartilage has been found to decrease approximately 3-fold with postnatal age from 3 to 36 months (10). The postnatal decrease in enzyme activity was not evaluated in terms of rat postnatal fractional growth rate; however, the relative decrease in enzyme activity in rat cartilage is consistent with the slower decrease in fractional rate of growth of the skeleton after birth compared to the rate during the fetal period. The amount of galactosamine-containing pro- teoglycan that can be extracted from rat cartilage also decreases with age, an observation that suggests that proteoglycans are synthesized postnatally by a mechanism that reflects reduced ezyme activity (10). More specifically, there is evidence that the xylosyltransferase interacts with the first galactosyltransferase in the subsequent steps in the synthesis of chondroitin sulfate to regulate the rate of synthesis (1). Additional evidence of the critical nature of this enzyme in fetal chondrocyte proteoglycan synthesis has been described for cultured embryonic chick chon- drocytes: xylosyltransferase activity levels increased coordinately with the exponential growth of the cells, reaching maximum levels approximately 2 days prior to the maximum rate of proteoglycan synthesis (1 1). The mediator for the decrease in fetal sheep cartilage xylosyl- transferase specific activity is unknown. The enzyme isolated from murine melanoma can be inhibited by dialdehyde nucleo- CARTILAGE XYLOSYLTRANSFERASE ACTIVITY, ng xylose-rng proteini.40 mini sides (12). When embryonic chick cartilage is treated in vitro Fig. 3. Fractional increase in length of fetal vertebral column as a with somatomedin or cortisol, the uptake of radiolabeled sulfate function of the specific activity in cartilage of core protein @-~xylosyl- is stimulated or depressed, respectively; these hormonal effects transferase. The dependency of growth rate on enzyme activity may be appear to be the results of acceptor protein changes, however, expressed by the equation: Y = 0.0079 f 0.00015X (r = 0.96; p < rather than xylosyltransferase activity changes (13). Moreover, 0.0001). in the ovine fetus, cortisol increases in the serum very late in gestation (14), and thus is unlikely to be responsible for either the decrease in radiolabeled sulfate uptake or the decrease in shown in Figure 3. The activity of the enzyme and the fractional enzyme activity that begins at midgestation (2). It is unknown if growth rate of the vertebral column are highly correlated; the UDP-xylose available for fetal cartilage metabolism becomes enzyme activity explains 92% of the variance in the fractional limiting during fetal growth, and it is unknown if inhibitors of rate of growth of the vertebral column. xylosyltransferase, such as dialdehyde nucleosides, increase in cartilage as gestation progresses. Klagsbrun and coworkers (15, DISCUSSION 16) have purified a factor extracted from bovine cartilage which inhibited sulfate uptake into bovine chondrocytes; however, We have demonstrated a decrease in the specific activity of neither the effect of the factor on xylosyltransferase activity nor xylosyltransferase in fetal sheep costal cartilage as gestation pro- gestational variation Was investigated. It is possible that the ceeds from approximately midgestation to nearterm. The frac- decrease in enzyme activity with gestational age is in part the tional growth rate of the fetal vertebral column, i.e. the incre- compounded result of less enzyme activity being present in mental increase in length divided by the attained length, may be cartilage that is farther from the growth plate coupled with a late- expressed as a dependent function of xylosyltransferase activity. gestation bias in sampling technique in which cartilage farther Whether the fetal cartilage xylosyltransferase activity, in fact, is from the growth plate is preferentially sampled. We have shown, rate-determining for the synthesis of cartilage proteoglycans and however, in earlier studies in which the same sampling technique for fetal skeletal growth during the last half of fetal life is not was used for a study of cartilage uptake of radiolabeled sulfate, known, but these results support such an hypothesis. Moreover, no detectable difference among samples obtained from the center the decrease in xylosyltransferase specific activity parallels the of a cartilage slice compared with those obtained from the decrease in fetal sheep costal cartilage uptake of radiolabeled periphery and no detectable difference among slices along a given sulfate in vitro for incorporation into proteoglycans during the rib or among ribs from a single fetus of this species (2). Moreover, same span of gestation (2). The enzyme activity decreased ap- cell density of costal cartilage sampled in this manner decreased proximately 10-fold while the uptake of radiolabeled sulfate, only 19% during the last one-half of gestation in 17 fetuses under submaximal stimulation by serum growth factors in vitro, studied. decreased approximately 8-fold. These observations do not ex- As an index of fetal skeletal growth, we measured the vertebral clude the possibility, indeed likelihood, that ovine fetal skeletal column length. This measurement, one first described for fetal growth rate may be found to correlate with other enzymes that sheep by Cloete (17), has served investigators as a standard, regulate cartilage synthesis of proteoglycans in particular or reproducible index of fetal gestational age, and is characteristic cellular metabolism in general, or with hormones or other agents. of fetal sheep skeletal growth. The equation for Figure 2 (bottom), The fetal skeleton grows principally by chondrocyte replica- which describes the fractional rate of increase in vertebral column tion, matrix secretion, and increase in cell volume. Synthesis of length derived from these 20 animals, is of the same form as the proteoglycans contributes to the latter two growth processes. equation Cloete (1 7) derived, and it is quantitatively similar. Polysaccharide chains are attached to core proteins of proteogly- In summary, we have demonstrated that xylosyltransferase- cans via a glycosidic linkage between ~xyloseand the hydroxyl specific activity in ovine fetal cartilage decreases during the last group of serine residues (1). The enzyme that catalyzes the one-half of gestation and that the decrease is highly correlated linkage, UDP-~xy1ose:proteoglycancore protein P-~xylosyl- with the decrement in ovine fetal skeletal fractional growth rate. transferase, has been purified from connective tissue sources (3, A subsequent metabolic event in cartilage growth processes, XYLOSYLTRANSFERASE ACTIVITY AND GROWTH 1243 sulfation of proteoglycans, as estimated in vitro also decreases as new substrate for UDP-D-xylose: proteoglycan core protein 8-D-xylosyltrans- ferase. Anal Biochem 137:505-516 gestation proceeds to term at a decremental rate equal to that 7. Bradford MM 1976 A rapid and sensitive method for the quantitation of which would be expected if the synthesis of unsulfated proteogly- microgram quantities of protein utilizing the principle of proteindye bind- cans were limited by enzyme activity. These related observations ing. Anal Biochem 72248-254 support the hypothesis that fetal skeletal growth during the last 8. Bio-Rad Technical Bulletin 1051 1977 Bio-Rad Laboratories, Richmond, CA 9. Hoffmann HP, Schwartz NB, Rodin L, Prockop DJ 1984 Location of xylosyl- one-half of gestation is dependent upon the activity of cartilage transferase in the cisternae of the rough endoplasmic reticulum of emblyonic xylosyltransferase. The decrease in the specific activity of the cartilage cells. Connect Tissue Res 12: 15 1- 163 enzyme during gestation is unexplained. 10. Wolf B, Gressner AM, Nevo 2,Greiling H 1982 Age-related decrease in the activity of UDP-xylose: core protein xylosyltransferase in rat costal cartilage. Mech Ageing Dev 19:181-190 I I. Schwartz NB 1976 Chondroitin sulfate glycosyltransferases in cultured chon- REFERENCES drocytes: turnover, oscillatory change during growth and suppression by 5'- bromodeoxyuridine.J Biol Chem 25 1:3346-3351 1. Schwartz NB 1982 Regulatory mechanisms in proteoglycan biosynthesis. In: 12. Lee MH, Lazo JS, Li CD, Hadfield AF, Maniglia CA, Sartorelli AC 1982 Varma RS, Varma R (eds) Glycosaminoglycans and Proteoglycans in Phys- Solubilization of murine melanoma xylosyltransferase and galactosyltrans- iological and Pathological Processes of Body Systems. Karger, Basel, pp 41- ferase activities and their inactivation by dialdehyde nucleosides. Chem Biol 54 Interact 42141-153 2. Moniss FH. Marshall RN. Crandell SS, Fitzerald BJ, Riddle LM 1985 13. Kilgore BS, McNan ML, Meador S, Lee JA, Hughes ER, Elders MJ 1979 ~ecreases'inovine fetal cartilage sulfate uptake and serum sulfate during Alteration of cartilage glycosaminoglycan protein acceptor by somatomedin gestation. Am J Physiol249:E115-El20 and cortisol. Pediatr Res 13:96-99 3. Schwartz NB, Rodin L 1974 Biosynthesis of chondroitin sulfate. Purification 14. Mescher EJ, Platzker ACG, Ballard PL, Kitterman JA, Clements JA, Tooley of UDP-~xylose:core protein 8-D-xylosyltransferase by affinity chromatog- WH 1975 Ontogeny of tracheal fluid, pulmonary surfactant, and plasma raphy. Carbohyd Res 37:167-180 corticoids in the fetal lamb. J Appl Physiol39: 10 17- 102 1 4. Schwartz NB, Dorfman A 1975 Purification of rat chondrosarcoma xylosyl- 15. Klagsbrun M, Langer R, Levenson R, Smith S, Lillehei C 1977 The stimulation transferase. Arch Biochem Biophys 17 1: 136-144 of DNA synthesis and cell division in chondrocytes and 3T3 cells by a 5. Rodin L, Baker JR, Helting T, Schwartz NB, Stoolmiller AC, Yamagata S, growth factor isolated from cartilage. Exp Cell Res 105:99-108 Yamagata T 1972 Biosynthesis of chondroitin sulfate. In: Ginsburg V (ed) 16. Klagsburn M, Smith S 1980 Purification of a cartilagederived growth factor. Vol XXVIII Complex carbohydrates, Part B. In: Colowick SP, Kaplan NO J Biol Chem 255:10859-10866 (eds) Methods in Enzymology. Academic Press, New York, pp 638-676 17. Cloete JHL 1939 Prenatal growth in Merino sheep. Onderstepoort J Vet Sci 6. Campbell P, Jacobsson I, Benzing-Purdie L, Rod6n L, Fessler JH 1984 Silk-a Anim Ind 13:417-557