Proc. Natl. Acad. Sci. USA Vol. 87, pp. 1342-1346, February 1990 Medical Sciences A genetic defect in the biosynthesis of dermatan sulfate : Galactosyltransferase I deficiency in fibroblasts from a patient with a progeroid syndrome (galactosyltransferase iI//p-nitrophenyl (8-D-xyloside) EDELGARD QUENTIN*, ACHIM GLADEN*, LENNART RODENt, AND HANS KRESSE* *Institute of Physiological Chemistry and Pathobiochemistry, University of Munster, Munster, Federal Republic of Germany; and tDepartments of Medicine and Biochemistry, University of Alabama at Birmingham, Birmingham, AL 35294 Communicated by Elizabeth F. Neufeld, November 28, 1989

ABSTRACT A small proteoglycan that contains only a at most 80% and in some experiments as little as 20% into single dermatan sulfate chain is the main proteoglycan synthe- mature proteoglycan molecules. The remainder was secreted sized by skin fibroblasts. Fibroblasts from a patient with in a -free form. progeroidal appearance and symptoms of the Ehlers-Danlos We had discussed previously that the patient could carry a syndrome have a reduced ability of converting the core protein mutant allele yielding a core protein with an absent or buried of this proteoglycan into a mature glycosaminoglycan chain- recognition site for the attachment of the glycosaminoglycan bearing species. This abnormality is the consequence of a chain. In light of the limited induction of glycosaminoglycan deficiency in galactosyltransferase I (xylosylprotein 4- biosynthesis in the presence of low concentrations of p- ,B-galactosyltransferase; EC 2.4.1.133), which catalyzes the nitrophenyl ,B-D-xyloside, which serves as an artificial stimu- second glycosyl transfer reaction in the assembly of the der- lator ofglycosaminoglycan synthesis, an abnormality in one of matan sulfate chain. The glycosaminoglycan-free core protein the enzymes involved in the synthesis of the polysaccharide- secreted by the patient's fibroblasts bears an unsubstituted protein linkage region, GlcA(J31-3)Gal(J31-3)Gal(J31-4)Xyl, was xylose residue. The mutant enzyme is abnormally thermo- also considered. The results presented in this paper provide labile. Preincubation of fibroblasts at 41°C leads to a further evidence that the second explanation is the correct one. reduction in the production ofmature proteoglycan and affects Fibroblasts from the patient contain reduced activities of an the capacity for glycosaminoglycan synthesis on p-nitrophenyl abnormally thermolabile galactosyltransferase I (xylosylpro- f8-D-xyloside more strongly in the mutant than in control cells. tein 4-,B-galactosyltransferase; EC 2.4.1.133), which catalyzes the second glycosyl transfer reaction in the assembly of the The occurrence of inborn errors in the metabolism of con- xylose/serine-linked (see ref. 13 for a review). nective tissue proteoglycans in humans has long been rec- ognized, and over the past two decades the basic genetic MATERIALS AND METHODS defects in many ofthese disorders have been elucidated. The majority of the diseases are caused by faulty degradation of Materials. O-,8-D-Xylopyranosyl-L-serine and O-,/-D-ga- the polysaccharide components ofthe proteoglycans (see ref. lactopyranosyl-(1-4)-O-/3-D-xylopyranosyl-L-serine were 1 for a review). In a few instances, defects in the biosynthesis synthesized as described (14). UDP[4,5-3H]galactose (spe- of proteoglycans have been demonstrated in experimental cific radioactivity, 1.5 GBq/mol) was obtained from DuPont. animals, such as impaired posttranslational processing of the Sodium boro[3H]hydride (929 TBq/mol), L-[4,5-3H]leucine cartilage proteoglycan core protein in nanomelic chicken (2) (1.8 TBq/mol), and sodium [35S]sulfate (carrier-free) were and a tissue-specific deficiency in the biosynthesis of 3'- from Amersham. A high-performance carbohydrate analysis phosphoadenylylsulfate in brachymorphic mice (3). Defec- column was purchased from Millipore Waters; Aminex HPX- tive biosynthesis of proteoglycans is probably the cause of 87 H and TSK DEAE-5PW Bio-Gel columns were from the connective tissue abnormalities observed in several dis- Bio-Rad. eases in humans, including macular corneal dystrophy (4) and Assay of Galactosyltransferases I and II. Human skin fibro- a form of spondyloepiphyseal dysplasia (5), though the de- blasts were cultured as described (15). To minimize the fects in these diseases have not yet been elucidated. concentration of endogenous galactose acceptors, confluent We have previously described a patient who represented a cultures were incubated for 3 hr with 20 ,uM cycloheximide progeroid variant with signs of the Ehlers-Danlos syndrome prior to harvesting according to Esko et al. (16). During the (6). In addition to aged appearance, developmental delay, course of the experiments, however, it became apparent that dwarfism, craniofacial dysproportion, and generalized os- this pretreatment had been unnecessary. Cells were homog- teopenia, this patient suffered from defective wound healing, enized by ultrasonication in 50 mM 2-(N-morpholino)ethane- hypermobilejoints, hypotonic muscles, and loose but elastic sulfonic acid/200 mM KCI/0.05% Triton X-100/20 ,M phen- skin. Biochemically, his cultured skin fibroblasts were de- ylmethylsulfonyl fluoride/2 ,M leupeptin, pH 5.5. Fifty fective in the biosynthesis of a ubiquitous proteoglycan microliters of the suspension containing 50-100 ,g of cell named small dermatan sulfate proteoglycan II (DS-PG II; protein (17) was mixed with 30 ,ul of 16.25 mM Tris/acetate refs. 7 and 8) or decorin (9), which consists of an Mr 36,319 buffer (pH 7.5) containing 37 kBq (86 pmol) of UDP- core protein (10), a single glycosaminoglycan chain on the [3H]galactose, 16.25 mM KCI, 12.5 mM 2,3-dimercaptopro- serine residue at position 4 (11), and either two or three panol, 31 mM MnC12, 0.25 mM ATP, 52 mM CDP-choline, asparagine-bound oligosaccharides (12). The fibroblasts syn- and either 13 mM xylosylserine (galactosyltransferase I) or thesized normal amounts of this core protein but converted 21.8 mM galactosylxylosylserine (galactosyltransferase II, EC 2.4.1.134). After 1 hr at 37°C, the incubation mixture was processed exactly as described (14). Briefly, proteins were The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact. Abbreviation: DS-PG I1, small dermatan sulfate proteoglycan II. 1342 Downloaded by guest on September 29, 2021 Medical Sciences: Quentin et al. Proc. Natl. Acad. Sci. USA 87 (1990) 1343 precipitated with ethanol, and the soluble material was sub- preincubated for 48 hr at 37°C or 41°C, respectively, and jected to cation-exchange chromatography for the purifica- incubated with [35S]sulfate in the presence of p-nitrophenyl tion ofthe radioactive product. Without exception, the purity ,-D-xyloside for 4 hr at 37°C. Proteoglycans were removed of the radioactive products, Gal-Xyl-Ser or Gal-Gal-Xyl-Ser, by an (NH4)2SO4 precipitation step, and induced chains were was subsequently checked by high-voltage electrophoresis purified by chromatography on Dowex AG 1X2 (200-400 (14) and determination of the 3H radioactivity in 1-cm paper mesh). For the determination ofthe glycosaminoglycan chain segments. By-products accounted for up to 10% in the length, DS-PG II was isolated after an incubation period of 4 control cell lines and 50o in the patient's fibroblasts, and the hr in the presence of 0.37 MBq of [35S]sulfate per ml. ion-exchange data were corrected correspondingly. Dermatan sulfate chains were obtained by a 8-elimination Preparation of DS-PG II and Its Core Protein. Conditioned reaction and chromatographed on a calibrated Sephacryl medium that contained only insulin and transferrin as exog- S-300 column as described (18). enously supplied proteins (18) served as the source of unla- beled DS-PG II and core protein. Glycosaminoglycan-free core protein and DS-PG II were separated from each other by RESULTS anion-exchange chromatography (12). Core protein-con- Presence of Unsubstituted Xylose Residues. Intact DS-PG II taining fractions were dialyzed against 0.1% Triton X-100, and its glycosaminoglycan-free core protein from the secre- lyophilized, and immune precipitated with an affinity- tions of the patient's fibroblasts were separately purified for purified antiserum against DS-PG II core protein (12). DS-PG an analysis of constituents of the polysaccharide protein II-containing fractions were similarly dialyzed and Iyo- linkage region. Alkaline borotritide reduction of both prep- philized but were then exhaustively digested with chondroitin arations led to the incorporation of similar amounts of ABC Iyase (19). Both preparations were subjected to prepar- radioactivity into glycosaminoglycan-free core protein and ative SDS/polyacrylamide gel electrophoresis (20). Proteins chondroitin ABC lyase-digested DS-PG II, the former con- were visualized by treating the gels with 1 M KCI (21), and taining 77% of the radioactivity of the latter. After further bands ofinterest were electroeluted. After drying the samples purification by descending paper chromatography, material under reduced pressure, salts were removed by washing with behaving as authentic [1-3H]xylitol was found upon chroma- methanol. The secretion of [3H]leucine-labeled DS-PG II and tography on either an Aminex HPX-87 H or a carbohydrate core protein was followed as described (12). analysis HPLC column when the glycosaminoglycan-free Identification of Xylose Residues. Core protein and chon- core protein was analyzed (Fig. 1). Acid hydrolysis prior to droitin ABC Iyase-digested DS-PG II from 240 ml of condi- the chromatographic separations did not result in an in- tioned medium were dissolved in 20 ILI of 0.92 M Na3HB4 creased recovery of [3H]xylitol, indicating that the alditol was (specific activity, 929 GBq/mol) in 0.1 M NaOH and incu- not substituted with galactose residues. Analogous treat- bated for 21 hr at 50'C. After acidification to pH 5 with acetic ments ofenzyme-digested DS-PG II resulted in the formation acid, each sample was concentrated to dryness, and the of [3H]xylitol after acid hydrolysis only. residue was evaporated three times with methanol. The Deficiency of Galacytosyltransferase I. The presence of samples were then subjected to descending paper chroma- unsubstituted xylose residues on the glycosaminoglycan-free tography for 14 hr on Schleicher & Schuell paper no. 2043 core protein could result from a deficiency of galactosyl- BMGL in butanol/acetone/H20, 2:7:1. Samples that had transferase I. Indeed, the activity of this enzyme in homoge- been hydrolyzed under nitrogen with 1 M HCl for 4 hr at nates of cultured fibroblasts from the patient was only about 105'C were analyzed in parallel. The radioactive peak mi- 5% of that in normal control cells (Table 1). The low activity grating as authentic [3H]xylitol was eluted with H20 and was not due to enhanced degradation of the enzyme during applied either to a carbohydrate analysis column or to an the 3-hr preincubation ofthe cultures with cycloheximide that Aminex HPX-87 H column. The first column was eluted at was carried out routinely, as no decrease in activity occurred ambient temperature with 80% acetonitrile in H20 at a flow during this period in control experiments with normal and rate of 1 ml/min; the second column was operated at 40°C and deficient cells. Mixing experiments excluded the presence of was eluted with 3 mM H2SO4 in 15% acetonitrile in H20 at a excessive amounts of an inhibitor of galactosyltransferase I. flow rate of 0.5 ml/min. The cells of the parents of the patient contained half the Other Methods. The secretion ofglycosaminoglycan chains normal activity, in accordance with the assumption that the induced on exogenously added p-nitrophenyl /3-D-xyloside disorder is inherited in an autosomal recessive manner. It will was measured as described (6). Briefly, fibroblasts were be shown below that the normal and the mutant enzymes

15- xylitol 1 2 core |core,HCI treated |PG, HCI treated

E 10 C.) x

co 5

Elution volume (ml)

FIG. 1. Identification ofxylitol by chromatography on a carbohydrate analysis column. Glycosaminoglycan-free core protein (core) and intact DS-PG 1I (PG), both from the secretions of the patient's fibroblasts, were subjected to alkaline borotritide reduction, and the products were purified. Similar amounts of radioactivity were applied for each run. The arrow marked 1 indicates the elution maximum of N- acetyl[1-3H]galactosaminitol; the arrow marked 2 indicates the elution maximum of [1-3H]galactitol. Downloaded by guest on September 29, 2021 1344 Medical Sciences: Quentin et al. Proc. Natl. Acad. Sci. USA 87 (1990)

Table 1. Activities of galactosyltransferases I and II in 100- A cultured fibroblasts Galactosyl- Galactosyl- transferase I, transferase II, 80s- fmol/min per fmol/min per x Enzyme source mg of protein mg of protein 22 4 Patient 60 - Deviation 13-53* (9) 2-8* (4) a Mother 277 10 E Father 274 15 559 20 40 - Controls 4- Range of 1 standard deviationt 334-794 16-26 0 No. of cell strains tested 9 5 - Formation of 1 fmol of product is represented by the incorporation 20 A of 26 dpm of 3H radioactivity. *The number of separate determinations is given in parentheses. tCalculated for logarithmic normal distribution. differ in their thermostability. On incubation at 370C, product 20 30 40 50 formation by either enzyme did not increase linearly with Temperature ( 0 C) time (Fig. 2). The low activity of the mutant enzyme made it necessary, however, to choose 1 hr as the standard incuba- FIG. 3. Effect of temperature on galactosyltransferase I activity. tion period. Extracts from normal cells (e) and from the patient's fibroblasts (A) Surprisingly, the patient's fibroblasts were also deficient in were incubated for 1 hr at the temperatures indicated. The extracts galactosyltransferase II, the activity being only 20% of the contained 90 ,ug of protein in both cases. of 1 standard mean normal value. Activities below the range twice as high as the control value (3.4 mmol/liter). With deviation were found in the cells from his parents (Table 1). respect to UDP-galactose, Km values of 0.026 mmol/liter Galactosyltransferase I activity was also determined by (patient) and 0.17 mmol/liter (control) were calculated and using p-nitrophenyl 83-D-xyloside (5 mM) as substrate under the activity of the mutant was only about 1% at substrate otherwise identical conditions. The radioactive product was saturation. Similar pH activity profiles with sharp maxima at quantitated by chromatography on a 1 x 7 cm Bio-Gel P2 pH 6.0 were observed for the normal and the mutant enzyme column in H20 where its elution volume was 1.75 times that (results not shown). of 3H20. Two control cell strains exhibited activities of 480 In a fibroblast homogenate, galactosyltransferase I did not and 600 fmol/min per mg ofprotein, respectively, whereas in function optimally at 370C. Instead, the normal enzyme the patient's cells the activity was below the limit ofdetection exhibited the highest activity at 30°C, and the mutant enzyme (<50 fmol/min per mg). was most active at 250C, the lowest temperature tested (Fig. Properties of Mutant Galactosyltransferase I. Some differ- 3). There were no differences in the temperature optima of ences were noted between the Km values of normal and galactosyltransferase II, which were at 300C in both cases. mutant galactosyltransferase I. For xylosylserine the Km The effect of temperature on glycosaminoglycan chain value of the patient's enzyme (6.4 mmol/liter) was about formation in intact cells is illustrated in Fig. 4. After prein- cubation at different temperatures for 2 days, fibroblast cultures were pulse-labeled with [3H]leucine for 30 min at 370C and then analyzed by SDS/polyacrylamide gel electro- phoresis. Following preincubation at 41°C, 80% of the im- 20 - munoreactive radioactive products synthesized by the pa- tient's cells was glycosaminoglycan-free core protein. In 0~~~~~~~

E 300C 0- 13°C

E .02 9754- .6 10 U 69- _0 0~ 46--

20 40 60 FIG. 4. Influence of temperature on DS-PG II core protein secretion. Fibroblasts from a control person (C) and from the patient Time (min) (P) were kept for 48 hr at the temperatures indicated and then labeled with [3H]leucine (3.7 MBq/ml) for 30 min at 37°C. The immune FIG. 2. Time dependence of galactosyltransferase I activity. precipitate from secreted material was directly applied to SDS/ Extracts from normal cells (e, 78 jg ofprotein) and from the patient's polyacrylamide gels for electrophoresis. The position of 14C- fibroblasts (A, 83 ,ug of protein) were incubated at 37°C for the times methylated molecular weight standards (Mr X 10-3) is shown on the indicated. left margin. Downloaded by guest on September 29, 2021 Medical Sciences: Quentin et al. Proc. Nati. Acad. Sci. USA 87 (1990) 1345 Table 2. Influence of temperature on biosynthesis of induced observed in vitro should have reduced substantially the glycosaminoglycan chains capacity of intact fibroblasts to synthesize mature DS-PG II. 35S, cpm x 10-3 per mg of protein A possible explanation of the fact that this did not occur is p-Nitrophenyl Patient Control that the physiological acceptor-the xylosylated core pro- p-D-xyloside, tein-may be a much better substrate than the two artificial mM 370C 410C 370C 410C substrates used in this study. Attempts were therefore made to use the core protein secreted by the patient's fibroblasts as 0.1 140 41 310 180 a galactose acceptor for partially purified galactosyltrans- 1.0 490 170 ND ND ferase I. Although galactose was indeed incorporated, char- Fibroblasts were preincubated for 48 hr at the temperatures acterization of the product indicated that the sugar had been indicated and then incubated with [35S]sulfate (0.37 MBq/ml) for 4 hr transferred only to asparagine-bound oligosaccharides (E.Q., at 370C. ND, not determined. unpublished work). A second explanation of the high level of proteoglycan contrast, mature proteoglycan molecules were the major synthesis in the intact cells is that the mutant enzyme may be products after preincubation at 370C and 300C, and free core more stable in its native environment than in the cell-free protein accounted for only 18% and 13%, respectively, ofthe extract. Galactosyltransferase I in normal cells is firmly total radioactivity in the immunoreactive material. Further- bound to intracellular membranes, and disruption of the more, it should be noted that the total incorporation of membrane structure-e.g., by digestion with phospholipase [3H]leucine, in control as well as patient cells, was reduced C-results in partial inactivation (22). Whereas solubilized, by about 60% when the preincubation was carried out at 410C purified preparations of the normal enzyme are fairly stable rather than at 300C or 370C. In addition to the temperature in the presence of appropriate phospholipids or nonionic sensitivity, Fig. 4 also indicated that DS-PG II from the detergents, the mutant enzyme may be more sensitive to patient's fibroblasts migrated more slowly than the product of perturbations of its native environment and may have lost normal cells. Labeling with [35S]sulfate at 370C showed that activity during preparation of the cell-free extract. this abnormality was caused by an increased glycosamino- The possibility should also be considered that a "nonspe- glycan chain length in the mutant (Mr 42,000 versus Mr cific" galactosyltransferase may augment the activity of 35,000), as determined by gel filtration on a calibrated Seph- galactosyltransferase I in the mutant cells in vivo. Lactose acryl S-300 column. Dermatan sulfate chains from the parents synthase is known to catalyze galactose transfer to xylose, were of normal size. albeit with low efficiency. However, the increased thermal Temperature Dependence of the Synthesis of Induced Gly- sensitivity of the patient's fibroblasts with respect to forma- cosaminoglycan Chains. We have shown previously that the tion of DS-PG II argued against the involvement of a non- patient's fibroblasts did not respond normally to induction of specific normal galactosyltransferase. The same conclusion glycosaminoglycan chain formation by p-nitrophenyl /8- emerges from the results of experiments with the artificial D-Xyloside, an artificial substrate for galactosyltransferase I substrate, p-nitrophenyl 8-D-xyloside. Again illustrating the (6). Whereas a xyloside concentration of 0.1 mM was suffi- discrepancy between the results of in vitro and in vivo cient for maximal response in normal cells, the mutant cells studies, the experiments showed no measurable galactosyl- produced maximal amounts of glycosaminoglycan chains transferase I activity in extracts of the patient's cells at a only at a concentration of 1 mM (6). Examination ofthe effect xyloside concentration of 5 mM, but a marked stimulation of of temperature on the induction of glycosaminoglycan chain glycosaminoglycan synthesis occurred in vivo in the presence formation (Table 2) showed that mutant and control cells had of 1 mM xyloside. At elevated culture temperature, the a reduced ability to synthesize free chains after preincubation decrease in glycosaminoglycan chain formation was substan- at 41°C rather than 37°C. However, the patient's fibroblasts tially greater in the mutant cells than in normal cells, sup- were about twice as sensitive as normal cells. porting the view that the mutant enzyme and not a nonspe- cific galactosyltransferase was the target of the thermal effect. DISCUSSION Interestingly, a deficiency in galactosyltransferase II, This investigation has shown that cultured fibroblasts from a which catalyzes the third glycosyl transfer step in polysac- patient with a progeria-like syndrome and signs ofthe Ehlers- charide chain assembly, was also observed. Although an Danlos syndrome are severely deficient in galactosyltrans- explanation of this finding is not yet apparent, the possibility ferase I, which catalyzes the second glycosyl transfer step in may be suggested that the two galactosyltransferases share a the biosynthesis of several connective tissue polysaccha- common subunit, which is the target ofthe mutation. It is also rides. Additional findings indicate that this deficiency is the possible that galactosyltransferase II participates in the for- primary genetic defect in this syndrome and that it is inherited mation of a multienzyme complex containing all glycosyl- in an autosomal recessive manner. Thus, the patient's en- transferases required for synthesis ofthe glycosaminoglycan- zyme was more thermolabile than its normal counterpart, a protein linkage region and that the instability of galactosyl- property often observed in mutant enzymes. Furthermore, transferase I may render other components of the complex both parents exhibited only half of the mean normal enzyme more labile. The known ability of galactosyltransferase I to activity in their fibroblasts. form complexes with xylosyltransferase in vitro (23, 24) is Consonant with the profound enzyme deficiency was the consistent with this notion. On the other hand, no deficiency accumulation in the culture medium of incomplete proteo- in xylosyltransferase was observed in the patient's cells, and glycan molecules without a glycosaminoglycan chain but the finding that xylosyltransferase and galactosyltransferase possessing an unsubstituted xylose residue attached to the I are not coordinately synthesized and degraded (22) leaves core protein of DS-PG II. However, mature proteoglycan doubts as to the physiological importance of the complex molecules with glycosaminoglycan chains ofincreased length formation observed in vitro. were also present and represented, in several experiments, The different /dermatan sulfate and the majority of the material reacting with an antibody to the heparan sulfate proteoglycans contain the same glycosami- DS-PG II core protein. Although we do not know what noglycan-protein linkage region (13). Studies with Chinese fraction of the normal galactosyltransferase I activity is hamster ovary cell mutants suggested that a single xylosyl- sufficient to sustain a normal level of proteoglycan biosyn- transferase (16) and a single galactosyltransferase I (25) are thesis in vivo, it seems likely that the profound deficiency involved in the biosynthesis of these proteoglycans. It is not Downloaded by guest on September 29, 2021 1346 Medical Sciences: Quentin et al. Proc. Natl. Acad. Sci. USA 87 (1990) known whether human fibroblasts express one or several 6. Kresse H., Rosth0j, S., Quentin, E., Hollmann, J., Glossl, J., proteins with galactosyltransferase I activity. If human fi- Okada, S. & T0nnesen, T. (1987) Am. J. Hum. Genet. 41, broblasts behave like Chinese hamster ovary cells, defects in 436-453. the biosynthesis of the large dermatan sulfate proteoglycan 7. Heinegard, D., Bjorne-Persson, A., Coster, L., Franzen, A., and of heparan sulfate proteoglycans would be expected in Gardell, S., Mamstrom, A., Paulsson, M., Sandfalk, R. & the patient. However, normal urinary excretion of heparan Vogel, K. (1985) Biochem. J. 230, 181-194. sulfate and normal [35S]sulfate incorporation into 8. Rosenberg, L. C., Choi, H. U., Tang, L.-H., Johnson, T. L., fibroblast Pal, S., Webber, C., Reiner, A. & Poole, A. R. (1985) J. Biol. proteoglycans other than DS-PG II have been observed Chem. 260, 6304-6313. previously (6). In light of the fact that the mature DS-PG II 9. Yamaguchi, Y. & Ruoslahti, E. (1988) Nature (London) 336, from the patient contains abnormally long dermatan sulfate 244-246. chains, P'S]sulfate incorporation studies do not exclude the 10. Krusius, T. & Ruoslahti, E. (1986) Proc. Nat!. Acad. Sci. USA possibility that the core proteins ofthe large dermatan sulfate 83, 7683-7687. proteoglycan and/or of heparan sulfate proteoglycans are 11. Chopra, R. K., Pearson, C. H., Pringle, G. A., Fackre, D. S. substituted with a reduced number of glycosaminoglycan & Scott, P. G. (1985) Biochem. J. 232, 277-279. chains. Further studies are required to exclude with certainty 12. Glossl, J., Beck, M. & Kresse, H. (1984) J. Biol. Chem. 259, a defect in the biosynthesis of these proteoglycans. 14144-14150. In spite ofthe many unanswered questions outlined above, 13. Roden, L. (1980) in The Biochemistry of Glycoproteins and it appears certain that we have observed a human defect in Proteoglycans, ed. Lennarz, W. J. (Plenum, New York), pp. the biosynthesis of the carbohydrate backbone of a connec- 267-371. tive tissue and the 14. Ekborg, G., Klinger, M., Roden, L., Jensen, J. W., Schutz- polysaccharide, diploid human-cell line bach, J. S., Huang, D. H., Krishna, N. R. & Anantharamaiah, carrying the defect should be a valuable tool in future studies G. M. (1987) Glycoconjugate J. 4, 255-266. of various cell/matrix interactions. 15. Cantz, M., Kresse, H., Barton, R. W. & Neufeld, E. F. (1972) Methods Enzymol. 28, 884-897. We are indebted to Mrs. M. Bahl for skillful technical assistance 16. Esko, J. D., Stewart, T. G. & Taylor, W. H. (1985) Proc. Nat!. and to Dr. G. Ekborg for gifts of xylosylserine and galactosylxylo- Acad. Sci. USA 82, 3197-3201. sylserine. This work was financially supported by the Deutsche 17. Lowry, 0. H., Rosebrough, N. J., Farr, A. L. & Randall, R. J. Forschungsgemeinschaft and by National Institutes ofHealth Grants (1951) J. Biol. Chem. 193, 265-267. DE 08252 and NS 27353. 18. Rauch, U., Glossl, J. & Kresse, H. (1986) Biochem. J. 238, 465-474. 1. Neufeld, E. F. & Muenzer, J. (1989) in The Metabolic Basis of 19. Saito, H., Yamagata, T. & Suzuki, S. (1968) J. Biol. Chem. 243, Inherited Disease, eds. Scriver, C. R., Beaudet, A. L., Sly, 1536-1542. W. S. & Valle, D. (McGraw-Hill, New York), 6th Ed., pp. 20. Laemmli, U. K. (1970) Nature (London) 277, 680-685. 1565-1588. 21. Nelles, L. P. & Blomburg, J. R. (1976) Anal. Biochem. 73, 2. O'Donnell, C. M., Kaczman-Daniel, K., Goetinck, P. F. & 522-531. Vertel, B. M. (1988) J. Biol. Chem. 263, 17749-17754. 22. Schwartz, N. B. (1976) J. Biol. Chem. 251, 285-291. 3. Sugahara, K. & Schwartz, N. B. (1982) Arch. Biochem. Bio- 23. Schwartz, N. B., Roden, L. & Dorfman, A. (1974) Biochem. phys. 214, 602-609. Biophys. Res. Commun. 56, 717-724. 4. Nakazawa, K., Hassell, J. R., Hascall, V, C., Lohmander, 24. Schwartz, N. B. & Roden, L. (1975) J. Biol. Chem. 250, L. S., Newsome, D A. & Krachmer, J. (1984) J. Biol. Chem. 5200-5207. 259, 13751-13757. 25. Esko, J. D., Weinke, J. L., Taylor, W. H., Ekborg, G., Roden, 5. Mourao, P. A. S., Kato, S. & Donnelly, P. V. (1981) Biochem. L., Anantharamaiah, G. & Gawish, A. (1987) J. Biol. Chem. Biophys. Res. Commun. 98, 388-396. 262, 12189-12195. Downloaded by guest on September 29, 2021