Identification of the site of sulfation of the fourth component of human complement

The alpha-chain of the fourth component of complement (C4) contains tyrosine sulfate (Karp, D.R. (1983) J. Biol. Chem. 258, 12745-12748). Here we have determined the site and stoichiometry of sulfation of C4 secreted by the human hepatoma-derived cell line Hep G2. C4 was labeled with [35S]sulfate and isolated from culture medium by immunoprecipitation. C4 digested with trypsin and chymotrypsin and analyzed by reverse-phase high-performance liquid chromatography contained a single sulfate-labeled peptide. Digestion of C4 with trypsin alone yielded two major sulfate-labeled peptides, suggesting that there may be some sequence variability in C4 near the site of sulfation. Sequential Edman degradation of tryptic peptides labeled with [3H]tyrosine and [35S]sulfate detected tyrosine residues at positions 5, 13, 16, and 18. Chymotrypsin cleaved 5 residues off the NH2-terminal end of tryptic peptides, yielding a peptide with tyrosine at positions 8, 11, and 13. Comparison of the position of tyrosine residues with the reported sequence of C4 identified the sites of sulfation as tyrosine residues at positions 738, 741, and 743 in the alpha-chain of C4. All 3 of these tyrosine residues appeared to be sulfated. When sulfation of C4 was partially inhibited by addition of catechol to culture medium, three different forms of the peptide were resolved by high-performance liquid chromatography, consistent with peptides containing 1, 2, or 3 sulfates. Comparison of the quantities of tyrosine and tyrosine sulfate in C4 which had been labeled with [3H]tyrosine and digested with Pronase also indicated that C4 contained an average of 2-3 residues of tyrosine sulfate/molecule. These results suggest that the biologically active form of the protein is sulfated.     

A target-specific approach for the identification of tyrosine-sulfated hemostatic proteins

A simple methodology for the identification of hemostatic proteins that are subjected to posttranslational tyrosine sulfation was developed. The procedure involves sequence analysis of members of the three hemostatic pathways using the Sulfinator prediction algorithm, followed by [³⁵S]sulfate labeling of cultured HepG2 human hepatoma cells, immunoprecipitation of targeted [³⁵S]sulfate-labeled hemostatic proteins, and tyrosine O-[³⁵S]sulfate analysis of immunoprecipitated proteins. Three new tyrosine-sulfated hemostatic proteins—protein S, prekallikrein, and plasminogen—were identified. Such a target-specific approach will allow investigation of tyrosine-sulfated proteins of other biochemical/physiological pathways/processes and contribute to a better understanding of the functional role of posttranslational tyrosine sulfation.     

The hormonogenic tyrosine 5 of porcine thyroglobulin is sulfated

Our previous results showed that sulfated tyrosines of thyroglobulin (Tg), the molecular support of thyroid hormonosynthesis, are involved in the hormonogenic process. Moreover, the consensus sequence required for tyrosine sulfation is present in most of the hormonogenic sites. These observations suggest that tyrosine sulfation might play a critical role in the hormonogenic process. In this paper we studied the putative sulfation of tyrosine 5 contained in the preferential hormonogenic site. Porcine thyrocytes were cultured with thyrotropin but without iodide to preserve the sulfation state of tyrosine 5 and then incubated or not with [35S]sulfate. Secreted Tg was purified and submitted to peptide sequence analysis which confirmed the known peptide sequence of the NH(2) extremity of Tg:NIFEYQV. The treatment of [35S]sulfate-labeled Tg by leucine aminopeptidase, which sequentially digested its amino-terminal extremity, released the same amino acids and further analysis by thin layer chromatography showed that the tyrosine was sulfated. We concluded that tyrosine 5 is sulfated but the role of sulfate group in the hormonogenic process remains to be elucidated.     

Tyrosine O-sulfate ester in proteoglycans

Tyrosine O-sulfate residues were detected in the protein core of sulfated proteoglycans. When cultured skin fibroblasts and arterial smooth muscle cells were incubated in the presence of [35S]sulfate, dermatan sulfate proteoglycan and chondroitin sulfate proteoglycan isolated from the culture medium contained tyrosine [35S]sulfate ester which accounted for 0.03%-0.82% of total 35S radioactivity incorporated into the sulfated proteoglycans. This corresponds to a tyrosine sulfation of every second (fibroblasts) and every 10th (smooth muscle cells) dermatan sulfate proteoglycan molecule. [3H]Tyrosine labeling of fibroblast dermatan sulfate proteoglycan gave a similar stoichiometry. However, the relative proportion of tyrosine [35S]sulfate in proteoglycans from arterial tissue was about 10 times higher than in that from cultured arterial cells. Pulse chase experiments with [35S]sulfate revealed that tyrosine sulfation is a late event in the biosynthesis of dermatan sulfate proteoglycan from fibroblasts and occurs immediately prior to secretion. Cultured skin fibroblasts from a patient with a progeroid variant (Kresse et al. 1987, Am. J. Hum. Gen. 41, 436-453) which exhibit a partial deficiency to synthesize dermatan sulfate proteoglycan were shown to form and to secrete a tyrosine-sulfated but glycosaminoglycan-free protein core, thus confirming a selective and independent [35S]sulfate labeling of the protein core.     

Tyrosine sulfation in precursors of collagen V

Radioactive labeling of p-collagens V, collagens V, and, to a small extent, of procollagen V occurred when [35S]sulfate was incubated with tendons or primary tendon cell cultures, or blood vessels and crops of 17- to 19-day-old chick embryos, or with lung slices from neonatal rats. Most or all of this label is in the form of 1 or more sulfated tyrosine residues/chain of p alpha 1(V), alpha 1(V), p alpha 1'(V), alpha 1'(V), p alpha 2(V), and alpha 2(V), and it remains attached through purification by dialysis, ammonium sulfate precipitation, CsCl-GdnCl2 equilibrium buoyant density and velocity sedimentations, ion-exchange chromatography, and sodium dodecyl sulfate gel electrophoresis. Radioactive tyrosine sulfate was identified in alkaline hydrolysates of these collagen V chains, after labeling the tissues with either [35S]sulfate or [3H]tyrosine, by electrophoretic and chromatographic comigration with a tyrosine sulfate standard. Tunicamycin A1, which inhibits the attachment of N-linked complex carbohydrate, did not interfere with the sulfation process. The tyrosine sulfate is located in a noncollagenous domain, which is probably adjacent to the amino end of the collagen helix, and is retained throughout the physiological proteolytic processing of procollagens V. After digestion with Staphylococcus aureus V8 protease, 35S-labeled p alpha 1(V) and alpha 1(V) chains gave the same map of labeled peptides, and this differed from the map given by p alpha 1'(V) and alpha 1'(V) chains. Little sulfation of p alpha 2(V) and alpha 2(V) chains occurs. The implications of these observations for the structure and properties of procollagens V and their derivatives are considered.     

Tyrosine O-sulfation of the fibrinogen gamma B chain in primary cultures of rat hepatocytes

Sulfation of fibrinogen was studied in a primary culture of rat hepatocytes. After cells were incubated with [35S]sulfate, 35S-labeled fibrinogen was obtained from the medium by immunoprecipitation and analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis/fluorography. It was demonstrated that [35S]sulfate is exclusively incorporated into the gamma B chain, which is a minor variant form found in rat fibrinogen, in addition to a major gamma A chain. When the purified 35S-gamma B chain was digested with carboxypeptidase Y, the radioactivity was almost completely released from the protein, and the labeled product released was identified as tyrosine O-sulfate. Based on the available primary structure of the gamma B chain, the results suggest that sulfation occurs on the tyrosine residue at the second position from its COOH terminus. Pulse-chase experiments using both [3H]leucine and [35S]sulfate showed that 35S-labeled fibrinogen is secreted into the medium much faster than the 3H-labeled molecule. Incubation of cells with monensin, an inhibitor of Golgi function, strongly inhibited the sulfation of fibrinogen. In addition, in vitro sulfation experiments demonstrated that sulfotransferase activity is localized in the Golgi fraction. These results indicate that the sulfation of fibrinogen takes place in the Golgi complex, especially in the trans Golgi region, just before its secretion.     

Sulfated tyrosines of thyroglobulin are involved in thyroid hormone synthesis.

Thyroid hormone synthesis is under the control of thyrotropin (TSH), which also regulates the sulfation of tyrosines in thyroglobulin (Tg). We hypothesized that sulfated tyrosine (Tyr[S]) might be involved in the hormonogenic process, since the consensus sequence required for tyrosine sulfation to occur was observed at the hormonogenic sites. Porcine thyrocytes, cultured with TSH but without iodide in the presence of [(35)S]sulfate, secreted Tg which was subjected to in vitro hormonosynthesis with increasing concentrations of iodide. A 63% consumption of Tyr[S] (1 residue) was observed at 40 atoms of iodine incorporated into Tg, corresponding to a 40% hormonosynthesis efficiency. In addition, hyposulfated Tg secreted by cells incubated with sodium chlorate was subjected to in vitro hormonosynthesis. With 0.5 Tyr[S] residue (31% of the initial content), the efficiency of the hormonosynthesis was 29%. In comparison, when hormonosynthesis was performed by cells, with only 0.25 Tyr[S] residue (16% of the initial content), the hormonosynthesis efficiency fell to 18%. These results show that there exists a close correlation between the sulfated tyrosine content of Tg and the production of thyroid hormones.     

Tyrosine sulfation site is located in the C-terminal fibrin-binding domain in secreted fibronectin from rat embryo fibroblasts, line 3Y1

Tryptic fragments of [35S]sulfate-labeled 3Y1 secreted fibronectin were fractionated by hydroxylapatite column chromatography and examined using sodium dodecyl sulfate gel electrophoresis, followed by autoradiography. Radioactive bands containing tyrosine-O-[35S]sulfate were detected at 17- and 40-kDa positions under reducing conditions. Under nonreducing conditions, the 17-kDa band was no longer present and new bands at 57- and 80-kDa positions appeared, indicating a disulfide linkage between the two smaller fragments in the native state. These fragments exhibited binding affinity toward fibrin and could be immunoprecipitated by the monoclonal antifibronectin Fib-2 domain antibody. These results suggested that the tyrosine sulfation site in 3Y1 secreted fibronectin is located in the C-terminal fibrin-binding (Fib-2) domain, being within 17 kDa of the C-terminus.     

Identification of two sites of sulfation of human heparin cofactor II

Tyrosine sulfate was identified as a constituent of human heparin cofactor II by analysis of sulfate-labeled protein secreted by a human hepatoma-derived cell line and of purified protein from human plasma. Alkaline hydrolysis of heparin cofactor II released tyrosine sulfate as demonstrated by anion-exchange high performance liquid chromatography of hydrolysates. Two sites of sulfation were identified, and the amino acid sequences of the sites were established by sequential Edman degradation of sulfate-containing tryptic peptides that were isolated by reverse-phase high performance liquid chromatography. Each peptide contains only a single tyrosine residue so that the sites of sulfation can be assigned unambiguously. The two sites of sulfation are separated by 13 residues and represent an internal sequence repeat in the heparin cofactor II molecule. The two sites have the following sequences. Glu56-Asp-Asp-Asp-Tyr(SO4)-Leu-Asp62 Glu69-Asp-Asp-Asp-Tyr(SO4)-Ile-Asp75 Sulfate-labeled heparin cofactor II formed a covalent complex with thrombin in a heparin-dependent manner. Thus, the sulfate-containing form of the protein was shown to be biologically active. The characteristic sulfate-containing segment of heparin cofactor II, which contains 17 acidic amino acid residues over a span of 30 residues, may contribute to the unique properties of this thrombin inhibitor.     

Sulfation of a gamma-chain variant of human fibrinogen

Biosynthetic sulfation of human fibrinogen was investigated using a hepatoma-derived cell line in culture. Very little [35]sulfate was incorporated into the major forms of the A alpha, B beta, or gamma-chains of fibrinogen, but there was a labeled peptide chain with electrophoretic mobility intermediate between the B beta and gamma-chains. Base hydrolysis of the sulfate-labeled product released tyrosine sulfate. The labeled peptide is identified as a form of gamma-chain by its resistance to proteolysis during extended periods of incubation, in contrast with A alpha and B beta-chains which are converted to smaller forms. The results indicate that human fibrinogen contains tyrosine sulfate primarily within a variant form of the gamma-chain.     

Chlorate: a reversible inhibitor of proteoglycan sulfation

Bovine aorta endothelial cells were cultured in medium containing [3H]glucosamine, [35S]sulfate, and various concentrations of chlorate. Cell growth was not affected by 10 mM chlorate, while 30 mM chlorate had a slight inhibitory effect. Chlorate concentrations greater than 10 mM resulted in significant undersulfation of chondroitin. With 30 mM chlorate, sulfation of chondroitin was reduced to 10% and heparan to 35% of controls, but [3H]glucosamine incorporation on a per cell basis did not appear to be inhibited. Removal of chlorate from the culture medium of cells resulted in the rapid resumption of sulfation.     

Sulfation of two tyrosine-residues in human complement S-protein (vitronectin)

Human S-protein (vitronectin) and hemopexin, two structurally related plasma proteins of similar molecular mass and abundance, were analyzed for tyrosine sulfation. Both proteins were synthesized and secreted by the human hepatoma-derived cell line Hep G2, as shown by immunoprecipitation from the culture medium of [35S]methionine-labelled cells. When Hep G2 cells were labelled with [35S]sulfate, S-protein, but not hemopexin, was found to be sulfated. Half of the [35S]sulfate incorporated into S-protein was recovered as tyrosine sulfate. The stoichiometry of tyrosine sulfation was approximately two mol tyrosine sulfate/mol S-protein. Examination of the S-protein sequence for the presence of the known consensus features for tyrosine sulfation revealed three potential sulfation sites at positions 56, 59 and 401. Tyrosine 56 is the most probable site for stoichiometric sulfation, followed by tyrosine 59 which appears more likely to become sulfated than tyrosine 401. Tyrosines 56 and 59 are located in the anionic region of S-protein which has no homologous counterpart in hemopexin. We discuss the possibility that tyrosine sulfation of the anionic region of S-protein may stabilize the conformation of S-protein in the absence of thrombin-antithrombin III complexes and may play a role in its binding to thrombin-antithrombin III complexes during coagulation.     

Biosynthesis of gastric mucus glycoprotein of the rat

We have studied the biosynthesis of rat gastric mucin in stomach segments using an antiserum against rat gastric mucin specific for peptide epitopes. Pulse-chase experiments were performed with [35S]methionine, [3H]galactose, and [35S]sulfate to label mucin precursors in different stages of biosynthesis, which were analyzed after immunoprecipitation. The earliest mucin precursor that could be detected by sodium dodecyl sulfate-polyacrylamide gel electrophoresis was a 300-kDa protein. The occurrence of N-linked "high-mannose" oligosaccharides on this protein was shown by susceptibility to degradation by endo-beta-N-acetylglucosaminidase H. This precursor could be labeled with [35S]methionine and not with [3H]galactose or [35S]sulfate. The 300-kDa precursor was converted into mature mucin after extensive glycosylation and sulfation. The mature mucin but not the 300-kDa precursor was in part secreted into the medium. Specific inhibition of sulfation with sodium chlorate had no effect on rate and amount of mucin secretion. In addition, we show that two core proteins are expressed in rats, slightly varying in Mr among individual animals.     

Gonadotropin-releasing hormone action upon luteinizing hormone bioactivity in pituitary gland: role of sulfation

The regulation of rat luteinizing hormone (rLH) bioactivity was studied in an in vitro system using isolated pituitaries from male rats. Stored and released rLH was evaluated in terms of mass (I-LH), bioactivity (B-LH), mobility in nonequilibrium pH gradient electrophoresis, and mannose and sulfate incorporation either in the presence or absence of gonadotropin-releasing hormone (GnRH). GnRH increased the biological potency of stored and released rLH. The pituitary content revealed seven I-LH species (pH 7.2, 7.8, 8.5, 9.0, 9.1, 9.3, and 9.7) and five B-LH species (pH 8.5, 9.0, 9.2, 9.4, and 9.7). The major I-LH and B-LH peaks were at pH 9.0 and 9.2, respectively. I-LH peaks at pH 7.2 and 7.8 are devoid of bioactivity; at these pH values, free rLH subunits are detectable. GnRH increases the amount of both I-LH and B-LH material secreted into the medium, and the major component migrates at pH 8.5 and is probably the alpha beta dimer. [3H]Mannose and [35S]sulfate can be incorporated into stored and released rLH (pH 7.2, 7.8, 9.0, 9.1, and 9.3 and 7.2, 7.8, 8.5, and 9.0, respectively). GnRH decreases [2-3H]mannose incorporation into secreted rLH. [35S]Sulfate was incorporated into I-LH released spontaneously into the medium; the form at pH 7.2 has no biological activity and is probably the free alpha subunit. GnRH decreases the [35S]sulfate-labeled rLH content of the pituitary concomitantly with a 500% increase in [35S]sulfate-labeled released rLH, suggesting that, soon after [35S]sulfate is incorporated, sulfated rLH is released. Sulfatase action on released rLH reveals that sulfation may be related to release of rLH but that sulfate residues are not involved in the expression of rLH bioactivity. In conclusion, GnRH stimulates carbohydrate incorporation and processing of the oligosaccharide residues giving the highest biological potent rLH molecule and also increases sulfation; this step is closely related to the step limiting the appearance of LH in the medium in the absence of GnRH.     

Sulfation of Tyr1680 of human blood coagulation factor VIII is essential for the interaction of factor VIII with von Willebrand factor

The acidic region of the Factor VIII light chain was studied with regard to structural requirements for the formation of a functional von Willebrand factor (vWF)-binding site. Factor VIII mutants lacking the B domain, with additional deletions and an amino acid replacement within the sequence 1649-1689 were constructed using site-directed mutagenesis and expressed in Cos-1 cells. These mutants, which were recovered as single-chain molecules with similar specific activities, were compared in their binding to immobilized vWF. Deletion of amino acids 741-1648 or 741-1668 did not affect the binding of Factor VIII to vWF. However, a mutant with a deletion of residues 741-1689 was no longer capable of interacting with vWF. This indicates a role for residues within the sequence 1669-1689 in the formation of a vWF-binding site. When recombinant Factor VIII was expressed in the presence of chlorate, an inhibitor of protein sulfation, the resulting Factor VIII displayed strongly reduced binding to vWF. vWF binding was completely abolished when within the sequence 1669-1689 the tyrosine residue Tyr1680, which is part of a consensus tyrosine sulfation sequence, was replaced by phenylalanine. The Factor VIII sequence 1673-1689 was identified as a high affinity substrate for tyrosylprotein sulfotransferase (Km = 57 microM) in cell-free sulfation studies. It is concluded that sulfation of Tyr1680 is required for the interaction of Factor VIII with vWF. Two synthetic peptides that represent the sequence 1673-1689, but differ with respect to sulfation of Tyr1680 are shown to have vWF binding affinity that is considerably lower than the Factor VIII protein. Several models to accommodate our findings are discussed.     

Role of sulfation in post-translational processing of gastric mucins.

1. Gastric mucosal segments were incubated in MEM supplemented with various sulfate concentrations in the presence of [3H]glucosamine, [3H]proline and [35S]Na2SO4, with and without chlorate, an inhibitor of 3'-phosphoadenosine-5'-phosphosulfate formation. 2. Incorporation of glucosamine and sulfate depended upon the sulfate content of the medium and reached a maximum at 300 microM sulfate. Introduction of chlorate into the medium, while having no effect on protein synthesis as evidenced by proline incorporation, caused, at its optimal concentration of 2 mM, a 90% decrease in mucin sulfation and a 40% drop in glycosylation. 3. At low sulfate content in the medium and in the presence of chlorate, the incorporation of sulfate and glucosamine was mainly into the low molecular-weight form of mucin. An increase in sulfate in the medium caused an increase in the high molecular-weight form of mucin and in the extent of sulfation in its carbohydrate chain. 4. The results suggest that the sulfation process is an early event taking place at the stage of mucin subunit assembly and that sulfate availability is essential for the formation of the high molecular-weight mucin polymer.     

Tyrosine sulfation of yolk proteins 1, 2, and 3 in Drosophila melanogaster

Protein sulfation was studied in Drosophila melanogaster after in vivo labeling of flies with inorganic [35S]sulfate. After separation of total fly protein by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, proteins with sulfated carbohydrates and proteins containing tyrosine sulfate were found in all the molecular weight ranges analyzed. When female and male fly proteins were compared with each other, the electrophoretic patterns of protein-bound carbohydrate sulfate were found to be similar, whereas those of protein-bound tyrosine sulfate were distinct. The most prominent difference was the exclusive presence in female flies of three major tyrosine-sulfated proteins with apparent molecular masses between 48 and 45 kDa. Radioimmunolabeling after two-dimensional polyacrylamide gel electrophoresis was used to identify these proteins as yolk proteins 1, 2, and 3. Each of the three yolk proteins existed in several isoelectric forms, all of which were sulfated. Since the number of tyrosine residues in the yolk proteins is known, the stoichiometry of tyrosine sulfation could be determined by a novel method and was found to be 2.2, 0.9, and 1.2 mol of tyrosine sulfate per mol of yolk protein 1, 2, and 3, respectively. The present results, together with the recently reported molecular cloning of the yolk protein genes, make the yolk proteins suitable objects for genetic approaches to investigate the biological role(s) of tyrosine sulfation of secretory proteins.     

Tyramine-O-sulfate, in addition to tyrosine-O-sulfate, is produced and secreted by HepG2 human hepatoma cells, but not by 3Y1 rat embryo fibroblasts

The spent media of HepG2 human hepatoma cells and 3Y1 rat embryo fibroblasts labeled with [35S]sulfate, upon ultrafiltration, were analyzed by a two-dimensional thin-layer separation procedure. Autoradiographs of the cellulose thin-layer plate revealed the presence of tyramine-O-[35S]sulfate in addition to tyrosine-O-[35S]sulfate in spent medium from human hepatoma cells. In contrast, only tyrosine-O-[35S]sulfate was observed in spent medium of 3Y1 rat fibroblasts. Using adenosine, 3'-phosphate, 5'-phospho[35S]sulfate as the sulfate donor, sulfotransferase(s) present in HepG2 cell homogenate catalyzed the sulfation of tyramine to tyramine-O-[35S]sulfate, but not the sulfation of tyrosine to tyrosine-O-[35S]sulfate. Endogenous aromatic amino acid decarboxylase present in HepG2 homogenate was shown to catalyze the decarboxylation of [3H]tyrosine to form [3H]tyramine while attempts to use it for the decarboxylation of tyrosine-O-sulfate to form tyramine-O-sulfate were unsuccessful. These results suggest that tyramine-O-sulfate may be derived from the de novo sulfation of tyramine, instead of the decarboxylation of tyrosine-O-sulfate.     

Heterogeneity in the tyrosine sulfation of Chinese hamster ovary cell produced recombinant FVIII

By the use of recombinant technology, several stable Chinese hamster ovary (CHO) cell lines expressing human FVIII were established. Thrombin treatment and SDS-PAGE analysis of the purified recombinant FVIII (rFVIII) revealed a striking difference from plasma-derived FVIII (pFVIII). A 43-kDa fragment of the FVIII heavy chain appears as a double band from rFVIII, while a single band from pFVIII is observed. All other fragments from the two samples appeared similar by SDS-PAGE. The heterogeneity is caused by incomplete tyrosine sulfation of one or more of the three potential tyrosine sulfation sites (Tyr718, Tyr719, Tyr723). To investigate if there is a general limitation and heterogeneity in the tyrosine sulfation of rFVIII, two other potential tyrosine sulfation sites on the FVIII light chain (Tyr1664, Tyr1680) were analyzed. The results show that both sites on the pFVIII light chain and on the rFVIII light chain are completely sulfated. The limitation of CHO cells to tyrosine sulfate rFVIII is therefore only restricted to a few sites. The two sulfated forms of rFVIII can easily be separated by ion-exchange chromatography, indicating the importance of the sulfate groups on the charge and/or conformation of FVIII. Both forms of rFVIII possess identical in vitro coagulation activity, von Willebrand factor binding, and thrombin activation profile. However, the difference in tyrosine sulfation may change other biological properties of FVIII.     

Formation of dermatan sulfate by cultured human skin fibroblasts. Effects of sulfate concentration on proportions of dermatan/chondroitin

[3H,35S]Dermatan/chondroitin sulfate glycosaminoglycans produced during culture of fibroblasts in medium containing varying concentrations of sulfate were tested for their susceptibility to chondroitin ABC lyase and chondroitin AC lyase. Chondroitin ABC lyase completely degraded [3H]hexosamine-labeled and [35S] sulfate-labeled dermatan/chondroitin sulfate to disaccharides. Chondroitin AC lyase treatment of the labeled glycosaminoglycans produced different results. With this enzyme, dermatan/chondroitin sulfate formed at high concentrations of sulfate yielded small glycosaminoglycans and larger oligosaccharides but almost no disaccharide. This indicated that the dermatan/chondroitin sulfate co-polymer contained mostly iduronic acid with only an occasional glucuronic acid. As the medium sulfate concentration was progressively lowered, there was a concomitant increase in the susceptibility to degradation by chondroitin AC lyase. Thus, the labeled glycosaminoglycans formed at the lowest concentration of sulfate yielded small oligosaccharides including substantial amounts of disaccharide. The smaller chondroitin AC lyase-resistant [3H,35S]dermatan/chondroitin sulfate oligosaccharides were analyzed by gel filtration. Results indicated that, in general, the iduronic acid-containing disaccharide residues present in the undersulfated [3H,35S]glycosaminoglycan were sulfated, whereas the glucuronic acid-containing disaccharide residues were non-sulfated. This work confirms earlier reports that there is a relationship between epimerization and sulfation. Moreover, it demonstrates that medium sulfate concentration is critical in determining the proportions of dermatan to chondroitin (iduronic/glucuronic acid) produced by cultured cells.