The Occurrence of a Wide Variety of Dermatan Sulfate-Chondroitin Sulfate Copolymers in Fibrous Cartilage*

Abstract A large fraction with the properties of a dermatan sulfate-chondroitin sulfate copolymer was obtained from the meniscus (semilunar cartilage) of the human knee joint. This fraction appeared to be homogeneous with respect to charge density, as judged by its behavior on DEAE-cellulose chromatography, but it was further resolved into a number of subfractions by ethanol fractionation, by electrophoresis on cellulose acetate, and by chromatography on Sephadex G-200. Analyses of these subfractions showed that the copolymer was composed of a series of variants whose properties could be accounted for by a gradual variation in the proportion of l-iduronosyl-N-acetylgalactosamine 4-sulfate units and d-glucuronosyl-N-acetylgalactosamine 6-sulfate units. These results indicate that meniscus mucopolysaccharides represent a separate family of mucopolysaccharides that may be specifically adapted to the function of fibrous (collagen-rich) cartilages, and further suggest that any attempt to obtain a mucopolysaccharide sample for structural study must take into consideration this type of polymorphism as well as the scarcity of reliable methods for purification of such mixtures.

HIROKO HABUCHT, TATSUYA YAMAGATA, HISASHI IWATA,$ AND SAKARU SUZUKI From the Department of Chemistry, Faculty of Science, Nagoya University, Nagoya 464, Japan SUMMARY A large fraction with the properties of a dermatan sulfatechondroitin sulfate copolymer was obtained from the meniscus (semilunar cartilage) of the human knee joint. This fraction appeared to be homogeneous with respect to charge density, as judged by its behavior on DEAE-cellulose chromatography, but it was further resolved into a number of subfractions by ethanol fractionation, by electrophoresis on cellulose acetate, and Ly chromatography on Sephadex G-200.
Analyses of these subfractions showed that the copolymer was composed of a series of variants whose properties could be accounted for by a gradual variation in the proportion of L-iduronosyl-N-acetylgalactosamine 4sulfate units and D-glucuronosyl-N-acetylgalactosamine 6-stlfate units.
These results indicate that meniscus mucopolysaccharides represent a separate family of mucopolysaccharides that may be specifically adapted to the function of fibrous (collagenrich) cartilages, and further suggest that any attempt to obtain a mucopolysaccharide sample for structural study must take into consideration this type of polymorphism as well as the scarcity of reliable methods for purification of such mixtures.
Among the sulfated mucopolysaccharides of connective tissue, a family of polysaccharides containing cr-1,3-Cduronic acid and /3-l ,4-N-acetyl-n-galactosamine 4-sulfate as the main repeating units is designated as chondroitin sulfate I3 (1) or dermatan sulfate (2). It has further been shown that multiple minor variations exist in the composition of dermatan sulfate chains of proteoglycans isolated from different tissue sources. Some examples of such variations are the presence of additional sulfate groups on uranic acid residues (3) or on hexosamine residues (4), and the presence of n-glucuronic acid besides L-iduronic acid (5-12). We do not know, however, how many of the variations are random, as, for example, caused by variable epimerization and sulfa-* This research was supported in part by a research grant from the Ministry of Education of Japan and by a grant from the Takeda Science Foundation.
$ Present address, tion and which, if any, are oriented in an orderly fashion. Fransson et al. (12) have pointed out that some regularities and tissue specificities in the distribution of n-glucuronic acid can be shown when dermatan sulfates from various sources are subjected to degradation with testicular hyaluronidase. Based on the chromatographic profiles of both intact samples and their hyaluronidase digests, the authors have postulated four models showing some principal differences in the length of the n-glucuronic acidcontaining sections and in the distribution of these sections along the chains. However, most procedures for the purification of dermatan sulfate consist of digesting the defatted tissues with proteolytic enzymes, followed by fractionation of the mucopolysaccharide material with cetylpyridinium chloride, or with ethanol, or both, and it is questionable whether any of the procedures can result in the production of a homogeneous preparation that has escaped closely related substances.
In fact, the fractionation profiles of dermatan sulfates from pig skin, human umbilical cord, and several other tissues are too disperse, and even contradictory, to permit any definite statements regarding regularities in uranic acid distribution to be made. It would thus appear that the dermatan sulfate fractions obtained from these tissues are still far from homogeneous and the development of more accurate and efficient methods of fractionation and analysis is necessary to further the study of dermatan sulfate structure.
Previous work in our laboratory (13,14) showed a tissuespecific synthesis of dermatan sulfate in the meniscus' of the human knee joint, which was manifested by the presence of a large proportion of GlcUA-GalNAc(6S)2 units. The present investigation was therefore designed to obtain sufficient data to Chemical A&y+-Analyses for total hexuronic acid were carried out by the carbazole (23) and orcinol (24) methods; the latter method was used with a 30-min boiling time. Differential determinations of n-glucuronic and L-iduronic acid in mucopolysaccharide samples were performed by the method of Saito et al. (25). This method is based on the findings that both L-iduronic and n-glucuronic acid residues were converted to A4-glucuronic acid residues (in disaccharide forms) under the influence of chondroitinase-ABC, while only n-glucuronic acid residues were converted to the unsaturated residues (in disaccharide and oligosaccharide forms) with chondroitinase-AC. The amounts of 3 Commercial preparations of chondroitinase-AC, chondroitinase-ABC, AGlcUA-GalNAc, AGlcUA-GalNAc (4S), and AGlcUA-GalNAc(6S) were also used in some of the experiments in this paper. More recently, an enzyme of the chondroitinase-AC type has been purified from the culture fluid of Arthrobacter aureus (19). A comparison was made of the specificities and activities of the Arthrobacter enzyme (kindly given by the above authors) and the Plavobacterium enzyme. Essentially identical results were obtained with both preparations.
The principal advantage of the Arthrobacter enzyme is a simplicity of the purification procedure. The Arthrobacter enzyme has become available from Seikagaku Kogyo CO., 2-9, Nihonbashi-Honcho, Tokyo 103, Japan (Dr. M. Nomoto, private communication).
A4-glucuronic acid residues can be estimated from ultraviolet absorption measurements at 232 nm with an average molar absorption coefficient of 5500. The increase in absorpt'ion under the influence of chondroitinase-AC and of chondroitinase-ABC is a measure of n-glucuronic acid and of the sum of u-glucuronic acid and L-iduronic acid, respectively.
Conditions for the digestion of mucopolysaccharides and methods for the assay of reactions were described previously (25).
The hexosamine content of preparations was determined, after hydrolysis of samples with 6 x HCl for 8 hours in a boiling water bath, by the Elson-h#lorgan method as modified by Strominger et al. (26) and corrected for losses by comparison with a standard sample run concurrently.
Sulfate was estimated according to the method of Kawai et al. (27). Amino acids were determined with the use of a Hitachi automatic amino acid analyzer after hydrolysis with 6 N WC1 at 110" for 20 hours. The authors would like to express their appreciation to Dr. N. Takahashi and Dr. T. Murachi, Nagoya City University, for assistance with this determination.
For determination of galactose and xylose, each mucopolysaccharide sample (1 pmole as hexosamine) was hydrolyzed for 3 hours at 100" in a vacuum-sealed tube with 1 N I-ICI. The hydrolysate was taken to dryness in a desiccator over NaOH, dissolved in a little water, and passed through columns (0.7 X 3 cm) of Dowex 50 (II+) and Dowex 2 (COs=). The effluent was concentrated to a small volume and subjected to descending paper chromatography with Solvent C (see below) for 20 hours. Authentic samples of n-xylose and n-galactose were run parallel with the test sample as external standards.
The zones corresponding to xylose (Rgrucose = 1.59) and galactose (RFrucose = 0.88) were located by staining the external standards with AgN03 (28), cut out, and eluted with water.
The content of reducing sugar in each of the eluates was determined by the method of Park and Johnson (29). In order to determine the losses of the neutral sugars that occurred during hydrolysis and chromatography, known amounts of n-xylose and n-galactose were added to one set of the mucopolysaccharide sample before hydrolysis.
The final recovery of n-galactose (98%) and that of n-xylose (80%) then provided internal standards with which the sugars liberated from the polysaccharide sample were compared.
Optical rotation was determined with a Jasco photoelectric spectropolarimeter, model ORD/UV-5. Electrophoresis and Chromatography-Electrophoresis of mucopolysaccharides was carried out on 6-cm long strips of Separax (cellulose acetate film) in a glass apparatus (JokG Sangyo CO., Tokyo) at a constant current of about 1 ma per cm for 30 min. The buffer used was pyridine-acetic acid-water (1:9: 115, by volume), pH 3.5. The strips were stained, according to the method of Seno et al. (30), with 0.5% Alcian Blue in 3y' acetic acid for about 20 min. After washing with water for 10 min, the strips were blotted dry, let stand at room temperature for several minutes, and then pressed between sheets of filter paper.
Electrophoresis of unsaturated oligosaccharides was carried out on 60.cm long strips of Toyo No. 51A filter paper in the apparatus described by Markham and Smith (31) at a potential gradient of 30 volts per cm for 45 min. The buffer used was 0.05 in ammonium acetate-acetic acid, pH 5.0. The oligosaccharides containing A4-glucuronic acid were detected by viewing under ultraviolet light or by staining with AgN03 (28).
Preparation of Mucopolysaccharide Mixture from Human Meniscus-Human menisci were collected from normal subjects aged 60 to 75 years in the autopsy room of the Medical School of this university (we are indebted to Dr. J. Kito and Dr. Y. Terashima for their assistance).
In each case, about 290 g of menisci were finely chopped with scissors, treated with several changes of acetone, and air-dried.
The dried material, 72 g, was suspended in 1 liter of 0.05 M Tris-HCI, pH 8.0, and heated at 70" for 1 hour.
After cooling, the mixture was covered with a layer of toluene and incubated at 37" for 3 days in the presence of 180 mg of Pronase.
During that time, the pH of the mixture was readjusted to pH 8 with 1 N NaOH and additional Pronase (180 mg) was added twice.
Insoluble material was removed by centrifugation at 12,000 x g for 30 min; the supernatant was dialyzed against running tap water for 2 days. The dialyzed solution was reduced to about 0.2 volume on a rotating evaporator. Potassium acetate was added to l%, and the mucopolysaccharide material was precipitated with 2 volumes of ethanol. After standing overnight in the refrigerator, the precipitate was centrifuged and dissolved in 80 ml of 0.05 M Tris-HCl, pH 8.0. The solution was treated with 90 mg each of Pronase as described above (three cycles) and dialyzed against running tap water for M, a mixture of 0.8 nmole each of 1 (a), 2 (b), 5 (c), and heparin (d). 2 days. Trichloroacetic acid was added at 4" to a concentration of 5%. The precipitate was removed by centrifugation at 12,000 x g for 10 min. The supernatant solution was dialyzed against distilled water at 4". Potassium acetate was added to the dialyzed solution to a concentration of 1%. The mucopolysaccharide material was precipitated with 2 volumes of ethanol, centrifuged, washed with ethanol, ether, and dried over PZ05 in a vacuum.
This preparation will be referred to as "crude mixture."

Survey of Mucopolysaccharides
in Crude Mixture-It has been shown previously4 that a separation of small amounts of mucopolysaccharides is possible by electrophoresis on cellulose acetate strips with appropriate buffer systems. Indeed, by the use of pyridine-acetic acid buffer, pH 3.5, it was possible to separate hyaluronic acid (human umbilical cord), dermatan sulfate (hog intestinal mucosa), chondroitin ii-sulfate (whale cartilage), or chondroitin B-sulfate (shark cartilage), and heparin from one another ( Fig. 1). Apparently, the profiles reflected the extent of sulfate substitution and the uranic acid composition, but not the difference in the location of sulfate.
In order to evaluate the contribution of molecular weight, partially depolymerized dermatan sulfate (average molecular weight = 16,000) and chondroitin 6-sulfate (average molecular weight = 15,000) were also examined by the same technique.
As shown in Fig. 1, the samples showed no tendency to separate according to molecular weight.
This technique was used to survey the mucopolysaccharides from human meniscus (Fig. 2 The other component appeared on the electrophoretogram as a broad band extending from the dermatan sulfate zone to the chondroitin sulfate zone (Fig. 2). The profile was not changed by treating the sample with 1.0 N NaOH at 4" for 15 hours. This component proved resistant to Xtreptomyces hyaluronidase, but susceptible to degradation with chondroitinase-ABC." Digestion with chondroitinase-AC, on the other hand, caused a partial reduction in the metachromatic activity of this component; those portions near the chondroitin sulfate area began to exhibit exceedingly feeble Alcian Blue reactions whereas the portions near the dermatan sulfate area were less affected by the enzyme digestion.
It was assumed from these properties that the fastmoving component of the crude mixture must contain, in addition to typical dermatan sulfate and chondroitin 4-or 6-sulfate, dermatan sulfate-chondroitin sulfate copolymers with a wide range of chemical heterogeneity.
It has been established previously (18) that chondroitinase-ABC cleaves dermatan sulfate and chondroitin 4-and g-sulfate at their fl-hexosaminidic bonds to either L-iduronic or D-glucuronic acid residues to produce unsaturated disaccharides, while chondroitinase-AC cleaves only the bonds to D-glucuronic acid residues. Degradation of a dermatan sulfate-chondroitin sulfate copolymer with chondroitinase-AC should yield oligosaccharides with one or more L-iduronic acid residues and one nonreducing terminal A4-glucuronic acid. Previous studies on the mucopolysaccharides from HeLa-S and L-929 cells (35) as well as those from horse aorta (11) have shown that this was indeed the case.
This method was applied to the analysis of the crude mucopolysaccharide sample from meniscus. Aliquots of the test sample were separately subjected to digestion with chondroitinase-ABC and chondroitinase-AC. The digests were chromatographed on paper in Solvent A (Fig. 3). The chromatogram of the chondroitinase-Al3C system showed the formation of four main spots; comparison of their mobilities with standard disaccharides permitted the conclusion that they, in the order of their Rp values on the chromatogram, were AGlcUA-GlcNAc, AGlc-UA-GalNAc, AGlcUA-GalNAc(4S), and AGlcUA-GalNAc(6S). gest that the mixture contains a high proportion of N-acetylgalactosamine g-sulfate residues linked to L-iduronic acid. Also notable in the chromatogram of the chondroitinase-AC system was the appearance of five slow spots (referred to as Oligo-I, Oligo-II, Oligo-III, Oligo-IV, and Oligo-V in the order of their RF values).
The chromatographic mobilities of Oligo-I and Oligo-II were similar to those of the known disulfated disaccharides bearing an extra sulfate on position 6 of the hexosamine moiety (4) or on the uranic acid moiety (3), respectively, but chemical analyses (see below) indicated that none of these products from the meniscus contains such extra sulfates.
Therefore, this pattern of chondroitinase products seems reasonable only if it, is assumed that such slower moving spots as those produced with chondroitinase-AC, but not with chondroitinase-ABC, represent oligosaccharides which are derived from hybrids containing both D-glucuronic and L-iduronic acid residues in the same polysaccharide chain. This interpretat.ion is in agreement with the prediction made from the electrophoretic analysis (see above) and further suggests that studies of these oligosaccharides will permit a formulation of the general hybrid properties of the parent polysaccharidcs.
Characterization of Oligosaccharides Obtained ajter Chondro-i&use-AC Degradation-Each oligosaccharide was prepared on a large scale as outlined in Table I. The results of chemical analyses shown in Table I   assumption t.hat each has 1 A4-glucuronic acid residue, which can be quantitatively determined by ultraviolet absorption, the degree of polymerization of each compound u-as calculated as indicated in Table I. Since the separation of oligosaccharides higher than decasaccharides has been difficult, the proposed structure of Oligo-V may represent only a statistical model of the higher oligosaccharides.
The presence of L-iduronic acid as well as the absence of Dglucuronic acid (saturated form) were confirmed with all the samples by mild acid hydrolysis followed by paper chromatography in Solvent E (34) and by chromatography on a column of AG LX8 (32).
The oligosaccharides were subjected to chondroitinase-ABC degradation and the resulting disaccharides were determined as described in the previous paper (25). As shown in Table I, each of the oligosaccharides gave AGlcUA-GalNAc(4S) and either AGlcUA-GalNAc or AGlcUA-GalNAc(6S) in yields which were consistent with those calculated from the proposed structures.
In order to obtain information regarding the sequence of the disaccharide units, each oligosaccharide was subjected to digestion with Flavobacterium glucuronidase, an enzyme which catalyzes a hydrolytic release of the Ad-glucuronic acid adjacent to N-acetylgalactosamine or N-acetylgalactosamine B-sulfate, but not the A4-glucuronic acid adjacent to N-acetylgalactosamine 4sulfate (3,18). The resulting mixture was then cleaved with chondroitinase-ABC and chromatographed as described above. The chromatogram revealed two reducing sugars corresponding to AGlcUA-GalNAc(4S) (from all the oligosaccharides) and either N-acetylgalactosamine (from Oligo-I) or N-acetylgalactosamine 6-sulfate (from Oligo-II, III, IV, and V) (for RF values of these sugars, see Reference 3). It should be pointed out that no spots corresponding to either AGlcUA-GalNAc or AGlcUA-GalNAc-(6s) could be observed.
Since such disaccharides as those pro- duced from the intact oligosaccharides, but not from the glucuronidase-treated oligosaccsharidrs should originate from the nonreducing portion, it was deduced that the intact oligosaccharides contained disaccharide units in the sequences shown in Table I.6 Preparation of IIyaluronic k&tree Mixture-As described above, electrophoresis of the crude mucopolysaccharide mixture showed two major components, with the slow component corresponding to standard hyaluronic acid. To purify the copolymer fraction for further characterization, the crude mixture was fractionated with DEAE-cellulose, and this enabled the mixture to be separated into two major fractions (Fig. 4).
The 0.4 M NaCl fraction was indistinguishable from standard hyaluronic acid in its chemical composition, electrophoretic mobility (Fig. 5), and behaviors toward chondroitinase-AC and Sfreptomyces hyaluronidase.
Therefore, no further investigation of this fraction has been carried out,.
The 0.7 M NaCl fraction, by electrophoresis, was shown to correspond to the fast component in the crude mixture (Fig. 5). Those compounds yielding oligosaccharides by chondroitinase-AC digestion were exclusively recovered in this fraction.  The 0.7 M NaCl fraction was dialyzed against distilled water, precipitated with 2 volumes of ethanol in the presence of 1% potassium acetate, centrifuged, washed with ethanol, ether, and dried over PzOh in a vacuum. The yield was 0.8 g from 1.35 g of the crude mixture. This material was redissolved in 100 ml of 0.02 M Tris-I-ICI, pH 7.2, and applied to a DEAE-cellulose column (2.8 X 41 cm) equilibrated with the same buffer. After washes with the buffer, the column was developed by linear gradient elution with 1.5 liters of 0.02 M Tris-HCl, pH 7.2, in the mixing flask and 1.5 liters of 1.0 M NaCl in the same buffer in the reservoir. A single peak of uranic acid-reacting material was observed (see the inset of Fig. 4). The fractions containing the major portion of this peak (tubes 89 to 142) were pooled, washed, and dried as described above. The yield was 0.67 g (0.92 mmole as carbazole-positive glucuronic acid). This preparation will be referred to as "hyaluronic acid-free mixture." The profile in Fig. 4 (see the inset) showed that this fraction is fairly homogeneous with respect to charge density. Nevertheless, the same fraction displayed a considerable heterogeneity in the electrophoretic profile (the profile was similar to that of Sample d in Fig. 5 and is not shown here), suggesting that heterogeneity of uranic acid composition in the copolymers may be reflected in their electrophoretic properties.
Alcohol Fractionation of Copolymers-It has been shown previously (37) that dermatan sulfate and chondroitin 6-sulfate (calcium salts) need different concentrations of ethanol to be precipitated. Both the sulfate location as well as the uranic acid composition might be expected to play a role in determining the solubility characteristics. Indeed, it was possible to separate standard dermatan sulfate (hog mucosa) from standard chondroitin 6-sulfate (shark cartilage) by the ethanol fractionation method; the former was quantitatively precipitated at ethanol concentrations as low as 2OQ/ the bulk of the latter between 30 and 40%.
This approach was used to study the properties of copolymers in the hyaluronic acid-free mixture. The results of the ethanol fractionation are listed in Table II. It can be seen that the sample of human meniscus is not simply a mixture of dermatan sulfate and chondroitin B-sulfate, i.e. the sample gave a large fraction between 20 and 30% ethanol concentration, where neither standard dermatan sulfate nor standard chondroitin 6sulfate are precipitated.
The 20% ethanol precipitate which from now on will be called "Fraction 20" produced on electrophoresis a band with the same mobility as standard dermatan sulfate, and the 30 to 4Ooje precipitate ('LFraction 40") a somewhat extended band close to the chondroitin g-sulfate zone (Fig. 6). However, the 20 to 30y0 precipitate ("Fraction 30") differed markedly from these standard polysaccharides, i.e. it occupied an intermediary position in the electrophoretic profile.
Fractions 20, 30, and 40 had different rotations and carbazole to orcinol ratios (Table II). As far as these properties are concerned, Fractions 20 and 40 are similar to standard dermatan sulfate and chondroitin 6-sulfate, respectively, and Fraction 30 appears to be a copolymer with intermediate properties.
Taking advantage of the selective action of chondroitinase-ABC and chondroitinase-AC on galactosaminide linkages to Liduronic and n-glucuronic acid residues, it is possible to determine the molar ratio of n-glucuronic acid to n-iduronic acid in a given mucopolysaccharide sample (25). The chondroitinase method was applied to the analysis of the three ethanol fractions. The results, shown in Table II, showed that Fraction 30 contains 62% of total uranic acid as n-iduronic acid and 38oj, as n-glucuronic acid, whereas Fractions 20 and 40 are much richer in n-iduronic acid and in D-&CUrOni C acid, respectively. The yields of unsaturated di-and tetrasaccharides in the three ethanol fractions were compared after chromatographic separation (Table II). Most notable in this comparison is the high yield of the tetrasaccharide from Fraction 30, i.e. about 18% of the total uranic acid residues (corresponding to about 2401, of the total n-glucuronic acid residues) was recovered as AGlcUA-GalNAc(6S)-MUA-GalNAc(4S) after chondroitinase-AC treatment. This indicates that in Fraction 30 about one-half of the total n-glucuronic acid residues is located in a sequence of -GlcUA-GalNAc(6S)-IdUA-GalNAc(4S)-GlcUA-.     ular size distribution of the fragments.
The difference may reflect the fact that the iduronic acid-containing sections of the Fraction 20 polysaccharides are larger in average size than those of the Fraction 30 polysaccharides.
Heterogeneity of Copolymers-The heterogeneity of meniscus copolymers was further illustrated by the behavior of the three ethanol fractions on fractionation with Sephadex G-200 (Fig. 8). For each experiment, human -y-globulin and bovine serum albumin were run as internal markers.
The materials in each ethanol fraction displayed a considerable polydispersity.
Furthermore, the elution profiles of the three fractions were significantly different; the peak of Fraction 30 was located at an earlier position than that of Fraction 20, while the peak of Fraction 40 was located at a later position.
In cvcry case, the behavior on gel chromatography was not affected by treatment of the material with 1 N KaOH at 4" for 15 hours, indicating that neither mucopolysaccharides associated with large peptides noi multichain mucopeptides were present in the samples.
The question was raised lyhether these profiles reflect a chain length polydispcrsity or a diversity of molecular hybrids. In order to evaluate the contribution of these factors, each peak IT-as subdivided into four fractions as iudicated in Fig. 8; these IT-ere dialyzed against distilled Ivater to rcmovc salt, and subjected to analyses for the determination of chain length and uranic acid composition.
Qualitative analysis of neutral sugars by paper chromatography in Solvent C showed the presence of galactose and sylose in all of the subfractions.
The quantitative analysis of Subfractions 20-IV and 30-I (Table III) showed that the molar ratio of serine to xylose to galactose was approximately 1: 1:2. Treatment of the preparations &h 1 s SaOH at 4" for 15 hours resulted in 45010 destruction of the serine residues.
Ko significant destruction of other amino acids \vas observed. The failure with these mucopcptides to attain complete destruction of the serine residues suggests that approximately one-half of the serine residues occupied COOH-or NHp-terminal positions in the peptides, since the work of Stern et al. (38) has indicated that cleavage of the xylosidic linkage does not occur if the amino or carbosyl groups of scrine are free.
Estimates based on the molar ratios of total uranic acid to total xylose gave values close to 91 (the number of dissacharide units per molecule) for Subfraction 20-IV and 95 for Subfraction 30-I. Since the difference in degree of polymerization between the earliest component and the latest component is within the limits of precision of the methods employed, it seems fair to conclude that the extent of hybridization, and not chain length, is the principal parameter involved in these elution profiles.
The analytical data for n-glucuronic acid composition (Fig. 9) indicate that the elution profiles of Fractions 20 and 30 reflect the n-glucuronic to L-iduronic acid ratio; i.e. the higher the Dglucuronic acid content, the faster the elution rate. For Fraction 40, however, the reverse is true. It seems likely, therefore, that the elution rate in gel chromatography may mirror the degree of hybridization in the sense that the clution rate becomes faster as the n-glucuronic to L-iduronic acid ratio become closer to unity.
The arguments presented so far indicate that the material in (Subfraction 40-IV) (Fig. 9). In addition, all the subfractions on gel chromatography differed with respect to their contents of the GlcUh-GalNAc(6S)-IdUA-GalNAc(4S)-GlcUA unit (estimated from the yields of Oligo-II after chondroitinase-AC digestion) : the values (per cent of total uranic acid) being increased progressively from about 3 70 for Subfractions 20-IV and 77, for 40-IV to about 407, for Subfraction 30-I.
The multiplicity of hybrid molecules was further illustrated by the electrophoretic profiles of these subfractions (Fig. 10). From 20.IV to 30-1, the subfractions could be arranged serially; the higher the n-glucuronic acid content, the greater the mobility. Subfractions 40-I to 40-IV, on the other hand, did not show such a separation on the electrophoretogram. This is to be expected since their ratios of glucuronic to iduronic acid did not seem to differ to any large extent as compared to those of Subfractions 20-IV t,o 30-I. Mucopolysacclzaride Patterns of Xenisci from Subjects Aged 20 and LS Years-Changes in the proportion of 4-and B-sulfate units in cartilage chondroitin sulfates have been noted during growth of higher vertebrates (39) as well as in many pathological conditions.
The question is then raised as to whet.her the structures of human meniscus copolymers may change with age or with some other physiological factors. We, therefore, examined the materials obtained from two young patients (20.year-old female and 23-year-old female dying without any known skeletal dis-by guest on March 24, 2020 http://www.jbc.org/ Downloaded from ease). Both samples were shown to have electrophoretic profiles indistinguishable from the profile of Sample 1 shown in Fig. 2. Furthermore, no significant differences in the paper chromatographic profiles of chondroitinase digests were found between the specimens from subjects aged 20 and 23 years and the specimens from subjects aged 60 to 75 years.

DISCUSSION
It is clear from the results presented in this paper that the human meniscus contains a great variety of dermatan sulfatechondroitin sulfate copolymers. Thus, electrophoresis in pyridine-acetic acid as well as chromatography on Sephadex G-200 were shown to separate the copolymer into a series of variants whose mobilities could be accounted for by a gradual variation in the proportion of GlcUA-GalNAc (6X) and IdUA-GalNAc (4s) units. Sephadex G-200 chromatography provided the resolution required to separate copolymers differing in extent of hybridization. There was a general trend toward a decrease in the extent of hybridization as the elution volume increased. A simple theory to account for this behavior is to assume that a copolymer chain would tend to take up a more complex conformation as compared with the same length of homopolymer chains. At any rate, these findings leave little doubt that the meniscus contains various hybrid molecules with different proportion and distribution of n-glucuronic acid-and L-iduronic acid-containing sections. It should be stressed, therefore, that any attempt to obtain a dermatan sulfate sample for structural study must take into consideration this type of polymorphism as well as the scarcity of reliable methods for purification of such mixtures.
When examined from the point of view how cells synthesize mucopolysaccharide chains, the occurrence of a series of dermatan sulfate-chondroitin sulfate copolymers with varying degrees of hybridization appears to be a direct consequence of how Liduronic acid residues are introduced into the molecule. Two hypotheses have been advanced to explain the occurrence of L-iduronic acid in mucopolysaccharide chains. One hypothesis is that UDP-n-glucuronic acid is converted to UDP-n-iduronic acid which might serve as the source of the L-iduronic acid in mucopolysaccharide (40). The second hypothesis is that initially a polymeric intermediate containing n-glucuronic acid is synthesized and that subsequently some of the n-glucuronic acid is converted enzymatically to L-iduronic acid (41). The impressive microheterogeneity revealed by analysis of the uranic acid compositions of dermatan sulfates from various sources may argue in favor of the second hypothesis, that the formation of L-iduronic acid is governed primarily by the availability of an enzyme that catalyzes the 5-epimerization of o-glu- curonic acid in the polymer intermediate. If the conditions in a cell do not provide enough active enzyme or time to complete the epimerization, most of the molecules synthesized by the cell should contain regions that are only partially epimerized and highly heterogeneous. However, no data are presently available concerning the occurrence of such an epimerase or the mechanism of such type of epimerization.
Previous studies on mucopolysaccharides from various connective tissues indicate that different tissues have met the challenges of selection and adaptation through the elaboration of multiple genes for mucopolysaccharide synthesis. For example, in most of the mammalian skins thus far examined, dermatan sulfate is a major constituent, whereas it is absent in the hyaline cartilages of the same organism (37). The latter tissues contain in most cases chondroitin 4-, or, 6-sulfate, or both, as major constituents. In the meniscus, however, the major mucopolysaccharide fraction has been shown by the present study to contain both dermatan sulfate and chondroitin B-sulfate elements in the form of a copolymer. One may ask then what advantage accrued to this type of cartilage by the introduction into chondroitin g-sulfate of dermatan sulfate elements. It may be suspected from the known distribution of dermatan sulfate that dermatan sulfate is associated with coarse collagen bundles (5). Thus, where collagen consists well ordered coarse bundles, dermatan sulfate is present as in the bottom layer of skin. In contrast, in hyaline cartilages, where the collagen bundles are fine and show no 640 A periodicity (42), dermatan sulfate is absent (37). It is to be expected from these observations that the introduction of dermatan sulfate elements into chondroitin 6-sulfate enables the fibrous cartilage to form collagen bundles which are much more like the skin type than those of hyaline cartilages. An additional finding worthy of comment is the observation of Miller (43,44) that the collagen of hyaline cartilage differs from that found in skin both in amino acid sequence of the a chains and in their distribution within the triple helical collagen molecule. Although no data are available regarding the chemistry of meniscus collagen, the finding of the unique copolymers in this tissue poses the interesting possibility that the fibrous cartilage may contain a unique type of collagen polypeptide which has not been detected in hyaline cartilage.