/. Embryol. exp. Morph. Vol. 70, pp. 61-14, 1982 61 Printed in Great Britain © Company of Biologists Limited 1982 Analysis of the intermediate size proteoglycans from the developing chick limb buds1 By NAGASWAM1SRI VASAN 2 From the Department of Anatomy\ New Jersey Medical School SUMMARY Limb-bud proteoglycans are heterogeneous molecules which vary in their chemical and physical properties with development. This report describes proteoglycan intermediates (PG-I) that predominate in stage-34 limbs, and compares them with proteoglycan aggregates (PG-A) in stage-38 limbs. We analysed proteoglycans and their components extracted with guanidinium chloride by subjecting them to density gradient centrifugation, molecular sieve chromatography, electrophoretic separation, and selective enzymatic degradation. PG-I and PG-A have similar chondroitin sulphate composition, amino sugars, chondroitin sulphate side-chain length, glycoprotein link factors, and hyaluronic acid binding capacity, and both cross react with antisera prepared against cartilage-specific chick sternal proteoglycans. However, PG-I has lower molecular weight, lower buoyant density, and fewer chondroitin sulphate side chains on the protein core. The PG-I in the developing limb can be considered a mixture of smaller aggregates and cartilage-specific large monomers in which the former predominate. INTRODUCTION Cartilage proteoglycans are complex heteropolysaccharides in which a large number of chondroitin sulphate and keratan sulphate chains are convalently linked to a core protein. This core protein has a separate hyaluronic acidbinding region, and a keratan sulphate-rich and chondroitin sulphate-rich region (Heinegard & Hascall, 1974; Hascall & Heinegard, 1974; Heinegard & Axelsson, 1977). Molecular sieve chromatography of proteoglycan from different stages of developing chick limb buds and sterna showed three distinct populations, proteoglycan aggregates (PG-A), intermediates (PG-I), and monomers (PG-M) (Lash, Ovadia & Vasan, 1978; Ovadia, Parker & Lash, 1980). Proteoglycans in the extracellular matrix of normal cartilage are present mainly in the form of aggregates (Hascall & Heinegard, 1974), while monomers 1 Presented in part at 72nd Annual Meeting of the American Society of Biological Chemists, St Louis, Missouri, U.S.A. 31 May to 4 June, 1981. 2 Author's address: Department of Anatomy, Cell Biology Division, College of Medicine and Dentistry of New Jersey, New Jersey Medical School, 100 Bergen Street, Newark, New Jersey 07103, U.S.A. „ E M B 7O 62 N. VASAN and non-aggregates form smaller percentages of the total (Hardingham & Muir, 1974). In developing chick embryonic limb buds (Vasan & Lash, 1977, 1979; Lash et al. 1978; Ovadia et al. 1980) and in chick limb buds chondrocytes grown in vitro (Goetinck, Pennypacker & Royal, 1974; Hascall, Oegema, Brown & Caplan, 1976; DeLuca et al. 1977; Royal & Goetinck, 1977), the proteoglycan size and chemical composition varied with age. A dissociative solvent like 4-0 M guanidinium chloride, containing proteolytic inhibitors, is widely used to extract proteoglycan from cartilage. Physical and chemical characterization is done by molecular sieve chromatography, density gradient centrifugation, and selective enzymatic and chemical degradation followed by separation and analysis of the products. This report describes the PG-I of stage-34 (day 8) chick tibia and compares them to the PG-A of stage-38 (day 12) chick tibia in which they form respectively 60-70% and 70-85 % of the total proteoglycan. The small molecular proteoglycan eluted in the total volume of the column (referred to earlier as PG-M) has been recognized and redesignated as PG-M ubiquitous. We can, consequently, differentiate the cartilage-specific PG-M and ubiquitous PG-M from other tissues. There is a semantic confusion regarding the intermediate-size molecule (PG-I). Some describe it as a new class of proteoglycan found in the developing limb buds (Lash et al. 1978; Ovadia et al. 1980), while others consider it a mixture of cartilage-specific large monomers (Reddi, personal communication; Vasan, 1981). The results of the present study show that the intermediate size proteoglycans are a mixture of smaller aggregates and large cartilage-specific monomers. MATERIALS AND METHODS Ultrapure guanidinium chloride and CsCl were purchased from Schwarz/ Mann Biochemicals, Orangeburg, NY, USA; 6-amino-hexanoic acid and benzamidine hydrochloride from Eastman Kodak Co., Rochester, NY, USA; controlled pore glass beads (CPG-10-2500) from Electronucleonic, Fairfield, NJ, USA; Sephadex G-200 from Pharmacia Fine Chemicals, Piscataway, NJ, USA; and tissue culture feeding supplies from Grand Island Biological Company, Grand Island, NY, USA. High molecular weight rooster comb hyaluronic acid was a gift from Dr Endre Balazs, Columbia University, New York, USA. Chondroitinase ABC and AC were obtained from Miles Laboratories, Elkhart, Ind., USA. ANALYSIS PROCEDURE Uronic acid analysis was done by the procedure of Bitter & Muir (1962), using glucuronolactone as a standard. Protein content was determined using the Folin reagent (Lowry, Rosenbrough, Farr & Randall, 1951), with bovine serum albumin as a standard. Hexosamines were measured by the method of Developing limb-bud proteoglycans 63 Elson & Morgan (1933) modified by Antonopoulos, Gardell, Sziramai & DeTyssonk (1964). Limb buds of stage 34 and 38 were from White Leghorn chick embryos whose development was determined by the staging series of Hamburger & Hamilton (1951). The cartilages, dissected free of adhering tissue, were cut into small pieces. Proteoglycans were labeled by incubating the limb buds in F12X nutrient medium containing 20/^Ci/ml of carrier-free Na^SC^ (New England Nuclear, Boston, MA., USA). After 16 h in a humidified incubator in 95 % air and 5 % CO2, the tissues were rinsed twice with Simms & Saunders (1942) salt saline solution. Proteoglycans were extracted with 4-0 M guanidinium chloride buffered at pH5-8 with 0 0 5 M sodium acetate which contained 0-01 M EDTA, 0-10M-6 aminohexanoic acid, and 0-005 M benzamidine hydrochloride to inhibit proteolysis (Oegema, Hascall & Dziewiatkowski, 1975). The tissues were agitated in the solvent on a rotary shaker for 48 h at 4 °C. The extract was dialysed for 48 h against cold distilled water, and cleared by centrifugation at 1O85O# for 30 min at 4 °C, and the supernatants lyophilized. Controlled pore glass-bead {CPG-10-2500) chromatography of proteoglycan extracts Proteoglycans in 0-5 M-NaCl were applied to a CPG-10-2500 column (100 x 0-9 cm) and eluted with 0-5 M-NaCl at the rate of 0-4 ml/min (Vasan & Lash, 1978). A portion from each 1 ml fraction was used for determining radioactivity or uronic acid analysis (if unlabelled). The peaks (PG-A, PG-I & PG-M-ubiquitous) were pooled, dialysed and concentrated for further study. Associative guanidinium chlorideI CsCl-density-gradient centrifugation A portion of 4-0 M guanidinium chloride extract was dialysed against 1000 volumes of 0-5 M guanidinium chloride/sodium acetate buffer, pH 5-8 for 24 h at 4 °C. Solid CsCl was added to the residues to a density of 1-60 g/ml. These residues were centrifuged in a Beckman L2-65B preparative ultracentrifuge with type 40 fixed-angle rotor at 15 °C and 39000 rev/min for 44 h. The tube contents were collected in 1 ml fractions, and a portion was used for determining density and uronic acid assay after dialysis. The lower 5 ml in stage 34 and 3 ml in stage 38 (from associative gradient centrifugation) were pooled separately (Al fraction) and processed as described below. The Al fraction was mixed with an equal volume of 7-5 M guanidinium chloride, brought to a density of 1-5 g/ml with solid CsCl and centrifuged as above. The tube content was collected in 1 ml fractions, and a portion analysed for uronic acid and density. The lower 4 ml in stage 34 and 38 were collected separately (Al-Dl fractions), dialysed, lyophilized, and used for further studies. The top 2 ml containing 3-2 64 N. VASAN link proteins were pooled (A1-D4 fraction) in each sample and used for polyacrylamide gel electrophoresis. Interaction of A1-D1 fractions with hyaluronic acid A l - D l fractions (35S-labelled) in 0-5 ml of 4 0 M guanidinium chloride were mixed with rooster comb hyaluronate (0-2 ml) containing 40 /*g of uronic acid. After 60 min at room temperature the mixture was dialysed against the elution medium (0-5 M sodium chloride) before molecular sieve chromatography on a CPG-10-2500 column (100 x 0-9 cm) at 4 °C. Dialysis against the low ionic medium facilitates interaction of proteoglycans with hyaluronic acid. The portion of proteoglycan bound to hyaluronate was determined from a comparison of the amount of radioactive-labelled proteoglycan eluted in the column's void volume region with the amount of control proteoglycan chromatographed in the absence of hyaluronate (Hardingham & Muir, 1974). Enzymatic treatments Samples were treated with chondroitinase ABC and AC (Saito, Yamagata & Suzuki, 1968) to determine chondroitin 4- and 6-sulphate content. Papain digestion was done at 58 °C for 10 h in 0-1 M sodium acetate, pH 6-8, containing 0-005 M EDTA (disodium salt) and 0005 M cystein hydrochloride (Vasan, 1980). Polyacrylamide gel electrophoresis The A1-D4 fraction (rich in glycoprotein link factors) was dialysed against distilled water, then lyophilized. Portions of the dried samples subsequently dissolved in 0-025 M Tris-HCl (pH 6-8), 1 % SDS, 10% glycerol, and 0-02% bromphenol blue (as the marker dye), and were subjected to electrophoresis (Weber & Osborn, 1969) on sodium dodecyl sulphated-polyacrylamide (7-5%) gels and scanned. Direct dissociative guanidinium chloride/CsCl centrifugation Proteoglycan aggregate fraction (PG-A) from stage-38 limb and intermediate fraction (PG-I) from stage-34 limb buds were isolated as described earlier. These fractions were dissolved in 4-0 M guanidinium chloride buffer, and brought to a density of 1-5 g/ml by the addition of solid CsCl (0-6 g/ml of solution). A direct dissociative gradient was established by centrifugation at 15 °C and 39000 rev/min for 44 h (Oegema et al. 1975). The gradients were partitioned into a lower two-fifths (Dl) fraction and an upper three-fifths in the case of PG-A (stage 38), and lower half (Dl) and upper half in the case of PG-I (stage 34). Because the surface gel did not contain a significant amount of uronic acid, it was discarded. The Dl fraction was dialysed, lyophilized and subjected to CPG-10-2500 column chromatography to determine molecular size. Developing limb-bud proteoglycans 65 Papain-digested Dl fractions were subjected to Sephadex G-200 column chromatography (80 x 0-8 cm) to determine the length of glycosaminoglycan side chains (Vasan, 1980). Radiolabelled antigen-binding assay was done by the method of Farr (1958). Twenty microlitres of immune serum and 20 /i\ of [35S]proteoglycan were mixed and incubated at 4 °C overnight. Subsequently 50 /*1 of 80 % ammonium sulphate was added to each tube, mixed and left at 4 °C for 60 min. The tubes were then centrifuged and the supernatant fluids were removed. The pellets were washed twice with 50% ammonium sulphate, resuspended in 0-1 ml of phosphate-buffered saline and counted in a liquid scintillation counter. Each sample was analysed in duplicate. The control tube contained normal (nonimmune) rabbit serum in place of immune serum and was processed similarly. The total count in the amount of antigen added was determined by counting a 20 /A sample in a liquid scintillation counter. Efficient precipitation concentration for each proteoglycan fraction was determined separately. Specific binding is defined as the amount of radioactivity precipitated with immune serum, while nonspecific binding is the radioactivity precipitated with non-immune serum. Corrected specific binding is determined using the formula described by Ho, Levitt & Dorfman (1977): percentage bound = „ total c.p.m. added-c.p.m. bound by immune serum total c.p.m. added-c.p.m. bound by non-immune serum . ~~ The antiserum prepared against chick sternal cartilage specific proteoglycan aggregate was a generous gift from Dr Paul Goetinck (Sparks, Lever & Goetinck, 1980). RESULTS Chondrogenesis in limb development is marked by changes in the size of proteoglycans synthesized. For instance, the total amount of larger aggregate increases while smaller monomers decrease. Chromatography of proteoglycans from stage 34 showed 18-25 % PG-A; 65-70 % PG-I and 8-10 % PG-M ubiquitous (Fig. 1). Stage 38 contained 70-80 % PG-A, 20-30 % PG-I and 5-8 % PG-M ubiquitous (Fig. 1). The proteoglycan molecule excluded from the column is designated as PG-A. Among the included, PG-I represents the molecule eluted very close to the void volume, and the PG-M ubiquitous refers to proteoglycan eluted at the total volume of the column. Proteoglycan monomers of cartilage are chemically different from ubiquitous proteoglycan monomers which are present in cartilage or in other tissues (Vasan, unpublished work in progress). Table 1 shows the distribution of the disaccharide digestion products from various proteoglycans resolved on the CPG-10-2500 column (Fig. 1). Radioactive sulphate counting indicated that in stage-34 PG-I 46 and 4 7 % are 4-sulphated and 6-sulphated disaccharides and 6% are resistant material; in stage-38 PG-A the corresponding percentages are 44, 49, and 7 %. 66 N. VASAN PG-A PG-M ., PG-I 0-25 0-20 0-15 §010 005 10 20fK 30 40 50 60 f 70 Fraction no. Fig. 1. Controlled pore glass-bead (CPG-10-2500) column chromatogram (100 x 0-9 cm) of the proteoglycan extracted from the limb buds. Proteoglycan extract was dialysed against the elution buffer (0-5 M-NaCl pH 6-5) before application on the column. The elution resulted in a profile which contained PG-A (excluded peak), PG-I (included peak close to the void volume), and PG-M (eluted at the total volume). O, Stage 38: # , Stage 34. The void volume (fraction 22) and total volume (fraction 64) are shown by arrows. Table 1. Relative amounts (percent) of the various 35S-Iabelled glycosaminoglycans present in the PG-A and PG-I fractions Type of glycosaminoglycans Chondroitin 4-sulphate* Chondroitin 6-sulphatef Resistant materialsf Stage 34/PG-I Stage 38/PG-A 46 47 7 44 49 7 * Determined as the relative amount of radioactivity associated with 4-sulphated disaccharide after chondroitinase enzyme treatment. t Relative amount of radioactivity associated with 6-sulphated disaccharide. t Labelled material not degraded by chondroitinase enzyme. Protein and uronic acid ratios (weight/weight) were 1:7-7 for stage-34 PG-I and 1:11-3 for stage-38 PG-A. Glucosamine/galactosamine molar ratios were 1:8-8 and 1:9-3 for PG-A and PG-I, respectively. Figure 2A shows the results of associative guanidinium chloride/CsCl gradient centrifugation. The Al fractions from stage-34 PG-I were distributed in the bottom four-tenths of the tube, while those from stage-38 PG-A were in the bottom three-tenths. The distribution pattern in associative gradient Developing limb-bud proteoglycans 67 0-20 i - 3 4 5 6 Density (g/ml) 7 9 10 <l-39 Fig. 2. Guanidinium chloride/CsCl-density gradient centrifugation profile of proteoglycans extracted from the limb buds. (A) Sedimentation under associative condition (0-5 M guanidinium chloride). (B) Al fraction from above (see materials and methods) centrifuged under dissociative condition (4 M guanidinium chloride) and Al-Dl and A1-D4 were isolated as described in the methods. D, Stage 34, PG-I: m, Stage 38, PG-A. indicates that the stage-34 PG-I molecules are of lower buoyant density compared to those of stage-38 PG-A. Pooled Al fractions were subjected to dissociative guanidinium chloride/ CsCl gradient centrifugation, and disaggregated (Al-Dl) molecules (Fig. 2B) were obtained in the lower two-fifths of the gradients. Mixtures of A l - D l with high-molecular-weight hyaluronic acid were prepared in dissociative conditions, dialysed into associative conditions and chromatographed on CPG-10-2500 volume. A l - D l from stage 38 showed a small proportion of proteoglycan aggregate excluded from the column, while stage 34 showed a bimodal distribution. This shows that A l - D l obtained from stage-34 PG-I is a heterogeneous molecule that appears in the included portion of the CPG-10-2500 column. The elution profiles of A l - D l in the presence of hyaluronic acid for both PG-A and PG-I indicate that a large proportion of the labelled proteoglycan molecules interact with hyaluronic acid (Fig. 3). The aggregate fraction (Al) from stage-38 PG-A and stage-34 PG-I were centrifuged under dissociative conditions. The glycoprotein link factor rich fraction (A1-D4) isolated from the upper one-fifth of the gradient was analysed by sodium dodecyl sulphate polyacrylamide gels, and showed no difference between the two samples (Fig. 4). One major protein band present was at the 68 N. VASAN Fraction no. Fig. 3. Controlled pore glass-bead (CPG-10-2500) column (100x0-9 cm) chromatogram of material from Al-Dl in the presence of rooster comb hyaluronate. • , Stage 34 A l - D l ; D, Stage 34 Al-Dl + hyaluronate • , Stage 38 A l - D l ; O, Stage 38 Al-Dl + hyaluronate. For details see the text. The arrows indicate the void volume (fraction 22) and total volume (fraction 64). Fig. 4. SDS-polyacrylamide gel electrophoresis pattern of A1-D4 fraction (glycoprotein link factors). (A) Fraction isolated from stage 38 PG-A. (B) Fraction isolated from Stage 34 PG-I. Note: The smaller fast moving peak has been exaggerated 10 times the original to indicate the definite presence of small molecular weight molecule. Developing limb-bud proteoglycans 69 position characteristic for the larger molecular weight link protein. Much smaller amounts of lower molecular weight link protein were also noted. Disaggregated proteoglycan monomer fractions (Dl) were isolated from PG-A and PG-I of stage-38 and -34 limb buds respectively. The uronic acid elution profile on CPG-10-2500 column (Fig. 5) shows the Dl from stage-38 PG-A is larger (Kav -0-41) than the Dl stage-34 PG-I (Kav-0-57). Also the 0-20 r 015 010 005 10 20fVo 30 40 60 \ v 50 70 Fraction no. Fig. 5. Controlled pore glass-bead (CPG-10-2500) column (100x0-9 cm) chromatogram of material from Dl fraction. O, PG-A from stage 38; • , PG-I from stage 34. 0-20 r 015 010 005 10 30 40 50 f. 60 70 Fraction no. Fig. 6. Sephadex G-200 column (80 x 0-8 cm) chromatogram of the papain-digest fraction. The column waseluted by using0-4 sodium buffer, pH 6-8, and lml fractions were collected. The pattern represents the glycosaminoglycan side chains. O, stage 38 PG-A; • , stage 34 PG-I. 70 N. VASAN stage-34 PG-I monomer (Dl) preparation did not exhibit large molecules that eluted in the void volumes in small quantities, as seen in monomers (Dl) prepared from stage-38 PG-A. Papain digest of Dl samples isolated from PG-I (stage 34) and PG-A (stage 38) were chromatographed on sephadex G-200 to determine the length of the released glycosaminoglycan chains (Fig. 6). For the stage-38 PG-A sample, the released chains were of slightly smaller average molecular size (Kav 0-30). The sizes synthesized in the stage-34 PG-I were a little larger (Kav-0-23). Table 2. ™S-labelled antigen fractions resolved on a CPG-10-2500 column bound by immune serum and non-immune serum Proteoglycan fractions Stage 34 PG-A PG-I PG-M (ubiq.) Stage 38 PG-A PG-I PG-M (ubiq.) Bound by immune serum (%)• Bound by non-immune serum (%)f Percent bound corrected for non-immune serum (%)} 81-3 70-7 21-2 22-6 21-1 27-7 81-2 681 00 88-6 74-2 19-7 20-6 22-4 29-4 85-6 69-8 00 • % bound = c.p.m. antigen bound by immune serum/c.p.m. antigen added x 100. f % bound = c.p.m. antigen bound by non-immune serum/c.p.m. antigen added x 100. fc.p.m. antigen added . .. , , ... . — c.p.m. antigen bound by immune serum xlOO. % % bound = 1 0 0 - ' c.p.m. antigen added -c.p.m. antigen bound by non-immune serumy Immune serum binding of proteoglycan fractions This batch of antisera obtained from Dr Goetinck's laboratory has been shown to contain IgG with a specificity directed against large proteoglycan from chick sternum, and has no affinity for glycosaminoglycan chains (Sparks, Lever & Goetinck, 1980). Table 2 shows that the immune serum has very high affinity for PG-A (81-89 %) and moderate reactivity for PG-I (70%). Furthermore, the present immune serum failed to exhibit any affinity for PG-M (ubiquitous) from either stage. In all the proteoglycan fractions studied, there was no 100% precipitation of antigen, and the non-immune serum showed 20-30% precipitation with the labelled antigen. Developing limb-bud proteoglycans 71 DISCUSSION Developmental significance of intermediate size PG (PG-I) was studied, since these molecules were reported to be present at certain stages of limb growth (Lash et al. 1978; Ovadia et al. 1980). When resolved on a CPG-10-2500 column (Fig. 1) the stage-38 proteoglycan molecules were mostly large (Kav 0-00-0-23), eluted in the void volume (PG-A), and also contained a small amount of intermediate (PG-I) size molecules (Kav 0-23-0-38). The stage-34 proteoglycan molecules were mostly intermediate in size (PG-I), eluted very close to the void volume with a Kav -0-32, and also contained a small amount of larger size molecules (Kav - 0-23) excluded from the column. Both stages contained equal amounts of smaller molecular weight ubiquitous components eluted in the total volume. Goetinck et al. (1974) chromatographed proteoglycan synthesized by 9-day-old chick limb-bud chondrocyte culture on 1 % agarose and reported an excluded peak and a partially included peak. The decrease in size of the PG (PG-I) could be due to one or more of the following: lack of aggregation with hyaluronate, lack of stabilizing glycoprotein link factor(s), shorter core protein, thus fewer GAG side chains; normal core protein with fewer GAG side chains or decrease in GAG side chain length. PG-A and PG-I in our study exhibited the same amount of chondroitin 6-sulphate and 4-sulphate and glucosamine/galactosamine ratios. However, the protein/uronic acid ratio was 30% higher for PG-A (1:11-3) than for PG-I (1:7-7). Furthermore, papain digests of Dl samples isolated from PG-A and PG-I showed that released polysaccharide chains (Fig. 6) of PG-A were slightly smaller (Kav -0-30) than those of PG-I (Kav -0-24). PG-I subjected to dissociated gradient centrifugation banded as a low buoyant fraction (see Materials and Methods). The average length of the chains prepared by papain digestion (Fig. 6) showed that the PG-1 chains were slightly larger than the PG-A chains. However, the protein/uronic acid ratio clearly indicates fewer chondroitin sulphate chains attached to the protein core. The distribution of Al fractions in the associative gradient (Fig. 2 A) and the A l - D l fraction in the dissociative gradient (Fig. 2B) also suggest that PG-I molecules have lower buoyant density than PG-A molecules. But the average molecular sizes of the proteoglycan synthesized by fully expressed chondrocytes vary from tissue to tissue. For example, the sedimentation coefficients (So) of bovine nasal and tracheal proteoglycan are about 25 s (Hascall & Heinegard, 1974), bovine articular cartilage proteoglycan, 14-3 s (Rosenberg et al. 1976), chick limb-bud chondrocyte day-8 culture proteoglycan, 19 s (Hascall, Oegema, Brown & Caplan, 1976), and rat epiphyseal proteoglycan, 13 s (Pita, Muller & Howell, 1975). Electron microscopic measurements of proteoglycans correlate with differences in the length of the molecule. Several investigations suggest, moreover, that within any single cartilage large variations in size and 72 N. VASAN length of proteoglycan correlate to the number of chondroitin sulphate chains attached to the core protein and to the length of that portion of core protein to which the chondroitin sulphate chains are attached (Hardingham, Ewins & Muir, 1976; Heinegard, 1977; Hascall & Heinegard, 1975). The differences in average sizes of proteoglycan synthesized at stage 34 and stage 38 may well reflect the properties of molecules synthesized by the chondrocytes. The multiple-size proteoglycan observed in this study may also be due to the size of the core protein as suggested earlier (Heinegard, 1972; Baxter & Muir, 1975; Palmoski, Khosla & Brandt, 1974). Sparks, Lever & Goetinck (1980) also suggest that post-translational modifications of the core protein could also contribute to the differences in the size of proteoglycan. Proteoglycan monomers ( A l - D l ) from both the samples were able to interact with exogenously provided hyaluronic acid (Fig. 3), providing evidence that the core protein of PG-I and PG-A is similar at the interactive end. Thus it is possible that in vivo, the hyaluronic acid molecule is not present in sufficient quantity. Increase in hyaluronidase and decrease in hyaluronate has been described as a phenomenon of limb differentiation (Toole, 1973, also lends support to the present result). It is also equally possible that the hyaluronate present at this stage is of low molecular weight and thus inadequate to form aggregate. While such a situation has not been reported during chondrogenesis, decrease in molecular weight of hyaluronic acid has been observed in osteoarthritis (Vasan, unpublished work in progress) causing decrease in the proteoglycan aggregates. Polyacrylamide gel electrophoresis showed the larger link protein molecules predominate (Fig. 4) in both PG-A and PG-I samples, but very small amounts of slower moving smaller link protein also occur. The larger link molecules were also found in proteoglycan aggregate from hyaline cartilages (Keiser, Shulman & Sandson, 1972; Hascall & Heinegard, 1974; Baker & Caterson, 1979), chick limb buds (Vasan & Lash, 1977), chick limb chondrocytes (Hascall et al. 1976; Lohmander, Hascall & Caplan, 1979), and rat epiphyseal cartilage (Pita et al. 1979). Utilizing the immunological methods, further evidence has been obtained which indicates that the PG-A and PG-I are cartilage-specific molecules which appear during the differentiation of limb buds to cartilage. In contrast, the sulphated proteoglycan in PG-M, an ubiquitous molecule, failed to react with the immune serum (Table 2). This agrees with the earlier report (Sparks et al. 1980) where proteoglycan from skin and minor proteoglycan component of cartilages (PGS-II) did not react with this antiserum. The results of the experiment described here clearly show that the intermediate-size proteoglycans (PG-I) resemble the cartilage-specific large aggregates (PG-A) in a number of physical and chemical characters and also differ in some respects. The presence of intermediate-size proteoglycan has also been observed in somite chondrogenesis (Lash & Vasan, 1978). Skeletal muscle forming cartilage on bone matrix in vitro also synthesizes proteoglycan molecules that migrate similarly on molecular sieve column (Nathanson, 1981). The Developing limb-bud protectglycans 73 presence of intermediate-size proteoglycan during chondrogenic expression could also be due to degradation of large aggregates. While this is a speculation, the developmental significance of this molecule is being further investigated. I am grateful to Dr Paul Goetinck for his helpful suggestions and discussion. The author wishes to thank Marcus Meyenhofer for photographic assistance and Marge Pascavage for typing the manuscripts. This investigation was supported by a grant from National Institute of Health 5 S07 RRO 5393. REFERENCES ANTONOPOULOS, C. A., GARDELL, S., SZIRAMAI, J. A. & DE TYSSONK, E. (1964). Determination of glycosaminoglycans (mucopolysaccharides) from tissues on the microgram scale. Biochim. biophys. Acta 83, 1-9. BAKER, J. R. & CATERSON, B. (1979). The isolation and characterization of the link proteins from proteoglycan aggregates of bovine nasal cartilage. / . biol. Chem. 254, 2387-2393. BAXTER, E. & MUIR, H. (1975). The nature of the protein moieties of cartilage proteoglycans of pig and ox. Biochem. J. 149, 657-668. BITTER, T. & Mum, H. (1962). A modified Uronic Acid carbazole reaction. Anal. Biochem. 4, 330-334. DE LUCA, S., HEINEGARD, D., HASCALL, V. C , KIMURA, J. H. & CAPLAN, A. I. (1977). Chemical and physical changes in proteoglycans during development of chick limb bud chondrocytes grown in vitro. J. biol. Chem. 252, 6600-6608. ELSON, L. A. & MORGAN, W. T. J. (1933). A colorimetric method for the determination of glucosamine and chondrosamine. Biochem. J. 27, 1824-1828. FARR, R. S. (1958). A quantitative immunochemical measure of the primary interaction between I*BSA and antibody. J. infect. Dis. 103, 239-262. GOETINCK, P. F., PENNYPACKER, J. P. & ROYAL, P. D. (1974). Proteochondroitin sulfate synthesis and chondrogenic expression. Expl Cell Res. 87, 244-248. HAMBURGER, V. & HAMILTON, H. C. (1951). A series of normal stages in the development of the chick embryo. / . Morph. 88, 49-92. HARDINGHAM, T. E., EWINS, R. J. F. & MUIR, H. (1976). Cartilage proteoglycans: structure and heterogeneity of the protein core and the effects of specific protein modifications on the binding of hyaluronate. Biochem. J. 157, 127-143. HARDINGHAM, T. E. & MUIR, H. (1974). Hyaluronic acid in cartilage and proteoglycan aggregation. Biochem. J. 139, 565-581. HASCALL, V. C. & HFINEGARD, D. (1974). Aggregation of cartilage proteoglycans. I. The role of hyaluronic acid. / . biol. Chem. 249, 4232-4241. HASCALL, V. C. & HEINEGARD, D. (1975). The structure of cartilage proteoglycans. In Extracellular Matrix Influences on Gene Expression (ed. H. C. Slavkin & R. C. Greulich), pp. 423-434. New York: Academic Press. HASCALL, V. C , OEGEMA, T. R., BROWN, M. & CAPLAN, A. I. (1976). Isolation and characterization of proteoglycans from chick limb bud chondrocytes grown in vitro. J. biol. Chem. 251, 3511-3519. HEINEGARD, D. (1972). Hyaluronidase digestion and alkaline treatment of bovine tracheal cartilage proteoglycans. Isolation and characterization of different keratan sulfate proteins. Biochim. biophys. Acta 285, 193-207. HEINEGARD, D. (1977). Polydispersity of cartilage proteoglycans. Structural variations with size and buoyant density of the molecules. / . biol. Chem. 252, 1908-1989. HEINEGARD, D. & AXELSSON, I. (1977). Distribution of keratan sulfate in cartilage proteoglycans. / . biol. Chem. 252, 1971-1979. HEINEGARD, D. & HASCALL, V. C. (1974). Aggregation of cartilage proteoglycans. III. Characteristics of the proteins isolated from trypsin digests of aggregates. / . biol. Chem. 249, 4250-4256. Ho, P. L., LEVITT, D. & DORFMAN, A. (1977). A radioimmune study of the effect of bromo- 74 N. VASAN deoxyuridine on the synthesis of proteoglycan by differentiating limb bud cultures. Devi Biol. 55, 233-243. KEISER, H., SHULMAN, H. J. & SANDSON, J. I. (1972). Immunochemistry of cartilage proteoglycan. Immunodiffusion and gel-electrophoretic studies. Biochem. J. 126, 163-169. LASH, J. W., OVADIA, M. & VASAN, N. S. (1978). Microheterogeneities, non-equivalence, and embryonic induction. Med. Biol. 56, 333-338. LASH, J. W. & VASAN, N. S. (1978). Somite chondrogenesis in vitro: stimulation by exogenous extracellular matrix components. Devi Biol. 66, 151-171. LOHMANDER, L. S., HASCALL, V. C. & CAPLAN, A. I. (1979). Effects of 4-methyl umbelliferyl-B-D-xylopranoside on chondrogenesis and proteoglycan synthesis in chick limb bud mesenchymal cell cultures. / . biol. Chem. 254, 10551-10561. LOWRY, O. H., ROSENBROUGH, N. J., FARR, A. L. & RANDALL, R. J. (1951). Protein measurement with the folin phenol reagent. /. biol. Chem. 193, 265-275. NATHANSON, M. A. (1981). Proteoglycan synthesis by skeletal muscle forming cartilage on bone matrix in vitro. J. Cell Biol. 91, 160a. OEGEMA, T. R., HASCALL, V. C. & DZIEWIATKOWSKI, D. D. (1975). Isolation and characterization of proteoglycans from the swarm rat chondrosarcoma. / . biol. Chem. 250, 6151-6159. OVADIA, N., PARKER, C. H. & LASH, J. W. (1980). Changing patterns of proteoglycan synthesis during chondrogenic differentiation. /. Embryol. exp. Morph. 56, 59-70. PALMOSKI, M., KHOSLA, R. & BRANDT, K. (1974). Small proteoglycans of cartilage: confirmation of their presence by non-disruptive extraction. Biochim. biophys. Acta 372, 171-175. PITA, J. C , MULLER, F. & HOWELL, D. S. (1975). Disaggregation of proteoglycan aggregate during endochondral calcification: Physiological role of cartilage lysozyme. In Dynamics of Connective Tissue Macromolecules (ed. P. M. C. Burleight & A. R. Poole), pp. 247-258. Amsterdam: North-Holland. ROSENBERG, L., WOLFENSTEIN-TODEL, C , MARGOLIS, R., PAS, S. & STRIDER, W. (1976). Proteoglycans from bovine proximal humeral articular cartilage. Structural basis for the polydispersity of proteoglycan subunit. / . biol. Chem. 251, 6439-6444. ROYAL, P. D. & GOETINCK, P. F. (1977). In vitro chondrogenesis by mouse limb mesenchymal cells; changes in ultra-structure and proteoglycan synthesis. J. Embryol. exp. Morph. 39, 79-95. SAITO, H., YAMAGATA, T. S. & SUZUKI, S. (1968). Enzymatic methods for the determination of small quantities of isomeric chondroitin sulfate. /. biol. Chem. 243, 1536-1542. SIMMS, H. S. & SAUNDERS, M. (1942). The use of serum ultrafiltrate in tissue culture for studying fat deposition and virus propagation. Archs Path. 33, 619-635. SPARKS, K. J., LEVER, P. L. & GOETINCK, P. F. (1980). Antibody binding of cartilage-specific proteoglycans. Archs Biochem. Biophys. 199, 579-590. TOOLE, B. (1973). Hyaluronate and hyaluronidase in morphogenesis and differentiation. Am. Zool 13, 1061-1065. VASAN, N. S. (1980). Proteoglycan in normal and severely osteoarthritic human cartilage. Biochem. J. 187, 781-787. VASAN, N. S. (1981). Intermediate size proteoglycans in limb development: physical and chemical analysis. Fedn Proc. 40, 1809 abst. VASAN, N. S. & LASH, J. W. (1977). Heterogeneity of proteoglycans in developing chick limb cartilage. Biochem. J. 164, 179-183. VASAN, N. S. & LASH, J. W. (1978). Proteoglycan heterogeneity in embryonic chick articular and epiphyseal cartilages. Connect. Tissue Res. 6, 191-199. VASAN, N. S. & LASH, J. W. (1979). Monomeric and aggregate proteoglycans in the chondrogenic differentiation of embryonic chick limb buds. J. Embryol. exp. Morph. 49, 47-59. WEBER, K. & OSBORN, M. (1969). The reliability of molecular weight determination by dodecyl sulfate-polyacrylamide gel electrophoresis. /. biol. Chem. 244, 4406-4412. (Received 27 July 1981, revised 18 January 1982)