Connective Tissue Research, 1992, Vol. 27, pp. 265-277 Reprints available directly from the publisher
Photocopying permitted by license only 0 1992 Gordon and Breach Science Publishers S.A. Printed in the United States of America
CHARACTERIZATION OF A CHONDROITIN SULFATE PROTEOGLYCAN SYNTHESIZED BY MONKEY ARTERIAL SMOOTH MUSCLE CELLS IN VZTRO
Connect Tissue Res Downloaded from informahealthcare.com by UB Giessen on 12/10/14 For personal use only.
TET-KIN YEO, STEPHEN MACFARLANE, and THOMAS N. WIGHT Department of Pathology SM-30, School of Medicine, University of Washington, Seattle, WA 98195 (Received June I , 1990; in revised form December 10, 1990; accepted October 16, 1991)
A monoclonal antibody against arterial smooth muscle cell chondroitin sulfate proteoglycan has been developed. Incubation of [3sS]-methionine labeled proteoglycans with MAb 941 quantitatively immunoprecipitated all the chondroitin sulfate proteoglycan (CSPG) synthesized by these cells. Digestion of the immunoprecipitate with chondroitin AC lyase revealed one major protein band (M, 420,000) and two minor bands (M, 509,000 and 390,000) on SDS-PAGE that are composed of very similar peptides when analyzed by limited peptide digestion by S. aureus V8 protease. Additional studies demonstrated that this monoclonal antibody recognized an epitope on the chondroitin sulfate chains. However, only a minor subpopulation (5-12%) of the alkaline-borohydride released glycosaminoglycan chains was immunoprecipitated and this subset of chains was slightly larger than the nonimmunoprecipitated chains. High pressure liquid chromatography analysis of the disaccharides generated from the immunoprecipitated glycosaminoglycan chains demonstrated that these chains were enriched in chondroitin-6sulfate relative to chondroitin-4-sulfate (2:l) while that of the non-immunoprecipitated chains had a ratio of I:1. These studies indicate that at least two distinct pools of chondroitin sulfate chains are present on all the chondroitin sulfate proteoglycan synthesized by arterial smooth muscle cells: a major population (89-95%) containing 6-sulfate and 4-sulfate in relatively equal proportion and a minor population (5-12%) which is hydrodynamically larger with a 6-sulfate to 4-sulfate ratio of 2: 1. KEYWORDS: antibody, chondroitin sulfate proteoglycan, smooth muscle cells, chondroitin sulfate chains
INTRODUCTION Although proteoglycans are minor components in blood vessels, a number of studies have demonstrated that these macromolecules influence a number of arterial properties such as viscoelasticity, permeability, lipid metabolism, hemostasis, and thrombosis.1 In order to fully understand the role that proteoglycans play in these processes as well as in the pathogenesis of atherosclerosis,2 it is important to structurally characterize the proteoglycans synthesized by vascular cells. Cultured monkey arterial smooth muscle cells synthesize at least three populations of proteoglycans, chondroitin sulfate proteoglycan (CSPG), dermatan sulfate proteoglycan (DSPG), and heparan sulfate proteoglycan (HSPG), of which CSPG predominates.3.4 Similar populations of proteoglycans have been described for intact aortic tissue.5fj Biochemical and ultrastructural characterizationof CSPG synthesized by monkey arterial smooth muscle cells indicates that it is a large molecule consisting of a core glycoprotein Address for correspondence: Tet-Kin Yeo, Department of Pathology, Harvard Medical School, Beth Israel Hospital, 330 Brookline Avenue, Boston, MA 02215
265
Connect Tissue Res Downloaded from informahealthcare.com by UB Giessen on 12/10/14 For personal use only.
266
T. YE0 et al.
measuring 220-320 nm in length, to which are attached 15-20 chondroitin sulfate (CS) chains, each with a M, = 43-60kd.3.4 This molecule contains a high content of 0-linked oligosaccharides, approximately 300 monomers per molecule resulting in a ratio of 15:l 0-linked oligosaccharides per glycosaminoglycan (GAG) chain. In addition, the molecule contains few N-linked oligosaccharides (20 monomers per m0lecule).3~4Although the majority of the CSPG secreted by monkey arterial smooth muscle cells is not aggregated under cell culture conditions, they are capable of forming link stabilized aggregates when supplied with hyaluronic acid.3.4 Using a monoclonal antibody (MAb 941) generated against the arterial CSPG,7 Lark et al demonstrated that this CSPG was enriched in developing atherosclerotic lesions. Therefore, the present study has been undertaken to define the epitope recognized by this antibody and to further characterize the molecular characteristics of the core glycoprotein and GAG chains of this CSPG. MATERIALS AND METHODS Materials Guanidine HCl (grade I), Tris base, EDTA, papain (Crude powder), N-ethylmaleimide and phenylmethylsufonyl fluoride (PMSF) were all purchased from Sigma Chemical Co., St. Louis MO; 6-aminohexanoicacid and benzamidine were from Eastman Kodak Co., Rochester NY; chondroitin AC lyase, ABC lyase and rabbit anti-mouse and Pansorbin were from ICN Biomedicals, Inc., Costa Mesa CA; sodium borohydride was from Wilshire Chemical Co., Gardena CA; Spectrapor dialysis tubing was from Spectrum Medical Ind., Inc., Los Angeles CA; Pronase and Triton X-100 were from Calbiochem, San Diego CA; Anti-DRC-1 was from Accurate Chem. & Scientific Corp., Westbury NY; DEAE-Sephacel and Sepharose CL-2B, and CLdB were from Pharmacia Fine Chemicals, Piscataway NJ; Biogel p-10, Acrylamide,N,N’-Methylene-bis-acrylamide,N,N,Nr,N’-Tetramethylethylenediamine and ammonium persulfate were from Biorad Laboratories, Richmond CA; Na, [35S04=](carrier-free), [35S]-methionine (1198 Ci/mmole) and D-[6-3H]-glucosamine HCl (31 Ci/mM) were from New England Nuclear, Boston MA; GelBond PAG film was from FMC Corp., Rockland ME; all cell culture supplies were from Grand Island Biological Co., Grand Island NY. All other chemicals were reagent grade. Cell Culture and Labeling Arterial smooth muscle cell cultures were established from pigtail monkeys (Macaca nernestrina) as described.3.4 Subconfluent cultures were labeled with 50 pCi/ml of [35S}methionine, or [35S04=]and 5 pCi/ml of [3H]-glucosamine for 24-48 h. Proteoglycan Isolation Upon termination of the labeling period, the medium from the cell cultures was removed and solid urea, Triton X-100, and protease inhibitors were added to give a final concentrationof 8 M urea, 0.3%Triton X-100, 2.5 mM EDTA, 5 mM benzamidine hydrochloride, 100 mM 6-aminohexanoic acid, and 1 mM PMSE4 Medium (20 ml) was passed through a 1.0 ml DEAE-Sephacel column which was pre-equilibrated with 8.0 M urea, 0.25 M NaCl buffer
CHARACTERIZATION OF CHONDROITIN SULFATE PROTEOGLYCAN
267
(8 M urea, 2 mM EDTA, 0.3% Triton X-100 and 50 mM Tris, pH 7.5). The DEAE column was washed with 20 ml of the above buffer to remove non-bound material and was eluted stepwise with 0.25 ml, 0.6 ml, 0.6 ml and a 1.0 ml of 8 M urea, 1.0 M NaCl buffer. After determining the radioactivity present in each eluted fraction, the fractions containing the radiolabel (usually fractions 2 and 3) were pooled and dialyzed.
Connect Tissue Res Downloaded from informahealthcare.com by UB Giessen on 12/10/14 For personal use only.
Immunoprecipitation Proteoglycans from the medium were first partially purified as described above and the pooled fractions exhaustively dialyzed against Ip buffer (50 mMTris-HC1, pH 7.15,150 mM NaC1, 1% Triton X-100,2.5 mM EDTA, 5 mM benzamidine hydrochloride, 100 mM 6-aminohexanoic acid, 1 mM PMSF). The antigen was pre-immunoprecipitatedwith M709 (Dako mouse monoclonal anti-DRC1) for 1 h followed by incubation with fixed and heat killed S. aureus (Pansorbin) which had been preabsorbed with rabbit anti-mouse Ig for 1 h. The sample was centrifuged in the microfuge for 45 s and the supernatant was incubated with MAb 941 ( I ml of the MAb 941 was used against 1.O X 106cpm radiolabeled antigen) for 1 h. S. uureus preabsorbed with rabbit anti-mouse Ig was added (1 h incubation on rotator) to precipitate the antigen-antibody complex. The pellet was washed five times in Ip wash buffer (Ip buffer containing 0.1% SDS and 1 mg/ml bovine serum albumin) and the immunoprecipitate was analyzed by gel electrophoresis.
Characterization of the Proteoglycan
To study the core glycoprotein of the proteoglycan, chondroitin AC lyase digestions was performed on the immunoprecipitate by resuspending the S. aureus antibody-antigen pellet in 30 p1 of 0.1 M Tris-HCL, pH 6.8, 0.03 M Na acetate, 10 mM EDTA, 10 mM N-ethylamide, 100 pg/ml bovine serum albumin and 1 mM PMSE Chondroitin AC lyase was added at 0.15 U and the incubation was allowed for 2-3 h at 37°C water bath. At the end of the incubation, SDS sample buffer was added to the mixture, boiled for 3 min and analyzed by SDS-PAGE. The GAG chains of the proteoglycan were obtained by alkaline-borohydridetreatment of an ethanol-precipitated sample by reconstituting the dried precipitate in a solution containing 1 M Na borohydride in 0.05 M NaOH and incubating the sample in an oven at 45°C for 24 h. After the reaction was complete, glacial acetic acid was used to neutralize the solution and water was added to dissolve any salt formed. The solution was dialyzed against water and precipitated by incubation with fourfold volume of ice-cold 95% ETOH, 1.3% K acetate containing 100 pg/ml heparin as a carrier for 2 h at 4°C. GAG chains were also prepared by sequential pronase and papain digestion.4 Pronase digestion was performed by incubating the proteoglycans with 0.5 mg/ml of pronase in 0.2 M Tris-HC1, pH 8 at 55-60°C overnight. The reaction was terminated by boiling the mixture for 5 min. Papain (30 pg/ml in a solution containing 0.1 M Na acetate, 5 mMNa EDTA and 5 mM cysteine, pH 7) was then added to the mixture and incubated for 4 h in a 65°C water bath. Upon termination of the digestion, the mixture was cooled and cold trichloroacetic acid was added to a final concentration of 10%. The sample containing the GAG chains was centrifuged at 10,000 g at 4°C for 30 min and the supernatant dialyzed and precipitated by ethanol as described above.
268
T. YE0 et al.
Analysis of the proteoglycan, the core glycoproteins and GAG chains were performed on SDS gel electrophoresis according to the method of Laemmli9 with upper stacking gels of 2.5% or 3% polyacrylamide and lower resolving gradient gels of 3-7.5% polyacrylamide. Gel bond PAG film was used for these gels to provide strength and ease of handling. Determination of the chondroitin-6- and 4-sulfate was carried out on the HPLC using a Whatman Partisil PAC-15 column. The chondroitin-6-sulfatase reaction was accomplished as described by Saito et al, 1968.8
Connect Tissue Res Downloaded from informahealthcare.com by UB Giessen on 12/10/14 For personal use only.
RESULTS Immunoprecipitation Smooth muscle cell cultures were radiolabeled with [35S]-methionineand the proteoglycans extracted as described in Methods. When this material was immunoprecipitated with MAb 941, approximately 2% of the total [35S]-methionine radiolabeled protein was immunoprecipitated. Analysis of the immunoprecipitate on 3-7.5% SDS-PAGE showed a large M, band which ran slightly into the resolving gel, similar to the large [35SO,']-labeled CSPG isolated from CL-2B gel chromatography (Fig. 1, lanes 2 and 4). Treatment of this immunoprecipitate with chondroitin AC lyase prior to electrophoresis consistently revealed 3-4 radiolabeled bands ranging from 390-500 kd, with the predominant band electrophoresing with M, = 420 kd (Fig. 1, lane 3). Limited peptide analysis performed on the three radiolabeled bands using S.aureus V8 protease digestion yielded several peptides, most of which are common for all the three radiolabeled core proteins (Fig. 2). Though quantitative differences are observed in the yielded peptides after V8 digestion, the three radiolabeled core proteins share several similar peptides, suggesting that they are related to one another.
Characterization of the Epitope of MAb 941 [35S]-methionine labeled proteoglycans were isolated form the medium of smooth muscle cells by ion exchange chromatography and digested with chondroitin AC lyase. The resulting material which contains radiolabeled core protein with oligosaccharide stubs of CSPG as well as other proteoglycans were incubated with MAb 941. No significant radioactivity was precipitated and no radiolabeled band was seen on the SDS-PAGE (Fig. 1, lane 5 ) , indicating that the core proteins or oligosaccharides of CSPG are not the epitope of MAb 941. To further demonstrate that the MAb 941 does not recognize the core proteins, immunoprecipitationof the [35S]-methioninelabeled proteoglycans were carried out and the S. aureus antibody-antigencomplex was incubated with chondroitin AC lyase. At the end of the digestion, the mixture was centrifuged to separate the supernatant from the pellet and the two fractions were analyzed. Ninety-seven percent of the immunoprecipitatedradioactivity was associated with the supernatant while only 3% of the radioactivity was associated with the pellet, indicating that upon digestion of the GAG chains, the radiolabeled core protein was released into the supernatant (Fig. 1, lanes 7 and 8). Thus, the antigen was not bound to the antibody via its protein core or its oligosaccharide stubs. [35S]-sulfatelabeled CSPG was prepared by preparative Sepharose CL4B chromatography and the CS chains were isolated by alkaline-borohydride treatment. Immunoprecipita-
269
Connect Tissue Res Downloaded from informahealthcare.com by UB Giessen on 12/10/14 For personal use only.
CHARACTERIZATION OF CHONDROITIN SULFATE PROTEOGLYCAN
FIGURE 1 Analysis of the imrnunoprecipitate by SDS-gel electrophoresis. DEAE concentrated [35S]-methionine labeled proteoglycans were immunoprecipitated with control Ab M709 (lane I), MAh 941 (lane 2) and the immunoprecipitate treated with chondroitin AC lyase to yield the core proteins (lane 3). [35SO,=] labeled CSPG was isolated from a Sepharose CL-2B column under dissociative conditions (4 M GuHC1),7 dialyzed exhaustively against Tris buffer, precipitated by ethanol and was analyzed on 3-7.5% SDS-PAGE (lane 4). To determine the epitope of MAh 941, the [35S]-methionine labeled proteoglycans were first treated with chondroitin AC lyase and the resulting core proteins immunoprecipitated with MAh 941 (lane 5). In another experiment, immunoprecipitated [35S]-methionine labeled CSPG (lane 6) was treated with chondroitin AC lyase when the proteoglycan was still attached to the Pansorbin pellet. The resulting mixture was microfuged and the pellet (lane 7) and the supernatant (lane 8) were analyzed on 3-7.5% SDS-PAGE. The gel was treated with Enlightning and a Ruorograph obtained. The arrow represents the start of the resolving gel, while the numbers on the right represents the M, in kd.
270
T. Y E 0 et al.
Connect Tissue Res Downloaded from informahealthcare.com by UB Giessen on 12/10/14 For personal use only.
A
B.
FIGURE 2 Limited peptide mapping of the core proteins. Immunoprecipitated CSPGs were incubated with chondroitin AC lyase and analyzed on a 4-7.5% SDS-PAGE without the use of a gel bond. The gel was fixed with acetic acid:MeOH:H,O (1.5:1:17.5) for 1 h, dried and exposed to XAR film. The three bands corresponding to the protein core bands (A) M200,OOO; (B) Mr420,000; (C) M,390,000), similar to that seen in Fig. I , lane 3, were cut and equilibrated with gel equilibration buffer (0.125 M Ris-HCI, pH 6.7, 1 mM EDTA, 1% SDS) for 1 h prior to loading the gel slices onto the 15% SDS-PAGE. S. aureus V8 protease (100 pg in 10% glycerol solution) was overlayed above the gel slices and electrophoresis was performed at 5 mA. The current was turned off for 30 min when the proteins were halfway into the stacking gel, then proceeded as usual. The gel was treated with Enlightning and the fluorograph was scanned. The numbers above indicate the relative mobility of the peptides on the gel.
Connect Tissue Res Downloaded from informahealthcare.com by UB Giessen on 12/10/14 For personal use only.
CHARACTERIZATION OF CHONDROITIN SULFATE PROTEOGLYCAN
27 1
tion of the [35S]-sulfate labeled CSPG and CS chains with MAb 941 indicated that although MAb 941 precipitated greater than 90% of the intact CSPG, only 5-9% of the alkalineborohydride derived CS chains were immunoprecipitated. These immunoprecipitated chains are larger in size as analyzed by Sepharose CL-6B gel chromatography (Fig. 3) and electrophoresed more slowly on SDS-PAGE when compared to the total heterogenous population of alkaline-borohydride released chains (Fig. 4A). However, the chains were of similar apparent charge density when chromatographed on a DEAE Sephacel ion exchange column (data not shown). To ascertain if these results were due to the chemical reactions of alkaline-borohydride treatment, thereby altering the epitope of MAb 941, CS chains were also prepared by sequential papain and pronase digestion. MAb 941 immunoprecipitated 18% of the [35S]sulfate labeled CS chains. Sepharose CL-6B chromatographyof the CS chains indicated that the immunoprecipitated chains were larger than the total population of CS chains (data not shown). Analysis by SDS-PAGE indicated that papain-pronase treatment resulted in two populations of CS chains: i) a higher M, population, apparently present due to incomplete protease digestion, which disappeared when the sample was treated with alkaline-borohydride, and ii) the remainder of the heterogenous population of CS chains (Fig. 4B). MAb 941 immunoprecipitated 18% of the radiolabeled chains which is comprised of all the incompletely digested population of CS chains (population (i)), as well as a larger M, fraction of the heterogenous chondroitin sulfate chains (population (ii), Fig. 4B). Analysis of
n
2000
2
1500
a
U
Y
>,
c, d
2
1000
+J
0
m
0 d
U
500
O
20
25
30
35
40
45
50
5i
F r a c t i o n Number Gel Jilrrurion chromatography of CS chains. Immunoprecipitated ( 0 ) and non-immunoprecipitated chains from alkaline-borohydride treatment were analyzed by Sepharose CL-6B gel chromatography. The void volume (V,) is at fraction number 20 while total volume (V,, determined by adding free [35SO,=] to the sample) is eluted at fraction number 50. This difference in the GAG elution profile, though small, is consistently observed in at least 4 different runs.
FIGURE 3 ( 0 )CS
T. YE0 et al.
Connect Tissue Res Downloaded from informahealthcare.com by UB Giessen on 12/10/14 For personal use only.
272
FIGURE 4 SDS-gel electrophoresis of GAG chains. [35SO,=] labeled CSPG was isolated from a preparative Sepharose CL-4B column chromatography and exhaustively dialyzed against Tris buffer. The CS chains were prepared by isolation of alkaline-borohydride treatment (A) and by pronase-papain digestion (B). Upon immunoprecipitation with MAb 941,5-9% of the alkaline-borohydride derived GAG chains and 18% of the pronase-papain derived chains were immunoprecipitated. The total, immunoprecipitated (Ip) and non-immunoprecipitated (NonIp) samples were analyzed on a 7.5-12% SDS-PAGE. The fluorograph was scanned.
Connect Tissue Res Downloaded from informahealthcare.com by UB Giessen on 12/10/14 For personal use only.
CHARACTERIZATION OF CHONDROITIN SULFATE PROTEOGLYCAN
213
the autoradiograph by a gel scanner indicated that the incompletely digested fraction represented 42% of the total CS chains immunoprecipitated by MAb 941. However, when this fraction was treated with alkaline-borohydride,only 47% of the resulting [35S]-sulfate radiolabeled chains was immunoprecipitated. Thus, the incompletely digested population contained more than one GAG chain attached to the undigested small peptide and the MAb 941 actually immunoprecipitated 12% of the total radiolabeled CS chains. The results thus far suggest that there is a minor subpopulation (5-12%) of the CS chains that is larger than the rest of the CS chains and that this minor subpopulation contains the epitope for MAb 941. To test if there are any differencesin the degree of 6- versus 4-sulfation in the immunoprecipitated CS chains, the affinities for commercial preparations of CS C (90% 6-5) and CS A (80% 4-S) to MAb 941 were compared. Competition binding experiments show that MAb 941 has a 50-fold higher affinity for CS C than CS A (Fig. 5). To compare the ratio of 4-S to 6-S in the immunoprecipitated and non-immunoprecipitated CS chains, smooth muscle cells were double-labeled with [35SO,=] and [3H]-glucosamine. The double-labeled CSPG was isolated from a preparative CL-4B column and the alkaline-borohydride released CS chains were immunoprecipitated with MAb 941. The immunoprecipitated and non-immunoprecipitated CS chains were digested with chon-
(Chondroitin Sulfate] f Ml FIGURE 5 Competition of rhe immunoprecipitate with CS-C and CS-A. [35S]-methionine labeled CSPG was immunoprecipitated with MAb 941 in the presence of increasing concentrations of CS C (C-6-S)or CS A (C-4-S). The immunoprecipitate was then washed several times and the radioactivity precipitated in the Pansorbin pellet was determined. The x-axis is plotted on log scale and the y-axis represents the percentage of [35S]-methionine labeled CSPG that was immunoprecipitated when increasing amounts of cold CS chains were added during immunoprecipitation with 100%representing the total CSPG immunoprecipitated when no cold CS chains were added.
Connect Tissue Res Downloaded from informahealthcare.com by UB Giessen on 12/10/14 For personal use only.
274
T.YE0 et al.
droitin AC lyase and subsequent analysis on Bio P-10 gel chromatography showed that 90% of the radioactivity was recovered in the disaccharide elution position while 10% eluted as tetrasaccharides (data not shown). This observation is consistent with the small amount of chondroitinAC lyase resistant material when the large CSPG was digested.4 Since the CSPG peak was completely sensitive to chondroitin ABC lyase digestion (data not shown), the tetrasaccharides indicates that both non- and immunoprecipitated GAG chains contain approximately 5% iduronic acid. Analysis of the disaccharides on HPLC, indicates that the immunoprecipitated chains are composed of 63% 6-S and 37% 4-S (Fig. 6A), while the majority of the CS chains (non-immunoprecipitatedchains) have a 1:1 ratio of 6-S:4-S (Fig. 6B). These results were confirmed by treatment of the disaccharides with chondroitin-6sulfatase, which removes specifically the sulfate moiety on 6-C position of galactosamine residues,*resulting in peaks which eluted at the 0-S region (data not shown). Therefore, the epitope recognized by MAb 941 appears to be localized to a large, chondroitin-6-S rich chain of the proteoglycan. Furthermore, all the CSPG synthesized and secreted by arterial smooth muscle cells appear to contain this chain.
DISCUSSION Monoclonal antibody against bovine arterial CSPG has been used to study the large CSPG secreted by monkey smooth muscle cell cultures. Chondroitin AC lyase digestion of the immunoprecipitate consistently yielded 3-4 [35S]-methionineradiolabeled bands with M, ranging from 390-500 kd with a predominant band migrating at about 420 kd. This core glycoprotein size is similar to the core glycoprotein from CSPG derived from intact bovine aorta10 indicating that cultured arterial smooth muscle cells synthesize and secrete a class of CSPG which is similar to what is found in viv0.5 At present it is not clear why three major protein components are generated following chondroitinasedigestion of the smooth muscle cell CSPG. It is of interest that others have observed the generation of multiple protein components following chondroitinase digestion of various proteoglycansl l-13 and have attributed at least some of the differences to protein processing (for example, differences in N- vs 0-linked glycosylationl2). Treatment with N-glycanase or endoglycosidase Fl4 (two enzymes that cleave N-linked sugars) did not shift the three protein cores significantly when analyzed on SDS-PAGE (data not shown). These results indicate that these bands did not arise from differences in the N-linked oligosaccharide substitution. Previous studies have indicated that this arterial smooth muscle cell derived CSPG contains a large number of 0-linked oligosaccharides when compared to other large CSPGs such as cartilage. Therefore, these bands may be due to differences in the amounts of 0-linked oligosaccharides present on the protein cores. Digestion of the major cartilage CSPG which contains considerably fewer 0-linked oligosaccharides but proportionately more CS chains with chondroitin AC lyase produces only one protein band at around 200 kd.11 At present, it is not clear whether this large aggregatingCSPG produced by aortic smooth muscle cell is similar to the cartilage CSPG or other aggregating CSPGs such as versican which is synthesized by cultured fibroblasts.15 Antibodies against the large aggregating proteoglycan core protein from aorta do not cross react with sclera, tendon or cartilage PG.15 Aorta CSPG core protein also produces a different tryptic peptide map when compared to cartilage CSPG.16 Recent
CHARACTERIZATION OF CHONDROITIN SULFATE PROTEOGLYCAN I-
1
C-6s
A
Connect Tissue Res Downloaded from informahealthcare.com by UB Giessen on 12/10/14 For personal use only.
215
0 0
A 5
10
15
20
c-4s
25
30
35
F r a c t i o n Number
1
B 5000 0
E
a
U
>.
L) 4
2 3000 +J
u
m 0
4
U
m U IOOC
(
5
10
15
20
25
30
35
F r a c t i o n Number FIGURE 6 HPLC analysis of the immunoprecipitated CS disaccharides. [35SO,=] (filled symbols) and [3H]glucosarnine (open symbols) labeled CSPG was treated with alkaline-borohydride to release the GAG chains. The chains were then immunoprecipitated with MAb 941 and the immunoprecipitated (A) and non-immunoprecipitated (B) CS chains were treated with chondroitin AC lyase, yielding disaccharides. The disaccharides, purified over BioGel P-10 column, were analyzed by HPLC. Chondroitin-0-sulfate (C-0 S) eluted at fraction 5 , chondroitin-6sulfate (C-6s) eluted at fraction 21, and chondroitin-4-sulfate (C-4 S) eluted at fraction 29 in Figure A, while in Figure B, chondroitin-0-sulfate (C-0 S) eluted at fraction 4, chondroitin-6-sulfate (C-6s) eluted at fraction 17, and chondroitin-4-sulfate (C-4 S) eluted at fraction 24.
Connect Tissue Res Downloaded from informahealthcare.com by UB Giessen on 12/10/14 For personal use only.
216
T. YE0 er a1
Northern blot analysis has indicated that monkey arterial smooth muscle cells contain mRNA for versican (Jarvalainen and Wight, unpublished observations). The polydispersed nature of this CSPG is further demonstrated by the presence of variable sized GAG chains which run as broad bands on both molecular sieve chromatography and SDS-PAGE. This heterogenous population of CS chains is made up of at least two distinct populations which are distinguishable by: 1) their ability to be immunoprecipitatedby MAb 941, 2) their average chain size, and 3) their 6- to 4-sulfate ratio. Both the immunoprecipitable and non-immunoprecipitableCS chains contain identical proportions (5%)of iduronic acid indicating that iduronic acids do not contribute to the epitope recognized by MAb 941. The presence of iduronic acid in the large CSPG from bovine aorta has also been identified.10 Release of the chains by alkaline-borohydride and pronase-papain treatment yielded a heterogenous peak on Sepharose CL-6B chromatography analysis with a K,, = 0.39. On SDS-PAGE, the radiolabeled chains also ran as a very broad species with the pronase-papain released chains migrating as two populations, the larger of which disappeared upon alkalineborohydride treatment. Since analysis by Sepharose CL-6B chromatography is not adequately sensitive to resolve the two populations, care must be taken if papain or pronase digestion is used to release and analyze the GAG chains as there could be more than one GAG chain attached to a small peptide. MAb 941 immunoprecipitates all intact CSPG secreted by the smooth muscle cells, but upon releasing the GAG chains from the PG, only 5-12% of the CS chains were immunoprecipitated. Furthermore, since MAb 941 is able to immunoprecipitate all the peptides with more than one attached GAG chain (Fig. 4B), it is possible that the antibody recognizes an epitope which is present in the intact CSPG but only in part in the isolated GAG chains. On the other hand, since the immunoprecipitable CS chains can be segregated from the remaining GAG chains by their average length and their 6 - 9 4 4 ratio, and the fact that the CS A and CS C can quantitatively compete for all the CSPG bound to MAb 941 (Fig. 5), it is more probable that the epitope of MAb 941 dwells in this minor subpopulation of the CS chains which are present in clusters, resulting in more complete immunoprecipitationwhen more than one GAG chain is present on the peptide. A similar situation exists in bovine aorta chondroitiddermatan sulfate, where the chondroitin chains can be subdivided into two fractions: the first of which contains solely 6-sulfate and has an average M, of 49,000, while the other fraction contains 4-sulfate and dermatan sulfate in a ratio of 3: 1 on a M, 37,000 chain size.10 The previous paper7 described the use of the MAb 941 in staining the smooth muscle cells and rabbit aorta. One point to remember is that this antibody only recognizes a subpopulation of the CS chains. Unless this species of chain is present in all the CSPG in aorta (as is the case of cultured smooth muscle cells), the antibody will not stain all the CSPGs. The preferential staining of the intima>media> adventitia of the rabbit aorta could indicate that there is a preponderance of the CSPG bearing the chondroitin-6-S epitope of the MAb 941 in the intima. This possibility is supported by the results obtained using monoclonal antibodies, which only recognize chondroitin-6-S or chondroitin-4-S stubs after chondroitin ABC lyase digestion.17 This demonstrates that the intima and media portion of the aorta contained proteoglycans with chondroitin-6-S stubs whereas the adventitia contained proteoglycans with chondroitin-4-S stubs.17 These results indicate that proteoglycan which differ in their chemical characteristics may also differ in their spatial distribution in the extracellular matrix.
CHARACTERIZATION OF CHONDROITIN SULFATE PROTEOGLYCAN
277
ACKNOWLEDGMENTS The authors wish to acknowledge support from National Institutes of Health Research Grants HL-18645. Valuable assistance from M. G. Kinsella is gratefully acknowledged.
Connect Tissue Res Downloaded from informahealthcare.com by UB Giessen on 12/10/14 For personal use only.
REFERENCES 1. Wight, T. N., Lark, M. W., and Kinsella, M. G. (1987). Biology of Proteoglycans. Academic Press, Orlando, pp. 267. 2. Wight, T. N. (1985). Fed. Proc. 44, 381. 3. Wight, T. N. and Hascall, V. C. (1983). J. CeZZ Biol. 96, 167. 4. Chang, Y., Yanagishita, M., Hascall, V. C., and Wight, T. N. (1983). J. Biol. Chem. 258, 5679. 5. Salisbury, B. G. J. and Wagner, W. D. (1981). J. Biol. Chem. 256, 8050. 6. Rowe, H. A. and Wagner, W. D. (1985). Arteriosclerosis 5, 101. 7. Lark, M. W., Yeo, T-K., Mar, H., Lara, S., Hellstrom, I., Hellstrom, K. E., and Wight, T. N. (1988). J. Histochem. Cytochem. 36, 1211. 8. Saito, H., Yamagata, T., and Suzuki, S . (1968). J. Biol. Chem. 243, 1536. 9. Laemmli, U. K. (1970). Nature 227, 680. 10. Radhakrishnamurthy, B., Jeansomme, N., and Berenson, G. S. (1986). Biochem. Biophys. Aczu 882, 85. 11. Hassell, J. R., Kimura, J. H., and Hascall, V. C. (1986). Ann. Rev. Biochem. 55, 539. 12. Glossl, J., Beck, M., and Kresse, H. (1984). J. Biol. Chem. 259, 14144. 13. Sobue, M., Nakashima, N., Fukatsu, T., Nagasaka, T., Katoh, T., Odura, T., and Takeuchi, T. (1988). J. Histochem. Cytochem. 36, 479. 14. Tarentino, A. L., Gomez, C. M., and Plummer, T. H. (1985). Biochem. 24, 4665. 15. Zimmermann, D. R. and Ruoslahti, E. (1989). Embo J. 8, 2975. 16. Heinegird, D., Bjorne-Person, A., Franzen, L., Gardell, S., Malmstrom, A,, Paulsson, M., Sandfolk, R., and Vogel, K. (1985). Biochem. J. 230, 1812. 17. Couchman, J. R., Caterson, B., Christner, J. E., and Baker, J. R. (1984). Nature 307, 650.
Photocopying permitted by license only 0 1992 Gordon and Breach Science Publishers S.A. Printed in the United States of America
CHARACTERIZATION OF A CHONDROITIN SULFATE PROTEOGLYCAN SYNTHESIZED BY MONKEY ARTERIAL SMOOTH MUSCLE CELLS IN VZTRO
Connect Tissue Res Downloaded from informahealthcare.com by UB Giessen on 12/10/14 For personal use only.
TET-KIN YEO, STEPHEN MACFARLANE, and THOMAS N. WIGHT Department of Pathology SM-30, School of Medicine, University of Washington, Seattle, WA 98195 (Received June I , 1990; in revised form December 10, 1990; accepted October 16, 1991)
A monoclonal antibody against arterial smooth muscle cell chondroitin sulfate proteoglycan has been developed. Incubation of [3sS]-methionine labeled proteoglycans with MAb 941 quantitatively immunoprecipitated all the chondroitin sulfate proteoglycan (CSPG) synthesized by these cells. Digestion of the immunoprecipitate with chondroitin AC lyase revealed one major protein band (M, 420,000) and two minor bands (M, 509,000 and 390,000) on SDS-PAGE that are composed of very similar peptides when analyzed by limited peptide digestion by S. aureus V8 protease. Additional studies demonstrated that this monoclonal antibody recognized an epitope on the chondroitin sulfate chains. However, only a minor subpopulation (5-12%) of the alkaline-borohydride released glycosaminoglycan chains was immunoprecipitated and this subset of chains was slightly larger than the nonimmunoprecipitated chains. High pressure liquid chromatography analysis of the disaccharides generated from the immunoprecipitated glycosaminoglycan chains demonstrated that these chains were enriched in chondroitin-6sulfate relative to chondroitin-4-sulfate (2:l) while that of the non-immunoprecipitated chains had a ratio of I:1. These studies indicate that at least two distinct pools of chondroitin sulfate chains are present on all the chondroitin sulfate proteoglycan synthesized by arterial smooth muscle cells: a major population (89-95%) containing 6-sulfate and 4-sulfate in relatively equal proportion and a minor population (5-12%) which is hydrodynamically larger with a 6-sulfate to 4-sulfate ratio of 2: 1. KEYWORDS: antibody, chondroitin sulfate proteoglycan, smooth muscle cells, chondroitin sulfate chains
INTRODUCTION Although proteoglycans are minor components in blood vessels, a number of studies have demonstrated that these macromolecules influence a number of arterial properties such as viscoelasticity, permeability, lipid metabolism, hemostasis, and thrombosis.1 In order to fully understand the role that proteoglycans play in these processes as well as in the pathogenesis of atherosclerosis,2 it is important to structurally characterize the proteoglycans synthesized by vascular cells. Cultured monkey arterial smooth muscle cells synthesize at least three populations of proteoglycans, chondroitin sulfate proteoglycan (CSPG), dermatan sulfate proteoglycan (DSPG), and heparan sulfate proteoglycan (HSPG), of which CSPG predominates.3.4 Similar populations of proteoglycans have been described for intact aortic tissue.5fj Biochemical and ultrastructural characterizationof CSPG synthesized by monkey arterial smooth muscle cells indicates that it is a large molecule consisting of a core glycoprotein Address for correspondence: Tet-Kin Yeo, Department of Pathology, Harvard Medical School, Beth Israel Hospital, 330 Brookline Avenue, Boston, MA 02215
265
Connect Tissue Res Downloaded from informahealthcare.com by UB Giessen on 12/10/14 For personal use only.
266
T. YE0 et al.
measuring 220-320 nm in length, to which are attached 15-20 chondroitin sulfate (CS) chains, each with a M, = 43-60kd.3.4 This molecule contains a high content of 0-linked oligosaccharides, approximately 300 monomers per molecule resulting in a ratio of 15:l 0-linked oligosaccharides per glycosaminoglycan (GAG) chain. In addition, the molecule contains few N-linked oligosaccharides (20 monomers per m0lecule).3~4Although the majority of the CSPG secreted by monkey arterial smooth muscle cells is not aggregated under cell culture conditions, they are capable of forming link stabilized aggregates when supplied with hyaluronic acid.3.4 Using a monoclonal antibody (MAb 941) generated against the arterial CSPG,7 Lark et al demonstrated that this CSPG was enriched in developing atherosclerotic lesions. Therefore, the present study has been undertaken to define the epitope recognized by this antibody and to further characterize the molecular characteristics of the core glycoprotein and GAG chains of this CSPG. MATERIALS AND METHODS Materials Guanidine HCl (grade I), Tris base, EDTA, papain (Crude powder), N-ethylmaleimide and phenylmethylsufonyl fluoride (PMSF) were all purchased from Sigma Chemical Co., St. Louis MO; 6-aminohexanoicacid and benzamidine were from Eastman Kodak Co., Rochester NY; chondroitin AC lyase, ABC lyase and rabbit anti-mouse and Pansorbin were from ICN Biomedicals, Inc., Costa Mesa CA; sodium borohydride was from Wilshire Chemical Co., Gardena CA; Spectrapor dialysis tubing was from Spectrum Medical Ind., Inc., Los Angeles CA; Pronase and Triton X-100 were from Calbiochem, San Diego CA; Anti-DRC-1 was from Accurate Chem. & Scientific Corp., Westbury NY; DEAE-Sephacel and Sepharose CL-2B, and CLdB were from Pharmacia Fine Chemicals, Piscataway NJ; Biogel p-10, Acrylamide,N,N’-Methylene-bis-acrylamide,N,N,Nr,N’-Tetramethylethylenediamine and ammonium persulfate were from Biorad Laboratories, Richmond CA; Na, [35S04=](carrier-free), [35S]-methionine (1198 Ci/mmole) and D-[6-3H]-glucosamine HCl (31 Ci/mM) were from New England Nuclear, Boston MA; GelBond PAG film was from FMC Corp., Rockland ME; all cell culture supplies were from Grand Island Biological Co., Grand Island NY. All other chemicals were reagent grade. Cell Culture and Labeling Arterial smooth muscle cell cultures were established from pigtail monkeys (Macaca nernestrina) as described.3.4 Subconfluent cultures were labeled with 50 pCi/ml of [35S}methionine, or [35S04=]and 5 pCi/ml of [3H]-glucosamine for 24-48 h. Proteoglycan Isolation Upon termination of the labeling period, the medium from the cell cultures was removed and solid urea, Triton X-100, and protease inhibitors were added to give a final concentrationof 8 M urea, 0.3%Triton X-100, 2.5 mM EDTA, 5 mM benzamidine hydrochloride, 100 mM 6-aminohexanoic acid, and 1 mM PMSE4 Medium (20 ml) was passed through a 1.0 ml DEAE-Sephacel column which was pre-equilibrated with 8.0 M urea, 0.25 M NaCl buffer
CHARACTERIZATION OF CHONDROITIN SULFATE PROTEOGLYCAN
267
(8 M urea, 2 mM EDTA, 0.3% Triton X-100 and 50 mM Tris, pH 7.5). The DEAE column was washed with 20 ml of the above buffer to remove non-bound material and was eluted stepwise with 0.25 ml, 0.6 ml, 0.6 ml and a 1.0 ml of 8 M urea, 1.0 M NaCl buffer. After determining the radioactivity present in each eluted fraction, the fractions containing the radiolabel (usually fractions 2 and 3) were pooled and dialyzed.
Connect Tissue Res Downloaded from informahealthcare.com by UB Giessen on 12/10/14 For personal use only.
Immunoprecipitation Proteoglycans from the medium were first partially purified as described above and the pooled fractions exhaustively dialyzed against Ip buffer (50 mMTris-HC1, pH 7.15,150 mM NaC1, 1% Triton X-100,2.5 mM EDTA, 5 mM benzamidine hydrochloride, 100 mM 6-aminohexanoic acid, 1 mM PMSF). The antigen was pre-immunoprecipitatedwith M709 (Dako mouse monoclonal anti-DRC1) for 1 h followed by incubation with fixed and heat killed S. aureus (Pansorbin) which had been preabsorbed with rabbit anti-mouse Ig for 1 h. The sample was centrifuged in the microfuge for 45 s and the supernatant was incubated with MAb 941 ( I ml of the MAb 941 was used against 1.O X 106cpm radiolabeled antigen) for 1 h. S. uureus preabsorbed with rabbit anti-mouse Ig was added (1 h incubation on rotator) to precipitate the antigen-antibody complex. The pellet was washed five times in Ip wash buffer (Ip buffer containing 0.1% SDS and 1 mg/ml bovine serum albumin) and the immunoprecipitate was analyzed by gel electrophoresis.
Characterization of the Proteoglycan
To study the core glycoprotein of the proteoglycan, chondroitin AC lyase digestions was performed on the immunoprecipitate by resuspending the S. aureus antibody-antigen pellet in 30 p1 of 0.1 M Tris-HCL, pH 6.8, 0.03 M Na acetate, 10 mM EDTA, 10 mM N-ethylamide, 100 pg/ml bovine serum albumin and 1 mM PMSE Chondroitin AC lyase was added at 0.15 U and the incubation was allowed for 2-3 h at 37°C water bath. At the end of the incubation, SDS sample buffer was added to the mixture, boiled for 3 min and analyzed by SDS-PAGE. The GAG chains of the proteoglycan were obtained by alkaline-borohydridetreatment of an ethanol-precipitated sample by reconstituting the dried precipitate in a solution containing 1 M Na borohydride in 0.05 M NaOH and incubating the sample in an oven at 45°C for 24 h. After the reaction was complete, glacial acetic acid was used to neutralize the solution and water was added to dissolve any salt formed. The solution was dialyzed against water and precipitated by incubation with fourfold volume of ice-cold 95% ETOH, 1.3% K acetate containing 100 pg/ml heparin as a carrier for 2 h at 4°C. GAG chains were also prepared by sequential pronase and papain digestion.4 Pronase digestion was performed by incubating the proteoglycans with 0.5 mg/ml of pronase in 0.2 M Tris-HC1, pH 8 at 55-60°C overnight. The reaction was terminated by boiling the mixture for 5 min. Papain (30 pg/ml in a solution containing 0.1 M Na acetate, 5 mMNa EDTA and 5 mM cysteine, pH 7) was then added to the mixture and incubated for 4 h in a 65°C water bath. Upon termination of the digestion, the mixture was cooled and cold trichloroacetic acid was added to a final concentration of 10%. The sample containing the GAG chains was centrifuged at 10,000 g at 4°C for 30 min and the supernatant dialyzed and precipitated by ethanol as described above.
268
T. YE0 et al.
Analysis of the proteoglycan, the core glycoproteins and GAG chains were performed on SDS gel electrophoresis according to the method of Laemmli9 with upper stacking gels of 2.5% or 3% polyacrylamide and lower resolving gradient gels of 3-7.5% polyacrylamide. Gel bond PAG film was used for these gels to provide strength and ease of handling. Determination of the chondroitin-6- and 4-sulfate was carried out on the HPLC using a Whatman Partisil PAC-15 column. The chondroitin-6-sulfatase reaction was accomplished as described by Saito et al, 1968.8
Connect Tissue Res Downloaded from informahealthcare.com by UB Giessen on 12/10/14 For personal use only.
RESULTS Immunoprecipitation Smooth muscle cell cultures were radiolabeled with [35S]-methionineand the proteoglycans extracted as described in Methods. When this material was immunoprecipitated with MAb 941, approximately 2% of the total [35S]-methionine radiolabeled protein was immunoprecipitated. Analysis of the immunoprecipitate on 3-7.5% SDS-PAGE showed a large M, band which ran slightly into the resolving gel, similar to the large [35SO,']-labeled CSPG isolated from CL-2B gel chromatography (Fig. 1, lanes 2 and 4). Treatment of this immunoprecipitate with chondroitin AC lyase prior to electrophoresis consistently revealed 3-4 radiolabeled bands ranging from 390-500 kd, with the predominant band electrophoresing with M, = 420 kd (Fig. 1, lane 3). Limited peptide analysis performed on the three radiolabeled bands using S.aureus V8 protease digestion yielded several peptides, most of which are common for all the three radiolabeled core proteins (Fig. 2). Though quantitative differences are observed in the yielded peptides after V8 digestion, the three radiolabeled core proteins share several similar peptides, suggesting that they are related to one another.
Characterization of the Epitope of MAb 941 [35S]-methionine labeled proteoglycans were isolated form the medium of smooth muscle cells by ion exchange chromatography and digested with chondroitin AC lyase. The resulting material which contains radiolabeled core protein with oligosaccharide stubs of CSPG as well as other proteoglycans were incubated with MAb 941. No significant radioactivity was precipitated and no radiolabeled band was seen on the SDS-PAGE (Fig. 1, lane 5 ) , indicating that the core proteins or oligosaccharides of CSPG are not the epitope of MAb 941. To further demonstrate that the MAb 941 does not recognize the core proteins, immunoprecipitationof the [35S]-methioninelabeled proteoglycans were carried out and the S. aureus antibody-antigencomplex was incubated with chondroitin AC lyase. At the end of the digestion, the mixture was centrifuged to separate the supernatant from the pellet and the two fractions were analyzed. Ninety-seven percent of the immunoprecipitatedradioactivity was associated with the supernatant while only 3% of the radioactivity was associated with the pellet, indicating that upon digestion of the GAG chains, the radiolabeled core protein was released into the supernatant (Fig. 1, lanes 7 and 8). Thus, the antigen was not bound to the antibody via its protein core or its oligosaccharide stubs. [35S]-sulfatelabeled CSPG was prepared by preparative Sepharose CL4B chromatography and the CS chains were isolated by alkaline-borohydride treatment. Immunoprecipita-
269
Connect Tissue Res Downloaded from informahealthcare.com by UB Giessen on 12/10/14 For personal use only.
CHARACTERIZATION OF CHONDROITIN SULFATE PROTEOGLYCAN
FIGURE 1 Analysis of the imrnunoprecipitate by SDS-gel electrophoresis. DEAE concentrated [35S]-methionine labeled proteoglycans were immunoprecipitated with control Ab M709 (lane I), MAh 941 (lane 2) and the immunoprecipitate treated with chondroitin AC lyase to yield the core proteins (lane 3). [35SO,=] labeled CSPG was isolated from a Sepharose CL-2B column under dissociative conditions (4 M GuHC1),7 dialyzed exhaustively against Tris buffer, precipitated by ethanol and was analyzed on 3-7.5% SDS-PAGE (lane 4). To determine the epitope of MAh 941, the [35S]-methionine labeled proteoglycans were first treated with chondroitin AC lyase and the resulting core proteins immunoprecipitated with MAh 941 (lane 5). In another experiment, immunoprecipitated [35S]-methionine labeled CSPG (lane 6) was treated with chondroitin AC lyase when the proteoglycan was still attached to the Pansorbin pellet. The resulting mixture was microfuged and the pellet (lane 7) and the supernatant (lane 8) were analyzed on 3-7.5% SDS-PAGE. The gel was treated with Enlightning and a Ruorograph obtained. The arrow represents the start of the resolving gel, while the numbers on the right represents the M, in kd.
270
T. Y E 0 et al.
Connect Tissue Res Downloaded from informahealthcare.com by UB Giessen on 12/10/14 For personal use only.
A
B.
FIGURE 2 Limited peptide mapping of the core proteins. Immunoprecipitated CSPGs were incubated with chondroitin AC lyase and analyzed on a 4-7.5% SDS-PAGE without the use of a gel bond. The gel was fixed with acetic acid:MeOH:H,O (1.5:1:17.5) for 1 h, dried and exposed to XAR film. The three bands corresponding to the protein core bands (A) M200,OOO; (B) Mr420,000; (C) M,390,000), similar to that seen in Fig. I , lane 3, were cut and equilibrated with gel equilibration buffer (0.125 M Ris-HCI, pH 6.7, 1 mM EDTA, 1% SDS) for 1 h prior to loading the gel slices onto the 15% SDS-PAGE. S. aureus V8 protease (100 pg in 10% glycerol solution) was overlayed above the gel slices and electrophoresis was performed at 5 mA. The current was turned off for 30 min when the proteins were halfway into the stacking gel, then proceeded as usual. The gel was treated with Enlightning and the fluorograph was scanned. The numbers above indicate the relative mobility of the peptides on the gel.
Connect Tissue Res Downloaded from informahealthcare.com by UB Giessen on 12/10/14 For personal use only.
CHARACTERIZATION OF CHONDROITIN SULFATE PROTEOGLYCAN
27 1
tion of the [35S]-sulfate labeled CSPG and CS chains with MAb 941 indicated that although MAb 941 precipitated greater than 90% of the intact CSPG, only 5-9% of the alkalineborohydride derived CS chains were immunoprecipitated. These immunoprecipitated chains are larger in size as analyzed by Sepharose CL-6B gel chromatography (Fig. 3) and electrophoresed more slowly on SDS-PAGE when compared to the total heterogenous population of alkaline-borohydride released chains (Fig. 4A). However, the chains were of similar apparent charge density when chromatographed on a DEAE Sephacel ion exchange column (data not shown). To ascertain if these results were due to the chemical reactions of alkaline-borohydride treatment, thereby altering the epitope of MAb 941, CS chains were also prepared by sequential papain and pronase digestion. MAb 941 immunoprecipitated 18% of the [35S]sulfate labeled CS chains. Sepharose CL-6B chromatographyof the CS chains indicated that the immunoprecipitated chains were larger than the total population of CS chains (data not shown). Analysis by SDS-PAGE indicated that papain-pronase treatment resulted in two populations of CS chains: i) a higher M, population, apparently present due to incomplete protease digestion, which disappeared when the sample was treated with alkaline-borohydride, and ii) the remainder of the heterogenous population of CS chains (Fig. 4B). MAb 941 immunoprecipitated 18% of the radiolabeled chains which is comprised of all the incompletely digested population of CS chains (population (i)), as well as a larger M, fraction of the heterogenous chondroitin sulfate chains (population (ii), Fig. 4B). Analysis of
n
2000
2
1500
a
U
Y
>,
c, d
2
1000
+J
0
m
0 d
U
500
O
20
25
30
35
40
45
50
5i
F r a c t i o n Number Gel Jilrrurion chromatography of CS chains. Immunoprecipitated ( 0 ) and non-immunoprecipitated chains from alkaline-borohydride treatment were analyzed by Sepharose CL-6B gel chromatography. The void volume (V,) is at fraction number 20 while total volume (V,, determined by adding free [35SO,=] to the sample) is eluted at fraction number 50. This difference in the GAG elution profile, though small, is consistently observed in at least 4 different runs.
FIGURE 3 ( 0 )CS
T. YE0 et al.
Connect Tissue Res Downloaded from informahealthcare.com by UB Giessen on 12/10/14 For personal use only.
272
FIGURE 4 SDS-gel electrophoresis of GAG chains. [35SO,=] labeled CSPG was isolated from a preparative Sepharose CL-4B column chromatography and exhaustively dialyzed against Tris buffer. The CS chains were prepared by isolation of alkaline-borohydride treatment (A) and by pronase-papain digestion (B). Upon immunoprecipitation with MAb 941,5-9% of the alkaline-borohydride derived GAG chains and 18% of the pronase-papain derived chains were immunoprecipitated. The total, immunoprecipitated (Ip) and non-immunoprecipitated (NonIp) samples were analyzed on a 7.5-12% SDS-PAGE. The fluorograph was scanned.
Connect Tissue Res Downloaded from informahealthcare.com by UB Giessen on 12/10/14 For personal use only.
CHARACTERIZATION OF CHONDROITIN SULFATE PROTEOGLYCAN
213
the autoradiograph by a gel scanner indicated that the incompletely digested fraction represented 42% of the total CS chains immunoprecipitated by MAb 941. However, when this fraction was treated with alkaline-borohydride,only 47% of the resulting [35S]-sulfate radiolabeled chains was immunoprecipitated. Thus, the incompletely digested population contained more than one GAG chain attached to the undigested small peptide and the MAb 941 actually immunoprecipitated 12% of the total radiolabeled CS chains. The results thus far suggest that there is a minor subpopulation (5-12%) of the CS chains that is larger than the rest of the CS chains and that this minor subpopulation contains the epitope for MAb 941. To test if there are any differencesin the degree of 6- versus 4-sulfation in the immunoprecipitated CS chains, the affinities for commercial preparations of CS C (90% 6-5) and CS A (80% 4-S) to MAb 941 were compared. Competition binding experiments show that MAb 941 has a 50-fold higher affinity for CS C than CS A (Fig. 5). To compare the ratio of 4-S to 6-S in the immunoprecipitated and non-immunoprecipitated CS chains, smooth muscle cells were double-labeled with [35SO,=] and [3H]-glucosamine. The double-labeled CSPG was isolated from a preparative CL-4B column and the alkaline-borohydride released CS chains were immunoprecipitated with MAb 941. The immunoprecipitated and non-immunoprecipitated CS chains were digested with chon-
(Chondroitin Sulfate] f Ml FIGURE 5 Competition of rhe immunoprecipitate with CS-C and CS-A. [35S]-methionine labeled CSPG was immunoprecipitated with MAb 941 in the presence of increasing concentrations of CS C (C-6-S)or CS A (C-4-S). The immunoprecipitate was then washed several times and the radioactivity precipitated in the Pansorbin pellet was determined. The x-axis is plotted on log scale and the y-axis represents the percentage of [35S]-methionine labeled CSPG that was immunoprecipitated when increasing amounts of cold CS chains were added during immunoprecipitation with 100%representing the total CSPG immunoprecipitated when no cold CS chains were added.
Connect Tissue Res Downloaded from informahealthcare.com by UB Giessen on 12/10/14 For personal use only.
274
T.YE0 et al.
droitin AC lyase and subsequent analysis on Bio P-10 gel chromatography showed that 90% of the radioactivity was recovered in the disaccharide elution position while 10% eluted as tetrasaccharides (data not shown). This observation is consistent with the small amount of chondroitinAC lyase resistant material when the large CSPG was digested.4 Since the CSPG peak was completely sensitive to chondroitin ABC lyase digestion (data not shown), the tetrasaccharides indicates that both non- and immunoprecipitated GAG chains contain approximately 5% iduronic acid. Analysis of the disaccharides on HPLC, indicates that the immunoprecipitated chains are composed of 63% 6-S and 37% 4-S (Fig. 6A), while the majority of the CS chains (non-immunoprecipitatedchains) have a 1:1 ratio of 6-S:4-S (Fig. 6B). These results were confirmed by treatment of the disaccharides with chondroitin-6sulfatase, which removes specifically the sulfate moiety on 6-C position of galactosamine residues,*resulting in peaks which eluted at the 0-S region (data not shown). Therefore, the epitope recognized by MAb 941 appears to be localized to a large, chondroitin-6-S rich chain of the proteoglycan. Furthermore, all the CSPG synthesized and secreted by arterial smooth muscle cells appear to contain this chain.
DISCUSSION Monoclonal antibody against bovine arterial CSPG has been used to study the large CSPG secreted by monkey smooth muscle cell cultures. Chondroitin AC lyase digestion of the immunoprecipitate consistently yielded 3-4 [35S]-methionineradiolabeled bands with M, ranging from 390-500 kd with a predominant band migrating at about 420 kd. This core glycoprotein size is similar to the core glycoprotein from CSPG derived from intact bovine aorta10 indicating that cultured arterial smooth muscle cells synthesize and secrete a class of CSPG which is similar to what is found in viv0.5 At present it is not clear why three major protein components are generated following chondroitinasedigestion of the smooth muscle cell CSPG. It is of interest that others have observed the generation of multiple protein components following chondroitinase digestion of various proteoglycansl l-13 and have attributed at least some of the differences to protein processing (for example, differences in N- vs 0-linked glycosylationl2). Treatment with N-glycanase or endoglycosidase Fl4 (two enzymes that cleave N-linked sugars) did not shift the three protein cores significantly when analyzed on SDS-PAGE (data not shown). These results indicate that these bands did not arise from differences in the N-linked oligosaccharide substitution. Previous studies have indicated that this arterial smooth muscle cell derived CSPG contains a large number of 0-linked oligosaccharides when compared to other large CSPGs such as cartilage. Therefore, these bands may be due to differences in the amounts of 0-linked oligosaccharides present on the protein cores. Digestion of the major cartilage CSPG which contains considerably fewer 0-linked oligosaccharides but proportionately more CS chains with chondroitin AC lyase produces only one protein band at around 200 kd.11 At present, it is not clear whether this large aggregatingCSPG produced by aortic smooth muscle cell is similar to the cartilage CSPG or other aggregating CSPGs such as versican which is synthesized by cultured fibroblasts.15 Antibodies against the large aggregating proteoglycan core protein from aorta do not cross react with sclera, tendon or cartilage PG.15 Aorta CSPG core protein also produces a different tryptic peptide map when compared to cartilage CSPG.16 Recent
CHARACTERIZATION OF CHONDROITIN SULFATE PROTEOGLYCAN I-
1
C-6s
A
Connect Tissue Res Downloaded from informahealthcare.com by UB Giessen on 12/10/14 For personal use only.
215
0 0
A 5
10
15
20
c-4s
25
30
35
F r a c t i o n Number
1
B 5000 0
E
a
U
>.
L) 4
2 3000 +J
u
m 0
4
U
m U IOOC
(
5
10
15
20
25
30
35
F r a c t i o n Number FIGURE 6 HPLC analysis of the immunoprecipitated CS disaccharides. [35SO,=] (filled symbols) and [3H]glucosarnine (open symbols) labeled CSPG was treated with alkaline-borohydride to release the GAG chains. The chains were then immunoprecipitated with MAb 941 and the immunoprecipitated (A) and non-immunoprecipitated (B) CS chains were treated with chondroitin AC lyase, yielding disaccharides. The disaccharides, purified over BioGel P-10 column, were analyzed by HPLC. Chondroitin-0-sulfate (C-0 S) eluted at fraction 5 , chondroitin-6sulfate (C-6s) eluted at fraction 21, and chondroitin-4-sulfate (C-4 S) eluted at fraction 29 in Figure A, while in Figure B, chondroitin-0-sulfate (C-0 S) eluted at fraction 4, chondroitin-6-sulfate (C-6s) eluted at fraction 17, and chondroitin-4-sulfate (C-4 S) eluted at fraction 24.
Connect Tissue Res Downloaded from informahealthcare.com by UB Giessen on 12/10/14 For personal use only.
216
T. YE0 er a1
Northern blot analysis has indicated that monkey arterial smooth muscle cells contain mRNA for versican (Jarvalainen and Wight, unpublished observations). The polydispersed nature of this CSPG is further demonstrated by the presence of variable sized GAG chains which run as broad bands on both molecular sieve chromatography and SDS-PAGE. This heterogenous population of CS chains is made up of at least two distinct populations which are distinguishable by: 1) their ability to be immunoprecipitatedby MAb 941, 2) their average chain size, and 3) their 6- to 4-sulfate ratio. Both the immunoprecipitable and non-immunoprecipitableCS chains contain identical proportions (5%)of iduronic acid indicating that iduronic acids do not contribute to the epitope recognized by MAb 941. The presence of iduronic acid in the large CSPG from bovine aorta has also been identified.10 Release of the chains by alkaline-borohydride and pronase-papain treatment yielded a heterogenous peak on Sepharose CL-6B chromatography analysis with a K,, = 0.39. On SDS-PAGE, the radiolabeled chains also ran as a very broad species with the pronase-papain released chains migrating as two populations, the larger of which disappeared upon alkalineborohydride treatment. Since analysis by Sepharose CL-6B chromatography is not adequately sensitive to resolve the two populations, care must be taken if papain or pronase digestion is used to release and analyze the GAG chains as there could be more than one GAG chain attached to a small peptide. MAb 941 immunoprecipitates all intact CSPG secreted by the smooth muscle cells, but upon releasing the GAG chains from the PG, only 5-12% of the CS chains were immunoprecipitated. Furthermore, since MAb 941 is able to immunoprecipitate all the peptides with more than one attached GAG chain (Fig. 4B), it is possible that the antibody recognizes an epitope which is present in the intact CSPG but only in part in the isolated GAG chains. On the other hand, since the immunoprecipitable CS chains can be segregated from the remaining GAG chains by their average length and their 6 - 9 4 4 ratio, and the fact that the CS A and CS C can quantitatively compete for all the CSPG bound to MAb 941 (Fig. 5), it is more probable that the epitope of MAb 941 dwells in this minor subpopulation of the CS chains which are present in clusters, resulting in more complete immunoprecipitationwhen more than one GAG chain is present on the peptide. A similar situation exists in bovine aorta chondroitiddermatan sulfate, where the chondroitin chains can be subdivided into two fractions: the first of which contains solely 6-sulfate and has an average M, of 49,000, while the other fraction contains 4-sulfate and dermatan sulfate in a ratio of 3: 1 on a M, 37,000 chain size.10 The previous paper7 described the use of the MAb 941 in staining the smooth muscle cells and rabbit aorta. One point to remember is that this antibody only recognizes a subpopulation of the CS chains. Unless this species of chain is present in all the CSPG in aorta (as is the case of cultured smooth muscle cells), the antibody will not stain all the CSPGs. The preferential staining of the intima>media> adventitia of the rabbit aorta could indicate that there is a preponderance of the CSPG bearing the chondroitin-6-S epitope of the MAb 941 in the intima. This possibility is supported by the results obtained using monoclonal antibodies, which only recognize chondroitin-6-S or chondroitin-4-S stubs after chondroitin ABC lyase digestion.17 This demonstrates that the intima and media portion of the aorta contained proteoglycans with chondroitin-6-S stubs whereas the adventitia contained proteoglycans with chondroitin-4-S stubs.17 These results indicate that proteoglycan which differ in their chemical characteristics may also differ in their spatial distribution in the extracellular matrix.
CHARACTERIZATION OF CHONDROITIN SULFATE PROTEOGLYCAN
277
ACKNOWLEDGMENTS The authors wish to acknowledge support from National Institutes of Health Research Grants HL-18645. Valuable assistance from M. G. Kinsella is gratefully acknowledged.
Connect Tissue Res Downloaded from informahealthcare.com by UB Giessen on 12/10/14 For personal use only.
REFERENCES 1. Wight, T. N., Lark, M. W., and Kinsella, M. G. (1987). Biology of Proteoglycans. Academic Press, Orlando, pp. 267. 2. Wight, T. N. (1985). Fed. Proc. 44, 381. 3. Wight, T. N. and Hascall, V. C. (1983). J. CeZZ Biol. 96, 167. 4. Chang, Y., Yanagishita, M., Hascall, V. C., and Wight, T. N. (1983). J. Biol. Chem. 258, 5679. 5. Salisbury, B. G. J. and Wagner, W. D. (1981). J. Biol. Chem. 256, 8050. 6. Rowe, H. A. and Wagner, W. D. (1985). Arteriosclerosis 5, 101. 7. Lark, M. W., Yeo, T-K., Mar, H., Lara, S., Hellstrom, I., Hellstrom, K. E., and Wight, T. N. (1988). J. Histochem. Cytochem. 36, 1211. 8. Saito, H., Yamagata, T., and Suzuki, S . (1968). J. Biol. Chem. 243, 1536. 9. Laemmli, U. K. (1970). Nature 227, 680. 10. Radhakrishnamurthy, B., Jeansomme, N., and Berenson, G. S. (1986). Biochem. Biophys. Aczu 882, 85. 11. Hassell, J. R., Kimura, J. H., and Hascall, V. C. (1986). Ann. Rev. Biochem. 55, 539. 12. Glossl, J., Beck, M., and Kresse, H. (1984). J. Biol. Chem. 259, 14144. 13. Sobue, M., Nakashima, N., Fukatsu, T., Nagasaka, T., Katoh, T., Odura, T., and Takeuchi, T. (1988). J. Histochem. Cytochem. 36, 479. 14. Tarentino, A. L., Gomez, C. M., and Plummer, T. H. (1985). Biochem. 24, 4665. 15. Zimmermann, D. R. and Ruoslahti, E. (1989). Embo J. 8, 2975. 16. Heinegird, D., Bjorne-Person, A., Franzen, L., Gardell, S., Malmstrom, A,, Paulsson, M., Sandfolk, R., and Vogel, K. (1985). Biochem. J. 230, 1812. 17. Couchman, J. R., Caterson, B., Christner, J. E., and Baker, J. R. (1984). Nature 307, 650.
