Non-collagen protein and proteoglycan in renal glomerular basement membrane

322

Biochimica et Biophysica Acta, 678 (1981) 322-328

Elsevier/North-HollandBiomedicalPress BBA 29822 NON-COLLAGEN PROTEIN AND PROTEOGLYCAN IN RENAL GLOMERULAR BASEMENT MEMBRANE

MARGO PANUSHCOHEN, VAN-YUWU and MARIA LINDA SURMA Department of Medicine, Wayne State University School of Medicine, 540 East Confield, Detroit, M148201 (U.S.A.)

(Received May 19th, 1981) (Revised manuscript receivedAugust 18th, 1981)

Key words: Non-collagen protein; Proteoglycan; (Kidney glomerular basement membrane)

Extraction of rat glomerular basement membrane, purified by osmotic lysis and sequential detergent treatment,

with 8 M urea containing protease inhibitors solubilizes protein that is devoid of hydroxyproline and hydroxylysine. This material represents 8-12% of total membrane protein, elutes mainly as two high molecular weight peaks on agarose gel f'dtration, and is associated with glycosaminoglycans. Isolated rat renal glomeruli incorporate [ 3SS]sulfate into basement membrane from which this non-collagenous 3ss-labeled fraction can be subsequently solubilized. The radioactivity incorporated into urea-soluble glomerular basement membrane eluted primarily with the higher molecular weight peak (Mr greater than 250 000). Cellulose acetate electrophoresis after pronase digestion of the urea-soluble fraction revealed glycosaminoglyean that was resistant to digestion with Streptomyces hyaluronidase and chondroitinase ABC, sensitive to nitrous acid treatment, and contained [3Ss]sulfate. The findings indicate that one of the non-coUagenous components of glomerular basement membrane is a proteoglycan containing heparan sulfate.

(EHS sarcoma) that produces a basement membrane matrix from which macromolecular components can be isolated without proteolytic enzymes have demonstrated the presence of proteoglycan in this basement membrane [16]. Other studies with the EHS sarcoma have shown that the basement membrane of this tissue also contains a large non-collagen glycoprotein which has been designated laminin [17]. A sulfated glycoprotein (Entactin) has been purified from the extracellular basement membrane-like matrix elaborated by an embryonal carcinoma endodermal cell line [18]. Immunofluorescent techniques with antibodies to laminin, entactin, and the heparan sulfate-containing proteoglycan purified from mouse tumor suggest that many basement membranes contain these components [16-20]. Although definitive biochemical confirmation of their presence in natural basement membranes is lacking, Parthasarathy and Spiro [21] recently reported that proteolytic digests of sonicated bovine glomerular basement membrane contain a peptide-linked glycosaminoglycan consistent with a heparan sulfate proteoglycan. 0304-4165/81/0000-0000/$02.50 © 1981 Elsevier/North-HollandBiomedicalPress Basement membranes are electron dense, amorphous connective tissue matrices that lie outside the plasma membrane and separate cells from adjacent tissue. They are relatively rich in carbohydrate and contain a form of collagen which resembles procollagen [1-5]. Although early compositional studies suggested the presence of both non-collagen and collagenous proteins [ 6 - 8 ] , the former have not been as intensively studied and their characterization has been even more difficult than that of the collagen components, in part because of the relatively insoluble nature of basement membranes. Until very recently, in fact, the presence of non-collagenous components other than procollagen extension sequences has been questioned [9]. Current evidence, however, indicates that basement membranes contain unique non-collagenous moieties. For example, glycosaminoglycans have been identified in the basement membranes isolated from renal glomeruli [10-12], lens capsule [13,14] and retinal microvessels [15]. Recent studies with a transplantable mouse tumor

323 The present report describes the identification of non-collagen protein in association with heparan sulfate as normal constituents of a naturally occurring basement membrane. For these studies, we used techniques to isolate renal glomerular basement membrane that retain architectural integrity [22-25[ and preserve anionic sites [10] which contain glycosaminoglycans [11,12]. In view of the susceptibility of basement membrane collagen to limited pepsin digestion [2,26,27], we employed methods for separation and solubilization of non-collagen components that did not entail pepsin treatment. Materials and Methods Preparation of tissue and isolation of basement membranes Renal cortex was separated by gross dissection from the kidneys of adult male white rats immediately after being killed in a carbon dioxide chamber. Details of the sieving process for isolation and collection of the glomeruli were as previously described [28,29]. Basement membranes were purified by osmotic lysis, followed by selective solubilization of the cell membranes, intracellular protein and adherent plasma proteins with the detergents Triton X-100 and sodium deoxycholate, leaving behind the basement membranes which are insoluble in these reagents [25]. Details of this procedure have been reported [23]. Osmotic lysis was performed in distilled water containing protease inhibitors (25 mM EDTA, 10 mM N-ethylmaleimide, 1 mM phenylmethylsulfonylfluoride, and 1 mM benzamidine hydrochloride). The basement membranes were washed with a solution of protease inhibitors in water before lyophilization. Analytical studies Lyophilized samples of basement membrane were stirred twice for 2 h at 4°C in 0.05 M Tris-HC1, pH 8.6, containing 8 M urea and protease inhibitors in the concentrations given above. The urea-soluble material was separated by centrifugation at 15 000 × g for 30 min, and applied to an agrose gel column or dialyzed in the cold and lyophilized until further analysis. This fraction represented approx. 8-12% of total membrane protein, determined by measuring protein (using the method of Lowry et al. [46]) in

unfractionated glomerular basement membrane and in the urea-soluble and urea-insoluble fractions. Gel filtration was performed on a 2.5 X 90 cm calibrated column of Agarose A-5m (200-400 mesh) which was equilibrated and eluted with 0.05 M TrisHC1 (pH 8.6) containing 8M urea. The void volume was determined by the elution position of Blue Dextran, and the column was standardized with several radioactively-labeled proteins of known molecular weight. Fractions of 5 ml were collected and were monitored for absorbance at 280 nm and, with radioactively labeled samples, for radioactivity with a liquid scintillation counter. Polyacrylamide gel electrophoresis was performed using the method of Goldberg et al. [30] at constant current. Gels were 7.5% in acrylamide and 0.07% in methylenebisacrylamide and were polymerized in electrophoresis buffer (0.1 M phosphate buffer, pH 7.0, containing 0.1% SDS and 0.5 M urea). Lyophilized samples were suspended in a solution of 1% SDS and 0.05 M urea containing 10% glycerol, with or without 0.05 M dithiothreitol, and 0.002% bromophenol blue as the tracking dye. Gels were stained for 90 min with a solution of 0.25% Coomassie blue, 20% trichloroacetic acid, 45% methanol and 9% acetic acid, and destained with several changes of a solution of 5% methanol and 7% acetic acid. Amino acid analysis was performed on a Beckman Model l l 8 B L single column analyzer, packed with W-2 resin and eluted with a three buffer system. Samples were reconstituted in buffer for analysis after hydrolysis in 6 M HC1 for 22 h at 110°C in sealed glass ampules, followed by evaporation in vacuo. Hexosamine was measured by the Elson-Morgan method [31] after hydrolysis for 4 h at 110°C in 4 N HC1, and uronic acid was determined by the carbazole reaction [32]. For glycosaminoglycan isolation, samples were digested with pronase (1 mg/ml with an enzyme: protein ratio of approx. 1 : 5) in 0.12 M Tris-HC1, pH 7.2, containing 0.01 M CaC12 for 24h at 37°C [33,34]. Digested protein was removed by precipitation with 5% trichloroacetic acid, and excess trichloroacetic acid was extracted from the supernatant with ether. Aliquots of the aqueous samples were taken for measurement of hexosamine, uronic acid and, in biosynthesis experiments (see below), for determination

324 of radioactivity in a liquid scintillation counter. After concentration, the rest of the samples was used for glycosaminoglycan separation on cellulose acetate strips with 0.1 M calcium acetate buffer, pH 4.0 [33]. The strips were stained with 0.05% alcian blue in 50 mM sodium acetate buffer, pH 5.8, containing 50 mM MgC12 and destained in 50 mM acetate, 50 mM MgC12 [35]. Known standards (chondroitin sulfates A and C, heparan sulfate, hyaluronic acid) were run simultaneously. Duplicate samples were subjected to nitrous acid treatment or to digestion with Streptomyces hyaluronidase or chondroitinase ABC before application to the electrophoresis strips. Specificity of the enzymatic digestion was corroborated by subjecting standards to the same procedure, with appropriate controls incubated in buffer alone. Volumes were adjusted so that equal amounts of material were applied before and after enzymatic digestion or treatment with nitrous acid.

Biosyn thesis studies Isolated rat renal glomeruli were incubated for 2 h at 37°C in modified Krebs solution containing 0.15 mg/ml glutamine, 50 /.tg/ml ascorbate, 10 mM glucose, and 100/aCi/ml [3SS]sulfate (New England Nuclear, 880 mCi/mM) in an atmosphere of 95% 02/ 5% CO2. Incubations were terminated on ice with the addition of 1 mM puromycin to inhibit proteoglycan (protein) synthesis in the immediate post-incubation period. Samples were then centrifuged at 1200 Xg for 10 min to separate the media from the glomeruli, and the pellets were washed several times by suspension in Krebs buffer and recentrifugation. The washed glomeruli were immediately placed in distilled water containing protease inhibitors for isolation of the glomerular basement membranes by the method described above. Results

Table I presents the amino acid analysis of the glomerular basement membrane protein solubilized in 8 M urea. It contains neither hydroxylysine nor hydroxyproline. While some early preparations contained approx. 2 residues/1000 of hydroxylysine and/ or small amounts of hydroxyproline, careful extraction and recentrifugation eliminated virtually all hydroxylysine and hydroxyproline and hence their pres-

TABLE I AMINO ACID COMPOSITION OF UREA SOLUBLE PROTEIN IN RAT GLOMERULAR BASEMENT MEMBRANE Values are uncorrected for hydrolytic losses and represent average of four separate preparations, n.d., not detected. Amino acid

Residues/1 000

Hydroxyprolinc Aspartic Acid Threonine Serinc Glutamic Acid Proline Glycine Alanine Half-cystine Valine Methionine Isoleucine Leucine Tyrosine Phenylalanine Hydroxylysine Histidine Lysine Arginine

n.d. 88.0 54.3 91.9 126.6 64.3 129.0 79.3 14.6 49.6 11.8 28.1 70.5 24.9 30.4 n.d 27.2 42.6 51.2

ence was interpreted as reflecting contamination with small amounts of basement membrane collagen. Uronic acid and hexosamine contents in this fraction were each approx. 50 /ag/mg protein (method of Lowry et al. [46]); these values are higher than those reported by Kanwar and Farquhar [12] for digests of unfractionated basement membrane, and apparently reflect relative enrichment of the urea-soluble fraction with glycosaminoglycan [11]. The glomerular basement membrane which remains insoluble after extraction with 8 M urea retains the compositional features generally associated with basement membranes [36,371. On agarose gel filtration, the material solubilized with urea eluted primarily as two peaks in the higher molecular weight regions (Fig. 1). The second major peak was followed by a minor shoulder of protein and a small peak later in the chromatogram which were not further studied. The elution positions of known radioactively-labeled standards indicated that these peaks, which were identical with or without

325 TABLE II --a

--b

1

2

3

--d

I I I o.lo bA-i

25

B--~

45

I'

FRACTION

65

3SS-INCORPORATION INTO GLOMERULAR BASEMENT MEMBRANE In each experiment, approx. 1.6 • 106 glomeruli were incubated for 2 h with [3SS]sulfate and the basement membranes were subsequently purified. Samples were extracted with 8 M urea containing protease inhibitors to separate non-coUagen (urea-soluble) from collagen (urea-insoluble) fractions, and then each fraction was subjected to pronase digestion for isolation of the glycosaminoglycans. Uronic acid was measured by the modified carbazole reaction [32]. Experiment

Fraction

cpm

% Total cpm

% Total uronic acid

1

Urea-soluble Urea-insoluble

9 880 1 140

90.0 10.0

81.7 18.3

2.

Urea-soluble Urea-insoluble

6 680 850

89.2 10.8

71.8 29.2

!

Fig. 1. Agarose gel filtration of urea-soluble protein of rat glomerular basement membrane. Numbers indicate the elution positions of radioactively-labeled proteins of known molecular weight: 1, Filamin (Mr 240000); 2, globulin (Mr 150000); 3, phosphorylase b (Mr 92500). The column (Agarose A-5M) was 2.5 × 90 cm and was equilibrated and eluted with 0.05 M Tris-HCl, pH 8.6, containing 8 M urea. Void volume was at fraction 24 and total volume at fraction 96. The cluant was monitored for absorbance at 280 nm. 5 ml fractions were collected and peaks were pooled as indicated by the dotted lines. Insert: Gel electrophoresis (7.5% in acrylamide) of Peak B eluted from the agarose column. Migration positions of standard proteins were: (a) myosin (212 000) ; (b) fl-galactosidase (135 000); (c) phosphorylase b (92 500); (d) bovine serum albumin (69 000); (e) ovalbumin (43 000).

reduction with dithiothreitol, corresponded to molecular weights in the range of not less than 250 000 (Peak A) and approx. 120000 (Peak B). On polyacrylamide gel electrophoresis, Peak A remained as a single band just penetrating the gel, while Peak B contained four main polypeptide bands with electrophoretic mobilities corresponding to apparent molecular weights of 85000, 105000, 120000 and 140000 (Insert, Fig. 1). The electrophoretic mobilities of proteins in Peaks A and B were identical with and without reduction with dithiothreitol, suggesting that the cysteine content represented intra- rather than inter-chain disulfide bonds or that other types of crosslinks in addition to disulfide bonds prevented dissociation of Peak A into subunits and its penetration into gels in the presence of reducing agent. Isolated rat renal glomeruli incorporated [3SS]-

sulfate into the urea-soluble fraction of the subsequently purified basement membrane (Table II). Approx. 75% of the incorporated counts were recovered after pronase digestion of this fraction; approx. 90% of the total counts recovered after pronase digestion of the urea-soluble and urea-insoluble fractions of glomerular basement membrane were

8OO

0.2.

i

I I

I

0.1

~ ;

[///I

400 \

25

f--.

I

I 45 FRACTION

!

615

Fig. 2. Agarose gel filtration of urea-soluble protein from basement membrane purified after incubation of isolated glomeruli with [as S]sulfate. Column conditions are as depicted in Fig. 1. The eluant was monitored for radioactivity ( . . . . . . , cpm) and for absorbance at 280 n m ( ).

326

I I

1

2

3

Fig. 3. Cellulose acetate electrophoresis of glycosaminoglycans prepared from urea-soluble fraction of rat glomerular basement membrane. Lane 1, heparan sulfate standard; lane 2, basement membrane sample, untreated, lane 3, basement membrane sample, hyaluronidase and chondroitinase ABC digested.

present in the non-collagen (urea-soluble)material (Table II). The major peak of 3SS radioactivity incorporated into urea-soluble basement membrane coeluted with the higher molecular weight protein in Peak A on agarose gel filtration (Fig. 2). Cellulose acetate electrophoresis of pronase-digested urea-soluble material revealed glycosaminoglycan that was resistant to digestion with Streptomyces hyaluronidase and chondroitinase ABC (Fig. 3), but was sensitive to nitrous acid treatment. Identical treatment of standards of hyaluronic acid and chondroitin sulfates A and C but not heparan sulfate with the corresponding enzymes resulted in their digestion. These electrophoretic features were similar to those reported for glycosaminoglycans isolated after pronase digestion of unfractionated glomerular basement membrane [38[, and are compatible with the presence of heparan sulfate. Discussion

Although still somewhat controversial, most investigations indicate that basement membranes contain a family of collagen-like peptides that resemble procollagen in size [2-5,39,40], in the presence of

a-chain-like segments associated with non-collagenous sequences linked by disulfide bonds [2,3,9], and in their partial sensitivity to limited pepsin digestion [26,41,42]. While the existence of unique non-collagen proteins, distinct from the procollagen extension sequences, has been questioned [9[, results from several recent studies indicate that such components are indeed integral constituents of glomerular and other basement membranes [10,11,16,17]. The present study provides biochemical confirmation that rat renal glomerular basement membrane contains noncollagen protein, at least some of which represents a high molecular weight species consisting of non-collagen protein associated with heparan sulfate. The relationship between components in peaks A and B is not yet clear; while preliminary analyses suggest that they contain different proteins, a precursor-product or subunit relationship has not been excluded. Previous studies have not identified this component as such among the multiple proteins that are observed on SDS-polyacrylamide gel electrophoresis of reduced and alkylated glomerular basement membrane purified by sonic disruption. In fact, all of these components which have been analyzed, although differing markedly in their molecular weights and amino acid compositions, contain some hydroxyproline and hydroxylysine [21,43] suggest ing that they contain pieces of collagen-like molecules. While other investigators have used urea extraction of glomerular basement membrane in attempt to fractionate component polypeptides, these studies employed membrane purified following sonic disruption as the starting material and the urea-soluble fraction contained substantial hydroxyproline and hydroxylysine [6,44]. The urea-soluble extract of glomerular basement membrane purified by the method described herein, in contrast, contains no hydroxyproline or hydroxylysine and is otherwise compositionally distinct from the unfractionated membrane. We believe this finding supports the interpretation that there is less disruption of the collagenous components of glomerular basement membrane with this technique. Parthasarathy and Spiro [21] recently isolated a high molecular weight peptide-linked glycosaminoglycan after collagenase and pronase digestion of sonicated glomerular basement membrane. This material contained a collagen-like sequence which was

327 cleaved with performic acid, suggesting that it was linked by disulfide bonds to another peptide to which the glycosaminoglycan chains were attached. The high molecular weight non-collagen proteins and 3SS-containing glycosaminoglycan identified in the present report were solubilized from the collagenous fraction without reduction, but methodologic differences between the two studies make direct comparison difficult, It is likely that different sites and types of collagen-non-collagen peptide interaction are affected by denaturation in urea versus solubilization with proteolytic treatment. The present study demonstrates that rat glomerular basement membrane contains non-collagen protein associated with glycosaminoglycan, and hence represents proteoglycan. Heparan sulfate appears to be the major glycosaminoglycan species, and this material thus resembles the proteoglycans recently isolated from the EHS sarcoma [16] and bovine glomerular basement membrane [21]. In view of the contribution of heparan sulfate to glomerular basement membrane permeability [451 and the decrease in 3SSincorporation into glomerular basement membrane glycosaminoglycans in experimental diabetes [38], the role of this proteoglycan in certain diseases affecting the renal glomerular filtration barrier warrants further investigation. Acknowledgments Supported in part by NIH Grant No. AM-26435, the Kroc Foundation, and the Affiliated Internists Corporation. The secretarial skills of Ms. Joan Quinones are acknowledged with pleasure.

References 1 Kefalides, N.A. (1973) Int. Rev. Connect. Tissue Res. 6, 63-104 2 Daniels, J.R. and Chu, G.H. (1975) J. Biol. Chem. 250, 3531-3537 3 Dehm, P. and Kefalides, N.A. (1978) J. Biol. Chem. 253, 6680-6686 4 Minor, R.R., Clark, C.C., Strause, E.L., Koszalka, T.R., Brent, R.L. and Kefalides, N.A. (1976) J. Biol. Chem. 251, 1789-1794 5 Heathcote, J.G., Sear, C.H.J. and Grant, M.E. (1978) Biochem. J. 176,283-294 6 Kefalides, N.A. and Winzler, R.J. (1966) Biochem. 5, 702 713

7 Kefalides, N.A. and Denduchis, B. (1969) Biochem. 8, 4613-4621 8 Spiro, R.J. (1967) J. Biol. Chem. 242, 1915-1922 9 Kefalides, N,A. and Clark, C.C. (1979) Int. Review Cytol. 61,167-228 10 Kanwar, Y.S. and Farquhar, M.G. (1979) Proc. Natl. Acad. Sci. USA 76, 1303-1307 11 Cohen, M.P. (1980) Biochem. Biophys. Res. Commun. 92,343-348 12 Kanwar, Y.S. and Farquhar, M.G. (1979) Proc. Natl. Acad. Sci. USA 76, 4493-4497 13 Laurent, M., Ronquin, N. and Regnault, F. (1978) Interdiscipl. Topics Geront. 12, 71-79 14 Hay, E.D. and Meier, S. (1974) J. Cell Biol. 62,889-898 15 Cohen, M.P. and Ciborowski, C.J. (1981) Biochim. Biophys. Acta 674,400-406 16 Hassell, J.R., Gehron-Robey, P., Barrach, H.J., Wilezek, J., Rennard, S.I. and Martin, G.R. (1980) Proe. Natl. Acad. Sci. USA 77, 4494-4498 17 Timpl, R., Rohde, H., Gehron,-Robey, P., Rennard, S.I., Foidart, J.M. and Martin, G.R. (1979) J. Biol. Chem. 254, 9933-9937 18 Carlin, B., Jaffe, R., Bender, B. and Chung, A.E. (1981) J. Biol. Chem. 256, 5209-5214 19 Rohde, H., Wick, G. and Timpl, R. (1979) Eur. J. Biochem. 102, 195-201 20 Foidart, J.M., Bere, E.W., Gaar, M., Rennard, S.I., Gullino, M., Martin, G.R. and Katz, S.I. (1980) Lab. Invest. 42, 336-342 21 Parthasarathy, N. and Spiro, R.G. (1981) J. Biol. Chem. 256,507-513 22 Cohen, M.P. and Wu, V.Y. (1980) J. Lab. Clin. Med. 95, 282-288 23 Cohen, M.P. and Surma, M. (1980) J. Biol. Chem. 255, 1767-1770 24 Cohen, M.P., Wu, V.Y. and Dasmahapatra, A. (1980) Nephron 27,146-151 25 Carlson, E.C., Brendel, K., Hjelle, J.T. and Meezan, E. (1978) 1. Ultrastruct. Res. 62, 26-53 26 Timpl, R., Bruckner, P. and Fietzek, P. (1979) Eur. J. Biochem. 95,255-263 27 Schuppan, D., Timpl, R. and Glanville, R.W. (1980) FEBS Lett. 115,297-300 28 Cohen, M.P. and Vogt, C.A. (1975) Bioehim. Biophys. Acta 393, 78-87 29 Cohen, M.P. and Klein, C.V. (1979) J. Exp. Med. 149, 623-631 30 Goldberg, B., Epstein, E.H. and Sherr, C.J. (1972) Proc. Natl. Acad. Sci. USA 69, 3655-3659 31 Boas, N.F. (1953) J. Biol. Chem. 264, 553-563 32 Bitter, T. and Muir, H. (1962) Analyt. Biochem. 4 , 3 3 0 334 33 Linker, A. and Hovingh, P. (1972) Methods Enzymol. 28, 902-911 34 Linker, A. and Hovingh, P. (1973) Carbohyd. Res. 29, 41-62 35 Whiteman, P, (1973) Biochem, J. 131,351-357

328 36 Cohen, M.P., Urdanivia, E., Surma, M. and Ciborowski, C.J. (1981) Diabetes 30, 367-371 37 Cohen, M.P., Urdanivia, E., Surma, M. and Wu, V.Y. (1980) Bioehem. Biophys. Res. Commun. 95,765-769 38 Cohen, M.P. and Surma, M.L. (1981) J. Lab. Clin. Med. 98, 715-722 39 Cohen, M.P. and Wu, V.Y. (1981) Renal Physiol. 4 , 1 1 2 120 40 Timpl, R., Martin, G.R., Bruckner, P., Wick, G. and Wiedeman, H. (1978) Eur. J. Biochem. 84, 43-52 41 Orkin, R.W., Gehron, P., McGoodwin, E.B., Martin, G.R., Valentine, T. and Swarm, R. (1977) J. Exp. Med. 145, 204 -220

42 Tryggvason, K. and Kivirikko, K.I. (1978) Nephron 21, 230-235 43 Hudson, B.G. and Spiro, R.G. (1972) J. Biol. Chem. 247, 4239-4247 44 Hudson, B.G. and Spiro, R.G. (1972) J. Biol. Chem. 247, 4229-4238 45 Kanwar, Y.S., Linker, A. and Farquhar, M.G. (1980) J. Cell Biol. 86,688-693 46 Lowry, C.H., Rosebrough, W.J., Farr, A.L. and Randall, R.J. (1951) J. Biol. Chem.193,263-275