Chondroitin. ABC lyase, chondroitin AC lyase, and hyaluronidase (Streptococcus hyalurolyticus) were purchased from Miles. Hyaluronic acid, heparan sulfate ...
THEJOURNALO F BIOLOGICAL CHEMISTRY Vol. 256. No. 15, Issue of August IO. pp. 8050-8057. Printed in lJ S A .
1981
Isolation and Preliminary Characterizationof Proteoglycans Dissociatively Extractedfrom Human Aorta* (Received for publication, January 12, 1981, and in revised form, May 12, 1981)
Brian G. J. Salisbury$ and William D. Wagner8 From the Department of Comparative Medicine and the Arteriosclerosis Research Center, Bowman Gray School of Medicine of Wake Forest University, Winston-Salem, NorthCarolina 27103
Aortic proteoglycans(PG) were isolated from human aorta intima-media preparations with 4 M guanidine hydrochloride in the presence of protease inhibitors. The extracted PG mixture comprised 67% of the total aortic PG and was composed of65%chondroitin sulfate, 22% dermatan sulfate, 8% heparan sulfate, and 4% hyaluronate. Attempts at isolation and purification of PG monomers using isopycnic CsCl gradient centrifugation under associative and dissociative conditions resulted in appreciable loss of PG through associations with co-extracted aortic proteins. The addition of a gel chromatographic step on Sepharose CL-4B under dissociative conditions resulted in separation of PG from the majority of co-extracted proteins. In addition, the procedure resulted in a separation of the PG into a population (PG-I) eluting near the column Vo and one (PG-11) included with a K,, of 0.38. Hyaluronic acid coeluted with PG-I. The major glycosaminoglycan in PGI was chondroitin sulfate (85 to 95%). No dermatan sulfate was detected in PG-I, but this glycosaminoglycan was the predominant glycosaminoglycan in PG-I1 (50 to 70%). Heparansulfate was present in small a m o u n t s in both PG-I and PG-11. D a t a presented support the proposal of at least three species of PG monomers in the aortic wall. Chromatographic studies under dissociative and associative conditions indicated that PG comprising PG-I but not PG-I1 were capable of associations with hyaluronic acid.
Proteoglycans are polysaccharide-protein macromolecules localized predominately in theintercellularmatrix of the arterial wall. The polysaccharide moieties, the GAG,’ have been studied extensively in arterial tissue (1, 2). Those GAG identified as the major constituents include hyaluronic acid, 4- and 6-sulfate dermatan sulfate, heparan sulfate, chondroitin * This workwas supported by grants HL-25161,HL-14164, and Institutional National Research Service Award HL-07115 from the National Heart, Lung, and Blood Institute, and by an educational grant from R. J. Reynolds Industries, Inc. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Recipient of R. J. Reynolds Industries Research Fellowship in Atherosclerosis during the completion of this work. This work was in partial fulfiiment of the requirements for the degree of Doctor of Philosophy in Comparative and Experimental Pathology from the Bowman Gray School of Medicine of Wake Forest University. Present address, Department of Pathology, The New York Hospital, Cornel1 Medical Center, New York, NY 10021. § Person to whom reprint requests should be addressed. The abbreviations and symbols used are: GAG, glycosaminoglycan(s); PG, proteoglycan(s); GuHC1, guanidine hydrochloride; CPC, hexadecylpyridinium chloride; Vo,void volume; V,, total volume; p , density (g/ml); po, loading density (g/ml).
+
(3-6). Although scant information is available on the physicochemical nature of the intact PG monomers or subunits, several reports have provided a partial characterization of some aortic PG(7-13). Initial studiesof PG isolated with high concentrations of dissociating agents such as GuHCl, MgC12, and CaC12 demonstrated the presence predominately of a dermatan sulfate-chondroitin sulfate PG which had a lower buoyant density, was smaller in size, and was more polydisperse thanthechondroitinsulfatePGfrom bovine nasal cartilage (8-9). Recently, Oegema et al. (13) characterized a dermatan sulfate-chondroitin sulfate PG extracted from bovine aorta with 4.0 M GuHCl in the presence of protease inhibitors. Sedimentation equilibrium centrifugation indicated the PG preparation was homogeneous but polydisperse with a M , = 1.5 to 2.0 X lo6.The authors estimated that there were between 19 and 25 GAG chains each with a M, = 4.0 X lo4attached toa core protein of M, = 2.0 to 2.4 X lo5.Oegema et al. further demonstrated that a portion of the PG monomers associated in vitro with hyaluronate. The present reportdescribes the isolation of PG present in normal and atherosclerotic human aortas using procedures combining gel chromatography under dissociative conditions and buoyant density isopycnic gradient centrifugation. Purification of the aortic PG was initially attempted using the methods described for cartilage PG (14, 15). Comparing isolation of aortic PG using procedures originally designed for cartilage PG (14, 15), the new procedure was found more advantageous mainly in separating artery PG monomers into discrete populations. The characteristics of these PG populations constitute thebulk of the present report. MATERIALS AND METHODS
Tissue Collection and Preparation Human thoracic and abdominal aortas were obtained from the Autopsy Service of the Department of Pathology of the North Carolina Baptist Hospital. Aortic samples were not used from individuals who had documented clinical histories of any connective tissue disorder, chronic hypertension, chronic obstructive pulmonary disease, or diabetes mellitus. Aortas were transported to thelaboratory on ice, and the adventitia and outer media were removed using a common plane of dissection through the media. Areas of normal tissue, fatty streak, and raised plaque (fatty and fibrous) were dissected and each normal or lesion type was pooled for a given aorta. No differentiation was made between tissue derived from thoracic or abdominal aortic segments. Complicated plaques with variable degrees of hemorrhage, ulceration, or calcification were not used. The preparations of intima-inner media were minced into pieces (2 X 3 mm), weighed and frozen in distilled water at -20 “C until extraction. Throughout the procedure the samples were kept on ice as much as possible. In one study, ear cartilage was obtained from two normal adult male rabbits. The perichondrium was removed and the cartilage was minced into pieces (2 X 3 mm). Cartilage and aortic tissues were extracted and treated identically.
8050
Human Aortic Proteoglycans Chemicals Ultrapure GuHCl was purchased from Schwarz/Mann and used without further purification as the PG extraction solvent. GuHCl (98%min) for chromatographic purposes was obtained from Eastman. Alcian Blue (98% min), m-hydroxydiphenyl, benzamidine-HCI, 6aminohexanoic acid, and CPC were also obtained from Eastman. The CPC was purified twice by recrystallization from acetone-water. Papain (twice crystallized), D-glucuronic acid (grade I), and bovine serum albumin (fraction V) were obtained from Sigma. Chondroitin ABC lyase, chondroitin AC lyase, and hyaluronidase (Streptococcus hyalurolyticus) were purchased from Miles. Hyaluronic acid, heparan sulfate, dermatan sulfate, chondroitin 4-sulfate, and chondroitin 6sulfate were NIH standards obtained from Drs. M. B. Mathews and J. A. Cifonelli of the University of Chicago (Contract NOl-AM-52205 from the National Institute of Arthritis, Metabolism, and Digestive Diseases, National Institutes of Health). All other chemicals and solvents used were reagent grade and obtained from Fisher. Analytical Procedures Quantification of uronic acids was carried out using the procedure of Blumenkrantz and Asboe-Hansen (16) with D-glucuronic acid as standard. Protein was measured according to themethod of Lowry et al. (17) with bovine serum albumin as standard. Absorbance at 280 nm was measured and used to estimate the protein distribution in the gel chromatographic and density gradient fractions. Densities were determined by pycnometry. Electrophoresis-Identification and quantification of the GAG were accomplished using a cellulose acetate electrophoretic procedure as detailed previously (18).The procedure involves the use of different buffer systems in two separate runs. Hyaluronic acid and heparan sulfate were separated using 0.3M cadmium acetate, pH 4.1, while 0.3 M calcium acetate, pH 10.0, was used to separate dermatan sulfate and chondroitin sulfate. Chondroitin 4-sulfate and chondroitin 6sulfate were not separated and are reported as chondroitin sulfate without regard to the isomeric form. Following electrophoresis, the cellulose acetate strips were stained with 0.5% Alcian blue, decolorized, cleared, and dried as described previously (18).The dried strips were scanned a t 615 nm in a Gelman ACD-18 densitometer. The method of Saito et al. (19) was used for combinations of selective enzymatic digestion in order to c o n f m the identity of the individual arterial GAG. Isopycnic CsCl Centrifugation-Isopycnic (gradient) centrifugation was carried out at 100,OOO X g (r,,,.,) at 15 "C, for 50 h in a Beckman preparative ultracentrifuge and Beckman type 60Ti or type 65 rotors. Gel Chromatography-Samples to be chromatographed were concentrated using an Amicon ultrafiltration cell and PM-30 Diaflo membranes. Columns were calibrated using Blue Dextran 2000 (Pharmacia) as a marker for the Voand L-tryptophan as a marker for the V!. Associative chromatography was carried out on a column (1.4 X 90 cm) of Bio-Gel A-50m (Bio-Rad), eluted with 0.5 M sodium acetate, pH 5.8, at a flow rate of10 d / h . Fractions of approximately 2.5 ml were collected. Dissociative chromatography was done using a column (1.4 X 90 or 4.4 X 60 cm) of Sepharose CL-2B or CL-4B (Pharmacia), eluted with 4.0 M GuHC.1 in 0.05 M sodium acetate,pH 5.8. The smaller column was eluted at a flow rate of 10 ml/h, thelarger column at 25 ml/h. Alkali Degradationof PG-PG in GuHCl were obtained following dissociative chromatography on Sepharose CL-4B, divided equally, and dialyzed against cold running tap water overnight. Following dialysis, one sample was brought to final concentrations of 0.5 M potassium hydroxide and 0.02 M sodium borohydride. Both alkalitreated and non-alkali-treated samples were incubated for 7 days at 4 "C. Glacial acetic acid was added and the samples dialyzed overnight, then brought to a final concentration of 4.0 M GuHCI. After the samples equilibrated under dissociative conditions for 24 h, each sample was chromatographed on a column of Sepharose CL-2B using 4.0 M GuHCl (0.05 M sodium acetate, pH 5.8) as the eluting buffer. As a reference point, 5.0 mg of chondroitin 6-sulfate was dissolved in 5.0 ml of 4.0 M GuHCl and chromatographed. Extraction Procedure-Tissues were thawed rapidly, blotted, and extracted in 15 volumes (v/w) of solvent a t 4 "C for 72 h under constant agitation. The extraction solvent included 4.0 M GuHCl in 0.05 M sodium acetate, pH 4.5, with three protease inhibitors: 0.01 M disodium EDTA, 0.1 M 6-aminohexanoic acid, and 0.005 M benzamidine hydrochloride. Following extraction, the PG-containing extract
8051
was decanted and the tissue residue rinsed twice with fresh extraction solvent. The extract and rinses were combined and filtered through Whatman No. 4 filter paper. The PG remaining in the tissue following extraction were referred to as extraction-resistant PG. The GAGof the extraction-resistant PGwere isolated following papain digestion of the tissue residue according to a procedure described previously (18) and outlined in Fig. 1.Initial attempts atisolating and purifying the extracted human aortic PG were carried out using the procedures originally described for cartilage PG (14, 15). This two-step procedure involved sequential associative and dissociative density gradient centrifugations of the PG extract. Subsequent to this attempt theprocedure outlined in Fig. 1 was employed in order to isolate and purify the extracted PG free of co-extracted proteins. RESULTS
PG Extraction-Intima-media minces of three grossly normalhumanaortas wereused todeterminetheoptimium extraction time for PG. Following extraction for 1,2,3, 5, 7, or 9 days, the uronate concentrations of the extractedPC andof the extracted tissue were determined. The majority (61%) of the PGwere extracted in 24 h. Moderate increases were seen after an additional 2 days, but no further increase occurred beyond 3 days. All subsequent extractions were carried out for 3 days to ensurea maximum yield of PG. The percentage of total tissue PG extracted for nine samples of aorta (from six individuals) was 66.6 f 1.7% (mean 2 S.E.).Three samples of aorta with predominantly fatty streak lesions were included in the calculation of this averagesince the extractionefficiencies (68.6 2 3.1%) were the same as those of grossly normal aorta (66.0 f 2.2%). The average concentration of extracted uronic acid for the same nine samples was 856 & 56 pg of uronic acid/g wet aorta. The GAG compositions of the extracted PG and extractionresistant PG for four samples of aorta are shown in Table I. Four GAG wereidentifiedin the PG extract. Chondroitin sulfate was the predominant GAG and comprised about 65R of the extractedGAG. This represented approximately 70% of the total aortic chondroitin sulfate based upon the extraction efficiency. Hyaluronic acid, heparan sulfate, and dermatan sulfate were also present in the PG extract in approximate proportions of 4%, 8%,and 22%, respectively. No hyaluronic acid or dermatan sulfate was detected in the extraction-resistINTIMA-MEDIA PREPARATIONS
I
I
I
I Residue delipidated E dehydrated with acetonelethanol
density gradient
Of
each PC population
Digertmn with
lxliation of GAG by
FIG. 1. Flow diagram of PG extraction and isolation designed to incorporate dissociative gel chromatography and dissociative density gradient centrifugation.Procedure for the isolation of GuHCl extraction-resistant PG (as GAG)isshownfor tissue residue.
Human Aortic Proteoglycans
5052
TABLE I GAG composition of the P G extracted with 4.0 M GuHCl a n d the P G resistant to extraction Each individual GAG is expressedas a percentage of the totalGAG following cellulose acetate electrophoresis and densitometric scanning. Sample 281 prepared from 26-year-old male Caucasian; postmortem interval 12.5 h. Sample 287 prepared from 21-year-old male Caucasian; postmortem interval 15.0 h. Samples
HyaluHeparan Dermatan Chondroironic acid sulfate sulfate tin sulfate
Extracted PG 2 10 28 60 281-Normal aorta 3 67 23 281-Fatty streak8 678 20 6 287-Normal aorta 21 66 6 7 287-Fatty streak Extraction-resistant PG ND 70 281-Normal aorta ND" 30 ND 72 ND 28 281-Fatty streak ND 60 ND 40 287-Normal aorta ND 66 ND 34 287-Fatty streak 1.735 ND = not detected. Hyaluronic acid and dermatan sulfatewould have been detected if they had been present a t a level of a1.621 t least 5% of the total sampleGAG.
associative and dissociative density gradient procedureswere not effective in isolating the majority of the extracted aortic PG, inlarge part due to their complexation with co-extracted proteins and perhaps lipids. To circumvent thiscomplexation, the extracted P C were isolated by a procedure that included agarose gel chromatographyunder dissociativeconditions. The elutionprofile of the PG extract aoncolumnof Sepharose CL-2B is shown in Fig. 2a. A single, broad asymmetric peak TABLEI1 Comparison of aortic a n d cartilage P G isolated accordingto Sajdera and Hascall(14, 15) Human aortic
PG
Rabbit ear cartilage PG
Fraction" Density
Uronic acidb
Density
Uronic acid
%
A1 A2 A3 A4
%
Associative density gradient centrifugation 1.767 38 1.766 24 1.719 13 1.693 1.673 1.630 25
(PO
1.69) 80 8 4 8
Dissociative density gradient centrifugation
(PO
1.50)
67 1.575 95 AlDl 1.604 ant PG. Chondroitin sulfatewas the predominantGAG (60 to 1.533 14 2 1.552 72%) present in the extraction-resistant PG. In these samples, A1D2 5 1.503 1 A1D3 1.518 heparansulfate,the only other GAG resistanttoGuHCl 1 6 1.455 A1D4 1.474 extraction, accountedfor approximately 62% ofthe total aortic A1D5 0 1 1.427 1.447 heparan sulfate. No major differences were observed in the 1 7 1.409 1.421 A1D6 GAG compositions of normal aorta and fatty streak. Notation according to HeinegCd (21). Fraction A1 corresponded Initial Studiesof P G Isolation a n d Purification-GuHC1 to approximately 24%of the extracted uronic acid and 14% of the extracts of normal aortic tissue and rabbit ear cartilagewere totalaortic uronic acid. Aortic sampleprepared from 35-year-old dialyzed to associative conditions and subjected toisopycnic Negro; postmortem interval 10 h. "Gelatinous floating pellicles were present in bothaorticand CsCl centrifugation (po 1.69). Dialysis of the aortic GuHCl extracts toassociative conditions resultedin the formationof cartilage associative density gradients but were not chemically analyzed. a white flocculent precipitate as has been observed by others (10, 12, 20). The uronicacid distributions of the resulting a densitygradientsare shownin Table 11. Floatinggummy pellicles were present in both gradients but were not analyzed ^o 35; 25 for uronate. Only 63% of the total startinguronic acid for the a aortic sample was recovered in the soluble portions of the gradient compared to81%for cartilage. The uronic acid pres15ent in the A1 fraction was estimated to representonly 24% of a IO the extracted uronate. Comparison of the uronic acid distri0 5bution of entire gradient indicated thata large percentage of $ 3 0the aortic PGwas more buoyant than the cartilage PG (Table 11). "a 0.5 "t Kav In an attempt toprovide more information on the buoyant densities of monomers inA l , this fraction from both aorta and cartilagewas broughtto dissociative conditions (4.0 M b. - 0.9 GuHCl), a loading density of 1.50 g/ml, and was subjected to Xr i -0.8 a second isopycnic CsCl centrifugation step. The resulting - 0.7 uronic acid distributions of these density gradients are shown - 0.6 in Table 11. For cartilage PG approximately 95% of the total uronic acid contained in the gradient (i.e. fraction A l ) was - 0.5 W found in the bottom fraction( p 1.57), whereas only 67% of the 0.4 y uronic acid was found in the corresponding fraction ( p 1.60) - 0.3 ; for the aortic PG. We used four additional samples of aorta - 0.2 a8 (two normal and two fatty streak) to examine the uronate - 0.1m $ I 5 a association with the free floating pellicle. Following associative density gradient centrifugation( p 1.69), 51, 55, 56, and 58% of 0.0 0 the gradient uronic acid was found in the floating pellicle of VO 0.5 vt each gradient. Although there was insufficient GAG to allow Kav quantification, the uronate containing material was identified FIG. 2. Dissociative gel chromatography of PG-containing as hyaluronic acid, heparan sulfate, dermatan sulfate, and 4.0 M GuHCl extracts. Extracts were concentrated by ultrafdtration chondroitin sulfate. Protein and cholesterol were also identi- and chromatographed oncolumns (1.4 X 90 cm) of Sepharose CL-2B fied as major constituentsof the pellicle. (a) or Sepharose CL-4B ( b ) using 4.0 M GuHCl in 0.05 M acetate, pH P G Isolation a n d Purification Incorporating Gel Chro5.8, as eluant. Roman numerals in b identify two major PG populamatography-The preceding results indicated that sequentialtions, PG-I and PG-11.
I
g
I I
-'
-
'9 L
Proteoglycans
Aortic
Human
8053
of uronate-containing material was observed. Chromatogra- yses of individual fractions from cesium chloride gradients phy of the extract on a Sepharose CL-4Bcolumn under indicated that with increasing buoyant densities the percentdissociative conditions gave an elution profiie that is illus- age ofGAG increased for both PG-I and PG-I1 (Table 111).At trated in Fig. 2b. The extracted PGwere separated under any given fraction PG-I1 was characterized by having lower dissociative conditions into two populations, based upon their GAG/protein ratios in comparison to PG-I. hydrodynamic size. One population (PG-I) eluted near the VO of the column (Kav0.06 f 0.01, N = 13), while the smaller - 1.50 population (PG-11) was included on the column (Kay0.38 f 250 0.01, N = 13). Separation of the extracted PG into two 1.45 populations constituted a reproducible finding in all of the aortic samples examined, regardless of whether the samples I 40 were normal aorta or aorta with fatty streaks. Under the a -I. conditions employed, we were not able to determine any major -4 loo: G differences in PG sizeor PG-I/PG-I1 ratios in normal ( N = 6 ) c, 7 5 2 50and atherosclerotic aortas ( N = 7). 25- I 30 All PG samples consistently displayed three peaks of ma0terial that absorbed at 280 nm (Fig. 2b). The K,, values of 1 2 3 4 5 6 7 8 9 1 0 each of these three peaks (Kav:0.02 f 0.01, 0.64 f 0.01, 0.95 FRACTION NUMBER f 0.01; N = 8) were consistent among the various types of samples examined. On the basis of the uronate and A280 elution profiles, the column procedure was useful in separating the - 1.50 majority of the co-extracted aortic proteins from the PG. 250 rIn another study PG populations were isolated from normal, fatty streak, and plaque tissues from several aortas. PG-I and PG-I1 populations from Sepharose CL-4B were subjected to dissociative density gradient centrifugation ( p o 1.40 g/ml) to 11.40 ," further purify monomers. the PG Representative density gra- 2 . dient profiles for both PG populations are shown in Fig. 3. 1.35 2 c, 75The majority (87 to 100%)of the loaded uronate was recovered s 50b in the soluble gradient fractions. All samples of aorta displayed 2 25- 1.30 similar uronate profiles. For PG-I, 75% f 1%(mean f S.E., N 3 0 -, , , , , I , , , L = 7) of the total gradient uronic acid was present in the I 2 3 4 5 6 7 8 9 1 0 FRACTION NUMBER bottom half of the density gradient ( p 2 1.41) while 59 f 2% of the wonatefor PG-II was found in a density range. FIG. 3. Dissociative density gradient centrifugation of PG from normal human aorta. PG-I ( a ) and PG-I1 ( b ) Less than lo%Of the acid was found in the top gradient populations were following dissociative gel chromatography on Sephafraction of both Populations (PG-1: 6 f 1%; PG-11: 9 f 1%). CL-4B. Fractions indicated by the bars in ~ i2 were ~ , concenSmall amounts of translucent amber Pellicleswere found trated, brought to a loading density of 1.40 g/ml, and centrifuged at floating on the gradients of both populations. Chemical anal- 100,000 x g at 15 "C for 50 h.
g
I
I
I
I
I
I
L
TABLEI11 Chemical composition of cesium chloride gradient fractions of PG-I and PG-II PG-I Fraction P
!#
019' Normal aorta
1 2
3 4 5 6 7 8 9 10 11 1 2 3 4 5 6 7 8 9 10 11 12
1.485 1.470 1.449 1.437 1.422 1.410 1.390 1.376 1.367 1.357 1.341 1.493 1.482 1.457 1.440 1.422 1.414 1.407 1.394 1.376 1.369 1.364
PG-I1
Uronic acid
540 304 195 149 106 97 89 75 64 78 125 344 180 106 85 62 48 42 75 32 24 19 110
M 509 346 268 252 224 227 45 217 222 231 364 -3,600 408 226 154 142 37 139 120 114 150 171 65 174 254 -6,000
w
B
75 72
25 28 32 37 42
68
3.03 2.50 2.08 1.69 1.36 1.21 1.17 0.96 0.79 0.61 0.10 2.41 2.27 1.96 1.71 1.27 1.13 1.05 1.43 0.53 0.39 0.22 0.05
P
Uronic acid
Protein
M
wg
B
B
1.480 1.471 1.450 1.422 1.413 1.401 1.387 1.377 1.363 1.351 1.342 1.489 1.476 1.455 1.446 1.432 1.411 1.407 1.388 1.377 1.365 1.360 1.339
237 145 125 107 95 86 79 71 52 50 98 329 237 157 133 128 114 123 83 89 91 95 163
408 277 257 280 278 292 363 413 781 54 878 -24,000 381 353 299 340 368 405 443 594 869 1253 3391 -43,000
62 60 58 52 49 46 38 33 22 14 1 71 66 60 53 50 44 44 29 23 17 8 1
38 40 42 48 51 54 62 67
63 58 55 54 46 49 51 44 56 38 62 9 91 019 Plaque 71 29 70 30 66 34 63 56 44 53 47 51 49 59 41 35 28 72 18 82 1.343 5 95 a GAG content was estimated by dividing uronic acid by 0.35. Sample 019 was prepared from 52-year-old male Caucasian; postmortem interval 7.5 h.
~
GAG"
Protein
86 99 29 34 40 47 50 56 56 71 77 83 92 99
1.66 1.49 1.39 1.09 0.97 0.84 0.63 0.49 0.28 0.16 0.01 2.46 1.92 1.50 1.11 0.99 0.80 0.87 0.40 0.29 0.21 0.08 0.01
Aortic
8054
Proteoglycans
Human
Proteoglycan Characterization-The existence of a polysaccharide-protein structure was demonstrated by subjecting the PG population to alkali cleavage. The elution profiies of PG-I andPG-11, chromatographed on Sepharose CL-2B under dissociative conditions, are shown in Fig. 4.Both the untreated (control)andalkali-treated samples are depicted. The K,, values for PG-I andPG-I1 control samples were 0.33 and 0.59, respectively. Both populations displayed small peaks of excluded uronate-containingmaterial which probably represented aggregated PG not disaggregated following addition of GuHCl to the dialyzed samples. The K,, values of alkali treated PG-I and PG-I1 were 0.78 and 0.75, respectively, which was comparable to the K,, of 0.75 observed for a commercial preparation of chondroitin 6-sulfate. Each of the PG populations from normal aorta, fattystreak, and plaque tissue was examined with regard toits GAG composition. In these studies, we pooled fractions of PG-I or PG-I1 from Sepharose CL-4B and isolated and identified the GAG present. Hyaluronic acid was found primarily with the larger PG population(PG-I) (TableIV) indicating that human aorta hyaluronate is of high molecular weight. The major GAG present in PG-I was chondroitin sulfate (85 to 95%). Heparan sulfate was present in lesser proportions. No dermatan sulfate was detected in the PG-Ipopulation. However, dermatan sulfate was the predominant GAG present in PG-I1 25
t
I"
\
CI
3 a GY
l5
$
IO
s
PG-I 307-Normal aorta Fatty streak Plaque PG-I1 307-Normal aorta Fatty streak Plaque9
Hyaluronic acid %
Heparan sulfate
11 10
14 9 9
14
I
I
1
I
0.5
VO
Vi
Hyaluronic acid
20
-
3
15
-
t
\
a 2 Y
10-
3
5 -
59 50 65
Heparan sulfate
Ln
3 10 24 15 12
2 3 01 Column Position: K"" : -0.030.01 0.07 0.
1i
KO,
b.
75 81 77
TABLEV GAG composition of PG-I and PG-II a t selectedpoints of the elution profile Values are expressed as percentage of the total GAG present in the aliquot taken at the given point. Sample prepared from 24-year-old male Caucasian; postmortem interval 6.5 h.
-i"
I
ND" ND ND
24 15 2 2 35 13 22 4 a ND = not detected. See Table I.
5
0
Dermatan Chondroitin sulfate sulfate % %
%
9 PG-I (left shoulder) 8 PG-I (peak) ND PG-1/11 (trough) ND PG-I1 (peak) ND PG-I1 (right shoulder) a ND = not detected. See Table I.
a.
20
TABLEIV GAG composition of PG-I and PG-II isolated from normal and atherosclerotic aortic tissue Values are expressed as percentage of the total GAG present in the population. Sample 307 prepared from 50-year-old male Caucasian: postmortem interval 24 h.
Dermatan Chondroisulfate tin sulfate
ND" ND 12 34 34
88 82 64 51 54
5 6 7 0.16 0.22 0.28 C
0
GAG HS DS
cs
0.04 I 0.00 2
g
3 4 5 6 0.03 0.20 0.27 0.40
FIG. 5. Sulfated GAG compositions of selected PG at various elution positions following dissociative gel chromatography on Sepharose CL-4B. The PG of (a) normal tissue and ( b ) athero-
0 I
I
1
0.5
VO
I
I
I
I
vt
KO,
FIG. 4. Alkali degradation of normal human aortic PG populations isolated following dissociative gel chromatography on Sepharose CL-4B. PG-I (a) and PG-I1 ( b ) were subjected to alkali degradation using 0.5 M KOH and 0.02 M NaBH4, at 4 "C for 7 days. Following incubation, nondegraded (control) populations (-) and treated populations (- - -) were chromatographed on a column of Sepharose CL-2B using 4.0M GuHCl in 0.05 M acetate, pH 5.8, as eluant. The arrows indicate the elution position of a commercial preparation of chondroitin 6-sulfate (Kav0.74).
sclerotic plaque of the same aorta were analyzed for GAG composition at the indicated Kavvalues. Abbreviations: HS, heparan sulfate; DS, dermatan sulfate; CS, chondroitin sulfate. Hyaluronic acid was present in fractions 1,2, and 3for normal aorta and fraction 1 for plaque. Sample prepared from a 52-year-old male Caucasian; postmortem interval 7.5 h.
and comprised between 50 to 70% of the PG-associated GAG. Both heparan sulfate and chondroitin sulfate were also present in PG-I1 in amounts lower than dermatan sulfate. As we later determined from other samples, the small amount of hyaluronic acid detected in PG-I1 was due to contamination from PG-I since we sampled broad cuts of PG-I and PG-I1
Human Aortic Proteoglycans
8055
Preliminary studies were carried out to determine whether from the Sepharose CL-4B columns (see Table V). In two human aortic PG shared this property. A sample of PG-I, additionalaorticsamples, we examined the PG-associated purified by dissociative density GAG distribution across the Sepharose CL-4B column run isolated from normal aorta and (Fig. 5). The K,, values are provided for a comparison of PG gradient centrifugation, was chromatographed on columns of Sepharose CL-2B. When chromatographed underdissociative from the normal aortic PG and the PG extracted from plaque conditions a single broad peak with a Kakof 0.36 was present tissue fromthesameaorta.Inbothsamplestherelative proportion of chondroitin sulfate decreased steadily with de- whereas chromatography under associativeconditions procreasing PG size (increasing Kay).Heparan sulfate increased duced two peaks of uronic acid (Fig. 6). The predominant peak (Kav0.39) was similarto thatobserved in the dissociative slightly, followed by a decrease as the PG became smaller while the proportion of the sulfated GAG that was dermatan sample. However, an additional peak was present in the V,, and accounted for approximately 20%of the eluted uronic sulfate increased significantly with decreasing PG size. The elution profiles also suggestthat atherosclerotic plaque acid. In another sample of PG from adifferent sample of may contain dermatan sulfate in PG of larger size than the normal aorta, approximately 17% of the PG-I aliquot was PG of normal aortaas suggested by the presence of dermatan found in the Vounder associative conditions. The PG-I1 preparation displayed K,, values of 0.54 and 0.53 sulfate at position 3 (KaV0.09) intheplaquesamplebut absence at positions 3 and 4 (Kav0.07 and 0.11) in the normal under dissociative and associativeconditions,respectively (Fig. 7). No uronic acid was detected in the V,, in either case. aorta sample. Association with hyaluronic acid has been shown to be a The additionof exogenous hyaluronic acid (from human umcharacteristic property of cartilage PG monomers (22, 23). bilical cord and having a K,, of 0.35 on Sepharose CL-2B) to samples of PG-I1 at a concentration of 0.7% (w/w on the basis of uronic acid) did not alter theelution profiles.
P 2ot
DISCUSSION
The classical cartilage PG isolation and purification procedures are based on the ability of a large proportion of the PG monomers to aggregate specifically with hyaluronic acid in a reversible manner (14, 15). The results of the present study have indicated someproblems in the use of this procedurefor arteries. First, using 4.0 M GuHC1, there are large amounts of Kav FIG. 6. Aggregation of PG-I without the addition of exoge- extraneous proteins, largely collagen and actin (24), co-exnous hyaluronic acid. Purified PG-I was obtained following disso- tracted with the PG. As demonstrated in the present report, ciative density gradient centrifugation. The sample was divided when the GuHClis removed following dialysis, these proteins equally and chromatographed on columns of Sepharose CL-2B. One aggregate with the PC, thus modifying their buoyant density aliquot (-) was eluted under dissociative conditions with 4.0 M and effecting subsequent losses following centrifugation. GuHClin 0.05 M sodium acetate, pH 5.8, while the other aliquot Whether this loss was a general loss or was specific for an (- - -) was eluted under associative conditions using 0.5 M sodium be a acetate, pH 5.8. PG were prepared froma 24-year-oldmale Caucasian; individual PG monomer was not determined. There may minimum level of GuHCl which can be maintained to avoid postmortem interval, 6.5 h. association of extraneous proteins with the PG still but obtain a. association of the PG with hyaluronic acid (20). Even if this \ optimum level of GuHCl canbe determined, itis possible that h 10 the dermatan sulfate-PG may not be isolated totally. In the present study, we demonstrated that PG-I1in the presence of exogenous hyaluronate did not form anaggregate as was seen b for PG-I. In other studies in our laboratory, we have deter0 0 mined that dermatan sulfate containing PG (PG-11) isolated 3 I I I I I from pigeon aortas do not aggregate with hyaluronic acid.' I I I I 0.5 Those investigators who have isolated dermatan sulfate conVO Vt taining PG by using the procedureof Sajdera andHascall (14, 15) may have been successful as a result of an association of b. the PG species with hyaluronic acid through a link glycoprotein which probably was removed duricg our procedure. We have not examined whether the association of PG-I1with hyaluronic acid requires any link glycoproteins but these are not requisites for cartilage PG association with hyaluronate (25). The two-step dissociative purification procedure described was developed in order to circumvent interactions between 0.5 vt the extracted PG and extraneous co-extracted aortic proteins. The initial step in this procedure, dissociative chromatograCL-4B resulted in the FIG. 7. Associative and dissociative chromatography of PG- phy of the GuHCl extract on Sepharose I1 without the addition of exogenous hyaluronic acid. Purified isolation of two populations of aorticPG. A second step, PG-I1 was obtained following dissociative density gradient centrifu- isopycnic centrifugation under dissociative conditions, is necgation. The sample was divided equally and both aliquots were essary to furtherpurify each PG population,especially PG-11. chromatographed on columns of Sepharose CL-2B. One aliquot (a) The sulfated GAG compositions of PG-I and PG-I1 were was eluted under dissociative conditions with 4.0 M GuHCl in 0.5 M different and also provided evidence to suggest that these PG
; s
sodium acetate, pH 5.8, while the other aliquot ( b ) was eluted under associative conditions using 0.5 M sodium acetate, pH 5.8.
'W. D. Wagner, personal communication
8056
Aortic
Human
Proteoglycans
populations were each heterogeneous. The data presented teinshavebeenpostulatedtoaccount for the extractionsupport a proposal fora t least three, and possibly four, species resistant PG (11). Radhakrishnamurthy et al. ( l l ) ,by examof PG monomers in the aortic wall. We demonstrated that the ining the PG solubilized from bovine aorta with collagenase major PG monomer species of human aorta is a chondroitin and elastase treatments, have postulated that dermatan sulsulfate-containing PG that can be separated on the basis of fate-chondroitin sulfate PGexisted in a specific binding with size from the PG monomer containing dermatan sulfate. Sev- collagen. Heparan sulfate PG were assumed to be associated eral investigators (7-13) have reported characteristics of a with elasticfibers. The data presented here indicated that the dermatan sulfate-chondroitin sulfate PG from bovine aorta. It extraction-resistant PGwere enriched in heparan sulfate and is possible that these authors(7-13) have.reported ona mixed chondroitin sulfate. Whether these extraction-resistant PG population of monomers thatwe have separated and described monomers containing heparan sulfate and chondroitin sulfate as PG-I and PG-11. are structurally and chemically distinct from those extracted In our study, we found dermatan sulfate to be a major remains to be determined. The postmortem interval for collection of the aortictissues constituent of the smaller PGpopulation. In this PGpopulation it remains to be proven if dermatan sulfate exists as a used in this study was kept as short as possible to control for single class of PG or wasa dermatansulfate-chondroitin any effect of postmortem autolysis on size the and composition sulfate hybrid PG. It does appear that chondroitin sulfate of the extracted PG. In other studies, significant no effects on exists primarily as a single PG species, at least in PG-I since human artery GAG composition have been reported as a we found only chondroitin sulfate in the very early fractions result of prolonged postmortem intervals of up to 90 h (28). from Sepharose CL-4B. Whether the chondroitin sulfate in Furthermore, Berberian and Fowler (29) provided evidence PG-I1 is as a separate PG monomer remains betodetermined. that a 48-h postmortem interval, with storageat 4 "C, had no Aortic heparan sulfate hasbeen identified to exist as a PG significant effecton thespecific activities of several lysosomal (13,26).There maybe two different species of heparan sulfate enzymes in human aortictissue. Pearson and Mason (30) concluded that prefreezing of boPG, one associated with cell surfaces (27) and one associated with elastic fibers (11). In the present report we have shown vine nasal cartilage at -20 "C for 2 h denatured some protethe presence in artery of a very polydisperse heparan sulfate olytic enzymes that apparently produced smaller degraded PG under control (nonfrozen) conditions. Finally, Oegema et PG, which is present in both PG-I and PG-I1populations. Direct comparisons of the size of human aortic PG with al. (13) indicated that extractionof bovine aortic PG without those reported for bovine aortic PG is difficult in view of the protease inhibitors resulted in PG with lower sedimentation various methods and conditions employed. However, on the coefficients. Although, in the current study, we were unable basis of gel chromatographic elution profiles, it appears that to determine any modifications in PG size as influenced by both PG populations of human aorta were larger than those length of postmortem interval,we included protease inhibitors at 4 "C to reduce the possibility reported by Ehrlich et al. (10) and McMurtreyet al. (12).PG- and carried out the extractions I did appear tobe similar insize to the PG preparation isolated of any proteolysis of the PGmonomers. by Antonopoulos et al. (8). Both PG-I and PG-I1were analogous in size to a portion of the heterogeneous PG preparation REFERENCES described by Oegema et al. (13). 1. Berenson, G. S., Radhakrishnamurthy, B., Dalferes, E. R., Jr., The ability of bovine aortic PG to associate or aggregate and Srinivasan, S. R. (1971) Human Pathol. 2, 57-79 with hyaluronic acid has been described previously (12, 13). 2. Robinson, R. W., Likar, I. N., and Likar, L. J. (1975) in MonoIn addition, Gardell et al. (20) have reported the presence of graphs on Atherosclerosis (Kirk,J. E., Kritchevsky, D., Pollak, antigenic determinants in bovine aortic extracts that reacted 0. J., and Simms, H. S., eds) Vol. 5, pp. 1-134, S.Karger, Base1 with antibodies to the link glycoproteins and the hyaluronic 3. Stevens, R. L., Colombo, M., Gonzales, J. J., Hollander, W., and acid binding regionsof cartilage PC. Results from the present Schmid, K. (1976) J.Clin. Znuest. 58,470-481 4. Kaplan, D., and Meyer, K. (1960) Proc. SOC.Exp. Biol. Med. 105, 20% of PG-I were capable of studyindicatedthatabout 78-81 forming a larger PG species under associative conditions. This 5. Kruger, C., and Teller, W. M. (1975) 2. Kinderheilkd 119, 253phenomenon was presumed to represent the aggregation of 259 PG with the endogenously present hyaluronic acid. The per6. Murata, K., Harada, T., andOkubo,K. (1968) J. Atheroscler. centage of PG capable of aggregationwassimilar to that Res. 8,951-958 reported for bovine aortic PGby Oegema et al. (13). Whereas 7. Kresse, H., Heidel, H., and Buddecke, E. (1971) Eur. J.Biochem. 22,557-562 70 to 80% of most cartilage PG are capable of aggregating with hyaluronic acid, the proportionof extracted human aortic PG 8. Antonopoulos, C. A., Axelsson, I., Heinegird, D., and Gardell, S. (1974) Biochim. Biophy~.Acta 338, 108-119 with this property appears to be less. The inability of some 9. Eisenstein, R., Larsson, S.-E., Kuettner, K. E., Sorgente, N., and aortic PG, specifically PG-11, to aggregate may very well be Hascall, V. C. (1975) Atherosclerosis 22, 1-17 related to their functional roles orinterrelationshipswith 10. Ehrlich, K. C., Radhakrishnamurthy, B., and Berenson, G. S. collagen and elastic fibers in arterial tissue. (1975) Arch. Biochem. Biophys. 171,361-369 The average extractionefficiency of 69% for human aortic 11. Radhakrishnamurthy, B., Ruiz, H. A,, Jr., and Berenson, G. S. (1977) J.Biol. Chem. 252,4831-4841 PG was less than that reported for most cartilage PG (80 to 85%), but the majority of PG were isolated effectively from 12. McMurtrey, J., Radhakrishnamurthy, B., Dalferes, E.R., Jr., Berenson, G. S., and Glegory, J. D. (1979) J.Biol. Chem. 254, the tissue. The PGyields (34 to 87%)that have been reported 1621-1626 for bovine aortic PG using 4.0 M GuHCl have been quite 13. Oegema, T. R., Jr., Hascall, V. C., and Eisenstein, R. (1979) J. variable. Although such factors as choice of extractant, tissue Biol. Chem. 254, 1312-1318 preparation, andinclusion of protease inhibitors are important 14. Sajdera, S.W., and Hascall, V. C. (1969) J.Biol. Chem. 244, 7787 points of consideration, no single factor appears to account for the diversity of extraction efficiencies (8-13). In the present 15. Hascall, V. C., and Sajdera, S. W. (1969) J. Biol. Chem. 244, 2384-2396 study, approximately 30% of the total aortic PG content was 16. Blumenkrantz, N., and Asboe-Hansen, G. (1973) Anal. Biochem. not extracted with 4.0 M GuHC1. Reasons such as simple 54,484-489 physical entrapment by the extracellular fibrous network or 17. Lowry, 0. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951) J.Biol. Chem. 193, 265-275 specific complexation between PG and the fibrous aortic pro-
Human Aortic Proteoglycans 18. Wagner, W. D., and Salisbury, B. G . J . (1978) Lab. Invest. 39, 322-328 19. Saito, H., Yamagata, T., and Suzuki, S. (1968) J . Biol. Chem. 243, 1536-1542 20. Gardell, S., Baker, J., Caterson, B., Heinegird, D., and Roden, L. (1980) Biochem. Biophys. Res. Cornmun. 95, 1823-1831 21. Heinegird, D. (1972) Biochirn. Biophys. Acta 285, 181-192 22. Hardingham, T. E., and Muir, H. (1972) Biochim. Biophys. Acta 279,401-405 23. Hascall, V. C. (1977) J . Supramol. Struct. 7, 101-120 24. Bach, P. R., and Bentley, J. P. (1980) Connect. Tissue Res. 7 , 185-196
8057
25. Swann, D. A., Powell, S., Broadhurst, J., Sordillo, E., and Sotman, S. (1976) Biochem. J. 157, 503-506 26. Jansson, L., and Lindahl, U.(1970) Biochem. J . 117,699-702 27. Rosenberg, L. (1975) in Dynamics of Connective Tissue Macromolecules (Burleigh, P. M. C., and Poole, A. R., eds) pp. 105124, North-Holland, Amsterdam 28. Manley, G . (1965) Br. J. Exp. Pathol. 46, 125-134 29. Berberian, P. A., and Fowler, S. (1979) Exp. Mol. Pathol. 30, 2740 30. Pearson, J. P., and Mason, R. M. (1977) Biochim. Biophys. Acta 498, 176-188
