Heterogeneity of heparan sulfate proteoglycans synthesized by PYS-2 cells

ARCHIVES OF BIOCHEMISTRY AND BIOPHYSICS Vol. 231, No. 2, June, pp. 328-335, 1984

Heterogeneity of Heparan Sulfate Proteoglycans Synthesized by PYS-2 Cells BERNADETTE TYREE,**p’ ELIZABETH A. HORIGAN,? DONALD L. KLIPPENSTEIN,? AND JOHN R. HA&SELL? *Howard

University Cancer Center, Washington, D. C. 20060, and ~Laboratm-y Biology and Anoma&?s, National Institute of Dental Research, National Institutes of Health, Bethesda, Marykznd 20205 Received September

22, 1983, and in revised form January

of Dmelcqrnwntal

30, 1984

Antibodies to the basement membrane proteoglycan produced by the EHS tumor were used to immunoprecipitate [35S]sulfate-labeled protoglycans produced by PYS-2 cells. The immunoprecipitated proteoglycans were subsequently fractionated by CsCl density gradient centrifugation and Sepharose CL-4B chromatography. The culture medium contained a low-density proteoglycan eluting from Sepharose CL-4B at K,, = 0.18, containing heparan sulfate side chains of M, = 3540,000. The medium also contained a high-density proteoglycan eluting from Sepharose CL-4B at K,, = 0.23, containing heparan sulfate side chains of M, = 30,000. The corresponding proteoglycans of the cell layer were all smaller than those in the medium. Since the antibodies used to precipitate those proteoglycans were directed against the protein core, this suggests that these proteoglycans share common antigenic features, and may be derived from a common precursor which undergoes modification by the removal of protein segments and a portion of each heparan sulfate chain. Numerous investigators have reported the existence of high- and low-density forms of heparan sulfate proteoglycans in cells and tissues. These proteoglycans have been shown to be associated with cell surfaces (l-6), basement membranes (7-12), and various extracellular matrices (4, 5, 13-16). Recently, Kjellen et al have isolated two heparan sulfate proteoglycans from rat liver with similar polyanionic properties, but which differ in buoyant density (3). The high-density form of the proteoglycan was associated with cell surface receptors, whereas the larger, low-density proteoglycan was suggested to be embedded in the lipid bilayer. This low-density proteoglycan could be converted to a smaller, high-density proteoglycan by treatment with detergents, suggesting that the low-density form is a mixture of pro1 To whom correspondence Howard University Cancer D. C. 20060.

should be addressed: Center, Washington,

0003-9861/84 $3.00 Copyright All rights

0 1984 by Academic Press, Inc. of reproduction in any form reserved.

328

teoglycan and lipid. Similar phenomena have been observed in various cultured cells by Norling et al. (4). Low-density heparan sulfate proteoglycans which do not contain lipid have also been isolated from cultured fibroblasts by Carlstedt et al. (13) and from the medium of cultured vascular endothelial cells by Oohira et aZ. (12). These low-density proteoglycans have been reported to be larger and to contain a higher ratio of protein to carbohydrate than the high-density proteoglycan produced by the same cells. Here, we have characterized the heparan sulfate proteoglycans synthesized by mouse PYS (parietal yolk sac carcinoma) cells, a cell line which we show synthesizes low- and high-density heparan sulfate proteoglycans. Antibodies against the protein core of a high-density proteoglycan isolated from a basement membrane tumor (EHS) immunoprecipitated both high- and low-density heparan sulfate proteoglycans. This suggests that high- and low-density

HEPARAN

SULFATE

PROTEOGLYCANS

heparan sulfate proteoglycans contain similar regions in their protein core and may be genetically related. MATERIALS

AND

METHODS

Cell culture and labeling conditiuns. PYS-2 cells were obtained from Dr. Andrew Kraft, National Cancer Institute, Bethesda, Maryland, designated PYS-2 (K); Dr. Ilse Oberbaumer, Max Plank Institute fur Biochemie, Martinsried, West Germany, designated PYS-2 (0); and Dr. John Lehman, University of Colorado, Denver, Colorado. PYS-1 cells were obtained from Dr. Leo Furcht, University of Minnesota Medical School, Minneapolis, Minnesota. Cells were maintained in Dulbecco’s modified Eagle’s medium containing 10% fetal bovine serum, 6 mg% glutamine, 200 U/ml penicillin, and 200 pg/ml streptomycin in an humidified atmosphere of 95% air and 5% COa. Cells were fed every 2 days and passaged twice per week. Confluent cultures in 35-mm dishes were labeled at 37°C in 1 ml of medium containing 76 pg/ml ascorbic acid and 100 pCi/ml Na, ?S04 (Amersham). E&-action procedures. After 18-24 h of incubation the culture medium was removed, the cell layer was washed with Iphosphate-buffered saline, and the wash was added to the medium. The medium was centrifuged to remove any floating cells and debris, and the pellet was discarded. The cell layer and medium were then processed separately. The cell layer was extracted at 4°C for 1 :h with 1.0 ml 5% Zwittergent 312 (Calbiochem) containing 0.05 M sodium acetate, 0.01 M EDTA, 0.005 M benzamidine-HCl, and 0.1 M B-aminohexanoic a.cid to inhibit proteolysis (17). An additional 1.0 ml of extraction solution containing 8 M guanidine-HCl was added and the extraction was continued for 1 h. This procedure extracted greater than 99% of the incorporated [?SJsulfate from the cell layer. Sol.id guanidine-HCI was added to the culture medium to 4 M and protease inhibitors were added to the concentrations mentioned above. Unincorporated [“Slsulfate was removed by chromatographing the samples on PD-10 Sephadex G-25 columns (Pharmacia). Radioactivity was measured in a Beckman LS-233 liquid scintillation counter using ACS scintillation cocktail (Amersham). Preparation of antisera, A high-density heparan sulfate proteoglycan was isolated from the EHS tumor, and antibodies were raised in rabbits by procedures described elsewhere (9). To determine the specificity of the antiserum, samples of the proteoglycan were pretreated with reagents which specifically degrade protein or glycosaminoglycan, and then centrifuged in gradients of CsCl. The location of the antigen in the gradient fractions was determined by an enzyme-linked immunoassay (18).

FROM

PYS-2

329

CELLS

Immunoprecipitation of proteogl~cans. A 20- to lOO-pl aliquot of culture medium or cell layer extract was added to 2.0 ml of Tris-buffered saline (TBS)r containing 0.3% Nonidet P-40 (NP40). followed by 100 pl of antiserum. Nonimmune serum was used as a control. The mixture was incubated for 30 min at room temperature. Protein A (IgG sorb, The Enzyme Center) was added to samples and incubated an additional 2 h at room temperature. The mixture was centrifuged at 5000g for 5 min. The pellet was washed three times with TBS containing 0.3% NP-40, with centrifugation after each wash. The washes were combined with the supernatant fraction. The immunoprecipitated material was obtained by extraction of the pellet with 0.5 ml of 0.01 M phosphate buffer, pH 7.0, containing 2% sodium dodecyl sulfate and 15 mg/ml dithiothreitol in a boiling-water bath for 3 min. The mix was then centrifuged at 5000g for 5 min, and the supernatant fraction was counted. Molecular sieve chromatography. Chromatography on Sepharose CL-4B or Sepharose CL-6B (Pharmacia) (1 X 150 cm) columns was performed in 4 M guanidineHCl in 0.02 M Tris-HCI, pH 7.0. In some cases 0.5% Triton X-100 was included in the solution. Molecular weight markers for Sepharose CL-4B included cartilage proteoglycan obtained from chick sterna labeled with [35S]sulfate, high-density heparan sulfate proteoglycan isolated from the EHS tumor (9), rhesus monkey cornea1 chondroitin/dermatan sulfate proteoglycan (19), and rhesus monkey cornea1 keratan sulfate proteoglycan (19). Molecular weight markers for Sepharose CL-6B included the glycosaminoglycans prepared from these proteoglycans by papain digestion of the protein cores. Density gradient eentrijkgation in CsCI Samples were adjusted to contain either 0.5 g CsCl/g 4 M guanidine-HCl or 1.0 g CsCl/g 0.5 M guanidine-HCl and transferred to Beckman Quick Seal high-speed tubes. Centrifugation was performed in a Beckman Model L5-50 ultracentrifuge using a 50 Ti rotor at 10°C for 72 h at 100,OOOg. Degradative procedures. Chondroitinase ABC digestion (20) and nitrous acid degradation (21) were performed as described elsewhere. /3 Elimination of O-linked oligosaccharides was performed in 0.05 M NaOH containing 1 M NaBH, at 45’C for 24 h. Papain digestion was performed for 18 h at 55°C in 1 M sodium acetate, pH 6.5, containing 5 X 1O-3 M cysteine and 5 X 10m3M EDTA. The reaction was stopped by the addition of a fivefold molar excess of iodoacetamide. Papain and chondrotinase ABC were purchased from Sigma. Heparitinase isolated from Flavobacbium heparinum (22) was a generous gift from Dr. Alfred Linker, Veterans Hospital, Salt Lake City, Utah. Heparitinase digestion was performed in 0.02 M Tris-

a Abbreviation

used: TBS, Tris-buffered

saline.

330

TYREE

ET AL.

HCl, pH 7.0, containing 1 X 10m6M CaClz at 30°C for 1 h. The proportion degraded was determined by chromatography on columns of Sephadex G-50 (Pharmacia) (1 X 50 cm) eluted with 0.05 M Tris-HCl, pH ‘7.0.

114250 111250 11250 l/50

RESULTS

The specificity of the antiserum was determined by subjecting the high-density heparan sulfate proteoglycan from the EHS tumor to chemical and enzymatic treatments, centrifuging the samples in gradients of CsCl, and then measuring immunoreactivity. The gradient was divided into four equal segments, and each was analyzed for the presence of immunoreactive material. As shown in Fig. la the intact proteoglycan localized to the bottom of the gradient. This profile was virtually unchanged by treatment to the proteoglycan with chondrointinase ABC (Fig. lb). Treatment of the proteoglycan with crude heparitinase (Fig. lc), however, caused the immunoreactive material to localize in the top of the gradient, suggesting that the antigen was the protein core freed of heparan sulfate by these treatments. Nitrous acid (Fig. Id) also released the protein core but reduced the amount of immunoreactive material. Treatment with papain (Fig. le) completely abolished the immunoreactivity of the material. Finally, treatment of the proteoglycan with dilute base (Fig. lf) resulted in immunoreactive material being present in all fractions of the gradient. This distribution was probably due to the incomplete release of heparan sulfate side chains, giving rise to protein cores containing varying amounts of these side chains. These results strongly suggest that the antibody is directed against the protein core of the heparan sulfate proteoglycan. We next tested by immunoprecipitation whether the high-density proteoglycan was produced by several mouse cell lines (Table I). Cultures of epidermal cells, limb bud chondrocytes, fibroblasts, and lens epithelial cells produced only a low proportion of immunoreactive proteoglycan. It is possible that these cells were producing antigenically unrelated proteoglycans. The PYS lines tested, however, produced substantially higher proportions of immunoreactive proteoglycans, most of which were

l/IO b 114250 111250 11250 1150 5

l/IO

(3 5 Q

11250

8

1150

$F

1110

3 E

d l/50 l/IO I,,

,

e 1150 l/IO

1150 l/IO 4

3

TOP

FRACTION

2

1 BOTTOM

NUMBER

FIG. 1. Sensitivity of the antigenic sites in a heparan sulfate proteaglycan to various treatments. Aliquots of purified EHS tumor high-density heparan sulfate proteoglycan were pretreated as indicated, followed by centrifugation in 0.5 g CsCl/g 4 M guanidine-HCI at 100,000~. Tubes were sliced into four equal fractions and assayed for reactivity with antiserum. Tube No. 1 = top, tube No. 4 = bottom of gradient. (a) Untreated control sample; (b) chondroitinase ABC treated; (c) crude heparitinase treated; (d) nitrous acid treated; (e) papain treated, (f) NaOH/NaBHI treated.

found in the culture medium. The greatest percentage of immunoreactive proteoglycans were produced by the PYS-2 cell line

HEPARAN TABLE

SULFATE

PROTEOGLYCANS

I

IMMLINOPRECIPITATION OF PROTEOGLYCANS Percentage of [~S]sulfak precipitated by antibody to the high-density proteoglycan

Cell lines Primary cell lines Mouse epidermal cells Mouse limb blud chondrocytes Established cell lines Mouse fibrobkts Mouse lens epithelial cells PYS-1 PYS-2 (K) PYS-2 (0)

Cell layer

Culture medium

0 0

0 N.D.

2 3 22 28 27

2 2 53 70 35

Note. Values are expressed as a percentage of the total incorporated pS)sulfate immunopreeipitated by the antibody corrected for nonspecific precipitation with nonimmune serum. N.D., Not determined.

(K), and thLese proteoglycans were subsequently isolated and characterized. Medium

FROM

PYS-2

CELLS

331

(Fig. 3a). The elution profiles of the proteoglycans were unchanged if 0.5% Triton X-100 was included in the elution buffer, suggesting the absence of lipid-mediated aggregation in these preparations. The size of glycosaminoglycans present in these two proteoglycans were estimated by molecular sieve chromatography (Fig. 4). The glycosaminoglycans from the high-density proteoglycan were smaller (Mr = 30,000) than the glycosaminoglycans from the lowdensity proteoglycan (Mr = 35-40,000). The glycosaminoglycan sizes were considerably smaller than those of the EHS high-density proteoglycan (Mr = 65,000), but were within the molecular weight range reported for other heparan sulfate glycosaminoglycans (12, 14). Approximately half of the glycosaminoglycan from the high-density proteoglycans were sensitive to nitrous acid, and nearly all of the glycosaminoglycan from the low-density proteoglycans were degraded by nitrous acid, indicating that half of the high-density and all of the low-density proteoglycans contain sulfate. The nitrous acid-resistant material in the high-density pool was chondroitin/dermatan sulfate glycosaminoglycans based

Proteoglycans

The immunoprecipitated proteoglycans from the culture medium were initially characterized by CsCl density gradient centrifugation (1 g CsCl/g 0.5 M guanidineHCl). These proteoglycans resolved into high-density (al.54 g/ml) and low-density (G1.32 g/ml) pools in approximately equal proportions (Fig. 2). In a separate study the medium was first centrifuged in CsCl to separate the high- and low-density proteoglycans, and each fraction was tested for immunoreactivity. Both low- and highdensity proteoglycans were found to react with the antibody. This suggests that the two proteoglycans share common structural features in their protein cores. The high- and low-density pools were chromatographed on Sepharose CL-4B. Most of the high-density immunoreactive pool was similar in size to the heparan sulfate proteoglycan produced by the EHS tumor, K,, = 0.23 (Fig. 3b), while the low-density material eluted somewhat earlier, Kay = 0.18

TOP

BOTTOM

FRACTION

NUMBER

FIG. 2. Separation of high- and low-density forms of heparan sulfate proteoglycans from the medium. Culture medium was immunoprecipitated with antiserum to the EHS tumor high-density heparan sulfate proteoglycan and protein A. The hound fraction was centrifuged in 1 g CsCl/g 0.5 M guanidine-HCl at 100,OOOg.The tubes were sliced into five equal segments, and their radioactivities were determined.

332

15

-

10

: 0

-

5

2 E 2p 40 2 9 "

30 20 10

3 L1 A-

TYREE

ET AL.

density than the high-density [?S]sulfate (Fig. 5b), and, as a result, the distribution of antibody-precipitable proteoglycans more closely resembled the distribution seen in the medium (compare Figs. 5a and 2a). The [35S]sulfate-labeled macromolecules not recognized by the antibody distributed in CsCl gradients as depicted in Fig. 5~. Molecular sieve chromatography on Sepharose CL-4B of the predominant cell layer fractions is shown in Fig. 6. Two size classes of immunoreactive proteoglycans (K,” = 0.37 and 0.61), were observed from the high-density pool (Fig. 6a), both eluting

),

& 0.2 0.4 0.6 0.8 ( I I I I

II

Ill

FIG. 3. Molecular sieve chromatography of proteoglycans immunoprecipitated from culture medium on Sepharose CL-4B. The top (low-density) and bottom (high-density) fractions from the CsCl gradients were chromatographed on a 1 X K&cm column of Sepharose CL-4B. The column was eluted with 4 M guanidine-HCl in 0.02 M Tris, pH 7.0, and 3.5-ml fractions were collected. Pools were made as indicated. The molecular weight markers were I, EHS tumor highdensity heparan sulfate proteoglycan; II, rhesus monkey cornea1 chondroitin/dermatan sulfate proteoglycan; and III, rhesus monkey cornea1 keratan sulfate proteoglycan. (a) Low-density proteoglycan; (b) highdensity proteoglycan.

on chondroitinase ABC digestion. The nonimmunoreactive proteoglycans were similar in size and composition to those precipitated by the antibody. It is not clear even under conditions of antibody excess why precipitation was not complete; it is possible that the nonprecipitated forms were partially degraded and lacked some antigenic determinates. Cell Layer Proteoglycans Approximately 30% of the proteoglycans associated with the cell layer were immunoreactive. Compared to the medium, a smaller proportion of the [35S]sulfate of the cell layer extract was of the low-density type (Fig. 5a). The antibody, however, precipitated a larger proportion of the low-

FIG. 4. Molecular sieve chromatography of glycosaminoglycans derived from high- and low-density proteoglycans on Sepharose CL-6B. Glycosaminoglycans prepared from the proteoglycan pools by @elimination with 0.1 M NaOH and 0.05 M sodium borohydride were chromatographed on a 1 X 150-cm column of Sepharose CL-6B. The column was eluted with 4 M guanidine-HCl in 0.02 M Tris-HCl, pH 7.0, and 3.5-ml fractions were collected. The molecular weight markers were 1, heparan sulfate prepared from the EHS tumor high-density proteoglycan; 2, chondroitin sulfate prepared from chick sternal cartilage proteoglycan; 3, chondroitin sulfate prepared from rhesus cornea1 chondroitin sulfate proteoglycan; and 4, keratan sulfate prepared from rhesus cornea1 keratan sulfate proteoglycan. (a) Glycosaminoglycan of lowdensity proteoglycan; (b) glycosaminoglycan of highdensity proteoglycan.

HEPARAN

SULFATE

PROTEOGLYCANS

FROM

PYS-2

333

CELLS

reactive proteoglycans. This fraction may represent free glycosaminogl$can or glycosaminoglycan attached to an antigenitally unrelated protein core. All but a few percent of the remaining glycosaminoglycans in all of the high- and low-density proteoglycans were sensitive to chondroitinase ABC. DISCUSSION

Heparan sulfate proteoglycans have been shown to be an integral part of the filtration barrier of basement membranes

2

a

1

)L 4 I-

b

-3 :

,-

0 i2 0 ? d 21

BOTTOM

TOP FRACTION

NUMBER

FIG. 5. Separ.ation of high- and low-density

heparan sulfate proteoglycans from the cell layer. The cell layer extract was immunoprecipitated and centrifuged in CsCl as in Fig. 2. (a) Unfractionated extract; (b) immunoprecipitated fraction; (c) fraction which did not bind to antibody.

later than the EHS high-density proteoglycan. The higher molecular weight material contained glycosaminoglycan chains of which 70% were sensitive to nitrous acid, and approximately 40% of the glycosaminoglycans i:n the later-eluting peak were sensitive to nitrous acid, indicating that there was heparan sulfate in both fractions. The immunoreactive low-density pool was composed of two peaks, K,, = 0.25 and 0.40 (Fig. 6b). The glycosaminoglycan contained in the high-density fraction not recognized by the antibody (Fig. 6~) were about 50% nitrous acid sensitive and smaller in size than any of the immuno-

ui I

6

c

4

2

t VO

I 0.2

, 0.4 I I

I II

1 0.6 I Ill

I 0.6

t “1

FIG. 6. Molecular sieve chromatography of the major proteoglycans extracted from the cell layer on Sepharose CL-4B. The high- and low-density fractions from the C&l gradients of the extract were chromatographed on Sepharose CL-4B as described in Fig. 3. (a) Immunoprecipitated low-density proteoglycans; (b) immunoprecipitated high-density proteoglycans; (c) proteoglycan which did not bind to antibody (high density).

334

TYREE

(7), and may be altered in disease states affecting basement membranes such as diabetes mellitus (23). We have shown that a cell line known to synthesize authentic basement membrane components, PYS-2 (24-28), produces both high- and low-density forms of heparan sulfate proteoglycans. It is interesting that PYS-2 cell lines maintained in different laboratories differ in their capacity to produce heparan sulfate proteoglycans (see Table I). PYS-2 cells have been known to produce heparan sulfate proteoglycans (ll), but not such a diverse array of high- and low-density forms. It is possible that this cell line is unstable and that current stocks do not resemble the cell line originally isolated (24). The proteoglycans we have isolated show immunological cross-reactivity with an heparan sulfate proteoglycan previously isolated from the EHS tumor (9). This suggests that high- and low-density proteoglycans contain common antigenic features in their protein cores and may be genetically or biosynthetically related. Most of the [35S]sulfate-labeled material produced by cultured PYS cells is in the form of heparan sulfate proteoglycans. Utilizing CsCl density gradient centrifugation and molecular sieve chromatography, these proteoglycans were separable into discreet classes. The cell layer contained high-density (al.54 g/ml) components which eluted from Sepharose CL-48 at Kay = 0.71, 0.61, and 0.37. The largest two of these, Kay = 0.37 and 0.61, were precipitated by the antibody (Fig. 6b). Glycosaminoglycan analysis indicated that the earlier peak was about 70% heparan sulfate, and that the latter peak was about 40% heparan sulfate (determined by nitrous acid sensitivity), while the remainder of the [35S]sulfate was in chondroitin/dermatan sulfate. Since the antibody does not precipitate the chondroitin sulfate proteoglycan produced by cartilage (see Table I; limb bud chondrocytes), it is unlikely that the PYS-2 cells are synthesizing proteoglycans characteristic of cartilage. It is possible that PYS cells produce heparan sulfate and chondroitin sulfate proteoglycans on antigenically similar proteincores, but distinct from the cartilage proteogly-

ET AL.

can protein core. The appearance of chondroitin sulfate in these immunoprecipitates may also be accounted for by interactions between chondroitin sulfate and heparan sulfate proteoglycans. Finally, it may be possible that chondroitin/dermatan sulfate and heparan sulfate glycosaminoglycans are attached to the same protein core. The smallest of the [35S]sulfate-labeled macromolecules in the cell layer, K,, = 0.71, was not recognized by the antibody (Fig. 6~). This fraction may contain free glycosaminoglycans or glycosaminoglycans attached to a fragment of the protein core lacking antigenic determinates. It is also possible that the proteoglycan is derived from a different gene product. The antibody also recognized low-density (~1.32 g/ml) cell layer proteoglycans eluting from Sepharose CL-4B at K,, = 0.28 and 0.39. Most of the immunoreactive proteoglycans were released into the culture medium (approximately 7080% of the total incorporated [35S]sulfate). The immunoreactive high-density proteoglycan eluted from Sepharose CL-4B in a similar position as the high-density proteoglycan isolated from the EHS tumor, K,, = 0.23 (Fig. 3b). Based on sensitivity to nitrous acid, 50% of the sulfate is in heparan sulfate (compared to about 85% for the EHS tumor proteoglycan preparations). Also, the size of the heparan sulfate side chains on the PYS proteoglycan are smaller than those on the EHS proteoglycan (ikf, 30,000 for PYS high-density proteoglycan as compared to ikf, 65,000 for the EHS high-density proteoglycan). The antibody also reacted with a low-density, medium proteoglycan which eluted earlier from Sepharose CL-4B, Kav = 0.18 (Fig. 3a). The low-density proteoglycan contained heparan sulfate side chains which were consistently a little larger (ik& approximately 35-40,000) than the heparan sulfate side chains on the high-density form. It is interesting that the largest of the proteoglycans produced by the PYS cells is recovered from the tissue culture medium and not the cell layer. Low-density proteoglycans have been reported to contain hydrophobic domains capable of in-

HEPARAN

SULFATE

PROTEOGLYCANS

tercalating within the lipid bilayer (3-5). The extracelllular location of the low-density proteoglycan from PYS cells suggests it to be a component of the matrix rather than the cell. membrane. It is possible that the low-density form in the medium is derived from the smaller cell layer forms by the formation of protein-proteoglycan complexes. Alternatively, it is possible that the smaller proteoglycans with higher density are degradation fragments of the largest proteoglycans from which segments of protein and glycosaminoglycan have been removed, as suggested by Oohira et al. (12). Consequentially, secretion of the low-density form into the medium may allow it to escape from the degradation processes. REFERENCES 1. UNDERHILL,

C. B.,

B&hem. 2. OLDBERG,

J. Biol 3. KJELLEN,

AND

KELLER,

Biophys. Res. Commun Ii.,

KJELLEN,

J. M.

(1975)

63,448-454.

I., AND HBBK,

M. (1979)

Chem. 256, 13235-13242. L., PETTERSSON,

Proc. N&l

I., AND, HOOK, M. (1981)

Acad Sci. USA 78, 5371-5375.

4. NORLING, El., GLIMELIUS, B., AND WASTESON, A. (1981) B&hem Biophys. Res. Commun. 103, 1265-1272. 5. VOGEL, K. G., AND PETERSON, D. W. (1981) J. Biol. Chem. 256, 13235-13242. 6. RAPRAEGEF:, A. C., AND BERNFIELD, M. (1983) J.

Biol. Chem. 258, 3632-3636. 7. KANWAR,

Y. S., AND FARQUHAR, M. G. (1979) Proc. Sci USA 76, 1303-1307. Y. S., AND FARQUHAR, M. G. (1979) Proc

Nat1 Aad 8. KANWAR,

Natl. Aod

Sci. USA 76, 4493-4497.

9. HASSELL, J. R., ROBEY, P. G., BARRACH, H. J., WILCZEK., J., RENNARD, S. I., AND MARTIN, Sk USA 77,4494G. R. (19:&l) Proc Natl Ad 4498.

FROM

10. PARTHASARATHY,

PYS-2

CELLS

335

N., AND SPIRO, R. G. (1981)

J.

Biol. Chem. 256,507-513. J., AND 11. OOHIRA, A., WIGHT, T. N., MCPHERSON, BORNSTEIN, P. (1982) J. Cell Biol 92, 357-367. 12. OOHIRA, A., WIGHT, T. N., AND BORNSTEIN, P. (1983) J. Biol Chem 258, 2014-2021. A. 13. CARLSTEDT, I., CUSTER, L., AND MALMSTR~~M, (1981) Biochem. J. 197, 217-225. E. 14. OLDBERG, ii., SCHWARTZ, C., AND RUOSLAHTI, (1982) Arch. B&hem. Biophys. 000, 400-406. 15. HEDMAN, K., JOHANSSON, S., VARTILO, T., KJELLEN, L., VAHERI, A., AND H~~BK, M. (1982) Cell 28, 663-671. 16. HAYMAN, E. G., OLDBERG, A. O., MARTIN, G. R., AND RUOSLAHTI, E. (1982) J. Cell Biol 94, 2835. J. H., CAPUTO, C. B., AND HASCALL, 17. KIMURA, V. C. (1981) J. Biol Chem 256, 4376-4381. 18. RENNARD, S. I., BERG, R., MARTIN, G. R., FOIDART, J. M., AND ROBEY, P. G. (1980) Anal. B&hem. 104, 205-214. 19. HASSELL, J. R., NEWSOME, D. A., AND HASCALL, V. C. (1979) J. Biol. Chem. 254, 12346-12354. T., AND SUZUKI, S. (1968) 20. SAITO, H., YAMAGATA, .I Biol Chem. 243, 1536-1542. 21. SHIVELY, J. E., AND CONRAD, H. E. (1976) Biochemistry 15. 3932-3942. in Enzymology (Co22 LINKER, A. (1966) in Methods lowick, S. P., and Kaplan, N. O., eds.), Vol. 7, pp. 652, Academic Press, New York. 23. ROHRBACH, D. H., HASSELL, J. R., KLEINMAN, H. K., AND MARTIN, G. R. (1982) Diabetes 31, 185-188. 24. LEHMAN, J. M., SPEERS, W. C., SWARTZENDRUBER, D. E., AND PIERCE, G. B. (1974) J. Cell Physiol 84.13-28. 25. TRYGGVASON, K., ROBEY, P. G., AND MARTIN, G. R. (1980) Biochemistry 19,1284-1289. 26. HOGAN, B. L. M. (1980) Dev. BioL 76, 275-285. 27. COOPER, A. R., KURKINEN, M., TAYLOR, A., AND HOGAN, B. L. M. (1981) Eur. J B&hem. 119. 189-197. 28. HOGAN, B. L. M., TAYLOR, A., AND COOPER, A. R. (1982) De-v. Biol 90, 210-214.