Biochimica et Biopg~ica Acta, 1137(1992)287-297

287

© 1992Elseuier Science Publishers B.V. All rights resop,-ed 0169-488g/g2/$tLS.01)

BBAMCR 13299

Effects of cycloheximide, brefeldin A, suramin, hcparin and primaquine on proteoglycan and glycosaminoglycan biosynthesis in human embryonic skin fibroblasts Lars-~ke

Fransson,

Pernilla Karlsson and Artur Schmidtchen

IX'purteple~£ of Medical and Ph~'si~Jo~ical CT~t,~;islry, ~J~i~'~,rsil),,of Lfflld, Ltl/ld (Slt~den)

(Rc~iwd 24 Fcbruard 1992,' (Revised rllalluscrlpl received I July IO~)2)

Key words: Brcf~ldin A: Fibrobla:~t:OlI~cosaminoglyean:Prmeog[yearl;Surarain (I) We have isolated radiolabelled proteoglycans and glycosaminogly,~ans produced by human embryonic skin fibroblasls in the presence of to) cyclohcximidc to inhibit protein Synthesis or (b) brcfcldin A to impede transport between the endaplasmic rcticulum and the Golgi complex or ie) suramin, hcparln or primaquine to interfere "~ith internalization, recycling and degradation. Eff¢ct,s on glycosaminoglycan synthesis wcrc assayed separately by using exogenous p-nitrophenyl #-D-Xylopyratioside (and [3H]galaeto~e) or I'~l, labelled p-hydroxyphenyl ~t-o-x-ylepyranosidc as initiators. (2) Inhibition of protein synthesis or blocking of transport 1o the Golgi complex prc~¢nled production of most of tile protgoglyeans with one exception: Ccll-assceiated hcparan sulphate-proteoglycan was still produced at 20% of the control Icveh (3) Treatment with suramin or heparin rcsulmd in decreased deposition of proteoglycan in the p~rlcc]iular matrix but increased accumulation of cell-a'~sociatcd proteoglyCun. Primaquine blocked all pro!eoglycan synthesis. (4) In the presence of eyclohoximide, exogenou~ /LD-xyloside initiated galactosaminoglycan production. In contrast, in hrefeldin A-treated cells, synthesis was completely abolished. Not even formation of the linkage-region trisaccharide could be delccled. (5) These results suggest that exogenous ~loside enters the endoplasm~c rcticulum and is subsequently transported ¢o the wcms-Golgi complex where all further steps involved in glycosaminnglycan assembly takes place. (6) Hcparan sulphate proteoglyean produced by brcfcldin Aqroated cells could hc derived from (a) an ~ntracellular pool of preformed core protein located to the traus-Golgl complex, or (b) re,ideal prnteoglycau that was either dcglyeanated/reglycanated or chain-extcodod. As ¢omhincd treatment with suramin and brefeldin A markedly reduced cell-associated pro'cogtyean production, the latter possibility is favoured.

Introduction Proteoglycans (PG, i . e , glycosaminoglycan ( G A G ) substituted glycopmteins) are complex macromolecoles which occur in connective tissues, in basement membrangs, at cell surfacas and in iatracellular storage granules (for recent reviews, see Rcfs. 1-8). Cultured human fibroblasts synthesise a variety o f chondroitin sulphate (CS), d e r m a t a n sulphate (DS) and heparan sulphate (HS) P G s [5,9-12]. T h e C S / D S - c o n t a i n i n g

Correspondence to: L.-,~. Fransson, Department of Medical and Pbysiotnsical Chemistry, University of Lurid, P.O. Box 9~., $-ZZI O0 Land. Sweden. Abbreviations: CS, chondroilin sulphate; DS, derm~tan sulphate: ER, endoplasmlc r,:ticulum; GAG, slyoosaminoglycan; GlcA, Dglucuronalc; HS, ht:p~ran :~ulphale; IdeA. L-iduronate: PG. pmteogtycan.

PGs include a large glucuronate (GIcA)-rich P G [9], with a core protein o f 4 0 0 - 6 0 0 kDa and related to versican [2,4]. as well as 1we small iduronate (ldoA)-rich D S P G s [9,11], both with 45 k D a core proteins (PG-S1 oT biglycan and PG-S2 or decorin, respectively 12-4]L ali of which are mainly secreted into the ¢xtrac¢llola¢ space. T h e major H S P G is relatively large, has a 350kDa core protein and is deposited in the extracnllular matrix [1 I,L2]. T h e r e are also o t h e r small H S P G s which represent separate gone products: (a) a cell membraneintercalated H S P G with a 4 5 - 4 8 - k D a core protein, also termed fibroglycan [10,11], (b) a 6 4 - 7 0 - k D a core HSPG, termed glypican, which is found both in the culture medium and anchored to phosphatidyl-inositol at the c¢11 surface [13,14] and (e) por~ibly a 90-kDa cor¢ HSPG- [10] which is exclusively associated with the• cells [11]. The latter is accompanied by (d) a C S / D S P G also with a 90-kDa core [11,15]. HSPGs wlth core proteins of 250 [10], 130, and 35 kDa have also been

288 detecled [10,11]. At present, it is not known whether ,*hey are distinct gene products or degradation products of larser HSPGs. PG core proteins are synthesized on ribosomes bound to the endoplasmic retlculum (ER). The polypeptides acquire both N-and G-linked oligosaceharidcs as well as GAGs as they are transported from the rough ER to the cis-Golgi complex and then via the mediaf~compartment to the trans-Golgi complex and its network where synthesis is completed [7,8,16]. The first step in the synthesis of a G A G is the xylosylation of specific sefine residues in the core protein. The xylose is then extended step-wise with the scqocncc GIcAGaI-Gal-, the linkage-region trisaccharide. Whether these initial steps take place in the rough ER, in the Golgi complex, or both, is still controversial [8,16-20]. The iinkage-reglon seives as primer for the synthesis of both C S / D S and HS chains. It is generally assumed that the polymerases, sulpbotransferases and modifying enzymes involved in these final steps are located to the lrans-Goigi complex or its associated network [7,8]. The final PG products are secreted into the extraccliutar space or deposi*~ed in pericellular matrices or they remain attached to the plasma membrane. Exogenous /]-D-xylosidus can compcta with native, xylosylated core proteins and serve as alternative accepters for the assembly of G A G chains [6,21]. Xyloside-primed chains are usually secreted into the extracellnlar space. PGs are also subject to degradation and metabolic turn-over [1,5,6,8]. Partial degradation may lead to shedding from cell surface- or matrix-sltes and to the generation of oligosaccharide fragments of GAGs. Some PGs may be internalized and rapidly desraded. To gain insight :ate the subcellular location of P G / G A G biosynthesis and degradation, we have investigated the effects of various drugs that interfere with various steps in these processes. Cyclohaximide was used to block synthesis of new core proteins and brefeldin A to prevent their exit from the ER. Brefeldin A causes disassembly of the Golgi complex and aceu-. mulation of secreto~ proteins in the ER (see Ref. 22 and references therein). The mechanism involves disruption of a membrane-recycling pathway between the ER and the early Golgi-complex compartments, resulting in a fusion of these organelles. Brefeldin A can also affect endosomes which results in the formation of a mixed ttaus-Golgi.complex network/endo~mal compartment. This compartment can still receive endocylosed material and tee'role receptors to the ~ l l surface. The lysosomes, however, appear to be kept out of this traffic [23]. Suramin blocks cell-surface binding of various growth factors [24,25] and inhibits HS-degrading enzymes [26]. Heparin is also a potential inhibitor of HS-uptak¢ and dcgr,~dation. Primaquine was tested as a potential inh~itor of endosomal/lysosomal acidification.

Experimental Procedures

Materials HS, CS, DS and hcparin were obtained as described earlier [11,27.28] and dcalran 1"-500 was from Pharmaeia LKB, Sweden. p-Hydroxy'phenyl [LD-xylopyrauoside (XyI-PhcOH) was a gift from Dr. S. Suzuki, Aichi Medical University, Japan and p-nitrophenyl B-oxylopyraneside (Xy[-PheNO2) was a product of Sigma. Cell culture media were purchased from NordVacc. Sweden and the enzymes used were chondroltin ABC lyase (chondroitinase ABC, EC. 4.2.2.4), heparan sulphate lyasc 1 (fieparan sulphate lyase, heparitinase, hcparinase Ill, EC. 4.2.2.8) and hcparin !yase (heparinase, heparinase i, EC. 4.2.2.7) from Scikagaku Kogyo, Japan and heparan sulphate I~se It (hcparinase I[, no assigned EC. number) from Sigma. Brefcldin A was generously supplied by Sander, Switzerland, suramin (Germanine) by Bayer and cyclohcximide and primaquino were purchased from Sigma. The radiochemicals used were: Na~Sl (!.7' 104 Ci/g, Cimichcm, Tuxedo, NY, USA), L-[4,5-JH]leucine (50 Ci/mmol), o-[6-~H]glucosamlne (40 Ci/mmol) and o-[1--1H]galaclose (20 Ci/mmol, all from American Radiolabelled Chemicals, St. Louis, Me, USA) and Na~SSO4(1310 Ci/mmoi, Amctsham International, LIK). The propacked columns and column media were: fast-desalt~ng (FD) Sephadex G-25 10/10, Superose 6 HR 10/30, Sepharose CL-4B, Sephacryi S-500 HR and Mona Q HR 5/5 (all from Pharmacla-LKB), Bio-Gel P-.2 and P-6 (Bio-Rad) and DE 53 DEAE-collulose (Whatman).

Cell culotte and rodiolabelling Fibsoblasts from human embryoalc skin were grown as monolayers in Earle's minimal essential medium supplemented with 10% (v/v) donor calf serum, 2 mM L-glutamine, pen';cillin (100 nnits/ml) and streptomycin (100 p.g/ml). Confluent cultures between passages 5 and 15 were used in the e~cperimems. Incorporation of ~5SO4 was performed in sulphate-deficient medium (0.11 mM total sulphate), [JH]lencine in a Iow-leucinemedium [11] and [JH]galactos¢ in regular medium. Previous studies from this laboratory have shown (see, e.g., Ref. 27)that the degree of sulphation of GAGs is not affected by using sulphate-deficient medium. Drugs and xylosides were added as described in the appropriate figures.

lsoinliou of total cellular protein Confluent c u l t m ~ grOWn in 2-cm 2 dishes were preineubated with [¢ncine-defieient medium for 1 h and then labeled with [JH]leucine (5 /zCi/ml) in the same medium for various time-periods in the absence or presence of cycloheximide (0.5 raM). The medium was removed, cells wcrc rinsed in 0,15 M NaCi, t0 mM

289 KH2PO 4 (PBS (pH 7.5)). twpsinised (5 mg of enzyme in l0 ml of saline per 2-106 cells at 20°C for approx_ l min) and collected by centrifugatimt (.2000xg for 5 min) in the presence of serum-containing medium. Cells were lysed in l ml of 10% (w/v) trichloroacetie acid at 4°C overnight, centrifuged and washed three times with 5% (w/v) tcichloroacetie acid. The protein pellet was solubilised in 0.2 ml of 0.5 M NaOH at 37°C for 1 h and radioactivity was measured by liquid scintillation in an LKB-Wallach RackBeta counter with automatic quench correction using a i0-foid exce~ of Readygafe (Beckman) as scintillator.

Extraction and isolation of proteoglycam and glycogarainoglycan$

Radiolabelled, polyanionie maeromoleeulcs were isolated from the ealtttre medium, a detergent extract of the ceii~ and i'fom ii~¢. tcntt~ining matrix. (Incubations with radiolah¢llcd precurso~ were performed as described in the appropriate figures). After removal of the medium, cell layers were washed with ice.cold PBS (3 times 2-5 ml) and extrgcted with 0.2 ml/em z of 29"0 (v/v) Triton X-100 in PBS containing 10 mM EDTA. 10 mM N-ethylmaleimido (NEM) and 1 mM di.isopmpyl fluorophosphate (DFP) at 4°C for 10 rain. Extracts wore immediately mixext with 1.3 volumes of 7 M urea, 10 mM Tris (pH 7.5), containing 0.1% (v/v) Ttiton X-IO0 and 10 told NEM. The residual matrix was solubilised in the same volume of 4 M guanidinium chloride, 50 mM NaOAc (pH 5.8), containing 0.2% (v/v) Triton X-100 and l0 mM NEM, at 4°C overnight. In ~ome experiments, the cdl layer was treated with trypsin (see above) and the trypsinate and the cell pellet were separated by centfifugafion. The pellet was then extracted sequentially with detergent and guanldine as above. All the guanidine extracts were centrifuged and dialyzed against five changes of 6 M urea, 0.2 M NaOAc (p H 5.8), containing 0.1% Triton X-100. The medium, the detergent retracts and the non-dialyzable material From the guanidine extracts were separately chromatographed, in the cold-room, on DEAE-cellulose columns (1-2 mD whiclt were equilibrated with the same urea-solvent used for dialysis, but also containined 10 mM NEM and 5 ,gg/ml of ovalburain. After sample npplieation and washing the columns with 10 bed volumes of (a) the equilibrating buffer, (b) 6 M urea remaining 0.5 M NaOAe (pH 5.8), (e) 6 M urea containing 10 mM Ttis-HCI (pH 8,0), all containing 0.1% Triton X-100 and (d) 50 mM Tris-HCl (pH 7.5), poiyanionic material was displaced with five l-ml portions of ~, M guanidinium chloride, 50 mM NaOAc (pH 5.8), 0.2% Tciton X-100, 10 mM NEM and 5 ~g/trd of ovalbumin. Aliquots were analyzed for radioactivi~ by lio,uld scintillation. Radioactive polyanionie maeromolecules from the detergent extracts were recovered by precipitation with

6 volumes of ethanol, centrifuged and dissolved in 0.2 ml of 4 M guanidinium chloride, 50 mM NaOAc (pH 5.8) and applied, at room temperature, to l-ml oe~lSepharose CL-4B columns equilibrated with the same solvent. The applied samples were allowed to bind to the gel far 30 rain before elmioa with five l-ml portions of the equilibrating buffer was stared. Bound (hydrophobic) material was displaced with lh l-ml portions of 4 M guanidinium chloride, 50 mM NaOAe (pH 5.8), conmlning 1.0% (v/,a) Triton X.100. In some cases, samples were precipitated with ethanol, centrifuged, redissolved in 0.2 ml of 4 M gaanidinium chloride, 50 mM NaOAe (pH 5.8), containing 0.2% (v/v) Triton X-100 and chromatographed on Snperose 6 (see belo~) to separate PGs from free GAG s.

Preparation of s2al . l a b e l l e d Con)Founds p-HydroxTphenyl B-o-xWIoside (10 ~mol) was labelled with t2~l (0.8 mCi) using the ehloramine-T procedure [28] and separated from free tz~I by gel chromatography on Bin-Gel P-2 (packed in a IB.ml disposable pipette) which was eluted with 0.5 M pyridine acetate (pH 5.3). The effluent was analyzed for radioactivity by using an LKB 1271 RiaGamma Counter. The isolated product contained 1.3- 10u cpm/mmol. To prepare Gal_Gai.Xyl-[tXSl]PheOH and Xyl[m~flPbeOH cells were incubated with [3H]galaetose in the presence of 1 mM Xyl-eheOH for 24 h. Free ~H-labellcd GAG was isolated from the medium after ion exchange and gel ehromatogropby as described in detail elsewhere ['Jg]. After 1251-labellingof the PheOH group, the G A G preparation was digested ~hanstively with chondroitin AC-I lyase and chromatographed on Bio-Gel P-2 [29]. The linkage-region fragment AHegAGal.Gai.Xyl_[12sI]PheOH was isolated, treated with HgCI 2 to remove aHexA and the resulting Gal-OalXyb[t~I]PheOH was isobtted by fast-desalting gel chromatography on Sephadex G-2~. "File two terminal Gal were removed by digestion with ~,galacmsida~ [291,

Degradation methods

GAG chains were released from PGs by (i-elimination (0.5 M NaOH, 50 mM NaBH, at room temperatree overuightL Digestion with ehondroitin ABC lyase was conducted in 0.l M Tris-HOAe (pH 7.3)) at 37"C for 4 h using 50 mU of enzyme/mL For the degradation of HS, we used heparan sulphate (HS) lyases I and II, as well as hcpadn l~,a~ (4 mO of each e ~ e / m l ) in 3 mM Ca(OAr)2, 0.1% Triton X-100) 10 mM Hepes-NaOH (pH 7.0), at 37°(2 overnight. In all eases, the proteiuase inhibitors EDTA, NEM and DFP were added from stock solutions to final concentrations of 10 raM, 10 mM and l raM, resl~etiv¢Iv.

29O Chromatographic methods POx and GAGs were separated and fractionated by gel chromatography in 4 M guanidinium chloride, 0.2% Triton X-100, 50 mM NaOAc (pH 5.8), on 18 r a m × 1000 mm columns of either Sepharos¢ CL-4B or Sephacryl S-500 HR in the LC mode (flow-rate 5 m l / h ) or on Superose 6 HR 10/30 in the FPLC mode (flow-rate 0.4 ml/min) using LKB-Pharmaeia equipment. Ion-exchange FPLC was performed on Menu O as described [30]. Fragments of GAGs were resolved by gel chromatography on a pro-packed column of fastdesalting Sephadex G-25 or on a column (18 m m x 1000 ram) of BiG-Gel P-6. The former was eluted with 0.2 M NH~HCO~ at a flow-rate of 1 m l / m i n (FPLC mode) and the latter with 0.5 M NH4HCO 3 at a flow-rate of 5 m l / h (LC mode). The effluents we~c analyzed for the presence of 3H- or 3~S-radioactiv~t'y by liquid scintillation, using either a Radiomatie bloOneBeta (for FPLC effluents) or as described above (for LC effluents) and for t~Si-radioactivity by gamma-radlation spectrometry. Pooled PG or GAG material was recovered by chromatography on 0.2-ml columns of DEAE-cc!lulose (see above) after dilution to an ionic strength below 0.4. The material was eluted in a small volume (less than 1 ml) and recovered by precipitation with 6 volumes of ethanol, centrifuged, washed with 95% (v/v) ethanol saturated with NaOAc, dried and dissolved in the appropriate buffer or solvent.

SDS-PAGE This was performed on 3 - 1 2 % gradient gels, with a 3% stacking gel, as described [11,30]. Samples were precipitated with 9 volumes of ethanol, centrifuged, dried, dissolved in SDS-containing sample buffer and boiled for 3 rain. After eleetrophoresls radiolabclled material was identified by fluorography (for details, see RoL 11). Results

Effects of cycloh~rimide o~ proteoglycan production To examine PG production in the absence of de novo protein synthesis, fibroblasts were treated with cyeloheximide for various timc-period~ and then pulsed with ~5S-snlphate for l h (with ~eloheximide still present). Radiosulphatcd, polyanionic macromolecules, i.e., PG or GAG, isolate~ from the culture medium, a detergent extract of the cells ('membrane" fraction) and a guanidine extract of the remainder ('matrix" fraction) were chromatographed on Sopharose CL-4B (Fig. 1). In the control cells, detcrgent-solubilized material consisted of both a large-slzed (Ka~ 0.1) and a small-sized (K~ 0.4) pool (Fig. la). Inhibition of protein synthesis by cycloheximide, resulted in rapid reduction of PG production (Fig. lb-e); within 1 h the amount of detergent-so~ubilized, radiolabetled PG was

[o) I~E~RANE

I.G~uZohe.ir~el.(i.511|]

[

(a]

~h

o~

/--,.. tte3 u e M s ~ ^ ~ s~ [0 ~Eolt~ s~

o

o

i

[0] MJ,TmX ~n

(b3 ~ u B ~ ~,NE

v~h

r~

025

;

6

o:e

;

Kay

Fig. 1. Gel chromatography~n Sepbarose CL-4:Bof 3SS.pulse-lahelled polyanionic macromol¢¢ules produced by ftbmblasff at variol~ limes after the addition of cy¢loheximide.Skin fibroblasts,grown in 10-cm2 dishes, wereincubated with sulphate-deficient medium (2 nil) fur 1 b followed by the same medium containingflJ mM cycloheximidefor the indicatedtime-perlods(a-g/, Then the cells were pulse-labelled for t h by addition of 500 /tCi X~S-~ulphat¢.Tile medium (tEl. a det©rgent-eTnrac!of the eell~ (membrane fraction,a-e) altd a ~,mnidine-extract of the residue (matrix fractk .; g) were prepared as described in Exl~fimeetalProcedures. Polyanioaicmacromoleculcs were isolaled by ion-e~hang¢ ebTomatography on DEAE-c:ellulo~e and subjected to chromatography,

approx. 20% of the control. At the same time, incorporation of [3H]leuelne into cell protein was 5.8% of the control. The PG produced in the presence of cyeloheximide was ~latively large (K~v 0.2) and continued to be produced as long as the experiments were conductcd, (Cells were generally viable for 6 - 7 h), After 6 h of cyclnheximide treatment, when protein synthesis was 1.4% of the control, production of the detergent~lubilized PG was approx. 50% of the level observed after 0.5 It. There was no secretion of PG into the medium (Fig. IO and little deposition in the matrix (Fig. lg). These results indicate that some membraneassociated PG can be produced despite almost complete inhibition of de nero protein synthesis. Hence, the ceils must contain a pool of cole- protein or proteo. glycan that can be made available for further PG synthesis.

Effects of brefefdin A, snramin, primaqnine and heparbi on proteoglycun production To assess the location of the p r o c u r e r pool, a number of drugs with reasonably well-known effects on

291

protein trafficking a n d receptor recycling w e r e tested for their effects o n P G production. 3sS-iabelled, macrom o l e c u l a r a n d poiyanionic product s derived from the m e d i u m , the matrix a n d a d e t e r g e n t lysate of the ceils ( s e p a r a t e d into h y d r o p h o b i e a n d non-hydrophobic m a terial) were analysed by S D S - P A G E (Fig. 2). As s hown in t h e contcol Jazzes (lanes ! in Fig. 2), the secreted P G s e n c o m p a s s e d a w i d e variety of sizes (Fig. 2a), the (m~

M~diu

m

m e m b r a n e - d e r l v e d a n d hydrophobie P G s w e r e generally l a r g e r t h a n 200 k D a (Fig. 2c) a n d the matrix-bound P G was the largest (Fig. 2b). T h e cell iysate (Fig. 2 d ) also c o n t a i n e d n o n - h y d r o p h o b i c [my-molecular weight d e g r a d a t i o n p r o d u c t s ( m o s t o f v, hich are free d e r m a * a n sulphate chains, see also R e L i1). l a ceils treated with brefeldin A, which p r e v e n t s egress from the E R , scct'ct i e s of P G t o the m e d i u m a n d deposition in the matrix (b)

1 2 3 4 5 6

(c)

1

Membrane/

(d)

Hydrol)hoblc 1

2

3

4

Matrix 2

C~ 4-

Membrane/ No n--hydr

5

R

..~ 6

1

2

3

4

ophobic 5

6

Fig. 2. SDS-PAGE ( md aclnS conditions) of 3~g-labelled polyanlonlc macromolecules derived f~m the spent euhtlre medium (a), the I~erlcellalar matrix (b) or the cellular merabra~t¢ fraction (cgl) of cells treated with various substances. Confluem cu~u~s were preincubalcd as in Fig. I and then labelled with 50 ~uCi./ml of 13~S]sulphate for 24 h. Lame t, no addition; lane 2,10 P.8/ml of brcf~!~*.inA; latt~ ~., 0.2 raM suramin; lane 4, 0,5 mM ptlmaquine; lane .5, L0 mg/ral of hepali~ lane 6, both 10 .ug/ml of brefeldin A and 0.2 mM sltramin. Poly'anlonlc macmmolecu~s ~ isolated from the three comparlments by ion-exchange chromatography as described in Expel'imental Procedure~. Material from the delerBent e:ctzact (raembrane fraction, u,d) was separated isle hrdlopbobi¢ (¢) and non-bydmphobic maleriat (d] by chmrae.toMaphy on octyi-Sepharos¢ as described in Experimental Praeedarcs. Similar allquots from each fraction v,rere precipitated with ethanol and subjected to electrophecesis, Molecular mass marke~ a~ indicated on the left. The expurlmcnt was reproduced twice.

292 were completely abolished (Fig. 2a and b, lanes 2). However, membrane-bound, hydrophobic FG was still produced (Fig. 2c, lane 2), but no degradation products were seen (Fig. 2d, lane 2). $mamin, which inhibits both endoheparanase and recycling of membrane receptors, had no apparent effect on secretory PGs (Fig. 2a, lane 3), hut ceased a marked reduction in deposition of matrix PG (Fig. 2h, lane 3). However, all forms of radiolabelled raacmmolecules in the cell lysate (Fig. 2e and d, lanes 3) apgearcd to be increased. Primaquine, which was expected to inhibit lysosomal/endnsomal acidification and thereby degradation of PO, resulted in complete cessation of P(:3 production (Fig. 2, lanes 4), probably ImeauSe it raises the pH in the trans-Golgi-complex network. Heparin could potentially inhibit complex formation between HSPG sidechains and various matrix-proteins, HS-binding growth factors or HS-degrading enzFraes. The effect of this drug was limited to a reduced deposition of PG in the matrix (Fig. 2b, lane 5). As brefeldin A inhibited all PG-produeti~n except membrane-bound forms, whereas suramin increased the formation of such PGs+ it was of interest to treat cells with both drugs simultaneously. As shown in Fig. 2c (lane 6), the result was a much reduced formation of membrane PG. A quantitative estimation of the effects of some of the drugs is presented in Figs. 3 and 4. In Fig. 3, the yield of [35S]sulphate- (a) or ['~H]glucosamine-labelled (b), polyanionic macromolceules in the cell lysate is presented (as % of control), in the control (see Fig. 3a, experiment t), the major portion of the ceii-ags~,2ciated -~sS-iabolled material consisted of non-hydrophobie GAG-chains (mainly dermatan sulphate) and heparan

sulphate oligesaccharides (see also ReL ll). In brefeldin A-treated cells (experiment 2), membranebound matcria| was reduced to approx. 20% of the control and most of this material was hydrophobic. Suram,:n increased production of both hydraphobic and non-hydrophoblc material by nearly 50% (experiment 3). In accordance with the results shown above (Fig. 2), primaqnine (experiment 4) and the combination brefeldin A/snramL~ (experiment 6) larl~ely abolished production of ~Sqabcllcd macromolccules. Although beparin did not markedly alter the total amount of membrane-bound material, the formation of degradation products (non-hydrophobic material) was decreased (experiment 5). By using [~H]glucosamine as che GAG.precursor similar results were obtained (Fig. 3b:,. Also incorporation into secreted material was affected in the same manner as shown in Fig. 2a and b (results not shown). Incorporation of [~sS~suiphate into PO and G A G in the absence or presence of brefeldin A was also followed Over a 24-h time-period (Fig. 4). It is seen that, in the presence of brefeldin A, secretion o[ [XSS]suiphate-labelled PG into the medium (Fig. 4a) and deposition in the matrix (Fig. 4c) was totally inhibited. However, production of radiolai0clled dctergentsolubilized P(3 was reduced to apprux. 20% of the control (Fig. 4b). Incorporation of radiolabel appeared to reach a steady-state after approx. [5 h. We also treated intact ceils with trypsin to defermine if the PG produced in the presence of brefeldin A ~'ds located at the cell surface. Approx. 60% of the PG produced during this treatment was released by trypsin digestion.

Fh~. 3. lncorpv~alion of [3sS~sulphute (a) and i3Fl~]u¢osurninc (b) into pol~aniottic~ detergent-extracted Crncrnb[anc-bound) macromolccut~:s p~ctucea by umr~azed cells (I) or cclh trcmcd with brefel~in A (2L suramin (3), pHrnaquinc {4). bcparin (5) or bl~feldin A and sv,rarnin (6). The ~pcdmenls ~ r c carried mlt u.'~ ¢Icscdbed ill Fig, 2 and in [a) Ikc isulatcd matcfiM was separated into bydn0phobic {Dpl~n bar~) ~ild rton-hydrophohic material (sniped bars). In (b). the material wa~ al~z) pmlfied by io*I exchange FPLC on MonoO In remnve cont~minalLqg hyaluronan and Sb'Cop~o~¢ins.

293 Characterization o f the-~-~S-labelled macromolecul~ sy~z. ~esized in file presence o f brute/din A T o examine the nature of the P G formed in the presence of br~fcldin A it was subjected to gel- o r ion-exchange chromatography before and after various specific degradations (see Figs. S-7). Intact P G prodoced in the presence of brefcldin A was larger than many o f the radiolabelled molecules made in the c~ntrol culture (Fig. 5a and b). It had a Ear of 0.1 ou Sul~:rose 6 (Fig. b-b), 0.2 ou Sepharose CL-4B (data not shown) and 0.4 on Sephacryl S-500 H R (data not shown). T h e PG was degraded by alkali (Fig, 5b, dashed line) and the side-chains released were larger than HS-chains derived from the major, matrix-associated H S P G from untseated fibrohlasfs (Re,l'. LI and Schmidtehen, A., unpublished results). Similar results (not shown) were obtained with t h e PG produced b y cveloheximide-treated cells, except that the side-chains also included

a pepulation

[a]

Control

S~:A-.~ac

(¢I

~

tb) -

egA ~ d l 'I [

.~_ BF,&-HS'~g1"41

- - untrealed - - - alka-

o f t h e s a m o size as t h e

I-IS-chalns derived from the matrix PG. T h e side-chains released from the PG pyoduced in the presence of brefeldiu A were treated with enzTmes that specifically degrade C S / D S (chondroitin A B C lyase,) o r H S / Kay Fig. 5. Gel FPL'C or= Superos¢ 6 of "~sS-[abcl[cd polyanioni¢ mocrornolcculca ~ynthesized by control and brefeldin A-t~ale'd cells. Fi-

,¢

N

brohlasls were incubated with [3"~S]sulphate in the absence tat or presence of ]h izg/nri of breteldin A (b-d) [or 24 h as described in the Icgelld IO Fig. 4. Polyanionic mac~ul¢culc~ were [solaled front the dele~e~z extra¢l {see Fig. 4b) by ion-exchange chromalography nn DEAE-ccllu[ose as described in the Experimental procedRre and finally subjected to :gel chromatosmphy. The materiels were chro-

/.-"

t

1

•

"~'"

7/f.

matographed dineclly in (a) and lb. solid liner, afier alkaline Belimination (b, dashed line) and alter dig~lion with cbOm[rOilh~ ABC lyasc nr hcparan sulphate | and II lytLse in (c) and (d), rosl~ctiv~ly. HS, ¢lution position of chains from the major fiblobl~t HSPG (matrix-associated and wilh a core protein tlf 35[I kDa [I I]).

J"

/ .

I 2 4

/

10

15

~4

TiME (hi Fig. 4. |ncorporatlon of t3AShulpha~c into polya~onlc macro-

molecules in the absence or presence or bfeFeldin A. FibroHastg we¢¢ grown and pic-iacubalcd as in Fig, I and then labelled with 100 b',Ci [~f'S]~ulphat¢ in the al~,ence t o n - - O) or presence (~ o) ~t Ill .ag]rnl of I)rmfcldln A for the indicated lime-periodLm. Polyan/,ln[c m ~ o ~ e ~ u l e s W¢~ i~olated from 1he Ihme ~mpartmflll5.

medium taL membrane (b} and rnalrix (c) by iOnrexcbmng¢ chromalOsr=pIWon DEAE-eellulc~e am¢lescribcdht Eapexirr~tllal Procedu+es. The experiment was do~¢ in triplicates three times,

heparin (HS [yases I and I l k As shown in Fig. 5c, the vast majority of the side-chains were resistant to the A B C lyase, but they were extensively d e p o l ~ e r i z e d by the H g / h c p a r i n - d c g r a d i n g ¢rtzytnes (Fig. 5d). T h e PG produced in the presence o f bzefeldin A could have bccn labeled with -tSS in a d e s u l p h a t i o n / r e sulphation process. T o determine irchain synthesis had taken place we incubated cells with both [SH]g[ncosamine and radlosulphate in the presence o f brefeldin A. As shown in Fig. 6, a -1H/SSS4ab¢lled P G could b¢ isolated by ion-exchange FPLC on MonoQ. It eluted in a position corresponding to that o f D$. suggesting that it had azl unusually high negative charge density. Theso results indicate that the P G proda¢od had acquired newly-synthesized, highly-charged HS-chalns. T o further examine the charge-density of the HS, chains, ion-exchange FPLC o f alkali-treated P G was performed (Fig. 7). The. rcsufts indicated that also the HS-chalns had a relatively high negative charge den-

294 HS

DS

1

1

mM Xyl

&

1.0 mM x¥1

/ u

ii =