Degradation of Endocytosed Dermatan Sulfate Proteoglycan in Human Fibroblasts*

Endocytosis and subsequent degradation of iduronic acid-rich small dermatan sulfate proteoglycan from fibroblast secretions were studied in human fibro- blasts. Upon endocytosis of [3H]leucine- and [35S]su1-fate-labeled proteoglycan release of free leucine was 10 to 15 times more rapid than that of inorganic sulfate. Within approximately 3 h a steady state was approached between transport of proteoglycan to the compartment of core protein degradation and release of free leucine. No such steady state could be found with respect to the dermatan sulfate chains. In the presence of benzyloxycarbonyl-Phe-Ala-diazome- thylketone or of other SH-protease inhibitors the degradation of the protein moiety of endocytosed proteo- glycan was much less inhibited than the degradation of the polysaccharide chain. Benzyloxycarbonyl-Phe- Ala-diazomethylketone did not affect the degradation of dermatan sulfate chains taken up by fluid phase endocytosis and the activities of all known dermatan sulfate-degrading enzymes. Percoll gradient centrifu- gation indicated that also in the presence of the protease inhibitor the partially degraded proteoglycan ac- cumulated in dense lysosomes. The isolation of intracellular dermatan sulfate peptides and molecular size determinations of endocytosed dermatan sulfate proteoglycan supported the conclusion that a critical pro- teolytic step is required before the dermatan dermatan sulfate-degrading enzymes p-N- acetylhexosaminidase (25), a-iduronidase (26), @-glucuronidase (25), N-acetylgalactosamine 6-sulfatase (27), N-acetylgalactosamine 4-sul- fatase (arylsulfatase B (28)), and iduronate sulfatase (26) were assayed in the absence or presence of 1 pM Cbz-Phe-Ala-CHNn as previously described. SDS-polyacrylamide gel electrophoresis followed by fluorography was done as quoted earlier (17). Protein was quantitated as described (29) using bovine serum albumin as standard. During the pulse period between 4400 and 5150 cpm were endocytosed without significant degradation, +, intracellular radioactivity in control cultures; A, intracellular radioactivity in Cbz-Phe-Ala-CHN,-treated cultures.

Endocytosis and subsequent degradation of iduronic acid-rich small dermatan sulfate proteoglycan from fibroblast secretions were studied in human fibroblasts. Upon endocytosis of [3H]leucine-and [35S]su1fate-labeled proteoglycan release of free leucine was 10 to 15 times more rapid than that of inorganic sulfate. Within approximately 3 h a steady state was approached between transport of proteoglycan to the compartment of core protein degradation and release of free leucine. No such steady state could be found with respect to the dermatan sulfate chains. In the presence of benzyloxycarbonyl-Phe-Ala-diazomethylketone or of other SH-protease inhibitors the degradation of the protein moiety of endocytosed proteoglycan was much less inhibited than the degradation of the polysaccharide chain. Benzyloxycarbonyl-Phe-Ala-diazomethylketone did not affect the degradation of dermatan sulfate chains taken up by fluid phase endocytosis and the activities of all known dermatan sulfate-degrading enzymes. Percoll gradient centrifugation indicated that also in the presence of the protease inhibitor the partially degraded proteoglycan accumulated in dense lysosomes. The isolation of intracellular dermatan sulfate peptides and molecular size determinations of endocytosed dermatan sulfate proteoglycan supported the conclusion that a critical proteolytic step is required before the dermatan sulfate chain becomes accessible to hydrolytic enzymes.
Degradation of proteoglycans to their monomeric constituents is an intralysosomal event as concluded from the intralysosomal accumulation of partially degraded glycosaminoglycans in the mucopolysaccharidoses, in which one of the glycosaminoglycan-degrading enzymes has been rendered inactive by mutation (1,2). Partial proteolytic or endoglycosidic degradation, however, may occur extracellularly (see Ref. 3 for a review) or in a prelysosomal intracellular compartment (4, 5). Different ways of initiating breakdown may be considered. First, the intact proteoglycan is internalized by receptormediated endocytosis. This is proposed to be the fate of a ubiquitous small dermatan sulfate proteoglycan (DS-PG)' the *This work was supported by the Deutsche Forschungsgemeinschaft (SFB 310, Project B2) and by the Braun-Stiftung. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. protein moiety of which is considered to be recognized by a specific receptor (6)(7)(8)(9). Second, a complex of proteoglycan and a second constituent of the extracellular matrix (e.g. fibronectin) which is bound to its membrane receptor is internalized as a result of normal membrane flow. Such a mechanism may operate in the internalization of heparan sulfate proteoglycans (10). Third, the proteoglycan diffuses out of the tissue of origin, maybe after limited proteolysis. It could then be degraded by liver endothelial cells which are equipped with a receptor recognizing hyaluronate and chondroitin sulfate (11,12).
The transport of proteoglycans from the plasma membrane to the lysosomes has not been studied in detail. Electron microscopic studies showed that after application to cultured smooth muscle cells proteoglycans from arterial tissue were present in coated vesicles, noncoated vesicles, multivesicular bodies, and secondary lysosomes (13). Two intracellular pathways for the degradation of membrane-bound proteoglycans were proposed to operate in rat ovarian granulosa cells (14). One pathway leads to a rapid and complete intralysosomal degradation (tlh of 30 min). In the second more slowly operating pathway (tlh of 4 h) degradation starts with extensive proteolysis and endoglycosidic breakdown in the case of heparan sulfate before final hydrolysis takes place.
Studies on the turnover of proteoglycans are complicated even in the simple model of cultured cells by the fact that cell-associated proteoglycans are localized not only in the different biosynthetic and degradative compartments but also in the nucleus (15) and as integral and peripheral components in the plasma membrane (16). Thus, the disappearance of proteoglycans from the cell comprises exocytosis and shedding as well as intracellular degradation. To get some insight into the interdependence of proteolytic and glycosidic degradation of proteoglycans we have, therefore, used an experimental system in which the degradation of endocytosed iduronic acidrich small DS-PG could be studied with minimal interference by the other processes of proteoglycan metabolism.
This DS-PG which is the main secretory product of cultured skin fibroblasts (17) is one of the two species of small proteoglycans that carry galactosaminoglycan chains (18). It contains a single dermatan sulfate chain (Mr approximately 37,000) which is linked to a serine residue near the N terminus (19). From cloned cDNA a molecular weight of 36,319 for the mature core protein was calculated (20). Additionally, either two or three asparagine-bound oligosaccharides are present (17). It had been shown previously that the core protein of DS-PG is required for receptor-mediated endocytosis, and that lysine and arginine residues are partly responsible for efficient uptake (9).
Preparation of Labeled Proteoglycans and Glycosaminoglycam-Skin fibroblasts from healthy juvenile and adult donors were maintained in culture as described (22). For the preparation of labeled proteoglycans cells grown to confluency in 75-cm2 Falcon plastic flasks were used. They were incubated with [3'S]sulfate in sulfatedepleted and modified Eagle's minimum essential medium (22) and with [3H]leucine and/or [3'S]sulfate in modified leucine-free Waymouth MAB 87/3 medium (17) as quoted earlier. For the preparation of [3H]arginine-labeled proteoglycans the cultures received 1 MBq/ ml [3H]arginine in arginine-deficient modified Waymouth MAB 87/ 3 medium for a period of 18 h. Labeled proteoglycans were isolated from the spent media by chromatography on DEAE-Trisacryl as described (9). After addition of fetal calf serum to a final concentration of lo%, they were dialyzed against serum-free Eagle's minimum essential medium.
[35S]Sulfate-labeled glycosaminoglycan chains were liberated from the core proteins by treatment with 0.15 M NaOH for 6 h at 37 "C and used for uptake experiments after neutralization with acetic acid, addition of serum, and dialysis.
Determination of Endocytosis and Degradation of Proteoglycans-Three to five days before endocytosis experiments the recipient cells were densely plated in 60 X 15-mm culture dishes or 25-cmZ plastic flasks (Falcon). Confluent cultures were preincubated for 24 h with the protease inhibitors indicated or the solvent of the drug before radioactively labeled proteoglycans were added. Preincubation and incubation media contained additionally 10 mM Hepes; the pH was adjusted to 7.3. Endocytosis and degradation were measured as described previously (6). Briefly, the radioactivity of the culture medium and of the ethanol-insoluble material in the cell pellet obtained after trypsinization was determined, as well as the ethanol-soluble radioactivity of culture medium and cell pellet. Endocytosed material represents the sum of intracellular and nonprecipitable radioactivity of the medium minus nonprecipitable radioactivity of the control medium. When expressed as clearance rate, the volume of medium cleared from labeled material/h and mg of cell protein are given. Degradation refers to the nonprecipitable part of endocytosed radioactivity.
In studies on the uptake of [3H]arginine-labeled proteoglycans recipient cells were grown in 75-cm2 plastic flasks and challenged with 5 ml of medium containing [3H]arginine-labeled DS-PG. A t the end of the incubation, the cell layer was washed twice with 3 ml each of 0.1% trypsin in Hanks' salt solution and incubated with 100 p1 of 0.1% trypsin for 14 min at 37 "C. 4 milliliters of Hanks' salt solution and 1 mg of cq-antitrypsin were added, and the cells were collected by centrifugation. After washing with serum-containing medium, the cells were solubilized by sonication in 1 ml of buffer A (20 mM Tris/ HCl buffer, pH 7.4, containing 0.1% Triton X-100 and protease inhibitors) containing 0.15 M NaCl. The extract was loaded on a 1.5ml DEAE-Trisacryl column prepared in a Pasteur pipette and equilibrated with this buffer. The column was eluted stepwise with 1.5 ml each of 0.15 M NaCl, 0.3 M NaCl, 0.4 M NaC1, and 1.0 M NaC1, all in buffer A. Appropriate fractions from the last step were dialyzed against 0.1% Triton X-100, concentrated to about 300 pl in a Speed-Vac concentrator (Bachofen), and treated with chondroitin ABC lyase (23). The digest was dried in vacuo, washed sequentially with methanol, 20% (w/v) trichloroacetic acid, and methanol prior to SDSpolyacrylamide gel electrophoresis.
Molecular Size Analysis of Endocytosed Proteoglycam-For molecular size analysis of endocytosed proteoglycans, the cell pellet obtained after trypsinization was extracted with 400 p1 of 1% SDS containing 0.1 M 6-aminohexanoic acid, 10 mM EDTA, disodium salt, 10 mM N-ethylmaleimide, and 5 mM benzamidine hydrochloride. Solubilized material prior and after the alkaline treatment described above was analyzed by high-performance liquid chromatography on a TSK G 3000 SW column (7.5 X 500 mm; Varian) in 0.1% SDS, 50 mM sodium phosphate, pH 6.0. Fractions of 0.5 ml were collected at a flow rate of 0.25 ml/min. Subcellular Fractionation on Percoll Gradients-After endocytosis, the cells were harvested, and postnuclear supernatants were applied on top of a Percoll gradient medium ( p = 1.065 g/cm3) as described (24) except that the sucrose cushion was omitted. Centrifugation was for 40 min at 20,000 rpm in a VTi 50 rotor (Beckman Instruments).
Fractions of 2 ml were collected and assayed for radioactivity, P-Nacetylhexosaminidase, and density.

RESULTS
Kinetics of Degradation of Endocytosed DS-PG-Proteoglycans secreted by human skin fibroblasts represent a mixture of various proteoglycans. 80-90% of the total amount consist of DS-PG. It had been shown previously that large chondroitin sulfate/dermatan sulfate proteoglycans are internalized by bulk phase endocytosis only (30). The clearance rate of heparan sulfate proteoglycan is about 10-fold lower than that of DS-PG (10). In consideration of the proportions of different proteoglycans about 98% of endocytosed proteoglycans reflect the uptake of DS-PG. With respect to endocytosis, the terms proteoglycan and DS-PG will, therefore, be used synonymously in the present study. The

Degradation of Dermatan
Sulfate Proteoglycan leucine and [35S]sulfate (9). Incubation of DS-PG with con-protein degradation (Fig. 2D), but after 3 h of endocytosis the ditioned medium in the absence of fibroblasts did not result degradation rate was only about 70% of that of the untreated in measurable degradation of the proteoglycan. Fig. 1  docytosed DS-PG-The observation that core protein degra-By employing longer incubation periods, it could be shown dation precedes glycosaminoglycan degradation could point that after approximately 3 h a steady state was approached to the existence of a critical proteolytic step taking place prior between entry of labeled DS-PG core protein into the com-to dermatan sulfate hydrolysis. We have, therefore, investipartment of its degradation and release of free [3H]leucine gated the influence of several protease inhibitors on the (Fig. 2C). Within 6 h no such steady state for the degradation degradation of endocytosed DS-PG. It is seen in Table I     Cbz-Phe-Ala-CHN, was used to study the dose dependence of the drug on the degradation of endocytosed [3H]leucineand [35S]sulfate-labeled DS-PG (Fig. 3). Unexpectedly, the degradation of the protein moiety of the proteoglycan was much less affected than the degradation of the carbohydrate portion. When fibroblasts were exposed to [35S]sulfate-labeled DS-PG and Cbz-Phe-Ala-CHN, for up to 30 h the proportion of degraded material on the endocytosed amount remained nearly constant (Fig. 4) indicating that the inhibitory effect did not change during this time period.
The half-life of the glycosaminoglycan moiety of endocytosed DS-PG was determined as follows. presence of 10 mM NH,Cl. During a subsequent chase in the absence of NH,Cl the amount of intracellularly remaining radioactivity was determined (Fig. 5). A tlh of about 12 h was found in control cultures, whereas Cbz-Phe-Ala-CHN, treatment resulted in a doubling of the half-life time.

Influence of Cbz-Phe-Ala-CHN, on the Enzymatic Degradation of Free Dermatan Sulfate
Chains-The inhibitory effect of SH-protease inhibitors on the degradation of endocytosed dermatan sulfate chains could theoretically be caused by a reduced activity of glycosaminoglycan-degrading hydrolases, by a delayed transport of endocytosed proteoglycan to lysosomes, or by an inaccessibility of the partially degraded substrate to the enzymes required for dermatan sulfate catab-

Effect of cycloheximide on endocytosis and degradation of P'S] sulfate-labeled DS-PG
Fibroblasts were pretreated with 20 PM cycloheximide or 1 pM Cbz-Phe-Ala-CHN, for the times indicated and then incubated with DS-PG (88,000 cpm) for 12 (Table 11). Uptake of protein-free dermatan sulfate chains occurs most likely by bulk endocytosis only, which explains the relatively low rate of degradation within the endocytosis period of 18 h as a result of the low clearance rate. However, in several experiments Cbz-Phe-Ala-CHN, did not exert a significant influence on the catabolism of free dermatan sulfate chains. The conclusion that dermatan sulfate-hydrolyzing enzymes were normal after Cbz-Phe-Ala-CHN, treatment was supported by in vitro assays of P-glucuronidase, a-L-iduronidase, 6-N-acetylhexosaminidase, iduronide 2-sulfatase, and N-acetylgalactosamine 4-and 6-sulfatases. Normal activities (*lo% of control values) were found regardless of whether homogenates from pretreated fibroblasts were used or whether the inhibitor was added to the assay mixtures (results not shown).
Influence of Cycloheximide on Endocytosis and Degradation of DS-PG-Almost all lysosomal enzymes undergo proteolytic maturation during transport to the lysosome. It seems conceivable that Cbz-Phe-Ala-CHN, could interfere with this processing and possibly with the transport of such enzymes. Since we could not investigate all dermatan sulfate-degrading enzymes for these events, we studied the consequences of cycloheximide treatment on DS-PG degradation in comparison with the effect of Cbz-Phe-Ala-CHN,. It can be deduced from Table I11 that cycloheximide did not cause an inhibition of endocytosis or degradation of DS-PG though incorporation of [3H]leucine into protein was reduced by more than 90%. This result indicates that even an almost complete inhibition of protein synthesis does not affect the capacity for DS-PG degradation under the present experimental conditions. Transport of DS-PG to Lysosomes-The second possibility to explain the inhibitory effect of SH-protease inhibitors, delayed transport of endocytosed DS-PG to lysosomes, has been investigated by subcellular fractionation of fibroblasts after various time periods of endocytosis ( Fig. 6). At least within 3 h similar amounts of radioactive material could be found in the region of dense lysosomes (density approximately 1.08 g/cm3) in control and Cbz-Phe-Ala-CHN3-treated fibroblasts. As expected, there was accumulation of radioactivity in dense lysosomes during the subsequent endocytosis period in treated cultures. Only minor amounts of radioactivity were detected in the region of light lysosomes (density approximately 1.05 g/cm3). This fraction represents a heterogenous mixture of microsomal vesicles including endosomes (31). Its content of "SO4 radioactivity was similar in preparations from treated and control cultures. It should also be noted that only minor amounts of radioactivity were found in the nuclear pellets.
Molecular Size of Endocytosed DS-PC-To test for the accessibility of endocytosed DS-PG for DS-PG-degrading enzymes fibroblasts were allowed to take up labeled proteoglycan in the presence of 10 mM NH4C1 during a 4-h pulse, and the molecular size of endocytosed material was followed during a subsequent chase in the absence of NH4C1. Cell extracts were chromatographed on a TSK 3000 SW column prior and after a @-elimination reaction (Fig. 6). Within the first 2 h of chase, there was a shift of the peak of the radioactivity in the control culture; thereafter only peak broadening could be observed. Peaks of progressively lower molecular size have not been  found. In Cbz-Phe-Ala-CHN*-treated cultures, the shift of the radioactivity peak was delayed. The size of the glycosaminoglycan chains was almost identical in extracts from untreated and treated cells after chase periods of 2 and 4 h; a maximum difference of 3% was found when the radioactivity of a given fraction was expressed as percent of total radioactivity. It is, therefore, likely that the size differences shown in Fig. 7 result from differences in the proteolytic degradation of the core protein.
Intracellular Accumulation of DS-PG Peptides-The results described in the preceding sections suggest that DS-PG peptides are accumulating in Cbz-Phe-Ala-CHN,-treated cultures. Since the sequence around serine 4, the attachment site of the dermatan sulfate chain, is rich in arginine residues (20) and since the most prominent lysosomal SH-protease, cathepsin B, is very active toward arginine carboxyl bonds (32), ['HI arginine-labeled DS-PG was used to search for such peptides. Fig. 8 shows that after endocytosis of this proteoglycan only trace amounts of dermatan sulfate peptides could be isolated from the control culture. In contrast, an only partially degraded DS-PG species was found in the presence of the thiol protease inhibitor.

DISCUSSION
The results described in this paper provide evidence that during the course of intracellular degradation of endocytosed DS-PG proteolysis of the core protein precedes the breakdown of the dermatan sulfate chains. Release of inorganic sulfate from DS-PG represents a measure of polysaccharide degradation since in fibroblasts hyaluronoglucosaminidase is absent (33, 34) and only exoglycosidases and exosulfatases are involved in dermatan sulfate catabolism (3). The time course of the degradation of endocytosed DS-PG suggests that high molecular peptidoglycans are accumulating as intermediary products. Oligosaccharides smaller than decasaccharides and inorganic sulfate are not accumulating intracellularly. The half-life (about 12 h) of the glycosaminoglycan moiety of endocytosed DS-PG is similar to that of total [3'SS]sulfatelabeled intracellular glycosaminoglycans in fibroblasts (1). Evidence for the existence of two kinetically distinct pathways as shown for DS-PG degradation in rat ovarian granulosa cells (14) has not been obtained.
An unexpected finding of this study was the observation that SH-protease inhibitors have a much more dramatic influence on the dermatan sulfate than on the core protein degradation of endocytosed DS-PG. Leupeptin has been shown by Yanagishita (35) to slow down the degradation of cell-associated heparan sulfate proteoglycans in one of the two kinetically distinct degradation pathways in rat ovarian granulosa cells. The main inhibitory effect of leupeptin, however, was considered to result from the interference with the translocation of glycosaminoglycans to the final degradation site. In Cbz-Phe-Ala-CHN*-treated fibroblasts transport of endocytosed DS-PG to dense lysosomes was apparently not delayed, and there was no indication of DS-PG accumulation in a nonlysosomal compartment which could have been caused by the drug. Several lines of evidence indicated that Cbz-Phe-Ala-CHN, did not interfere with the activity of dermatan sulfate-degrading enzymes. All enzymes known to be involved in dermatan sulfate catabolism exhibited normal activities in vitro. Furthermore, the degradation of endocytosed free dermatan sulfate chains was unaffected by the protease inhibitor. The latter observation also supports the conclusion of an unaltered endocytotic transport, albeit different routes of transport of endocytosed DS-PG and of free dermatan sulfate cannot be excluded (36).
The most plausible interpretation of our data is, therefore, the assumption that the inhibition of a critical proteolytic step leads to an inaccessibility of partially degraded DS-PG for dermatan sulfate-degrading enzymes. It would be conceivable that the dissociation of the complex between DS-PG and its putative receptor involves the action of an SH-protease. Upon inhibition of this step receptor-bound DS-PG would arrive in lysosomes, and degradation would be impaired as the proteoglycan is still associated with the lysosomal membrane. However, the kinetics of DS-PG endocytosis suggest a reutilization of DS-PG receptors (8). The lack of an inhibition of DS-PG uptake by cycloheximide treatment for 24 h is also inconsistent with this hypothesis. It seems, therefore, more likely that the very basic DS-PG core protein (PI 9.8,Ref. 20) interacts with the single polyanionic dermatan sulfate chain attached to serine 4 at the N terminus of the core protein (19) in a fashion that renders the glycosaminoglycan chains largely inaccessible for the degrading exoenzymes. This hypothesis implies that extensive proteolysis has to precede dermatan sulfate degradation and, furthermore, that the inhibition of some critical proteolytic steps may lead to a reduced glycosaminoglycan breakdown as a secondary effect.
The inhibition of dermatan sulfate degradation under the influence of Cbz-Phe-Ala-CHN, does not concern endocytosed DS-PG only. It had been shown previously that fibroblasts divert a portion of proteoglycans to the lysosomes immediately after biosynthesis (1). Upon treatment with the protease inhibitor there is a %fold intracellular accumulation of [35S]sulfate-labeled proteoglycans during a 24-h pulse experiment resembling the behavior of mucopolysaccharidosis fibroblasts.' Whether treatment with the drug could serve as a mucopolysaccharidosis model, remains to be investigated.