Proteoglycan synthesis is increased in cells with impaired clathrin-dependent endocytosis

Endocytosis of membrane, ligands and fluid occurs both by clathrin-dependent and −independent mechanisms (for review see Schmid, 1992; Smythe and Warren, 1991; Lamaze and Schmid, 1995; Sandvig and van Deurs, 1994; van Deurs et al., 1989). During the last years clathrin-dependent endocytosis has been extensively characterized. Also several approaches have been used to block this process, thus facilitating the study of clathrin-independent endocytosis, the endocytic pathway used by different molecules, and the role of clathrin-dependent endocytosis for the physiology of the cells. Initially, K⁺-depletion (Moya et al., 1985) and cytosol acidification (Sandvig et al., 1987) were used to block clathrin-dependent endocytosis. More recently the 100-kDa GTPase dynamin was shown to play an important role in clathrin-dependent endocytosis (for review see Damke, 1996; De Camilli et al., 1995; Urrutia et al., 1997), and transiently and stably transfected cell lines overexpressing GTPase defective dynamin mutants are now commonly used to block this endocytic mechanism (Damke et al., 1994; Herskovits et al., 1993; van der Bliek et al., 1993). The role of endocytosis in signal transduction has recently been investigated in stably transfected HeLa cells where the overexpression of one of these mutants, dynK44A, is regulated by tetracycline. The results from these studies suggest that endocytosis is important

not only to attenuate signalling, but also for certain aspects of the signalling process itself (Ceresa and Schmid, 2000).

We have previously observed that HeLa dynK44A cells incubated with radioactive sulfate to label a modified ricin containing a sulfation site (Llorente et al., 1998) incorporate a larger amount of radioactive sulfate into high molecular mass molecules, presumably proteoglycans (PG)s, when clathrin-dependent endocytosis is impaired. However, the incorporation of sulfate into several newly synthetized proteins is unchanged. PGs are formed by addition of one or more glycosaminoglycan (GAG) chains to core proteins. The GAG chains are built of repeating dissaccharide units and are classified according to the nature of these units and by the degree and position of sulfation. PGs seem to be synthesized by all vertebrate cell types and have been found at cell surfaces, in vesicles, and in the extracellular matrix (for review see Hardingham and Fosang, 1992; Kjellén and Lindahl, 1991). These molecules have been ascribed a large variety of functions that are often mediated by electrostatic interactions of the GAG chains with other molecules such us growth factors, extracellular matrix molecules, or enzymes.

In this paper we have investigated the nature of the high molecular mass molecules giving rise to increased incorporation of radioactive sulfate upon inhibition of clathrin-dependent endocytosis. We show that the synthesis of PG core proteins is increased when clathrin-dependent endocytosis is inhibited, whereas the total protein synthesis is not affected. Protein kinase C (PKC) regulates PG synthesis in a number of cell lines (Tao et al., 1997; Fagnen et al., 1999; Thiébot et al., 1999), and this seems also to be the case in HeLa dynK44A cells. Expression of mutant dynamin and reduction of endocytosis in HeLa dynK44A cells leads to activation of PKC. Evidence is presented that PKC activation is responsible for the increased synthesis of PGs.

Tetracycline, bovine serum albumin (BSA), puromycin, guanidine, Tris-HCl, bisindolylmaleimide (BIM), phorbol 12-myristate 13-acetate (TPA) and Triton X-100 were from Sigma Chemical Co., St Louis, MO, USA. Chondroitinase ABC was purchased from Seikagaku Corp., Tokyo, Japan. Ba(NO₂)₂ was from Merck, Darmstadt, Germany. Na₂³⁵SO₄, [³²P]dCTP, Na¹²⁵I Sephadex G-50 Fine, Superose 6, and Protein A-Sepharose CL-4B were obtained from Amersham Pharmacia Biotech, Uppsala, Sweden. Geneticin was obtained from Saveen Biotech, Malmö, Sweden. [³H]Leucine was from NEN Life Science Products, Boston, MA, USA. Econo-pac high Q Cartridges were from Bio-Rad Laboratories, Hercules, CA. Acidic fibroblast growth factor (FGF) was produced in bacteria, purified on a heparin-Sepharose column (Więdlocha et al., 1996), and iodinated by the iodogen method (Fraker and Speck, 1978). Transferrin was iodinated by the same method.

The HeLa cell lines stably transformed with the cDNAs for dynWT or dynK44A were kindly provided by Dr S. L. Schmid, The Scripps Research Institute, La Jolla, CA, USA (Damke et al., 1994). The cells were grown in Falcon (Franklin Lakes, NJ, USA) or Nunc (Naperville, IL, USA) flasks and maintained in DMEM (Flow Laboratories, Irvine, Scotland) supplemented with 10% FCS, 100 units/ml penicillin, 100 μg/ml streptomycin, 2 mM glutamine, 400 μg/ml geneticin, 200 ng/ml puromycin and 1 μg/ml tetracycline. BHK21 cells which in an inducible manner produce antisense mRNA clathrin heavy chain (CHC) (BHK21-tTa/anti-CHC) (G. Skretting, unpublished) were grown in DMEM supplemented with 7.5% FCS, 100 units/ml penicillin, 100 μg/ml streptomycin, 2 mM L-glutamine, 200 μg/ml geneticin, 200 ng/ml puromycin and 2 μg/ml tetracycline. For experiments, these cells were grown with or without tetracycline for the indicated times.

Cells were washed with sulfate-free DMEM before [³⁵S]sulfate (0.1 mCi/ml) was added to the cells in sulfate-free DMEM supplemented with 5% FCS, 2 mM L-glutamine, 1% non-essential amino acids and 1 mM CaCl₂. 4 hours or 10 hours later the medium fractions (approx. 1 ml) were briefly centrifuged to remove detached cells, and an equal volume of 8 M guanidine/4% Triton X-100 in 0.2 M sodium acetate buffer, pH 6.0 was added. The cells were washed with cold PBS on ice, and then lyzed in 1 ml of 4 M guanidine/2% Triton X-100 in 0.2 M sodium acetate buffer, pH 6.0. To remove free [³⁵S]sulfate, 1 ml of each fraction was applied to 4 ml columns of Sephadex G-50 Fine in 0.05 M Tris-HCl, pH 8.0. Elution was carried out with a 0.05 M Tris-HCl, pH 8.0, buffer containing 0.15 M NaCl. The first ml eluted after application was discarded, and the next 1.5 ml was collected. The radioactivity was then analyzed in a scintillation counter. The same procedure was followed for [³H]leucine (50 μCi/ml) labelling, but using leucine-free Hepes containing 5% FCS and 2 mM L-glutamine, and for removal of free [³H]leucine. In some experiments, free [³⁵S]sulfate was removed from labelled cells by washing them with PBS and then twice with TCA (5%) at room temperature for 10 minutes. Finally, the cells were solubilized in KOH (0.1 M). The efficiency of the two methods to remove free [³⁵S]sulfate was the same.

To remove chondroitin sulfate chains from core proteins, samples from medium and cell fractions pooled after Sephadex G-50 Fine chromatography (approx. 10,000 cpm) were treated with chondroitinase ABC (0.01 units) in 0.05 M Tris-HCl, pH 8.0, containing 0.05 M sodium acetate and 0.1% BSA at 37°C overnight. To break down heparan sulfate chains, equal volumes of 0.5 M H₂SO₄ and 0.5 M Ba(NO₂)₂ were mixed. After centrifugation, 15 μl of the supernatant was added to approx. 10,000 cpm of sample (Shively and Conrad, 1976). 10 minutes later the reaction was stopped by adding 1 M Tris-HCl (pH ≥5).

After removal of free [³⁵S]sulfate or [³H]leucine, a polyclonal antibody against mouse perlecan (a gift from Dr J. R. Hassell, University of Pittsburg, PA, USA), a monoclonal antibody against human versican (Seikagaku Corp., Tokyo, Japan), or a monoclonal antibody against human syndecan (a gift from Dr M. Jalkanen, Center for Biotechnology, Turku, Finland) (5 μg/ml) was incubated with samples of medium and cell lysates overnight at 4°C in the presence of 5 mM sulfate or 5 mM leucine. Protein A-Sepharose preblocked with PBS containing 1% BSA was then added to the samples. After 2 hours at 4°C the beads were washed 5 times in 50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 0.05% Triton X-100, 5 mM MgSO₄ solution with 1% BSA, and 3 times without BSA. The adsorbed material was analyzed by SDS-PAGE.

Sephadex G-50 fine columns were used to remove free [³⁵S]sulfate from the medium and the cellular fractions. Then, similar volumes of these fractions were treated with 0.5 M NaOH overnight at room temperature to release GAG chains from core proteins. The reaction was stopped by adding HCl until the pH was between 7 and 8. The fractions were then subjected to Superose 6 gel chromatography with a Pharmacia FPLC pump system, and eluted with 0.15 M NaCl in 0.05 M Tris-HCl with 0.5% Triton X-100, pH 8.0. Fractions (0.5 ml) were collected with a flow rate of 0.5 ml/minute, and the radioactivity was analyzed in a scintillation counter.

After removal of free [³⁵S]sulfate or [³H]leucine, medium and cellular samples (more than 15,000 cpm in 0.7 ml) were subjected to Econo-pac high capacity ion-exchange columns (5 ml) connected to a Bio-Rad Econo System. Fractions (1 ml) were collected with a flow rate of 2 ml/minute. Elution was carried out with a NaCl gradient (0.15-1.5 M) in 0.05 M Tris-HCl. The radioactivity in the fractions was analyzed in a scintillation counter.

mRNA was isolated using the Dynabeads mRNA Direct Kit (Dynal, Oslo, Norway), run on a 1% agarose gel with formaldehyde, and finally blotted onto a Hybond-N membrane (Amersham Pharmacia Biotech). cDNA probes for syndecan-1 (kindly provided by Dr Lars Uhlin-Hansen, University of Tromsø, Norway) and glyceraldehyde-3-phosphate dehydrogenase (Clonetech Laboratories, CA, USA) were labelled with Redivue [α-³²P]dCTP using Rediprime™ II (Amersham Pharmacia Biotech). Hybridization was carried out overnight at 65°C.

Cells were incubated for 2 hours on ice in Hepes medium containing ¹²⁵I-aFGF (50 ng/ml). In some cases heparin (10 U/ml) was added. The cells were then washed 3 times with ice-cold Hepes and lyzed in 0.1 M KOH. The radioactivity in the lysate was measured.

Endocytosis of ¹²⁵I-transferrin was measured as described previously (Ciechanover et al., 1983). In principle, cells were incubated with ¹²⁵I-transferrin (50-150 ng/ml; 10,000-20,000 cpm/ng) at 37°C for 5 minutes, washed 3 times with ice-cold Hepes, and treated for 1 hour at 0°C with Hepes containing 0.3% (w/v) pronase. The cells and the medium were then transferred to Eppendorf tubes and centrifuged for 2 minutes. The radioactivity in the pellet and in the supernatant was measured.

The content of cAMP in cells was measured with a [8-³H] cyclic AMP assay system from Amersham Pharmacia Biotech. In principle, cells were washed twice in PBS and then dissolved in ice-cold HCl (10 mM) in ethanol (96%). After a 5 minutes incubation at 0°C, the cells were removed with a rubber policeman and the cell suspension was centrifuged for 10 minutes in an Eppendorf centrifuge. The supernatant was freeze-dried and the pellet was dissolved in KOH (0.2 M). The OD (280 nm) of this solution was used as a measurement of the amount of cells used. The freeze-dried supernatant was dissolved in 250 μl Na-acetate (0.5 M, pH 6.2). This solution was then used in the cAMP kit to measure the concentration of cAMP according to the manufacturer’s instructions.

PKC and protein kinase A (PKA) activities were assayed according to GibcoBRL instructions.

SDS-PAGE was performed as described (Laemmli, 1970). Labelled PGs were run in precast 4-20% Tris-HCl-glycine gels (Novex, San Diego, CA, USA). The gels were then fixed in 4% acetic acid and 27% methanol for at least 30 minutes, and then treated with 1 M Na-salicylate, pH 5.8 in 2% glycerol for 15-30 minutes. Dried gels were exposed for fluorography at −80°C.

Removal of tetracycline from the growth medium of HeLa dynK44A cells induces the overexpression of dynK44A, a dynamin I molecule where lysine 44 has been substituted by alanine (Damke et al., 1994). DynK44A is defective in GTP binding and hydrolysis, and causes a selective block in clathrin-dependent endocytosis, whereas clathrin-independent endocytosis continues (Damke et al., 1994; Vieira et al., 1996; Llorente et al., 1998). Initial experiments revealed a difference in the level of [³⁵S]sulfate incorporation into high molecular mass molecules between HeLa dynK44A cells grown in the presence or in the absence of tetracycline. To quantify this difference, cells were metabolically labelled with [³⁵S]sulfate for 10 hours, and free sulfate was then removed from cell lysates and medium fractions by Sephadex G-50 Fine chromatography. Interestingly, the amount of [³⁵S]sulfate incorporated into both cellular and secreted molecules in cells overexpressing mutant dynamin was increased (Fig. 1). The largest increase was observed in the secreted molecules (Fig. 1). Furthermore, control experiments showed that the amount of incorporated [³⁵S]sulfate was unchanged by overexpression of wild-type dynamin I in HeLa dynWT cells (Fig. 1). The experiment shown in Fig. 1 was performed with cells grown with or without tetracycline for 48 hours. High levels of expressed dynK44A are obtained after this period of time (Damke et al., 1995). However, dynK44A expression is already detected 4 hours after tetracycline removal (Damke et al., 1995), and clathrin-dependent endocytosis is strongly inhibited 24 hours after removal of tetracycline, and it remains inhibited after 48 and 72 hours (Fig. 2A). We therefore decided to investigate the [³⁵S]sulfate incorporation into macromolecules in cells grown with or without tetracycline for shorter and longer periods of time than 48 hours. As soon as 24 hours after removal of tetracycline a small increase in cells with mutant dynamin was observed (Fig. 2B, open bars). The increase was more marked 48 hours after removal of tetracycline, and it was maintained at the same level after 72 hours (Fig. 2B, open bars). Control experiments showed that the differences in [³⁵S]sulfate incorporation were not due to differences in cell numbers between cells grown with or without tetracycline for 24, 48 and 72 hours (Fig. 2B, hatched bars).

As already mentioned, dynK44A overexpression in HeLa dynK44A cells causes a selective block in clathrin-dependent endocytosis. However, it has been shown that dynK44A overexpression may also alter other intracellular traffic steps (Llorente et al., 1998; Nicoziani et al., 2000). To investigate whether the effect of dynK44A overexpression on macromolecules [³⁵S]sulfate incorporation was in fact due to a block in clathrin-dependent endocytosis, similar experiments were performed with BHK21-tTa/anti-CHC cells. In a tetracycline-inducible manner, these cells produce CHC antisense mRNA, which inactivates endogenous mRNA and prevents the synthesis of CHC. This results in an inhibition of clathrin-dependent endocytosis, but not of fluid-phase endocytosis. Also inhibition of clathrin-dependent endocytosis in these cells caused an increased incorporation of [³⁵S]sulfate into cellular and released high molecular mass molecules (Fig. 1), thus suggesting that the effect of dynK44A is due to a block in clathrin-dependent endocytosis.

Sulfate groups can be linked to tyrosine residues in some proteins, and to sugars in glycoproteins, glycolipids, and free or core protein-linked GAG chains. In HeLa dynK44A cells, [³⁵S]sulfated-macromolecules mainly appear as smears on the top of 4-20% gradient gels (Fig. 3), thus suggesting that most of the [³⁵S]sulfate is incorporated into the GAG chains of PGs. To investigate the nature of the GAG chains, [³⁵S]sulfate-labelled cellular and medium fractions were subjected to depolymerization treatments before analysis by SDS-PAGE. As shown in Fig. 3, the [³⁵S]sulfate-labelled macromolecules disappear almost completely from the cellular fraction upon treatment with nitrous acid, which degrades HSPGs and heparin (Shively and Conrad, 1976). This result indicates that the cellular fraction contains mainly heparan sulfate PGs (HSPG)s. However, the medium fraction contains mainly chondroitin sulfate PGs (CSPG)s, as suggested by the disappearance of [³⁵S]sulfate-labelled PGs after chondroitinase ABC treatment (Fig. 3).

To further characterize the PGs present in HeLa dynK44A cells, specific PGs were immunoprecipitated. For this purpose, [³⁵S]sulfate-labelled fractions were incubated with antibodies against the HSPG perlecan, the CSPG versican, and syndecan-1, an HS/CS hybrid PG. Both when perlecan and syndecan immunoprecipated from the cellular fraction were run on SDS-PAGE, a high molecular mass band was visible (Fig. 4, see fraction C). Furthermore, the bands were stronger when dynK44A expression was induced (Fig. 4, lanes without tet). No bands were visible when medium fractions were immunoprecipitated with these antibodies (Fig. 4, see M). Finally, since the PGs found in the medium fraction are CSPG (Fig. 3), we considered the possibility that versican was one of them. However, no bands were detected on SDS-PAGE gels when [³⁵S]sulfate-labelled medium fractions were immunoprecipitated with an antibody against versican (data not shown).

Since sulfate is mainly incorporated into sugar units, the increased [³⁵S]sulfate incorporation in cells overexpressing dynK44A could be due to alterations in the GAG length and/or negative charge density (higher concentration of sulfate groups) of PGs in these cells. To determine the size of the GAG chains, [³⁵S]sulfate-labelled cellular and medium fractions were treated with NaOH to release GAG chains from core proteins, and then analyzed by Superose 6 FPLC chromatography. Only minor differences in the GAG hydrodynamic volume (K_av) between cells with endogenous dynamin and cells expressing dynK44A were found (data not shown), thus suggesting that in both cases the GAG chains have the same length.

To investigate if there are any differences in the density of negative charge of the GAG chains from control cells and cells expressing mutant dynamin, [³⁵S]sulfate-labelled cellular and medium samples were analyzed by ion-exchange chromatography. If there is a difference in the GAG density of charge, the concentrations of NaCl required to elute the samples would be different (Safaiyan et al., 1999). Fig. 5 shows the elution pattern of medium and cellular samples labelled with [³⁵S]sulfate 2 days after tetracycline removal. After the salt gradient starts (fraction 12) two peaks are observed: one presumably containing sulfated proteins (around fraction 19), and the other containing PGs (around fraction 36). The PGs of the cellular and medium fractions of cells grown with or without tetracycline were eluted with the same concentration of NaCl, thus suggesting that they have similar densities of negative charge. Furthermore, a large amount of the total incorporated radioactive sulfate was found in PGs both in cells grown with and without tetracycline.

In an attempt to investigate whether the increase in sulfate incorporation observed after dynK44A overexpression was due to an increased synthesis of PG core proteins, HeLa dynK44A grown with or without tetracycline for 48 hours were metabolically labelled with [³H]leucine for 10 hours. After removal of free [³H]leucine, similar amounts of [³H]leucine-labelled proteins were detected in both cellular and medium fractions from HeLa cells with endogenous or mutant dynamin (Fig. 6A). However, larger amounts of perlecan were immunoprecipitated from cellular [³H]leucine labelled fractions of cells containing mutant dynamin (Fig. 6B).

The possibility that the increase in sulfate incorporation in cells overexpressing dynK44A was due to an increased synthesis of PG core proteins was further investigated by northern blot analysis of the mRNA levels of syndecan-1. A glyceraldehyde-3-phosphate dehydrogenase probe was used to control the amount of mRNA applied in the different lanes. Interestingly, cells grown without tetracycline for 48 hours showed approx. 3-fold increased levels of syndecan-1 mRNA (Fig. 7). This experiment, as the [³H]leucine-labelled perlecan immunoprecipitation experiment (Fig. 6B), suggests that in cells where clathrin-dependent endocytosis has been inhibited by overexpressing dynK44A, the increased incorporation of radioactive sulfate into macromolecules can be explained by an increased synthesis of PG core proteins.

PGs bind to several growth factors (Ruoslahti and Yamaguchi, 1991). It has been shown that HSPGs, the PGs found in the cellular fraction of HeLa dynK44A (Fig. 3), bind FGFs (Burgess and Maciag, 1989). Since cells overexpressing mutant dynamin synthesize more PGs, we tested the possibility that the binding of the acidic FGF was increased. When HeLa dynK44A cells grown with or without tetracycline for 48 hours were incubated with ¹²⁵I-acidic FGF for 2 hours at 4°C, larger amounts of acidic FGF bound to the plasma membrane of cells overexpressing mutant dynamin (Fig. 8). Heparin strongly inhibited acidic FGF binding to the cells (data not shown). Furthermore, no significant difference in binding was observed when cells were grown for 24 hours with or without tetracycline (Fig. 8). This result is in agreement with the observation that only minor differences in sulfate incorporation into macromolecules in cells with or without mutant dynamin are observed after 24 hours (Fig. 2B).

Activation of PKC by phorbol esters leads to an increased PG synthesis in several cell lines (Tao et al., 1997; Fagnen et al., 1999; Thiébot et al., 1999). In particular, TPA has been shown to upregulate the expression of syndecan in rat immature Sertoli cells (Brucato et al., 2000), and of perlecan in the erythroleukemia cell line K562 (Grassel et al., 1995). We found that addition of TPA (0.1 μM) increased the incorporation of [³⁵S]sulfate in HeLa dynK44A cells (data not shown), thus suggesting that also in these cells activation of PKC leads to an increased PG synthesis. We therefore decided to measure the levels of activated PKC in cells with or without mutant dynamin. Interestingly, cells grown without tetracycline for 48 hours showed higher levels of activated PKC than cells grown with tetracycline (Fig. 9A). Since differences in PKC activity may be responsible for the differences in PG synthesis between cells with or without dynK44A, we measured the sulfate incorporation into PGs in the presence of BIM, a PKC inhibitor (Toullec et al., 1991). BIM (10 μM) was added 24 hours after removal of tetracycline, when the activity of PKC is only partially stimulated (Fig. 9A). 24 hours later the incorporation of radioactive sulfate into macromolecules was measured. As shown in Fig. 9B, the differences between cells with or without mutant dynamin were reduced by 50%. These results clearly indicate that the activity of PKC is important for the level of PG synthesis. Finally, we could not detect changes in the activity of PKA or the levels of cAMP when clathrin-dependent endocytosis was inhibited (data not shown).

The main finding in the present work is that overexpression of the dynamin mutant dynK44A and the subsequent inhibition of clathrin-dependent endocytosis leads to a selective increase in PG synthesis, probably due to activation of PKC.

The increased incorporation of [³⁵S]sulfate into cellular and secreted PGs of cells with impaired clathrin-dependent endocytosis was not due to differences in cell growth. Moreover, it was shown by gel filtration and ion-exchange chromatography that cells with or without dynK44A have PGs with GAG chains of similar size and density of negative charges, thus suggesting that the core proteins of PGs in cells with endogenous dynamin II or expressing the mutant of dynamin I, dynK44A, are modified in an identical manner with respect to these criteria in the Golgi apparatus.

Characterization of the PGs present in HeLa dynK44A by immunoprecipitation of lyzed [³⁵S]sulfate-labelled cells revealed the presence of perlecan and syndecan in the cellular fraction but not in the secreted fraction. Indeed, although syndecan has been found in the medium (Jalkanen et al., 1987), it is usually present as a plasma membrane PG (for review see Carey, 1997). Concerning perlecan, it is found both in the cellular fraction and in the medium of MDCK cells (Svennevig et al., 1995), whereas in colon carcinoma cells it is closely associated with the plasma membrane (Iozzo et al., 1994). The fact that in HeLa dynK44A cells perlecan was not immunoprecipitated from the medium fraction is in agreement with our results that HSPGs are not found in the medium of HeLa dynK44A cells.

Interestingly, our results suggest that the increased amount of radioactive sulfate found in cells overexpressing mutant dynamin may be due to a selective increase in the synthesis of PG core proteins. Higher amounts of [³H]leucine-labelled perlecan were immunoprecipitated from cells containing mutant dynamin, although the overall protein synthesis was unchanged. Furthermore, the mRNA level of the PG syndecan-1 was higher in cells overexpressing mutant dynamin. In contrast, the mRNA level of glyceraldehyde-3-phosphate dehydrogenase was not increased, thus indicating that inhibition of clathrin-dependent endocytosis does not increase the levels of mRNA expression in general.

Dynamin II has been localized on the plasma membrane (Damke et al., 1994) and at the Golgi apparatus (Cao et al., 1998), and in vitro studies show that dynamin seems to be required for the formation of clathrin-coated and constitutive secretory vesicles from the TGN (Jones et al., 1998). Furthermore, a dynamin-related protein in yeast, Vsp1p, has been implicated in trafficking from the Golgi apparatus to the vacuole (Rothman et al., 1990). Overexpression of mutant dynamin I clearly changes the distribution of the mannose-6-phosphate receptor intracellularly, suggesting a role for dynamin or a dynamin-like molecule in intracellular transport (Llorente et al., 1998; Nicoziani et al., 2000). Furthermore, we have previously shown that in HeLa dynK44A cells overexpressing mutant dynamin not only clathrin-dependent endocytosis, but also transport of internalized ricin to the Golgi apparatus is inhibited (Llorente et al., 1998). However, experiments performed in BHK21-tTa/anti-CHC cells where clathrin-dependent endocytosis had been inhibited supported the idea that the increase in the synthesis of PG core proteins results from a block in clathrin-dependent endocytosis. Although both dynamin and clathrin affect uptake from clathrin-coated pits at the cell surface, they do not always act together. For instance, dynamin but not clathrin is important for caveolae function (Henley et al., 1998).

Why do cells synthesize more PGs when clathrin-dependent endocytosis is inhibited? It has recently been shown that clathrin-dependent endocytosis is necessary for the activation of kinases such as phosphatidyl inositol 3-kinase and ERK 1/2 (Vieira et al., 1996; Ceresa et al., 1998). However, clathrin-dependent endocytosis can also serve as a mechanism for signal transduction attenuation since it removes receptors from the cell surface (Wells et al., 1990; Vieira et al., 1996). PKC has been shown to be involved in PG synthesis (Tao et al., 1997; Fagnen et al., 1999; Thiébot et al., 1999), and inhibition of endocytosis and degradation of activated receptors could result in an increased and prolonged activity of this enzyme. We find that the PKC activity is increased when clathrin-dependent endocytosis is inhibited, and that PKC also seems to regulate PG synthesis in HeLa dynK44A cells. Moreover, in the presence of a PKC inhibitor the differences in sulfate incorporation between cells with and without mutant dynamin are reduced. Therefore, we suggest that the continuous presence at the cell surface of activated receptors signalling through the PKC pathway, leads to an increase in PKC activity that causes the stimulation of PG core protein synthesis. Growth factors present in the serum may be potential ligands for these receptors since it is known that these molecules regulate PG expression (Elenius et al., 1992).

Phospholipase Cγ, an enzyme that produces the PKC activator diacylglycerol, is hyperphosphorylated in cells overexpressing dynK44A (Vieira et al., 1996). However, phospholipase Cγ is probably not responsible for the activation of PKC here shown since it has been recently reported that its activity is not increased in these cells in spite of the change in phosphorylation (Ringerike et al., 1998).

PGs are involved in the binding and internalization of several growth factors (Ruoslahti and Yamaguchi, 1991). Therefore, the increased PG synthesis in the cellular fraction observed in cells overexpressing dynK44A was likely to result in an increased binding of growth factors to the cell surface. As here shown, more acidic FGF was bound to cells overexpressing mutant dynamin than to cells with endogenous dynamin. Since HeLa dynK44A cells expressing mutant dynamin are used to study endocytosis and its relevance for signalling of different ligands, it is important for the interpretation of such results to be aware of the change in PG expression. A shift in endocytosis or a shift in response to a certain ligand is not necessarily caused by the inhibition of clathrin-dependent endocytosis of the ligand. PGs can be important for signalling (Spivak-Kroizman et al., 1994; Rapraeger et al., 1994), and a change in the subset of these molecules might also affect the pathway of endocytosis followed by a ligand.

We thank Jorunn Jacobsen, Anne-Grethe Myrann, Mette Sværen and Klaus Magnus Johansen for expert technical assistance. This work was supported by the Norwegian Research Council for Science and the Humanities (NAVF), The Norwegian Cancer Society, the Novo-Nordisk Foundation, Blix legacy, Torsteds legacy, the Jahre foundation, Jeanette and Søren Bothners legacy. Mieke Sprangers was a Leonardo programme student from the University of Professional Education in Rotterdam.