HEPARAN SULFATE IS A CELLULAR RECEPTOR FOR PURIFIED INFECTIOUS PRIONS

Prions replicate in the host cell by the self-propagating refolding of the normal cell surface protein, PrP(C), into a beta-sheet-rich conformer, PrP(Sc). Exposure of cells to prion-infected material and subsequent endocytosis can sometimes result in the establishment of an infected culture. However, the relevant cell surface receptors have remained unknown. We have previously shown that cellular heparan sulfates (HS) are involved in the ongoing formation of scrapie prion protein (PrP(Sc)) in chronically infected cells. Here we studied the initial steps in the internalization of prions and in the infection of cells. Purified prion "rods" are arguably the purest prion preparation available. The only proteinaceous component of rods is PrP(Sc). Mouse neuroblastoma N2a, hypothalamus GT1-1, and Chinese hamster ovary cells efficiently bound both hamster and mouse prion rods (at 4 degrees C) and internalized them (at 37 degrees C). Treating cells with bacterial heparinase III or chlorate (a general inhibitor of sulfation) strongly reduced both binding and uptake of rods, whereas chondroitinase ABC was inactive. These results suggested that the cell surface receptor of prion rods involves sulfated HS chains. Sulfated glycans inhibited both binding and uptake of rods, probably by competing with the binding of rods to cellular HS. Treatments that prevented endocytosis of rods also prevented the de novo infection of GT1-1 cells when applied during their initial exposure to prions. These results indicate that HS are an essential part of the cellular receptor used both for prion uptake and for cell infection. Cellular HS thus play a dual role in prion propagation, both as a cofactor for PrP(Sc) synthesis and as a receptor for productive prion uptake.


Summary
Prions replicate in the host cell by the selfpropagating refolding of the normal cell surface protein, PrP C , into a ß-sheet rich conformer, PrP Sc . Exposure of cells to prion-infected material and subsequent endocytosis can sometimes result in the establishment of an infected culture. However, the relevant cell surface receptors have remained unknown. We have previously shown that cellular heparan sulfates (HS) are involved in the ongoing formation of PrP Sc in chronically infected cells.
Here we studied the initial steps in the internalization of prions and in the infection of cells. Purified prion "rods" are arguably the purest prion preparation available. The only proteinaceous component of rods is PrP Sc . Mouse neuroblastoma N2a, hypothalamus GT1-1, and Chinese hamster ovary CHO cells efficiently bound both hamster and mouse prion rods (at 4º C) and internalized them (at 37º C). Treating cells with bacterial heparinase III or chlorate (a general inhibitor of sulfation) strongly reduced both rods binding and uptake, whereas chondroitinase ABC was inactive. These results suggested that the cell surface receptor of prion rods involves sulfated HS chains. Sulfated glycans inhibited both rods binding and uptake, probably by competing with the binding of rods to cellular HS. Treatments that prevented rods endocytosis also prevented the de novo infection of GT1-1 cells when applied during their initial exposure to prions. These results indicate that HS are an essential part of the cellular receptor used both for prion uptake and for cell infection. Cellular HS thus play a dual role in prion propagation, both as a co-factor for PrP Sc synthesis and as a receptor for productive prion uptake.

Introduction
The transmissible spongiform encephalopathies (TSE), which comprise infectious, familial, and sporadic neurodegenerations such as Creutzfeldt-Jakob disease (CJD) of humans (1), scrapie of sheep and bovine spongiform encephalopathy (BSE) (2), are caused by prions (3). These proteinaceous agents are thought to propagate by refolding a normal cell surface glycoprotein of the host, the cellular prion protein PrP C , into an abnorma l ß-sheet rich (4,5) conformation (reviewed in 6). The resulting pathological conformer, PrP Sc , is in turn the only known component of the infectious prion. The formation of PrP Sc is thought to involve a direct contact between "seed" PrP Sc and "substrate" PrP C (7,8) , and probably involves cellular co-factors (9) including the laminin receptors (10)(11)(12) and cellular heparan sulfate proteoglycans (HSPG) (13)(14)(15).
Although several cell lines are susceptible to prion infection (16, reviewed in 17), the molecular mechanisms involved remain largely obscure. Infection is usually started by exposing cells to prion-infected material, such as brain homogenate. Many cell types (including cell lines (18) and primary dendritic cells (19)) can internalize prion-infected material, but the cellular receptors for prions have not been identified. One factor that is likely to complicate the study of how prions enter cells is their notorious association with heterogeneous cellular membranes and aggregates (which contain in addition other cellular components). Thus, prions in crude tissue homogenates are likely to 'hitch-hike' their way into the cell using aggregates and microsomes as vehicles, via a variety of cell-surface receptors. However, prions can be extracted from membranes to yield purer preparations. Prion "rods" are arguably the purest form of prions known (20,21). These infectious, unbranched amyloidic structures are prepared from prioninfected tissues by the combined action of detergents and proteases (22) (often supplemented by nucleases) and their only proteinaceous component is PrP27- 30 (23), the protease-resistant core of PrP Sc . The size of rods is very heterogeneous, and they may contain up to several thousands PrP molecules (20). Although purified prion rods are efficiently taken up by a variety of cells and often lead to productive infections (18), even in this highly simplified situation the relevant cellular receptors have not yet been identified. Here we set out to characterize receptors for rods in two infectible mouse cell lines, the neurobla stoma N2a (24) and the hypothalamic cell line GT1-1 (25)(26)(27), and Chinese hamster ovary (CHO) cells, which seem to be refractive to prion infection.
Glycosaminoglycans ( GAGs) such as heparan sulfate (HS) are long, unbranched side chains of proteoglycans that are found in several cellular compartments including the cell surface and endosomes (reviewed in 28). Three arguments put forward GAGs as candidates for cellular receptors for prion rods. First, PrP has several heparin-binding sites both in the N -terminal unstructured region and in the PrP27-30 core (13,29,30). Second, cellular HS are required for PrP Sc formation in persistently infected mouse neuroblastoma ScN2a cells (15). Since these HS prion co-factors probably perform their task by binding the endogenous PrP isoforms, it is plausible that they can also bind PrP27-30 molecules found in exogenous rods. Another indication that GAGs may serve as prion receptors is the finding that certain soluble dextran-based heparan mimetics (HMs) (31) reduce the internalization of prion rods in both N2a and CHO cells (32). Conceivably, this inhibition could result from the competition of soluble HM molecules with putative HS receptors of rods.
To evaluate the role of cellular HS in the binding and endocytosis of prion rods, we used GAG-degrading enzymes, the sulfation inhibitor chlorate, and soluble glycans including several HM species. Our results indicate that heparinase III-sensitive HS on the cell surface are involved in both the binding and the uptake of rods in N2a, CHO and GT1-1 cells. Treatments that prevented the binding and internalization of rods also prevented the de novo infection of GT1-1 cells. Cellular HS are thus an essential component of cellular receptors for the uptake of prions and the infection of cells. Preparation of prion rods. Prion rods were purified from the brains of Syrian hamsters and of C57/bl mic e infected with experimental Sc237 and RML scrapie, respectively, u sing a procedure modified from Prusiner et al. (33) and Diringer et al. (34). All steps were performed at 4° C. One hamster brain or three mouse brains (a total of about 1g) were homogenized in 20 ml of 10% sucrose in PBS (buffer A). A 3220 g av , 10 min pellet was rehomogenized in 10 ml buffer A and then re-pelleted, and the supernatants were united and cleared by a 3220 g av , 30 min spin. The supernatant was made 1mM with each EDTA and dit hiothreitol (DTT), and Triton X -100 and Nadeoxycholate were added to 4% and 2% final concentrations, respectively. The supernatant was then stirred for 30 min and the following reagents were added dropwise while stirring: Tris acetate (pH 8.3) to 30 mM, KCl to 100 mM, glycerol to 20%, and PEG10000 to 8% w/v. After an additional 30 min stirring, the homogenate was pelleted at 4500 g av , 30 min. The pellet was resuspended in 20 mM Tris acetate, pH 8.3, 0.02% TX-100, 1 mM DTT, and 2 mM CaCl 2 , and then digested with micrococcal nuclease (12.5 units/ml) for 16 h at 4° C. The nuclease reaction was stopped by the addition of 2 mM EDTA and 0.2% Sarkosyl. The homogenate was then subjected to proteolysis with proteinase K (100 µg/ml, 8 h, 4° C), and the reaction was stopped by incubating for 30 min with 100 mM phenylmethylsulfonyl fluoride (PMSF). Sarkosyl was added to 1%, and after 30 min incubation on ice, the homogenate was spun at 100,000 g av for 1 h. The pellet, which contains the prion rods, was resuspended in 500 µl TNS (10 mM Tris pH 7.5, 150 mM NaCl, 1% Sarkosyl) using a probe sonicator (Sonopuls, Dandelin Electronics, Germany; full power, 4x1 sec) and then repelleted (3x). Sarkosyl was removed by rinsing the pellet twice with 70% EtOH, (100,000 g av , 30 min), and then resuspended by sonication in 600 µl of 30% sucrose in TN. A 5 µl sample of this rods preparation was analyzed by electrophoresis and silver staining (Fig. 1A, right lane), and compared to 10 ng recombinant mouse PrP (Fig. 1A Cells Mouse neuroblastoma ScN2a-M are ScN2a cells (24) that stably expresses the MHM2-PrP chimera that reacts with the mAb 3F4 (35). An uninfected version (N2a-M) was obtained by curing ScN2a-M cells with pentosan polysulfate (5 µg/ml, 5 d) (36) and subsequently maintaining them without inhibitors for at least 1 month prior to use. GT1-1 are mouse hypothalamus cells (27). GT1-1-M stably express MHM2-PrP and react with 3F4. Cells were grown at 37° C in low glucose DMEM-16 (N2a and GT1-1) or F12 (CHO-K1, (37)) containing 10% fetal calf serum. In some experiments, cells were maintained in a 1:1 mixture of the above media and OptiMem (Gibco BRL). To assess the susceptibility of GT1-1-MHM2 cells to infection, cells growing on 12 well trays were treated with the relevant inhibitors for 24 h . Thereafter, the cells were inoculated either with purified mouse rods for 24h or with cell supernatant of ScGT1-1 (the medium was frozen and thawed at least 3 times before addition to the cells) for 48 h at 37º C in DMEM-OptiMem (1:1) in the presence of the inhibitors. The cells were then rinsed and further grown in fresh medium (without inoculu m or treatments) for either 5 or 10 d, as indicated in Fig.  5. Since the mouse inocula are not recognized by the mAb 3F4, successful infection was identified by the appearance in the cells of 3F4-reactive PrP Sc (as depicted in Fig. 5B).

PrP isoforms and analysis
PrP Sc was defined as the PrP fraction resistant to proteinase K (PK, 20 µg/ml, 37° C, 30 min). Western blots (WB) were carried out as described (38). The protein content of paralle l samples was normalized using a Bradford kit (500-006) from Bio-Rad (Hercules, CA) prior to electrophoresis. Cell lysates (in ice-cold lysis buffer: 0.5% Triton X-100, 0.25% Na-deoxycholate, 150 mM NaCl, 10 mM Tris-Cl, pH 7.5, 10 mM EDTA), were immediately centrifuged for 40 sec at 14,000 rpm in a microfuge and biochemical analyses were performed on the post-nuclear supernatant.

Rods internalization and binding assays
were carried out on confluent monolayers grown in 24 wells trays. Cells were treated with enzymes or inhibitors for 24 h at 37° C prior to the assays, unless otherwise indicated. Rods (1µl/well, ca 5 ng of PrP, Fig. 1B) were then added either for 20 h at 37º C (internalization) or for 3 h at 4º C (binding assay; performed in PBS or in F12/OptiMem, 1:1, with identical results). The cells were then rinsed with ice-cold PBS, lysed, and the protein content was normalized using the Bradford method. The samples were incubated with 2 µg/ml PK (1 h, 37° C, stopped by 2 mM PMSF), and the level of cell-associated PrP Sc (from the exogenous rods) was assayed by WB developed with 3F4 mAb (for Syrian hamster rods) or D13 Fab (for mouse rods).

Fluorescence microscopy
For internalization experiments, cells growing on 8well slides (Nunc, (Roskilde, Denmark) were pretreated with the relevant inhibitors for 24 h, and then exposed for 20 h to purified Syrian hamster prion rods with or without inhibitors. At the end of the incubation the cells were rinsed and further incubated in normal medium for 4 h to reduce the rods signal on the cell surface. The cells were then fixed (8% formalin in PBS, 30 min , RT), denatured in situ (3 M GdnSCN, 0.1% TX-100, 50 mM Tris-HCl, pH 7.5; RT, 5 min) (18) to visualize rods, immunostained with 3F4, and examined by fluorescence microscopy.

Inhibition of rods uptake by soluble glycans correlates with their anti-prion activity
It is well established that certain glycans reduce the ongoing formation of PrP Sc in chronically infected cells. To see if cellular GAGs might be part of a cellular receptor for prions, we thus first studied the extent to which several glycans reduce the uptake of prion rods, as compared to their anti-prion efficacy. To this end we chose the "classical" anti-prion dextran sulfate (DS500) (36), and 3 members of the HM library of substituted dextrans with vastly different antiprion potencies (32) : B103, HM2102, and HM2602 (Fig. 2). Chronically infected ScN2a-M cells were treated for 5 days with these compounds, and protease-resistant PrP Sc was then analyzed by WB developed with 3F4 (panel A). As expected, the more sulfated and/or benzylaminated compounds (DS500, HM2602 and B103) reduced PrP Sc more efficiently than the non sulfated HM2102 (32). Next, we turned to determine the extent to which these glycans decrease rods uptake in uninfected N2a-M cells. The cells were pre-incubated for 24 h with the polyanions, as in panel A. Purified Syrian hamster rods were then added to the cell medium for an additional 20 h in the presence of the inhibitors. The cells were then rinsed thoroughly, and their protease-resistant PrP Sc was monitored in WBs with 3F4 (panel B). The ability of these compounds to reduce rods internalization correlated well with their anti-PrP Sc potency (compare panels A and B, and Fig. 6; we have previously reported that HMs are not general endocytosis inhibitors (32)). Interestingly, higher concentrations of inhibitors were required to reduce rods uptake than to reduce endogenous PrP Sc (see Discussion).

Anti-prion heparinase III and chlorate inhibit rods uptake
This correlation suggested that there may be a mechanistic relationship between the two anti-prion activities of these polyanions, namely that the cellular HS that are needed for the ongoing synthesis of PrP Sc (15) also form parts of internalization receptors for rods. To verify that HS are involved in rods uptake, we explored other treatments that inhibit or digest cellular HS (Fig.   3). N2a-M (panel A) and CHO-K1 (panels B and C) cells were pre-treated for 24 h with either heparinase I or heparinase III to digest away HS, or they were incubated with the metabolic sulfation i nhibitor, Na chlorate (30 mM) (39), or with soluble glycans, as indicated in panel B . Rods were added to the cell medium and were allowed to enter the cells for 20 h. The cells were then harvested and PK-resistant PrP was detected using either WB ( Fig.  3A-B) or immunofluorescence (IF, Fig. 3C). Heparinase III, Na-chlorate, dextran sulfate DS500 and HM2602 almost completely abolished rods uptake, while heparinase I was only slightly inhibitory and the inactive HM compound HM2102 failed to reduce the uptake of rods altogether. These results correlate perfectly with the anti-prion potency of these treatments ( Fig. 2 and (15)). Similar inhibitory results were obtained when mouse rods were applied to GT1-1 cells (Fig. 5A). In contrast to heparinase III, chondroitinase ABC failed to reduce rods uptake (Fig. 3B), correlating with its inability to reduce PrP Sc in chronically infected ScN2a (15). Whether this is caused by the paucity of chondroitin sulfate in these cells or by other mechanisms remains to be determined.

Heparinase III-sensitive HS mediate both binding and internalization of rods
These results strongly suggested that heparinase III-sensitive cellular HS are part of endocytosis receptors for purified prion rods in N2a, CHO-K1 and GT1-1 cells. Prions also bound to cell surface HS at 4ºC. N2a-M (not shown) or CHO-K1 cells (Fig. 4) were pre-treated for 24 h with heparinase I, heparinase III, Na-chlorate (panel A), or dextran-based polyanions (panel B). The plates were then cooled on ice (to prevent endocytosis) and further incubated with purified rods in the presence of the inhibitors (3 h on ice). At the end of the incubation, the cells were rinsed thoroughly and cell surface bound rods were monitored by WB. All the efficient anti-PrP Sc treatments also effectively reduced the binding of purified rods to the surface of these cells. Thus, only heparinase I and HM2102 left this binding intact (Fig. 4A, B).
In addition to removing HS GAGs from the cell surface, heparinase III is also likely to release HS fragments to the cell medium. By analogy with other sulfated glycans, these degradation products could also inhibit rods binding by competing with surface receptors. It was therefore necessary to verify directly that the presence of HS chains on the cell surface perform as a binding receptor for prion rods. For this purpose, we incubated CHO-K1 cells with heparinase III (0.1, 1 or 2 units/ml) for 24 h (37° C). Then the cells were rinsed with ice-cold PBS to remove any soluble heparinase products, supplemented with fresh medium without heparinase, and rods were added to all cells (3 h on ice). At the end of the incubation the cells were rinsed thoroughly and the cell surface bound rods were monitored in WB (Fig. 4C). The cells treatment with heparinase III inhibited rods binding to the cells (lane 3-4). Taken together, the results described above indicate that: (i) heparinase III-sensitive HS play a crucial role both in the binding and in the internalization of prion rods in N2a and in CHO-K1 cells, and (ii) that these HS molecules are similar or identical to those that serve as co-factors for the ongoing formation of PrP Sc in ScN2a cells.
We next asked whether endogenous HS are required for rods binding, or w hether exogenous HS can mediate binding of rods to another cellular receptor. For this purpose CHO-K1 cells were treated with chlorate for 24 h to stop the sulfation of cellular HS. Exogenous HS (50 µg/ml) were then added to the cell medium along with prion rods , the cells were further incubated for 3 h on ice, and then cell-bound rods were analyzed by WB. As shown in Fig. 4D, exogenous HS did not mediate rods binding to other cell surface receptors. These results demonstrate conclusively that cellular HS are needed for prion rods binding.

Infection of GT1-1 cells requires cellular HS during the exposure to the inoculum
Having found that cellular HS are required for the binding and uptake of prions in N2a and CHO-K1 cells, we next examined their implication for the establishment of a productive infection. To this end, we chose GT1-1 cells (27) since they are infected more readily than N2a, but this time we used mouse prions in order to avoid the hamster/mouse species barrier. Mouse PrP Sc is easily recognized by its characteristic glycoform triplet (Fig. 5A). First, we confirmed that heparinase III, chlorate and the anti-prion polyanions also prevent the uptake of rods in these cells. GT1-1 cells were exposed to mouse rods for 20 h, and the uptake of rods was monitored by WB (Fig. 5A).
Next, we examined whether HS mediate long term infection by prion rods. In order to distinguish more easily PrP Sc made de novo in the cells (which denote a productive infection) from the mouse prion inoculum, (which is endocytosed by the cells), we prepared GT1-1-MHM2 cells whic h stably express a 3F4-tagged mouse PrP. First, we exposed these cells to mouse rods (Fig.  5C, left panel). The cells were treated with either chlorate or HM2602 for 24 h, and then rods were added to the cell medium for another 24 h incubation (in the presence of the inhibitors). The cells were then rinsed, and further incubated in fresh medium (without rods or inhibitors) for 5 d to allow PrP Sc to form in the event of an infection. Cells were then proteolyzed prior to WB analysis with 3F4. While untreated cells were clearly infected by the mouse rods ( Fig. 5C, leftmost lane), the presence of both chlorate and HM2602 at the time of exposure prevented this infection.
While rods are arguably the purest preparation of PrP Sc available, cell-to-cell propagation of prions is likely to involve more complex prion containing structures, such as vesicles. Prions can propagate among GT1-1 cells in culture, in part using the medium as vehicle (27). To see if cellula r HS also play a role in the initiation of infection by this route, we thus repeated the experiment above using the cell medium of ScGT1-1 cells as a source of infectious prions, but this time the cells were exposed for 48 h and further incubated in fresh medium without inhibitors for 10 d. This medium efficiently infected GT1-1-MHM2 cells, as shown by the appearance of protease-resistant, 3F4-reactive PrP Sc . In contrast, little or no infection was established in cells that were treated with chlorate or DS500 at the time of exposure to the infected medium.
Taken together, the results of this section show that establishing an infection in GT1-1 cells requires that cellular HS be present during exposure to prio ns.

Discussion
The initial interactions of pathogens with the cellular receptors are an essential aspect of their life cycle. Prion receptors have remained unknown. We now show that cell surface HS participate in the binding and uptake of purified prion rods in N2a, CHO-K1 and GT1-1 cells, and that they are involved in the establishment of a prion infection in GT1-1 cells (Fig. 5). Purified prion rods thus join a long series of pathogens (such as herpes simplex and adenoviruses) and toxins that use HS to penetrate their target cells (see 40 for a review). Our studies reveal that unlike most pathogens, however, prions require HS both as a receptor (this work) and as a cofactor for ongoing PrP Sc formation and prion replication (in ScN2a and ScGT1-1 cells , (15)). The exact HS motifs that bind rods and additional putative molecular components of these receptors remain to be identified.

Heparan sulfate is a binding and uptake receptor for rods and is required for cell infection
Removing cellular HS with heparinase III, preventing its sulfation with chlorate, or competing with sulfated glycans prevented both the binding of rods (at 4 º C) and their internalization (at 37º C) in N2a, CHO-K1, and GT1-1 cells. The biological significance of HSmediated prion uptake is shown convincingly by the finding that chlorate or glycans, that inhibit prion binding, also prevented the infection of GT1-1 cells when applied during their exposure to either purified rods or to the conditioned medium of ScGT1-1 cells. Taken together, these results argue convincingly that HS are an essential component of prion receptors in these cells. Whether additional cellular molecules are also involved in this receptor remains to be seen. Among the few known cell surface proteins that bind PrP, the 37-kDa/67-kDa laminin receptor is a prominent candidate to take part in a prion receptor since it binds both PrP and HS (10). Because there is only one heparin-binding region in PrP27-30 (aa-110-118 (29)), our results suggests that this region is exposed on the surface of rods.

Rods receptors and prion co-factors
Our results broaden the role of cellular HS in prion biology, as these GAGs are now implicated in two aspects of prion metabolism which are not necessarily related mechanistically : (i) HS serve as co-factors in the propagation of PrP Sc (15) and (ii) they perform as receptors for both Syrian hamster Sc237 and mouse RML rods (Fig. 6). The exact sequences within HS that are involved in these two tasks remain to be determined and need not be identical in the two cases. Heparan sulfate chains have a complex structure that includes regions of hypersulfation that are separated by undersulfated spacers (reviewed in 28). Many functions of HS are performed by very specific oligosaccharide sequences within the chain. The tools used in this study were unable to address this specificity. Thus, heparinase III, which cleaves undersulfated regions, often removes the entire HS chain only to leave the tetrasaccharide linker attached to the core protein (41). Therefore, the finding that heparinase III inhibits both rods binding and PrP Sc formation does not prove that the same HS sequences are involved in both tasks. More detailed tools, such as synthetic low Mr oligosaccharides, will have to be used to determine the specific HS sequences involved in each of these two functions. Of note, these two roles may also be performed in different subcellular locations since, in contrast to cell surface binding of rods, the de novo formation of PrP Sc may take place in endocytic compartments (42).
Although there was a strict correlation between the types of exogenous glycans that inhibited the two prion-related functions of HS, higher concentrations were needed to inhibit rods uptake. Several mechanisms could contribute to this result. First, since rods are very large assemblies of PrP27-30 molecules (up to several thousands (20)), their binding to the cell surface is likely to be mediated by many parallel interactions with HS chains, which may act synergistically and thus may need higher levels of soluble glycan to be competed out. Another mechanism that could explain the greater sensitivity of the de novo formation of PrP Sc (Fig. 6, right panel) to soluble inhibitors is that the l atter process involves full length PrP molecules that contain the two Nproximal HS binding regions in addition to the one region within PrP27-30 found in rods (29). Further studies will be needed to answer this question.

Other receptors
Although this work pinpoints HS as a major receptor component for prion rods in three cell types (N2a, CHO-K1 and GT1-1), it is probable that other cell-surface molecules will be found that bind PrP27-30 in other cells. In preliminary experiments (not shown), we found that CHO pgs A-745 mutants cells (37) , that fail to initiate synthesis of GAGs on proteoglycan core proteins, do still bind and internalize purified rods (albeit less efficiently than their parent line CHO-K1). As expected, this binding was insensitive to heparinases, and is thus mediated by a non-HS molecule which we are now trying to identify. Of note, many cell surface molecules, including scavenger receptors and integrins, have been shown to bind other amyloids such as Aß (43).

Biological relevance
Whether infection of cells with prions requires the internalization of these pathogens, or if only contact is sufficient has not yet been established. The provocative findings that prions associated to steel wires (44) are infectious suggest that productive interactions between membrane-bound PrP C and exogenous PrP Sc might perhaps also occur on the surface of the target cell. In any case, our GT1-1 infection studies indicate that cellular HS are required for the establishment of an infection.
The finding that cell surface HS molecules bind PrP Sc molecules may also have important implications for the subcellular trafficking of this pathogenic isoform. Metabolic studies have shown that PrP Sc traffics from its synthesis site (either the cell surface or endosomes (4,42)) to an acid hydrolytic compartment were it looses its Nterminus (45,46) , and then back to the cell surface. The molecular details of this trafficking, such as the identity of endocytic adaptors, remain unknown. Our findings raise the possibility that transmembrane HS molecules may serve as such trafficking co-factors for both PrP isoforms.
It will be interesting to see if the initial interaction of prions with cells through HS plays a role in targeting prions to particular anatomical targets within the host. Prions strains, for instance, seem to target specific regions within the brain (47). While heparan sulfates are found in all tissues, the variations of their fine structure from cell to cell (reviewed in 48, see also [49][50][51] might conceivably contribute to the specific targeting of prion strains. Further studies will be required to see if various strains interact with different HS molecules. To study the importance of HS in the development of prion diseases, we have undertaken bioassays using transgenic mice that overexpress the mammalian heparanase (52). Preliminary results indicate that scrapie incubation time is prolonged in these mice.    or 0.1 unit /ml chondroitinase ABC, were also included. Prion rods were then added to the medium and the cells were further incubated for 20 h. Cells were then harvested and treated with PK (2 µg/ml, 30 min, 37º C) to analyze the amount of internalized prion rods. Samples were normalized by protein content and then analyzed by WB developed with the mAb 3F4. M r markers are 26, and 19 kDa. (C) Immunofluorescence. CHO-K1 cells were treated with chlorate (c), heparinase I (d), heparinase III (e), DS500 (f), HM2602 (1 µg/ml) (g), HM2102 (h) at the same concentrations used in panels A and B for 24 h, or left untreated (a,b). Rods were then added to the cell medium (b-h) for 20 h. At the end of the incubation the cells were thoroughly rinsed and chased in fresh medium for 4 h (to reduce the signal from surface-bound cells) and then hamster rods were revealed by immunofluorescence using mAb 3F4 after in situ denaturation with 3 M guanidine thiocyanate.   Because of their large dimension, rods are likely to bind cooperatively to a very large number of HS molecules. These HS receptors are similar to or identical with the cellular HS co-factors (right panel). However, whereas rods binding take place on the cell surface, PrP Sc synthesis (right panel) may also occur in intracellular compartments. (B) Digesting with heparinase or preventing sulfation with chlorate prevents both rods binding and internalization (left) and PrP Sc formation (right). (C) Identical polyanions prevent the binding of rods and the formation of PrP Sc , but a larger concentration is required to compete out the presumably more numerous binding sites of rods (left). by guest on March 25, 2020