Portal de Periódicos da CAPES

Identification of interaction domains of the prion protein with its 37-kDa/67-kDa laminin receptor

2001; Springer Nature; Volume: 20; Issue: 21 Linguagem: Inglês

10.1093/emboj/20.21.5876

1460-2075

C. Hundt,

RNA regulation and disease

Article1 November 2001free access Identification of interaction domains of the prion protein with its 37-kDa/67-kDa laminin receptor Christoph Hundt Christoph Hundt Laboratorium für Molekulare Biologie-Genzentrum-Institut für Biochemie der LMU München, Feodor-Lynen Strasse 25, D-81377 Munich, Germany CEA, Service de Neurovirologie, DRM/DSV, CRSSA, 18, Route du Panorama, BP.6, F-92265 Fontenay-aux-Roses Cedex, France Search for more papers by this author Jean-Michel Peyrin Jean-Michel Peyrin CEA, Service de Neurovirologie, DRM/DSV, CRSSA, 18, Route du Panorama, BP.6, F-92265 Fontenay-aux-Roses Cedex, France Search for more papers by this author Stéphane Haïk Stéphane Haïk CEA, Service de Neurovirologie, DRM/DSV, CRSSA, 18, Route du Panorama, BP.6, F-92265 Fontenay-aux-Roses Cedex, France Search for more papers by this author Sabine Gauczynski Sabine Gauczynski Laboratorium für Molekulare Biologie-Genzentrum-Institut für Biochemie der LMU München, Feodor-Lynen Strasse 25, D-81377 Munich, Germany Search for more papers by this author Christoph Leucht Christoph Leucht Laboratorium für Molekulare Biologie-Genzentrum-Institut für Biochemie der LMU München, Feodor-Lynen Strasse 25, D-81377 Munich, Germany Search for more papers by this author Roman Rieger Roman Rieger Laboratorium für Molekulare Biologie-Genzentrum-Institut für Biochemie der LMU München, Feodor-Lynen Strasse 25, D-81377 Munich, Germany Search for more papers by this author Maria Louise Riley Maria Louise Riley Laboratorium für Molekulare Biologie-Genzentrum-Institut für Biochemie der LMU München, Feodor-Lynen Strasse 25, D-81377 Munich, Germany Search for more papers by this author Jean-Philippe Deslys Jean-Philippe Deslys CEA, Service de Neurovirologie, DRM/DSV, CRSSA, 18, Route du Panorama, BP.6, F-92265 Fontenay-aux-Roses Cedex, France Search for more papers by this author Dominique Dormont Dominique Dormont CEA, Service de Neurovirologie, DRM/DSV, CRSSA, 18, Route du Panorama, BP.6, F-92265 Fontenay-aux-Roses Cedex, France Search for more papers by this author Corinne Ida Lasmézas Corresponding Author Corinne Ida Lasmézas CEA, Service de Neurovirologie, DRM/DSV, CRSSA, 18, Route du Panorama, BP.6, F-92265 Fontenay-aux-Roses Cedex, France Search for more papers by this author Stefan Weiss Corresponding Author Stefan Weiss Laboratorium für Molekulare Biologie-Genzentrum-Institut für Biochemie der LMU München, Feodor-Lynen Strasse 25, D-81377 Munich, Germany Search for more papers by this author Christoph Hundt Christoph Hundt Laboratorium für Molekulare Biologie-Genzentrum-Institut für Biochemie der LMU München, Feodor-Lynen Strasse 25, D-81377 Munich, Germany CEA, Service de Neurovirologie, DRM/DSV, CRSSA, 18, Route du Panorama, BP.6, F-92265 Fontenay-aux-Roses Cedex, France Search for more papers by this author Jean-Michel Peyrin Jean-Michel Peyrin CEA, Service de Neurovirologie, DRM/DSV, CRSSA, 18, Route du Panorama, BP.6, F-92265 Fontenay-aux-Roses Cedex, France Search for more papers by this author Stéphane Haïk Stéphane Haïk CEA, Service de Neurovirologie, DRM/DSV, CRSSA, 18, Route du Panorama, BP.6, F-92265 Fontenay-aux-Roses Cedex, France Search for more papers by this author Sabine Gauczynski Sabine Gauczynski Laboratorium für Molekulare Biologie-Genzentrum-Institut für Biochemie der LMU München, Feodor-Lynen Strasse 25, D-81377 Munich, Germany Search for more papers by this author Christoph Leucht Christoph Leucht Laboratorium für Molekulare Biologie-Genzentrum-Institut für Biochemie der LMU München, Feodor-Lynen Strasse 25, D-81377 Munich, Germany Search for more papers by this author Roman Rieger Roman Rieger Laboratorium für Molekulare Biologie-Genzentrum-Institut für Biochemie der LMU München, Feodor-Lynen Strasse 25, D-81377 Munich, Germany Search for more papers by this author Maria Louise Riley Maria Louise Riley Laboratorium für Molekulare Biologie-Genzentrum-Institut für Biochemie der LMU München, Feodor-Lynen Strasse 25, D-81377 Munich, Germany Search for more papers by this author Jean-Philippe Deslys Jean-Philippe Deslys CEA, Service de Neurovirologie, DRM/DSV, CRSSA, 18, Route du Panorama, BP.6, F-92265 Fontenay-aux-Roses Cedex, France Search for more papers by this author Dominique Dormont Dominique Dormont CEA, Service de Neurovirologie, DRM/DSV, CRSSA, 18, Route du Panorama, BP.6, F-92265 Fontenay-aux-Roses Cedex, France Search for more papers by this author Corinne Ida Lasmézas Corresponding Author Corinne Ida Lasmézas CEA, Service de Neurovirologie, DRM/DSV, CRSSA, 18, Route du Panorama, BP.6, F-92265 Fontenay-aux-Roses Cedex, France Search for more papers by this author Stefan Weiss Corresponding Author Stefan Weiss Laboratorium für Molekulare Biologie-Genzentrum-Institut für Biochemie der LMU München, Feodor-Lynen Strasse 25, D-81377 Munich, Germany Search for more papers by this author Author Information Christoph Hundt1,2, Jean-Michel Peyrin2, Stéphane Haïk2, Sabine Gauczynski1, Christoph Leucht1, Roman Rieger1, Maria Louise Riley1, Jean-Philippe Deslys2, Dominique Dormont2, Corinne Ida Lasmézas 2 and Stefan Weiss 1 1Laboratorium für Molekulare Biologie-Genzentrum-Institut für Biochemie der LMU München, Feodor-Lynen Strasse 25, D-81377 Munich, Germany 2CEA, Service de Neurovirologie, DRM/DSV, CRSSA, 18, Route du Panorama, BP.6, F-92265 Fontenay-aux-Roses Cedex, France ‡C.Hundt, J.-M.Peyrin and S.Haïk contributed equally to this work *Corresponding authors. E-mail: [email protected] or E-mail: [email protected] The EMBO Journal (2001)20:5876-5886https://doi.org/10.1093/emboj/20.21.5876 C.I.Lasmézas and S.Weiss should be considered as the senior authors of this work PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info Cell-binding and internalization studies on neuronal and non-neuronal cells have demonstrated that the 37-kDa/67-kDa laminin receptor (LRP/LR) acts as the receptor for the cellular prion protein (PrP). Here we identify direct and heparan sulfate proteoglycan (HSPG)-dependent interaction sites mediating the binding of the cellular PrP to its receptor, which we demonstrated in vitro on recombinant proteins. Mapping analyses in the yeast two-hybrid system and cell-binding assays identified PrPLRPbd1 [amino acids (aa) 144–179] as a direct and PrPLRPbd2 (aa 53–93) as an indirect HSPG-dependent laminin receptor precursor (LRP)-binding site on PrP. The yeast two-hybrid system localized the direct PrP-binding domain on LRP between aa 161 and 179. Expression of an LRP mutant lacking the direct PrP-binding domain in wild-type and mutant HSPG-deficient Chinese hamster ovary cells by the Semliki Forest virus system demonstrates a second HSPG-dependent PrP-binding site on LRP. Considering the absence of LRP homodimerization and the direct and indirect LRP–PrP interaction sites, we propose a comprehensive model for the LRP–PrP–HSPG complex. Introduction We recently identified the 37-kDa laminin receptor precursor (LRP) as an interactor for the prion protein (PrP) (Rieger et al., 1997; for reviews see Rieger et al., 1999; Gauczynski et al., 2001a). Employing a series of neuronal and non-neuronal cells, we proved that the 37-kDa LRP/67-kDa high-affinity laminin receptor (LR) acts as the receptor for the cellular PrP (Gauczynski et al., 2001b). In the present manuscript we used the yeast two-hybrid system and cell-binding studies on neuronal as well as non-neuronal cells involving the Semliki Forest virus (SFV) system (for reviews see Liljestrom and Garoff, 1991; Tubulekas et al., 1997) to identify domains on the PrP and the LRP involved in the PrP–LRP interaction on the cell surface. We identified two binding domains for LRP on PrP termed PrPLRPbd1 and PrPLRPbd2. The first one binds directly to LRP, whereas the second one depends on the presence of heparan sulfate proteoglycans (HSPGs) on the cell surface. The yeast two-hybrid system and cell-binding assays on wild-type and mutant HSPG-deficient Chinese hamster ovary (CHO) cells also identified two binding domains for PrP on LRP. The relationship between 37-kDa LRP and 67-kDa LR is not yet fully understood and has been explained with homodimerization of 37-kDa LRP (Landowski et al., 1995) or an additional factor, such as a polypeptide (Castronovo et al., 1991), which might bind to 37-kDa LRP to form the 67-kDa form of the receptor. The 67-kDa heterodimer might be stabilized by hydrophobic interactions mediated by fatty acids such as palmitate, oleate and stearate bound to 37-kDa LRP and to a galectin-3 (gal-3) cross reacting polypeptide (Landowski et al., 1995; Buto et al., 1998). However, we recently proved that the β-galactoside lectin gal-3 is not present on the surface of neuronal or non-neuronal cells used for PrP-binding/internalization studies (Gauczynski et al., 2001b) and anti-gal-3 antibodies failed to compete for the 37-kDa LRP/67-kDa LR-mediated binding and internalization of the cellular PrP (Gauczynski et al., 2001b), suggesting that gal-3 is not a partner of the 37-kDa LRP in this context. In this study we investigated by a yeast two-hybrid system analysis whether gal-3 interacts with 37-kDa LRP and/or the cellular PrP. In addition, we investigated whether 37-kDa LRP interacts with itself in the yeast two-hybrid and analysed the monomer/dimer status of the receptor by size-exclusion chromatography. Both PrP (Gabizon et al., 1993; Caughey et al., 1994; Chen,S.G. et al., 1995; Brimacombe et al., 1999) and the 37-kDa/67-kDa LR (Guo et al., 1992; Kazmin et al., 2000) bind to heparan sulfates. HSPGs are required for the binding of the fibroblast growth factor (FGF) to its FGFR receptor (Yayon et al., 1991; Spivak et al., 1994; Venkataraman et al., 1999) and act as initial attachment receptors for bacteria (Chen,T. et al., 1995) and viruses including alphaviruses (Byrnes and Griffin, 1998), human immunodeficiency virus (HIV) type 1 (Mondor et al., 1998) and vaccinia virus (Chung et al., 1998). Heparan sulfates are components of amyloid plaques in prion diseases (Gabizon et al., 1993). We investigated the role of HSPGs as possible co-factors for 37-kDa LRP mediating PrP binding. We also constructed recombinant (rec.) SFV vectors leading to the expression of an LRP mutant termed LRPdelBD::FLAG lacking the direct-binding domain for PrP in wild-type and mutant HSPG-deficient CHO cells. We compared the PrP-binding capacity of these cells with wild-type and mutant CHO cells hyperexpressing wild-type LRP::FLAG. In light of our findings that 37-kDa LRP fails to form homodimers, and that HSPGs mediate the binding of PrP to 37-kDa LRP, the relationship between 37-kDa LRP and 67-kDa LR might be explained by the association of LRP with HSPGs as outlined in a proposed model for the LRP–PrP–HSPG complex on the cell surface. Results Identification of a direct LRP interaction domain on PrP by the yeast two-hybrid system To determine the domains of PrP interacting directly with LRP, we employed a yeast two-hybrid analysis with truncated PrP molecules in the bait position and LRP44–295 in the prey position. Only truncated PrP retaining the regions amino acids (aa) 144–179 (Figure 1A, rows 6 and 7) interacted with LRP. This region contains domains corresponding to the first α-helix (aa 144–154), the second β-strand (aa 161–164) and the first amino acid of the second α-helix (aa 179–193) of the human PrP (Zahn et al., 2000). Regions from aa 23 to 143 of the human PrP are not sufficient for binding to LRP (Figure 1A, rows 1–5). Regions from aa 180 to 230 of human PrP (row 8) are not required for the direct interaction between PrP and LRP. We termed this LRP interaction domain on PrP PrPLRPbd1. Figure 1.Identification of direct PrP–LRP-interaction domains. (A) Identification of the direct PrP–LRP interaction domain on PrP. HuPrP23–93 (row 1), huPrP23–118 (row 2), huPrP23–127 (row 3), huPrP23–131 (row 4), huPrP23–143 (row 5), huPrP23–154 (row 6), huPrP23–181 (row 7) and huPrP180–230 (row 8) were co-expressed in fusion with GST in the bait position together with LRP in prey position of the yeast two-hybrid system. (B) Retrenchment of the direct PrP–LRP interaction domain on LRP (Rieger et al., 1997). LRP44–101 (row 1), LRP44–160 (row 2) and LRP44–295 (row 3) were co-expressed in prey position together with huPrP23–230 fused to GST in bait position of the yeast two-hybrid system. *Interactions between LRP157–295 and LRP180–295 (in prey position) versus huPrP23–230 fused to GST (in bait position) have been investigated previously (Rieger et al., 1997). (C) PrP144–179 interacts directly with LRP161–179 in the yeast two-hybrid system. PrP144–179 fused to GST in bait position was co-expressed with LRP161–179 (row 1), LRP180–295 (row 2) and LRP44–160 (row 3) in prey position of the yeast two-hybrid system. (A–C) All interactions were monitored by the β-galactosidase reporter system. Download figure Download PowerPoint Retrenchment of the direct PrP-binding domain on LRP by the yeast two-hybrid system Recently, we mapped a direct PrP-binding domain on LRP between aa 157 and 180 (Rieger et al., 1997) employing N-terminally truncated LRP molecules. In order to re trench this binding domain precisely we co-expressed the C-terminally truncated LRP molecules LRP44–101 and LRP44–160, respectively, together with full-length PrP in the yeast two-hybrid system. Both truncations failed to interact with PrP (Figure 1B) confirming that this direct PrP-binding site coincides with the laminin-binding domain (aa 161–180). Expression of an LRP mutant lacking this direct PrP-binding domain (LRPdelBD161–180) in CHO cells (Figure 4K) showed that LRPdelBD161–180::FLAG was still able to bind to PrP, indicating the presence of a second binding site for PrP on LRP, which locates either between aa 101 and 160 or 181 and 295 of LRP. PrP144–179 interacts directly with LRP161–179 in the yeast two-hybrid system In order to prove a direct interaction between the PrP and the 37-kDa/67-kDa LRP/LR via PrP144–179 and LRP161–179, we co-expressed both protein domains in bait and prey position, respectively, resulting in a strong interaction (Figure 1C, row 1). In contrast, huPrP144–179 failed to interact with LRP180–295 (row 2) or LRP44–160 (row 3). β-galactoside gal-3 does not interact with PrP or LRP in the yeast two-hybrid system The association of gal-3 with the LRP has been suggested (Landowski et al., 1995; Buto et al., 1998). However, gal-3 antibodies do not influence the LRP-dependent binding/internalization of PrP on the cell surface, suggesting that this molecule does not act as a co-receptor for LRP (Gauczynski et al., 2001b). For confirmation that gal-3 does not interact with PrP, we expressed gal-3 in bait and PrP in prey position of the yeast two-hybrid system resulting in no interaction between the two proteins (Figure 2A, row 2). Gal-3 also failed to interact with LRP in the yeast two-hybrid system (Figure 2A, row 3). Figure 2.LRP fails to interact with itself in the yeast two-hybrid system and appears monomeric by native size-exclusion chromatography. Gal-3 fails to interact with PrP and LRP. (A) huPrP23–230 and LRP fail to interact with the β-galactoside lectin gal-3 in the yeast two-hybrid system. huPrP23–230 fused to GST in bait position was co-expressed with LRP44–295 in prey position (row 1), gal-3 in bait position was co-expressed with GST::huPrP23–230 (row 2) and LRP44–295 (row 3) in prey position. (B) LRP fails to interact with itself in the yeast two-hybrid system. GST (row 1), LRP44–295 (row 2) and GST::huPrP23–230 (row 3) were expressed in bait and GST (row 1) and LRP44–295 (row 2 and 3) in prey position of the yeast two-hybrid system. Detection (A and B) by the β-galactosidase reporter system. (C) Analysis of rec. LRP::FLAG on SDS–PAGE. One microgram of rec. LRP::FLAG purified under native conditions from the SFV system was analysed on a 12.5% SDS–PA gel stained with silver (lane 1). Marker proteins are indicated. (D) Analysis of rec. native LRP::FLAG by size-exclusion chromatography. Homogeneous LRP::FLAG (2.2 μg) were analysed by size-exclusion chromatography on a Superose 12PC3.2/30 column (Amersham Pharmacia). Marker proteins are indicated. Download figure Download PowerPoint 37-kDa LRP fails to interact with itself in the yeast two-hybrid system and appears monomeric by size-exclusion chromatography The polymorphism of the LR is still unclear. In order to test whether homodimerization of LRP could account for this 37-kDa/67-kDa polymorphism and to understand better the configuration of the LRP–PrP-binding complex, we cloned the cDNA encoding for LRP in the bait and prey position of the yeast two-hybrid system. LRP fails to interact with itself (Figure 2B, row 2), suggesting that LRP is unable to directly form homodimers. For confirmation we purified LRP::FLAG from SFV RNA-LRP::FLAG-transfected BHK cells by anti-FLAG antibody chromatography to homogeneity and analysed the native protein by SDS–PAGE and size-exclusion chromatography. The protein migrated as a 37-kDa protein on an SDS–polyacrylamide gel (Figure 2C, lane 1) and eluted as a 40-kDa protein from a native size-exclusion column (Figure 2D), confirming that 37-kDa LRP is monomeric under native conditions. Thus, the 67-kDa form of the LR may result from the association of the LRP with other molecules such as HSPGs. Identification of the PrP interaction domains PrPLRPbd1 and PrPLRPbd2 by binding assays with prion peptides to NT2 and N2a cells The yeast two-hybrid system identified the domain aa 144–179 of PrP as a direct binding site for LRP termed PrPLRPbd1. To identify other domains of PrP, which might bind indirectly to LRP, we exposed NT2 and N2a cells to glutathione S-transferase (GST)-fused PrP peptides covering the entire PrP sequence. Besides peptide 129–175 encompassing the direct binding domain PrPLRPbd1, GST::PrP53–93 bound to the cells in an LRP-dependent fashion. The binding of GST::PrP53–93 (Figure 3A) and GST::PrP129–175 (Figure 3C) is shown on NT2 cells in comparison with GST::PrP90–109 (Figure 3B). This binding can be inhibited by addition of the LRP antibody W3 (insets in Figure 3A and C). The binding properties of the whole array of peptides are also shown on N2a cells (Figure 3D–J). Only GST::PrP53–93 (Figure 3E) and GST::huPrP129–175 (Figure 3H) bound to the cells dependent on LRP–LR (LRP antibody competition is shown in the bottom insets of Figure 3). The integrin laminin receptor VLA6 does not co-localize with PrP or LRP–LR on the surface of neuroblastoma cells (Gauczynski et al., 2001b). The addition of an anti-VLA6 antibody, failed also to compete for the binding of GST::PrP53–93 or GST::PrP129–175 to N2a cells (Figure 3E and H, top insets), confirming that VLA6 does not act as a receptor for PrP. We termed the indirect binding domain PrPLRPbd2. As the two binding domains are located N- and C-terminally of the proteinase K cleavage site of PrPres, we tested longer peptides in our binding assay corresponding to the two fragments that result from proteolytic cleavage of PrP, i.e. PrP23–89 and PrP90–230. Both peptides bound to both cell types in an LRP-dependent manner (Table I). Combining these data with the results from the yeast two-hybrid system (Figure 1) we conclude that two binding sites on PrP for LRP termed PrPLRPbd1 (aa 144–179) and PrPLRPbd2 (aa 53–93) do exist. Results of the PrP peptide binding studies to N2a and NT2 cells including antibody competitions are summarized in Table I. Figure 3.Identification of PrP-interaction domains for LRP/LR by binding assays with rec. prion peptides on NT2 and N2a cells. (A–C) Binding to NT2 cells. NT2 cells were incubated with PrP peptides fused to GST in the absence (A–C) and presence (insets in A–C) of the pre-incubated pAb LRP W3 (dilution 1:50). The following peptides (4 μg/ml) were used: GST::PrP53–93 (A), GST::PrP90–109 (B) and GST::PrP129–175 (C). Immunofluorescence analysis was performed by triple labelling involving actin staining (phalloidin, red), nuclear staining [4′,6-diamidino-2-phenylindole (DAPI), blue] and GST staining (sec. Ab FITC, green) (magnification ×630). (D–J) Binding to N2a cells. N2a cells were incubated with PrP peptides fused to GST in the absence (D–J) or presence of either pre-incubated pAb LRP (dilution 1:50) (bottom insets in D–J) or pre-incubated mAb VLA6 (dilution 1:50) (top inset in E and H). The following peptides (4 μg/ml) were used: GST::PrP23–52 (D), GST::PrP53–93 (E), GST::PrP90–109 (F), GST::PrP110–128 (G), GST::PrP129–175 (H), GST::PrP180–210 (I) and GST::PrP218–230 (J). Immunofluorescence was performed with mAb GST, sec. Ab Texas Red, DAPI staining (magnification ×400). Download figure Download PowerPoint Figure 4.Influence of HSPGs on the LRP–LR–PrP-binding reaction analysed by wild-type and mutant HSPG-deficient CHO cells; use of an LRP deletion mutant lacking the direct PrP-binding domain on LRP; PrP–LRP in vitro interaction studies. (A–F) Binding of GST::huPrP23–230 to CHO cells and to HSPG-deficient CHO cells (S745) (Esko et al., 1985) in the absence and presence of antibodies. Binding of 4 μg/ml GST::huPrP23–230 to CHO wild-type cells in the absence of any antibody (A), in the presence of the pAb LRP W3 (1:50) (C), mAb SAF 70 (aa 140–180 of PrP) (E). Binding of GST::huPrP23–230 to CHO-S745 cells in the absence of any antibody (B), in the presence of the pAb LRP W3 (D) and in the presence of the mAb SAF 70 (F). The mAb SAF70 was used at a 1:1000 dilution to saturate GST::huPrP23–230 while pAb LRP was pre-incubated with the cells at a 1:50 dilution prior to addition of the rec. PrP. Immunofluorescence analysis: pAb LRP (sec. Ab FITC), mAb 3F4 (sec. Ab Texas Red; DAPI staining; magnification ×400). (G–J) Binding of GST::PrP53–93 to CHO wild-type and CHO-S745 cells. Binding of 4 μg/ml GST::huPrP53–93 to CHO wild-type (G) and CHO-S745 cells in the absence of HSPGs (H), and in the presence of 10 (I) and 40 μg/ml (J) HSPGs, respectively. Immunofluorescence analysis: mAb GST (sec. Ab Texas Red; DAPI staining; magnification ×400). (K) Binding of PrP by wild-type CHO and mutant HSPG-deficient CHO-S745 cells hyperexpressing LRP::FLAG or LRPdelBD::FLAG (lacking aa 161–180). CHO cells (lanes 1–6) either hyperexpressing LRP::FLAG (lanes 1 and 2), LRPdelBD::FLAG (lanes 3 and 4) by the SFV system or non-tranfected (lanes 5 and 6) were incubated with 5 μg/ml GST::huPrP (lanes 2, 4 and 6). CHO-S745 cells (lanes 7–13) either hyperexpressing LRP::FLAG (lanes 7 and 8), LRPdelBD::FLAG (lanes 9–11) by the SFV system or non-transfected (lanes 12 and 13) were incubated with 5 μg/ml GST::huPrP (lanes 8, 10, 11 and 13). HSPGs (40 μg/ml) were added simultaneously with GST::huPrP to the CHO-S745 cells overexpressing LRPdelBD::FLAG (lane 11). Total cell extracts were analysed by western blotting employing the mAb 3B5 (upper panels) or pAb LRP W3 (lower panels). (L) Binding of PrP by non-transfected wild-type CHO and mutant HSPG-deficient CHO-S745 cells. Non-transfected CHO wild-type cells (lanes 1–3) and non-transfected CHO-S745 cells (lanes 4–6) were incubated with 5 μg/ml GST::huPrP (lanes 2, 3, 5 and 6). HSPGs (40 μg/ml) were added simultaneously with GST::huPrP to both cell types (lanes 3 and 6, respectively). Total cell extracts were analysed as described in (K). Interaction of rec. FLAG::PrP and rec. GST::LRP in vitro (M). FLAG::huPrP23–230 (0.5 μg) immobilized on anti-FLAG Sepharose beads analysed on a 12% SDS–PA gel stained with silver (lane 1) and by western blotting employing the pAb JB007 (lane 2) were incubated with 1 μg of GST::LRP in the absence (lane 4) or in the presence of 1.5 μg/μl HSPGs (lane 3), 0.5 μg of GST in the absence (lane 6) or the presence of 1.5 μg/μl HSPGs (lane 5). Unloaded beads were incubated with 1 μg of GST::LRP in the presence of 1.5 μg/μl HSPGs (lane 7). Beads after washing were analysed by western blotting on a 12% SDS–PA gel employing mAb GST (sec. antibody POD). Download figure Download PowerPoint Table 1. Summary of the binding behaviour of individual GST-fused PrP peptides to NT2 and N2a cells including LRP–LR and VLA6 antibody competition Peptide (aa) Binding to N2a cells LRP–LR antibody competition Binding to NT2 cells LRP–LR antibody competition PrP53–93 +++ +++a +++ +++ PrP90–109 − − − − PrP110–128 − − − − PrP129–175 +++ +++a +++ +++ PrP180–210 − − − − PrP218–230 − − − − PrP23–89 +++ +++ +++ +++ PrP90–230 +++ +++ +++ +++ a No competition with a VLA6 antibody. +++, strong binding/competition. −, no binding/competition. Binding of PrP to LRP via PrPLRPbd2 is dependent on HSPGs The cell-binding assay led to the identification of an additional binding domain for LRP on PrP which was not identified in the yeast two-hybrid system, indicating that a third molecule is necessary to mediate the binding of LRP to PrPLRPbd2. It has been reported that 60% of the binding of rec. chicken PrP to CHO cells depends on the presence of endogenous heparan sulfates (Shyng et al., 1994). Mutant CHO cells (S745) are severely deficient in HSPGs because of an altered xylose transferase activity, the first enzyme required for glycosylaminoglycan (GAG) synthesis (Esko et al., 1985) and, therefore, represent an appropriate model system to investigate whether HSPGs might represent the third interactor in the binding of PrPLRPbd2 to LRP. First, we proved that the binding of rec. GST::huPrP23–230 to wild-type CHO cells (Figure 4A) and to the mutant CHO cells (Figure 4B) was LRP–LR-dependent (Figure 4C and D). We then saturated selectively PrPLRPbd1 by incubating GST::huPrP23–230 with a monoclonal PrP antibody directed against the domain 140–180 of PrP. Obstructing PrPLRPbd1 inhibited the binding of the rec. PrP to HSPG-deficient (Figure 4F) but not to wild-type CHO cells (Figure 4E), demonstrating that binding of PrP to LRP via PrPLRPbd2 needs the presence of HSPGs. The peptide GST::PrP53–93 corresponding to PrPLRPbd2, which bound to normal cells (Figure 4G), did not bind to HSPG-deficient cells (Figure 4H). However, the binding was restored in a dose-dependent manner after addition of soluble HSPGs (Figure 4I and J), confirming that the interaction of PrP to LRP via PrPLRPbd2 is HSPG dependent. Identification of an HSPG-dependent second binding site for PrP on LRP The two binding sites on PrP, one of which is direct and the other indirect, suggested that two ‘acceptor’ sites might also exist on LRP. To test this hypothesis we adapted our SFV expression system to CHO cells. We then expressed LRP::FLAG and a mutant LRP lacking the direct PrP-binding domain (aa 161–180), termed LRPdelBD::FLAG, in wild-type CHO cells and the mutant CHO-S745 cell line lacking HSPGs (Figure 4K, lower panels). In wild-type CHO cells (Figure 4K, lanes 1–6) the binding of GST::huPrP was enhanced, when LRP::FLAG was hyperexpressed (Figure 4K, lanes 2 versus 6). Hyper expression of LRPdelBD::FLAG did not reduce the amount of the bound GST::huPrP (Figure 4K, lane 4), indicating that the binding was HSPG mediated. In mutant CHO-S745 cells lacking HSPGs (Figure 4K, lanes 7–13) the amount of bound GST::huPrP was also enhanced in LRP::FLAG hyperexpressing cells compared with non-transfected cells (Figure 4K, lanes 8 and 13). CHO-S745 cells expressing LRPdelBD::FLAG, however, showed a reduced binding of GST::huPrP (Figure 4K, lane 10), similar to non-transfected cells (Figure 4K, lane 13) suggesting that the second indirect PrP-binding site was not functioning in the absence of HSPGs. In order to confirm that HSPGs are responsible for the binding of PrP to the second binding domain on LRP, we added HSPGs to cells overexpressing LRPdelBD::FLAG resulting in a total restoration of the PrP binding (Figure 4K, lane 11). HSPGs failed to increase the binding of GST::huPrP to CHO wild-type (Figure 4L, lanes 2 and 3) and CHO-S745 cells (Figure 4L, lanes 5 and 6) due to the presence of the direct binding domains on PrP and LRP. We conclude from these data that two binding sites for PrP on LRP exist: a direct one, which is located from aa 161–179, and a second indirect one, which resides either between aa 101 and 160 or between aa 180 and 295 of LRP. Interaction of PrP and LRP in vitro The direct binding domains on LRP (aa 161–179) and PrP (aa 144–179) should allow the two proteins to interact with each other in vitro. GST-fused LRP (Rieger et al., 1997) and immobilized FLAG::huPrP (Figure 4M, lanes 1 and 2) were able to interact with each other in vitro as shown in the pull down assay depicted in Figure 4M. As already observed on wild-type CHO and CHO-S745 cells, HSPGs did not influence the interaction due to the presence of the direct interaction domains (Figure 4M, lanes 2 and 3). However, HSPGs did affect the LRP–PrP53–93 interaction (HSPG-dependent binding domain on PrP) and the LRPdelBD–PrP interaction (lacking the direct binding domain on LRP) in CHO-S745 cells (Figure 4G–J and K, respectively). Discussion Cell-binding and internalization studies proved that the 37-kDa LRP/67-kDa LR acts as the receptor for the cellular PrP, PrPc, on the cell surface (Gauczynski et al., 2001b). In order to investigate the interaction domains on PrP and LRP–LR mediating the binding of these two proteins a series of interaction studies employing the yeast two-hybrid system as well as PrP-binding assays with neuronal and non-neuronal cells including the SFV system have been performed. Mapping of LRP interaction sites on PrP First, we aimed to determine which part of the PrP interacts with LRP.