Adenovirus dodecahedron cell attachment and entry are ... - CORE

1 downloads 0 Views 1MB Size Report
Available online at www.sciencedirect.com. Virology 371 (2008) 155–164 ... over-expressed along the viral cell cycle (Fender et al., 2005), and these extra Ad3 ...
Available online at www.sciencedirect.com

Virology 371 (2008) 155 – 164 www.elsevier.com/locate/yviro

Adenovirus dodecahedron cell attachment and entry are mediated by heparan sulfate and integrins and vary along the cell cycle Pascal Fender a,⁎, Guy Schoehn b , Françoise Perron-Sierra c , Gordon C. Tucker c , Hugues Lortat-Jacob a a CNRS, CEA, UJF: Institut de Biologie Structurale, 41 rue Jules Horowitz 38027 Grenoble, France CNRS, UJF, EMBL: Unit of Virus Host Cell Interactions, 6 rue Jules Horowitz B.P.181 38042 Grenoble Cedex9, France Institut de Recherches Servier: Medicinal Chemistry and Cancer Drug Discovery Division, 125 Chemin de ronde, 78290 Croissy sur seine, France b

c

Received 5 July 2007; returned to author for revision 22 August 2007; accepted 7 September 2007 Available online 22 October 2007

Abstract The adenovirus penton base is a strategic protein involved in the virus internalisation pathway through interaction between its RGD sequences and integrin. In some human adenovirus serotypes, this pentameric protein features the ability of interacting together by twelve, leading to the formation of a symmetric nanoparticle called dodecahedron (Dd). This non-infectious adenovirus-like particle exhibiting sixty RGD sequences interacts with integrin but also with heparan sulfate proteoglycans (HSPGs) expressed at the cell surface. In this study, we discriminate the respective importance of HSPGs and integrin on human adenovirus serotype 3 dodecahedron attachment and entry. Using different cell lines and a specific integrin inhibitor, we have determined that HSPGs are mainly responsible for particle attachment to the cell surface, favouring a strictly required interaction with integrin that triggers internalisation. No other receptors are involved in Dd entry and integrins on their own can mediate the particle entry in HSPGs-deficient cells. Moreover, integrin recognition by Dd is highly susceptible to cations and particularly to manganese that enhances particle binding by 4- to 7-fold compared to calcium. Interestingly, investigations on Dd receptors along the cell cycle revealed an enhanced particle targeting to mitotic cells and a loss of internalisation at this stage. This phenomenon observed with both HeLa- and HSPGs-deficient cells, depends on integrin remodelling during mitosis. This provides new clues for the use of this adenovirus nanoparticle as a delivery vector and sheds light on the integrin and HSPGs relationship in both resting and dividing cells. © 2007 Elsevier Inc. All rights reserved. Keywords: Adenovirus; Dodecahedron; Integrin; Heparan sulfate proteoglycan; Receptor; Entry; Mitosis; Cell cycle

Introduction Heparan sulfate proteoglycans (HSPGs) are glycoproteins ubiquitously expressed at the surface of mammalian cells and in extracellular matrices. HSPGs properties are mostly mediated by the O-linked heparan sulfate (HS) polysaccharide chains present on the protein core that bind and regulate a wide range of proteins, through motifs of specific saccharide sequence (Lindahl et al., 1998). Pathogens binding to HS are generally believed to promote infectivity by increasing the local concentration of pathogens at the cell surface and thus favouring access to specific entry receptors. Hence, HS can play the role of a viral ⁎ Corresponding author. Fax: +33 4 38 78 54 95. E-mail address: [email protected] (P. Fender). 0042-6822/$ - see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.virol.2007.09.026

attachment receptor, as described for adenovirus types 2 and 5 (Dechecchi et al., 2000, 2001), adeno-associated parvovirus-2 (AAV-2) (Summerford and Samulski, 1998), several members of the herpes virus family (Shukla and Spear, 2001), Flavivirus including hepatitis C (Germi et al., 2002), Dengue virus (Chen et al., 1997) and Sindbis virus (Byrnes and Griffin, 1998). Besides attachment functions, HSPGs can also be used for virus storage and are involved in in trans infection. This has been shown for HIV whose interaction with non-permissive epithelial cells protects the virion awaiting a secondary interaction with permissive cells (reviewed in Vives et al., 2006). Integrins are heterodimeric type I transmembrane receptors composed of an α and β subunit with over 20 members belonging to their family. Many of them recognise an arginine, glycine, aspartic acid (RGD) sequence in host extracellular matrix

156

P. Fender et al. / Virology 371 (2008) 155–164

proteins such as vitronectin, fibronectin and tenascin. The crystal structure of the extracellular segment of integrin αvβ3 with an RGD ligand has been solved and has confirmed the role of divalent cations, such as Ca2+ or Mn2+, in the integrin function (Mould, 1996; Xiong et al., 2002). These heterodimeric receptors are involved in many versatile cellular processes including cell adhesion, differentiation, motility or pathogen invasion eliciting a major interest in their study (Hynes, 1992). Many viruses utilise integrins in their entry strategy such as echovirus, foot and mouth disease virus (FMDV), cocksackie virus or adeno-associated virus (AAV) (Summerford et al., 1999; White, 1993). Adenovirus internalisation is a two step mechanism involving a high affinity attachment receptor recognised by the apical domain of an elongated capsid protein called the fibre and a secondary interaction with an entry receptor mediated by the penton base protein. A 46-kDa protein called CAR and HSPGs have been identified as the major attachment receptors for a majority of human adenovirus serotype (Bergelson et al., 1997; Dechecchi et al., 2000) with the notable exception of subgroup B adenoviruses such as human adenovirus of serotype3 (Ad3). Interaction of adenoviruses to attachment receptors facilitates a secondary interaction with integrin that triggers internalisation. The latter is mediated through a specific interaction of the conserved RGD sequences exposed at the surface of the penton base that are located at the twelve virion vertices (Belin and Boulanger, 1993; Wickham et al., 1993; Mathias et al., 1998; Nemerow and Stewart, 1999). Interestingly, Ad3 penton base is over-expressed along the viral cell cycle (Fender et al., 2005), and these extra Ad3 penton base proteins feature a self-assembly capacity, leading to the formation of dodecahedric particle with fibres protruding outside called the penton-dodecahedron: “PtDd”. Dodecamerisation is not dependent on fibres as attested by the formation in the baculovirus system of a dodecahedron made only of the base proteins called base-dodecahedron “Bs-Dd” (Schoehn et al., 1996; Fender et al., 1997; Fuschiotti et al., 2006b). Similarly to the entire 90-nm icosahedric adenovirus capsid, these 29-nm- and 49-nm-diameter particles (without and with fibres, respectively) also expose sixty RGD sequences, but in different spatial arrangements (Schoehn et al., 1996; Stewart et al., 1997). As observed with Pt-Dd, the fibreless Bs-Dd has also been reported to efficiently enter a wide variety of cell types such as epithelial and endothelial cells, fibroblasts, or human primary astrocytes (Fender et al., 1997; Garcel et al., 2006). Its great internalisation efficacy may rely upon the high-affinity interaction of Bs-Dd to HSPGs (Vives et al., 2004). Nevertheless, the exact role of HSPGs and integrin in Bs-Dd attachment and entry remains to be investigated. We have recently demonstrated that dodecahedra could constitute a good tool for gene, chemical compound, or protein delivery (Fender et al., 1997; Garcel et al., 2006; Fuschiotti et al., 2006a). To go further with the potential applications of this particle, the clarification of the fine mechanisms involved in cell attachment and entry is then required. In the following study, the respective role of HSPGs and integrin on Bs-Dd binding and internalisation has been studied. To discriminate between these two receptor functions, different

tools have been used such as cells expressing integrin but not HSPGs (CHO-2241) or the S 36578 chemical compound which specifically inhibits ligand biding to αvβ3 and αvβ5 integrins at low nanomolar concentrations without affecting other integrins such as the platelet αIIbβ3 integrin (Perron-Sierra et al., 2002). The mechanism of Bs-Dd attachment and entry has been determined on cells in interphase and on mitotic cells revealing an interesting feature of integrin during the cell cycle. Results The RGD loops are tightened together in the dodecahedron and possess a typical HSPG-binding site Bs-Dd strongly interacts with HSPGs although it has been reported that the Ad3 virion does not bind to proteoglycans. It is

Fig. 1. RGD localisation on Ad3 virus particles and base dodecahedron. (a) Schematic representation of an icosahedra (left) and a dodecahedron (right) viewed down their 3-fold axis. For the icosahedra, one facet shows the spatial organisation that hexons (different gray level) and penton base (white) adopt in the adenovirus capsid. (b) Three-dimensional reconstructions of an adenovirus structure and of the Bs-Dd structure calculated from electron microscopy images and filtered to 25 Å resolution (the maps are accessible on the Macromolecular Structure Database under the accession number EMD 1113 (adenovirus) and EMD 1178 (Bs-Dd)). RGD loops are highlighted in blue. (c) Electron micrographs of a mixture of human adenovirus serotype 3 and Bs-Dd particles. The scale bar represents 100 nm.

P. Fender et al. / Virology 371 (2008) 155–164

interesting to notice that Ad3 and Bs-Dd both contain twelve penton bases in their structure but with a different structural arrangement (Fig. 1). Indeed, whereas the pentamer organisation is similar on these two particles (4.5 nm distance between intrapentameric RGD loops), the distance between the penton bases dramatically changes. This results on the clustering of RGD loops (depicted in blue on Fig. 1b) on the Bs-Dd that are tightened together with only 7.5 nm between RGD loops from adjacent pentamers whereas they are separated by four hexon capsomers lying on each edge of the virion capsid (Figs. 1b and c). It is worth noting that a short basic sequence 336KQKR339 known as a typical heparan sulfate binding site is located within the RGD loop of the Ad3 penton base (Fig. 1b). Knowing that the HSPGs binding site has already been described close to the integrin recognition motif in other viruses such as AAV-2 or FMDV (Fry et al., 1999; Kern et al., 2003), either a sequential or a cooperative use of these two dodecahedron receptors might be envisaged. The respective role of HSPGs and integrin has thus been studied in both cell attachment and entry. Integrins are the sole entry receptor Our first goal was to determine whether HSPGs are involved in endocytosis or if integrins work on their own. To investigate this point, an integrin-blocking drug (S 36578) described to block selectively αvβ3 and αvβ5 was used (Perron-Sierra et al., 2002). While HeLa cells (i.e., expressing both HSPGs and integrin) efficiently internalised Bs-Dd (Fig. 2a, inset), preincubation with the S 36578 drug resulted in a total abolishment of Dd entry (Fig. 2a). With respect to attachment, the bright green signal surrounding S 36578-treated HeLa cells indicated that even if integrins are blocked, Bs-Dd attachment to HSPGs is not significantly affected by the drug. To determine whether HSPGs are required in integrinmediated endocytosis, a similar experiment was performed using HSPGs-deficient CHO-2241 cells. Internalisation of the

157

adenoviral pseudo-particle was observed in CHO-2241 cells (Fig. 2b), showing that integrins are sufficient to mediate endocytosis. Moreover, neither attachment nor entry could be observed in CHO-2241 cells pre-incubated with the S 36578 drug prior to Bs-Dd addition (Fig. 2b, inset), reinforcing the critical role of integrins in Bs-Dd internalisation, as previously observed with HeLa cells. Nuclei are counterstained in red by propidium iodide. Respective role of HSPGs and integrin in Bs-Dd attachment: importance of cations Bs-Dd internalisation into CHO-2241 cells (i.e., deficient for HSPGs expression) suggests that integrins are sufficient for BsDd entry, but at this stage, we cannot exclude that another receptor than integrin is used in the absence of HSPGs. To address this question, CHO-2241 cells have been incubated at non-permissive temperature (4 °C), with Bs-Dd in PBS containing either Ca2+, the integrin-activating cation Mn2+ or the integrin-blocking drug S 36578. The FACS curve obtained in presence of calcium confirms that Bs-Dd binds to the surface of these HSPGs deficient cells (Fig. 3a). Interestingly, when cells were pre-incubated with the S 36578 drug (blue line), no binding can be observed as the curve shift to the antibody background level (dashed line), thus indicating that no other Bs-Dd receptors than integrins are expressed in this cell line. Moreover, confirming the role of integrins in Bs-Dd attachment, addition of Mn2+ that is known to trigger their activation, directs a 4-fold increase in Bs-Dd binding (green curve, 104 arbitrary units instead of 24). Altogether, these data show that in the absence of HSPGs, integrins can be used for particle attachment and that no other Bs-Dd' receptors are required (at least on the cell line used). A comparable experiment performed with HeLa cells shows similar integrin behaviour with a major difference coming from the presence of HSPGs at the surface of this cell line (Fig. 3b). The expression of the latter receptor explains the higher binding

Fig. 2. Integrins inhibition blocks Bs-Dd entry into HeLa and CHO-2241 cells. (a) HeLa cells pre-treated with S 36578 for 30 min are incubated for 1 h at 37 °C with Bs-Dd. Particle detection is performed on permeabilised cells using appropriate FITC antibodies. Inset: Control in the absence of drug. (b) Similar experiment performed with CHO-2241. Bs-Dd internalisation is observed in non-treated cells while pre-incubation with S 36578 prior to Bs-Dd incubation abolishes particles attachment and internalisation (no green fluorescence, see inset). Nuclei are counterstained in red by propidium iodide.

158

P. Fender et al. / Virology 371 (2008) 155–164

Fig. 3. Relative importance of HSPGs and integrin in Bs-Dd attachment to HeLa and CHO-2241 cells. (a) CHO-2241 cells are incubated at 4 °C with Bs-Dd in the presence of 1 mM CaCl2 (red curve), 1 mM MnCl2 (green curve) or 3 nM S 36578 (blue curve). The dashed brown curve represents antibody background in the absence of Bs-Dd. The relative average binding is calculated from FACS statistics and indicated in arbitrary unit (a.u.). (b) Similar experiment with HeLa cells (same colour code). (c) Relative contribution of HSPG (blue), integrins in the presence of calcium (red) and integrins in the presence of manganese (green, slightly separated from the binding of reference in presence of Ca2+) on Bs-Dd binding to both HeLa and CHO-2241 cells.

intensity of Bs-Dd to this cell line compared to the one observed with CHO-2241. This demonstrates that most of the particles are captured on the HeLa cell surface by HSPGs, which is therefore the main attachment receptor for Bs-Dd. The HSPGs contribution can be estimated using the data obtained for cells pre-incubated with the integrin-blocking drug (blue curve). Indeed, this signal reflects the particle attachment mediated by HSPGs while the little shift toward the higher binding intensity in non-treated cell gives information on the integrin contribution in presence of calcium. Considering the Bs-Dd binding in presence of calcium as the reference (645 units), it is possible to assess from the FACS statistics that, on average, HSPGs are responsible for 571 units (89%) of the particle binding whereas direct integrin recognition accounts for the extra 74 units (11%). Interestingly, when cells were incubated with Mn2+, the BsDd attachment to HeLa cell is greatly enhanced (Fig. 3b, green curve). This presumably reflects the cation effect on integrin previously observed with CHO-2241 cells. The total particle binding under this condition is 1085 units compared to 645 in the presence of calcium. Based on BIAcore experiments (Bs-Dd injection at 100 μg/μl for 5 min at 5 μl/min on a heparin-coated censorship yielded 1064 RU and 1006 RU in absence or presence of Mn2+, respectively), it can be assumed that Bs-Dd binding to HSPG is not affected by the presence of Mn2+. It can therefore be deduced that integrins contribution to Bs-Dd attachment shifts from 74 to 514 units in the presence of Mn2+ (Fig. 3c). This 7-fold integrin activation by Mn2+ makes this receptor almost as important as HSPGs under this condition (514 and 571 units, respectively; see Fig. 3c).

Data collected from CHO-2241 and HeLa have been summarized (Fig. 3c). Blue bar represents the relative HSPGs contribution in Bs-Dd binding explaining the higher signal obtained with HeLa cells. Red and green bars show the contribution of integrin under two different conditions (Ca2+ and Mn2+ incubation, respectively). Strong binding of Bs-Dd to mitotic cells Visual inspection of cells incubated with Bs-Dd at 37 °C shows that internalisation does not occur similarly in all cells. As expected from our previous work, particle internalisation is clearly seen in HeLa cells at 37 °C, as reflected by the punctuated signal covering all the cell (Fig. 4a), but surprisingly, other cells exhibit a strong green signal specifically at their surface (indicated by arrowheads). Thus, internalisation does not occur on the latter cells while attachment is still efficient. The differences between these two features can be explained by looking at the DNA repartition in the cell. Indeed, propidium iodide staining of cells surrounded by Bs-Dd shows DNA condensation indicating that they are engaged in the division process. Different stages of mitosis can be seen in Fig. 4a: (i) chromosomes condensation (prophase); (ii) chromosomes alignment onto the mitotic spindle (metaphase); (iii) chromosomes segregation at the opposite poles of the dividing cell (anaphase). Cells decorated with Bs-Dd at these different stages of the cell cycle are clearly seen in the same picture (Fig. 4a, lower picture) and contrast strongly with cells in interphase for both DNA staining and Bs-Dd localisation.

P. Fender et al. / Virology 371 (2008) 155–164

159

Role of integrins in Bs-Dd mitotic cell targeting To further study the role of integrins in mitotic targeting by the adenovirus pseudo-particle, two complementary experiments were performed. One has been done on CHO-2241 cell blocked in mitosis by nocodazole and incubated or not with the S 36578 drug. The second one was performed on HeLa cell treated by sodium chlorate in order to prevent HSPGs sulfation (Cardin and Weintraub, 1989) and subsequently synchronised in mitosis by nocodazole treatment. For both experiments, cell cycle synchronisation by nocodazole was checked by FACS, using DNA content quantification

Fig. 4. Bs-Dd targeting to mitotic HeLa and CHO-2241 cells. (a) Nonsynchronised HeLa are incubated for 1 h at 37 °C with Bs-Dd in PBS containing 1 mM CaCl2. Bs-Dd is detected by the appropriate antibodies (left panels) and DNA is stained in red with propidium iodide (median panels). Overlays of BsDd (green) and DNA staining (red) are shown on the right panels with arrowheads pointing to mitotic cells. (b) Overlaid picture of a similar experiment performed with CHO-2241 cell on the left with detailed views of cells in metaphase, anaphase and telophase on the right.

This surprising behaviour of mitotic cells has also been investigated in HSPG-deficient cells in order to determine whether integrins are involved in this process. A similar experiment performed with CHO-2241 cells also shows a very effective mitotic cell targeting by the adenoviral pseudo-particle (Fig. 4b). This suggests that the enhanced targeting of Bs-Dd to dividing cells does not rely upon HSPGs.

Fig. 5. Integrin involvement in mitotic cell recognition. (a) Nocodazole-blocked CHO-2241 cells are assessed for Bs-Dd binding at 4 °C in the presence (purple curve) or in the absence (blue curve) of the S 36578 drug. Inset: Mitotic status of the cells checked by DNA content determination (red curve: nocodazolesynchronised cells; filled green curve: non-synchronised control cells). (b) BsDd binding to non-treated HeLa cells is shown in orange (100% binding). Bs-Dd binding to chlorate-treated HeLa cell is represented for both non-synchronised (red curve) or nocodazole-synchronised cells (green curve): Inset: The relative binding percentage of Bs-Dd under these conditions is depicted with the same colour code.

160

P. Fender et al. / Virology 371 (2008) 155–164

by propidium iodide (Fig. 5a, inset). Nocodazole-treated cells all harboured a 4n DNA content (red curve) as compared to the typical profile of non-synchronised cells (green curve) with a majority of cells in interphase (2n DNA content) and few cell in S-phase or mitosis (4n DNA content). A similar control was used for HeLa cells (data not shown). For the CHO-2241 experiment, synchronised cells were pre-incubated or not with the S 36578 integrin blocking drug, prior to addition of Bs-Dd for 1 h at 4 °C. If particle binding could be clearly detected on these mitotic cells in the absence of the drug (Fig. 5a, blue curve), pre-incubation with S 36578 nearly totally abolished Bs-Dd attachment as reflected by the purple curve shift toward the background level. A parallel experiment using immunofluorescence confocal microscopy confirmed that nocodazole-synchronised CHO-2241 cells all harboured a strong Bs-Dd binding while pre-treatment with S 36578 prevented attachment of this particle, thus confirming the receptor role of integrins on CHO-2241 mitotic cells (data not shown). With HeLa cells, investigation on the importance of integrins in mitotic cell recognition is hindered by the strong binding of Bs-Dd to HSPGs. To overcome this, HeLa cells were treated with sodium chlorate in order to prevent HSPGs sulfation. As expected, this treatment greatly reduced Bs-Dd binding (Fig. 5b, red curve). The 13% remaining signal (Fig. 5b, inset), in keeping with the previous experiment (Fig. 3, 11% of integrins recognition by Bs-Dd) strongly suggested that this binding mainly reflects direct integrins recognition by Bs-Dd. Interestingly, when sodium chlorate treated cells were blocked in mitosis by nocodazole, the FACS curve shifted toward higher intensity (Fig. 5b, green curve). This 4-fold increase in signal indicated that mitotic cells were more susceptible to Bs-Dd binding. Hence, mitotic cell over-targeting by Bs-Dd is mediated by integrins. This observation is of interest for both the mechanistic understanding of Bs-Dd attachment and entry and also for the development of adenoviral-derived vectors. During mitosis, integrins are overexpressed rather than activated The increased Bs-Dd binding to mitotic cells could be due either to integrin overexpression or to their activation during cell cycle. To provide clues on the mechanism taking place, two complementary experiments have been performed. First, nocodazole-blocked HeLa cells were incubated with Bs-Dd in the presence of either Ca2+ or Mn2+. Although this experiment is not quantitative, FACS analysis (Fig. 6a) indicates that manganese activation of integrin still occurred on mitotic cells thus invalidating the hypothesis of a spontaneous activation of integrin during the cell cycle. Reinforcing the idea of integrin overexpression at the mitotic cell surface, the second experiment using the LM609 antibody showed that αvβ3 integrin was clearly detected at the surface of non-synchronised HeLa mitotic cells, while its detection was weak on cells in interphase (Fig. 6b). Altogether, the data suggest that integrins play a major role on Bs-Dd binding to mitotic cells whether in “spontaneous” mitosis or in a nocodazole-synchronised cell population. This is linked

Fig. 6. Integrins are overexpressed at the mitotic cell surface rather than activated. (a) Nocodazole-synchronised HeLa cells were incubated at 4 °C with Bs-Dd in PBS supplemented with either Ca2+ (red curve) or Mn2+ (green curve). The shift towards a higher intensity suggests that integrin activation by Mn2+ still occurs on cells in mitosis. (b) The expression of αvβ3 integrins was assessed at the surface of non-synchronised HeLa cells with the LM609 antibody, showing different intensities in staining between cells in interphase and mitotic cells revealed by Hoechst DNA labelling (depicted by arrow).

to integrin overexpression at the surface of cells engaged in the mitotic process. Discussion Deciphering virus entry mechanisms is critical for both fundamental research and biotechnological applications in vectorology. In the present study, the respective role of two receptors has been investigated for both the attachment and entry of an adenovirus-derived nanoparticle. Interestingly, this particle is made of only one protein of the human adenovirus of serotype 3, the pentameric penton base, which is able to self-assemble by twelve leading to the formation of a highly symmetric structure called the base-dodecahedron (Bs-Dd) (Fender et al., 1997; Fuschiotti et al., 2006b; Schoehn et al., 1996). The penton base

P. Fender et al. / Virology 371 (2008) 155–164

protein is normally located at the twelve vertices of the Ad3 icosahedral virion structure and is known to trigger virus entry through interaction with integrins (Belin and Boulanger, 1993; Mathias et al., 1998; Wickham et al., 1993). Contrary to the Ad3 virion that does not interact with HSPGs (Dechecchi et al., 2000), Bs-Dd binds with high affinity to this sulfated polysaccharidic receptor (Vives et al., 2004). It can be hypothesized that the architecture of Bs-Dd creates HS binding sites that do not exist in Ad3 virions (Fig. 1). In accordance with this hypothesis, structural information on others HSPGs-interacting viruses such as AAV-2 or FMDV has also shown that the HS binding sites are made of a spatial arrangement of basic residues coming from different monomers (Fry et al., 1999; Kern et al., 2003). Moreover, a putative HS binding site, 337KQKR340 , is located on the RGD loop, suggesting a close relationship between HSPGs and integrin receptors. This has been recently reported for AAV2 in which the NGR integrin recognition sequence is located at the vicinity of the HSPG recognition sequence on the VP3 trimer (Asokan et al., 2006). In the latter case, similarly to our observation (Fig. 3), HSPGs mediate primary attachment to the cell surface until efficient engagement of integrins is achieved. Nevertheless, the proximity of the putative HSPGs binding site to the RGD sequence in Bs-Dd might suggest a more intricate role in the binding and entry strategy of this particle. This is particularly interesting knowing that some integrins, such as α5β1, can act as facultative proteoglycans with extensive heparan sulfate glycosylation (Veiga et al., 1997). HSPGs as well as integrins have been reported to act as primary attachment receptors for several viruses. FMDV and AAV can even use both these receptors to promote infection (Berinstein et al., 1995; Jackson et al., 1996; Summerford et al., 1999; Summerford and Samulski, 1998). In the adenovirus case, the most studied, Ad2 and Ad5 but not Ad3, are known to interact with HSPGs through the fibre (Dechecchi et al., 2001, 2000) thus facilitating a subsequent interaction of the penton base protein with integrins that triggers endocytosis (Belin and Boulanger, 1993; Wickham et al., 1993). Concerning Ad3 BsDd, the penton base protein supports recognition of these two receptors. In our study, their respective role has been determined. One important finding is that, despite the secondary role played by integrin in particle attachment to HSPGs expressing cells (Fig. 3), the function of this receptor cannot be overcome for entry, as shown through the use of the anti-αv integrin S 36578 drug which completely prevents internalisation without affecting binding to HSPGs (Fig. 2). Additionally, experiments performed with CHO-2241 cells lacking HSPGs show that the internalisation process triggered by integrin in the presence of calcium does not strictly require the presence of HSPGs at the surface of the cell (Fig. 2). Moreover, in these HSPGs-deficient cells, no other receptors besides integrins are involved in Bs-Dd attachment, as S 36578 treatment leads to the total abolishment of Bs-Dd binding (Fig. 3). Taken altogether, the results obtained with HeLa cells in interphase (Figs. 2 and 3) suggest that integrins are a strictly required entry receptor, playing an accessory role in Bs-Dd attachment (11%) that is predominantly mediated by HSPGs (89%). This explains the weak Bs-Dd docking seen with CHO-

161

2241 cells compared to HeLa cells (Fig. 3). However, Mn2+ dramatically changes the integrin behaviour. Indeed, Bs-Dd binding to this receptor in the presence of Ca2+ (Fig. 3c, red) is increased 4- and 7-fold by Mn2+ (Fig. 3c, green) in CHO-2241 and HeLa cells, respectively. This indicates that this receptor is likely under an inactive conformation hiding its ligand binding site that is reversed by Mn2+ addition (Mould, 1996). This hypothesis is in agreement with previous work reporting structural and functional changes of several integrins (such as α5β1 or αvβ3) mediated by different cations (Dransfield et al., 1992; Lee et al., 1995; Mould et al., 1995; Smith et al., 1994). In summary, it can be concluded that the Bs-Dd entry mechanism in nondividing cells is a two-step process where HSPGs concentrates the particle at the cell surface, thus presumably facilitating interaction with the entry receptor. Both differences in receptor expression at the cell surface and ligand affinity explained the major part taken by HSPGs in attachment. It has been assessed that on average, 104 to 105 integrins are expressed on HeLa cells with a 10 to 20 nM affinity for Ad2 virions (Belin and Boulanger, 1993), whereas HSPG-binding sites are in the 105 to 106 range (Yanagishita and Hascall, 1992) and with an affinity for Bs-Dd in the nanomolar range (Vives et al., 2004). Nevertheless, in HSPGs-deficient cells, direct interaction with integrins is sufficient to support the entire entry mechanism, though with lower efficiency (Fig. 2b). During mitosis, a dramatic rearrangement of the cell occurs leading to major differences compared to cells in interphase. First, Bs-Dd internalisation is completely abolished, despite a very strong binding that contrasts with the punctuated signal observed with non-mitotic cells (Fig. 4). Lack of internalisation might be explained by the cytoskeleton remodelling taking place during cell division (Glotzer, 2001; Maddox and Burridge, 2003). Indeed, it has been previously reported that, as for adenovirus virions, both actin filaments and microtubules are required to trigger adenovirus penton base endocytosis (Rentsendorj et al., 2006; Suomalainen et al., 1999; Wang et al., 1998). Second, the binding efficiency of Bs-Dd to mitotic cells is higher than expected (Fig. 4). This phenomenon involves integrins, as treatment with S 36578 nearly completely abolished the Bs-Dd binding to CHO-2241 (i.e., HSPGs-deficient) mitotic cells (Fig. 5). At the light of the confocal microscopy study (Fig. 4), it cannot be excluded that cell rounding during mitosis increases the exposure of integrins that are normally located at the basal side of the cell, in contact with the extracellular matrix. Nevertheless, whatever the part taken by this mechanical effect, a change in integrin status occurs during mitosis. Indeed, in an experiment performed in suspension, chlorated HeLa cells (to prevent HSPGs receptor functionality) exhibit a 4-fold increase in particle binding upon nocodazole synchronisation compared to non synchronised cells (52% and 13%, respectively, in Fig. 5). Regarding our previous observation on integrin activation by manganese on cells in interphase (Fig. 3b), it could be envisaged that integrin shifts from low to high affinity during mitosis. In this case, integrin activation by manganese observed with cell in interphase should not be observed with mitotic cells. The experiment performed with nocodazole-blocked HeLa cells clearly showed that manganese activation still takes place, thus

162

P. Fender et al. / Virology 371 (2008) 155–164

invalidating this hypothesis (Fig. 6a). Thus, the enhanced particle binding at the surface of mitotic cells is due to integrin overexpression, as observed for αvβ3 using the LM609 antibody which specifically stained brightly the surface of HeLa mitotic cell (Fig. 6b). This observation is in agreement with a previous work reporting that integrin expression at the cell surface changes between cells in interphase and mitotic cells (Anilkumar et al., 1996). Indeed, this published work shows that despite an unchanged total expression level (i.e., extracellular and intracellular) of this receptor in the cell, αvβ3, which is the main heterodimeric receptor involved in adenovirus penton base recognition (Wickham et al., 1993), is up-regulated at the surface of the cell during mitosis by 2- to 3-fold. This result obtained with rat fibroblast cells has been confirmed with the human epithelial A549 cell line where a 3-fold increased expression of αv integrin is also found upon nocodazole treatment (Seidman et al., 2001). Our observation with HeLa cells is in the same range with a 4-fold enhancement of Bs-Dd binding to mitotic cells (Fig. 5b). Moreover, it has also been reported in another work that receptor remodelling on mitotic cells directs a 6-fold increase in adenovirus binding, leading to a 4-fold greater expression of recombinant adenovirus transgene (Seidman et al., 2001). Altogether, the results presented herein provide new clues for the understanding of both HSPGs and integrin function upon ligand binding and highlight a special feature of integrin during mitosis. These basic findings could also be exploited for biotechnological applications, as we have previously reported that Bs-Dd can be utilised as a versatile delivery vector (Fender et al., 1997, 2003; Fuschiotti et al., 2006a; Garcel et al., 2006). This work is a step forward towards the development of mitotic cell targeting vectors. Material and methods Cell culture Chinese hamster ovary (CHO-2241) cell lines (GAG negative, pgsB-618, ATCC code CRL-2241) were obtained from the European Collection of Cultured Cells. Cells were cultured at 37 °C, under 5% CO2 atmosphere, in Ham's F12 (Gibco BRL), supplemented with 10% fetal calf serum (FCS), penicillin (50 IU/ml) and streptomycin (50 μg/ml). HeLa cells were cultured in EMEM medium (Gibco BRL) supplemented with 10% FCS, penicillin (50 IU/ml) and streptomycin (50 μg/ml). Chlorate treatment was performed by adding 30 mM sodium chlorate in EMEM 2% FCS for 36 h, to trypsinised cells (Cardin and Weintraub, 1989). For mitotic cell synchronisation, nocodazole was added overnight at 40 μg/ml in the cell medium. Mitotic cells were harvested by gentle flushing. Production and purification of Bs-Dd Bs-Dd was expressed using the baculovirus system. Particles were purified by sucrose density gradient, as previously described (Fender et al., 1997) and stored in Hepes buffer pH 7, 4–150 mM NaCl.

S 36578 drug The S 36578-2 compound is the sodium salt of (5-(S)-7-{[4(pyridin-2-ylamino)-butyryl amino]-methyl}-6,9-dihydro-5Hbenzocyclohepten-5-yl) acetic acid and was separated on a chiral column from the racemate compound (compound 1.2 synthesized as described in Perron-Sierra et al., 2002). This RGD heterocyclic peptidomimetic blocks ligand binding to purified αvβ3 and αvβ5 at nanomolar concentrations in molecular and cellular assays with a good selectivity over the αIIbβ3 fibrinogen receptor and the α5β1 fibronectin receptor in cell assays (IC50 N 10 μM). Electron microscopy The purified Ad3 virus was mixed with purified Bs-Dd in PBS in order to obtain a particle ratio of 1 to 50. The mix was applied to a carbon film and negatively stained with 2% ammonium molybdate, pH 7.6. Micrographs were taken under lowdose conditions using a Jeol 1200 EX II microscope operating at 100 kV and with a nominal magnification of 30,000-fold. Flow cytometry analysis Cells were harvested by gentle scrapping and flushing, washed with PBS and re-suspended in cold PBS at 1 × 106 cells in 300 μl. Cations (CaCl2 or MnCl2) were added at a final concentration of 1 mM. The integrin-blocking drug S 36578 was added at a final concentration of 3 nM and pre-incubated for 30 min on ice before Bs-Dd addition. Bs-Dd was then added at 3 nM and incubated on ice for 1 h. After cold PBS washes, cells were incubated with the primary anti-Dd antibody (1/1000; 1 h on ice), washed with cold PBS then incubated with FITCconjugated secondary antibody (1/200; 1 h on ice). After final cold PBS washes, 10,000 cells were analysed by flow cytometry (Becton Dickinson). For DNA content analysis, cells were treated for 10 min at 37 °C in 500 μl of solution containing 1% Triton X-100, 1.2 mg/ml of trisodium citrate, 30 U/ml of RNAse (Roche) and 2.5 mg/ml of propidium iodide (Sigma). Reaction was stopped by addition of 14 μl of 5 M sodium chloride and samples were analysed by flow cytometry. Confocal microscopy Cells grown overnight on glass coverslips (about 105 cell/cm2) were rinsed with PBS and incubated for 1 h at 37 °C in 50 μl PBS–1 mM CaCl2 containing 3 nM Bs-Dd. For experiments using the S 36578 integrin-blocking drug, S 36578 was preincubated at 3 nM final concentration for 30 min prior to Bs-Dd addition. After PBS washes, cells were permeabilised with cold methanol for 10 min. Bs-Dd immunofluorescence detection was performed by incubation of coverslips with an anti-Bs-Dd rabbit polyclonal serum at 1/1000 in 50 μl PBS and a subsequent incubation with FITC-labelled goat anti-rabbit antibody (Jackson), diluted 1/250 in the same buffer. Cell nuclei were counterstained with propidium iodide (5 μg/ml). Laser

P. Fender et al. / Virology 371 (2008) 155–164

scanning confocal microscopy was performed on MRC600 (Bio-Rad). For integrin detection, live HeLa cells were incubated for 30 min at 37 °C with the anti-αvβ3 antibody (LM609, Chemicon) at 1/1000 in PBS–1 mM CaCl2 and stained with Hoechst. After cell washing, a Cy3-conjugated anti-mouse antibody (Jackson Immunoresearch) was added at 1/250 for 30 min. Cells were observed on an Olympus BX61 using Volocity software. Acknowledgments We are grateful to Evelyne Gout (CNRS) for her expertise in baculovirus expression. We acknowledge Charles Vragniau and Kathy Boardman, two very promising students. Thanks to Romain Vivès for his helpful corrections. We are indebted to Yves Rolland (Laboratoire Servier) for his encouragement during this integrin–HSPG project. References Anilkumar, N., Bhattacharya, A.K., Manogaran, P.S., Pande, G., 1996. Modulation of alpha 5 beta 1 and alpha V beta 3 integrins on the cell surface during mitosis. J. Cell. Biochem. 61 (3), 338–349. Asokan, A., Hamra, J.B., Govindasamy, L., Agbandje-McKenna, M., Samulski, R.J., 2006. Adeno-associated virus type 2 contains an integrin alpha5beta1 binding domain essential for viral cell entry. J. Virol. 80 (18), 8961–8969. Belin, M.T., Boulanger, P., 1993. Involvement of cellular adhesion sequences in the attachment of adenovirus to the HeLa cell surface. J. Gen. Virol. 74 (Pt 8), 1485–1497. Bergelson, J.M., Cunningham, J.A., Droguett, G., Kurt-Jones, E.A., Krithivas, A., Hong, J.S., Horwitz, M.S., Crowell, R.L., Finberg, R.W., 1997. Isolation of a common receptor for Coxsackie B viruses and adenoviruses 2 and 5. Science 275 (5304), 1320–1323. Berinstein, A., Roivainen, M., Hovi, T., Mason, P.W., Baxt, B., 1995. Antibodies to the vitronectin receptor (integrin alpha V beta 3) inhibit binding and infection of foot-and-mouth disease virus to cultured cells. J. Virol. 69 (4), 2664–2666. Byrnes, A.P., Griffin, D.E., 1998. Binding of Sindbis virus to cell surface heparan sulfate. J. Virol. 72 (9), 7349–7356. Cardin, A.D., Weintraub, H.J., 1989. Molecular modeling of protein– glycosaminoglycan interactions. Arteriosclerosis 9 (1), 21–32. Chen, Y., Maguire, T., Hileman, R.E., Fromm, J.R., Esko, J.D., Linhardt, R.J., Marks, R.M., 1997. Dengue virus infectivity depends on envelope protein binding to target cell heparan sulfate. Nat. Med. 3 (8), 866–871. Dechecchi, M.C., Tamanini, A., Bonizzato, A., Cabrini, G., 2000. Heparan sulfate glycosaminoglycans are involved in adenovirus type 5 and 2–host cell interactions. Virology 268 (2), 382–390. Dechecchi, M.C., Melotti, P., Bonizzato, A., Santacatterina, M., Chilosi, M., Cabrini, G., 2001. Heparan sulfate glycosaminoglycans are receptors sufficient to mediate the initial binding of adenovirus types 2 and 5. J. Virol. 75 (18), 8772–8780. Dransfield, I., Cabanas, C., Craig, A., Hogg, N., 1992. Divalent cation regulation of the function of the leukocyte integrin LFA-1. J. Cell Biol. 116 (1), 219–226. Fender, P., Ruigrok, R.W., Gout, E., Buffet, S., Chroboczek, J., 1997. Adenovirus dodecahedron, a new vector for human gene transfer. Nat. Biotechnol. 15 (1), 52–56. Fender, P., Schoehn, G., Foucaud-Gamen, J., Gout, E., Garcel, A., Drouet, E., Chroboczek, J., 2003. Adenovirus dodecahedron allows large multimeric protein transduction in human cells. J. Virol. 77 (8), 4960–4964. Fender, P., Boussaid, A., Mezin, P., Chroboczek, J., 2005. Synthesis, cellular localization, and quantification of penton-dodecahedron in serotype 3 adenovirus-infected cells. Virology 340 (2), 167–173.

163

Fry, E.E., Lea, S.M., Jackson, T., Newman, J.W., Ellard, F.M., Blakemore, W.E., Abu-Ghazaleh, R., Samuel, A., King, A.M., Stuart, D.I., 1999. The structure and function of a foot-and-mouth disease virus–oligosaccharide receptor complex. EMBO J. 18 (3), 543–554. Fuschiotti, P., Fender, P., Schoehn, G., Conway, J.F., 2006a. Development of the dodecahedral penton particle from adenovirus 3 for therapeutic application. J. Gen. Virol. 87 (Pt 10), 2901–2905. Fuschiotti, P., Schoehn, G., Fender, P., Fabry, C.M., Hewat, E.A., Chroboczek, J., Ruigrok, R.W., Conway, J.F., 2006b. Structure of the dodecahedral penton particle from human adenovirus type 3. J. Mol. Biol. 356 (2), 510–520. Garcel, A., Gout, E., Timmins, J., Chroboczek, J., Fender, P., 2006. Protein transduction into human cells by adenovirus dodecahedron using WW domains as universal adaptors. J. Gene Med. 8 (4), 524–531. Germi, R., Crance, J.M., Garin, D., Guimet, J., Lortat-Jacob, H., Ruigrok, R.W., Zarski, J.P., Drouet, E., 2002. Cellular glycosaminoglycans and low density lipoprotein receptor are involved in hepatitis C virus adsorption. J. Med. Virol. 68 (2), 206–215. Glotzer, M., 2001. Animal cell cytokinesis. Annu. Rev. Cell Dev. Biol. 17, 351–386. Hynes, R.O., 1992. Integrins: versatility, modulation, and signaling in cell adhesion. Cell 69 (1), 11–25. Jackson, T., Ellard, F.M., Ghazaleh, R.A., Brookes, S.M., Blakemore, W.E., Corteyn, A.H., Stuart, D.I., Newman, J.W., King, A.M., 1996. Efficient infection of cells in culture by type O foot-and-mouth disease virus requires binding to cell surface heparan sulfate. J. Virol. 70 (8), 5282–5287. Kern, A., Schmidt, K., Leder, C., Muller, O.J., Wobus, C.E., Bettinger, K., Von der Lieth, C.W., King, J.A., Kleinschmidt, J.A., 2003. Identification of a heparin-binding motif on adeno-associated virus type 2 capsids. J. Virol. 77 (20), 11072–11081. Lee, J.O., Bankston, L.A., Arnaout, M.A., Liddington, R.C., 1995. Two conformations of the integrin A-domain (I-domain): a pathway for activation? Structure 3 (12), 1333–1340. Lindahl, U., Kusche-Gullberg, M., Kjellen, L., 1998. Regulated diversity of heparan sulfate. J. Biol. Chem. 273 (39), 24979–24982. Maddox, A.S., Burridge, K., 2003. RhoA is required for cortical retraction and rigidity during mitotic cell rounding. J. Cell Biol. 160 (2), 255–265. Mathias, P., Galleno, M., Nemerow, G.R., 1998. Interactions of soluble recombinant integrin alphav beta5 with human adenoviruses. J. Virol. 72 (11), 8669–8675. Mould, A.P., 1996. Getting integrins into shape: recent insights into how integrin activity is regulated by conformational changes. J. Cell Sci. 109 (Pt 11), 2613–2618. Mould, A.P., Akiyama, S.K., Humphries, M.J., 1995. Regulation of integrin alpha 5 beta 1–fibronectin interactions by divalent cations. Evidence for distinct classes of binding sites for Mn2+, Mg2+, and Ca2+. J. Biol. Chem. 270 (44), 26270–26277. Nemerow, G.R., Stewart, P.L., 1999. Role of alpha(v) integrins in adenovirus cell entry and gene delivery. Microbiol. Mol. Biol. Rev. 63 (3), 725–734. Perron-Sierra, F., Saint Dizier, D., Bertrand, M., Genton, A., Tucker, G.C., Casara, P., 2002. Substituted benzocyloheptenes as potent and selective alpha(v) integrin antagonists. Bioorg. Med. Chem. Lett. 12 (22), 3291–3296. Rentsendorj, A., Xie, J., MacVeigh, M., Agadjanian, H., Bass, S., Kim, D.H., Rossi, J., Hamm-Alvarez, S.F., Medina-Kauwe, L.K., 2006. Typical and atypical trafficking pathways of Ad5 penton base recombinant protein: implications for gene transfer. Gene Ther. 13 (10), 821–836. Schoehn, G., Fender, P., Chroboczek, J., Hewat, E.A., 1996. Adenovirus 3 penton dodecahedron exhibits structural changes of the base on fibre binding. EMBO J. 15 (24), 6841–6846. Seidman, M.A., Hogan, S.M., Wendland, R.L., Worgall, S., Crystal, R.G., Leopold, P.L., 2001. Variation in adenovirus receptor expression and adenovirus vector-mediated transgene expression at defined stages of the cell cycle. Mol. Ther. 4 (1), 13–21. Shukla, D., Spear, P.G., 2001. Herpesviruses and heparan sulfate: an intimate relationship in aid of viral entry. J. Clin. Invest. 108 (4), 503–510. Smith, J.W., Piotrowicz, R.S., Mathis, D., 1994. A mechanism for divalent cation regulation of beta 3-integrins. J. Biol. Chem. 269 (2), 960–967.

164

P. Fender et al. / Virology 371 (2008) 155–164

Stewart, P.L., Chiu, C.Y., Huang, S., Muir, T., Zhao, Y., Chait, B., Mathias, P., Nemerow, G.R., 1997. Cryo-EM visualization of an exposed RGD epitope on adenovirus that escapes antibody neutralization. EMBO J. 16 (6), 1189–1198. Summerford, C., Samulski, R.J., 1998. Membrane-associated heparan sulfate proteoglycan is a receptor for adeno-associated virus type 2 virions. J. Virol. 72 (2), 1438–1445. Summerford, C., Bartlett, J.S., Samulski, R.J., 1999. AlphaVbeta5 integrin: a coreceptor for adeno-associated virus type 2 infection. Nat. Med. 5 (1), 78–82. Suomalainen, M., Nakano, M.Y., Keller, S., Boucke, K., Stidwill, R.P., Greber, U.F., 1999. Microtubule-dependent plus- and minus end-directed motilities are competing processes for nuclear targeting of adenovirus. J. Cell Biol. 144 (4), 657–672. Veiga, S.S., Elias, M., Gremski, W., Porcionatto, M.A., da Silva, R., Nader, H.B., Brentani, R.R., 1997. Post-translational modifications of alpha5beta1 integrin by glycosaminoglycan chains. The alpha5beta1 integrin is a facultative proteoglycan. J. Biol. Chem. 272 (19), 12529–12535.

Vives, R.R., Lortat-Jacob, H., Chroboczek, J., Fender, P., 2004. Heparan sulfate proteoglycan mediates the selective attachment and internalization of serotype 3 human adenovirus dodecahedron. Virology 321 (2), 332–340. Vives, R.R., Lortat-Jacob, H., Fender, P., 2006. Heparan sulphate proteoglycans and viral vectors: ally or foe? Curr. Gene Ther. 6 (1), 35–44. Wang, K., Huang, S., Kapoor-Munshi, A., Nemerow, G., 1998. Adenovirus internalization and infection require dynamin. J. Virol. 72 (4), 3455–3458. White, J.M., 1993. Integrins as virus receptors. Curr. Biol. 3 (9), 596–599. Wickham, T.J., Mathias, P., Cheresh, D.A., Nemerow, G.R., 1993. Integrins alpha v beta 3 and alpha v beta 5 promote adenovirus internalization but not virus attachment. Cell 73 (2), 309–319. Xiong, J.P., Stehle, T., Zhang, R., Joachimiak, A., Frech, M., Goodman, S.L., Arnaout, M.A., 2002. Crystal structure of the extracellular segment of integrin alpha Vbeta3 in complex with an Arg-Gly-Asp ligand. Science 296 (5565), 151–155. Yanagishita, M., Hascall, V.C., 1992. Cell surface heparan sulfate proteoglycans. J. Biol. Chem. 267 (14), 9451–9454.