Background information. Application of CPPs (cell-penetrating peptides) constitutes a promising strategy for the intracellular delivery of therapeutic molecules. The non-covalent approach based on the amphipathic peptide MPG has been successfully used to improve the delivery of biologically active macromolecules, both in cellulo and in vivo, through a mechanism independent of the endosomal pathway and mediated by the membrane potential.
Results. In the present study, we have investigated the first step of the cellular uptake mechanism of MPG and shown that both MPG and MPG—cargo complexes interact with the extracellular matrix through the negatively charged heparan sulfate proteoglycans. We demonstrated that initiation of cellular uptake constitutes a highly dynamic mechanism where the binding of MPG or the MPG—cargo to the extracellular matrix is rapidly followed by a remodelling of the actin network associated with the activation of the GTPase Rac1. We suggest that MPG-induced clustering of the glycosaminoglycan platform constitutes the ‘onset’ of the cellular uptake mechanism, thereby increasing membrane dynamics and membrane fusion processes. This process favours cell entry of MPG or MPG—DNA complexes, which is further controlled by the ability of MPG to induce a local membrane destabilization.
Conclusions. Although CPPs are taken up through different pathways and mechanisms, the initial step involves electrostatic interactions with the glycosaminoglycan platform, and the dynamics of associated membrane microdomains can be generalized to most non-viral delivery systems.
Over the past 5 years, a dramatic emergence of new potential therapeutic molecules has occurred, mainly due to the development of proteomics and genomics. However, these molecules still remain limited by their poor ability to enter cells. In order to render them more applicable for therapy in vivo, an increasing interest is being taken in the development of non-viral delivery methods (Niidome and Huang, 2002; Torchilin, 2005). For this aim, CPPs (cell-penetrating peptides) have been successfully used to improve the delivery of biologically active macromolecules, including nucleic acids, peptides and proteins, both in cell culture and in vivo (Järver and Langel, 2004; Gupta et al., 2005). Two major strategies have been developed. The first is based on natural or chimaeric PTDs (protein transduction domains) covalently linked to cargoes, the most representative of which include a peptide derived from the HIV-1 protein Tat (Fawell et al., 1994; Vivés et al., 1997; Frankel and Pabo, 1998; Schwarze et al., 1999), polyarginine (Wender et al., 2000; Futaki et al., 2001), the third helix of pAnt (antennapedia homeodomain protein) (Derossi et al., 1994) and transportan (Pooga et al., 1998). The second is based on primary amphiphatic peptides, such as MPG or Pep-1, which form stable non-covalent complexes with cargoes (Morris et al., 1997, 2001; Simeoni et al., 2003; Gros et al., 2006). The cellular uptake mechanism of PTDs has been shown to be essentially associated with endosomal pathways (Richard et al., 2003, 2005; Nakase et al., 2004; Wadia et al., 2004). However, clear evidence for distinct routes of cellular uptake have been reported, some of which are independent from the endosomal pathway and involve transmembrane potentials (Dom et al., 2003; Terrone et al., 2003; Thoren et al., 2003; Rothbard et al., 2004; Deshayes et al., 2005; Pujals et al., 2006).
The first contact that CPPs make with cells occurs through components of the extracellular matrix, the proteoglycans, which then trigger cellular uptake. Proteoglycans are heterogeneous proteins that carry one or more GAG (glycosaminoglycan) side chains, and that vary in size and shape (Kjellen and Lindahl, 1991; Esko and Selleck, 2002). Proteoglycans tend to form electrostatic interactions with molecules, which are primarily charge-mediated and dependent on the number of charges (Ruoslahti, 1998). Therefore they constitute a membrane ‘anchor’ through their GAG chains for a large variety of ligands (Sawitsky et al., 1996; Esclatine et al., 2001; Juliano, 2002; Couchman, 2003). Proteoglycans have been reported to be involved in different pathways controlling cell motility, shape or cell proliferation, which are directly associated with the dynamics of the cytoskeleton and actin network (Ruoslahti, 1988; Sawitsky et al., 1996; Esclatine et al., 2001; Juliano, 2002; Couchman, 2003; Beauvais and Rapraeger, 2004; Iozzo, 2005). The HSPGs (heparan sulfate proteoglycans) syndecan and glycan interact with the actin network via their cytoplasmic tail actin-binding protein and are both involved in the regulation of membrane ‘receptors’ and the cellular uptake of macromolecules (Couchman, 2003; Yoneda and Couchman, 2003; Beauvais and Rapraeger, 2004). GAG and HSPG play a central role in the translocation mechanism of polycationic carriers, liposomes (Mislick and Baldeschwieler, 1996; Kopatz et al., 2004) and PTDs (Belting, 2003). It has been suggested that GAG constitutes a cell-surface receptor for extracellular-peptide-carrier molecules that are associated or not with cargoes (Rusnati et al., 1999, 2001; Belting, 2003). The initial step for several CPPs, including the arginine-rich Tat and pAnt peptides, is associated with strong electrostatic interactions with negatively charged GAGs, which trigger their internalization via different endocytosis pathways depending on the presence and the nature of the cargo and the ability of the CPP to interact with lipids (Suzuki et al., 2002; Console et al., 2003; Nakase et al., 2004; Wadia et al., 2004; Richard et al., 2005).
MPG carriers are amphipathic peptides bearing a hydrophobic domain derived from the fusion domain of HIV-1 gp41 (glycoprotein 41), and a NLS (nuclear localization sequence) with 5 positive charges. MPGs are able to form stable complexes with nucleic acids and improve their cellular uptake, and therefore their associated biological response. They have been largely used to improve the delivery of antisense oligonucleotides (Morris et al., 1997), plasmid DNA (Morris et al., 1999a), siRNAs (small interferring RNAs) (Simeoni et al., 2003; Morris et al., 2004; Langlois et al., 2005) and peptides (Morris et al., 1999b). Two MPG peptides have been developed called MPG-α and MPG-β which differ in their secondary structure when within the lipid membrane (Deshayes et al., 2004a, 2004b). It has been reported that the uptake mechanism of biologically active MPG—cargo complexes is independent of the endosomal pathway and associated with the ability of MPG to interact with lipids and to induce local membrane destabilization (Deshayes et al., 2004a, 2004b). Nevertheless, the parameters associated with the initiation of the cellular uptake mechanism of these CPPs are still poorly understood. In the present work, we have investigated the first step of the cellular uptake mechanism of both MPG-β and MPG-α. We have demonstrated that both MPG and MPG—cargo complexes interact with the negatively charged GAG of the extracellular matrix. The binding of MPG to GAG triggers specific activation of Rac1 GTPase, which is associated with the remodelling of the actin network, thereby constituting the ‘onset’ of cellular uptake and promoting the entry into the cell of MPG or MPG—DNA complexes and, for most CPPs, by increased membrane fluidity. Our results suggest that, although cell entry of CPPs can follow different pathways, there are some common initial steps which involve the GAG platform and GTPase activation.
MPG peptides interact with HSPGs
An important step in the cellular uptake mechanism of CPPs is their interaction with the components of the cell surface. We first evaluated the capacity of MPG-β and MPG-α to interact with specific components of the extracellular matrix (Figure 1). As MPG delivery is based on the formation of stable complexes with cargoes, it was crucial to investigate the properties of complexes in parallel with those of free peptides. In order to investigate the impact of GAG on the stability of MPG—nucleic acid complexes, MPG-β and MPG-α peptides were associated, in water, with an 18-mer oligonucleotide at a charge ratio of 5, at which the complexes have been shown to be highly stable (Deshayes et al., 2004a). The complexes were then incubated in the presence of increasing concentrations of several soluble GAGs, including heparin, dextran sulfate and hyaluronic acid, which have different numbers of sulfate groups per disaccharide unit. The integrity of complex stability was analysed by electrophoresis on agarose gels (Figure 1). When DNA was associated with MPG-β (Figure 1A) or MPG-α (Figure 1B) stable complexes did not appear on the gel. The presence of heparin or dextran sulfate, both sugars bearing a high density of negative charges, induced a GAG-concentration-dependent dissociation of the complexes which allowed free DNA to migrate into the gel. In contrast, hyaluronic acid, which has a low-charge density and no sulfate group, had no effect on complex stability. These results indicate that both of the MPG peptides can interact with negatively charged GAG, a major component of proteoglycans of the cell surface and extracellular matrix, and that a high concentration of negatively charged GAG may affect the stability of the MPG—cargo complex.
MPG-β, MPG-α and TAT peptides induce actin network remodelling
In order to further understand the first step of cellular uptake, we investigated the impact of the presence of CPPs on the organization of the actin network. Recently, Nakase et al. (2004) established that arginine-rich peptides, as well as penetratin to a certain extent, were able to induce significant actin network rearrangement when incubated with HeLa cells. Therefore the effect of MPGs and TAT peptides on the actin network organization was evaluated using both the HeLa and C2C12 cell lines. Cells were treated with 5 μM of MPG-β, MPG-α or TAT peptides for 20 min, and the actin network was then revealed with rhodamine-conjugated phalloidin (Figure 2A). In control experiments, numerous stress fibres were detected in the cytosol and a few lamellipodia became visible at the cell periphery in both C2C12 and HeLa cells. In contrast, cells treated with either MPG-β, MPG-α or TAT peptides showed a marked modification of their actin network. The number of stress fibres was reduced, whereas F-actin accumulated around the cells and large lamellipodia were formed at the cell periphery. Similar patterns were observed with the NIH3T3 and MEF cell lines upon incubation with the different peptides (data not shown), suggesting that the mechanism was independent of cell lines. Therefore only HeLa cells were used for the rest of the present study.
To better characterize peptide-carrier-mediated actin remodelling, the kinetics of lamellipodia formation and stress fibre decrease were evaluated (Figures 2B and 2C). The percentage of cells showing remodelling of actin was estimated from the statistical analysis of 150 cells incubated with MPG-β, MPG-α or TAT peptides (5 μM concentration). The rate of actin network modification and lamellipodia formation were dependent on the concentration of peptide used and was maximal at the concentration used (5 μM). We demonstrated that the process started within the first 5 min and a maximal effect was obtained after 20 min, with more than 90% of the cells presenting lamellipodia (Figure 2B), which correlates well with the loss of stress fibres (Figure 2C). It was observed that actin network modifications were fully reversible and cells returned back to normal after of approx. 1 h, suggesting that changes in the actin network are not due to any toxicity associated with the carrier peptides.
MPG-β or MPG-α in complex with a nucleic acid induce actin network remodelling
In contrast with TAT, which is covalently linked to its cargo, MPG peptides form non-covalent complexes with cargoes, and it is therefore essential to investigate the effect of MPG—nucleic acid complexes on the actin network. MPG-β–DNA or MPG-α–DNA complexes (with a charge ratio of 5), which were previously shown to be biologically active (Simeoni et al., 2003), were added to cultured cells for 20 min in serum-free medium prior to actin network analysis. As shown in Figure 3(A) (panels b and c), both complexes altered actin organization to the same degree as the free MPG peptide, as the percentage of cells presenting lamellipodia was not affected by the interaction of MPG with the DNA (Figure 3B). Taken together, these results show that MPG-β–DNA and MPG-α–DNA complexes are still able to interact with cell-surface components and suggest that the MPG—DNA particle contains exposed charges. Moreover, that actin network remodelling was fully reversible with all of the MPG peptides used, thereby confirming the lack of toxicity of the peptides.
MPG-β- or MPG-α- and TAT-peptide-associated lamellipodia formation is mediated by interactions with cell-surface GAG
We demonstrated that both free and cargo-associated, MPG-β, MPG-α and TAT peptides induced lamellipodia formation in HeLa cells (Figure 2A). To evaluate the role of the interaction between these peptides and the cell surface, GAG, heparin or hyaluronic acid were incubated with cells concomitantly with the peptides. The percentages of cells showing lamellipodia were then evaluated. Figure 4(A) shows actin network modifications with MPG-β (panels d—f) and TAT (panels g—i). Heparin and hyaluronic acid alone have no effect on the actin network (Figure 4A, panels b and c). MPG can induce lamellipodia formation when complexed with DNA at the same extent as that observed for peptide alone (Figure 4B). As shown in Figure 4(B), the MPG-β, MPG-α and TAT peptides induced formation of lamellipodia in more than 90% of the cells. This was completely abolished by the presence of heparin (Figure 4B, black bars), whereas it was quite unaffected in the presence of the same amount of hyaluronic acid (Figure 4B, white bars). These data strongly suggest that the interaction of MPG-β, MPG-α (free or complexed) and TAT peptide with cell-surface sugars is required for lamellipodia formation and actin network modification. The effect of the interaction between CPPs and GAG on actin remodelling was further evaluated using CHO-745 cells (Figures 5A and 5B), which do not produce detectable levels of proteoglycans, because they lack xylosyltransferase, an enzyme required for the initiation of GAG synthesis (Rostand and Esko, 1997). We demonstrated again that MPG-β and TAT peptides are able to induce formation of lamellipodia in more than 90% of the wild-type CHO-K1 cells (Figure 5A, panels b and c). In contrast, no significant changes in the actin network were observed in CHO-745 cells in the presence of MPG-β and TAT peptides (Figure 5A, panels e and f), and the percentage of cells presenting lamellipodia was not significantly affected with respect to the control (Figure 5A, panels a—d). The fact that MPG was reported to co-localize with its cargo inside the cell and to improve its intracellular trafficking ruled out the possibility that GAG dissociated the MPG—cargo complex and strengthened the conclusion that actin remodelling favours the entry of MPG—cargo complexes into cells.
Activity of Rac1 GTPase is required for CPP-induced lamellipodia formation
The members of the Rho GTPase family are key regulatory molecules that link the cell surface to the organization of the actin cytoskeleton (Hall, 1998; Etienne-Manneville and Hall, 2005). The Rho GTPases RhoA, Cdc42 and Rac-1 regulate stress fibres, filopodia and lamellipodia formation respectively (Jaffe and Hall, 2005). In order to determine the involvement of these Rho GTPases in the CPP-induced actin remodelling, we studied the impact of their dominant-negative GDP-bound (Cdc42N17, RhoAN19 and Rac1N17) or dominant-positive GTP-bound (RhoAV14, Rac1V12 and Cdc42V12) forms on the cellular response to the presence of peptides. HeLa cells were transfected with GFP (green fluorescent protein)–Cdc42N17, GFP—RhoAN19, GFP—Rac1N17 or GFP—RhoAV14 using a transfection reagent that does not affect the actin network (FuGENE® 6) (Figures 6 and 7). After transfection (24 h), cells were treated with MPG-β, MPG-α and TAT peptides, and the changes to the F-actin cytoskeleton were observed by rhodamine-labelled phalloidin staining. Statistical analysis of 150 cells was carried out with GFP-positive cells treated with MPG-β, MPG-α or TAT (Figure 6D). As a control, we showed that more than 90% of HeLa cells transfected with pEGFP-N1 control vector and treated with MPG-β, MPG-α or TAT peptides produced lamellipodia (Figures 6 and 7). The expression of GFP—RhoAN19 and GFP—Cdc42N17 had no effect on the organization of the actin network in control cells and did not prevent peptide-induced lamellipodia formation (Figures 6A and 6B). As reported in Figure 7, the expression of the dominant-negative GFP—Rac1N17 completely suppressed peptide-induced actin network remodelling, and this effect is dependent on the level of expression of the dominant negative (Figure 7B). These data show that the activity of Rac1 is required for CPP-induced actin cytoskeleton reorganization. In contrast, expression of the dominant-positive GFP—RhoAV14 which induced actin condensation, completely suppressed peptide-carrier-induced actin network remodelling (Figure 6C).
MPG-β, MPG-α and TAT peptides increase Rac1 activity
Our data suggested that the Rac1 GTPase mediates actin network remodelling in response to MPG and TAT peptide. In order to confirm the involvement of Rac1, we have investigated its activity using an RBD (Rho-binding domain)-pull-down assay, following a 20 min incubation of the peptides in HeLa cells (Figure 8). Cell lysates were incubated with GST (glutathione S-transferase) fused to the CRIB (Cdc42/Rac interacting binding) domain of the Rac1 effector molecule Pak (p21-activated kinase) (GST—Pak-CRIB) and activated, then the amount of GTP-loaded Rac1 bound to GST—Pak-CRIB was analysed by Western blotting (Charrasse et al., 2006). Strong activation of endogenous Rac1 was detected in HeLa cells treated with MPG-β, MPG-α or TAT compared with control cells. Since the total amount of Rac1 was identical in all conditions, Rac1 appears to be activated by all of the peptides, prior to actin remodelling.
Lamellipodia formation induced by other CPPs and transfection reagents
Numerous non-viral delivery methods are currently being used by many laboratories. Therefore it is of interest to evaluate their putative effect on rearrangement of the actin network. In this respect, reagents, including CPPs (hCT, R4-R5 and R9) and commercially available transfection reagents (FuGENE® 6, Lipofectamine™ and Oligofectamine™), were assessed. As shown in Figure 9, the hCT, R4-R5 and NLS peptides had no effect on lamellipodia formation. In contrast, the R9 peptide promoted formation of actin protrusions, which agreed with a previous study (Nakase et al., 2004). In contrast, we demonstrated that cationic lipid-based transfection reagents, FuGENE® 6 and Oligofectamine™ did not alter the actin network in contrast with Lipofectamine™, which induced lamellipodia formation, again indicating a direct correlation between the charge of the transfection reagents and actin remodelling.
CPPs constitute a very promising technology to improve the cellular uptake of biologically active molecules (Schwarze et al., 1999; Gupta et al., 2004; Järver and Langel, 2004; Deshayes et al., 2005; Pujals et al., 2006). Although the cellular uptake mechanism for several CPPs has been shown to be mainly associated with the endosomal pathway (Richard et al., 2003, 2005; Wadia et al., 2004), evidence for several distinct cellular entry pathways have been reported (Deshayes et al., 2005). The cellular uptake mechanism of CPPs begins with interactions with the extracellular matrix, more specifically with the cell-surface proteoglycans which serve as a ‘capture platform’, thereby triggering the ‘onset’ of internalization (Rusnati et al., 1999). Therefore a full understanding of this first step is critical to further improve selectivity and efficiency of CPPs.
One of the major differences between CPPs resides in their mode of interaction with the biomolecule that they deliver, which may be either covalent or simply complexed. In the case of the covalent types of CPP, TAT peptide (Silhol et al., 2002; Console et al., 2003) polyarginine and pAnt (Nakase et al., 2004), their interactions with the extracellular matrix have been reported to be primarily electrostatic (Rusnati et al., 1999). In the present work, we have demonstrated that the MPG peptide family uptake mechanism is also initiated by electrostatic interactions with the extracellular matrix. One hypothesis for the mechanism of CPPs implies that, following binding to the extracellular matrix, CPPs exploit the turnover of the cell membrane to be internalized (Brooks et al., 2005). In the present study, we have showed that initiation of the uptake mechanism of MPG carriers, and probably of most of CPPs, is highly dynamic. The binding of MPG or MPG—cargo to the GAG platform is followed by a selective activation of the GTPase Rac1, which allows the remodelling of the actin network (Figure 10). We propose that both GTPase activation and actin remodelling constitute the ‘onset’ of the internalization mechanism and have a major impact on membrane fluidity, thereby promoting the cell entry of MPG or MPG—DNA complexes.
The initial contacts of MPGs in their free and cargo-complex forms with cell-surface GAG are mainly electrostatic and directly correlated with the number of charges borne by the peptide. A minimum of five charges is required for significant induction of GAG clustering. These initial contacts with the extracellular matrix do not depend on the secondary structure of the peptide, as shown for MPG-α and MPG-β which adopt different secondary structures within the lipid membrane (Deshayes et al., 2004b), and yet exhibit a similar propensity to interact with GAGs and to induce their clustering. A similar mechanism was observed with both TAT and penetratin peptides, which adopt either no specific conformation or a variety of different conformations, depending on the nature of the cargo during cellular uptake (Thoren et al., 2003).
Proteoglycans play an essential role in the regulation of cell-surface microdomains, and evidence of direct relationships between cytoskeletal organization and activation of small GTPases has been clearly established (Conner and Schmid, 2003; Couchman, 2003; Eitzen, 2003). HSPGs and syndecans, which are the major components of the extracellular matrix, constitute anchors for many external molecules and pathogens on the host cell surface (Esclatine et al., 2001; Yoneda and Couchman, 2003). Their clustering triggers cytoskeleton remodelling upon activation of PKC (protein kinase C) and Rho/Rac GTPases, which controls the dynamics of cholesterol-rich ‘raft’ microdomains, and therefore the ligand-binding and cellular uptake pathways (Dehio et al., 1998; Saoncella et al., 1999; Couchman, 2003; Beauvais and Rapraeger, 2004; Tkachenko et al., 2004).
Interestingly, the present study and previous work (Nakase et al., 2004; Wadia et al., 2004) have shown that interaction of other CPPs, or polycationic carriers, such as Lipofectamine™ and polyethylenimide (Belting and Petersson, 1999; Wiethoff et al., 2001), with the extracellular matrix also induces GAG clustering, followed by remodelling of the actin network. Accordingly, the initial step involving the GAG platform and the dynamics of cholesterol-rich microdomains can be generalized to apply to most non-viral delivery systems.
Actin network remodelling and the related dynamics of cholesterol-rich microdomains has a major impact on the mechanism of cellular penetration, as it may favour either endocytosis pathways, such as macropinocytosis, as previously reported for TAT and polyarginine peptides (Nakase et al., 2004; Wadia et al., 2004), or increase membrane fluidity to improve non-endosomal cellular uptake, as reported for MPG (Simeoni et al., 2003; Deshayes et al., 2004a). The involvement of Rac/Rho GTPase activity in the control of membrane fluidity, cytoskeleton remodelling and membrane fusion have been documented (Hall, 1998; Conner and Schmid, 2003; Eitzen, 2003; Beauvais and Rapraeger, 2004). We have demonstrated that MPGs induce specific activation of the GTPase Rac1. Rac1 appears to be required for several pathways involving membrane fusion; Rac1 activation upon HIV-1 or paramyxovirus infections stimulated actin network remodelling, which primarily increased the membrane fluidity and membrane fusion (Pontow et al., 2004; Schowalter et al., 2006). The fact that the kinetics of Rac1 activation and actin network remodelling occur within the first 5 min and are fully reversible after 60 min is in agreement with the reported cellular uptake of MPG and the biological response associated with the cargo (Simeoni et al., 2003). MPG was reported to enter the cell through a non-endosomal mechanism involving a transient membrane deorganization associated with its folding into a β-sheet structure upon interaction with phospholipids (Simeoni et al., 2003; Deshayes et al., 2004a), and we suggest that the increase of membrane fluidity associated to Rac1 activation favours MPG cellular uptake and therefore its efficiency. However, activation of Rac1 and cytoskeleton remodelling initiates membrane reorganization, but cannot be considered as the only control of the cell entry of CPPs. Indeed, the lack of correlation between the strong ability to induce actin remodelling and the efficiency of transduction or transfection points to the existence of other factors, involving membrane crossing and cargo release in the cell, as well as the associated biological response which is dependent on the intrinsic properties of the carrier. The divergence between cellular uptake mechanisms is dependent on the intrinsic properties of a given peptide carrier and driven by its ability to interact with the lipid membrane and to adopt a secondary structure within the membrane. For cationic CPPs, including polyarginine and TAT peptides, electrostatic interactions with the extracellular matrix and its clustering are prerequisites for cell entry (Wadia et al., 2004; Richard et al., 2005; Ziegler et al., 2005). In contrast, the MPG carrier was shown to interact with membrane lipids through both electrostatic and hydrophobic contacts and to form stable complexes with lipids (Deshayes et al., 2004a; Li et al., 2004). This property is shared with only a few CPPs, including penetratin (Dom et al., 2003; Thoren et al., 2003) and transportan (Magzoub and Gräslund, 2004), whose interaction with lipids induces folding of the peptide into α-helical or β-sheet structures that directly control their cellular uptake or escape from early endosomes. This is confirmed by the fact that MPG can enter cells devoid of GAG and that an MPG peptide containing only four charges (Simeoni et al., 2003) is still able to enter cells.
We have demonstrated a direct involvement of Rac1 in controlling or enhancing the rate of the initial step of the CPP uptake mechanism. Although the uptake mechanism is dependent on the inherent properties of the peptide, the role of Rac1 in the global process is crucial. Implication of Rac1 activation in the biological response of the cargoes and the release of the active form of the cargo is an essential point that is currently under further investigation in our laboratory.
Materials and methods
MPG-α (acetyl-GALFLAFLAAALSLMGLWSQPKKKRKV-cysteamide), MPG-β (acetyl-GALFLGFLGAAGSTMGAWSQPKKKRKV-cysteamide), peptide NLS (WKKKRKV-cysteamide) and polyarginine peptide R4-R5 were synthesized with the Fmoc/tBu strategy, as described previously (Morris et al., 1997; Deshayes et al., 2004b), on a Pioneer™ Peptide Synthesizer (Applied Biosystems, Foster City, CA, U.S.A.), starting from Fmoc-PAL-PEG-PS resin at a scale of 0.2 mmol. The coupling reactions were performed with 0.5 M HATU (O-hexafluorophospho-[7-azabenzotriazol-1-yl]-N,N,N′,N′-tetramethyluronium) in the presence of 1 M of DIEA (di-isopropylethylamine). Protecting group removal and final cleavage from the resin were carried out with trifluoroacetic acid/phenol/water/thio-anisol/ethanedithiol (82.5:5:5:5:2.5) for 3 h 30 min. Peptides were N-acetylated and bear a cysteamide group (-NH-CH2-CH2-SH) at their C-terminus. The crude peptide was purified by RP-HPLC on a C18 column (Interchrom UP5 WOD/25M Uptispere 300 5 ODB; 250 mm×21.2 mm). Electrospray ionization mass spectra were in agreement with the proposed structures. TAT (GRKKRRQRRRC) was obtained from NeoMPS (Strasbourg, France). The hCT peptide (LGTYTQDFNKFHTFAQTAIGVGAP) was kindly provided by Dr A. Beck-Sicklinger (Institute of Biochemistry, University of Leipzig, Leipzig, Germany), and the polyarginine R9 peptide was kindly provided by Dr E. Vivès (University of Montpellier, Montpellier, France). Dimethylformamide, piperidine, trifluoroacetic acid and CH3CN (HPLC grade) were purchased from SDS (Peypin, France). The protected amino acids were from SENN Chemicals (Dielsdorf, Switzerland). HATU, DIEA and resins were from Applied Biosystems, and thio-anisol and ethanedithiol were from Sigma—Aldrich.
Complex formation and electrophoretic mobility-shift assay
The oligonucleotide (5′-TCTCCCAGCGTGCGCCAT-3′) was obtained from MWG-Biotech (Courtaboeuf, France). One hundred pmol of 18-mer oligonucleotides were mixed with 1.8 μmol of MPG-β or MPG-α peptides in water, corresponding to a ‘+/−’ charge ratio (positive charges of the peptide versus negative charges of the DNA) of 5. After 30 min of incubation at room temperature, increasing concentrations of either heparin, dextran sulfate or hyaluronic acid (from 10 ng to 10 μg) were added to the complex solution. Samples were then loaded on to an agarose gel (1%) for the analysis of complexes. For all experiments with cell culture, complexes were formed as described above and were added directly to the serum-free culture medium.
Cell culture and transfection
HeLa and C2C12 cell lines were obtained from the European Collection of Cell Cultures (Salisbury, U.K.) and were maintained in DMEM (Dulbecco's modified Eagle's medium) (Invitrogen, Cregy Pontoise, France) supplemented with 10% fetal calf serum and 100 μg/ml penicillin and streptomycin. CHO-K1 and CHO-745 cell lines were obtained from Sigma and maintained in Ham's F12 medium supplemented as above with 10% fetal calf serum and 100 μg/ml penicillin and streptomycin.
Adherent HeLa cells, plated on glass coverslips, were transfected with pEGFP-N1 (BD Bioscience/Clontech, France), pEGFP-Cdc42N17, pEGFP-Rac1N17, pEGFP-RhoAN19, pEGFP-RhoAV14, pEGFP-Rac1V12 or pEGFP-Cdc42V12 using FuGENE® 6 (Roche, Meylan, France) according to the manufacturer's protocol. The different constructs encoding the GTPases have been described previously (Gauthier-Rouviere et al., 1998; Charrasse et al., 2006). After 10 h, the medium was removed and 14 h later cells were treated with MPG-β, MPG-α or TAT peptides, as indicated in the Figure legends. Heparin, hyaluronic acid and dextran sulfate were purchased from Sigma. Lipofectamine™ and Oligofectamine™ were obtained from Invitrogen.
HeLa, C2C12, CHO-K1 or CHO-745 cells lines were cultured on coverslips and treated in serum-free medium with the MPG-β, MPG-α or TAT peptides (5 μM) for 5, 20, 40, 60 and 120 min, or water as a negative control (the volume of water never exceeds 1% of the total volume). Cells were then fixed with 3.7% paraformaldhehyde for 20 min and permeabilized with 0.1% Triton X-100. F-actin was labelled with rhodamine-conjugated phalloidin (0.5 units/ml, Sigma).
In situ detection of the active Rho GTPase Rac1
HeLa cells lysis was performed in 25 mM Hepes, pH 7.5, 1% Nonidet P40, 10 mM MgCl2, 100 mM NaCl, 5 mM NaF, 5% glycerol, 1 mM PMSF and a cocktail of protease inhibitors (Sigma). Cleared lysates were incubated with GST—Pak-CRIB bound to glutathione-coupled Sepharose beads for 30 min at 4°C. Beads were washed three times in 25 mM Hepes, pH 7.5, 1% Nonidet P40, 30 mM MgCl2, 40 mM NaCl and 1 mM dithiothreitol (Charrasse et al., 2006). The bound proteins were eluted from beads by the addition of Laemmli sample buffer, resolved by SDS/PAGE (12.5% gel) and immunobloted using an anti-Rac1 antibody (Santa Cruz Biotechnology). The total amount of Rac1 in the whole lysates was determined by Western blotting.
This work was supported in part by the CNRS (Centre National de la Recherche Scientifique) and by grants from the ANRS (Agence Nationale de Recherche sur le SIDA) and the EU (grant QLK2-CT-2001-01451 and grant LSHB-CT-2003-503480/TRIoH). S.G.-C. and C.G. were supported by fellowships from the European Community and SIDACTION respectively. We thank Anne Blangy and Eric Vivés for fruitful discussion and supplying the peptides. We thank May C. Morris for critical reading and proofreading of the manuscript.