Poliovirus 2b insertion into lipid monolayers and pore formation in vesicles modulated by anionic phospholipids

Biochimica et Biophysica Acta 1778 (2008) 2621–2626

Contents lists available at ScienceDirect

Biochimica et Biophysica Acta j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / b b a m e m

Poliovirus 2b insertion into lipid monolayers and pore formation in vesicles modulated by anionic phospholipids Aitziber Agirre a,1, Maier Lorizate a,2, Shlomo Nir b, José L. Nieva a,⁎ a b

Unidad de Biofísica (CSIC-UPV/EHU) and Departamento de Bioquímica, Universidad del País Vasco, Aptdo. 644, 48080 Bilbao, Spain Seagram Center for Soil and Water Sciences, Faculty of Agricultural, Food and Environmental Quality Sciences, The Hebrew University of Jerusalem, Rehovot 76100, Israel

a r t i c l e

i n f o

Article history: Received 21 April 2008 Received in revised form 9 June 2008 Accepted 18 June 2008 Available online 25 June 2008 Keywords: Enterovirus 2B Pore-forming protein Viroporin Protein–lipid interaction

a b s t r a c t Enterovirus 2B viroporin has been involved in membrane permeabilization processes occurring late during cell infection. Even though 2B lacks an obvious signal sequence for translocation, the presence of a Lys-based amphipathic domain suggests that this product bears the intrinsic capacity for partitioning into negatively charged cytofacial membrane surfaces. Pore formation by poliovirus 2B attached to a maltose-binding protein (MBP) has been indeed demonstrated in pure lipid vesicles, a fact supporting spontaneous insertion into and direct permeabilization of membranes. Here, biochemical evidence is presented indicating that both processes are modulated by phosphatidylinositol and phosphatidylserine, the main anionic phospholipids existing in membranes of target organelles. Insertion into lipid monolayers and partitioning into phospholipid bilayers were sustained by both phospholipids. However, MBP-2B inserted into phosphatidylserine bilayers did not promote membrane permeabilization and addition of this lipid inhibited the leakage observed in phosphatidylinositol vesicles. Mathematical modelling of pore formation in membranes containing increasing phosphatidylserine percentages was consistent with its inhibitory effect arising from a higher reversibility of MBP-2B surface aggregation. These results support that 2B insertion and pore-opening are mechanistically distinguishable events modulated by the target membrane anionic phospholipids. © 2008 Elsevier B.V. All rights reserved.

1. Introduction Infection by certain members of the Picornaviridae family such as entero- and rhinoviruses causes cell plasma membrane and endomembrane permeabilization [1–3]. Both, the diffusion across plasma membrane of ions and small molecules under ca. 1000 Da and the increase in cytoplasmic calcium ion levels have been described to occur during late phases of the infection cycle [2,4–7]. The mechanism underlying this phenomenon has been studied particularly in detail for two enteroviruses: the poliovirus [2,4,8–13] and the coxsackievirus [3,6,7,14–16]. Membrane permeabilization requires ongoing synthesis of viral products to evolve and can be partly mimicked by the individual expression of the non-structural 2B and 2BC proteins [10,11,14–16]. Several lines of evidence including prediction and mutagenesis studies, as well as recent biophysical characterization in model membranes, suggest that the capacity to translocate into membranes and enhance permeability therein primarily resides in the 2B moiety [7,10,13,17–19].

⁎ Corresponding author. Tel.: +34 94 601 3353; fax: +34 94 601 3360. E-mail address: [email protected] (J.L. Nieva). 1 Present address: Elhuyar Foundation, 20170 Usurbil, Spain. 2 Present address: Abteilung Virologie, Universitätsklinikum Heidelberg, im Neuenheimer Feld 324, D-69120 Heidelberg, Germany. 0005-2736/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.bbamem.2008.06.013

Early hydrophobicity analyses of 2B sequence put forward the existence of two potential helical transmembrane domains (TMD) intervened by a short interfacial turn [7,18]. These elements are predicted to combine into a membrane-embedded “hairpin” (Fig. 1). When expressed alone in mammalian cells the 2B–2BC proteins have been described to localize at different endomembrane systems including the cytofacial side of the plasma membrane [6], the Golgi complex [12,16] and the endoplasmic reticulum [6,15]. These findings suggest that 2B–2BC have the potential to become integral to cell and model membranes. However, little is known on the factors governing biogenesis and sorting of membrane-anchored 2B species during infection. Viral genomic RNA is initially traduced into a single polyprotein precursor, which is afterwards proteolitically processed giving rise to individual mature 2B and 2BC products. No signal peptide has been described for 2B sequence and, therefore, it is likely that spontaneous insertion drives translocation of this protein into the diverse cell membranes. The aim of the present work is to advance our knowledge on the translocation and refolding mechanisms leading to 2B pore-opening in target membranes. In a recent review article van Meer et al. [20] survey the distribution of the main lipids among the mammalian cell membrane systems. Phosphatidylserine (PS) and phosphatidylinositol (PI) are the prevalent species of the anionic phospholipid fraction of 2B target membranes. Thus, we assayed the effect of these lipids using model membranes and the poliovirus 2B sequence attached to a

2622

A. Agirre et al. / Biochimica et Biophysica Acta 1778 (2008) 2621–2626

20 mM NaCl and 5 mM Hepes were obtained by separating the unencapsulated material by gel filtration in a Sephadex G-75 column eluted with 5 mM Hepes, and 100 mM NaCl (pH 7.4). Osmolarities were adjusted to 200 mOsm in a cryoscopic osmometer (Osmomat 030, Gonotec, Berlin, Germany). Fluorescence measurements were performed by setting the ANTS emission at 520 nm and the excitation at 355 nm. A cutoff filter (470 nm) was placed between the sample and the emission monochromator. The absence of leakage (0%) corresponded to fluorescence of the vesicles at time zero; 100% leakage was taken as the fluorescence value obtained after addition of Triton X-100 (0.5% v/v). The degree of permeabilization was then inferred from the following equation: kLeakage = ½ðFf −F0 Þ=ðF100 −F0 Þ  100

Fig. 1. Schematic diagram and sequence of poliovirus (PV-1) 2B viroporin membraneinserting helical hairpin. Encircled residues at the N-terminal helix denote the conserved Lys in picornaviruses encoding lytic 2B [3,18]. The Cys residue titrated with the NBD fluorophore is also indicated within the C-terminal helix.

maltose-binding protein (MBP), as an N-terminal solubilization partner. This fusion protein was previously shown to assemble into stable membrane pores that allowed the passage of solutes of MW under 1000 [17]. We report here biochemical data indicating that MBP-2B membrane insertion and pore-opening can be distinctly modulated by PI and PS anionic lipids. Although required for efficient insertion into lipid monolayers and water–bilayer partitioning, a negative surface charge was not sufficient to sustain membrane permeabilization by this construct. We conclude that 2B pore-forming activity can be regulated by the negatively charged phospholipid species existing at the cytofacial monolayers of the target membranes. 2. Materials and methods 2.1. Materials Phosphatidylcholine (PC), phosphatidylinositol (PI) and phosphatidylserine (PS) were purchased from Avanti Polar Lipids (Birmingham, AL, USA). N,N′-dimethyl-N-(iodoacetyl)-N′-(7-nitrobenz-2-oxa-1, 3-diazol-4-yl)ethylenediamine (IANBD), 8-aminonaphtalene-1,3,6-trisulfonic acid sodium salt (ANTS) and p-xylenebis(pyridinium)bromide (DPX) were from Molecular Probes (Junction City, OR, USA). 2.2. 2B protein expression and purification Plasmids encoding MBP-2B fusion proteins were constructed by sub-cloning poliovirus-2B encoding DNA into pMALc2 vector (New England Biolabs Inc., Beverly, MA). The constructs were expressed in Escherichia coli BL21 as the host strain and purified as previously described [17].

where Ff is the fluorescence determined after protein addition, F0 is the initial fluorescence of the intact LUV suspension, and F100 is the fluorescence value after the addition of Triton X-100. The extent of ANTS release was determined after fluorescence intensity reached a plateau (routinelyN60 min). Vesicles were stored at 4 °C and used within 1–2 days. No spontaneous leakage of entrapped material was observed for vesicles stored at 4 °C for at least 1 week. Parallel controls confirmed that MBP protein devoid of the 2B moiety did not induce by itself ANTS leakage under the experimental conditions used in our study. 2.4. Monolayer penetration Surface pressure was determined in a fixed-area circular trough (μTrough S system, Kibron, Helsinki). Measurements were carried out at room temperature and under constant stirring. The aqueous phase consisted of 1 ml 5 mM Hepes, and 100 mM NaCl (pH 7.4). Lipid mixtures, dissolved in chloroform, were spread over the surface and the desired initial surface pressure (π0) was attained by changing the amount of lipid applied to the air–water interface. Protein was injected into the subphase with a Hamilton microsyringe. At the used concentrations protein alone induced negligible increase in surface pressure at the air–water interface. 2.5. Binding to membranes MBP-2B protein contains two single Cys residues within the 2B sequence (positions 73 and 86). This allowed the specific labeling of 2B with thiol-reacting NBD fluorophore. Labeling of MBP-2B to render MBP-2B-NBD as described by Agirre et al. [17] did not appreciably affect the capacity of the protein to induce leakage of contents from PI vesicles. For MBP-2B-NBD excitation and emission maxima were observed at 481 and 529 nm respectively. Partitioning curves were produced by monitoring the fractional change in emitted NBD fluorescence from MBP-2B-NBD titrated with increasing lipid concentrations. Fluorescence was recorded in a Perkin Elmer MPF-66 spectrofluorimeter with excitation and emission monochromators set at 480 and 530 nm respectively and corresponding slits of 5 nm. Values of Kx were determined by fitting the experimental values to the hyperbolic function: F=F0 = 1 +

2.3. Permeability assays Large unilamellar vesicles (LUV) were prepared according to the extrusion method of Hope et al. [21] in 5 mM Hepes, and 100 mM NaCl (pH 7.4). The lipid concentrations of the liposome suspensions were determined by phosphate analysis. The average vesicle size did not change upon treatment with MBP-2B at any of the tested doses as revealed by quasielastic light scattering determinations. Release of vesicular contents into the medium was monitored by the ANTS/DPX assay [22]. LUV containing 12.5 mM ANTS, 45 mM DPX,

ð1Þ

½ðFmax =F0 Þ−1½L Kd + ½L

ð2Þ

where [L] is the lipid concentration and Kd is the apparent dissociation constant, or lipid concentration at which the bound protein fraction is 0.5. Thus, Kx = [W] / Kd where [W] is the molar concentration of water. 2.6. Analysis of leakage via pore formation The mathematical model employed for the analysis of MBP-2Binduced ANTS/DPX leakage has been previously described in detail [23,24]. The pore model assumes that vesicle-bound proteins self

A. Agirre et al. / Biochimica et Biophysica Acta 1778 (2008) 2621–2626

2623

aggregate. When an aggregate within a membrane has reached a critical size, i.e., it consists of M monomers, a pore can be created within the membrane, and leakage of encapsulated molecules can occur. It is also assumed that the process of protein binding is rapid and once a pore has been formed in a vesicle, all its contents will quickly leak out. MBP-2B proteins do not interchange among vesicles [17], therefore leakage must be characterized by an all-or-none mechanism, i.e., the vesicle population will be comprised of those not leaking and those losing their entire contents. Further, the leakage process must end after a certain period to give a final extent of leakage, which depends on protein-to-lipid ratios. The rate and extent of leakage are assumed to be limited by the rate and extent of formation of surface aggregates of M or more proteins that form the pore. According to the model, surface aggregation of the proteins to assemble the pore is not irreversible and thus depends on Ks =C /D, where C and D denote on and off rate constants of surface aggregation. 3. Results Fig. 1 displays the sequence of the poliovirus 2B helical hairpin, which is postulated to integrate spontaneously into membranes of target organelles. The amino-terminal helix contains three conserved Lys residues that confer partial amphipathicity to this domain [14,18]. Supporting its lytic role, we have recently demonstrated that a peptide designed to mimic the N-terminal helix can permeabilize the cell plasma membrane and induce therein channel currents as effectively as the peptaibol alamethicin [19]. In principle the positively charged face of this helix might also be involved in driving partitioning into negatively charged monolayers of target organelles, when 2B is synthesized intracellularly. As a criterion for testing 2B pore formation in membranes we first analyzed the capacity of the MBP-2B construct to permeabilize LUV at low doses (Fig. 2). Kinetic traces displayed in panel A indicate that

Fig. 2. Permeabilization of large unilamellar vesicles by MBP-2B (ANTS/DPX assay). (A) Leakage of internal contents as a function of time. The curves were obtained at 8000:1 lipid-to-protein molar ratio. The arrow indicates the time of protein addition. (B) Extent of MBP-2B-induced aqueous content release (percentage of maximum leakage at time = 60 min) from PI (circles and continuous line), PS (squares and dotted line) and PC (triangles and slashed line) vesicles as a function of the lipid-to-protein molar ratios. In all cases the lipid concentration was 50 μM. Data points correspond to mean values of two independent determinations in a representative experiment.

Fig. 3. Penetration of MBP-2B protein into phospholipid monolayers. (A) Surface pressure increase in time upon injection of the protein (indicated by the arrow) in the subphase (final protein concentration: 0.17 μM) with initial pressure fixed at 25 mN/m. (B) Maximum increase in surface pressure as a function of the initial surface pressure of PI (circles and continuous line), PS (squares and dotted line) and PC (triangles and slashed line) phospholipid monolayers.

upon MBP-2B addition at a lipid-to-protein ratio of 8000:1 (mol:mol) PI vesicles released their aqueous contents into the medium, while PC and PS vesicles remained intact. Induction of leakage at such low doses suggests a lysis mechanism not requiring massive amounts of protein adsorbed to vesicle membranes. Titration experiments displayed in Panel B revealed that PC and PS vesicles were barely affected by MBP-2B at lipid-to-protein ratios as low as 500:1, conditions that induced substantial permeabilization of PI vesicles. Control experiments also indicated that the MBP solubilization partner in isolation was unable to induce leakage from PI vesicles at any of the tested doses (not shown). We analyzed next MBP-2B capacity for inserting into lipid monolayers composed of the same phospholipids (Fig. 3). At a fixed surface pressure of 25 mN/m, MBP-2B inserted into PS and PI, but not into PC monolayers (Panel A). Furthermore, the kinetic traces were consistent with MBP-2B insertion being faster into PS than into PI monolayers (t1/2-s of 40 and 140 seconds, respectively). Estimation of the critical pressures for insertion (πc), i.e., the surface pressures at which the protein was excluded from the monolayer, rendered values above 30 mN/m for the anionic PI and PS lipid monolayers (Panel B). By comparison, the protein was unable to insert into PC monolayers compressed above 22.5 mN/m. Comparable πc values of ≈22 mN/m were obtained for insertion of MBP protein devoid of 2B moiety into PI or PS monolayers (not shown). This behavior supports the capacity of 2B to insert into PI or PS monolayers at lateral pressures existing in biological membranes (π0 ≥ 30–35 mN/m [25]). Thus, incapacity for inserting into unstressed electrically neutral membranes provides an explanation for the absence of 2B-induced leakage from PC LUV (Fig. 2). In contrast, monolayer assays reflected that MBP-2B should be capable of inserting into unstressed PS membranes, even though protein addition did not induce aqueous contents release from PS LUV. Results in Fig. 4 confirmed MBP-2B capacity for partitioning into PI or PS LUV. Partitioning curves in panel A were obtained from the fractional fluorescence intensity increase of NBD upon titration with

2624

A. Agirre et al. / Biochimica et Biophysica Acta 1778 (2008) 2621–2626

Fig. 5. Dependence of final extents of MBP-2B-induced leakage and water–membrane partitioning on PI density. Extents of leakage correspond to the level of ANTS released into the aqueous medium after incubating vesicles made of PI:PC mixtures with MBP-2B for 60 min at 37 °C at the 4000:1 lipid-to-protein mole ratio. Water–membrane partitioning constants (inset) were determined from the fractional change in MBP-2BNBD fluorescence upon titration with the different vesicle preparations. Protein concentration was 0.2 μM. Data points correspond to mean values of two independent determinations in a representative experiment.

Fig. 4. MBP-2B water–bilayer partitioning. (A) Partitioning curves calculated from the fractional change in MBP-2B-NBD fluorescence upon titration with PI (circles and continuous line), PS (squares and dotted line) and PC (triangles and slashed line) vesicles. Protein concentration was 0.2 μM. The solid line corresponds to best fit of the experimental values to the function in equation (2). (B) Inhibition of MBP-2B-induced leakage by pre-incubation with increasing amounts of empty PI (circles and continuous line), PS (squares and dotted line) and PC (triangles and slashed line) vesicles. The leakage activity remaining after 30 min incubation was determined by supplementing the mixtures with ANTS/DPX-containing PI vesicles (lipid concentration, 25 μM). The maximal leakage was measured in the absence of empty vesicles at the lipid-to-protein mole ratio of 4000:1.

lipid. NBD penetrated into the non-polar environment of the bilayer in the case of PS and PI vesicles, but not in the case of PC vesicles. Thus, water–bilayer partitioning data were consistent with the previous lipid monolayer insertion experiments. Data displayed in panel B reflect the degree of leakage inhibition observed after pre-incubating MBP-2B protein with increasing amounts of empty lipid vesicles. Concordant with the previous water–bilayer partitioning data, preincubation with PS or PI vesicles reduced MBP-2B leakage activity, an effect not observed when the protein was pre-incubated with PC vesicles. Together these data indicate that, as opposed to a putative reversible adsorption phenomenon, association with PS and PI vesicles involved the insertion of the NBD-labeled C-terminal hairpin helix into the hydrophobic milieu, a process that was in both cases essentially irreversible. In summary, the results presented so far indicate that MBP-2B inserts and partitions into anionic phospholipid bilayers, while it is unable to associate with electrically neutral PC vesicles (Figs. 2–4). This phenomenon suggests that the membrane surface potential might be important for sustaining 2B translocation into bilayers. Thus, we next evaluated partitioning and induced leakage in PI:PC mixtures (Fig. 5). Leakage in this case was assayed at the 4000:1 lipid-to-protein ratio, which results in ca. 60 % ANTS/DPX release from pure PI vesicles upon incubation with MBP-2B for 60 min. Results displayed in Fig. 5 show that increasing PI density improved more or less in parallel water–LUV partitioning (inset) and MBP-2B-induced leakage. PI:PC ratios above 7:3 were required to attain partitioning constants in the order of 106 or higher, under conditions sustaining effective vesicle permeabilization. Thus, from the water–membrane partitioning data (inset) it can be estimated that bilayer surface potentials higher than approximately −90 mV were required for stable protein–bilayer association.

The previous results also indicate that, although required for its translocation into membranes, a strong membrane surface potential might not be sufficient for sustaining 2B-induced membrane permeabilization. Specifically we found that MBP-2B was unable to induce permeabilization of PS vesicles. The data displayed in Fig. 6 further indicate that increasing PS density actually inhibited the process. The initial rates (not shown) and extents of leakage were reduced with increasing PS densities. However, the estimated partitioning constants reflected comparable degrees of protein association for all assayed PS: PI vesicles (inset). In line with the previous lipid monolayer insertion and water–membrane partitioning results, partitioning constants in the order of 106 were measured for pure PS vesicles under conditions not inducing vesicle permeabilization. In order to gain insights into the mechanism of this inhibitory process, PS effect on MBP-2B-induced leakage was analyzed quantitatively according to a lytic pore model [23,24]. To that end we determined the extents of leakage as a function of the protein dose in different PI:PS mixtures (Fig. 7A). In accordance with our previous observations, higher protein-to-lipid ratios (Ri-s) were required to permeabilize the vesicles that contained PS. Mathematical analysis of the experimental extents of leakage according to the pore model rendered optimal fits (R2 ≥ 0.98) for M = 4 and M = 6 in the case of pure

Fig. 6. Dependence of final extents of MBP-2B-induced leakage and water–membrane partitioning on PS density. Extents of leakage correspond to the level of ANTS released into the aqueous medium after incubating vesicles made of PI:PS mixtures with MBP-2B for 60 min at 37 °C at the 4000:1 lipid-to-protein mole ratio. Water–membrane partitioning constants (inset) were determined as described in the caption for the previous figure.

A. Agirre et al. / Biochimica et Biophysica Acta 1778 (2008) 2621–2626

Fig. 7. PS inhibitory effect on MBP-2B-induced LUV permeabilization. (A) Final extents of leakage (percentage after 60 min) in PI (empty circles), PI:PS (2:1) (filled circles), PI:PS (1:2) (empty squares) and PS (filled squares) as a function of the protein-to-lipid mole ratio (Ri). Lipid concentration was constant (50 μM). The curves correspond to the predicted values according to a pore model, in which the minimal number of monomers required for pore formation was M = 4. (B) Dependence of experimental and calculated leakage extents (filled and empty circles, respectively) and the surface aggregation constants (squares) on PS density. The calculated values were determined for M values of 6 (right) or 4 (left).

PI and PI:PS (2:1) vesicles, while extents of leakage from PI:PS (1:2) could not be fitted so accurately (R2 = 0.87). Extents of leakage from pure PS vesicles could not be explained by a pore model. Inhibition of maximal leakage induced by increasing the PS content of the vesicles also correlated with a reduction of the aggregation surface constants (Ks) for both M = 4 and M = 6 (Fig. 7B). As elaborated in [26,27] the parameter Ks gives a quantitative measure to the tendency of peptides or proteins to self aggregate within the membrane. For irreversible peptide aggregation Ks is infinite, although in practice Ks = 15 gave the same outcome as using an infinite value in GALA-induced leakage [26], where a correlation between Ks values and reversibility of peptide aggregation in the membrane was directly demonstrated by a fluorescent assay. The Ks values deduced here are intermediate between those in [26] and [27]. In conclusion, results in Figs. 6 and 7 suggest that the lower pore-forming efficiency in PS-containing membranes could be due to a lower tendency of the protein to selfassociate at the bilayer surface, rather than to an insertion defect. 4. Discussion Viroporins comprise a family of small, non-glycosylated, highly hydrophobic viral polypeptides produced intracellularly during infection [2,28]. An additional common feature of viroporins is their capacity to alter the permeability barrier of several cell membrane systems [28], a phenomenon that has led to the proposal that these proteins integrate a group of intracellularly produced pore-forming toxins of viral origin [19]. The molecular mechanisms underlying viroporin cytotoxic effects remain in great part unknown. In particular, the amount of experimental data obtained in chemically defined systems including purified protein and model membranes is to date very limited.

2625

Non structural 2B and 2BC products encoded by members of the enterovirus and rhinovirus genera belonging to the Picornaviridae family, constitute paradigms of viroporins [3]. Specifically, the membrane permeabilizing effects induced by poliovirus and coxsackievirus 2B/2BC have been extensively characterized in cultured cells by the groups of Carrasco and van Kuppeveld, respectively [2,3,6– 11,13–16]. The pioneering approach of protein solubilization by generating a fusion construct with MBP, provided some clues on the pore-forming activity of poliovirus 2B in model membranes [17]. In good agreement with cell permeabilization data, MBP-2B fusion protein was found to assemble into discrete structures allowing the passage of low molecular-weight solutes across vesicle bilayers. Using the same approach, in this work we further demonstrate that PI and PS, the main species of monovalent anionic phospholipids existing in target cell endomembranes [6,10,12,15,16,20], sustain MBP-2B insertion (Fig. 3) and partitioning (Fig. 4). Those processes are causally linked with pore-opening in the case of PI (Fig. 5) but not in the case of the PS (Figs. 6 and 7). Our results confirmed that the initial contact with the bilayer depended on a minimal surface electrostatic potential. Thus, inclusion of electrically neutral PC reduced significantly the amount of protein that associated with the vesicles (Fig. 5). This observation is in accordance with the insertion of 2B into the endoplasmic reticulum and Golgi membranes and the cytofacial plasma membrane monolayer [10,11,15,16], all of them enriched in anionic phospholipids [20,29]. The negative surface electrostatic potential originated by these phospholipids in natural membranes is probably not uniformly distributed [30]. It has been described that regions enriched in anionic phospholipids might form in mixtures with zwitterionic species [31,32], a phenomenon that might enable the insertion of proteins that interact electrostatically with cytofacial membrane surfaces [33]. Thus, segregated membrane regions bearing high charge density are conceivable selected for 2B translocation into target organelles. Nonetheless, we foresee that the specific environment within these regions may condition 2B pore-opening. MBP-2B was capable of permeabilizing PI vesicles at low protein-to-lipid ratios, under conditions that did not perturb the permeability barrier of PS LUV (Fig. 2). The absence of PS bilayer lysis did not correlate with lower levels of insertion (Fig. 3) or partitioning (Fig. 4) indicating that PSinserted versions were specifically defective in opening transbilayer aqueous connections. Furthermore, the presence of this lipid appeared to block pore-opening (Figs. 6 and 7). The faster kinetics of insertion into membrane monolayers (Fig. 3A) already suggested the existence of 2B-PS interactions that differed from those established with the other monovalent anionic phospholipids. As compared to monovalent phospholipids such as PI, specific chemical features of the trivalent PS polar head-group include the absence of hydroxyl groups, and the fact that the net -1 charge is the result of 3 ionized groups, 2 negatives (from the phosphate and the carboxyl groups, respectively), and 1 positive (from the ammonium group). Moreover, PS carboxyl group is highly accessible from the water phase, a phenomenon that contributes to facilitate the binding of soluble cationic molecules to the bilayer surface [34]. Hence, charge distribution within the PS interface might facilitate the interaction of 2B cationic helix with the bilayer surface. The reduction in the surface aggregation constant induced by PS (Fig. 7) suggests that this favorable interaction might reduce surface self-oligomerization and further translocation into the hydrocarbon core to form a pore. A similar possibility was proposed by Matsuzaki et al. [35] to explain inhibition of pore formation by antimicrobial magainin 2 peptide in PS vesicles. In comparison to PS vesicles, the peptide permeabilized phosphatidylglycerol (PG) vesicles at much lower membrane doses. However, it associated with both type of vesicles with comparable affinity. Incapacity to permeabilize PS vesicles was also described for cytolytic mellitin by Ohki et al. [36]. These authors

2626

A. Agirre et al. / Biochimica et Biophysica Acta 1778 (2008) 2621–2626

[7] [8]

[9]

[10] [11] [12]

[13] [14] Fig. 8. Modulation of MBP-2B pore formation by PI and PS. Initial association of the protein with the bilayer is not detected in liposomes composed of electrically neutral PC (a) because it requires net negative surface charge (b). Association with PI bilayers results in oligomeric pore formation (c). In contrast, association with PS bilayers favors inserted monomers that do not translocate across the hydrocarbon core.

[15]

[16]

suggested that the peptide remained adsorbed to the PS bilayer interface, unable to translocate across the bilayer. The model depicted in Fig. 8 summarizes the experimental observations in this work. MBP-2B protein interacted electrostatically with PI and PS bilayers, but only in the case of the former this interaction resulted in pore-opening. The characterization in our previous work and the kinetic analysis reported here support that poliovirus 2B may assemble into tetrameric aqueous pores of diameter in the range 5–10 Å in anionic PI bilayers [17,18]. Thus, the experimental data fit better with the assembly by 2B of a discrete permeating pore, either barrel-stave or toroidal, than with a carpet-like mechanism of general bilayer destabilization [37–39]. The model in Fig. 8 further emphasizes that MBP-2B monomers inserted at the interface of PS bilayers were probably defective in self-oligomerization and, therefore, unable to establish aqueous connections across the bilayer. This phenomenon might be the consequence of a defective translocation of the hydrophobic hairpin into the hydrocarbon core. In summary, this model suggests that 2B membrane association-insertion depends on anionic phospholipids, while pore-opening is subsequently regulated by the chemical nature of the polar-head-group moiety. We propose that such mechanism might be functional during poliovirus infection for: i) driving 2B–2BC translocation into target organelle membrane, and ii) regulating the membrane permeabilizing activity therein. Acknowledgements

[17]

[18] [19]

[20] [21]

[22] [23] [24] [25] [26] [27] [28] [29] [30]

This study was supported by Spanish Ministerio de Educación y Ciencia (BFU 2005-0695) and University of the Basque Country (042.310-13552/2001).

[32]

References

[33]

[1] L. Carrasco, Modification of membrane permeability induced by animal viruses early in infection, Virology 113 (1981) 623–629. [2] L. Carrasco, Modification of membrane permeability by animal viruses, Adv. Virus Res. 45 (1995) 61–112. [3] A.S. de Jong, F. de Mattia, M.M. Van Dommelen, K. Lanke, W.J. Melchers, P.H. Willems, F.J. van Kuppeveld, Functional analysis of picornavirus 2B proteins: effects on calcium homeostasis and intracellular protein trafficking, J. Virol. 82 (2008) 3782–3790. [4] L. Carrasco, A.E. Smith, Sodium ions and the shut-off of host cell protein synthesis by picornaviruses, Nature 264 (1976) 807–809. [5] L. Carrasco, Membrane leakiness after viral infection and a new approach to the development of antiviral agents, Nature 272 (1978) 694–699. [6] F.J. van Kuppeveld, J.G. Hoenderop, R.L. Smeets, P.H. Willems, H.B. Dijkman, J.M. Galama, W.J. Melchers, Coxsackievirus protein 2B modifies endoplasmic reticulum

[31]

[34] [35]

[36] [37]

[38] [39]

membrane and plasma membrane permeability and facilitates virus release, Embo J. 16 (1997) 3519–3532. F.J. van Kuppeveld, W.J. Melchers, K. Kirkegaard, J.R. Doedens, Structure–function analysis of coxsackie B3 virus protein 2B, Virology 227 (1997) 111–118. L. Carrasco, L. Perez, A. Irurzun, F. Martinez-Abarca, P. Rodriguez, R. Guinea, J.L. Castrillo, M.A. Sanz, M.J. Ayala, in: L. Carrasco, N. Sonenberg, E. Wimmer (Eds.), Regulation of Gene Expression in Animal Viruses, Plenum Press, London, 1993, pp. 283–305. A. Barco, L. Carrasco, A human virus protein, poliovirus protein 2BC, induces membrane proliferation and blocks the exocytic pathway in the yeast Saccharomyces cerevisiae, Embo J. 14 (1995) 3349–3364. R. Aldabe, A. Barco, L. Carrasco, Membrane permeabilization by poliovirus proteins 2B and 2BC, J. Biol. Chem. 271 (1996) 23134–23137. R. Aldabe, A. Irurzun, L. Carrasco, Poliovirus protein 2BC increases cytosolic free calcium concentrations, J. Virol. 71 (1997) 6214–6217. I.V. Sandoval, L. Carrasco, Poliovirus infection and expression of the poliovirus protein 2B provoke the disassembly of the Golgi complex, the organelle target for the antipoliovirus drug Ro-090179, J. Virol. 71 (1997) 4679–4693. A. Barco, L. Carrasco, Identification of regions of poliovirus 2BC protein that are involved in cytotoxicity, J. Virol. 72 (1998) 3560–3570. F.J. van Kuppeveld, J.M. Galama, J. Zoll, P.J. van den Hurk, W.J. Melchers, Coxsackie B3 virus protein 2B contains cationic amphipathic helix that is required for viral RNA replication, J. Virol. 70 (1996) 3876–3886. A.S. de Jong, H.J. Visch, F. de Mattia, M.M. van Dommelen, H.G. Swarts, T. Luyten, G. Callewaert, W.J. Melchers, P.H. Willems, F.J. van Kuppeveld, The coxsackievirus 2B protein increases efflux of ions from the endoplasmic reticulum and Golgi, thereby inhibiting protein trafficking through the Golgi, J. Biol. Chem. 281 (2006) 14144–14150. A.S. de Jong, W.J. Melchers, D.H. Glaudemans, P.H. Willems, F.J. van Kuppeveld, Mutational analysis of different regions in the coxsackievirus 2B protein: requirements for homo-multimerization, membrane permeabilization, subcellular localization, and virus replication, J. Biol. Chem. 279 (2004) 19924–19935. A. Agirre, A. Barco, L. Carrasco, J.L. Nieva, Viroporin-mediated membrane permeabilization. Pore formation by nonstructural poliovirus 2B protein, J. Biol. Chem. 277 (2002) 40434–40441. J.L. Nieva, A. Agirre, S. Nir, L. Carrasco, Mechanisms of membrane permeabilization by picornavirus 2B viroporin, FEBS. Lett. 552 (2003) 68–73. V. Madan, S. Sanchez-Martinez, N. Vedovato, G. Rispoli, L. Carrasco, J.L. Nieva, Plasma membrane-porating domain in poliovirus 2B protein. A short peptide mimics viroporin activity, J. Mol. Biol. 374 (2007) 951–964. G. van Meer, D.R. Voelker, G.W. Feigenson, Membrane lipids: where they are and how they behave, Nat. Rev. Mol. Cell. Biol. 9 (2008) 112–124. M.J. Hope, M.B. Bally, G. Webb, P.R. Cullis, Production of large unilamellar vesicles by a rapid extrusion procedure. Characterization of size distribution, trapped volume and ability to maintain a membrane potential, Biochim. Biophys. Acta 812 (1985) 55–65. H. Ellens, J. Bentz, F.C. Szoka, H+- and Ca2+-induced fusion and destabilization of liposomes, Biochemistry 24 (1985) 3099–3106. S. Nir, J.L. Nieva, Interactions of peptides with liposomes: pore formation and fusion, Prog. Lipid. Res. 39 (2000) 181–206. R.A. Parente, S. Nir, F.C. Szoka Jr., Mechanism of leakage of phospholipid vesicle contents induced by the peptide GALA, Biochemistry 29 (1990) 8720–8728. D. Marsh, Lateral pressure in membranes, Biochim. Biophys. Acta. 1286 (1996) 183–223. F. Nicol, S. Nir, F.C. Szoka Jr., Effect of cholesterol and charge on pore formation in bilayer vesicles by a pH-sensitive peptide, Biophys. J. 71 (1996) 3288–3301. D. Rapaport, R. Peled, S. Nir, Y. Shai, Reversible surface aggregation in pore formation by pardaxin, Biophys. J. 70 (1996) 2502–2512. M.E. Gonzalez, L. Carrasco, Viroporins, FEBS Lett. 552 (2003) 28–34. A. Zachowski, Phospholipids in animal eukaryotic membranes: transverse asymmetry and movement, Biochem. J. 294 (Pt 1) (1993) 1–14. D. Murray, A. Arbuzova, G. Hangyas-Mihalyne, A. Gambhir, N. Ben-Tal, B. Honig, S. McLaughlin, Electrostatic properties of membranes containing acidic lipids and adsorbed basic peptides: theory and experiment, Biophys. J. 77 (1999) 3176–3188. T. Ahn, F.P. Guengerich, C.H. Yun, Membrane insertion of cytochrome P450 1A2 promoted by anionic phospholipids, Biochemistry 37 (1998) 12860–12866. T. Ahn, D.B. Oh, B.C. Lee, C.H. Yun, Importance of phosphatidylethanolamine for the interaction of apocytochrome c with model membranes containing phosphatidylserine, Biochemistry 39 (2000) 10147–10153. A. Mulgrew-Nesbitt, K. Diraviyam, J. Wang, S. Singh, P. Murray, Z. Li, L. Rogers, N. Mirkovic, D. Murray, The role of electrostatics in protein–membrane interactions, Biochim. Biophys. Acta 1761 (2006) 812–826. G. Cevc, Membrane electrostatics, Biochim. Biophys. Acta 1031 (1990) 311–382. K. Matsuzaki, K. Sugishita, N. Ishibe, M. Ueha, S. Nakata, K. Miyajima, R.M. Epand, Relationship of membrane curvature to the formation of pores by magainin 2, Biochemistry 37 (1998) 11856–11863. S. Ohki, E. Marcus, D.K. Sukumaran, K. Arnold, Interaction of melittin with lipid membranes, Biochim. Biophys. Acta 1194 (1994) 223–232. Y. Shai, Mechanism of the binding, insertion and destabilization of phospholipid bilayer membranes by alpha-helical antimicrobial and cell non-selective membrane-lytic peptides, Biochim. Biophys. Acta 1462 (1999) 55–70. D.M. Ojcius, J.D. Young, Cytolytic pore-forming proteins and peptides: is there a common structural motif? Trends Biochem. Sci. 16 (1991) 225–229. H.W. Huang, Molecular mechanisms of peptide-induced pores in membranes, Phys. Rev. Lett. 92 (2004) 198304.