Mimicking Influenza Virus Fusion Using Supported Lipid Bilayers

Sep 3, 2014 - In this paper, we used a minimalist approach based on a lipid bilayer supported on mica, which we imaged by atomic force microscopy (AFM...
0 downloads 0 Views 10MB Size
Download PDF
Article pubs.acs.org/Langmuir

Mimicking Influenza Virus Fusion Using Supported Lipid Bilayers Cédric Godefroy,†,‡ Selma Dahmane,†,‡ Patrice Dosset,†,‡ Olivier Adam,§ Marie-Claire Nicolai,§ Frédéric Ronzon,§ and Pierre-Emmanuel Milhiet*,†,‡ †

Institut National de la Santé et de la Recherche Médicale, Unité 1054, 34090 Montpellier, France Centre National de la Recherche Scientifique, Unité Mixte Recherche 5048, Centre de Biochimie Structurale, Université de Montpellier, 34090 Montpellier, France § Sanofi Pasteur, Bâtiment X3, 1541 avenue Marcel Mérieux, 69280 Marcy l’Etoile, France ‡

S Supporting Information *

ABSTRACT: Influenza virus infection is a serious public health problem in the world, and understanding the molecular mechanisms involved in viral replication is crucial. In this paper, we used a minimalist approach based on a lipid bilayer supported on mica, which we imaged by atomic force microscopy (AFM) in a physiological buffer, to analyze the different steps of influenza fusion, from the interaction of intact viruses with the supported bilayer to their complete fusion. Our results show that sialic acid recognition and priming upon acidification are sufficient for a complete fusion with the host cell membrane. After fusion, a flat and continuous membrane was observed. Because of the fragility of the viral membrane that was removed by the tip, most probably due to the disorganization of the matrix layer at acidic pH, fine structural details of ribonucleoproteins (RNP) were obtained. In addition, AFM topography of intact virus in interaction with the supported lipid bilayer confirms that hemeagglutinin and neuraminidase can form isolated clusters within the viral membrane.

■

INTRODUCTION Influenza A viruses are among the causative agents of flu infection that are responsible for the worldwide death of at least 200 000 people per year. The virus infects the respiratory tract, and the viral entry is initiated by its binding to the plasma membrane of target cells, a process that is followed by its endocytosis. Influenza A viruses belong to the Orthomyxoviridae family and form particles of 80−120 nm. Its genome is a negative-stranded RNA bound to a single RNA-dependent RNA polymerase and decorated with viral nucleoproteins. These nucleoproteins complexes named RNPs form ribbons surrounded by a matrix coat composed of the M1 protein. This structure is contained within a lipid bilayer envelope, formed during the release of viral particles. This membrane originates from the plasma membrane of the host cell but also contains viral proteins like hemeagglutinin (HA), neuraminidase (NA), and the M2 ion protein channel. HA and NA are two large glycoproteins that protrude into the extraviral space. HA is a lectin composed of two disulfidelinked subunits, HA1 and HA2 (see the review1). HA1 mediates virus targeting to host cells by binding to sialic acid moieties present on the cell surface. This protein also contains the antigenic determinants.2 The HA2 subunit, also referred to as fusion peptide, is primarily involved in the membrane fusion process. Several hundred copies of HA are expressed in the viral membrane and, according to electron microscopy (EM) experiments,2 form “spikes” at the viral surface that protrude approximately 14 nm into the extraviral space. EM images further revealed that HA are not homogeneously distributed © 2014 American Chemical Society

within the viral forming clusters, which were also visualized by atomic force microscopy.3 The NA expression (15% of total membrane proteins) is much more lower than that of HA. This protein also forms clusters and has a long-stemmed “mushroom” shape easily recognizable in EM as objects that protrude approximately 16 nm into the extraviral space.2 NA is involved in the release of progeny virus from infected cells by cleaving sugars that bind the mature viral particles (reviewed in ref 4). HA binding to the host membrane via sialic acid moieties induces virus endocytosis in a clathrin-dependent manner, and the acidic pH in endosomal compartment makes the HA proteins fusogenic. At pH 55,6 viral and host membranes fuse, allowing the viral genome to enter the cell cytoplasm. X-ray crystallography has provided structural details of acid-induced conformational changes of HA which showed that the release of the amphipathic, membrane-active fusion peptide from its buried position inside the HA spike at neutral pH and the refolding of HA into an extended coil−coil conformation are prerequisites for fusion.7−9 On a larger scale, EM and AFM experiments have shown that the distribution of HA proteins at the viral surface was altered at acid pH.3,10,11 More recently, cryo-electron tomography (Cryo-ET) has been used to image the shape of influenza virus under acidic conditions.10 At pH 4.9, it was shown that HA spikes become disorganized and that the layer of M1 matrix proteins is no longer resolved on most Received: July 2, 2014 Revised: August 19, 2014 Published: September 3, 2014 11394

dx.doi.org/10.1021/la502591a | Langmuir 2014, 30, 11394−11400

Langmuir

Article

of the cantilever tip shape. However, we cannot exclude that diameter values are a little bit overestimated. The measurement of height above the bilayer is more quantitative (for more information, see Figure S2).

virions. In addition, the ribonucleoprotein complexes (RNPs) coagulate on the interior surface of the virion. More recently, another article of Steven’s group confirmed the disorganization of the M1 protein at acidic pH that led to its dissociation from the membrane which was hypothesized to be a priming mechanism for fusion.11 Quick-freezing EM has also been used to study the initial steps of the virus fusion.12 In addition to modification of virus morphology upon acidification, especially the regular polygonal arrangements of hemeagglutinin spikes on the viral surface, this report proposed that microprotrusion of the host membrane is a key event in the onset of fusion. In order to better understand the molecular mechanisms associated with the fusion process of H3N2 influenza virus with biological membranes, we have used here supported lipid bilayers (SLB) combined with atomic force microscopy (AFM) imaging. This technique is able to provide 3D structural information on biological samples in buffer with a lateral resolution in the nanometer range and has been successfully applied to viral structure analysis (see the review13). In addition, AFM is also an ideally suited technique for observing planar artificial membranes and lateral segregation of their components.14 In this paper, we have used a fluid SLB doped with the sialic acid-containing ganglioside GM3 as a mimic of the cell membrane recognized by the influenza virus. These artificial membranes have been incubated with H3N2 virus and their fusion observed upon acidification. Taken together, our results provided structural details at nanometer resolution of influenza viruses in interaction with a biological membrane mimic. Our study also demonstrated that SLB constitute a very interesting minimalist system allowing the study of virus fusion.

■

■

RESULTS As mentioned above, the influenza virus specifically recognizes the cell surface of the host cells through its binding to sialic acid moieties, to sugars present on proteins such as glycophorin,18 and to lipids such as gangliosides.19 In order to mimic the plasma membrane, we fabricated supported bilayers made of dioleoylphosphatidylcholine (DOPC) containing the ganglioside GM3. DOPC forms a fluid phase at room temperature and is a standard means to mimic eukaryotic membranes,14,20 whereas the polar heads of GM3 contain sialic acid residues that forms viral binding sites. Similar artificial membranes have successfully been used to investigate influenza virus binding.21 AFM experiments in tapping mode (intermittent contact) showed that this bilayer obtained by fusion of large unilamellar vesicles on mica was continuous and uniform over large areas. Some rare defects in the membrane made it possible to determine the thickness of the bilayer as 4.7 ± 0.4 nm (n = 10). In addition, no lipid phase separation or domains were observed (Figure 1A). The absence of ganglioside domains in artificial bilayers22,23 can be explained by the low GM3 concentration used and the fluidity of the DOPC at room temperature (Tm of −20 °C) that makes the membrane highly dynamic, which in turn prevents domain formation. The DOPC/GM3 bilayer was then incubated with intact H3N2 viruses at pH 7.4 in order to verify the ability of the viruses to bind gangliosides within the artificial membrane. As expected, round-shaped viruses interact with the SLB (inset in Figure 1A), and as observed in our previous report,3 the virus diameter was variable with an average value of 106 ± 23 nm (n = 102). The height measured between the buffer-exposed lipid polar heads and the virus top was of 77 ± 14 nm (n = 188), which demonstrates that the virus morphology is not dramatically modified after its binding to the SLB and that the force applied by the tip did not induce a significant deformation of viral particles (see the Material and Methods section). Zoom into virus-enriched areas allowed us to observe that the viral particles are corrugated (Figure 1B); these corrugations most probably correspond to the HA and NA proteins, the major components of the viral membrane. Interestingly, even if the tip could not delineate individual molecules, protein domains or assemblies that varied in shape were clearly observed at the membrane (see black arrowheads in Figure 1B). Lateral segregation of HA and NA was more pronounced for some particles allowing the observation of smoother areas (see the darker region in particles in Figure 1B and in the zoom in Figure 1D) that probably correspond to the lipid matrix devoid of proteins. This assumption is strongly supported by the phase image recorded in tapping mode (Figure 1C), which is related to the energy dissipated by the tip when it transiently interacts with the sample surface and characterizes its viscoelastic properties. The phase signal was indeed very different between the corrugations and the smooth areas. The latter was much higher, reminiscent of a region with high viscoelastic properties that is expected for a pure lipid bilayer as compared to a membrane containing proteins. This interpretation is also supported by the height difference (12 ± 1.6 nm) measured between smooth areas and corrugations. This value is in good agreement with the extensions of the HA and NA extracellular domains that were determined by EM to

MATERIALS AND METHODS

Virus Preparation. An egg-grown influenza virus preparation, influenza virus A/Victoria (H3N2), was obtained from allantoic fluid centrifuged on a sucrose gradient. The viral protein concentration, determined using a Micro BCA Protein Assay Reagent kit (Pierce), was 2 mg/mL in PBS at pH 7.5 after overnight dialysis. Supported lipid bilayers (SLB) were prepared as previously reported.15 Briefly, dioleoylphosphatidylcholine (DOPC)/GM3 (0.95:0.05 mol/mol) liposomes, prepared by extrusion at 70 °C of multivesicular vesicles at 0.125 mM in phosphate buffer, pH 7.4, were deposited onto freshly cleaved mica, allowed to adsorb and fuse during 2 h incubation at 70 °C, and rinsed with the same buffer. Atomic Force Microscopy Imaging (AFM). Lipid bilayer supported on mica was incubated with viruses suspended in PBS pH 7.4 for 30 min at room temperature in a humid chamber. Once the viruses bound to the SLB, the sample was rinsed and further incubated under the same conditions in the same buffer or in phosphate/citric acid buffer pH 5. The sample was then rinsed by buffer exchange and fixed using 5% glutaraldehyde in PBS at pH 7.4. Glutaraldehyde was used to inactivate influenza viruses and also to preserve the morphology of biological structures. It is also valuable in AFM to perform high-resolution imaging of viruses16 or of fibers.17 AFM imaging was performed in 20 mM Tris, 150 mM KCl, pH 7.4, in tapping (intermittent contact) mode using a Multimode AFM (NanoBruker, Palaiseau, France) equipped with a J or E scanner and oxidesharpened Si3N4 sharp cantilevers (k = 0.09 N/m, Olympus, Japan). Free amplitude oscillation was generally a few tens of nanometers and the set point adjusted to less than 10% amplitude damping. The scanning rate was between 0.3 and 1 Hz depending on the scan size. Importantly, tight packing of viruses as well as freshly thawed samples favored high-resolution imaging. Image Analysis. Height images were flattened using the software provided by Nano-Bruker. The diameter of round-shaped particles was determined by full width at half-maximum (fwhm) measurement in order to minimize the distortion of lateral distances due to convolution 11395

dx.doi.org/10.1021/la502591a | Langmuir 2014, 30, 11394−11400

Langmuir

Article

Figure 1. continued the SLB. The black frame in (B) indicates the zoom shown in (D) (height image). This zoom-in shows two types of morphology: one where proteins are homogeneously distributed at the virus surface (virus on the left) and the other where protein-enriched areas alternated with areas devoid of proteins (smoother area, on the right). The inset is a zoom-in as indicated by the white frame. Scale bars are 1 μm (A), 200 nm (inset in A), 50 nm (B, C), 25 nm (D), and 10 nm (inset in A); the z color scale is 25 nm (A, the color code is shown on the right of the picture) and 200 nm (inset in A, B, and D).

be 14 and 16 nm, respectively.2 Two different morphologies can be distinguished in the higher magnification shown in Figure 1D, namely virus with proteins homogeneously distributed at the viral surface (on the left) or virus displaying clusters or protein-enriched areas alternated with protein-free areas (on the right). In the latter case, an isolated 30 nm diameter cluster of proteins is clearly observed as well as a side view of another cluster (white arrowhead in Figure 1D). This inset even permits the observation of the bulky head of proteins that protrude 15.1 nm above the lipid matrix. Importantly, no viruses were observed in the control experiment with the GM3free bilayer. Once bound to the cell plasma membrane, the influenza virus is normally engulfed to form an endosome, and the acidic pH of this intracellular compartment is a key event inducing virus fusion with the host membrane. In order to visualize virus fusion, we then decreased the pH by exchanging the buffer. Under these conditions, we observed in certain areas a decrease of the apparent diameter of membrane bound viruses as highlighted by the black arrowhead in Figure 2A, which strongly suggests that the virus had started to fuse (the average value of the viral height with respect to the lipid bilayer was 55 ± 18 nm (n = 54), which corresponds to a reduction by 29% as compared to the viral height at neutral pH). The SLB surrounding viruses appeared homogeneous (no apparent difference in height), except some elongated structures that protruded 7−9 nm above the lipid bilayer (white arrowheads in Figure 2A). Interestingly, the phase signal associated with the lipid membrane was not uniform (Figure 2B). Indeed, the phase signal of the membrane surrounding the virus was higher than the signal recorded at a longer distance (the contour of the membrane area with a higher phase contrast is delineated by the thin black line in Figure 2A). This difference indicates that the viscoelastic properties of the membrane surrounding the viral particle are different from those of the prefusion membrane, also suggesting a partial fusion of the virus. Therefore, protein-enriched membranes of the virus should have different mechanical properties as compared to a pure lipid membrane. Intriguingly, a similar difference in the phase signal was also observed for areas devoid of protrusions (upper right white arrowhead in Figure 2A) that surround the elongated structures mentioned above. Such a pattern is clearly visible in Figure 3A where elongated structures are delineated by the AFM tip (Figure 3A), and even if the membrane appeared to have an uniform topography, the phase signal from areas surrounding these structures was higher than that for the rest of the membrane. Such a signal is consistent with a complete fusion of influenza viruses with the SLB, and the higher magnification images shown in Figure 3C−E strongly suggest that the elongated structures correspond to viral RNPs. The structures were indeed ribbon-like particles with lengths

Figure 1. Interaction of influenza virus with lipid membrane imaged at pH 7.4. (A) Topography of a supported lipid bilayer of DOPC containing 5% (w/v) GM3. The bilayer is mainly continuous and homogeneous except some defects on the right of the image. Dark areas correspond to mica and bright dots to nonfused vesicles. The inset corresponds round-shaped viruses that interact with the SLB after their incubation at pH 7.4 with the membrane. (B, C) Zoom-in of a virus-enriched area. The height image in (B) reveals isolated roundshaped particles with a corrugated surface composed of HA and NA proteins (not differentiated by the AFM tip). Black arrowheads indicate lateral segregation of membrane proteins at the viral surface. Smooth areas are observed at times (black asterisk) and probably correspond to pure lipid membranes devoid of proteins as suggested by the phase signal recorded in (C). The white asterisk in (B) pointed 11396

dx.doi.org/10.1021/la502591a | Langmuir 2014, 30, 11394−11400

Langmuir

Article

Figure 2. Shape modification of influenza virus in interaction with membrane after incubation at pH 5. Height and phase AFM images of virus in interaction with supported lipid membranes after 60 min incubation at pH 5. The height of viruses (black arrowhead in part A) is lower as compared to the inset in Figure 1A, and their shape is less delineated by the tip suggesting a partial disruption of their integrity when interacting with the bilayer at acidic pH. Interestingly, the phase signal in (B) indicates that viruses are surrounded by an area with viscoelastic properties different to those of the lipid bilayer (highlighted by the black line in part A) although no clear height difference was observed in (A). Ribbon-like structures are observed in the height image in areas with higher phase signal (white arrowhead in part A). Small virus debris are also delineated by the AFM tip on top of the DOPC/GM3 bilayer. Scale bars are 200 nm, and the z color scale is 260 nm in (A). The phase color scale is 20° in (B).

ranging from 25 to 80 nm, which is in good agreement with previous EM observations.24 Ribbons were either isolated, in close proximity to each other, or attached to each other by to a very large structure not properly delineated by the tip (white arrowhead in Figure 3C). This structure is probably a piece of the matrix layer that contacts RNPs in intact virus as described in EM.2 Objects with helical shape were also observed in highmagnification AFM images (Figure 3D,E, see the white arrowhead) and a right-handed helix pitch of 10.5 ± 2.5 nm (n = 22) was calculated. The observation of RNPs, either associated with matrix or not, is of interest to understand the fusion mechanism. Indeed, a classical fusion process should result in viral membrane components (HA and NA) facing the tip and the RNPs localized in the space between the viral membrane and the mica substrate. Because the force applied by the tip was very low (in the 100 pN range), it is unlikely that the tip indented that much into the sample as to delineate the topography of the RNPs below the membrane. Such an explanation is also inconsistent with the high-resolution images obtained in Figure 3D,E. A closer inspection of these high-resolution images in fact

Figure 3. Complete fusion of influenza virus with supported lipid bilayer. Longer incubation of influenza virus with the DOPC/GM3 SLB led to flat surface where only ribbon-like structures are observed (see black arrowheads in the height image in part A). This suggests a complete fusion of virus with the lipid membrane in good agreement with the phase signal shown in (B). Ribbon-like structures correspond to ribonucleoproteins (RNPs) that appear to be sometimes attached along their main axis or aggregated (white arrowheads in part C). Helical structures are sometimes observed (white arrowheads in part E). Scale bars are 200 nm (A, B), 100 nm (C), and 50 nm (D, E). The height images (D, E) correspond to 300 nm zooms. The z color scale is 200 nm (A) or 30 nm (C−E). 11397

dx.doi.org/10.1021/la502591a | Langmuir 2014, 30, 11394−11400

Langmuir

Article

decreased to 5, clearly indicating that sialic acid priming in acidic conditions is sufficient for viral fusion to a planar lipid bilayer. The viral fusion was possible using the SLB probably due to the existence of a layer of buffer of 2−3 nm thickness between the mica and lipid polar heads, already observed under these experimental conditions.15 Our study also questions somehow different previously published assumptions such as the requirement of high membrane curvature or the presence of microprotrusion of the lipid bilayer as requirements to initiate fusion.12 We are obviously aware that the lipid composition of the model membrane does not correspond to that of biological membranes but that the properties of DOPC suggest that it is a suitable mimic of the fluidity encountered within biological membranes. However, since AFM could not provide information on the contact zone between the virus and the bilayer, we cannot exclude the existence of SLB deformation that could prime the fusion process. Fusion of the viral membrane with the SLB was assessed by imaging the topography of the viruses in tapping mode AFM and by using the phase signal recorded in this mode, a signal that is related to the viscoelastic properties of the sample. Very difficult to interpret in liquid, this signal can only be used for probing the energy dissipated by the tip when entering in contact with the sample, but no accurate information on membrane mechanics can be deduced from such analyses. The topography clearly indicated a decrease of virus height measured as compared to the SLB surrounding the particle upon lowering the pH to 5. This observation could be related to a partial fusion of the virus within the SLB or to a change in virus stiffness at this pH. The latter interpretation is in good agreement with a loss of integrity of the matrix M1 that was observed in EM under acidic conditions where the M1 layer was no longer observed.10,26 This disorganization is the consequence of a M1 conformational change.11 However, we believe that the decrease in height observed here was mainly related to fusion of viral particles since we observed some membrane areas, surrounding viral particles, with very low corrugation but with viscoelastic properties different from the rest of the SLB (illustrated in Figure 3). Fusion of viruses with SLB is also supported by the delineation by the AFM tip of RNPs that are responsible for virus transcription. These structures were abundant in areas with a higher phase signal and easily identified as >10 nm diameter ribbons with a significant variation in length (25−80 nm) and an important curvature as already observed.32 Identification of these ribbons as RNPs is also supported by their tendency to form aggregates that appeared as a result of the assembly of several ribbons to a structure of significant height (white arrowheads in Figure 3C), similarly to RNPs coagulation on the interior surface of the virion as observed by EM.2 This structure is probably composed of M1 since it was described that each RNP makes contact with the matrix layer.10 However, we were unable to observe M1 molecules associated with each RNP under the experimental conditions used. Interestingly, the helical organization of RNPs was sometimes observed in highresolution AFM pictures and suggested a right-handed helix handeness. However, the resolution was not sufficient to surely discriminate between left- and right-handed pitch, both described in two separate EM studies.32,33 The clear identification of RNPs in these experiments however questions the commonly accepted mechanism of fusion. If fusion occurred according to this mechanism, the RNPs should have been masked from the tip by the viral

indicates that part of the viral membrane was removed, probably during the scanning (asterisks in Figure 3C−E). In the zoom in Figure 3E, the upper part clearly showed an intact membrane protruding 3 nm above the surrounding membrane. This value is very low compared to that of the expected height of HA extracellular domains but could be explained by the fact that HA ectodomains are generally disorganized at acidic pH.10 This image also revealed some particles protruding 2.5 nm above the mica that might correspond to several matrix proteins deposited on mica during the dissociation of the M1 layer observed at acidic pH.11 The dissociation of the M1 layer could also explain the fragility of the viral envelope that could have been damaged during tip scanning.

■

DISCUSSION Death by respiratory complications from influenza infections continues to be a major global health concern, and detailed knowledge on the molecular mechanism of viral entry and replication is indispensable. Viral entry is initiated by the recognition of sialic acid moieties present on cell surface by the HA viral proteins,25 which then leads to membrane fusion. To study both processes, a significant number of different experimental models have been developed that are primarily based on viral binding studies to glycophorin or ganglioside containing liposomes.12,26−28 In these models the fusion of two spherical objects, the virus and the liposome, is monitored but in situ target tissues are nonspherical and are formed by planar cellular monolayers of ganglioside exposing cells. However, there is also a significant number of studies that show that the lipid curvature of the liposome is a parameter that influences fusion efficiency of the influenza and other viruses.12,29−31 Here we report a new experimental model that is based on the viral binding to planar, mica supported membranes containing gangliosides. The observations made on this system therefore does not depend on the intrinsic curvature of target membranes. Our images show that the virus is able to bind to these membranes at neutral pH, which contrasts with several other studies where binding was observed at pH 5 but not at pH 7.28 The exposure of bound virus to pH 5 had several consequences: (1) significant reduction of the viral height with respect to the membrane, (2) change in the viscoelastic properties of the membrane surrounding the viral particle, and (3) the observation of elongated structures that correspond to viral RNPs. Taken together, these data suggest that under the experimental conditions used a complete membrane fusion has been achieved. The minimalistic model system described here is thus well suited to study viral binding at neutral pH and to investigate the fusion with a planar membrane. This way, any effects of intrinsic membrane curvature of liposomes are suppressed, and flat membranes are convenient for highresolution AFM imaging. Binding of influenza virus to host cells as well as fusion of viral membrane with the endosomal compartment has been largely studied, highlighting the importance of both sialic acid moieties on the host membrane and acidic pH for virus fusion and the release of its genetic material into the cytoplasm. In this paper we used planar artificial bilayers containing the ganglioside GM3 to get new insights into the different steps of replication of influenza virus. As expected, the presence of this lipid containing a polar head with sialic acid is required for virus binding to the SLB. Interestingly, even if the membrane was supported on mica, this minimal system was sufficient to allow viral fusion with the artificial membrane when the pH was 11398

dx.doi.org/10.1021/la502591a | Langmuir 2014, 30, 11394−11400

Langmuir

■

membrane. The observation of RNPs in these experiments is explained by the fact that, once the fusion completely achieved, the viral membrane was often removed by the tip, most likely due to its fragility that is supported by the impossibility to get high-resolution imaging under theses conditions (data not shown). Such fragility can be explained by the disruption of the M1 layer described above as well as structural modifications and rearrangements of HA upon acidification. This removal was clearly observed in the zoomed areas in Figure 3. Thus, the fusion process observed with SLB appears to be identical to what is observed with liposomes or in cells. Such viral membrane fragility could facilitate the release of genetic material into the host cell. Under neutral pH conditions that do not favor fusion of viruses, we also provided precise 3D structural details of the surface of intact influenza viruses in interaction with SLB and under physiological conditions. As already observed with AFM for viruses produced in eggs,34 viral particles were generally spherical but can adopt a diversity of shapes. The viral surface was corrugated, probably due to the presence of HA and NA proteins that differently protrude from the lipid envelope and to the fact that HA and NA are not always homogeneously distributed at the viral membrane at neutral pH. Corrugations could also be the consequence of the organization of proteins into clusters that we observed for several viruses. Topography of these clusters was indeed solved by the AFM tip with a lateral resolution similar to the one recorded in EM, but AFM provided additional information in the z direction. The consequence of this organization is the presence of proteinfree membrane areas, revealing the lipid matrix that was identified by the phase signal recorded in intermittent contact mode AFM. These protein-free areas nicely compare with gaps described in EM tomograms.2 According to these previous EM pictures, it is tempting to speculate that the isolated small clusters could be mainly composed of NA. Taken together, our study confirms that initial recognition of sialic acid moieties by influenza virus is sufficient for viral fusion. It demonstrates that a minimal system composed of a fluid lipid matrix is suitable for monitoring viral fusion at acidic pH, strongly suggesting that fusion priming is mainly due to membrane destabilization of the host cell by the viral membrane modified under acidic conditions. Our result also highlight that tapping mode AFM performed in liquid can provide fine structural details of viral membranes as well as RNPs at a nanometer lateral and vertical resolution.

■

REFERENCES

(1) Steinhauer, D. A. Role of Hemagglutinin Cleavage for the Pathogenicity of Influenza Virus. Virology 1999, 258, 1−20. (2) Harris, A.; Cardone, G.; Winkler, D. C.; Heymann, J. B.; Brecher, M.; White, J. M.; Steven, A. C. Influenza Virus Pleiomorphy Characterized by Cryoelectron Tomography. Proc. Natl. Acad. Sci. U. S. A. 2006, 103, 19123−19127. (3) Giocondi, M. C.; Ronzon, F.; Nicolai, M. C.; Dosset, P.; Milhiet, P. E.; Chevalier, M.; Le Grimellec, C. 0rganization of Influenza A Virus Envelope at Neutral and Low pH. J. Gen. Virol. 2009, 91, 329−338. (4) Rossman, J. S.; Lamb, R. A. Influenza Virus Assembly and Budding. Virology 2011, 411, 229−236. (5) Maeda, T.; Kawasaki, K.; Ohnishi, S. Interaction of Influenza Virus Hemagglutinin with Target Membrane Lipids Is a Key Step in Virus-Induced Hemolysis and Fusion at pH 5.2. Proc. Natl. Acad. Sci. U. S. A. 1981, 78, 4133−4137. (6) White, J.; Kartenbeck, J.; Helenius, A. Membrane Fusion Activity of Influenza Virus. EMBO J. 1982, 1, 217−222. (7) Godley, L.; Pfeifer, J.; Steinhauer, D.; Ely, B.; Shaw, G.; Kaufmann, R.; Suchanek, E.; Pabo, C.; Skehel, J. J.; Wiley, D. C. Introduction of Intersubunit Disulfide Bonds in the Membrane-Distal Region of the Influenza Hemagglutinin Abolishes Membrane Fusion Activity. Cell 1992, 68, 635−645. (8) Stegmann, T.; White, J. M.; Helenius, A. Intermediates in Influenza Induced Membrane Fusion. EMBO J. 1990, 9, 4231−4241. (9) White, J. M.; Wilson, I. A. Anti-Peptide Antibodies Detect Steps in a Protein Conformational Change: Low-pH Activation of the Influenza Virus Hemagglutinin. J. Cell Biol. 1987, 105, 2887−2896. (10) Fontana, J.; Cardone, G.; Heymann, J. B.; Winkler, D. C.; Steven, A. C. Structural Changes in Influenza Virus at Low pH Characterized by Cryo-Electron Tomography. J. Virol. 2012, 86, 2919−2929. (11) Fontana, J.; Steven, A. C. At Low pH, Influenza Virus Matrix Protein M1 Undergoes a Conformational Change prior to Dissociating from the Membrane. J. Virol. 2013, 87, 5621−5628. (12) Kanaseki, T.; Kawasaki, K.; Murata, M.; Ikeuchi, Y.; Ohnishi, S. Structural Features of Membrane Fusion between Influenza Virus and Liposome as Revealed by Quick-Freezing Electron Microscopy. J. Cell Biol. 1997, 137, 1041−1056. (13) McPherson, A.; Kuznetsov, Y. G. Atomic Force Microscopy Investigation of Viruses. Methods Mol. Biol. (Clifton, N. J.) 2011, 736, 171−195. (14) Seantier, B.; Giocondi, M.; Le Grimellec, C.; Milhiet, P. Probing Supporting Model and Native Membranes Using AFM. Curr. Opin. Colloid Interface Sci. 2008, 13, 326−337. (15) Berquand, A.; Levy, D.; Gubellini, F.; Le Grimellec, C.; Milhiet, P. E. Influence of Calcium on Direct Incorporation of Membrane Proteins into in-Plane Lipid Bilayer. Ultramicroscopy 2007, 107, 928− 933. (16) Kuznetsov, Y. G.; McPherson, A. Atomic Force Microscopy Investigation of Turnip Yellow Mosaic Virus Capsid Disruption and RNA Extrusion. Virology 2006, 352, 329−337. (17) Wegmann, S.; Muller, D. J.; Mandelkow, E. Investigating Fibrillar Aggregates of Tau Protein by Atomic Force Microscopy. Methods Mol. Biol. (Clifton, N. J.) 2012, 849, 169−183. (18) Burness, A. T.; Pardoe, I. U. A Sialoglycopeptide from Human Erythrocytes with Receptor-like Properties for Encephalomyocarditis and Influenza Viruses. J. Gen. Virol. 1983, 64, 1137−1148. (19) Slepushkin, V. A.; Starov, A. I.; Bukrinskaya, A. G.; Imbs, A. B.; Martynova, M. A.; Kogtev, L. S.; Vodovozova, E. L.; Timofeeva, N. G.; Molotkovsky, J. G.; Bergelson, L. D. Interaction of Influenza Virus with Gangliosides and Liposomes Containing Gangliosides. Eur. J. Biochem. FEBS 1988, 173, 599−605. (20) El Kirat, K.; Morandat, S.; Dufrene, Y. F. Nanoscale Analysis of Supported Lipid Bilayers Using Atomic Force Microscopy. Biochim. Biophys. Acta 2009, 1798, 750−765. (21) Bonnafous, P.; Nicolaï, M.-C.; Taveau, J.-C.; Chevalier, M.; Barrière, F.; Medina, J.; Le Bihan, O.; Adam, O.; Ronzon, F.; Lambert,

ASSOCIATED CONTENT

S Supporting Information *

AFM topography of the lipid bilayer after influenza virus binding (Figure S1); measurement of fwhm and heights of viral particles (Figure S2). This material is available free of charge via the Internet at http://pubs.acs.org.

■

Article

AUTHOR INFORMATION

Corresponding Author

*E-mail [email protected] (P.-E.M.). Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS S.D. is a recipient of the French Ministry of Education and Research. 11399

dx.doi.org/10.1021/la502591a | Langmuir 2014, 30, 11394−11400

Langmuir

Article

O. Treatment of Influenza Virus with Beta-Propiolactone Alters Viral Membrane Fusion. Biochim. Biophys. Acta 2013, 1838, 355−363. (22) De Jong, D. H.; Lopez, C. A.; Marrink, S. J. Molecular View on Protein Sorting into Liquid-Ordered Membrane Domains Mediated by Gangliosides and Lipid Anchors. Faraday Discuss. 2013, 161, 347−363 ; discussion 419−459. (23) Yuan, C.; Furlong, J.; Burgos, P.; Johnston, L. J. The Size of Lipid Rafts: An Atomic Force Microscopy Study of Ganglioside GM1 Domains in Sphingomyelin/DOPC/Cholesterol Membranes. Biophys. J. 2002, 82, 2526−2535. (24) Noda, T.; Sagara, H.; Yen, A.; Takada, A.; Kida, H.; Cheng, R. H.; Kawaoka, Y. Architecture of Ribonucleoprotein Complexes in Influenza A Virus Particles. Nature 2006, 439, 490−492. (25) Schauer, R. Sialic Acids as Regulators of Molecular and Cellular Interactions. Curr. Opin. Struct. Biol. 2009, 19, 507−514. (26) Lee, K. K. Architecture of a Nascent Viral Fusion Pore. EMBO J. 2010, 29, 1299−1311. (27) Nussbaum, O.; Rott, R.; Loyter, A. Fusion of Influenza Virus Particles with Liposomes: Requirement for Cholesterol and Virus Receptors to Allow Fusion with and Lysis of Neutral but Not of Negatively Charged Liposomes. J. Gen. Virol. 1992, 73 (Pt 11), 2831− 2837. (28) Wessels, L.; Elting, M. W.; Scimeca, D.; Weninger, K. Rapid Membrane Fusion of Individual Virus Particles with Supported Lipid Bilayers. Biophys. J. 2007, 93, 526−538. (29) Baljinnyam, B.; Schroth-Diez, B.; Korte, T.; Herrmann, A. Lysolipids Do Not Inhibit Influenza Virus Fusion by Interaction with Hemagglutinin. J. Biol. Chem. 2002, 277, 20461−20467. (30) Chernomordik, L. V.; Leikina, E.; Frolov, V.; Bronk, P.; Zimmerberg, J. An Early Stage of Membrane Fusion Mediated by the Low pH Conformation of Influenza Hemagglutinin Depends upon Membrane Lipids. J. Cell Biol. 1997, 136, 81−93. (31) Stiasny, K.; Heinz, F. X. Effect of Membrane CurvatureModifying Lipids on Membrane Fusion by Tick-Borne Encephalitis Virus. J. Virol. 2004, 78, 8536−8542. (32) Arranz, R.; Coloma, R.; Chichón, F. J.; Conesa, J. J.; Carrascosa, J. L.; Valpuesta, J. M.; Ortín, J.; Martín-Benito, J. The Structure of Native Influenza Virion Ribonucleoproteins. Science 2012, 338, 1634− 1637. (33) Moeller, A.; Kirchdoerfer, R. N.; Potter, C. S.; Carragher, B.; Wilson, I. A. Organization of the Influenza Virus Replication Machinery. Science 2012, 338, 1631−1634. (34) Wrigley, N. G. Electron Microscopy of Influenza Virus. Br. Med. Bull. 1979, 35, 35−38.

11400

dx.doi.org/10.1021/la502591a | Langmuir 2014, 30, 11394−11400