Visualization of the Two-Step Fusion Process of the Retrovirus AvianSarcoma/Leukosis Virus by Cryo-Electron Tomography

Giovanni Cardone,a* Matthew Brecher,b* Juan Fontana,a Dennis C. Winkler,a Carmen Butan,a* Judith M. White,b andAlasdair C. Stevena

Laboratory of Structural Biology, National Institute of Arthritis and Musculoskeletal and Skin Diseases, National Institutes of Health, Bethesda, Maryland, USA,a andDepartments of Cell Biology and Microbiology, University of Virginia, Charlottesville, Virginia, USAb

Retrovirus infection starts with the binding of envelope glycoproteins to host cell receptors. Subsequently, conformationalchanges in the glycoproteins trigger fusion of the viral and cellular membranes. Some retroviruses, such as avian sarcoma/leuko-sis virus (ASLV), employ a two-step mechanism in which receptor binding precedes low-pH activation and fusion. We used cryo-electron tomography to study virion/receptor/liposome complexes that simulate the interactions of ASLV virions with cells.Binding the soluble receptor at neutral pH resulted in virions capable of binding liposomes tightly enough to alter their curva-ture. At virion-liposome interfaces, the glycoproteins are �3-fold more concentrated than elsewhere in the viral envelope, indi-cating specific recruitment to these sites. Subtomogram averaging showed that the oblate globular domain in the prehairpin in-termediate (presumably the receptor-binding domain) is connected to both the target and the viral membrane by 2.5-nm-longstalks and is partially disordered, compared with its native conformation. Upon lowering the pH, fusion took place. Fusion is astochastic process that, once initiated, must be rapid, as only final (postfusion) products were observed. These fusion productsshowed glycoprotein spikes on their surface, with their interiors occupied by patches of dense material but without capsids, im-plying their disassembly. In addition, some of the products presented a density layer underlying and resolved from the viralmembrane, which may represent detachment of the matrix protein to facilitate the fusion process.

Entry of enveloped viruses into host cells occurs by means offusion between the membranes of the virus and the host, an

event mediated by one or more of the glycoproteins that stud thevirion surface. Although viral fusion proteins fall into three struc-tural classes (20, 53) and vary in their modes of activation, all arethought to undergo similar sets of conformational changes as fu-sion proceeds. Upon activation, fusion proteins first switch fromtheir relatively compact native state into an elongated intermedi-ate, in which a hydrophobic motif, the fusion peptide, is exposedand becomes embedded in the target membrane. Inferences con-cerning this “prehairpin” intermediate have mostly been indirect,based primarily on biochemical studies (17, 27, 30, 35, 42, 48).After engaging the target membrane, this intermediate is ready totransition into the postfusion state. In all classes of fusion pro-teins, the prehairpin conformer folds back on itself to form thehairpin. In class I proteins, which are trimeric, the predominantfeature of this end state is a six-helix bundle composed of threehairpins. Formation of the hairpins is coupled with merging of theviral and target membranes and is thought to occur in a concertedprocess involving multiple glycoprotein spikes. First, a zone ofhemifusion is created, i.e., a single bilayer comprising one leafletfrom each membrane; ultimately, a fusion pore opens, allowingthe viral nucleocapsid to pass into the host cell cytoplasm (20,26, 53).

Several recent studies have used cryo-electron microscopy(cryo-EM) to examine intermediates in the fusion process medi-ated by the fusogenic proteins of Moloney murine leukemia virus(32, 55), herpes simplex virus 1 (36), influenza virus (15, 29), andvesicular stomatitis virus (31). However, no cryo-EM study hasyet reported visualization of a prehairpin intermediate.

Avian sarcoma/leukosis virus (ASLV) is an alpharetrovirusthat fuses with target cells via a two-step mechanism. Interactionof ASLV Env, a class I fusion protein, with its receptor, Tva, in-

duces conformational changes in Env that expose its fusion loop(9, 21). Subsequent exposure to low pH induces further confor-mational changes in Env that mediate fusion (35, 38, 39, 51). Be-cause interaction with the receptor induces formation of the in-ferred prehairpin intermediate without progressing to fusion,ASLV provides an excellent system for visualizing the target mem-brane-binding (prehairpin) stage of fusion, as well as the finalfusion products, produced by subsequent exposure to low pH.

In this study, we analyzed successive stages of ASLV fusion withliposomes by using cryo-electron tomography (cryo-ET), a tech-nique that allows visualization of individual pleiomorphic parti-cles, such as ASLV virions, in three dimensions in their native state(19). This approach has already been used to characterize virionmorphology and capsid polymorphism of the closely related Roussarcoma virus (RSV) (2, 3). The experiments in this study werefacilitated by the ability of sTVA, a soluble form of the receptor(which is normally membrane bound via either a glycosylphos-

Received 19 July 2012 Accepted 20 August 2012

Published ahead of print 29 August 2012

Address correspondence to Alasdair C. Steven, stevena@mail.nih.gov, orJudith M. White, jw7g@virginia.edu.

* Present address: Giovanni Cardone, Department of Chemistry and Biochemistry,University of California, La Jolla, California, USA; Matthew Brecher, Computationaland Structural Biology Laboratory, Division of Genetics, New York StateDepartment of Health, Center for Medical Science, Albany, New York, USA;Carmen Butan, Department of Biological Chemistry and Molecular Pharmacology,Harvard Medical School, Boston, Massachusetts, USA.

G.C., M.B., and J.F. contributed equally to this article.

Supplemental material for this article may be found at http://jvi.asm.org/.

Copyright © 2012, American Society for Microbiology. All Rights Reserved.

doi:10.1128/JVI.01880-12

November 2012 Volume 86 Number 22 Journal of Virology p. 12129–12137 jvi.asm.org 12129

on February 13, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

phatidylinositol anchor, i.e., Tva800, or a transmembrane do-main, i.e., Tva950), to bind to Env, preparing it for acid-induciblefusion. These observations provide new insights into the mecha-nism of membrane fusion and retroviral cell entry.

MATERIALS AND METHODSProduction and purification of ASLV, sTva, and R99. ASLV was pro-duced by using DF-1 cells chronically infected with the RCASBP(A) vec-tor system as described previously (13). Virus-containing supernatantswere centrifuged at 2,500 rpm to clear debris and concentrated �50� bycentrifugation in Vivaspin-20 300-kDa cutoff concentrators (catalognumber VS2052; Sartorius-Stedim). Concentrated supernatants were pel-leted through a 1-ml cushion of 20% (wt/vol) sucrose by centrifugation at175,000 � g in an SW55 rotor (Beckman) for 2 h. The viral pellets wereresuspended in HM buffer (20 mM HEPES, 20 mM morpholineethane-sulfonic acid, 130 mM NaCl; pH 7.5), layered on top of a 25%-to-50%(wt/vol) sucrose step gradient, and centrifuged at 50,000 � g in an SW55rotor for 16 h. The virus-containing band at the 25%-50% sucrose inter-face was collected, resuspended in HM, and centrifuged at 100,000 � g for2 h to remove residual sucrose. The final pellet was resuspended in �200�l of HM buffer. Its protein concentration was determined in a bicin-choninic acid assay (catalog number 23228; Thermoscientific). Solublereceptor (sTva) was produced as described previously (52). C-helix pep-tide R99 was produced as described in reference 13.

Production of liposomes. To produce fluorescent liposomes, lipidmixtures of phosphatidylcholine (PC; number catalog 830051P; Avanti),phosphatidylethanolamine (PE; catalog number 841118P; Avanti), sph-ingomyelin (SPH; 860061P; Avanti), cholesterol (CH; C-3137; Sigma),rhodamine phosphatidylethanolamine (RH-PE; 810146; Avanti), and(N-7-nitrobenz-2-oxa-1,3-diazol-4-yl)-1,2-dihexadecanoyl-sn-glycero-3-phosphatidylethanolamine (NBD-PE; N-360; Molecular Probes) at aratio of 1:1:1:1.5:0.045:0.11 were dried with N2, lyophilized overnight, andresuspended in HM buffer to a 2 mM lipid concentration. The liposomesuspension was freeze-thawed five times, extruded 30 times through a100-nm filter, and stored at 4°C. Liposomes for electron microscopy wereproduced as described above, except that PC, PE, SPH, and CH weremixed in a 1:1:1:1.5 ratio and resuspended at a 20 mM lipid concentration.

Fluorescence resonance energy transfer (FRET) lipid mixing assay.Fusion was measured essentially as described in reference 10. For thestandard assay, 20 �l reaction mixtures (in HM buffer) were set up in wellsof a 384-well plate as follows: 20 to 30 �g of ASLV and sTva (at 1 �M) weremixed for 30 min at 4°C. Fluorescent liposomes (final concentration, 25 to50 �M) were added, and the plates were incubated for an additional 5 minat 4°C. At this time the samples were incubated for 15 min at 37°C (togenerate virus-liposome complexes bridged by ASLV Env in its prehairpinstate). The plate was then placed in a fluorescent plate reader, and thebaseline NBD fluorescence (excitation at 460 nm, emission at 540 nm)was measured for 5 min at 37°C. Unless otherwise specified, the sampleswere then acidified to pH 5.0 and the fluorescence was monitored for 10min at 37°C. The maximum possible NBD fluorescence was determinedby adding NP-40 to a final concentration of 1% and measuring fluores-cence at 37°C for 15 min. The percent fusion was calculated as follows:(FpH � F0)/(FT � F0) � 100, where F0 is the baseline fluorescence (pH7.5), FpH is the fluorescence value averaged over the plateau at pH 5.0, andFT is the fluorescence at infinite dilution (after disruption of the mem-branes with 1% NP-40).

Samples for electron microscopy. For EM analysis, 50 �g of ASLVwas incubated for 30 min on ice with sTva (0.3 �M). Liposomes wereadded (0.5 mM), and the samples were further incubated on ice for 20min. The mixture was then incubated at pH 7.4 for 30 min at 37°C to allowformation of virus-liposome complexes. Where indicated, samples werethen treated at pH 7.4, 6.1, or 5.0 for 5 min at 37°C. In the samples withoutreceptor, HM buffer was used instead of sTva.

Cryo-electron microscopy. Samples were mixed 1:1 with a suspen-sion of 10-nm colloidal gold particles (Ted Pella, Redding, CA) to serve as

fiducial markers, applied to Quantifoil R2/2 holey carbon grids (SPI, WestChester, PA) and plunge-frozen into liquid ethane by using a Vitrobot(FEI, Hillsboro, OR). Tilt series of projections were acquired on a Tec-nai-12 (FEI) microscope, operating at 120 keV, using SerialEM (34). Themicroscope was equipped with an energy filter set in zero-energy-lossmode (slit width, 20 eV). Data were recorded on a charge-coupled-devicecamera, at 2,048 by 2,048 pixels (Gatan, Warrendale, PA). At least 500viruses (�100 images) per condition were analyzed in two-dimensionalprojections. For tomography, a magnification of �38,500 was employed,corresponding to a pixel size of 0.78 nm. Images were recorded at �4 �mand �6 �m nominal defocus, in 1° or 2° angular increments, over anangular range of ��66° to 66°. The total electron dose for each tilt serieswas �80 electrons/Å2 or lower. The numbers of tilt series acquired fromthe samples containing either isolated virions or virions mixed with re-ceptors and liposomes at pH 7.4 and pH 5.0 were 4, 28, and 6, respectively.

Image processing. Tilt series projections were aligned using theIMOD program (28), with the gold particles serving as fiducial markers.Tomograms were reconstructed by a weighted back-projection method(46) and then noise-filtered using two different procedures in Bsoft (22).For visualization, bilateral filtering (24) was applied to the reconstruc-tions, with the following parameters: kernel size, 15 pixels; � of spatialfilter, 5 pixels; � of the range filter, equal to the standard deviation (SD) ofthe tomogram. Stronger deionizing by nonlinear anisotropic diffusion(16) was applied to the subtomogram data extracted for the statisticalanalysis of the receptor-primed interaction between virions and lipo-somes at neutral pH. The edge-enhancing diffusion parameter was set to1/100 of the standard deviation of the subtomograms, and 40 iterationswere performed. Tomogram quality was assessed by using the NLOOapproach (4) to estimate the in-plane resolution of three virions extractedfrom each tomogram. In each case the resolution was around 6 nm, basedon a threshold criterion of 0.5. Gray-scale sections were prepared usingthe tools available in Bsoft (22), while segmentation was performed withinAmira (Visage Imaging).

Statistical analysis of Env proteins. Glycoprotein spikes on 14 iso-lated virions and on 24 virions in receptor/virion/liposome complexeswere located manually, using the 3dmod utility in IMOD (28), and theircoordinates were recorded. Data for the junctional spikes and nonjunc-tional spikes were stored separately. For each spike, we measured thedistances between it and its three nearest neighbors. The same data wereused to calculate the nearest neighbor index (NNI) (5). Specifically, weused the K-order NNI, which is the ratio of observed k-th nearest neigh-bor distance to the average random distance, i.e., the distance that wouldbe observed if the points were randomly distributed (6). Thus, a randomdistribution of points gives an NNI value close to 1, while values less than1 are evidence of clustering and values greater than 1 are indicative of auniform pattern. Our NNI was measured as the average of the K-orderNNI, with K ranging from 1 to 3. The NNI for spikes located at the junc-tions was, on average, equal to 1.8, while the NNI for spikes distal to thejunctions was equal to 1.1. The same value of 1.1 was obtained for the NNImeasured on the spikes decorating isolated virions. All the analyses wereperformed with the statistical software R (45; http://www.R-project.org/)and its packages fractal (8), misc3d (14), geometry (1) and tripack (47),with custom scripts written in its programming language.

Env averaging. The position coordinates of the glycoproteins used forthe statistical analysis (see above) were also used for subtomogram aver-aging, after an initial screening. Junctional (n � 672) and nonjunctional(n � 2,392) spikes were extracted from raw tomograms in volumes of 42by 42 by 42 voxels and 34 by 34 by 34 voxels, respectively. A total of 1,197Env spikes in the native conformation were located manually on a tomo-gram of isolated virions without receptors and then extracted in volumesof 44 by 44 by 44 voxels. For each virion, an estimate of its center was usedto determine the initial orientations of all of its spikes. These subtomo-grams, each containing one spike and the adjacent regions, were subjectedto multiple (up to 8) iterations of alignment and averaging. As initialtemplates, a computationally generated cylinder on a plane or a cylinder

Cardone et al.

12130 jvi.asm.org Journal of Virology

on February 13, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

between two planes was used for nonjunctional spikes or junctionalspikes, respectively. The correlation metric used to determine the spikeorientations was compensated for the missing wedge, and after two iter-ations the noise in real space was masked from both the template and theraw particles, using their most recent solutions to properly orient themask. Alignment and averaging procedures were performed with a set ofPython programs that were interfaced to routines from Bsoft. Three-foldsymmetry was imposed on the averages after each iteration. After com-pleting the iterations, the aligned particles were classified using a maxi-mum likelihood approach, as implemented in xmipp (50), to discard lessconsistent particles, and the final averages were then calculated. The num-bers of contributing particles were 513 (junctional spikes), 2,000 (non-junctional spikes), and 680 (native Env spikes). Resolution was estimatedby calculating the Fourier shell correlation curve between the reconstruc-tions from two half-data sets (49), resulting in values between 4.0 and 4.5nm at a cutoff of 0.5.

RESULTSAt neutral pH, receptor-bound virions and liposomes form sta-ble fusion-competent complexes. Incubation with sTva at neu-tral pH and temperatures above 25°C induces Env on ASLV viri-ons to change their conformation (11, 18, 35) and bind toliposomes (9, 21). Subsequent shifting to a low pH triggers fusion(35, 39, 51). It has also been shown that ASLV that has becomeendocytosed into cells in the presence of NH4Cl (via interactionwith the receptor, Tva 800) traffics into endosomes, but after 6 h itcan still proceed to infect these cells when the NH4Cl is washed out(40). This suggests that Env can persist in its receptor-triggeredstate for extended periods.

In order to determine how long receptor-triggered virions re-main fusion competent, we used a fluorescence-based fusion assay(Fig. 1). ASLV was incubated with sTva in the presence or absence

of liposomes containing a pair of lipids capable of FRET (RH-PEand NBD-PE), at neutral pH at 37°C for various times and thenswitched to pH 5.0. Fusion was measured from the gain in NBDemission, as the FRET lipids were diluted upon merging of theliposomes with viral membranes. We found that ASLV/sTva couldbe maintained in the presence of liposomes for at least 8 h at 37°Cwithout a significant loss in fusion activity (Fig. 1A, black col-umns). If the C-helix peptide analog R99 was added just prior toacidification, fusion was strongly inhibited (Fig. 1A, hatched col-umns), even if the ASLV-liposome complexes had been incubatedfor 8 h at 37°C. (C-helix peptides are derived from Env and inhibitfusion by binding to the N-terminal heptad repeats of Env in itsextended prehairpin conformation, thereby blocking formationof the six-helix bundle [13, 17, 23, 37, 42, 54].)

The latter observation indicated that Env had remained in theprehairpin conformation throughout. However, if liposomeswere not added to sTva-ASLV complexes incubated for varioustimes (at pH 7.5 and 37°C) until 5 min prior to acidification, aprogressive loss of fusion activity was observed (Fig. 1A, gray col-umns). This inactivation might reflect aggregation of Env (in vi-rions) or insertion of the fusion loop into the viral membrane, ashas been observed for influenza virus hemagglutinin andparamyxovirus F proteins (7, 44). We conclude that sTva-ASLV-liposome fusion intermediates remained stable for at least 8 h at37°C (Fig. 1A). This time could be increased to at least 32 h if theprefusion incubation took place at 4°C (Fig. 1B).

Env spikes accumulate at virion-liposome junctions. The ob-served longevity of the virion-liposome complexes encouragedus to examine them by cryo-EM. After incubation with sTvaand liposomes at 37°C and neutral pH, more than 90% of thevirions were observed to be in contact with two or more lipo-somes (Fig. 2A). These interactions were accompanied by achange in membrane curvature in some liposomes, reflectingtheir greater malleability. In contrast, in the absence of sTva,fewer than 5% of virions were seen to be in contact with two ormore liposomes (Fig. 2B).

We then performed cryo-ET on receptor-virion-liposomecomplexes formed at neutral pH, whose susceptibility to fusionblocking by R99 (see above) indicated that Env was in the prehair-pin state. From a total of 11 tomograms, we identified 59 virions(out of 101 total) that were clearly engaged with one or moreliposomes (Fig. 2C to I). There were a total of 134 such interac-tions, and the highest number of interacting liposomes per virionwas 7. The interactions involved extensive interfaces, where mul-tiple Env spikes were observed to be present, as illustrated by the8-nm-thick slice through a tomogram in Fig. 2C (arrowheads).The spikes appeared closely packed, giving almost continuous rib-bons of density. This clustering was not observed for Env spikesoutside the virion-liposome interfaces (Fig. 2E and F), nor oncontrol virions that had not been mixed with receptors and lipo-somes (Fig. 2J and K). As already observed from the cryo-electronmicrographs (above), the interfaces exhibited alterations in lipo-some curvature. In most cases (�80%), the liposome membranewas visibly flattened where it bound to the virion (Fig. 2D). In afew cases (�8%) where the interface was larger, the liposome wasmolded around the virion (Fig. 2H and 3).

Next, we analyzed in greater detail the most clearly definedjunctions (n � 39, involving 24 virions). The intermembranespacing (virus to liposome) was uniformly 16 nm, center to center.Env spikes in these junctions appeared as globular densities, ap-

FIG 1 Membrane-associated prehairpins remain fusion competent for at least8 h at 37°C and for 32 h at 4°C. (A) Samples were processed as follows. Step 1:sTva and ASLV were mixed for 30 min at 4°C in 384-well plates. Step 2: lipo-somes containing NBD-PE and RH-PE were either added (black or hatchedbars) or omitted (gray), and incubations continued for 5 min at 4°C. Step 3:samples were warmed to 37°C for 15 min (to trigger Env to convert to itsprehairpin state). Step 4: samples were incubated for the indicated times at37°C. Step 5: liposomes were added to samples that had not previously receivedthem (gray), and R99 (100 �g/ml) was added to other samples (bars withhatches). Step 6: baseline NBD fluorescence was recorded for 5 min at 37°C.Step 7: samples were acidified to pH 5.0, and NBD fluorescence was monitoredfor 10 min at 37°C. The percent fusion was calculated as described in Materialsand Methods and normalized to the percent fusion at 0 min with the liposomesample. All samples were analyzed in triplicate, and error bars represent stan-dard deviations. ND, not done. One of two experiments with similar results isshown. (B) Results of an experiment performed as described for panel A,except that all samples contained liposomes from step 2 onwards, and step 4incubations were performed at 4°C.

Cryo-ET of ASLV Interacting with Liposomes

November 2012 Volume 86 Number 22 jvi.asm.org 12131

on February 13, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

proximately midway between the membranes. Env spikes in non-junctional regions had much the same appearance, although theyappeared slightly larger. The number of spikes per junction variedfrom 4 to 47 (median, 15), roughly in proportion to the size (area)

of the junction. (The total number of spikes per ASLV virion inthese preparations was determined to be 111 � 33 [SD], morethan for the related Rous sarcoma virus [3]. The spread representsalmost entirely particle-to-particle variability, not experimental

FIG 2 ASLV virions that were incubated with liposomes, with or without sTva, at pH 7.4. When the receptor is present (A and C to I), virion-liposomeinteractions are observed. (B) When the receptor is not present, virions and liposomes do not interact. The insert in panel A is an enlargement of the regionaround the virion (marked with a black arrow). (C) Central slice through a cryo-electron tomogram of ASLV virions incubated with sTva and liposomes atneutral pH. Arrowheads mark virus-liposome junctions. A polyhedral capsid (core) inside a virion is marked with a white arrow. (D to F) Examples of junctionsbetween virions and unilamellar liposomes. (G) Interaction of a virion with a bilamellar liposome; the curvature of the liposome was not altered. (H and I)Extensive interfaces where a liposome is molded around a virion. (J and K) Control virions at pH 7.4 in the absence of sTva and liposomes. The tomographic slicesare 7.8 nm thick. Bars, 100 nm.

Cardone et al.

12132 jvi.asm.org Journal of Virology

on February 13, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

error.) Of note, the concentrations of spikes in junctions appearedhigher than elsewhere in the viral envelope (Fig. 3; see also MoviesS1 and S2 in the supplemental material). To examine this obser-vation on a quantitative basis, we determined the distributions ofinterspike spacings in both junctional and nonjunctional regions(Fig. 4). This analysis yielded average interspike spacings of 11 � 3nm and 19 � 6 nm, respectively. For comparison, we repeated theanalysis on the glycoproteins identified on 14 virions in the ab-sence of receptors and liposomes. In this case, the interspike spac-ing was 19 � 7 nm. Use of the nearest neighbor index (see Mate-rials and Methods), which measures the degree of spatialdispersion, confirmed that the spikes tended to be more closelypacked in junctions than elsewhere. These data implied that Env isable to diffuse within the viral membrane and be recruited to thejunctions.

Visualization of Env in the prehairpin state. In order to com-pare native Env with the receptor-bound and membrane-embed-ded prehairpin conformations, we performed subtomogram av-eraging for each state. For native Env, we averaged spikes takenfrom virions that had not been mixed with receptors or liposomes.Thus visualized (Fig. 5A), the molecule appeared as an oblate den-sity, 8 nm in diameter and 5 nm thick, and significantly offset fromthe membrane. The observed offset implied that the globular do-main is connected to the membrane by a thin stalk that was notdirectly visualized at the current resolution (�4 nm) (see Discus-sion). Then we analyzed the spikes in receptor-virion-liposomecomplexes formed at neutral pH. The overall shape and dimen-

sions of Tva-bound spikes not interacting with liposomes(Fig. 5B) were similar to those of the native glycoprotein, exceptthat they appeared slightly less wide (diameter, 6.5 nm; thickness,5 nm) and less defined, despite the expected additional mass ofthree receptors (50 kDa for three sTVAs, added to the 150 kDa of

FIG 3 Tomographic reconstruction of an ASLV virion interacting with twoliposomes. (A to C) Tomographic slices at different depths. Bar, 50 nm. (D)Surface rendering with the outer surface of the viral membrane shown in gray,Env spikes in green, liposomes in yellow, and the capsid in orange.

FIG 4 Analysis of the distribution of ASLV glycoproteins. Nearest neighbordistance and the NNI are shown. (A) Distribution of ASLV glycoproteins fromcontrol virions not incubated with sTva or liposomes. (B and C) Distributionsof glycoproteins on virions incubated with sTva and engaged with one or moreliposomes; the analysis was performed separately for glycoproteins outsidevirus-liposome junctions (B) and in virus-liposome junctions (C).

Cryo-ET of ASLV Interacting with Liposomes

November 2012 Volume 86 Number 22 jvi.asm.org 12133

on February 13, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

three glycosylated Env ectodomains). This suggests that the con-formational change triggered by receptor binding involves somedisordering.

In junctions, the Env spikes are assumed to be in the prehairpinstate. Here, a globular domain remained visible, although it wassmaller still (diameter, �6 nm; thickness, �4 nm) and apparentlyembedded in a continuous sheet of density that presumably arosefrom averaging the close-packed Env ectodomains. This domainwas centered between the two membranes (Fig. 5C), which are 16nm apart (see above). Also, the liposome membrane density wasmarkedly lower than that of the virion envelope, presumably re-flecting that the latter lipid bilayer is backed by the closely apposedmatrix protein layer.

Visualization of fusion products after switching to acidic pH.According to FRET-based fusion assay results, lipid mixing be-tween virions and liposomes proceeds in two stages as the pH islowered: the first stage occurs at only a slightly acidic pH, startingat pH 6.7, and the second stage occurs when the pH falls below 5.8(10). With cryo-EM, we detected some fusion when the complexesthat formed at neutral pH were shifted to pH 6.1, but the incidencewas low (5 out of �500 observed). Despite their rarity, fusionproducts were readily recognizable as such (Fig. 6A, right insert),because their membranes were fringed with glycoprotein spikes(like virions) and their interiors were partially filled with densematerial (as for virions) and partially void (as for liposomes). Fu-sion was more prolific when the pH was lowered to pH 5.0. Underthis condition (Fig. 6C), approximately equal numbers of intactvirions and fusion products were observed; from this we inferredthat �50% of the virions underwent fusion. Fusion productstended to be larger than virions and to have a greater fraction oftheir interiors as voids (Fig. 6C; see also Movie S3 in the supple-mental material). If receptors (sTva) were omitted, no fusion wasobserved at pH 6.1 or 5.0, and virions and liposomes showed littletendency to interact, as at neutral pH (Fig. 6B and D).

The differences between virions and fusion products were bet-

ter observed in tomographic slices (Fig. 6E to K). The fusion prod-ucts analyzed (n � 19) had an average diameter of 146 � 20 nm,consistent with fusion of an average-sized virus (diameter, �130nm) with a smaller liposome. All postfusion complexes had gly-coproteins on their surface and patchy internal densities, whichwe took to be the contents, somewhat redistributed, of the fusedvirions (Fig. 6I and J, arrows). We did not observe more-advancedfusion intermediates or membrane dimpling at virus-liposomeinterfaces. As such, our data concurred with a recent report thatfusion follows rapidly upon acidification of viruses whose glyco-proteins are trapped in the prehairpin conformation (43).

A notable feature of fusion products is the absence of intactcapsids. In unfused virions, capsids are clearly visible in the tomo-graphic slices (Fig. 2D and E and 3A). These observations sug-gested that conditions in the postfusion lumen favor disassembly.

Another property of the postfusion products was that most oftheir envelopes appeared thinner than in virions (Fig. 6F to K; cf.Fig. 2D to K) and had essentially the same thickness as liposomes(Fig. 2C). (In native ASLV virions, the envelope appears thickerbecause the matrix layer is closely apposed to the bilayer and re-solved only marginally if at all in the tomograms.) Also of note,about 25% of the postfusion complexes (n � 5) showed a layer ofdensity that tracked but was clearly resolved from the interiorsurface of the membrane (Fig. 6J and K, arrowheads). This layercovered from �25 to 50% of the interior surface of the postfusionparticle.

DISCUSSION

Viral glycoproteins undergo conformational changes in order toinduce fusion of viral and cellular membranes (20, 53). One of theproposed intermediates in this process is the preharpin conforma-tion; however, although many reports point to its existence, basedon biochemical approaches (17, 35, 42, 48) and by negative-stain-ing EM (27, 30), a prehairpin has never been visualized directly bycryo-EM. Subtomogram averaging (Fig. 5) conveyed this state forASLV-Env spikes that we know, based on the inhibitory effect ofR99, to be in it. Although these data were of limited resolution (to�4 nm), they yielded two incisive observations.

(i) Native Env appears larger than the Env-sTva complex out-side junctions which, in turn, appear larger than the complex injunctions. The mass of Env-sTva should be �50 kDa larger thanthat of the Env ectodomains alone, although sTva has been ob-served to dissociate after ASLV binds to membranes (9, 21).Whether or not sTva is retained, a definite size disparity is ob-served, and it appears to arise from partial disordering of Envupon its binding to the receptor. This interpretation is consistentwith the observation by Delos et al. (11) that sTva binding by SU(the receptor-binding portion of Env) induces a conformationalchange in SU that involves a loss of the �-helical structure, occlu-sion of the hydrophobic surface(s), and occlusion of tryptophanresidues.

(ii) With junctional Env, the main globular density is clearlyobserved midway between the two membranes. By analogy withother class I fusion proteins, in particular with the Ebola virus GP2subunit (12), we expected transmembrane domains in the pre-hairpin conformation to adopt a trimeric coiled-coil stalk madeup of the N-heptad, chain reversal region, and the C-heptad. Ac-cordingly, the observed globular density (Fig. 5C) should havebeen contributed by the SU domains, which appeared to retainsome organized structure in the prehairpin conformation. The

FIG 5 Subtomogram averaging of Env glycoproteins in different states. Sag-ittal sections through the center of the spike are shown. (A) Env from controlvirions not incubated with sTva or liposomes. (B and C) Env from virionsincubated with sTva and engaged with one or more liposomes; separate aver-ages are shown for spikes outside virus-liposome junctions (B) and in virus-liposome junctions (C). Bar, 10 nm.

Cardone et al.

12134 jvi.asm.org Journal of Virology

on February 13, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

FIG 6 ASLV virions, incubated with liposomes, with or without sTva, at pH 6.1 and pH 5.0. At pH 6.1, virion-liposome interactions were similar to thoseobserved at neutral pH (A), although fusion products were occasionally found (A, right insert). (C) At pH 5.0, fewer intact virions and more fusion products(inset) were observed. (B and D) As at neutral pH, when the receptor was not present, virions and liposomes did not interact. (E to K) Tomographic slices ofASLV-liposome fusion products. (E) A fusion product (middle) is readily distinguishable from a virion (top right) or liposomes (bottom left). (F to K) Fusionproducts showed glycoproteins on their surfaces, contained patchy internal densities, and had diameters consistent with the fusion of an average-sized virus witha liposome. Arrows in panels I and J mark “compact” fusion-product contents. (J and K) Examples of fusion products that exhibited a partial layer of densitybeneath their membrane (arrowheads). Bars, 100 nm.

Cryo-ET of ASLV Interacting with Liposomes

November 2012 Volume 86 Number 22 jvi.asm.org 12135

on February 13, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

function of the SU domains after receptor binding remains un-clear, although they could be involved in SU-SU lateral interac-tions that recruit more Env spikes into a growing junction and/orfacilitate the coordinated action of the multiple trimers in produc-ing an effective fusion pore (see below).

Prehairpin Env molecules are recruited into junctions. It isgenerally held that multiple copies of Env are required for fusion,as with other viral fusion proteins, although the exact numbersrequired remain unclear (20, 33, 53). Our data indeed pointed inthis direction. We observed a 3-fold increase in the concentrationof Env spikes in virus-liposome junctions compared with non-junctional regions. Moreover, statistical analysis pointed to atighter packing of spikes within junctions. We can imagine twoscenarios that would have this concentrating effect. Importantly,both require Env to be mobile within the viral membrane. First,the buildup of Env in junctions could be due to trapping of thespikes once their fusion loops have inserted in the liposome mem-brane, causing a decrease in their lateral mobility. Second, theclose packing of Env in junctions could be due to attractive in-tertrimer SU-SU interactions. Such interactions could help to or-ganize spikes into higher-order assemblies, which appear to beneeded to generate an effective fusion pore.

In junctions, the globular domain of Env connects to bothmembranes by narrow stalks. Although the nominal resolutionof the averaged tomographs was modest, their high signal-to-noise ratio and a priori information about the components al-lowed certain more-detailed parameters to be extracted. Thethickness of the target membrane should be 4.5 nm, typical of alipid bilayer, and the position of its center is specified quite pre-cisely by the tomogram (Fig. 5C). Similarly, we estimated the viralenvelope thickness as 7 nm (4.5 nm for the bilayer plus an esti-mated 2.5 nm for the closely associated matrix protein layer). Asthe centers of the two membranes are 16 nm apart, there should beabout 10 nm between the surfaces of the respective bilayers. Thethickness of the globular Env domain is not known exactly, but 5nm would seem a reasonable approximation (an oblate ellipsoidwith a 7-nm major axis and 5-nm minor axis; the density of pro-tein would have a mass of about 160 kDa). This leaves 5 nm for thecombined separations of this domain from the two membranes,and as the domain is centered between them, the connectionsshould have the same length, about 2.5 nm. We performed com-putational simulations with model structures (Fig. 7) that sug-gested that these conclusions are fairly robust to variations in theactual thickness of the globular domain, supporting the supposi-tion that two narrow connections, most likely both coiled coils,connect the remaining globular domain to the viral membraneand to the fusion peptides embedded in the target membrane.Conversely, no placement of a larger globular domain closer to thetarget membrane could reproduce the observed structures.

Fusion products and their implications for the fusion pro-cess. The ALSV-liposome fusion products provided some insightinto this process. On addressing virus-liposome interactions atneutral pH, we observed that only liposome membranes were de-formed. Therefore, the viral envelope is more resistant to defor-mation than the liposome membrane, probably on account of itsbacking layer of matrix protein. (These observations also agreewith what has been seen for influenza virus under similar condi-tions [29].) It follows that, for fusion to take place, the matrixprotein layer should at some point lose its rigidity or detach fromthe viral membrane. In this context, we found the membranes of

postfusion vesicles to be thinner than viral envelopes, suggestingthat the matrix protein had already detached. If not solubilized,this protein should form part of the patchy internal density seen inthese particles (Fig. 6F to K).

Approximately 25% of these membranes showed a layer ofdensity tracking the membrane on the inside and resolved from it.It is plausible that this layer represents matrix protein at an inter-mediate stage of detachment. This could be an effect of pH alone,but if host lipids were to have a lower affinity for matrix proteinthan viral membrane lipids, detachment would be promoted bythe outward diffusion of host lipids from an incipient fusion pore.On the other hand, this layer of density is unlikely to representcapsid protein, as its shape does not resemble capsids as originallyassembled, and conditions inside postfusion products appear notto be conducive to capsid assembly.

ACKNOWLEDGMENTS

We thank Sue Delos for helpful discussions during the initial stages of thiswork and Bryan Fellman at the M.D. Anderson Cancer Center for helpwith the statistical analysis.

This research was supported in part by the Intramural Research Pro-gram of the National Institute of Arthritis and Musculoskeletal and SkinDiseases. Work in the White laboratory was supported by NIH AI22470.M.B. was supported in part by an Infectious Disease Training Grant (5T32AI07046-27) from the NIAID.

REFERENCES1. Barber CB, et al. 2012. Geometry: mesh generation and surface tessela-

tion. R package version 0.3-0. R Development Core Team, Vienna, Aus-tria.

2. Butan C, et al. 2010. Suppression of a morphogenic mutant in Roussarcoma virus capsid protein by a second-site mutation: a cryo-electrontomography study. J. Virol. 84:6377– 6386.

3. Butan C, Winkler DC, Heymann JB, Craven RC, Steven AC. 2008. RSVcapsid polymorphism correlates with polymerization efficiency and enve-lope glycoprotein content: implications that nucleation controls morpho-genesis. J. Mol. Biol. 376:1168 –1181.

4. Cardone G, Grünewald K, Steven AC. 2005. A resolution criterion forelectron tomography based on cross-validation. J. Struct. Biol. 151:117–129.

FIG 7 Computational simulation of the prehairpin conformation at the virus-liposome junction. The prehairpin was modeled as a globule connected to thetwo membranes by a stalk of �2 nm in diameter. Both virus (bottom) andliposome (top) were assumed to have a diameter of 110 nm. Each row shows asimulation for a different size of globule. The first column shows the modelused for each simulation, while the other columns show the simulations atdifferent levels of resolution. See the text for further details.

Cardone et al.

12136 jvi.asm.org Journal of Virology

on February 13, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

5. Clark PJ, Evans FC. 1954. Distance to nearest neighbor as a measure ofspatial relationships in populations. Ecology 35:445– 453.

6. Clark PJ, Evans FC. 1979. Generalization of a nearest neighbor measureof dispersion for use in K-dimensions. Ecology 60:316 –317.

7. Connolly SA, Leser GP, Yin HS, Jardetzky TS, Lamb RA. 2006. Refold-ing of a paramyxovirus F protein from prefusion to postfusion conforma-tions observed by liposome binding and electron microscopy. Proc. Natl.Acad. Sci. U. S. A. 103:17903–17908.

8. Constantine W, Percival D. 2011. Fractal:fractal time series modeling andanalysis. R package version 1.1-1. R Development Core Team, Vienna,Austria.

9. Damico RL, Crane J, Bates P. 1998. Receptor-triggered membrane as-sociation of a model retroviral glycoprotein. Proc. Natl. Acad. Sci. U. S. A.95:2580 –2585.

10. Delos SE, et al. 2008. Cysteines flanking the internal fusion peptide arerequired for the avian sarcoma/leukosis virus glycoprotein to mediate thelipid mixing stage of fusion with high efficiency. J. Virol. 82:3131–3134.

11. Delos SE, Godby JA, White JM. 2005. Receptor-induced conformationalchanges in the SU subunit of the avian sarcoma/leukosis virus A envelopeprotein: implications for fusion activation. J. Virol. 79:3488 –3499.

12. Delos SE, La B, Gilmartin A, White JM. 2010. Studies of the “chainreversal regions” of the avian sarcoma/leukosis virus (ASLV) and Ebola-virus fusion proteins: analogous residues are important, and a His residueunique to EnvA affects the pH dependence of ASLV entry. J. Virol. 84:5687–5694.

13. Earp LJ, Delos SE, Netter RC, Bates P, White JM. 2003. The avianretrovirus avian sarcoma/leukosis virus subtype A reaches the lipid mixingstage of fusion at neutral pH. J. Virol. 77:3058 –3066.

14. Feng D, Tierney L. 2008. Computing and displaying isosurfaces in R. J.Stat. Softw. 28:1–24.

15. Fontana J, Cardone G, Heymann JB, Winkler DC, Steven AC. 2012.Structural changes in influenza virus at low pH characterized by cryo-electron tomography. J. Virol. 86:2919 –2929.

16. Frangakis AS, Hegerl R. 2001. Noise reduction in electron tomographicreconstructions using nonlinear anisotropic diffusion. J. Struct. Biol. 135:239 –250.

17. Furuta RA, Wild CT, Weng Y, Weiss CD. 1998. Capture of an earlyfusion-active conformation of HIV-1 gp41. Nat. Struct. Biol. 5:276 –279.

18. Gilbert JM, Hernandez LD, Balliet JW, Bates P, White JM. 1995.Receptor-induced conformational changes in the subgroup A avian leu-kosis and sarcoma virus envelope glycoprotein. J. Virol. 69:7410 –7415.

19. Grünewald K, et al. 2003. Three-dimensional structure of herpes simplexvirus from cryo-electron tomography. Science 302:1396 –1398.

20. Harrison SC. 2008. Viral membrane fusion. Nat. Struct. Mol. Biol. 15:690 – 698.

21. Hernandez LD, et al. 1997. Activation of a retroviral membrane fusionprotein: soluble receptor-induced liposome binding of the ASLV envelopeglycoprotein. J. Cell Biol. 139:1455–1464.

22. Heymann JB, Cardone G, Winkler DC, Steven AC. 2008. Computa-tional resources for cryo-electron tomography in Bsoft. J. Struct. Biol.161:232–242.

23. Jiang S, Lin K, Strick N, Neurath AR. 1993. HIV-1 inhibition by apeptide. Nature 365:113.

24. Jiang W, Baker ML, Wu Q, Bajaj C, Chiu W. 2003. Applications of abilateral deionising filter in biological electron microscopy. J. Struct. Biol.144:114 –122.

25. Jun S, et al. 2011. Direct visualization of HIV-1 with correlative live-cellmicroscopy and cryo-electron tomography. Structure 19:1573–1581.

26. Kielian M. 2006. Class II virus membrane fusion proteins. Virology 344:38 – 47.

27. Kim YH, et al. 2011. Capture and imaging of a prehairpin fusion inter-mediate of the paramyxovirus PIV5. Proc. Natl. Acad. Sci. U. S. A. 108:20992–20997.

28. Kremer JR, Mastronarde DN, McIntosh JR. 1996. Computer visualizationof three-dimensional image data using IMOD. J. Struct. Biol. 116:71–76.

29. Lee KK. 2010. Architecture of a nascent viral fusion pore. EMBO J. 29:1299 –1311.

30. Lee KK, et al. 2011. Capturing a fusion intermediate of influenza hem-agglutinin with a cholesterol-conjugated peptide, a new antiviral strategyfor influenza virus. J. Biol. Chem. 286:42141– 42149.

31. Libersou S, et al. 2010. Distinct structural rearrangements of the VSVglycoprotein drive membrane fusion. J. Cell Biol. 191:199 –210.

32. Loving R, Wu SR, Sjoberg M, Lindqvist B, Garoff H. 2012. Maturation

cleavage of the murine leukemia virus Env precursor separates the trans-membrane subunits to prime it for receptor triggering. Proc. Natl. Acad.Sci. U. S. A. 109:7735–7740.

33. Magnus C, Rusert P, Bonhoeffer S, Trkola A, Regoes RR. 2009. Esti-mating the stoichiometry of human immunodeficiency virus entry. J. Vi-rol. 83:1523–1531.

34. Mastronarde DN. 2005. Automated electron microscope tomography usingrobust prediction of specimen movements. J. Struct. Biol. 152:36–51.

35. Matsuyama S, Delos SE, White JM. 2004. Sequential roles of receptorbinding and low pH in forming prehairpin and hairpin conformations ofa retroviral envelope glycoprotein. J. Virol. 78:8201– 8209.

36. Maurer UE, Sodeik B, Grünewald K. 2008. Native 3D intermediates ofmembrane fusion in herpes simplex virus 1 entry. Proc. Natl. Acad. Sci.U. S. A. 105:10559 –10564.

37. Medinas RJ, Lambert DM, Tompkins WA. 2002. C-terminal gp40 pep-tide analogs inhibit feline immunodeficiency virus: cell fusion and virusspread. J. Virol. 76:9079 –9086.

38. Melikyan GB, Barnard RJ, Abrahamyan LG, Mothes W, Young JA.2005. Imaging individual retroviral fusion events: from hemifusion topore formation and growth. Proc. Natl. Acad. Sci. U. S. A. 102:8728 – 8733.

39. Mothes W, Boerger AL, Narayan S, Cunningham JM, Young JA. 2000.Retroviral entry mediated by receptor priming and low pH triggering of anenvelope glycoprotein. Cell 103:679 – 689.

40. Narayan S, Barnard RJ, Young JA. 2003. Two retroviral entry pathwaysdistinguished by lipid raft association of the viral receptor and differencesin viral infectivity. J. Virol. 77:1977–1983.

41. Narayan S, Young JA. 2004. Reconstitution of retroviral fusion and un-coating in a cell-free system. Proc. Natl. Acad. Sci. U. S. A. 101:7721–7726.

42. Netter RC, et al. 2004. Heptad repeat 2-based peptides inhibit aviansarcoma and leukosis virus subgroup a infection and identify a fusionintermediate. J. Virol. 78:13430 –13439.

43. Padilla-Parra S, Marin M, Kondo N, Melikyan GB. 2012. Synchronizedretrovirus fusion in cells expressing alternative receptor isoforms releasesthe viral core into distinct sub-cellular compartments. PLoS Pathog.8:e1002694. doi:10.1371/journal.ppat.1002694.

44. Puri A, Booy FP, Doms RW, White JM, Blumenthal R. 1990. Confor-mational changes and fusion activity of influenza virus hemagglutinin ofthe H2 and H3 subtypes: effects of acid pretreatment. J. Virol. 64:3824 –3832.

45. R Development Core Team. 2012. R: a language and environment forstatistical computing. R Development Core Team, Vienna, Austria.

46. Radermacher M, Wagenknecht T, Verschoor A, Frank J. 1986. A new3-D reconstruction scheme applied to the 50S ribosomal subunit of E. coli.J. Microsc. 141:RP1–RP2.

47. Renka RJ, Gebhardt A, Eglen S, Zuyev S, White D. 2009. Tripack:triangulation of irregularly spaced data. R package version 1.3-4. R Devel-opment Core Team, Vienna, Austria.

48. Russell CJ, Jardetzky TS, Lamb RA. 2001. Membrane fusion machines ofparamyxoviruses: capture of intermediates of fusion. EMBO J. 20:4024 –4034.

49. Saxton WO, Baumeister W. 1982. The correlation averaging of a regu-larly arranged bacterial cell envelope protein. J. Microsc. 127:127–138.

50. Scheres SH, Melero R, Valle M, Carazo JM. 2009. Averaging of electronsubtomograms and random conical tilt reconstructions through likeli-hood optimization. Structure 17:1563–1572.

51. Smith JG, Mothes W, Blacklow SC, Cunningham JM. 2004. The matureavian leukosis virus subgroup A envelope glycoprotein is metastable, andrefolding induced by the synergistic effects of receptor binding and low pHis coupled to infection. J. Virol. 78:1403–1410.

52. Tonelli M, Peters RJ, James TL, Agard DA. 2001. The solution structureof the viral binding domain of Tva, the cellular receptor for subgroup Aavian leukosis and sarcoma virus. FEBS Lett. 509:161–168.

53. White JM, Delos SE, Brecher M, Schornberg K. 2008. Structures andmechanisms of viral membrane fusion proteins: multiple variations on acommon theme. Crit. Rev. Biochem. Mol. Biol. 43:189 –219.

54. Wild CT, Shugars DC, Greenwell TK, McDanal CB, Matthews TJ. 1994.Peptides corresponding to a predictive alpha-helical domain of humanimmunodeficiency virus type 1 gp41 are potent inhibitors of virus infec-tion. Proc. Natl. Acad. Sci. U. S. A. 91:9770 –9774.

55. Wu SR, et al. 2008. Turning of the receptor-binding domains opens upthe murine leukaemia virus Env for membrane fusion. EMBO J. 27:2799 –2808.

Cryo-ET of ASLV Interacting with Liposomes

November 2012 Volume 86 Number 22 jvi.asm.org 12137

on February 13, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from