Abstract
Background. Transport of ebolavirus (EBOV) nucleocapsids from perinuclear viral inclusions, where they are formed, to the site of budding at the plasma membrane represents an obligatory step of virus assembly. Until now, no live-cell studies on EBOV nucleocapsid transport have been performed, and participation of host cellular factors in this process, as well as the trajectories and speed of nucleocapsid transport, remain unknown.
Methods. Live-cell imaging of EBOV-infected cells treated with different inhibitors of cellular cytoskeleton was used for the identification of cellular proteins involved in the nucleocapsid transport. EBOV nucleocapsids were visualized by expression of green fluorescent protein (GFP)–labeled nucleocapsid viral protein 30 (VP30) in EBOV-infected cells.
Results. Incorporation of the fusion protein VP30-GFP into EBOV nucleocapsids was confirmed by Western blot and indirect immunofluorescence analyses. Importantly, VP30-GFP fluorescence was readily detectable in the densely packed nucleocapsids inside perinuclear viral inclusions and in the dispersed rod-like nucleocapsids located outside of viral inclusions. Live-cell imaging of EBOV-infected cells revealed exit of single nucleocapsids from the viral inclusions and their intricate transport within the cytoplasm before budding at the plasma membrane. Nucleocapsid transport was arrested upon depolymerization of actin filaments (F-actin) and inhibition of the actin-nucleating Arp2/3 complex, and it was not altered upon depolymerization of microtubules or inhibition of N-WASP. Actin comet tails were often detected at the rear end of nucleocapsids. Marginally located nucleocapsids entered filopodia, moved inside, and budded from the tip of these thin cellular protrusions.
Conclusions. Live-cell imaging of EBOV-infected cells revealed actin-dependent long-distance transport of EBOV nucleocapsids before budding at the cell surface. These findings provide useful insights into EBOV assembly and have potential application in the development of antivirals.
Ebolaviruses (EBOVs) and Marburg viruses (MARVs) belong to the family Filoviridae and cause severe hemorrhagic fever in human and nonhuman primates, with case-fatality rates for Zaire ebolavirus reaching 90% [1]. The current EBOV disease outbreak in West Africa, with case numbers exceeding 15 000 as of November 2014, demonstrates the necessity to develop vaccines and antivirals against filoviruses [2]. At present, no approved specific therapy is available [3–5], and therefore further understanding and characterization of the different steps of filovirus life cycle will be effectual in the development of novel therapeutic approaches.
EBOVs and MARVs have a nonsegmented single-stranded RNA genome, with a nucleotide sequence identity of 35% [6–9]. The genome of filoviruses is encapsidated with nucleoprotein (NP), the RNA-dependent RNA polymerase (L), viral polymerase cofactor viral protein (VP35), and also the viral proteins VP30 and VP24 [10–12]. These ribonucleoprotein complexes (RNPs), called “nucleocapsids,” have a rod-like filamentous shape and are about 50 nm in diameter and about 1000 nm in length [11, 13]. Formation of nucleocapsids occurs in the viral inclusions, which are located in the perinuclear region and represent sites of virus replication [14]. For the completion of the filovirus replication cycle, nucleocapsids need to be redistributed from the perinuclear region to the plasma membrane, where they interact with the viral matrix protein VP40 and the membrane viral surface glycoprotein. Virions are released by budding through the plasma membrane [15–18]. It is currently considered that redistribution of big viral RNPs cannot be provided by random motion (diffusion) within the cytoplasm but needs controlled active transport directed to the site of budding [19]. Indeed, a recent live-cell imaging study of MARV-infected cells revealed long-distance actin-dependent transport of MARV nucleocapsids from the viral inclusions to the plasma membrane [20]. What kind of cellular cytoskeleton is important for EBOV nucleocapsid transport remains unknown.
To elucidate EBOV nucleocapsid transport, we performed live-cell imaging analysis of EBOV-infected cells cotransfected with VP30-GFP fusion protein. This approach allowed visualization of the exit of single nucleocapsids from inclusion bodies. We found that nucleocapsids are transported within the cytoplasm over a long distance before budding. The mean speed (±standard deviation [SD]) of EBOV nucleocapsids was 250 ± 67 nm/second. Application of specific cytoskeleton inhibitors revealed that transport of EBOV nucleocapsids was dependent on actin filaments and the activity of actin nucleation factor Arp2/3. However, EBOV nucleocapsids continued to move in cells treated with microtubule depolymerizing agent or inhibitor of N-WASP. When nucleocapsids reach cell margins, they enter a single filopodium, move inside, and use this protrusion for the release. We consider that identification of cellular factors important for the EBOV nucleocapsid transport might contribute to the development of specific antiviral therapies.
METHODS
Cell Culture and Virus
Huh-7 (human hepatoma) cells were cultured in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal bovine serum (FBS; Pan Biotech), 5 mM of glutamine (Q; Life Technologies), and 50 U/mL each of penicillin and streptomycin (P/S; Life Technologies) at 37°C and 5% CO2. The recombinant virus used in this study was based on EBOV Zaire (strain Mayinga; GenBank accession number AF272001). Cloning and rescue of full-length EBOV was performed as described earlier [21, 22].
Infection of Cells and Live-Cell Microscopy
Huh-7 cells were infected with recombinant EBOV at a multiplicity of infection of 0.1%–1% of a 50% tissue culture infective dose in DMEM/P/S/Q without FBS for 1 hour at 37°C. After infection the inoculum was removed, the cells were washed once with phosphate-buffered saline, and appropriate cell-culture medium was added. For live-cell imaging, Huh-7 cells were seeded onto 35-mm µ-dishes (Ibidi) 24 hours before infection. Cells were infected in 400 µL of Opti-MEM without phenol red (Life Technologies) for 1 hour, the inoculum was removed, and, if needed, the cells were transfected according to the manufacturer's instructions (Mirus) with 1 µg of DNA in a 500-µL final volume of CO2-independent Leibovitz's medium (Life Technologies). Live-cell time-lapse experiments were recorded with a Leica DMI6000B device, using a 63× oil objective. Details are described elsewhere [20]. All work with infectious EBOV was performed under biosafety level 4 conditions at the Institute of Virology, Philipps University of Marburg, in Marburg, Germany.
Cloning of VP30-GFP
For the construction of the plasmid pCAGGS-VP30-GFP coding for the VP30-GFP fusion protein, the EBOV VP30 open reading frame was cloned in frame to the 5′ end of the gene encoding GFP, using primer-extension polymerase chain reaction analysis and the restriction enzymes EcoRI HF and AgeI HF (New England Biolabs).
Preparation of Whole-Mounted Cells for Electron Microscopy
This procedure was performed as described previously [23].
Sodium Dodecyl Sulfate Polyacrylamide Gel Electrophoresis (SDS/PAGE), Immunoblot Analysis, Immunofluorescence Analysis, and Antibodies
These methods were performed as described previously [20, 24]. Nuclei staining was performed using 4′,6-diamidino-2-phenylindole, and actin filament staining was performed with phalloidin–Alexa Fluor 594 (Life Technologies). For the detection of VP30 by Western blot or immunofluorescence, a specific rabbit anti-VP30 antibody was used. For detection of EBOV proteins by immunofluorescence microscopy, goat anti-EBOV serum was used. For the detection of NP, a monoclonal anti-NP antibody was used. As secondary antibodies for Western blot analysis, goat anti-rabbit–Alexa Fluor 680 (LI-COR Biosciences) were used. For immunofluorescence analysis, donkey anti-goat–Cy5 (Life Technologies), goat anti-rabbit–Alexa Fluor 488, and goat anti-mouse–Alexa Fluor 647 were used. Immunofluorescence samples were acquired on a DMI6000B TCS SP5 laser scanning microscope, using a 63× oil objective (Leica Microsystems), a 488-nm argon laser, a DPSS 561-nm laser, or a helium 633-nm laser.
Treatment of Cells with Cytoskeleton-Modulating Drugs
EBOV-infected cells were treated 24 hours after infection with 15 µM nocodazole (Sigma), 0.3 µM cytochalasin D (Sigma), 100 µM CK869 (Sigma), 5 µM wiskostatin (Sigma), or 0.15% dimethyl sulfoxide (Sigma). Chemicals were added to the cell culture medium 3 hours before fluorescence microscopy was performed. For each tested inhibitor, >10 different cells with >100 nucleocapsids inside were compared with control cells.
RESULTS AND DISCUSSION
Establishment of a Live-Cell Imaging Assay of Nucleocapsid Transport in EBOV-Infected Cells
To visualize EBOV nucleocapsids, we constructed a plasmid encoding VP30-GFP. To test whether VP30-GFP is incorporated into EBOV nucleocapsids, we infected cells with recombinant EBOV and transfected them 1 hour after infection with an expression plasmid encoding VP30-GFP. Western blot analysis of cell lysates and supernatants detected VP30-GFP in the cell lysate and in released virus particles (Figure 1A). Expression of VP30-GFP in the EBOV-infected cells resulted in a readily detectable fluorescence signal (Figure 1B). The distribution pattern of VP30-GFP was very similar to the pattern for nucleocapsid proteins. Confocal microscopy of cells stained with anti-EBOV antibody confirmed complete colocalization of VP30-GFP, NP-induced inclusions, and individual nucleocapsids (Figure 1B), supporting efficient incorporation of VP30-GFP into nucleocapsids. Live-cell imaging analysis of EBOV-infected cells transiently expressing VP30-GFP was performed 18 hours after infection, when budding of virions had already started. Remarkably, individual VP30-GFP–positive nucleocapsids were recognizable both inside and outside of viral inclusions, although a high density of nucleocapsid packing inside viral inclusions often resulted in bright merged signals (Figure 1C). The mean length (±SD) of VP30-GFP–positive nucleocapsids was 1050 ± 70 nm (n = 20), which is comparable with the length of EBOV nucleocapsids measured by other methods [11]. We observed that the size of the inclusion bodies and the intensity of the VP30-GFP signal inside inclusions increased over time. Furthermore, time-lapse microscopy showed that exit of single nucleocapsids from viral inclusions occurred within few seconds (Figure 1D and Supplementary Movie 1). Our data indicated that the addition of a 26-kDa GFP tag to the C-terminus of VP30 did not substantially affect expression and incorporation of the protein into virions. This finding is consistent with a similar approach for the tracing of nucleocapsids of MARV, a close relative of EBOV, in which transiently expressed MARV VP30-GFP was integrated into virus particles without affecting virus growth kinetics [20, 25]. In addition, previous studies demonstrated that transiently expressed EBOV VP30, which was integrated into the viral RNP complexes, supported EBOV transcription [26, 27]. These results showed that our use of fluorescently labeled VP30 is well suited for tracing EBOV nucleocapsids in infected cells.
Analysis of the suitability of nucleocapsid VP30-GFP fusion protein as a fluorescent tag for ebolavirus (EBOV) nucleocapsids. A–D, Huh-7 cells were infected with EBOV and transfected with a plasmid encoding VP30-GFP fusion protein 1 hour after infection. The time and method of analysis are indicated below. A, Western blot analysis of cell lysate (lane 1) and supernatant (lane 2) harvested 48 hours after infection was performed using antibody directed toward VP30. B, Confocal microscopy of cells fixed 24 hour after infection. The enlarged left panel shows viral inclusion stained with α-EBOV serum from goat, followed by donkey α-goat antibody coupled to Cy5. The enlarged right panel shows autofluorescence of VP30-GFP. C and D, Time-lapse microscopy of cells 19 hours after infection. The movie was acquired at 1 frame every 20 seconds. C, The left panel shows one frame of a time-lapse series. The middle panel shows a schematic drawing of the inclusion body shape, with rod-like nucleocapsid structures inside the inclusion indicated by black dotted lines and nucleocapsids located outside of the inclusion indicated by white rods. The right panel shows merged image of the scheme and fluorescent image. D, Several frames of a time-lapse series. The time in minutes and seconds is displayed in the upper left corner. A nucleocapsid detached from the viral inclusion is indicated by a thin white arrow. Sites of nucleocapsid exit from viral inclusion are indicated by black arrows. Abbreviations: MW, molecular weight; rEBOVwt, recombinant Ebolavirus wildtype.
EBOV Nucleocapsids Detached From Inclusion Bodies Are Transported Across the Cytoplasm Over Long Distances
Twenty-four hours after infection, EBOV-infected cells contained several viral inclusions and hundreds of individually distributed nucleocapsids (Figure 2A). To analyze the transport route of EBOV nucleocapsids, we used time-lapse microscopy of VP30-GFP–tagged nucleocapsids, which displayed different patterns of movement. We observed long-distance movement spanning >14 μm within a minute without stop (Figure 2B). In addition, nucleocapsids were detected that underwent stop-and-go movement with pauses lasting for several seconds and even minutes (Supplementary Movie 2). Remarkably, none of the monitored EBOV nucleocapsids used a direct path from the viral inclusions to the plasma membrane. In contrast, nucleocapsids had intricate routes that were highly variable in terms of direction and track length. Likewise, velocities of nucleocapsid movement varied from 100 nm/s to almost 400 nm/s with a mean speed (±SD) of 250 ± 67 nm/second (Figure 2C). The transport of EBOV nucleocapsids was thus considerably slower than MARV nucleocapsids [20]. While this difference is currently not understood, it might be that the larger size of the EBOV nucleocapsids, compared with MARV nucleocapsids (1000 nm vs 800 nm), contribute to the reduced velocity. Alternatively, differences in velocity might be caused by the use of different, unidentified motor proteins.
Characterization of ebolavirus (EBOV) nucleocapsid dynamics. A, Huh-7 cells were infected with EBOV and transfected with the plasmids encoding nucleocapsid VP30-GFP fusion protein and TagRFP-actin 1 hour after infection. One frame of time-lapse microscopy was acquired 24 hours after infection. B, Representative trajectories of EBOV nucleocapsids (n = 5) moving without stopping. Images were recorded within 60 seconds. C, Nucleocapsid speeds (n = 27) tracked in different infected cells during 4 independent experiments.
Movement of EBOV nucleocapsids was generally unidirectional with a defined leading tip and a rear end. The nucleocapsids never switched to a reverse gear with the rear end taking the lead. In addition, lateral movement was not observed. The molecular mechanisms supporting the polarity of the nucleocapsids are unclear.
Interestingly, we frequently detected the simultaneous movement of 2 or even 3 nucleocapsids, which resembled a piggyback mode (Supplementary Movie 3). The velocity of these cotransported nucleocapsids was in range of the average speed values. Three nucleocapsids were cotransported with speeds of 247 nm/second, 2 were cotransported with speeds of 234 nm/second, and interestingly 2 individually transported nucleocapsids had slower speeds (71 nm/s and 78 nm/s). Regarding observed speed values of cotransported nucleocapsids, one might assume that a dual or even triple cotransport event is of synergistic value for nucleocapsids. Owing to resolution limits, our technique could not resolve the spatial arrangement of structures closer than 300 nm. Therefore, acquisition with an improved resolution would help to understand this process in more detail.
Movement of Nucleocapsids Is Dependent on F-actin and Actin-Nucleation Activity of the Arp2/3 Complex
We then determined the components of the cellular cytoskeleton that are involved in transport of nucleocapsids. Using live-cell imaging, we analyzed the impact of several drugs affecting components of the cytoskeleton on the nucleocapsid movement. Incubation of EBOV-infected cells with the microtubule-depolymerizing drug nocodazole for 3 hours had no influence on nucleocapsid movement (Figure 3A). In contrast, incubation of EBOV-infected cells with the F-actin–depolymerizing drug cytochalasin D resulted in the immediate cessation of nucleocapsid movement (Figure 3A). We further analyzed the relevance of the actin-nucleating Arp2/3 complex and the actin-nucleation-promoting factor N-WASP for the transport of nucleocapsids. Incubation of EBOV-infected cells with the compound CK869, which inhibited the activity of the Arp2/3 complex, impaired nucleocapsid movement [28] (Figure 3A). However, incubation of EBOV cells with wiskostatin, an inhibitor of N-WASP [29], for at least 3 hours did not affect nucleocapsid transport (Figure 3A).
Live-cell imaging analyses of ebolavirus (EBOV) nucleocapsid transport in cells incubated with different cytoskeletal inhibitors. A, Huh-7 cells were infected with EBOV and transfected with a plasmid encoding nucleocapsid VP30-GFP fusion protein 1 hour after infection. Different drugs, such as 0.15 µM nocodazole (NOC), 0.3 µM cytochalasin D (CD), 100 µM CK869, 5 µM wiskostatin (WISK), or dimethyl sulfoxide (DMSO) as a control, were added to the culture medium 24 hours after infection, and time-lapse microscopy was performed. Upper panels show a representative maximal intensity projection of 30 frames, corresponding to 60 seconds and a 40 µm2 area of the cell. Lower panels show enlarged images of upper panels. B, Huh-7 cells were infected with EBOV and transfected with the plasmids encoding VP30-GFP fusion protein and TagRFP-actin 1 hour after infection. Cells were analyzed by time-lapse microscopy 24 hours after infection. The acquisition time in seconds is displayed in the upper left corner.
We also investigated whether nucleocapsid movement was associated with actin-based propulsion, using EBOV-infected cells expressing VP30-GFP and a fluorescently labeled form of actin (TagRFP-actin). Live-cell imaging analyses revealed actin comet tails at the rear end of nucleocapsids, suggesting a direct contribution of actin polymerization in driving the transport of the nucleocapsids (Figure 3B and Supplementary Movie 4).
Certain viruses and prokaryotes stimulate the actin-nucleation activity of the Arp2/3 complex by mimicking the actin nucleation-promoting factor N-WASP. For instance, RickA of Rickettsia species, ActA of Listeria species, and p78/83 of baculovirus can mimic the function of N-WASP–like proteins [30]. In our study, inhibition of N-WASP failed to stop nucleocapsid movement. It is therefore presumed that EBOV proteins directly act as N-WASP–like proteins promoting the actin nucleation or that cellular-nucleation-promoting proteins are recruited by viral proteins to activate the Arp2/3 complex [31, 32]. Stimulation of the Arp2/3 complex by WASP-like proteins is mediated by acidic domains [30, 33], which can also be detected in EBOV and MARV proteins. For example, NP or L sequences show acidic DD or WE/D motifs. Further studies will show whether these viral proteins mediate the stimulation of the ARP2/3 complex.
Nucleocapsids Move Inside and Bud From the Tips of Filopodia
Live-cell imaging analysis of EBOV nucleocapsids at the cell margins showed entry of a single nucleocapsid into a filopodium and its movement inside this thin cellular protrusion (Figure 4A and Supplementary Movie 5). Interestingly, inside one filopodium, we mostly detected several nucleocapsids, which compromised determination of transport velocity. In a single case in which the velocity of an individual nucleocapsid could be determined unequivocally, it was 18 nm/second. Finally, fission of virus particles took place at the tips or sides of filopodia (Figure 4B). Electron microscopy of EBOV-infected cells, using the whole-mount method, also detected close association between thin cellular protrusions and nucleocapsids located at the very margins of the infected cell (Figure 4C). This finding supported the idea that filopodia are used for the egress of EBOV. The detection of EBOV nucleocapsids in association with long filopodial protrusions is consistent with previous studies showing release of MARV from the side or tip of filopodia [11, 13, 34]. The use of filopodia for efficient egress and spread to neighboring cells has also been described for retroviruses [35, 36]. It is presumed that EBOV takes advantage of filopodia as the site of release because these delicate cellular protrusions often make connections to neighboring cells, thus facilitating access to the next target cell. The release of budding virions directly to the target cells minimizes the appearance of virions in the extracellular space, which avoids recognition by the hostile environment (eg, the immune system and host proteases) [20].
Analysis of ebolavirus (EBOV) nucleocapsids at the cell margins. A and B, Huh-7 cells were infected with EBOV and transfected with the plasmids encoding nucleocapsid VP30-GFP fusion protein and TagRFP-actin 1 hour after infection. Cells were analyzed by time-lapse microscopy 24 hours after infection. Left images show overview of cells. Higher magnification of framed areas is shown in right panels. Acquisition time in seconds is displayed in the upper left corner. A, Movement of nucleocapsid (white arrow) inside a filopodium. B, Movement of several nucleocapsids inside 1 filopodium and fission of virus particle (white arrow). C, Huh-7 cells were infected with EBOV, fixed 22 hours after infection, and underwent whole-mount electron microscopy. The framed area in the left panel is shown at a higher magnification in the right panel.
We found that the speed of nucleocapsid movement in EBOV-infected cells varied considerably, which can be explained by the use of different actin-dependent motor proteins, such as myosins, but also by the accessibility of actin-associated proteins necessary for polymerization. The slowest speed of nucleocapsid transport was detected inside filopodia, where intrafilopodial transport can be supported by myosin X [20, 37].
Supplementary Data
Supplementary materials are available at The Journal of Infectious Diseases online (http://jid.oxfordjournals.org). Supplementary materials consist of data provided by the author that are published to benefit the reader. The posted materials are not copyedited. The contents of all supplementary data are the sole responsibility of the authors. Questions or messages regarding errors should be addressed to the author.
Notes
Acknowledgments. We thank Katharina Kowalski and Dirk Becker, for excellent technical assistance; Dr Markus Eickmann, for supervision of, and Gotthard Ludwig and Michael Schmidt, for technical support with, biosafety level 4 work; and Dr Allison Groseth and Dr Thomas Hoenen, for providing ebolavirus constructs.
Financial support. This work was funded by Deutsche Forschungsgemeinschaft and Collaborative Research Centers 593 and 1021.
Potential conflicts of interest. All authors: No reported conflicts.
All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.
References
