The Large Marseillevirus Explores Different Entry Pathways by Forming Giant Infectious Vesicles

ABSTRACT

Triggering the amoebal phagocytosis process is a sine qua non condition for most giant viruses to initiate their replication cycle and consequently to promote their progeny formation. It is well known that the amoebal phagocytosis process requires the recognition of particles of >500 nm, and most amoebal giant viruses meet this requirement, such as mimivirus, pandoravirus, pithovirus, and mollivirus. However, in the context of the discovery of amoebal giant viruses in the last decade, Marseillevirus marseillevirus (MsV) has drawn our attention, because despite its ability to successfully replicate in Acanthamoeba, remarkably it does not fulfill the >500-nm condition, since it presents an ∼250-nm icosahedrally shaped capsid. We deeply investigated the MsV cycle by using a set of methods, including virological, molecular, and microscopic (immunofluorescence, scanning electron microscopy, and transmission electron microscopy) assays. Our results revealed that MsV is able to form giant vesicles containing dozens to thousands of viral particles wrapped by membranes derived from amoebal endoplasmic reticulum. Remarkably, our results strongly suggested that these giant vesicles are able to stimulate amoebal phagocytosis and to trigger the MsV replication cycle by an acidification-independent process. Also, we observed that MsV entry may occur by the phagocytosis of grouped particles (without surrounding membranes) and by an endosome-stimulated pathway triggered by single particles. Taken together, not only do our data deeply describe the main features of MsV replication cycle, but this is the first time, to our knowledge, that the formation of giant infective vesicles related to a DNA virus has been described.

IMPORTANCE Triggering the amoebal phagocytosis process is a sine qua non condition required by most giant viruses to initiate their replication cycle. This process requires the recognition of particles of >500 nm, and many giant viruses meet this requirement. However, MsV is unusual, as despite having particles of ∼250 nm it is able to replicate in Acanthamoeba. Our results revealed that MsV is able to form giant vesicles, containing dozens to thousands of viral particles, wrapped in membranes derived from amoebal endoplasmic reticulum. Remarkably, our results strongly suggest that these giant vesicles are able to stimulate phagocytosis using an acidification-independent process. Our work not only describes the main features of the MsV replication cycle but also describes, for the first time to our knowledge, the formation of huge infective vesicles in a large DNA viruses.

INTRODUCTION

The discovery of the amoebal giant virus Acanthamoeba polyphaga mimivirus (APMV) in 2003 (1) raised new and exciting questions regarding the virosphere and boosted the hunt for new giant viruses. Owing to these efforts, an increasing number of remarkable giant viruses have been described (2–6).

Recent data suggest that giant viruses initiate their replication cycles after being phagocytosed by amoebas or other phagocytic cells (1, 7). This conclusion is well supported by classical studies which show that the phagocytosis process is triggered in Acanthamoeba by particles of >500 nm (8). Therefore, most of the giant viruses, such as mimivirus, pandoravirus, pithovirus, and mollivirus, meet this requirement (1–3, 5). Marseillevirus marseillevirus (MsV) particles, on the other hand, are formed by 250-nm icosahedral capsids that do not reach the 500-nm size threshold (6). Nonetheless, MsV still is able to successfully replicate in Acanthamoeba, raising the question of how the virus enters its host cell. After entry, a large and diffuse viral factory is assembled (2 to 4 h), wherein genome replication and virion morphogenesis occurs (6 to 8 h), and the virions are released from the cell within 24 h. Just like the mimivirus, MsV has fibers with globular ends on the surface, although they are shorter in length (∼12 nm compared to ∼125 nm of mimiviruses), and both have an internal membrane surrounding the nucleocapsid (6). However, the origins of the MsV inner membrane remain unknown (6). The first marseillevirus was isolated from a water sample collected from a cooling tower in Paris, France, by culturing Acanthamoeba castellanii cells. Since then, marseillevirus-like organisms have been isolated from several environmental samples, with representatives from France, Tunisia, Senegal, Australia, and Brazil (6, 9–13).

Although these viruses were discovered almost a decade ago, little is known regarding their replication cycle. Thus, our objective was to perform a detailed study on the steps of the MsV replication cycle, specifically focusing on two fundamental questions: does MsV enter its amoeba host by phagocytosis, and if so, how does it stimulate the phagocytosis process while being smaller than 500 nm in size? Interestingly, our findings reveal that MsV is able to form giant vesicles containing dozens to thousands of viral particles. Using several approaches, it was possible to track the origins of these vesicles, as well as their important role in MsV biology and the maintenance of these viruses in the environment. Our results strongly suggest that these giant vesicles are remarkably able to stimulate amoeba phagocytosis by an acidification-independent process, successfully initiating the MsV replication cycle. Also, we observed that MsV entry may occur by phagocytosis of grouped particles (without surrounding membranes) or by an endosome-mediated pathway triggered by single particles. Taken together, our data not only thoroughly describe the main features of the MsV replication cycle but also provide the first report, to our knowledge, of the formation of giant infective vesicles as a way to boost the replicative success of a large DNA virus.

MATERIALS AND METHODS

Virus production, purification, and replication cycle analysis.

Acanthamoeba castellanii cells (ATCC 30010) were cultivated in PYG medium supplemented with 7% fetal calf serum (FCS; Cultilab, Brazil), 25 mg/ml amphotericin B (Fungizone; Cristalia, São Paulo, Brazil), 500 U/ml penicillin, and 50 mg/ml gentamicin (Schering-Plough, Brazil). A total of 7 × 10⁶ cells were infected with MsV at a multiplicity of infection (MOI) of 0.01 and incubated at 32°C. After the appearance of cytopathic effect, the cells and supernatant were collected and the viruses were then purified through ultracentrifugation with a 25% sucrose cushion at 36,000 × g for 2 h. After purification, the amoeba cells were infected with purified viruses at MOIs of 0.01 and 10 to evaluate the MsV replication cycle. Infected cells were fixed with 2.5% glutaraldehyde in a 0.1 M sodium phosphate buffer for 1 h at room temperature at various times postinfection. The amoebas were postfixed with 2% osmium tetroxide and embedded in EPON resin. Ultrathin sections then were analyzed under transmission electron microscopy (TEM; Spirit Biotwin FEI-120 kV).

Purification of vesicles.

For the purification of the vesicles, the supernatant obtained from A. castellanii cells infected with MsV at an MOI of 0.01 was submitted to ultracentrifugation at 20,000 × g for 5 min. The pellet containing vesicles was collected and the supernatant containing only naked particles was discarded. After this purification, the samples were prepared for analysis by scanning electron microscopy (SEM), as previously described (14). Furthermore, we also evaluated if these vesicles could be phagocytosed by amoebas. This analysis was conducted under TEM. For the infection, 1 × 10⁶ 50% tissue culture infective doses (TCID₅₀) (15) of purified vesicles were added to a culture flask containing 7 × 10⁶ cells of A. castellanii. One hour postadsorption the cells were fixed with 2.5% glutaraldehyde in a sodium phosphate buffer, 0.1 M, and analyzed by TEM as described above.

Entry assays.

In this set of experiments, we evaluated different entry pathways that MsV explores in order to infect amoeba cells. First, we evaluated the endocytic pathway using chloroquine treatment. For this experiment, A. castellanii cells were infected with single particles of MsV or vesicles at an MOI of 10, plus chloroquine (Sigma) at a concentration of 50 μg/ml. Twenty-four hours postinfection, the treated and untreated cells were titrated using TCID₅₀ methods. In addition, we evaluated whether the inhibition of actin filaments and microtubules (related to phagocytosis) could decrease the viral titer. For this, A. castellanii cells were infected with particles of the MsV or vesicles at an MOI of 10 and further treated with cytochalasin D (Sigma) at a concentration of 2 μM. Twenty-four hours postinfection, the treated and untreated cells were titrated by TCID₅₀ methods.

In order to verify the impact of membrane transportation on the viral replication cycle, A. castellanii cells were infected with MsV at an MOI of 10 and treated with 10 μg/ml of brefeldin A (BFA). We observed four different infection periods, 1, 2, 4, and 8 h, that correspond to four different stages of the MsV replication cycle, namely, entry, membrane recruitment, early viral factory formation, and mature viral factories. The amoebas were then transferred to 96-well microplates containing 100 μl of PYG medium and maintained at 32°C for 24 h before being further titrated by TCID₅₀.

Biological properties of the vesicles.

A. castellanii cells were infected with MsV at an MOI of 0.01. Forty-eight hours postinfection, the vesicles were purified as described above. A portion of the purified vesicles was treated with a lysis buffer at a ratio of 1:1 for 10 min for the release of viral particles present in the vesicles. The vesicles containing the viruses, and also the single particles, then were titrated using the TCID₅₀ method. After titration, 1 × 10⁵ TCID₅₀ of each was used to infect new A. castellanii cells. The cells were observed daily, and 100 μl was collected after 24, 48, and 72 h. In addition, we evaluated the temperature resistance of vesicles and single particles. For this, we used 1 × 10⁵ TCID₅₀ vesicles or single particles and submitted them to 70°C for 1, 3, 5, 7, and 10 min using an Eppendorf Thermomixer comfort apparatus. Seventy degrees was chosen due to previous data from our laboratory that indicated this temperature is a good choice for performing heat resistance assays of large/giant viruses. It is important to highlight that at room temperature (25°C), both MsV naked particles and vesicles remain infective for at least 2 months (data not published). The samples then were titrated in A. castellanii using TCID₅₀.

The origins of the internal membrane and the giant vesicles.

To investigate the origins of the internal membrane and the giant vesicles formed by marseilleviruses, immunofluorescence microscopy and an immunoblotting assay were performed.

(i) Immunofluorescence microscopy.

A. castellanii cells were grown on coverslips and infected with MsV (MOI of 10) for 1.5 and 4 h for endosome, endoplasmic reticulum (ER), and Golgi analysis. After infection, the cells were rinsed in cold phosphate-buffered saline (PBS) and fixed with 4% paraformaldehyde (PFA) in PBS for 10 min at room temperature (RT). After fixation, the cells were permeabilized with 0.2% Triton X-100 in 3% bovine serum albumin (BSA)–PBS for 5 min, followed by a rinse with 3% BSA–PBS three times. The cells then were stained for 1 h at room temperature with four specific primary antibodies: rabbit polyclonal anti-sorting nexin 2 (SNX2) (Santa Cruz Technology, USA) (16), goat polyclonal anti-GRP 78 (N-20) (Santa Cruz Technology, USA), mouse monoclonal anti-GM130 (BD Biosciences, USA), and mouse polyclonal anti-MsV (17). After incubation with secondary antibodies, fluorescently labeled cells were visualized using a Zeiss (LSM 510 META) microscope. The images were processed with LSM Image Browser and Adobe Photoshop (version 7.0) software.

(ii) Immunoblotting assay.

Briefly, for protein extraction, electrophoresis, and immunoblotting of the infected cells, vesicles, and purified single particles (described above), the samples were disrupted on ice with a lysis buffer and centrifuged at 18,000 × g for 15 min at 4°C. The protein concentration was determined using Bio-Rad protein assay dye reagent concentrate number 5000006. Thirty micrograms of protein per sample was separated by electrophoresis on a 10% SDS polyacrylamide gel and transferred to polyvinylidene difluoride membranes. The membranes were incubated overnight at 4°C with anti-sorting nexin 2 (1:1,000), anti-GRP 78 (1:1,000), and anti-GM130 (1:1,000) antibodies. After washing, the membranes were incubated with peroxidase-conjugated anti-rabbit (1:5,000) and anti-goat (1:5,000) secondary antibodies. Immunoreactive bands were visualized using a Luminata Forte Western horseradish peroxidase (HRP) substrate.

RESULTS

MsV viral factory observations.

It was possible to clarify the various stages of the MsV replication cycle, which had so far remained poorly understood, through observations of A. castellanii cells at different times postinfection using transmission electron microscopy (TEM). Around 3 to 4 h postinfection we observed large viral factories, occupying 1/3 to 1/2 of the amoeba cytoplasm, all containing MsV particles in different stages of morphogenesis (Fig. 1A). At the edge of the viral factories it was possible to observe membrane-rich areas (MRA) (Fig. 1A) in which most of the mature viral particles were wrapped into large vesicles. These vesicles varied widely in terms of size and number of membranes (Fig. 1B, C, and D). By means of TEM, we also could observe some cells undergoing lysis, releasing a number of MsV vesicles (Fig. 1E). We noticed that the vesicles could measure from 300 nm up to >1,000 nm in diameter, depending on the amount of virus found within them. Most vesicles had an approximate size of 300 to 500 nm, representing 45% of the produced vesicles (Fig. 1F). Interestingly, we observed that the vesicles consisted of various numbers of membranes. By performing an analysis of infected amoebas containing vesicles, we evaluated the number of membranes present in each one. There were vesicles with only one membrane but also vesicles with three or more membranes. Vesicles with two membranes were the most prevalent, making up around 40% of the total (Fig. 1G).

FIG 1.

MsV forms a large viral factory and produces giant vesicles. (A) A TEM image of an A. castellanii cell infected with MsV exhibiting a large viral factory surrounded by membranes and vesicles (asterisks), with several viral particles, both inside and outside, often being found in vesicles. The dotted red line surrounds the viral factory. (B to D) TEM images demonstrating that the vesicles may contain one or more membranes and variable numbers of viruses. (E) A TEM image demonstrating a cell undergoing lysis, releasing a number of MsV vesicles. (F) Analyses of the size of the vesicles, in nanometers, found in the cytoplasm of an A. castellanii cell. (G) The distribution of the number of membranes present in the vesicles. For panels F and G, a total of 200 circular vesicles were considered. *, viral particles in vesicles; VF, viral factory; MRA, membrane-rich area; N, nucleus. All of the images were obtained 24 h postinfection at low MOI (0.01).

The origins of vesicle membranes and their impact on viral replication.

During the replication of MsV, we observed the recruitment and rearranging of cell membranes toward vesicle formation. In addition, we noted that the vesicles could consist of one or several membranes. In order to investigate the origins of these membranes, we performed immunofluorescence (IF) assays targeting three distinct parts of cellular membranes, namely, endoplasmic reticulum (ER; anti-GRP78), Golgi complex (anti-GM130), and endosome (anti-SNX2), as these structures have been described as being involved in the morphogenesis of a number of other viruses. We also used an anti-MsV polyclonal antibody. A. castellanii cells were infected with MsV. Three hours postinfection we could observe colocalization between the anti-GRP78 and anti-MsV targets, suggesting that the membranes of the vesicles originated from the ER (Fig. 2A). The colocalization between MsV vesicles and other analyzed cellular targets was not observed (data not shown). Only structures outside the particles were targeted, as we believe that the immunofluorescence permeabilization process is not efficient enough to allow the binding of antibodies to membranes inside the viral particle (to answer this question, we performed immunoblotting assays of purified particles as described below). It is also clear from Fig. 2A that vesicle formation is polarized within the cell, coinciding with the polarization of ER membranes at this specific moment of infection. The formation of ER membrane-rich areas during MsV infection could be further visualized by TEM (Fig. 2B).

FIG 2.

Membranes of MsV giant vesicles likely originate from ER. (A) Immunofluorescence microscopy showing the colocalization between anti-GRP78 and anti-MsV targets, suggesting that the membranes of the vesicles originate from the ER. (B) A TEM image of an A. castellanii cell 3 h postinfection, demonstrating an ER-rich area. (C) The impact of BFA on viral replication. After infected A. castellanii cells were treated with BFA, an ∼2 log reduction in the viral titer was observed relative to untreated cells. The experiment was performed twice in duplicate. (D) TEM image of a cell not treated with brefeldin A, evidencing many viral particles within the viral factory. (E) TEM image of brefeldin A-treated cells (3 h of treatment) exhibiting a decrease in or absence of membranes around the viral factories as well as a significant reduction in the number of particles being assembled. **, P < 0.001; VF, viral factory.

After we confirmed the origins of the vesicle membranes, we evaluated the influence of BFA, a membrane transport inhibitor, during the MsV replication cycle and vesicle formation. A decrease of ∼2 logs in viral titer was observed in treated cells compared to untreated cells (Fig. 2C). In addition, using TEM images, we observed that treated cells exhibited a decrease/absence of membranes around the viral factories, as well as a significant reduction in the number of particles being assembled compared to untreated cells (Fig. 2D), suggesting that membranes are important not only to vesicle formation but also to particle morphogenesis (Fig. 2E).

Vesicle attachment and entry into amoebas.

After establishing the origin of the vesicles, we then evaluated whether MsV vesicles could attach to and be phagocytosed by amoebas. Initially, to demonstrate that the viruses could be released within vesicles, we performed a scanning electron microscopy (SEM) analysis of these structures after they were purified from amoebas infected with MsV. The SEM images corroborated the existence of the extracellular vesicles (Fig. 3A). In addition, the results also demonstrated the efficacy of the purification method used. Considering that MsV particles are icosahedral and that most of the vesicles are spherical, we estimated the number of virus particles present within the vesicles, as they varied in size (300 to >1,000 nm in diameter). Using a mathematical model, we sought to obtain the most probable number of viral particles present in each vesicle. A logarithmic profile was achieved, and it was possible to estimate that >1,000 viral particles could exist inside a vesicle (Fig. 3B). A vesicle 1,100 nm in diameter hypothetically could contain ∼100 MsV particles, for example (Fig. 3B).

FIG 3.

Giant vesicles contain several viral particles and promote phagocytosis. (A) SEM images of purified giant vesicles produced by MsV during its replication cycle. (Top right) Details of some particles close to a small vesicle. (B) A putative amount of viral particles inside vesicles of different sizes. Using a mathematical model, the number of particles per vesicle with diameters ranging from 300 to 3,500 nm was estimated. (C to E) Vesicles of different sizes attached to the surface of an A. castellanii cell. (F) A TEM image of intact giant vesicles inside an A. castellanii cell directly after penetration. The vesicles had more than one membrane. The external layer fused with the cell membrane, releasing the viruses surrounded by the inner membrane of the vesicles. (G) A TEM image of several dispersed viral particles inside an A. castellanii cell directly after penetration of a giant vesicle. In contrast to results shown in panel F, the giant vesicle likely had only one membrane that fused with the cell membrane, releasing the viruses directly into the cytoplasm of the cell. An asterisk indicates vesicles attached to amoebal surface; in panels C and E it is possible to see amoebal pseudopods attached to the vesicles. Both TEM images were obtained 1 h postinfection with the large vesicles.

In order to verify that the vesicles were able to attach to the host surface and stimulate phagocytosis, purified vesicles were incubated with A. castellanii cells for 10 min and further analyzed by SEM. The images demonstrated that an interaction between vesicles of different sizes and amoebae occurred (Fig. 3C to E). TEM analysis of the vesicles after 30 min of incubation with A. castellanii cells also was performed. It was possible to observe intact vesicles and many viruses spread throughout the cell cytoplasm (Fig. 3F and G). We believe that when a vesicle has only one membrane, it merges with the phagosome membrane and releases the viruses inside the cytoplasm of the amoeba (Fig. 3G). However, when a vesicle has several membranes, only the outer membrane merges with the phagosome membrane and the internal membranes remain intact, keeping the viruses inside the vesicle (Fig. 3F).

The origins of the inner membrane of MsV.

Considering the results described above, a rational next step would be to perform immunoblotting assays of purified vesicles to confirm their origins. However, Boyer and colleagues (6) have previously hypothesized the existence of an MsV inner membrane. Therefore, to design this assay we would have to consider the possibility that this inner membrane has a different origin from that described for the membranes of the giant vesicles. Indeed, as described previously (6), during MsV morphogenesis we observed the incorporation of structures resembling membranes (Fig. 4A).

FIG 4.

Marseilleviruses have an internal membrane that likely originates from the endosome. (A) TEM images of immature (inset) and mature MsV particles showing the crescent structure of the viral particles and the internal membrane surrounding the nucleocapsid (arrows). (B) The internal membrane originates from the endosome, while the giant vesicles originate from the ER. A specific endosome antibody (SNX2) recognized the full vesicles and the purified particles but not the vesicles' membrane, indicating that endosomal components are present only when the viral particles are present (60 kDa). Otherwise, a specific antibody to ER (GRP78) recognized the full vesicles and the vesicles' membrane but not the purified particles, demonstrating the ER origins of giant vesicles (100 kDa). Both antibodies recognized the A. castellanii structures. The anti-Golgi antibody (anti-GM130) recognized only the positive control. (C) Immunofluorescence microscopy corroborates the Western blotting data once MsV colocalizes with the endosomal components. (D) A TEM image of an A. castellanii cell infected with MsV, showing the presence of several endosomes (some containing viruses) and some membrane-rich areas containing viruses, probably originating from the endosome. (E and F) TEM images showing different membrane-rich areas containing viral particles. *, viral particles; M, mitochondria; E, endosome; Ω, membrane-rich area. TEM images were obtained 3 h postinfection.

In order to confirm the origins of the giant vesicles and to investigate the origins of the MsV inner membrane, we performed immunoblotting assays of vesicles (membrane and particles), of particles released from vesicles (after detergent treatment), and of membranes purified from vesicles. Distinct parts of the cellular membranes were probed (ER, Golgi complex, and endosome). As expected, purified vesicle membranes were recognized only by the anti-ER marker antibody (GRP78). Remarkably, purified particles were recognized only by the anti-endosome marker antibody (SNX2). Finally, vesicles containing MsV particles (and the A. castellanii protein extract-positive control) were recognized both by anti-ER and anti-endosome antibodies. The anti-Golgi antibody (anti-GM130) only recognized the positive control (Fig. 4B). Therefore, these data suggest that the origin of the MsV inner membrane is endosomal.

Endosomes are recruited to early viral factories.

To investigate how endosome membranes are incorporated in the viral factory, we infected A. castellanii cells with MsV, and after 90 min we performed IF assays using antibodies against MsV (anti-MsV) and an endosome marker (anti-SNX2). The colocalization between anti-SNX2 and anti-MsV reinforced our previous results and suggests that endosomes are recruited to the MsV viral factory during the early stages of infection (Fig. 4C).

Since the endosome participates in the formation of MsV particles, we performed TEM of infected cells in order to observe its location and participation in the MsV multiplication cycle. During the early stages (1, 2, and 3 h) of infection, we observed many endosomal vacuoles in the cytoplasm of the host cell surrounding an early membrane-rich viral factory (Fig. 4D, E, and F). Since these vacuoles are found near the viral factory region, they probably contribute to the morphogenesis of the viruses, likely contributing to the formation of the particle internal membrane, corroborating our previous data.

Vesicles versus particles: biological properties.

Our next step was to evaluate if MsV giant vesicles confer any advantage compared to single viral particles. In order to measure the efficacy of vesicles and single particles in infecting A. castellanii, 1 × 10⁵ TCID₅₀ of each (full vesicles and viruses, previously normalized as infectious entities) was incubated with A. castellanii cells. The cells were observed daily and collected 24, 48, and 72 h postinfection for further titration. It was possible to observe that the vesicles have a higher efficacy in terms of infection than the isolated viruses. For each evaluated time point the vesicles reached significantly higher titers. After 72 h the titer of the vesicles increased by ∼2 logs, while the particles exhibited a lower increase of ∼1.5 logs (Fig. 5A).

FIG 5.

MsV entry into the host cell through different pathways. (A) Vesicles containing viruses provide a faster beginning to the replication cycle than single particles. (B) Vesicles confer longer-duration temperature resistance to MsV than single particles. (C and D) The effects of chloroquine and cytochalasin D on viral replication when viruses are isolated or found within vesicles. (E) TEM images of the MsV penetrating an A. castellanii cell using the endocytic pathway. (F) TEM images of giant vesicles penetrating an A. castellanii cell by phagocytosis. (G) A TEM image of grouped viral particles being engulfed by A. castellanii (arrows). The experiments were performed twice in duplicate. **, P < 0.001; ***, P < 0.0001. TEM images were obtained 1 h postinfection.

In order to investigate if giant vesicles confer any physical advantage to MsV compared to isolated particles, we simulated an extreme heat condition. After exposure to 70°C on multiple occasions, our findings showed the giant vesicles were significantly more resistant to this physical stress than the particles. Exposure to 70°C for 7 and 10 min completely eliminated the infectivity of isolated particles of MsV; however, some vesicles remained infectious under the same conditions, reaching ∼1 × 10² TCID₅₀ (Fig. 5B). This result strongly suggests that giant vesicles confer a physical advantage to MsV.

Vesicles versus particles: entry.

Based on our findings (Fig. 3C to G), i.e., the discovery of giant vesicles, we raised an interesting question regarding MsV entry: would vesicles be able to stimulate phagocytosis? To address this question, we designed additional experiments and compared the MsV entry pathways of giant vesicles and isolated particles.

Thus, A. castellanii cells were treated with chloroquine or with cytochalasin D and then infected with giant vesicles or isolated particles. Twenty-four hours postinfection the infected and treated cells were collected and the MsV titer determined. Chloroquine is a 9-aminoquinoline that prevents the acidification of the endosome and has a negative impact on the replication of some viruses that depend on this pathway (18). Cytochalasin D is a mycotoxin that promotes a variety of changes in cellular functions, including the destabilization of the cytoskeleton, and has been associated with the inhibition of phagocytosis (19).

Our results showed that treatment with chloroquine significantly impacted the titers of isolated particles (P < 0.001) but not (significantly) those of giant vesicles (Fig. 5C). On the other hand, cytochalasin D significantly decreased (P < 0.01) the titers of the cells infected with giant vesicles but not (significantly) those infected with isolated particles (Fig. 5D). Together, these results indicate that while isolated particles use an endosomal acidification-dependent entry pathway, giant vesicle entry is mediated via phagocytosis.

In addition, TEM images from amoebas infected with MsV corroborated our hypothesis that particle entry into the amoeba occurs via the endosome. It was possible to observe the viral particle interacting with the surface of the amoebae (Fig. 5E), and such interaction might occur through the short viral fibers found on the viral capsid. Moreover, after entry there were particles within the amoeba cytoplasm (Fig. 5E). Also, the TEM images corroborated the results from the cytochalasin D condition that vesicles can enter via phagocytosis, as it was possible to observe them both outside and inside the cell (Fig. 5F). Furthermore, it was possible to observe an aggregate of viral particles being engulfed by the cell, suggesting that phagocytosis of grouped particles is another virus entry route (Fig. 5G).

DISCUSSION

In 2013, Feng and colleagues described a pathogenic picornavirus able to acquire an envelope by hijacking cellular membranes (20). This represented a breakthrough in virology as, until then, a virus was considered either enveloped or nonenveloped, without the possibility of interchangeable states by the acquisition of membranes by the virus. A few other RNA viruses also are able to form infectious vesicles, which act mainly as a mechanism for escape from host immunity (21). Here, we describe for the first time, to our knowledge, a DNA virus, marseillevirus, that is able to induce the formation of giant vesicles while replicating within amoebas, its natural host.

Acanthamoeba species are unicellular protozoans that obtain nutrients through phagocytosis of microorganisms, including giant viruses (1, 22). Here, we have demonstrated that MsV, which is different from other amoebal viruses where a single particle can penetrate by phagocytosis, has developed an alternative mechanism to trigger this process, the formation of giant vesicles. The dispersion of pathogenic RNA viruses by vesicles was described as useful in repelling specific antibodies, thus providing an important advantage to those viruses within their hosts (20, 21). Given the natural hosts of marseilleviruses, this does not seem to be applicable, as amoebas do not possess an adaptive immune system. However, this could be important if humans were their hosts, since marseilleviruses have already been described as putative human pathogens (17, 23). In addition, our results suggest that infection through vesicles evolved as a powerful mechanism to boost the replicative success of this virus within its natural hosts and/or its survival in the environment where they coexist. Indeed, we have demonstrated here that being wrapped inside vesicles may confer to the virus a number of advantages, including a greater efficiency of infection and the ability to stimulate phagocytosis. Therefore, these vesicles can facilitate the spread of the virus to other susceptible cells by collectively transferring multiple viral genomes into the cytoplasm (21). Also, we demonstrated that the vesicles confer to MsV resistance to high temperature. Considering the vesicle as an infectious entity and that each viral particle has the same heat stability, being wrapped within a membranous structure is an advantage for the species (in the case of MsV as the infectious agent), since more particles will remain infectious and grouped even under harsh conditions. In addition, the vesicle membranes could confer more resistance to wrapped particles. The giant vesicles observed in this study seem to have originated from the ER and can contain one or more membranes (Fig. 1) and up to 1,000 viral particles (Fig. 3). It seems that MsV modulates these membranes for the host cell in order to delimit a large viral factory where the vesicles are produced. The number of membrane layers in the vesicles can influence the entry of the virus into the cell. When the vesicles have only one layer, they can fuse with the cell membrane, releasing the viruses directly into the cytoplasm of the amoeba in a disorganized way. When more than one layer is present, fusion also occurs but the internal membranes remain intact, preserving the viruses within them. The large size of the vesicles makes them appropriate for phagocytosis. According to a study by Korn and Weisman, which evaluated the phagocytosis capacity of Acanthamoeba using latex beads of different sizes, particles smaller than 500 nm, such as those of MsV, were not phagocytosed unless they formed vesicles or particle agglomerations around the amoeba surface (8). Generally, giant viruses, such as mimivirus, pandoravirus, and pithovirus, enter into the host cell via phagocytosis, since they range in size from 750 nm to 1.5 μm (1–3). However, despite the relatively small size of marseilleviruses, they are able to enter into their host via phagocytosis using a mechanism never described before for large DNA viruses (Fig. 6).

FIG 6.

Models of MsV cycle replication. This model summarizes the replication cycle of MsV in Acanthamoeba cells. MsV can enter into the cell by different pathways, including endocytosis of single particles, phagocytosis of giant vesicles, and phagocytosis of grouped particles.

In addition to the phagocytic pathway, marseilleviruses can enter into their host via endocytosis (Fig. 6). Macropinocytosis might be involved as well, but this should be investigated further (Fig. 5G and 6). A similar mechanism was observed in vaccinia virus (VACV), which can penetrate the cell in different ways, e.g., membrane fusion, macropinocytosis, or a low-pH-dependent endocytic route (24, 25). Another similarity between VACV and MsV is the presence of an inner membrane. In this study, we suggested that the inner membrane of MsV is acquired from the endosomes, although we believe that other sources cannot be ruled out. After penetration, it was possible to observe the accumulation of endosomal vacuoles near the viral factories (Fig. 4); therefore, when the particle was formed a membrane layer was incorporated into the viral capsid. Other large DNA viruses, such as Paramecium bursaria chlorella virus, African swine fever virus, and APMV, also acquire their inner membranes during the morphogenesis step, but these membranes have different origins (26). Altogether, our results fill in some of the gaps regarding the replication cycle of the marseilleviruses, providing some important biological information about these large viruses. Using several different approaches, we have demonstrated that MsV either can interact with the cell surface and enter through endocytosis or can be found inside giant vesicles that stimulate the phagocytosis mechanism. Finally, the discovery of giant viruses not only has brought to light their genomic complexity, with hundreds of new genes/proteins that are able to perform activities never before attributable to a virus, but also allowed us a glimpse into a microcosm of ecological interactions, where viruses are able, for example, to actively seek new hosts and habitats by adsorbing into other, more complex life forms, such as fungi and arthropods (14). Here, we have identified yet another possible ecological strategy for the release of a giant virus from its last host in vesicles that have the potential to boost its entry into another host, as well as to help it endure in the environment during an “interhost” period. Although rare, we can trace a parallel between this strategy and the type B inclusion body of cowpox virus, which also may represent a similar method of enhancing interhost resistance and infectivity (27). The investigation and discovery of new giant viruses is fundamental to uncovering new information about the biology of these amazing viruses and to better understand their biological interactions, both cellular and molecular, with their hosts.