Characterization of viroplasm formation during the early stages of rotavirus infection

Background During rotavirus replication cycle, electron-dense cytoplasmic inclusions named viroplasms are formed, and two non-structural proteins, NSP2 and NSP5, have been shown to localize in these membrane-free structures. In these inclusions, replication of dsRNA and packaging of pre-virion particles occur. Despite the importance of viroplasms in the replication cycle of rotavirus, the information regarding their formation, and the possible sites of their nucleation during the early stages of infection is scarce. Here, we analyzed the formation of viroplasms after infection of MA104 cells with the rotavirus strain RRV, using different multiplicities of infection (MOI), and different times post-infection. The possibility that viroplasms formation is nucleated by the entering viral particles was investigated using fluorescently labeled purified rotavirus particles. Results The immunofluorescent detection of viroplasms, using antibodies specific to NSP2 showed that both the number and size of viroplasms increased during infection, and depend on the MOI used. Small-size viroplasms predominated independently of the MOI or time post-infection, although at MOI's of 2.5 and 10 the proportion of larger viroplasms increased. Purified RRV particles were successfully labeled with the Cy5 mono reactive dye, without decrease in virus infectivity, and the labeled viruses were clearly observed by confocal microscope. PAGE gel analysis showed that most viral proteins were labeled; including the intermediate capsid protein VP6. Only 2 out of 117 Cy5-labeled virus particles colocalized with newly formed viroplasms at 4 hours post-infection. Conclusions The results presented in this work suggest that during rotavirus infection the number and size of viroplasm increases in an MOI-dependent manner. The Cy5 in vitro labeled virus particles were not found to colocalize with newly formed viroplasms, suggesting that they are not involved in viroplasm nucleation.


Background
Rotaviruses are the major cause of severe diarrhea in children and young animals worldwide. As a members of the family Reoviridae, they have a genome of 11 segments of double-stranded RNA (dsRNA) enclosed in three protein layers, forming infectious triple-layered particles (TLP) [1]. During, or just after entering the cell's cytoplasm, the outer capsid, composed of VP4 and VP7, is released, yielding transcriptionally active doublelayered particles (DLP). The produced viral transcripts direct the synthesis of viral proteins and serve as templates for the synthesis of negative-RNA strands to form the genomic dsRNA. During the replication cycle of rotavirus electron-dense cytoplasmic inclusions, named viroplasms, are formed [2]. Such cytoplasmic inclusions are observed during infection with a number of animal viruses [3], including reoviruses, as other members of the Reoviridae family [4].
In rotaviruses two non-structural proteins, NSP2 and NSP5, have been shown to be sufficient to form membrane-free cytoplasmic inclusions, which are known as viroplasms-like structures [5]. In vivo immunofluorescence visualization of viroplasms shows they are heterogeneous in size [6,7]. It is in these structures where the synthesis of dsRNA and its packaging into pre-virion core particles take place [8]. Besides NSP2 and NSP5, other viral proteins accumulate in viroplasms -namely VP1, VP2, VP3, VP6, and NSP6 [7,[9][10][11]. The key role of NSP2 and NSP5 proteins in the formation of viroplasms has been demonstrated by knocking-down their expression by RNA interference, which results in the inhibition of viroplasm formation, genome replication, virion assembly, and a general decrease of viral protein synthesis [7,8,12]. Viroplasm formation has been studied using electron or fluorescence microscopy [6,[13][14][15], however, despite their importance in the replication cycle of rotavirus, little is know about their dynamics of formation. The observation that bromouridine-labeled RNA localizes to viroplasms suggested that the viral transcripts are synthesized within viroplasms, which led to the hypothesis that the entering viral particles could serve as points of nucleation for the formation of viroplasms [8]. In this work, the dynamics of viroplasm formation in MA104 cells infected with rotavirus strain RRV was studied as a function of time and multiplicity of infection (MOI). Using fluorescently labeled purified rotavirus particles; we showed that the incoming TLPs do not seem to be involved in the formation of viroplasms.
Identification, quantitation and size analysis of viroplasms MA104 cells grown in 10 mm coverslips were infected with rotavirus strain RRV at different MOI's for 1 hour at 4°C. After washing unbound virus, the cells were incubated at 37°C for different times post-infection. The cells were fixed with 2% paraformaldehyde, and permeabilized with 0.5% Triton X-100 in PBS containing 1% bovine serum albumin, as described previously [17]. Cells were then incubated with rabbit polyclonal sera to NSP2 protein, followed by staining with goat anti-rabbit IgG coupled to Alexa-488 or 568. The images were acquired using a Zeiss Axioskop 2 Mot Plus microscope and analyzed by Image Pro Plus 5.0.2.9 and Adobe Photoshop 7.0. All images were acquired with a 60× objective, with a real time CCD Camera in 256 grey scales, and the size of the images was 1392 × 1040 pixels, with 8 bits. The estimation of viroplasm size was done using the Analyze particle function of Image J 1.32j program (Wayne Rasband, NIH, USA).

Immunodetection of rotavirus NSP2 protein
MA104 cells grown in 24-well plates were infected with rotavirus strain RRV at different MOI's for 1 hour at 4 o C. After washing unbound virus, the cells were incubated at 37 o C for different times post-infection. At the indicated time points, the cells were washed twice with PBS and lysed with Laemmli sample buffer. Proteins were separated by 10% SDS-PAGE and transferred to nitrocellulose membranes (Millipore, Bedford, MA). Membranes were blocked with 5% non-fat dried milk in PBS, and incubated at 4 o C with primary anti NSP2 polyclonal antibody in PBS with 0.1% milk, followed by incubation with secondary, horseradish peroxidase-conjugated antibodies. The peroxidase activity was revealed using the Western Lightning™Chemiluminiscence Reagent Plus (Pelkin Elmer Life Sciences). The images obtained were scanned and the band densities analyzed using Image pro software.

Conjugation of virus with fluorophore and colocalization of labeled viruses with viroplasms
To label virus with fluorophores, RRV virions were purified by cesium chloride gradient centrifugation as described previously [18]. The purified TLP's of simian strain RRV were washed twice with 10 mM Hepes pH 8.2, 5 mM CaCl 2 , 140 mM NaCl, and labeled with Cy5 mono reactive dye (0.1, 0.5, 1, 2.5, and 5 nmol of fluorophore for 1 μg of purified virus) at room temperature for 1 hour with gentle agitation. The reaction was stopped by addition of Tris-HCl pH 8.8 to a final concentration of 50 mM. Labeled viruses were separated from unbound fluorophore by gel filtration on a G25 sepharose column. As control, the purified TLP's of strain RRV were processed in identical way without addition of fluorophore. Viral titres were determined by a standard immunoperoxydase assay as described previously [19]. DLP's were prepared by EDTA treatment of labeled TLP's. To determine which viral proteins were conjugated with fluorophore, labeled and nonlabeled TLP's and DLP's were resolved in PAGE gel, analyzed on Typhoon-Trio (Amersham Biosciences) and stained by silver nitrate. Labeled particles were compared with FluoSpheres [carboxylate-modified microspheres, 0.1 μm, yellow-green fluorescent (505/515)] using confocal microscope LSM-510 Zeiss, mounted on inverted microscope Zeiss Axiovert 200 M, with AIM software, using objective Plan-neofluor 100×/1.30 Oil Ph3 (Carl Zeiss). To detect green staining, excitation laser Argon 2 488 nm was used with emission filter BP 500-530 nm, and for far red staining laser Helio-Neon 633 nm was used with emission filter BP650-670 nm. To colocalize labeled TLP's with viroplasms, MA104 cells grown on coverslips were infected with Cy5 -labeled RRV TLP's (MOI of 2) for 1 hour at 37°C. After washing unbound virus, the infection was left to proceed for 4 hours, and then the cells were fixed and the viroplasms were detected as described above using a rabbit polyclonal antibody specific for rotavirus NSP2 protein, and a goat anti rabbit IgG coupled to Alexa 488. Images were acquired using a confocal microscope as described above, as stacks of 10 images 800 nm thick, with resolution of 1024 × 1024 pixels, and processed by nearest neighbor deconvolution using AIM software. Acquired images were processed by Image J 1.32j and Adobe Photoshop 7.0. Analyzing corresponding individual images ensured localization of all Cy-5 labeled viral particles inside cytoplasm.

Results
The number of viroplasms and the level of the NSP2 protein increase during rotavirus infection, in direct correlation with the multiplicity of infection It has been described that viroplasms can be visualized in rotavirus infected cells by immunofluorescence as early as 2 hours postinfection (hpi) [14]. Therefore, to learn about the kinetics of viroplasm formation at early stages of rotavirus infection, MA104 cells grown in coverslips were infected with rotavirus strain RRV at different MOI's, and at different times post-infection (2, 4, 6, or 8 hours) the cells were fixed and the viroplasms were detected by immunofluorescence using a mono-specific serum to NSP2. Viroplasms were detected as early as 2 hpi as discrete dots that were not observed in control, mock infected cells, and their number and size increased as the infection proceeded ( Figure 1). To quantitate the increase in the number of viroplasms during infection, the number of viroplasms in 400 infected cells from each condition was scored. It was observed that independently of the MOI used, the number of viroplasms per cell increased as the infection proceeded (P < 0.05, Student's t-test), almost duplicating every two hours up to 6 hours ( Figure 2).
To determine the size of viroplasms, their area was determined in pixels 2 in 360 infected cells per condition. While at higher MOI's (2.5 and 10) there was a constant increase in the average size of the viroplasms, at low MOI's (0.1 and 0.5) a fluctuation in the viroplasm size was observed ( Figure 3A). To analyze the size of the viroplasms in more detail, viroplasms were divided into three arbitrary groups: small (4 -33 pixels 2 ), medium (34 -69 pixels 2 ), and large (70 -200 pixels 2 ) ( Figure  3B). Throughout the course of infection, and independently of the MOI used, or the time post infection analyzed, the population of small viroplasms predominated in number and also in proportion of all viroplasms (Figure 3C). Differences were observed when the size of viroplasms was compared in cells infected with low (0.1 and 0.5) or high (2.5 or 10) MOI's. While at high MOI's there was a gradual decrease in the proportion of small viroplasms during the course of infection, from 90 and 92% (2 hpi) to 56 and 45% (8 hpi) respectively, with a concomitant increase in the medium and large size viroplasms, the proportion of small viroplasms at low MOI's was more stable ( Figure 3C). This suggests that while at high MOI's small viroplasms might convert to larger size viroplasms, probably due to large amount of protein synthesized in cell, at low MOI's the formation of new small viroplasms prevails, and they could become larger at later times post infection, however, this possibility was not investigated in this work.
To determine if the number and size of viroplasms correlate with the level of NSP2 synthesized during infection, cells infected at different MOI's were harvested at different times post-infection, and the level of NSP2 was determined by Western blot (Figure 4). While at high MOI's (2.5 and 10) the NSP2 protein was detected at 4 hours post-infection, and increased as infection proceeded, at low MOI's (0.1 and 0.5) the NSP2 protein was detected until 8 hours post-infection ( Figure 4A). Densitometric analyses showed that the dynamics of accumulation of NSP2 during infection with high MOI's ( Figure, 4B) was similar to that observed for the increase in the number of viroplasms ( Figure 2).

Viroplasms do not colocalize with fluorescently labeled particles
It has been previously described that virus-like particles produced in insect cells by the co-expression of rotavirus capsid proteins VP6, VP4, VP7, and a VP2 protein fused to GFP or to DsRed protein, can be visualized in living cells [20]. Other viruses have also been observed in cells after being directly labeled with fluorophores; among these are influenza A virus, poliovirus, dengue virus, and SV40 [21][22][23][24]. To determine if the formation of viroplasms is nucleated by the entering viral particles, purified infectious TLP's of RRV were conjugated with Cy5. Analysis of viral proteins by PAGE showed that all proteins were conjugated. Importantly the intermediate capsid protein VP6 was efficiently labeled, ensuring that viral particles will be visible even after loss of the outer capsid proteins VP7 and VP4 (or its trypsin cleavage products VP5 and VP8) ( Figure 5A and 5B). Viral titre after conjugation was similar to mock conjugated virus, observing a small decrease of infectivity when using 5 nmol of Cy5 for conjugation, suggesting that viral infectivity was not compromised ( Figure 5C), therefore, for the following experiments 1 nmol of Cy5/μg of virus was chosen. Most importantly, both TLP's and DLP's (prepared from TLP's by EDTA treatment), labeled with 1 nmol of Cy5, were comparable to 100 nm Fluorospheres when observed in confocal microscopy ( Figure 5D). Since it was possible to visualize the fluorescently labeled viral particles, we used them to observe their intracellular distribution with respect to the newly formed viroplasms. To do this, MA104 cells grown in coverslips were infected with Cy5-conjugated RRV TLP's at an MOI of 2, and 4 hpi the cells were fixed, the viroplasms were immunostained using a polyclonal sera to NSP2, and images were acquired using confocal microscopy, as described under material and methods. Fluorescently labeled viral particles were observed distributed in the cytoplasm as discrete spots ( Figure 6). The number of labeled viral particles, viroplasms, and their co-localization was counted independently by two persons in 31 cells. In these, 117 labeled virus particles

Discussion
The formation of viroplasms has been previously studied using electron and fluorescence microscopy, however, those studies have focused only on late (4 to 24 hpi) stages of infection [6,13,15]. Only Eichwald et al [14] have studied earlier stages of viroplasm formation, and in their work, following the expression of an NSP2 protein fused to EGFP in rotavirus SA-11 infected cells, they observed that the total number of viroplasms decreased with time, with a concomitant increase in their size, starting at 6 hpi. This observation was interpreted as fusion events between smaller viroplasms. Similar results were reported by Cabral-Romero and Padilla-Noriega [15] using the strain SA-11 in BSC1 cells, although at even later (10 hpi) stages of infection. Comparing the formation of viroplasms between SA-11 and OSU rotavirus strains, Campagna et al. [6] observed that the viroplasms formed in OSU infected cells did not increase in size as readily as those formed during infection with SA-11. In this work, after infection with rotavirus strain RRV, using different MOI's, an increase in the number of viroplasms and in the amount of the NSP2 protein was observed. The size of viroplasms was observed to increase when higher MOI's were used. There are several possibilities to explain the discrepancies reported. First, the decrease in the number of viroplasms was observed only during infection with strain SA-11 [14,15], but not with strains OSU [12], and RRV (this work). It is known that some viral functions (receptor specificity, plaque formation, extraintestinal spread, IRF3 degradation, etc) may vary among different rotavirus strains [25][26][27][28] what opens the possibility that there could also be strain-specific differences for viroplasm formation. In fact, an impaired phosphorylation of NSP5 affected differently the morphogenesis of viroplasms in cells infected with either SA-11 or OSU rotavirus strains [6]. The differences observed between our studies and those of other groups could also arise from the different methodologies used to detect viroplasms. While in our case the newly synthesized rotavirus proteins were immunodetected and analyzed in 400 cells, in the study by Eichwald et al. [14] the identification of viroplasms was based on the detection of NSP2-EGFP or NSP5-EGFP fusion proteins in 20 cells. It is possible that the large amount of recombinant fusion proteins that accumulated in the cytoplasm of transfected cells before rotavirus infection could change the kinetics of viroplasm formation, since upon rotavirus infection a rapid redistribution of the EGFP -proteins was observed. It was not possible to compare the exact number of viroplasms obtained in that study, since the MOI that was used to infect the transfected MA104 cells was not mentioned.
In this work, studying the kinetics of viroplasm formation during the infection of strain RRV, we observed an increase in both the number and size of viroplasms with time and this increment was dependent on the MOI used. At high MOI's (2.5 and 10) the increase correlated with the amount of NSP2 protein detected at a given time point, while at lower MOI's (0.1 and 0.5), the smaller increase in NSP2 protein correlated with a less variable viroplasm size. It is possible that when a critical concentration of NSP2 and NSP5 is reached, and as other viral proteins accumulate, viroplasms start to form, first as small entities, and then becoming larger at later stages of the replication cycle. Although, it is not possible to determine if the increase in size is caused by fusion of smaller viroplasms or by addition of newly produced rotavirus proteins to small viroplasms, our observations are more consistent with the idea that new small viroplasms are generated constantly during the replication cycle, since even at later stages of infection a large proportion of small viroplasms was observed. It remains to be determined if the small viroplasms, presumably generated by the aggregation of NSP2 and NSP5 require an additional priming signal, or if it is only the concentration of free NSP2 and NSP5 what dictates the formation of a new viroplasms.
The mechanism of viroplasms formation and its protein content is unknown. The fact that viroplasms are sites for rotavirus transcription at late stages of infection (8.5 hpi) led to the so far unproven hypothesis that incoming DLP's serve as focal points of viroplasm assembly [8]. In this work we tested this hypothesis by visualization of incom-  C   hpi   small  496  1024  1493  2438  medium  46  75  367  396  large  29  28  239  267  total  571  1127  2099  3101   small  1425  1993  3626  3371  medium  75  318  1125  2048  large  54  210  697  2128  total  1554  2521  5448  7547   %   of   viroplasm   small  737  1319 2523  4313  medium 105  88  746  768  large  26  64  415  503  total  868  1417  3684 Figure 2). These results suggest that at the onset of infection the entering viral particles do not serve as nucleation centers for the formation of viroplasms as suggested [8]. The fact that the plasmid expression of NSP2 and NSP5 proteins alone, in absence of infectious virus, are able to form viroplasm-like structures also supports this conclusion.
Recently it was suggested that rotavirus viroplasms could interact with microtubules [15]. NSP2 was also shown to interact with tubulin, inducing the collapse of the microtubule network, and viroplasms were shown to colocalize with tubulin granules [29]. Similar interaction of reovirus viral inclusion bodies with microtubules [30] suggests the possibility that tubulin could have a more general role in the replication cycle of viruses of the Reoviridae family.
Although viroplasms play a crucial role in rotavirus replication and assembly, the factors that govern their formation and function, are still not clearly understood. The development of live cell imaging tools should provide more detailed information about these processes.

Conclusions
Rotavirus replication takes place in electrodense structures known as viroplasms, however, little is known about their dynamics of formation, and the factors that drives viroplasm nucleation. The results presented in this work show that during rotavirus infection the number and size of viroplasms increases steadily with time, and depends on the MOI used. Using in vitro Cy5labeled infectious viral particles we observed that the entering viruses do not seem to be involved in viroplasm nucleation. It is possible that some cellular protein, like  Figure 6 Viroplasms do not colocalize with Cy5 labeled infectious rotavirus particles. MA104 cells grown in coverslips were infected with purified rotavirus strain RRV TLPs (left side panels) or purified rotavirus strain RRV TLP's labeled with Cy5 (right side panels). Four hours after infection, cells were fixed, and the viroplasms detected with an anti-NSP2 polyclonal antibody, and a secondary antibody coupled to Alexa-488. Images were obtained with confocal microscope LSM 510 and processed as described in material and methods. In merge, detected viroplasms are in green, and Cy5 labeled RRV particles are in red, pointed by arrows. Detail of viroplasm location and Cy5 labeled RRV are shown. tubulin, are required for this process, however much work is needed to characterize in detail this essential step of rotavirus replication.