Morphological and Taxonomic Properties of the Newly Isolated Cotonvirus japonicus, a New Lineage of the Subfamily Megavirinae

ABSTRACT Since 2003, various viruses from the subfamily Megavirinae in the family Mimiviridae have been isolated worldwide, including icosahedral mimiviruses and tailed tupanviruses. To date, the evolutionary relationship between tailed and nontailed mimiviruses has not been elucidated. Here, we present the genomic and morphological features of a newly isolated giant virus, Cotonvirus japonicus (cotonvirus), belonging to the family Mimiviridae. It contains a linear double-stranded DNA molecule of 1.47 Mb, the largest among the reported viruses in the subfamily Megavirinae, excluding tupanviruses. Among its 1,306 predicted open reading frames, 1,149 (88.0%) were homologous to those of the family Mimiviridae. Several nucleocytoplasmic large DNA virus (NCLDV) core genes, aminoacyl-tRNA synthetase genes, and the host specificity of cotonvirus were highly similar to those of Mimiviridae lineages A, B, and C; however, lineage A was slightly closer to cotonvirus than the others were. Moreover, based on its genome size, the presence of two copies of 18S rRNA-like sequences, and the period of its infection cycle, cotonvirus is the most similar to the tupanviruses among the icosahedral mimiviruses. Interestingly, the cotonvirus utilizes Golgi apparatus-like vesicles for virion factory (VF) formation. Overall, we showed that cotonvirus is a novel lineage of the subfamily Megavirinae. Our findings support the diversity of icosahedral mimiviruses and provide mechanistic insights into the replication, VF formation, and evolution of the subfamily Megavirinae. IMPORTANCE We have isolated a new virus of an independent lineage belonging to the family Mimiviridae, subfamily Megavirinae, from the fresh water of a canal in Japan, named Cotonvirus. In a proteomic tree, this new nucleocytoplasmic large DNA virus (NCLDV) is phylogenetically placed at the root of three lineages of the subfamily Megavirinae—lineages A (mimivirus), B (moumouvirus), and C (megavirus). Multiple genomic and phenotypic features of cotonvirus are more similar to those of tupanviruses than to those of the A, B, or C lineages, and other genomic features, while the host specificity of cotonvirus is more similar to those of the latter than of the former. These results suggest that cotonvirus is a unique virus that has chimeric features of existing viruses of Megavirinae and uses Golgi apparatus-like vesicles of the host cells for virion factory (VF) formation. Thus, cotonvirus can provide novel insights into the evolution of mimiviruses and the underlying mechanisms of VF formation.


RESULTS
Isolation of cotonvirus. Upon coculture with Acanthamoeba castellanii, a cytopathic effect (CPE) was observed. After isolation, we obtained three new viruses of the subfamily Megavirinae, including a novel virus (named cotonvirus herein) from Japan's Chiba Prefecture, an unidentified virus belonging to lineage A, and a mimivirus belonging to lineage C (identified as Megavirus musashi) from Japan's Saitama Prefecture; this was elucidated based on molecular phylogenetic analyses using B family DNA polymerase genes, as described below.
Morphological features. Transmission electron microscopy (TEM) and cryoelectron microscopy (cryo-EM) revealed that cotonvirus particles were morphologically very similar to other lineage A, B, and C mimiviruses, i.e., with three capsid layers, an inner membrane surrounding a core that stained densely, and surface fibrils (Fig. 1a and c). The cotonvirus particles exhibited an icosahedral capsid (approximately 400 nm in diameter), surface fibrils (approximately 100 nm), and a stargate structure, which is present at a single vertex of the particle for releasing the genome into the host cell, similar to those of other mimiviruses (Fig. 1a to d). The surface fibrils of the cotonvirus were denser and shorter than those of other mimiviruses (Fig. 1a, c, e, and g). Scanning electron microscope images revealed that the surface of cotonvirus particles was smoother than that of lineage A Mimivirus shirakomae particles ( Fig. 1c and g). Although the likelihood of interference in fiber visualization by electron microscope sample preparation procedures cannot be excluded, we named the new virus Cotonvirus japonicus because of the cotton-like appearance of the surface fibrils (Fig. 1a).
Host specificity. The genus Acanthamoeba is classified into three groups, namely, groups 1 to 3, based on the characteristics of their cysts (22). A. castellanii belongs to group 2, along with A. polyphaga (22). To identify cotonvirus host specificity, it was allowed to infect acanthamoeba cells of typical species of each group: Acanthamoeba comandoni (group 1), A. castellanii (group 2), and Acanthamoeba culbertsoni (group 3), respectively. All the viruses tested infected A. castellanii, but only the lineage A mimivirus infected A. culbertsoni ( Fig. 2 and Table 1). Other giant viruses, including cotonvirus, lineage C (megavirus) viruses, members of the family Marseilleviridae, and medusavirus, did not infect A. comandoni and A. culbertsoni, suggesting that the host specificity of cotonvirus is similar to that of lineage C viruses (we did not test for lineage B). Conversely, Vermamoeba vermiformis, which has already been reported to be infected by tupanviruses (19), was not infected with cotonvirus ( Fig. 2 and Table 1).
Genome. The cotonvirus genome is 1,476,527 bp. It is larger than those of lineages A to C of the subfamily Megavirinae and closer to the Tupanvirus soda lake (1,516,267 bp) and Tupanvirus deep ocean (1,439,508 bp) strains in terms of size ( Fig. 3 and Table 2). In addition, the G1C content of the cotonvirus genome was 25.3%, similar to those of the lineage B moumouvirus (24.6%) and lineage C Megavirus chilensis (25.2%).
We identified 1,306 predicted open reading frames (ORFs) (Data Set S1) and three tRNA-like sequences for tryptophan (CCA), cysteine (GCA), and leucine (TAA). The predicted number of ORFs is also similar to those of the tupanvirus soda lake and tupanvirus deep ocean strains (1,425 and 1,276, respectively) ( Table 2). Among all the ORFs, 1,149 (88.0%), including those encoding annotated hypothetical proteins, showed high homology to the genes of the family Mimiviridae ( Fig. 4a and b). The other best hits included 4 ORFs with homology to those of other viruses, 35 with homology to those of eukaryotes, 48 with homology to those of bacteria, and 1 with homology to those of archaea. A total of 69 were orphan genes (ORFans) (Fig. 4a and c). Among the ORFs that matched Mimiviridae, the best hits were from lineages A (42.0%), B (18.9%), and C (20.4%), tupanviruses (6.0%), klosneuviruses (0.6%), and Cafeteria roenbergensis virus (0.1%), suggesting that the genes of cotonvirus are homologous to the genes of the entire Mimiviridae family, rather than those of any particular lineages or viruses.
Furthermore, we calculated the proportion of each category of the genes in the subfamily Megavirinae that were best hits for cotonvirus genes (Fig. 5b). Among these categories, cotonvirus genes are similar to those of lineage A in the categories "DNA replication, recombination and repair," "signal transduction regulation," and "transcription and RNA processing." The cotonvirus genes under the category "mobile elements" showed the greatest similarity with those of lineage C, whereas those under "virion structure and morphogenesis"  were more similar to those of lineage B than to those of other lineages. In the "host-virus interaction" category, similarities among cotonvirus, lineage B viruses, and tupanvirus were observed.
These results suggest that cotonvirus is a new member and that it exhibits chimeric features of the subfamily Megavirinae. Our findings clearly demonstrate that cotonvirus is a novel independent lineage under Megavirinae, which includes icosahedral mimiviruses and tailed tupanviruses.
Molecular phylogenetic analysis. To estimate the phylogenetic position of cotonvirus in the subfamily Megavirinae, we performed molecular phylogenetic analysis based on NCLDV core genes, including those encoding B family DNA polymerase, major capsid  protein, D5-like ATPase, mRNA-capping enzyme, and virion packaging ATPase, and reconstructed their molecular phylogenetic trees (Fig. 6). Cotonvirus did not belong to any known lineage and was not included in existing sister groups in each of these phylogenetic trees. In a molecular phylogenetic tree based on the B family DNA polymerase gene, cotonvirus formed a new sister group with the clade comprising lineage A and tupanvirus (Fig. 6a), whereas in one based on the major capsid protein, D5-like ATPase, and mRNA-capping enzyme genes, cotonvirus formed a new sister group with the clade comprising lineage A only (Fig. 6b). Additionally, in a molecular phylogenetic tree based on the virion-packaging ATPase gene, cotonvirus formed a sister group with the whole subfamily Megavirinae. These results suggest that each cotonvirus NCLDV core gene is similar to those of different lineages, indicating that cotonvirus and the other existing lineages of the subfamily Megavirinae diverged at different times during their evolution. Furthermore, based on a concatenated gene sequence consisting of the five NCLDV core genes, the cotonvirus formed a sister group with lineage A (Fig. 7). Thus, cotonvirus is phylogenetically and evolutionarily related to lineage A. According to the constructed proteomic tree, the cotonvirus branched from the root of the clade consisting of 3 lineages (A, B, and C), and these viruses, along with cotonvirus, formed a sister group with the clade comprising tupanvirus (Fig. 8). These results suggest that cotonvirus represents an independent lineage under the subfamily Megavirinae.  Additionally, we performed molecular phylogenetic analyses based on the aa-RS genes ( Fig. 9). Similar to the results of phylogenetic analyses based on the NCLDV core gene, cotonvirus was found not to belong to any existing lineage and to instead form an independent lineage. The arginyl-RS and isoleucyl-RS genes of cotonvirus represent independent lineages in the clade comprising lineages A to C ( Fig. 9a and e). The cysteinyl-RS and methionyl-RS genes of cotonvirus formed sister groups with those of lineage A ( Fig. 9b and c), whereas the tyrosyl-RS gene of cotonvirus formed a sister group with that of lineage C, although the bootstrap support was relatively poor (Fig. 9d). Based on these results, we concluded that molecular phylogenetic analysis has provided clues to elucidate the evolutionary perspective of the subfamily Megavirinae, including cotonvirus.
Furthermore, the 18S rRNA intronic region is widely observed in the family Mimiviridae, and there have been reports of there being two copies of 18S rRNA intronic regions in tupanviruses, Borely moumouvirus (lineage B), and lineage C viruses, whereas there is only one copy in other lineages (19,23). In the cotonvirus genome, we detected two copies of 18S rRNA-like sequences, similar to tupanviruses and lineage C (Fig. 10), one of which exists in the intron of the DNA-directed RNA polymerase gene, similar to lineages A to C (19), whereas the other is in the neighboring region of the putative ATP-binding protein, unlike tupanviruses and lineage C viruses, whose 18S rRNA-like sequence is close to the capsid protein 1 gene (19). These data support the hypothesis that cotonvirus represents an independent lineage of the subfamily Megavirinae.
Thus, we propose a putative evolutionary model for cotonvirus and for the subfamily Megavirinae (Fig. 11). The common ancestor of the subfamily Megavirinae gave rise to the ancestors of the tupanvirus and cotonvirus, and the latter further gave rise to the ancestors of cotonvirus and of lineages A, B, and C. We hypothesized that each lineage of the subfamily Megavirinae evolved by acquiring genes via lateral gene transfer (LGT) (24)(25)(26). The putative evolutionary model is consistent with the genomic features of cotonvirus, whose genes are highly homologous with those of lineages A to C and whose genome complexity is similar to that of tupanvirus.
Sequences of virophages in cotonvirus genome. We tested whether virophage sequences, such as the 28-nucleotide-long Zamilon insert sequence (AATCTGATAATGAA TCTGATAATGAATC), which has been observed in reported APMV genomes (15), are present in the cotonvirus genome. We have detected this sequence in the cotonvirus genome at only one site: in a complementary strand of ORF1052 (annotated as "hypothetical protein") at genome position 1,204,030 to 1,204,061 (data not shown). This Zamilon insert-like sequence is 32 nucleotides long (GAATCTGATAATGAATCTGATAATGAATCCGA), which is longer than that in the APMV genome (15). Conversely, we did not detect the derived 15-nucleotide repeated unit, TGATAATGAATCTGA, and Cas-like gene, as presented in the genomes of lineage A viruses. Therefore, we have not presently confirmed that there is a MIMIVIRE-like system in the cotonvirus genome as in the lineage A mimivirus genome (15). Additionally, we did not detect sequences highly homologous with those of virophages Sputnik (27) or Guarani (28) in the cotonvirus genome.
Infection cycle by time-lapse imaging. Using time-lapse imaging for 36 h, we observed the kinetics of cotonvirus-infected A. castellanii cells and compared it with the kinetics of A. castellanii cells infected with Mimivirus shirakomae (lineage A) and Megavirus musashi (lineage C) ( Fig. 12; Movie S1 and Supplemental Data Set and Movie Legends). Figure 12 shows graphical representations of values for each virus infection (left) and the average value of each virus infection every 30 min (right). The estimated cell numbers did not increase in mimivirus-, megavirus-, and cotonvirus-infected A. castellanii cells ( Fig. 12a  and b), but they finally decreased due to cell lysis at the end of the infection (data not shown). On the other hand, there were differences in cell migration (which represents the distance of movement of A. castellanii cells from one frame to the next frame of time-lapse images) among the groups. The average numbers of mimivirus-and megavirus-infected A. castellanii cells decreased after 4 to 6 h postinfection (hpi), whereas the cotonvirus-infected A. castellanii cells decreased only after 10 hpi, suggesting that the CPE of cotonvirus is expressed later than that of lineage A and C viruses ( Fig. 12c and d).
Because lineage A mimivirus-infected cells could form a VF at 4 to 6 hpi (16), this suggests that the average step (average moving distance) of cotonvirus-infected A. castellanii cells delays the formation of VF. Based on these results, we estimated that cotonvirus-infected A. castellanii cells would begin to form cotonvirus VFs at 10 hpi. The average sizes of mimivirus-, megavirus-, and cotonvirus-infected A. castellanii cells decreased, plateaued, and then started to decrease again ( Fig. 12e and f). These cells had a round morphology due to CPE, and thus, their sizes were smaller than the original sizes of healthy trophozoites. Subsequently, they spread the new virions via cell lysis at the end of infection, resulting in a further reduction in size. Therefore, cell lysis was considered to begin during the second decrease ( Fig. 12e and f). The average size of cotonvirus-infected A. castellanii cells decreased again at 24 hpi, which was later than those infected with mimivirus and megavirus; therefore, the cotonvirus infection cycle was completed and numerous virions were released after approximately 24 hpi. Similar to the rounding, which was observed microscopically, the sphericality of the mimivirus-, megavirus-, and cotonvirus-infected A. castellanii cells gradually increased and plateaued at 10 (mimivirus and megavirus) or 18 (cotonvirus) hpi ( Fig. 12g and h). Maturation of virion factory using host Golgi apparatus-like vesicles. We then investigated the formation of the VF using TEM. At 2 hpi, cotonvirus was engulfed by the phagosome of A. castellanii cells, which led to the opening of the stargate structure and the release of its inner materials into the host cytoplasm (Fig. 13a). At 4 hpi, the viral core, which appeared to be surrounded by an intertwined membrane-like structure, was observed to be released into the host cytoplasm. At 8 hpi, VF started to form in the host cytoplasm; it was fully developed at 16 hpi. Interestingly, we observed electron-dense organelle-like structures, which represent a dot-and filament-like form around the early VFs ( Fig. 13a and b), that resembled the Golgi apparatus previously reported in the cysts of A. castellanii (22). At 16 and 24 hpi, new virions were assembled at the edge of the VF and accumulated in the host cytoplasm (Fig. 13a). Finally, cell debris and a piece of the VF were observed at 28 hpi, suggesting that cell lysis due to the many duplicated cotonvirus particles occurred only after 24 hpi. The duration of one cotonvirus infection cycle, from its addition to A. castellanii culture until cell lysis, is similar to that of tupanvirus (19).
To depict the conformation of the Golgi-like structures around the early VF, scanning transmission electron microscopy (STEM) tomography was performed on a 12-hpi sample. We detected flat and curved tubular structures similar to the previously reported Golgi stack (22) around the early VF ( Fig. 14; Movie S2). In addition, some early VFs were combined with the flat and curved tubular structures, which were then incorporated by early VFs (Fig. 14).
Because the cotonvirus VF was formed within Golgi apparatus-like vesicles (Fig. 12), we next performed immunofluorescence experiments using anti-GM130 antibodies, which target cis-Golgi matrix protein 130 (GM130) (Fig. 15). GM130 antibodies were clearly localized at the edge of matured VFs at 24 hpi compared with early VFs (Fig.  15a and b). This strong GM130 signal was not observed in the mature VFs of viruses of lineage A (mimivirus) or C (megavirus) (Fig. 15c). Thus, these results strongly suggest that the cotonvirus VF is derived from the membranes of the Golgi apparatus; this feature of VF formation has only been observed in cotonvirus-infected cells.
In APMV, it has been reported that the maturation of VF and production of viral particles require ATP and nucleotides for the synthesis of viral DNA, with VFs being surrounded by numerous mitochondria (16). In the cotonvirus infection cycle that we observed, we did not detect mitochondria surrounding VFs; therefore, in the present study, we do not confirm that cotonvirus VFs also require mitochondrial function similarly to APMV.

DISCUSSION
In this study, we isolated three new viruses belonging to the subfamily Megavirinae (family Mimiviridae), one of which was named cotonvirus and has several unique features not observed in other existing lineages of the subfamily Megavirinae, including APMV, moumouvirus, and megavirus.
Genome analysis showed that cotonvirus has the largest genome (1,476,527 bp long) among the existing icosahedral lineage A, B, and C viruses, and its genome size is similar to those of tupanviruses (18). A proteomic tree reconstructed using VipTree (29) suggested that the root of the cotonvirus genome was positioned after the branching of tupanviruses and other lineages. Furthermore, the cotonvirus genome has 1,306 ORFs, including ORFans, which is close to the number of ORFs of tupanviruses. The existence of two candidate 18S rRNA-like sequences in the cotonvirus genome is also similar to what has been observed in tupanviruses. The results suggest that the overall structure of the cotonvirus genome is closer to that of the tupanvirus genome than to those of other icosahedral mimivirus genomes. Conversely, several NCLDV core genes, including those encoding major capsid protein, D5-like ATPase, and mRNA-capping enzymes, are phylogenetically closer to those of icosahedral mimivirus lineage A viruses than of tupanvirus, and host specificity trends where cotonvirus only infects A. castellanii but not V. vermiformis are consistent with the trends in icosahedral mimiviruses and not tupanvirus.
Phylogenetic analysis of individual genes, as well as the molecular phylogenetic analysis of five NCLDV core genes, revealed that cotonvirus has a slightly higher number of genes homologous with those of lineage A of the subfamily Megavirinae than of genes homologous with those of other lineages, including tupanviruses, in the functional categories of DNA replication, recombination, and repair, transcription and RNA processing, and signal transduction. Additionally, molecular phylogenetic analysis of 5 aa-RS genes suggests that the cotonvirus may have independently acquired these genes during the evolutionary process. Three of the five aa-RS genes of cotonvirus are more similar to those of lineage A viruses than to others, but the remaining two genes represent independent lineages. Based on these results, we propose an evolutionary scenario for cotonvirus, stating that the common ancestor of the subfamily Megavirinae was first divided into the ancestral tupanvirus and the ancestral cotonvirus, which then further divided into the independent cotonvirus lineage and lineages A to C.
We attribute the relatively chimeric phylogenetic relationships between the individual genes of cotonvirus and their homologous genes of other lineages to the branch-  ing before lineages A to C, which have diverged. Several independent mimiviruses that can also infect acanthamoeba cells have been reported (4), suggesting that simultaneous infection by cotonvirus and other viruses may have occurred, followed by LGT between cotonvirus and other lineages that share their hosts.
Notably, we did not observe MIMIVIRE-like repeat sequences similar to that in APMV, but we did observe one 32-nucleotide-long Zamilon insert-like sequence in the complementary strand of ORF1052 in the cotonvirus genome. This suggests that an LGT between cotonvirus and Zamilon occurred in the past, and if it implies that cotonvirus has just recently acquired the defense against Zamilon, its molecular mechanism is distinct from that in APMV. Otherwise, we may be observing the evolution of the cotonvirus' defense system in action. Although we did not detect any MIMIVIRE-like systems similar to that in Zamilon in the present study, it is still possible that cotonvirus has unknown systems similar to MIMIVIRE. In future analyses, the infectivity of virophages like Zamilon, Sputnik, and Guarani in cotonvirus should be investigated.
The cotonvirus infection cycle was investigated in detail by analyzing the kinetics of cotonvirus-infected A. castellanii cells and VF formation using various microscopes. The PKA3 software recently developed in our laboratory can reveal the giant-virus-infected A. castellanii cell-specific motilities using time-lapse imaging under phase-contrast microscopy (30). Using this algorithm, we estimated two time points of note in the cotonvirus infection cycle: the starting point of VF formation (approximately 10 hpi) and the endpoint of the infection cycle, which represents the cell lysis of virus-infected A. castellanii cells (.24 hpi). In combination with TEM analysis, what occurs at these time points can be explained based on their actual intracellular status.
Interestingly, TEM and STEM analysis suggested that the VF formation of cotonvirus started in the Golgi apparatus-like vesicles of A. castellanii cells, which have been reported to change in morphology between trophozoite and cyst stages and to include more electron-dense materials in cysts than in trophozoites (22). These Golgi apparatus-like structures are filled with electron-dense materials typical of those observed in a cyst of A. castellanii cells, as previously reported (22), suggesting that cotonvirus infection results in a host cell response to cyst-forming conditions. Along with the observed colocalization of GM130, these observations strongly suggest that the membranes of the host Golgi apparatus are used for constructing VFs. GM130 is a peripheral membrane protein frequently found on the cis face of the Golgi apparatus and is involved in the sac mooring/dissociation of Golgi stack formation (31,32). A previous study has shown that membrane assembly of mimivirus particles occurs at the edge of VFs. The inner membrane of mimivirus particles has also been suggested to be derived from the ER (16)(17)(18). However, based on our observation of the structures around early VFs and the colocalization of GM130 to the edge of matured VFs, we propose that the membrane assembly or detachment of the VFs of cotonvirus particles involves or requires the function of the membranes of the Golgi apparatus, which is different from what has been observed for other icosahedral lineages. In other words, we propose that the inner membranes of cotonviruses are directly provided by the host Golgi apparatus.
The structural differences around the core and early VFs were presumably due to the morphological changes in the host Golgi apparatus during VF formation, i.e., when the host Golgi body became encysted. The higher fluorescence intensity of anti-GM130 antibody at 24 hpi than at other time points may be due to the aggregation of scattered Golgi apparatuses in the host cytoplasm or the high expression of GM130 at the end of the infection cycle. On the other hand, we have not yet obtained evidence of the transport of some cotonvirus-encoded proteins into the Golgi apparatus, the involvement of such proteins in VF formation within the Golgi apparatus, or the recruitment of Golgi-derived inner membranes of cotonviruses for subsequent viral infection processes. Further studies are needed to determine how the functions of the host Golgi apparatus are used for virus production. In addition, because this phenomenon was not observed in mimivirus-and megavirus-infected A. castellanii cells, how GM130 is used during cotonvirus replication should be investigated.
In summary, we isolated a new lineage of the subfamily Megavirinae named cotonvirus, which has not previously been identified by metagenomic analysis or isolated from environmental samples. Analysis of the infection cycle and genomic features of cotonvirus provided new insights into the subfamily Megavirinae. Our results have important implications for understanding the replication mechanisms and evolution of the subfamily Megavirinae and, consequently, the family Mimiviridae.

MATERIALS AND METHODS
Virus isolation, purification, and titration. Acanthamoeba castellanii (Douglas) strain Neff (ATCC 30010) was purchased from the American Type Culture Collection (ATCC, Manassas, VA, USA). A. castellanii cells were cultured in proteose peptone-yeast extract-glucose (PYG) medium at 26°C as previously described (7,11,33). Water samples (50 ml) were collected from a canal in Chiba, Japan (35°20906.20N, 139°50944.20E), and an irrigation ditch in Saitama, Japan (35°51937.00N, 139°48946.70E), and then stored at 4°C until inoculation. An aliquot (4.5 ml) was mixed with 2Â PYG medium (4.5 ml) and an antibiotic solution (360 ml) as described previously (33) and added to the A. castellanii cell suspension (50 ml). The mixture was inoculated into a 96-well microplate and incubated at 26°C. After 3 to 4 days, we observed cytopathic effects (CPE) in A. castellanii cells. The supernatant was seeded into fresh A. castellanii cells, and the virus was cloned by serial dilution. The supernatant of the most diluted well that showed CPE was seeded into a fresh A. castellanii cell suspension in a 25-cm 2 culture flask and incubated at 26°C. After 4 days, the supernatant was collected and centrifuged at 500 Â g for 5 min to remove cellular debris. Then, the supernatant was centrifuged at 8,000 Â g for 35 min to obtain the viral pellet. The viral pellet was washed with phosphate-buffered saline (PBS), filtered through a 1.2-mm filter (GVS Filter Technology, Inc., Indianapolis, IN, USA), and diluted in PBS. Ultimately, we isolated two new mimiviruses: Cotonvirus japonicus (a new lineage) from Chiba and Megavirus musashi (lineage C) from Saitama, which were identified using phylogenetic analysis. The viral titer was calculated using the endpoint method with TCID 50 (50% tissue culture infective dose) Calculator version 2.1 (Marco Binder, Department of Infectious Diseases, Molecular Virology, Heidelberg University).
Amoeba cells (A. castellanii or V. vermiformis; 2.0 Â 10 4 ) were mixed with PYG medium (100 ml) and viruses of the family Mimiviridae (MOI = 100) or other viruses, including medusavirus and those of the family Marseilleviridae (MOIs were not calculated). The mixture was inoculated into a 96-well microplate and incubated at 26°C. After 1 day, we determined the presence of CPE using an all-in-one fluorescence microscope (BZ-X800/X810; Keyence Co., Osaka, Japan) with a 20Â objective lens. sequence homology searches between cotonvirus and MIMIVIRE-related sequences or virophages (Sputnik and Guarani) were performed using BLASTn (NCBI Nucleotide collection (nt) database).
Phylogenetic analysis. Nucleotide sequences of the family Mimiviridae were obtained from NCBI (Data Set S2) and aligned using the default option of the MUltiple Sequence Comparison by Log-Expectation (MUSCLE) program. All positions containing gaps and missing data were removed from the alignments, and the best evolutionary models were estimated. Phylogenetic trees were constructed using the maximum-likelihood method with the estimated best evolutionary models and 1,000 bootstrap replicates, and figures were drawn using the MEGA X program (version 10.0.5) (42). Concatenated alignments of NCLDV core genes, including those encoding the B family DNA polymerase, major capsid protein, D5-like ATPase, mRNA-capping enzyme, and virion-packaging ATPase, were manually joined after eliminating the gap. Then, the concatenated tree was reconstructed similarly to the individual trees.
Data availability. The genome sequence of cotonvirus has been deposited in the DNA Data Bank of Japan (DDBJ) and GenBank (accession no. AP024483).