Diversity of Surface Fibril Patterns in Mimivirus Isolates

ABSTRACT

Among the most intriguing structural features in the known virosphere are mimivirus surface fibrils, proteinaceous filaments approximately 150 nm long, covering the mimivirus capsid surface. Fibrils are important to promote particle adhesion to host cells, triggering phagocytosis and cell infection. However, although mimiviruses are one of the most abundant viral entities in a plethora of biomes worldwide, there has been no comparative analysis on fibril organization and abundance among distinct mimivirus isolates. Here, we describe the isolation and characterization of Megavirus caiporensis, a novel lineage C mimivirus with surface fibrils organized as “clumps.” This intriguing feature led us to expand our analyses to other mimivirus isolates. By employing a combined approach including electron microscopy, image processing, genomic sequencing, and viral prospection, we obtained evidence of at least three main patterns of surface fibrils that can be found in mimiviruses: (i) isolates containing particles with abundant fibrils, distributed homogeneously on the capsid surface; (ii) isolates with particles almost fibrilless; and (iii) isolates with particles containing fibrils in abundance, but organized as clumps, as observed in Megavirus caiporensis. A total of 15 mimivirus isolates were analyzed by microscopy, and their DNA polymerase subunit B genes were sequenced for phylogenetic analysis. We observed a unique match between evolutionarily-related viruses and their fibril profiles. Biological assays suggested that patterns of fibrils can influence viral entry in host cells. Our data contribute to the knowledge of mimivirus fibril organization and abundance, as well as raising questions on the evolution of those intriguing structures.

IMPORTANCE Mimivirus fibrils are intriguing structures that have drawn attention since their discovery. Although still under investigation, the function of fibrils may be related to host cell adhesion. In this work, we isolated and characterized a new mimivirus, called Megavirus caiporensis, and we showed that mimivirus isolates can exhibit at least three different patterns related to fibril organization and abundance. In our study, evolutionarily-related viruses presented similar fibril profiles, and such fibrils may affect how those viruses trigger phagocytosis in amoebas. These data shed light on aspects of mimivirus particle morphology, virus-host interactions, and their evolution.

INTRODUCTION

In 2003, Acanthamoeba polyphaga mimivirus (APMV) was the first amoeba-associated giant virus to be described and studied (1). The APMV exhibits particles composed of a capsid with pseudo-icosahedral symmetry covered by a dense layer of external fibrils, approximately 750 nm in size. The capsid is formed by several protein layers and an inner lipid membrane, which encompasses the viral core, where the viral genome is located (2). Recently, a study suggested that the APMV genome is organized within a 30-nm helical protein shell composed primarily of 2-glucose-methanol-choline (GMC) oxidoreductases (genomic fiber) encoded by the R135 gene, which was also shown to be one of the most abundant elements of the surface fibrils that cover the capsid (3, 4). Fibrils are important structures for adhesion to the surface of amoebas (5). For instance, APMV fibrils are often found with one of their ends free and associated with a globular terminal, while the other is attached to a single central structure (6). In addition, they appear to be associated with a peptidoglycan-like structure, exhibiting successive rings of density (6). Moreover, in addition to R135, the L829 and L725 proteins have also been associated with APMV fibril composition, as the lack of those genes in the genome of mimivirus M4 was linked to an almost-fibrilless phenotype of M4 isolate particles (7). To date, the M4 isolate was obtained after successive passages of APMV in amoebas under allopatric conditions, resulting in genome shrinking and a decrease in fibril abundance (7). Furthermore, the adhesion of mimivirus fibrils to the host membrane seemed to be mediated by sugars found on both the host surface and in the fibrils (e.g., mannose and N-acetylglucosamine), allowing the phagocytosis of the virus and subsequent entry to start the cycle (5). Additionally, mimiviruses have a star-shaped structure called a stargate at one of the vertices of their particle, and this structure is responsible for releasing the genome at the beginning of the cycle within the host and is the only region at the capsid uncovered by fibrils (2, 6).

Currently, mimiviruses belong to the genus Mimivirus, family Mimiviridae, order Imitervirales, class Megaviricetes, and phylum Nucleocytoviricota (8). Metagenomics studies indicate that mimiviruses are among the most abundant viral entities in different biomes worldwide (9, 10). Phylogenetic analyses considering hallmark genes have shown that mimiviruses seem to be divided into three main lineages, A, B, and C (11). Lineage A is represented by APMV and other viruses similar to it, such as the Brazilian isolate sambavirus (12). Lineage B consists of the moumouviruses and related viruses, and lineage C is composed of megaviruses such as Megavirus chilensis and other similar viruses (13, 14). Genetic analyses show that mimivirus lineages, despite being evolutionarily related, present important differences between some of their genes. In this context, the possible differences in the morphological organization of mimivirus fibrils have not been explored in a comparative way, considering the three lineages and/or their isolates, and is an open field for study and new discoveries.

In this work, we describe the isolation of a novel mimivirus belonging to lineage C, named Megavirus caiporensis. The microscopic and genomic analyses of this virus raised questions on how surface fibrils are organized in different mimivirus isolates and encouraged us to expand this study to 14 other mimivirus isolates. Taken together, our data demonstrate at least three main patterns of fibril organization among isolates from different lineages. Considering the viruses analyzed here, evolutionarily-related isolates present similar fibril profiles. Our data also indicate that, in contrast to previous speculations (7), even the almost-fibrilless isolates (e.g., Borely moumouvirus) can have the R135, L829, and L725 genes in their genomes.

RESULTS

Megavirus caiporensis: isolation of a new mimivirus in Brazil.

As part of our continuous efforts to discover new giant viruses of amoebas, a new virus was isolated from a water sample collected from an urban lagoon in Belo Horizonte, Brazil. After observing cytopathic effect in the microplate well where this sample was inoculated, we proceeded with characterization of this isolate. By light microscopy, we observed the development and progression of the infection cycle as cytopathic effects emerged and intensified. At 6 h postinfection (hpi), the cells began to round and detach from the monolayer. At 12 hpi, fully rounded cells began to undergo lysis, and at 24 hpi almost all cells were lysed and cellular debris was observed in the supernatant and at the bottom of the flasks (Fig. 1A).

FIG 1.

Megavirus caiporensis viral particle and cycle. (A) Cytopathic effects on A. castellanii amoeba cells infected with Megavirus caiporensis. For the uninfected cell control, it is possible to observe the cells presenting irregular shapes, vacuoles, and good adhesion forming a monolayer, characteristics that allowed us to verify a healthy culture. From 6 h postinfection (hpi) with the megavirus that we isolated, the cells had already begun to undergo rounding and loosening of the monolayer, as seen in abundance in the supernatant. After 12 hpi, complete rounding and the beginning of cell lysis were observed, which significantly increased after 24 hpi, at which time few cells were still intact. Images were obtained using 100× magnification. (B and C) Characteristics of the Megavirus caiporensis viral particle. The Megavirus caiporensis capsid is composed of multiple layers involving the genome, a central and darker region, and a layer of fibrils covering the structure (B). The SEM image of a Megavirus caiporensis particle is shown (C). (D to G) Megavirus caiporensis multiplication cycle stages. (D) Assembly of new Megavirus caiporensis particles, beginning with the filling of the crescent-shaped form with the material present in the viral factory (VF). (E) Megavirus caiporensis viral particle receiving the internal contents through the opposite end (black arrow) to the stargate (orange arrow). (F) Viral factory producing new particles and the fibril acquisition area highlighted in purple. (G) The mature viral factory (VF) occupies a large part of the host cytoplasm of infected amoebas in full production of new Megavirus caiporensis particles.

Images obtained by transmission and scanning electron microscopy (TEM and SEM, respectively) allowed us to identify the isolate as a mimivirus (Fig. 1B and C), due to the morphologic characteristics of its particles, composed of capsids with an average size of 435 nm, covered by a layer of fibrils of 108 nm, and a total size of 651 nm (average). Interestingly, compared to what has been described for APMV (6), this isolate exhibited fibrils in a different organization, forming small clumps (Fig. 1B and C). Fibrils seemed to be grouped from the point of insertion into the capsid to their outermost portions (Fig. 1B). Here, as the fibrils of this virus are similar to those observed for Megavirus chilensis (13) and considering the sequencing results on this isolate (discussed below), we named it Megavirus caiporensis.

TEM images revealed that the entire cycle of Megavirus caiporensis is similar to what has been previously described for other mimiviruses (15). For instance, the immature viral factory is formed in the amoeba cytoplasm after virus entry, likely by phagocytosis, initiating the replication cycle. After the formation and maturation of the viral factory, mitochondria can be observed in its periphery, as well as formation of new particles assembling by crescent-shape structures, which gradually increase in size and are filled by the content present in the factory (Fig. 1D and E). During the acquisition of fibrils at the so-called fibril acquisition area, the particles receive their internal content from the opposite side of the stargate (Fig. 1D to F). At the end of the cycle, mature factories occupy a large area in the cell cytoplasm, as described for other mimiviruses (15). The newly formed particles agglomerate into the host cytoplasm, and the factories decrease in size throughout the replication cycle (Fig. 1G). Finally, particle release occurs by cell lysis.

Megavirus caiporensis sequencing and phylogenetic analysis confirmed it belongs to mimivirus lineage C.

The genome of Megavirus caiporensis is a linear, double-stranded DNA molecule of 1,199,709 bp in length (Fig. 2A). Its genomic content is predicted to encode 1,098 genes (568 located on the negative strand and 530 located on the positive strand) with a coding density of 91.3% and a GC content of 25.4%. Among the predicted proteins, 723 had no known function and were considered uncharacterized. Most functional genes fit into the categories miscellaneous (i.e., domain and repeat proteins; diverse enzymes) (171), other metabolic functions (65), signal transduction regulation (63), and DNA recombination, replication, and repair (35) (Fig. 2B). B-family DNA polymerase, chaperone, helicases, and replication factors are some of the genes that are incorporated inside the category of DNA recombination, replication, and repair. Seven open reading frames (ORFs) with no detectable homology to other ORFs in a database were detected and were considered ORFans. In addition, genes related to translation, a trademark of mimiviruses, were identified, including three aminoacyl-tRNA-synthetases, four RNA transporters (Fig. 2A), and one GTP binding translation elongation/initiation factor. The Megavirus caiporensis genome presents 13 genes related to transcription and RNA processing, including RNA ligase and different subunits of DNA-directed RNA polymerase. The majority of genes predicted matched those also predicted for Powai Lake megavirus and Megavirus chilensis, including ankyrin repeat proteins (miscellaneous category), putative lipoprotein (other metabolic functions), and putative protein kinases (signal transduction regulation category). Genes associated with host-virus interactions, nucleotide metabolism, and virion structure and morphogenesis were also predicted for Megavirus caiporensis. Additionally, in order to investigate the evolutionary and phylogenetic relationship of Megavirus caiporensis with other Nucleocytoviricota, including mimiviruses, we also performed a phylogenetic analysis using the conserved gene that encodes family B DNA polymerase as a target (Fig. 2C). Phylogenetic construction confirmed that Megavirus caiporensis clustered with lineage C mimiviruses.

FIG 2.

Genomic features of Megavirus caiporensis. (A) Circular representation of the Megavirus caiporensis genome. Rings, from innermost to outermost, correspond to genome coordinates in kilobases (kb), GC skew, GC content, and predicted protein-coding sequences (CDS) on both the forward and reverse strands. The green arrowheads correspond to transfer RNAs. A colored subtitle is provided on the right side of the figure. The blue and purple boxes at right indicate the three aminoacyl-tRNA synthetases (aaRS) and four tRNAs identified for this virus, respectively. (B) A set of genes from Megavirus caiporensis, classified according predicted gene categories. A colored subtitle is provided on the right side of the figure. (C) Phylogenetic analysis of subfamily B DNA polymerase sequences of Megavirus caiporensis and Nucleocytoviricota-related viruses. A maximum-likelihood phylogenetic tree was constructed with sequences of subfamily B DNA polymerase of mimiviruses and other Nucleocytoviricota-associated viruses. In this phylogeny, the new isolate described here, Megavirus caiporensis, clustered with lineage C mimiviruses (highlighted in red). The scale bar indicates the rate of evolution.

Mimivirus isolates have at least three patterns of surface fibrils.

The isolation and characterization of Megavirus caiporensis intrigued us regarding its structural aspects, as we noticed a different organization of its fibrils, further expanding these analyses to lineage A and B mimiviruses. Thus, we selected APMV (lineage A) and Borely moumouvirus (lineage B), as they are two viruses for which the whole genomes have been sequenced, as well as their being available in our laboratory. In order to analyze those isolates’ fibrils, particles of APMV, Borely moumouvirus, and Megavirus caiporensis were observed by TEM and SEM (Fig. 3A). As for APMV, it was possible to observe, as previously described in the literature, that its fibrils are long and abundantly surround the capsid (Fig. 3A), herein termed pattern A. In comparison, Borely moumouvirus (lineage B) appeared to have fewer fibrils around its capsid and these were less homogeneously distributed (Fig. 3A); here, we termed this pattern B. Interestingly, some particles belonging to pattern B were almost fibrilless. Finally, the isolate described here, Megavirus caiporensis, which belongs to lineage C, had an abundant number of fibrils, as seen for lineage A, but they were organized in small groups, as clumps (Fig. 3A); this pattern is herein termed pattern C. In SEM images, the patterns described here for fibrils were maintained, and we also observed very clearly the aforementioned differences among these mimiviruses (Fig. 3B). For instance, APMV exhibited its surface with fibrils homogeneously distributed, while for Megavirus caiporensis fibrils appeared to be organized in groups or clusters. As for Borely moumouvirus, it is possible to see that the capsid surface exhibits a geometric appearance, which probably is linked with the lower number of fibrils that is usually observed in EM images. We also developed a protocol to estimate the relative abundance of surface fibrils in the particles of mimiviruses, and then we applied it to compare APMV, Borely moumouvirus, and Megavirus caiporensis (Fig. 3C). After calibration of average values, the analysis revealed that Megavirus caiporensis exhibited the highest relative average contrast, 552-fold higher than Borely moumouvirus, followed by APMV, for which the contrast was 394-fold higher than Borely moumouvirus (Fig. 3D). Megavirus caiporensis had the highest relative average contrast and this was possibly related not only to the number of fibrils but also to the clump’s organization, while Borely moumouvirus having the lowest contrast was expected, since it has fewer fibrils surrounding the capsid (Fig. 3D).

FIG 3.

Morphological differences of mimivirus fibrils according to lineages A, B, and C. (A) Different patterns of fibril organization observed by TEM. Mimivirus extern fibrils in TEM images are highlighted with different colors to show the discrepancies of the fibril abundances and organizations (APMV in green, Borely moumouvirus in blue, and Megavirus caiporensis in red). (B) Different patterns of fibril organization observed by SEM. SEM images show the fibrils on the surfaces of the particles of APMV, Borely moumouvirus, and Megavirus caiporensis, respectively. Scale bars are indicated for each image. (C) Analysis of mimivirus fibril densities. For this illustrative representation of the protocol used to analyze the densities of fibrils of different mimiviruses, we have provided images obtained by TEM for each mimivirus analyzed (I). The fibrils were highlighted in the images manually by using the drawing tool available in Microsoft PowerPoint version 2021, and the background image was erased, leaving only the fibrils that were highlighted. (III) For each of the six sides of the particles, the fibril area was covered with rectangles of the same size and thickness in ImageJ software (version v1.53k). (IV) From each rectangle we generated a graph showing curves referring to pixel values read within the demarcated region. (V) The values obtained were plotted in the form of a table in Excel version 2021 and analyzed. (D) Graph representing the densities of fibrils found for each of the mimiviruses analyzed in this work. The x axis represents the analyzed lineages and the y axis shows the densities of fibrils in pixels. Values were calibrated based on the lowest observed value. Megavirus caiporensis presented the highest value identified among the three (552 pixels), followed by APMV (394 pixels). Finally, Borely moumouvirus presented the lowest number among the three (1 pixel), probably due to its lower number of these structures.

It is important to mention that preparation for TEM can influence fibril appearance. To understand whether the different fibril organizations and abundances could be the result of possible artifacts of preparation, we concomitantly analyzed different combinations of mixed purified particles of APMV, Borely moumouvirus, and Megavirus caiporensis by TEM. Therefore, those different viruses were mixed in the same sample in trio or pairs (Fig. 4) and then analyzed. All three fibril patterns were observed in the trio preparation (lineages A + B + C) (Fig. 4A). Patterns A and B were observed in preparations containing a mix of particles belonging to lineages A and B (Fig. 4B), and no particles with clumped fibrils were observed. Patterns B and C were the only ones observed in preparations containing lineage B and C viruses (Fig. 4C). Finally, patterns A and C were the only ones observed in preparations containing lineage A and C viruses (Fig. 4D). Taken together, these results confirmed that the distinct patterns of fibril organization and abundances were not likely to be an artifact of analysis related to TEM sample preparation.

FIG 4.

TEM images of trios and pairs of purified mimivirus particles related to lineages A, B, and C obtained in one same sample. (A) Mimivirus trio in the same sample; TEM sample images contain mimivirus A + B + C. (B to D) Mimivirus pairs in the same sample, with TEM sample images for pairs A + B (B), B + C (C), and A + C (D). For the preparations, only purified viral particles were used in combinations in trios or pairs, with APMV representing pattern A; Borely moumouvirus for pattern B; and Megavirus caiporensis representing pattern C. Scale bars are indicated in the images.

Considering these previous results regarding different mimivirus organization and abundance of fibril patterns, we expanded our analyses to other mimivirus isolates obtained by our group in the past 10 years. To date, all viruses isolated in our lab have been inspected by TEM, and their images are stored in a database. We searched for isolates with the three aforementioned fibril patterns (Fig. 2). We selected 15 isolates, 5 of each pattern (Fig. 5), and sequenced their family B DNA polymerase gene, which is considered a hallmark gene for differentiating the three mimivirus lineages. Sequences were used to construct a data set and a phylogenetic tree considering APMV, Borely moumouvirus, and Megavirus chilensis (13) as references for lineages A, B, and C, respectively. The results revealed a corresponding match among the three lineages and the predicted fibril morphotypes (Fig. 5 and 6): (i) lineage A, fully covered by fibrils (pattern A), including APMV, Mimivirus PU, Mimivirus capivarensis, Kroon mimivirus, and Amazonia mimivirus; (ii) lineage B, almost fibril-less (pattern B), including Borely moumouvirus, Moumouvirus crenensis, Moumouvirus dionensis, Moumouvirus limneidensis, and Moumouvirus naiadiensis; and (iii) lineage C, with fibrils in clumps (pattern C), including Megavirus caiporensis, Megavirus curupirensis, Megavirus botiensis, Megavirus boitataensis, and Megavirus muiraquitaensis. Taken together, our results suggested that, considering our collection of isolates, each lineage is related to a pattern of fibril morphotype. A comprehensive search for TEM images of mimivirus isolates in the literature also revealed that at least two of the three patterns of fibrils presented here have already been observed in natural isolates: those with particles fully covered by fibrils (pattern A) (16–18) and those with clumped fibrils (pattern C) (13, 19–21). Although TEM images of lineage B mimivirus particles are scarce in the literature, there is evidence that some moumouviruses have long abundant fibrils, similar to that described for APMV and other lineage A mimiviruses (22), such as those presented here. Therefore, although our isolates indicated a possible correspondence between patterns of fibrils and mimivirus lineages, this feature may not be expandable to all isolates and requires further investigation.

FIG 5.

Comparative panel of several mimivirus fibril morphologies related to patterns A, B, and C. Each column represents a pattern of fibril organization for each mimivirus lineage (A, B, or C), exhibited in five different isolates each. Pattern A mimiviruses (first column) fibrils are long and abundantly and homogeneously distributed around the capsid. Pattern B mimiviruses (second column) exhibit many fewer fibrils that are not uniformly distributed, and for some isolates the number of fibrils was too low to be captured by TEM images. Finally, for pattern C mimivirus (third column), the fibrils are organized surrounding the capsid, forming clumps. Some of those fibril clusters are so closely knit that the particle seems to have fewer fibrils (Megavirus muiraquitensis and Megavirus botiensis) compared to others, implicating an interspecific variance between pattern C viruses.

FIG 6.

Phylogenetic construction of mimiviruses of patterns A, B, and C. Phylogenetic trees were constructed based on the DNA polymerase gene of subfamily B of 15 mimiviruses, presented previously in a comparative panel of fibril organization patterns. To investigate the history of the relationship of each mimivirus with one lineages, A, B, or C, a representative previously related to each lineage was chosen: APMV for lineage A, Borely moumouvirus for lineage B, and Megavirus chilensis for lineage C. Our results suggested that different fibril patterns are lineage-associated, considering our analyzed isolates.

As aforementioned, the M4 isolate was obtained after successive passages of APMV in amoebas under allopatric conditions, resulting in a different set of phenotypes, including almost fibril-less particles. Proteomic analysis of M4 particles indicated the absence of a GMC-oxidoreductase encoded by the APMV R135 gene, while this gene has been reported as a main component of mimivirus fibrils (4, 6). To evaluate this correlation between fibrils and R135, we investigated the presence of R135 orthologues in the Borely moumouvirus genome. Interestingly, although Borely moumouvirus particles have a pattern similar to the M4 isolate, we found an R135 orthologue in its genome with 68.2% similarity to APMV R135, considering predicted amino acid sequences. Other APMV genes previously reported to be related to fibrils but absent in the M4 genome, such as L829 and L725, were also found in the Borely moumouvirus genome, with 62.9% and 54.7% similarity to the APMV predicted proteins, respectively (Table 1). We also observed that those three genes were widespread in most mimiviruses, which were clustered according to their respective lineages (A, B, or C) regardless of their fibrils pattern (Fig. 7A to C).

TABLE 1.

Similarity values of the best hits found by the BLASTp tool from amino acid sequences of proteins R135, L829, and L275^a

Virus	% similarity with protein:
	R135	L829	L725
APMV	100	100	100
Mimivirus reunion	99.6	99.8	100
Acanthamoeba castellanii mamavirus	99.6	99.8	100
Cotonvirus japonicus	68.4	61.3	66
Borely moumouvirus	68.2	62.9	54.7
Moumouvirus australiensis	68	61.1	55.1
Moumouvirus maliensis	68	62.5	54.7
Moumouvirus monve	68.5	62.2	55.2
Moumouvirus goulette	68.5	61.2	54.3
Acanthamoeba polyphaga moumouvirus	68.5	62.2	55.6
Saudi moumouvirus	68.5	62.3	55.6
Megavirus caiporensis	67.4	59	56.5
Bandra megavirus	68.3	58.5	56.5
Megavirus courdo7	68	58.8	56.5
Megavirus chilensis	68.1	58.3	56.5
Megavirus vitis	68	58.8	56.5
Megavirus lba	68	59	56.5
Powai Lake megavirus	67.8	60.1	56.5

^aAPMV was used as the reference.

FIG 7.

Phylogenetic constructions based on amino acid sequences of proteins associated with mimivirus fibril structures, based on proteins R135 (A), L829 (B), and (C) L275. Our results suggest that those genes are widespread in mimivirus isolates, regardless the lineage.

Fibrils may affect particle incorporation by amoebas.

Our group previously demonstrated that mimivirus fibrils may play an important role in viral adhesion to host amoebas (5). To understand whether the pattern of fibrils could affect the adhesion and entry of the particles in A. castellanii, we performed a set of biological assays with viruses presenting patterns A, B, and C. Analysis of TEM images showed pattern A and C particles binding to host plasma membrane, mediated by surface fibrils (Fig. 8A), but no pattern B particles were visualized attached to host cells. To quantitatively evaluate particles adhesion to amoebas, cells were infected at a multiplicity of infection (MOI) of 10, and after 60 min the supernatants were collected and titers were determined. The titration of the samples collected 1 hpi revealed approximately 10 times more Borely moumouvirus infectious particles than APMV and Megavirus caiporensis, suggesting that Borely moumouvirus particles (pattern B) were less able to be associated to the amoeba surface than pattern A and C viruses (Fig. 8B). To evaluate the incorporation of lineage A, B, or C isolate particles by amoebas, 50 cells were analyzed by TEM at 30 min pi for each isolate. It is important to mention that these results should be analyzed with care, considering that TEM was performed on a given section of biological samples. Nevertheless, these results are worthwhile to be presented since cells infected by APMV, Borely moumouvirus, or Megavirus caiporensis were analyzed under the same conditions. Here, 43 of 50 analyzed amoebas had at least two APMV particles internalized, and some cells incorporated a substantial number of APMV particles, up to 34 (Fig. 8C). As for Borely moumouvirus-inoculated sample, only 4 infected cells were observed, whereas 16 analyzed cells showed Megavirus caiporensis particles internalized (Fig. 8C). Despite those differences in adhesion and entry described for pattern A, B, and C viruses, we were able to observe particle uncoating (stargate opening) for APMV, Borely moumouvirus, and Megavirus caiporensis (Fig. 8D). Taken together, our results suggest that the pattern of fibrils may affect not only particles adhesion but also their incorporation by amoebas. Finally, the viral factories of viruses belonging to patterns A, B, and C were analyzed by TEM, and this revealed that, in spite of differences in fibril abundances and organization, factories from all viruses presented a fibrils acquisition area, suggesting that further studies on these structures are necessary (Fig. 8E).

FIG 8.

Fibril abundance seems to influence host-amoeba adherence and entry of pattern A, B, and C mimiviruses. (A) TEM images of pattern A (I) and C (II) mimiviruses attached to Acanthamoeba castellanii plasma membrane. No images were obtained for pattern B mimivirus. (B) Pattern B mimivirus particles are less incorporated by amoebas than mimivirus pattern A and C particles. Amoebas were infected by pattern A, B, or C at an MOI of 10. Sixty minutes postinfection, the supernatants were collected to verify the number of infectious particles not incorporated by amoebas, and then titers were determined. (C) Graphical representations of the relationship of the number of internalized particles within each amoeba (y axis) versus the number of cells (x axis) for APMV, Borely moumouvirus, and Megavirus caiporensis, respectively. The red circles indicate the number of cells with a specific number of particles internalized. Fifty TEM images of amoebas for each virus were selected randomly for analysis. (D) Mimivirus particles in denudation after entry into the host cell: APMV (I), Borely moumouvirus (II), and Megavirus caiporensis (III), indicating the continuation of the mimivirus cycle. (E) Fibril acquisition areas of mimiviruses related to lineages A, B, and C. Despite the differences in the patterns of fibrils among the three lineages, all three representatives presented a fibril acquisition area (I, APMV; II, Borely moumouvirus; (III), Megavirus caiporensis).

DISCUSSION

The isolation of a new lineage C mimivirus in Brazil highlights the diversity and ubiquity of mimiviruses. The role of fibrils for the adhesion of mimiviruses and their consequent success in invading their hosts to start the infection cycle has been reported previously (4–6). Findings from this current work corroborate these data, as we provided evidence that fibril abundance may be related to the entry of mimivirus into amoebas. It is important to highlight that lineage A includes the majority of isolates described worldwide, and most of the mimiviruses isolated by our team in Brazil are related to this lineage (23, 24). A possible explanation for this is the morphology of the fibrils presented by these viruses, which are found abundantly and homogeneously distributed throughout the viral capsid (Fig. 3 and 5). Additionally, the smaller number of isolates related to lineage B can also be related to the same hypothesis, since these viruses have fewer fibrils, which can decrease their abilities to adhere to their hosts and trigger phagocytosis and the consequent entry.

Additionally, regarding fibril morphology, it is also possible to notice an interspecific pattern of clustering of these structures within lineage C (Fig. 5, third column). The organization of the fibrils in clumps starts at the insertion from the capsid and continues until the tips facing the external environment. However, in some particles, these clumps were in smaller quantities and presented with more contrast in TEM images (Fig. 5, Megavirus botiensis and Megavirus muiraquitaensis), while they were more abundant and seemed to cluster fewer fibrils in others (Fig. 5, Megavirus caiporensis, Megavirus curupirensis, and Megavirus boitataensis). Understanding the reason for these differences within the same clade, searching for molecular differences that may explain them, and understanding the implications of these data are important perspectives to complement the novelties brought by this work.

Regarding their structure, we also showed that the fibril morphotypes were not the result of preparations for EM analysis but were an intrinsic characteristic of each viral particle pattern’s specificity (Fig. 4). Moreover, we found some evidence of molecular differences at important proteins for mimivirus fibrils, namely, R135, L829, and L725 (Fig. 7A to C), which may be closely related to the different phenotypes found among the three lineages. It is important to mention that the scarce number of images from lineage B mimivirus viral particles in the literature was a limiting factor in our study, making our findings concerning fibril patterns dependent on the isolates of each lineage to which we had access in our database. In addition, although it was clear that the pattern B remained among the isolates of our group, it is essential to attempt to cover future studies with more isolates of lineage B in order to affirm that this pattern can be maintained. Thereby, we propose that fibril morphological characteristics can be considered as indicators of mimivirus relationships to one of the A, B, or C lineages and that they can be complements to molecular traits for the characterization of these viral entities. Understanding the implications of different fibril morphotypes among mimivirus lineages can provide important information about the biology of these viruses and the way they relate to the environment and to their hosts, allowing further unraveling of these features within the family Mimiviridae.

MATERIALS AND METHODS

Virus isolation, multiplication, and purification.

In August 2017, 10 water samples were collected from Pampulha Lagoon, Belo Horizonte, Brazil. The collection was performed with sterile tubes, and the samples were stored at 5°C until the inoculation process. The samples were submitted to the process of prospecting and isolation of new giant viruses in Acanthamoeba castellanii amoebas, as previously described (25). From this collection of samples, we isolated nine amoeba viruses, including Megavirus caiporensis (mimivirus). These new isolates were registered at the Sistema Nacional de Gestão do Patrimônio Genético e do Conhecimento Tradicional Associado (SisGen), number ABF23CC. After the isolation, the virus was inoculated at a MOI of 0.01 in cell culture roller flasks containing 1.4 × 10⁷ Acanthamoeba castellanii cells and 35 mL of peptone-yeast extract-glucose (PYG) medium supplemented with penicillin (100 U/mL; Cellofarm, Brazil), streptomycin (100 μg/mL; Sigma-Aldrich, Burlington, MA, USA), and amphotericin B (0.25 μg/mL; Cultilab, Brazil). The cells were incubated at 30°C under slow rotation (0.2 rpm) on a roller. After the observation of cytopathic effect caused by viral infection (i.e., rounding cells and cellular lysis), the flask contents were collected. This content was subjected to freezing and thawing three times, to lyse the cells that remained intact and release viral particles. Then, it was ultracentrifuged (36,000 × g) in a 22% sucrose cushion for 50 min. The pellet containing purified viral particles was suspended in phosphate-buffered saline (PBS), and the viral titers were obtained by the endpoint method (26).

Electron microscopy.

Acanthamoeba castellanii cultures infected with different mimivirus strains (belonging to the lineages A, B, and C) and the respective isolated and purified mimiviruses were analyzed by SEM and TEM. Experiments and analyses were performed in the Center of Microscopy at the Universidade Federal de Minas Gerais, Belo Horizonte, Minas Gerais, Brazil (http://www.microscopia.ufmg.br).

For SEM assays, samples were fixed by immersion in a solution containing glutaraldehyde (2.5%) in 0.1 M sodium cacodylate buffer (pH 7.2) for 2 h. Following this, postfixation was performed for each sample with 2% osmium tetroxide (OsO₄) for 2 h at room temperature. Fixed samples were dehydrated using a growing series of ethanol dilutions (30%, 50%, 70%, 95%, and 100%) for 10 min in each step. Then, samples were dried with CO₂ at critical point using CPD 030 equipment (Bal-Tec, Liechtenstein). Next, samples were supported in aluminum stubs and metalized with a thin layer (5 nm) of gold particles using MED 020 equipment (Bal-Tec, Liechtenstein). Samples were observed in a FEG-Quanta 200 FEI microscope (FEI Co., Eindhoven, Netherlands) at 15 to 20 kV.

For TEM assays, samples were fixed by immersion in a solution containing glutaraldehyde (2.5%) in 0.1 M sodium phosphate buffer (pH 7.2) for 2 h. After fixation, postfixation was performed with a solution of 1% osmium tetroxide in sodium cacodylate buffer (0.1 M, pH 7.2) for 1 h followed by en bloc counterstaining with uranyl acetate (2% uranyl acetate in deionized water). Samples were gradually dehydrated by immersion in 70%, 80%, and 90% ethanol once for 15 min each and twice in 100% ethanol for 15 min. Next, samples were embedded in Epon resin. Ultrathin sections were obtained using an ultramicrotome with diamond knives (Leica Microsystems), and these sections had a thickness of 70 nm and were placed on a 200 mesh copper screen. The screens were counterstained with Reynold's lead citrate solution for 10 min. Images were obtained using a Tecnai G2-12-SpiritBiotwin FEI electron microscope (FEI Co., Eindhoven, Netherlands) at an acceleration voltage of 120 kV using a charge-coupled-device camera. In order to obtain the Megavirus caiporensis particle medium dimensions, the ImageJ software (version v1.53k, National Institutes of Health) was used to measure seven different particles during the acquisition of images. The measures were used to calculate the medium sizes of the particle and the fibrils.

Sequencing, assembly, and annotation.

The samples containing purified virus were sequenced using the Illumina MiSeq system, with a paired-end library and an Illumina DNA Prep kit (Illumina Inc., San Diego, CA, USA). The FastQC program was used to quality control of the obtained reads, which were trimmed using the Trimmomatic tool (27). For genome de novo assembly, we used Spades 3.12 with default parameters (28), and the obtained scaffolds were ordered based on a reference genome using MeDuSa online (29). The reference genome used was that of Powai Lake megavirus, obtained from the NCBI database (GenBank accession number KU877344.1). The GeneMarkS tool (30) was used to predict ORFs, considering only proteins that were larger than 50 amino acids. Additionally, the predictions of tRNA coding sequences were performed using Aragorn (31). The predicted ORFs were annotated using BLASTp (expect threshold, 10⁻³) against the NCBI non-redundant protein sequences (nr) database.

Phylogeny analysis.

Phylogenetic trees were constructed employing IQtree software (version 1.6.12) using the maximum-likelihood statistic method, with 1,000 bootstrap replicates as branch support (32). The mimivirus subfamily B DNA polymerase and the gene sequences for proteins R135, L829, and L725 were used. The data sets containing the sequences used for alignments were prepared using BLASTp (expected threshold, 10⁻³) against the NCBI non-redundant protein sequences (nr) database. The sequences selected were aligned by the Muscle algorithm executed by MEGAX (33, 34). The best-fit substitution models were selected by the ModelFinder algorithm implemented in IQtree (35), and the visualization and editing of the phylogenetic trees were carried out with MEGAX software and iTOL (34, 36).

Fibril morphology and density analysis.

In order to analyze differences between the fibrils’ morphology and density, we developed some specific protocols. To highlight the fibrils and their morphological differences, TEM images of three mimiviruses were used: APMV (representing lineage A), Borely moumouvirus (representing lineage B), and Megavirus caiporensis (the isolate described here, representing lineage C). First, fibrils were digitally highlighted with colors, emphasizing their peculiarities. For this, fibrils were marked with the aid of the “Marker” or “Watercolor” drawing tools of the Paint 3D (version 2021) program (Microsoft Corporation), with thickness settings between 2 and 4 pixels. Single colors were used for each mimivirus, with a maximum opacity of 35%, to highlight the fibrils with colors without losing their original shape.

For the density analysis, a protocol commonly used to measure the electrophoretic gel band densities was adapted (37). First, 70 TEM images for each one of the three mimiviruses representing lineages A, B, and C were individually highlighted with the tool “Scribble” of Microsoft PowerPoint version 2021 (Microsoft Corporation) very carefully (Fig. 3C, image I). Then, the background image was deleted, leaving only the outlines of the fibrils that were marked (Fig. 3C, image II). Those images were than used in the program ImageJ (version v1.53k, National Institutes of Health) to read the contrast and estimate the fibril densities for the 6 sides of each particle (previously selected, with many rectangles of same size), until the entire area presenting fibrils was fully covered by them (Fig. 3C, image III). After the selection of areas to be measured, the tool plot lanes were used to estimate the contrast from generated graphics (Fig. 3C, image IV). Each rectangle generated a graph with values related to the pixels detected in the analyzed images. These values were then used to interpret the mean contrast values for each of the six sides of the particles, and a total mean value was calculated, considering the standard deviation. The mean values were then normalized to be plotted in graph form (Fig. 3C, image V).

Evaluation of the relationship between fibrils and entry at the host cell.

To assess whether the number of fibrils was associated with the success of mimiviruses to adhere and consequently penetrate the host amoebas, we performed A. castellanii amoeba infection assays. The same representatives of lineages A, B, and C used previously were used in this experiment. Cell culture flasks (25 cm²; Thermo Fisher Scientific, USA) containing 1 × 10⁶ A. castellanii amoebas were infected with APMV, Borely moumouvirus, or Megavirus caiporensis at an MOI of 10. After infection, 30 min of adsorption was carried out at 30°C, with the flasks being gently shaken every 10 min to ensure that the inoculum covered the entire adhered cell monolayer. After adsorption, the inoculum was removed and the monolayer was washed 2 times with PBS, followed by the addition of 5 mL PYG medium to loosen the monolayer and collect the contents of the flasks. We proceeded with sample preparation for TEM analysis as described above. During imaging, we randomly selected 50 amoebas for each of the samples and counted the number of particles found inside each of the cells, to later analyze and compare the results.

We also tested the inoculums of infections performed in 96-well plates containing 4 × 10⁴ amoebas per well, with inoculums at an MOI of 10 using APMV, Borely moumouvirus, and Megavirus caiporensis. After 1 h of adsorption, the inoculums were collected and the viral titers were obtained by the endpoint method (26).

Data availability.

The genome of Megavirus caiporensis is available at GenBank under accession number OP925046.