Comparative Genomics of Chrysochromulina Ericina Virus and Other Microalga-Infecting Large DNA Viruses Highlights Their Intricate Evolutionary Relationship with the Established Mimiviridae Family

Abstract

Chrysochromulina ericina virus CeV-01B (CeV) was isolated from Norwegian coastal waters in 1998. Its icosahedral particle is 160 nm in diameter and encloses a 474-kb double-stranded DNA (dsDNA) genome. This virus, although infecting a microalga (the haptophyceae Haptolina ericina, formerly Chrysochromulina ericina), is phylogenetically related to members of the Mimiviridae family, initially established with the acanthamoeba-infecting mimivirus and megavirus as prototypes. This family was later split into two genera (Mimivirus and Cafeteriavirus) following the characterization of a virus infecting the heterotrophic stramenopile Cafeteria roenbergensis (CroV). CeV, as well as two of its close relatives, which infect the unicellular photosynthetic eukaryotes Phaeocystis globosa (Phaeocystis globosa virus [PgV]) and Aureococcus anophagefferens (Aureococcus anophagefferens virus [AaV]), are currently unclassified by the International Committee on Viral Taxonomy (ICTV). The detailed comparative analysis of the CeV genome presented here confirms the phylogenetic affinity of this emerging group of microalga-infecting viruses with the Mimiviridae but argues in favor of their classification inside a distinct clade within the family. Although CeV, PgV, and AaV share more common features among them than with the larger Mimiviridae, they also exhibit a large complement of unique genes, attesting to their complex evolutionary history. We identified several gene fusion events and cases of convergent evolution involving independent lateral gene acquisitions. Finally, CeV possesses an unusual number of inteins, some of which are closely related despite being inserted in nonhomologous genes. This appears to contradict the paradigm of allele-specific inteins and suggests that the Mimiviridae are especially efficient in spreading inteins while enlarging their repertoire of homing genes.IMPORTANCE Although it infects the microalga Chrysochromulina ericina, CeV is more closely related to acanthamoeba-infecting viruses of the Mimiviridae family than to any member of the Phycodnaviridae, the ICTV-approved family historically including all alga-infecting large dsDNA viruses. CeV, as well as its relatives that infect the microalgae Phaeocystic globosa (PgV) and Aureococcus anophagefferens (AaV), remains officially unclassified and a source of confusion in the literature. Our comparative analysis of the CeV genome in the context of this emerging group of alga-infecting viruses suggests that they belong to a distinct clade within the established Mimiviridae family. The presence of a large number of unique genes as well as specific gene fusion events, evolutionary convergences, and inteins integrated at unusual locations document the complex evolutionary history of the CeV lineage.

Keywords: Aureococcus anophagefferens virus; Chrysochromulina ericina virus; Haptolina ericina virus; Megamimivirinae; Mesomimivirinae; Mimiviridae; Phaeocystis globosa virus; nucleocytoplasmic virus.

FIG 1

Phylogeny of the concatenation of MCP, DNAPol B, and DNA packaging ATPase. The phylogenetic tree was built using PhyML (58) based on multiple alignments generated using Expresso (60). The tree was rooted with Ectocarpus siliculosus virus (Phaeovirus). Representatives of the genera Mimivirus (blue) and Cafeteriavirus (red) have been included, as well as the three fully sequenced alga-infecting Mimiviridae relatives (green). The tree was drawn using MEGA7 (61).

FIG 2

Venn diagram indicating the global proximity in gene content of CeV, its two closest relatives, PgV and AaV, and one member of each genus of the family Mimiviridae (Cafeteriavirus genus, CroV; Mimivirus genus, Mimi). The numbers in parentheses correspond to the raw number of encoded proteins without a homolog in the four other viruses. The numbers without parentheses indicate how many distinct clusters they constitute. The analysis was driven using OrthoMCL software (20), with a 10⁻⁵ E-value threshold and 1.5-mcl inflation parameter.

FIG 3

Duplication of the second largest subunit of the DNA-directed RNA polymerase II (RPB2) in CeV, PgV, and AaV. A maximum likelihood phylogenetic tree (58) of RPB2, aligned using Tcoffee (62) (Mcoffee mode), was constructed. Statistical branch supports (in percentages) for SH-like local support tests are given beside nodes. This phylogeny strongly supports a separate lineage leading to CeV, AaV, and PgV.

FIG 4

Relationship of the two major capsid protein homologs (MCP) found in CeV, PgV, and AaV. A maximum likelihood phylogenetic tree (58) of MCP, aligned using Tcoffee (62) (Expresso mode), was constructed. Statistical branch supports (in percentages) are given beside nodes.

FIG 5

Independent MIGE spreading in CeV and PgV. A maximum likelihood phylogenetic tree (58) of MIGE protein sequences, aligned using Muscle (57), was constructed. Statistical branch supports (in percentages) for SH-like local support tests are given beside nodes.

FIG 6

Helicase-type intein inserted in the CeV Lon peptidase. Multiple alignment, computed with Muscle (57), of the 8 blocks (numbered below the alignment) characteristic of inteins (Table 3). The intein hosted by the CeV GDP-mannose 4,6-dehydratase is included for comparison.

FIG 7

Extein DNA sequences in different genes hosting similar intein alleles (Fig. 6) are not similar.

FIG 8

Convergent acquisitions of host LIL proteins. A maximum likelihood phylogenetic tree (58) of LIL proteins, aligned using Muscle (57), was constructed. Statistical branch supports (in percentages) for the S-H-like local support tests are given beside nodes. Branches corresponding to viruses are colored red, with blue for haptophyceae and green for prasinophyceae. Sequences used for the analysis were selected using BLAST Explorer as implemented in Phylogeny.fr (63).

FIG 9

Convergent acquisitions of host HSP70 proteins. A maximum likelihood phylogenetic tree (58) of HSP70, aligned using Muscle (57), was constructed. Statistical branch supports (in percentages) for S-H-like local support test are given beside nodes. Viral branches are colored red, with blue for haptophyceae and green for prasinophyceae.