Functional Genomic Analyses Reveal an Open Pan-genome for the Chloroviruses and a Potential for Genetic Innovation in New Isolates

ABSTRACT

Chloroviruses (family Phycodnaviridae) are large double-stranded DNA (dsDNA) viruses that infect unicellular green algae present in inland waters. These viruses have been isolated using three main chlorella-like green algal host cells, traditionally called NC64A, SAG, and Pbi, revealing extensive genetic diversity. In this study, we performed a functional genomic analysis on 36 chloroviruses that infected the three different hosts. Phylogenetic reconstruction based on the DNA polymerase B family gene clustered the chloroviruses into three distinct clades. The viral pan-genome consists of 1,345 clusters of orthologous groups of genes (COGs), with 126 COGs conserved in all viruses. Totals of 368, 268, and 265 COGs are found exclusively in viruses that infect NC64A, SAG, and Pbi algal hosts, respectively. Two-thirds of the COGs have no known function, constituting the “dark pan-genome” of chloroviruses, and further studies focusing on these genes may identify important novelties. The proportions of functionally characterized COGs composing the pan-genome and the core-genome are similar, but those related to transcription and RNA processing, protein metabolism, and virion morphogenesis are at least 4-fold more represented in the core genome. Bipartite network construction evidencing the COG sharing among host-specific viruses identified 270 COGs shared by at least one virus from each of the different host groups. Finally, our results reveal an open pan-genome for chloroviruses and a well-established core genome, indicating that the isolation of new chloroviruses can be a valuable source of genetic discovery.

IMPORTANCE Chloroviruses are large dsDNA viruses that infect unicellular green algae distributed worldwide in freshwater environments. They comprise a genetically diverse group of viruses; however, a comprehensive investigation of the genomic evolution of these viruses is still missing. Here, we performed a functional pan-genome analysis comprising 36 chloroviruses associated with three different algal hosts in the family Chlorellaceae, referred to as zoochlorellae because of their endosymbiotic lifestyle. We identified a set of 126 highly conserved genes, most of which are related to essential functions in the viral replicative cycle. Several genes are unique to distinct isolates, resulting in an open pan-genome for chloroviruses. This profile is associated with generalist organisms, and new insights into the evolution and ecology of chloroviruses are presented. Ultimately, our results highlight the potential for genetic diversity in new isolates.

INTRODUCTION

Aquatic viruses are an integral part of Earth’s ecosystems, and they play a major role in maintaining global geochemical cycling. It is estimated that 10²³ virus infections occur each second in ocean waters. These infections are responsible for killing approximately 20% of the marine microbial biomass per day, cycling nutrients, and even changing local weather patterns (1). Metagenomic studies indicate that after bacteriophages, phycodnaviruses are one of the most abundant viral groups in the oceans (2–4). The Phycodnaviridae family currently comprises six genera (Coccolithovirus, Phaeovirus, Prasinovirus, Prymnesiovirus, Raphidovirus, and Chlorovirus) of genetically diverse, large, double-stranded DNA (dsDNA) viruses. Even though the majority of studies have been conducted on marine environments, a plethora of viruses that infect eukaryotic green algae are also present in terrestrial water environments worldwide (5–8). Among the phycodnaviruses identified in inland waters, representatives of the genus Chlorovirus have been isolated from different locations across five continents but mainly in North America and Europe, likely due to sampling bias. They infect unicellular eukaryotic algae, commonly referred to as zoochlorellae, that normally exist as symbionts associated with the protozoon Paramecium bursaria and other aquatic invertebrates, including Hydra viridis and Acanthocystis turfacea (9, 10). Chloroviruses possess icosahedral capsids with an internal lipid membrane and linear dsDNA genomes ranging from 290 to 370 kb in length that have up to 400 coding sequences (CDSs). Of the predicted CDSs, approximately 50% resemble proteins of annotated function; however, a large number of putative proteins remain uncharacterized (5, 9).

Phylogenetic analyses indicate that chloroviruses diverged into three monophyletic clades according to the chlorella-like algal host species that they infect, Chlorella variabilis NC64A (NC64A viruses), first isolated from a paramecium in the United States; Micractinium conductrix Pbi (Pbi viruses), first isolated from a paramecium in Europe; and Chlorella heliozoae SAG 3.83 (SAG viruses), isolated from a heliozoon in Europe (9, 11). The core protein family set consists of about 125 members common to all chloroviruses, indicating that they are essential for viral replication; however, many of these core proteins are of unknown function. Besides the conserved genes, some clusters of orthologous groups of genes (COGs) associated with protein families are restricted to each of the three individual clusters in one of the algal hosts; thus, they are not present in the other chloroviruses, and they could be responsible for host recognition and specificity (9). While little is known about their functions and the mechanisms underlying host specificity, it is known that the inability of viruses to attach to nonhost cells is a major factor in limiting chloroviruses’ host range (12).

The identification of conserved genes and those specific for each group of viruses is important for the advancement of genomic and evolutionary studies. Furthermore, such studies might provide important insights into the genetic innovation within each group. The Chlorovirus conserved genes include a subset that is in clusters (3 to 11 members) of collinear monocistronic genes, referred to as gene gangs (13). There are dozens of chlorovirus isolates described and genetically characterized so far; however, a better comprehension of the genome evolution of these viruses on a large scale is still missing. In this sense, pan-genomic analyses are valuable for understanding the genomic features under selective pressure and the evolutionary history of chloroviruses. In this study, we performed a functional pan-genomic analysis of chloroviruses, including viruses isolated from different places on Earth, infecting the three different algal hosts. We observed a distinct proportion of COG prevalences between the pan-genome and core genome for a few functional categories of genes, and a network construction evidenced COG sharing among host-specific viruses. Finally, our results revealed that chloroviruses have an open pan-genome and a well-established core genome; this observation indicates the potential for genetic innovation in new chlorovirus isolates.

RESULTS

Phycodnaviridae phylogeny.

The Phycodnaviridae family currently comprises 6 genera, and a total of 33 virus species are officially recognized by the International Committee on Taxonomy of Viruses (ICTV) (14). This family is the only one assigned to the newly created Algavirales order. Of the phycodnaviruses, chloroviruses are the most studied, and Chlorovirus is the genus with the largest number of sequenced representatives, currently comprising more than 50% of all viruses listed in the Phycodnaviridae. The phycodnaviruses are proposed to have a common origin with the whole group of nucleocytoplasmic large DNA viruses (NCLDVs) (15–17). Phylogenetic reconstruction of phycodnaviruses based on the DNA polymerase B gene, a well-conserved gene among members of the phylum Nucleocytoviricota, shows a clear separation of the six genera with high bootstrap support, with Chlorovirus forming a sister clade to Prasinovirus (Fig. 1). It is important to note that despite being monophyletic groups, viruses in different genera have high genetic diversity, as noted by the long branches of each clade, suggesting a particular evolutionary history for each group of viruses after its origin. A deeper analysis of the complete genome and DNA polymerase B gene shows intragenus conservation but great intergenus diversity (Fig. 2 and 3).

FIG 1.

Phylogeny of Phycodnaviridae. A phylogenetic reconstruction using amino acid sequences of the DNA polymerase B family from different members of the family Phycodnaviridae is shown. Representatives of all six currently established genera were included. Only bootstrap values of >70 are shown. Chloroviruses infecting different host cells are colored as follows: green, Chlorella variabilis; purple, Micractinium conductrix; yellow, Chlorella heliozoae. Gray bars indicate the number of CDSs for each chlorovirus isolate available in GenBank. The bar indicates the rate of evolution.

FIG 2.

Similarity range of chlorovirus genomes. Shown is a heat map of identity analysis of the complete genomes of chloroviruses included for phylogenetic reconstruction. Identity ranges from 60% (yellow) to 100% (blue).

FIG 3.

Similarity range of chlorovirus DNA polymerase B genes. Shown is a heat map of identity analysis of DNA polymerase B genes of all Phycodnaviridae members included for phylogenetic reconstruction. Genera are indicated in roman numerals. I, Chlorovirus; II, Prasinovirus; III, Coccolithovirus; IV, Raphidovirus; V, Phaeovirus; VI, Prymnesiovirus. Identity ranges from 20% (yellow) to 100% (blue).

Within the genus Chlorovirus, it was possible to observe high bootstrap support for the formation of three distinct clades (Fig. 1). These clades grouped viruses that infect the same host, namely, Chlorella variabilis NC64A, Micractinium conductrix Pbi, and Chlorella heliozoae SAG 3.83. These hosts were recently reclassified, and before that, they were commonly called NC64A, Pbi, and SAG, respectively, and are named as such throughout this work to allow for a more adequate comparison with other data in the literature. Interestingly, one of the viruses included in this analysis (NE-JV-1) appeared in a separate branch within the SAG clade, although it was isolated from Pbi (9). Finally, there was high gene diversity among the NC64A viruses because they separated into two clades, compared to the Pbi viruses and SAG viruses, which had an apparently lower evolution rate (Fig. 1); however, this may be biased due to a narrow sampling range of space and time.

Evolution of the chlorovirus pan-genome.

As mentioned above, chloroviruses have a large dsDNA genome, harboring hundreds of genes. A total of 42 chloroviruses have had their genomes sequenced and annotated and are available in public databases (last checked in May 2021). These viruses were isolated beginning in 1983 and have been collected from at least 11 countries (Fig. 4 and Table 1); however, the majority have been isolated from different sites in the United States. These viruses have genomes ranging from 287 to 369 kb and contain between 314 and 886 coding sequences (CDSs); they also contain 3 to 16 genes encoding tRNAs (Table 1). These data suggest large genomic diversity among chloroviruses. To gain a better understanding of which genes are well conserved in this group of viruses as well as a dimension on the evolution of viral genomic diversity, we gathered 36 viruses for a pan-genome analysis. Six of the 42 viruses with available genomes were excluded from the analysis because they have a much higher number of CDSs (Fig. 1 and Table 1), which would distort the analysis and may not adequately reflect the evolution of the viral pan-genome.

FIG 4.

Worldwide chlorovirus isolation sites. The size of the areas of the green circles corresponds to the number of chlorovirus isolates in each country, as identified in Table 1.

TABLE 1.

General information on chloroviruses with available genomes in a public database

Virus	Host species	Genome size (kb)	GC content (%)	No. of CDSs	No. of tRNAs	GenBank accession no.	Source of isolate	Yr of isolation
AN69C	Chlorella variabilis (NC64A)	332	40	362	10	JX997153	Canberra, Australia	1995
CviKI	Chlorella variabilis (NC64A)	308	40	336	14	JX997162	Kyoto, Japan	1990
CvsA1	Chlorella variabilis (NC64A)	310	40	342	14	JX997165	Sawara, Japan	1992
IL-3A	Chlorella variabilis (NC64A)	323	40	349	12	JX997169	Illinois, USA	1983
IL-5-2s1	Chlorella variabilis (NC64A)	344	41	379	8	JX997170	Illinois, USA	1986
KS1B	Chlorella variabilis (NC64A)	287	40	319	13	JX997171	Kansas, USA	2003
MA-1D	Chlorella variabilis (NC64A)	339	41	371	11	JX997172	Massachusetts, USA	1984
MA-1E	Chlorella variabilis (NC64A)	336	40	376	14	JX997173	Massachusetts, USA	1984
NE-JV-4	Chlorella variabilis (NC64A)	328	40	352	11	JX997179	Nebraska, USA	2008
NY-2B	Chlorella variabilis (NC64A)	344	41	371	8	JX997182	New York, USA	1986
NYs-1	Chlorella variabilis (NC64A)	348	41	381	7	JX997183	New York, USA	1984
AR158	Chlorella variabilis (NC64A)	344	41	814	6	NC_009899.1	Buenos Aires, Argentina	1997
NY-2A	Chlorella variabilis (NC64A)	369	41	886	7	NC_009898.1	New York, USA	1984
PBCV-1	Chlorella variabilis (NC64A)	330	40	802	11	NC_000852.5	New York, USA	1984
OSy-NE5	Chlorella variabilis (Syngen 2-3)	327	42	357	13	NC_032001.1	New York, USA	2012
AP110A	Micractinium conductrix (Pbi)	327	44	348	9	JX997154
Can18-4	Micractinium conductrix (Pbi)	329	45	357	10	JX997157	Canada	1995
CVA-1	Micractinium conductrix (Pbi)	326	45	346	9	JX997159	Amonau, Germany	1984
CVB-1	Micractinium conductrix (Pbi)	319	44	346	10	JX997160	Berlin, Germany	1984
CVG-1	Micractinium conductrix (Pbi)	318	45	333	9	JX997161	Gottingen, Germany	1984
CVM-1	Micractinium conductrix (Pbi)	327	44	341	9	JX997163	Marburg, Germany	1984
CVR-1	Micractinium conductrix (Pbi)	329	45	351	9	JX997164	Rauschenberg, Germany	1984
CZ-2	Micractinium conductrix (Pbi)	305	45	340	10	JX997166	Czech Republic	1995
Fr5L	Micractinium conductrix (Pbi)	302	45	345	11	JX997167	France	1995
NE-JV-1	Micractinium conductrix (Pbi)	326	45	337	3	JX997176	Nebraska, USA	2008
NW665.2	Micractinium conductrix (Pbi)	325	44	350	8	JX997181	Norway	1984
OR0704.2.2	Micractinium conductrix (Pbi)	313	45	344	7	JX997184	Oregon, USA	2007
FR483	Micractinium conductrix (Pbi)	321	44	849	9	NC_008603.1	France	1997
MT325	Micractinium conductrix (Pbi)	314	45	845	10	DQ491001.1	Montana, USA	1996
Br0604L	Chlorella heliozoae (SAG 3.83)	295	49	346	9	JX997155	Sao Paulo, Brazil	2006
Can0610SP	Chlorella heliozoae (SAG 3.83)	307	49	341	13	JX997156	British Columbia, Canada	2006
Canal-1	Chlorella heliozoae (SAG 3.83)	293	51	336	10	JX997158	Nebraska, USA	2008
GM0701.1	Chlorella heliozoae (SAG 3.83)	315	48	362	10	JX997168	Guatemala	2007
MN0810.1	Chlorella heliozoae (SAG 3.83)	327	52	343	9	JX997174	Minnesota, USA	2008
MO0605SPH	Chlorella heliozoae (SAG 3.83)	289	49	323	11	JX997175	Missouri, USA	2006
NE-JV-2	Chlorella heliozoae (SAG 3.83)	319	48	346	13	JX997177	Nebraska, USA	2008
NE-JV-3	Chlorella heliozoae (SAG 3.83)	298	49	334	12	JX997178	Nebraska, USA	2008
NTS-1	Chlorella heliozoae (SAG 3.83)	323	48	364	7	JX997180	Nebraska, USA	2008
OR0704.3	Chlorella heliozoae (SAG 3.83)	311	49	342	13	JX997185	Oregon, USA	2007
TN603.4.2	Chlorella heliozoae (SAG 3.83)	321	49	351	9	JX997186	Tennessee, USA	2006
WI0606	Chlorella heliozoae (SAG 3.83)	289	50	329	11	JX997187	Wisconsin, USA	2006
ACTV-1	Chlorella heliozoae (SAG 3.83)	288	49	860	11	NC_008724.1	Stuttgart, Germany	2002

It should be noted that all open reading frames (ORFs) with 64 or more codons were counted in the six viruses AR158, NY-2A, PBCV-1, FR483, MT325, and ATCV-1. These viruses have more than 800 CDSs, which is due to different strategies used for gene prediction and annotation (18–20). In contrast, the viruses with lower numbers of CDSs were selected by removing any ORFs that were inside larger ORFs because they were unlikely to be real protein-encoding genes. When this was done with the 6 viruses mentioned above, the number of likely authentic CDSs resembled those of the 36 viruses used in this study.

Our analysis of the 36 chloroviruses identified a total of 1,345 clusters of orthologous groups of genes (COGs) as components of the chlorovirus pan-genome (Fig. 5). With the addition of new viruses, new unique genes were identified, contributing to a constant increase in the viral pan-genome. At the same time, well-conserved genes were identified, making up the core genome of chloroviruses. A total of 543 COGs were singletons; that is, they had only one gene and were unique to a specific virus. Three hundred COGs were shared by 2 to 5 viruses, and 131 COGs were shared by between 6 and 10 viruses. Only 126 (9.36%) COGs were shared by all 36 chloroviruses, constituting the core genome of this group of viruses (Fig. 5; see also Data Set S1 in the supplemental material). Interestingly, the addition of genes identified in specific host viruses did not result in a significant increase or decrease in the pan-genome or core genome, respectively. A slightly greater slope in the pangenome curve was observed with the inclusion of Only-Syngen Nebraska virus 5 (Osy-NE5). Although this virus infects the same host cell at the species level, i.e., Chlorella variabilis, only the Syngen 2-3 strain supports virus replication, thus having a different virus-host relationship than the other NC64A viruses (21). Such specific interactions can be attributed to a greater genetic diversity of this virus, and this hypothesis will be tested in future studies. Our analysis showed an open pan-genome for chloroviruses and a well-conserved core genome for this group of viruses.

FIG 5.

Evolution of the pan-genome of chloroviruses. The stepwise inclusion of chloroviruses indicates an “open” pan-genome and a core genome of this group of viruses. Blue bars indicate the number of CDSs for each virus. The green line indicates the evolution of the gene content (pan-genome), and the red line indicates the evolution of conserved genes (core genome). Viruses are separated according to the host with which they are associated, indicated by shades of blue. The y axis shows the number of COGs, and the x axis shows the chlorovirus isolates.

Functional characterization of the pan-/core genome of chloroviruses.

Among the 1,345 COGs, 891 (66.2%) had no defined function, being considered hypothetical proteins with no known conserved domains (Fig. 6A). A total of 115 COGs do not currently have a well-established function (listed as miscellaneous), but conserved domains were identified in proteins from genes belonging to these clusters, such as domains containing glycoprotein repeats and ankyrin repeats, chitin binding domains, and PBCV-specific basic adaptor domain-containing proteins, among others. Thus, currently, approximately 75% of chlorovirus COGs do not have a well-defined function, constituting a vast field for future investigations. Among the COGs with known functions, most were related to DNA replication, recombination, and repair processes (123; 9.1%), including, among others, DNA polymerase B, DNA primase, DNA topoisomerase type II, DNA ligase, and various endonucleases containing GIY-YIG catalytic domains (Fig. 6A; Data Set S1). Forty-three COGs were associated with carbohydrate metabolism, including enzymes synthesizing extracellular matrix polysaccharides such as hyaluronan and chitin, enzymes synthesizing sugar nucleotides, enzymes responsible for the synthesis of the glycans attached to the virus major capsid proteins, and enzymes involved in algal cell wall degradation such as chitinase and chitosanase activities (22, 23). The other COGs were distributed into functions such as nucleotide (21; 1.6%), protein (16; 1.2%), and lipid (6; 0.4%) metabolism; transcription and RNA processing (20; 1.5%); signal transduction (16; 1.2%); integration and transposition (9; 0.7%); virus-host interaction (8; 0.6%); and virion structure and morphogenesis (25; 1.8%) (Fig. 6A).

FIG 6.

Functional characterization of the pan-genome and core genome of chloroviruses. (A) Distribution of all COGs according to different functional categories of genes; (B) distribution of COGs from the core genome (126 COGs) into functional categories, evidencing both raw numbers and percentages; (C) comparison of COG distributions in the pan-genome and core genome considering the different functional categories. Colors in panels B and C correspond to those depicted in panel A.

Regarding the core genome of chloroviruses composed of 126 COGs, almost half of these had no known function (60; 47.5%), evidencing a large number of genes with homologues present in all known chloroviruses that may have important functions for viral biology (Fig. 6B). Among the COGs with known function, 11 were related to DNA replication, recombination, and repair, including 2 genes conserved in other NCLDVs, such as DNA primase and DNA polymerase, and another 11 related to RNA transcription and processing, including, among others, the well-conserved VLTF3 (Fig. 6B; Data Set S1). Six COGs were related to nucleotide and protein metabolism. Sixteen COGs were associated with virion structure and morphogenesis, with genes corresponding to the major capsid protein, the A32-like ATPase, postulated to be involved in the encapsidation of the viral genome, and also the penton protein and 13 other minor capsid proteins, previously identified as P2 to P14 in PBCV-1 (24) (Fig. 6B; Data Set S1). Only two COGs related to carbohydrate metabolism were part of the core genome of chloroviruses, a chitinase and a multidomain glycosyltransferase, a protein with three distinct domains, each with a specific activity (25). Furthermore, a gene encoding a lipid hydrolase was present in all chloroviruses, being the only COG related to lipid metabolism composing the core genome of the viruses (Fig. 6B; Data Set S1). No genes related to integration and transposition or virus-host interaction were conserved in all chloroviruses, suggesting a high specificity of these genes for their respective viruses.

In general, there was a proportional correspondence in the numbers of COGs in the different functional categories between the pan-genome and the core genome (Fig. 6C). Interestingly, COGs related to RNA transcription and processing, protein metabolism, and virion structure and morphogenesis were at least 4-fold more represented in the core genome than in the pan-genome, suggesting that these genes and processes are essential for the chloroviruses (Fig. 6C). Finally, we identified 390 genes in the prototype chlorovirus PBCV-1 with homologues in at least one of the 36 chloroviruses included in this analysis. A total of 141 of the 148 virus-encoded proteins incorporated into the PBCV-1 particle were assigned to a COG (Data Set S1). Among these proteins, 54 (35.5%) fell into COGs shared by all 36 chloroviruses, including the major capsid protein (A430L), the penton protein (A310L), and 13 minor capsid proteins, named P2 to P14, identified by cryo-electron microscopy (cryo-EM) analysis (24) (Data Set S1).

Nearly 60% (754 of 1,345 COGs) of the chloroviruses’ pan-genome is unique to this group of viruses (i.e., the proteins have no homologues outside chloroviruses), while the other 40% possibly originated from Bacteria (293; 21.8%), Eukarya (160; 11.9%), viruses (128; 9.5%), and Archaea (10; 0.7%) (Fig. 7A). Most notably, more than 50% of the COGs with best hits with eukaryotic organisms have homologues in green algae from the Chlorellaceae and Chlorophyceae families (Fig. 7A). Among COGs of putative viral origins, most are likely from other members of the Phycodnaviridae or Mimiviridae family (Fig. 7A). When considering the core genome, 30.2% (38 of 126 COGs) is unique to chloroviruses, and most COGs have best hits with Bacteria (57; 45.2%), followed by viruses (16; 12.7%), Eukarya (13; 10.3%), and Archaea (2; 1.6%) (Fig. 7B). Only a small fraction of the most conserved COGs was inferred to originate from green algae (9), and more than 90% of the proteins with best hits with viruses have homologues in members of the Phycodnaviridae and Mimiviridae families (Fig. 7B).

FIG 7.

Putative origin of genes composing the pan-genome of chloroviruses. (A) Distribution of all COGs according to the best-hit analysis against the NCBI nr database excluding the chloroviruses; (B) distribution of COGs from the core genome (126 COGs) according to the best-hit analysis. The contributions of different groups of viruses and green algae to the evolution of the pan-genome are indicated. COGs with no hits in the analysis are referred to as “ORFans.”

COG sharing among host-specific viruses.

The 1,345 COGs were evenly shared among the 36 chloroviruses included in this analysis (Fig. 8). In our network graph, the presence of COGs shared among viruses from each of the three host groups was evident, and there were specific COGs present in only a single isolate (singletons), which, for the most part, had no known function (Fig. 8). A total of 270 COGs were shared by at least one specific virus isolated from each host species. Also, some COGs were found specifically in viruses that infect only one type of host, possibly being important in the direct interaction of these viruses with their hosts. NC64A viruses shared 73 COGs exclusively with Pbi viruses and shared only 32 with SAG viruses, while 69 COGs were shared exclusively between Pbi viruses and SAG viruses (Fig. 8). Totals of 368, 268, and 265 COGs were found exclusively in virus isolates that infect NC64A, SAG, and Pbi, respectively. Although the same number of isolates from each host was included in the analysis (12 isolates for each host), there were considerable differences in exclusive COGs among them, with a higher number for NC64A viruses, around 40% more, than for the viruses specific for the other two hosts. In total, 743 COGs were shared among NC64A viruses, 677 COGs were shared among Pbi viruses, and 639 were shared among SAG viruses.

FIG 8.

COG sharing among the group of chloroviruses. A bipartite network graph connects the 1,345 COGs to the 36 chloroviruses included in the pan-genome analysis. Larger nodes correspond to the viruses, with different colors corresponding to the specific host with which they are associated. Smaller nodes correspond to the COGs, and colors refer to different functional categories, as indicated in the legend of Fig. 3. The graph was generated using a force-based algorithm, and the nodes were then manually organized to allow better visualization of the connections, especially the singletons. Nodes in the middle of the graph are the COGs shared by at least one virus infecting one of the three distinct hosts.

Of the 891 uncharacterized COGs, 138 were present in at least one virus that infects each of the 3 host types, while 253, 208, and 200 COGs were shared exclusively among NC64A viruses, Pbi viruses, and SAG viruses, respectively (Fig. 9A). One-third of the COGs related to carbohydrate metabolism were shared among viruses of the 3 host species, including chitinase, chitosanase, and glycosyltransferase clusters, and it is particularly notable that around 30% of COGs were unique to NC64A viruses (Fig. 9B). Thirty-two COGs related to DNA replication, recombination, and repair were shared among viruses that infect the three types of hosts, while 61 were unique to viruses infecting a given host, with 32 shared exclusively among NC64A viruses, 15 shared among SAG viruses, and 14 shared among Pbi viruses (Fig. 5C). Unlike the functional categories of COGs mentioned above, for those related to RNA transcription and processing (20 COGs), 65% were shared among viruses of all hosts, 2 were unique to NC64A viruses, 3 were unique to SAG viruses, and there were no unique COGs shared among the Pbi viruses (Fig. 9D). Regarding the other functional categories, the sharing of COGs was slightly more homogeneous among the specific viruses for each host (Fig. 10). An exception would be the set of genes whose functions are not well elucidated, but there was the presence of conserved domains (miscellaneous), in which there was a greater proportion of these exclusive genes in NC64A viruses than in the others.

FIG 9.

Functional diversity of chlorovirus COGs according to specific host species. Venn diagrams evidence the number of COGs shared by viruses infecting different hosts according to distinct functional categories: uncharacterized (A); carbohydrate metabolism (B); DNA replication, recombination, and repair (C); and transcription and RNA processing (D).

FIG 10.

Venn diagrams evidencing the number of COGs shared by viruses infecting different hosts according to distinct functional categories. (A) Host-virus interaction; (B) integration and transposition; (C) lipid metabolism; (D) miscellaneous; (E) nucleotide metabolism; (F) other metabolic functions; (G) protein metabolism; (H) signal transduction regulation; (I) virion structure and morphogenesis.

DISCUSSION

The Phycodnaviridae family was officially created by the ICTV in 1990 to group the newly discovered chloroviruses, large dsDNA viruses associated with chlorella-like green algae (26). Currently, the Phycodnaviridae family is composed of 6 distinct genera, among which the genus Chlorovirus is the one with the largest number of recognized species and by far the most studied of the family. Recently, in a notable effort by the ICTV to create a viral megataxonomy (27), phycodnaviruses were added to a new order, Algavirales, which is included in the class Megaviricetes, phylum Nucleocytoviricota. Members of this phylum include other families of large DNA viruses that replicate in the cytoplasm and/or nucleus of their host cells, commonly called NCLDVs (15, 16). The NCLDVs have some well-conserved genes (e.g., DNA polymerase B), which point to a common origin, possibly prior to the emergence of eukaryotic organisms (28). Our phylogenetic analysis based on the DNA polymerase B gene, including distinct members of the Phycodnaviridae family, showed a clear separation between the different genera, with high statistical support. It is interesting to note, however, that there was a large evolutionary distance between viruses of different genera, which suggests a unique evolutionary path for each group of viruses. Identity analyses, both at the individual-gene level and at the complete-genome level, showed major differences between the different genera (below 60%). Such differences could be enough for a taxonomic restructuring of Phycodnaviridae, a point that has recently been analyzed for different members of the Nucleocytoviricota (29). Even within the genus Chlorovirus, it was possible to observe a separation into at least three distinct clades, which could be classified into different subgenera. Each clade corresponds to viruses infecting a specific host, which indicates that the most recent ancestor of each group already infected its respective host lineage before divergence (9).

Interestingly, our analysis placed the chlorovirus NE-JV-1 isolate in the SAG virus clade, although it was isolated from Micractinium conductrix Pbi (9). However, a phylogenetic reconstruction based on a concatenated alignment of 32 conserved chlorovirus genes places this isolate next to the Pbi virus clade (9). This suggests that the evolutionary history of a gene may not fully reflect the evolutionary history of a virus, something also observed for other biological entities (30).

Our analysis revealed an open pan-genome for chloroviruses, and unlike those observed for other virus groups such as Cedratvirus and Marseillevirus (31, 32), the addition of viruses from different genetic groups (possibly subgenera) did not result in an abrupt increase of the pan-genome or a significant decrease in the core genome, suggesting a more constant evolution for chloroviruses. This general theory assumes that the evolution of the pan-genome is shaped mainly by gene gain and loss by horizontal gene transfer (HGT) events (33, 34). Considering this scenario, one can assume that a considerable fraction of the chloroviruses’ pan-genome came from cellular organisms, especially Bacteria and Eukarya, based on best-hit analysis. Surprisingly, a considerable fraction of genes with best hits with eukaryotic organisms is related to green algae, a result in contrast to those previously reported (9). This could be due to differences in the parameters used in the analysis or in the database. The former is unlikely since similar strategies were used (i.e., a BLASTP search against the NCBI nr database with an E value of <10⁻⁵) (9). The latter is a better explanation, especially considering that new green alga genomes have been available recently. In fact, most of the best hits with green algae were related to Chlorella sorokiniana, Chlorella desiccata, and Micractinium conductrix, all organisms whose genomes became available after 2018. This update to the genomic database highlights the importance of new studies involving whole-genome sequencing of new green algae, which will profoundly benefit the limnology field but will also have a great impact on the virology field, allowing for an improvement in the understanding of chlorovirus evolution and virus-host interactions.

It is notable that nearly 60% of the genes that constitute the pan-genome are unique to chloroviruses, suggesting that these genes were possibly created de novo in the last common ancestor of chloroviruses (or in specific groups within Chlorovirus). De novo gene creation was proposed to make an important contribution to the evolution of the genomes of giant pandoraviruses, a group closely related to Phycodnaviridae (35). A more robust investigation in this regard based on the phylogenetic reconstruction of each of these genes could confirm this hypothesis and provide new insights into the origin and evolution of the chlorovirus pan-genome.

Interestingly, previous studies suggest that the evolution of large and giant DNA viruses occurs following an accordion-like model, wherein their ancestors were small viruses that evolved by gaining and losing genes throughout history through HGT events from their hosts and sympatric organisms (36–38). In this sense, our data supported this hypothesis for chloroviruses, wherein coevolution with their algal hosts would be critical for viral genome evolution. Moreover, the pan-genomic analysis assumes that each gene is separate (or interacts separately) from one another. However, certain chlorovirus genes exist in conserved clusters, referred to as “gene gangs,” so the entire cluster appears to be acting as a single unit (13). Therefore, the pan-genome evolution of chloroviruses may have been shaped by additional forces other than HGT, with some genes being conserved among distinct chloroviruses as gangs, given their pivotal importance to virus replicative success. In addition, organisms with an open pan-genome are usually characterized by having large population sizes, having diverse community interactions, and being niche generalists (34). Although it appears that the chloroviruses are strictly related to chlorella-like green algae, it is possible that these viruses are capable of infecting a broader host range, both now and in the past. This would allow diverse interactions with a plethora of organisms, thus contributing to the increase of the pan-genome, until the establishment of an optimal system for development, in this case, the modern chlorella-like green algal cells.

Chloroviruses have a large dsDNA genome encoding hundreds of proteins (5). On average, chloroviruses have around 350 protein-coding sequences. However, 6 of the 42 chloroviruses whose genomes are annotated and deposited in the databases have a far higher number of CDSs than this, ranging from 802 to 886 CDSs, which results in a major distortion in the gene annotation for some isolates. Although we did not include the prototype chlorovirus PBCV-1 in the pan-genome construction because of the inflated number of CDSs, we searched for homologues and associated them with each COG. Of the 802 CDSs predicted for PBCV-1, 390 were associated with a COG, some of them included in the same cluster, being considered paralogues. Most of the remaining CDSs in PBCV-1 probably do not encode proteins. It should be noted that the PBCV-1 genome was hand-curated after resequencing and comparing the virion proteome (39), allowing an appropriate adjustment for the initial gene prediction and annotation for this virus. However, the same approach has not been applied to the other virus genomes.

It is worth noting that of the 148 virus-encoded proteins present in the PBCV-1 virion (39), 141 had homologues in other chloroviruses and were assigned to a COG. Interestingly, among the remaining seven proteins, four were products of ORFan genes (i.e., genes found exclusively in PBCV-1). Among them is A256/257L, a 2,514-bp gene whose product is a 97-kDa protein (39). This unique protein must have an important role for the virus and should be investigated in further studies.

Much of the chlorovirus pan-genome can be considered a “dark pan-genome,” as nearly 70% of all COGs were composed of uncharacterized genes. This constitutes a major unexplored research field that could reveal important novelties of interest for the scientific community. A possible way to infer the functions of these genes would be through the “guilt by association” strategy (13). In that case, conserved collinearity of genes, or gene gangs, could provide useful insights for a functional classification of a gene that belongs to a functionally characterized gang, even though the exact function of that gene would require an in-depth experimental investigation. Considering that the biological function of this hypothetical gene could not be confidently assigned by this strategy, we decided to maintain the “hypothetical” status in our analysis. Even without a defined function, these genes must play an important role in the virus’s biology since they are expressed during the viral replication cycle, as evidenced by the transcriptomic data for PBCV-1, where 99% of the virus genome was covered by poly(A)⁺ mRNA reads (40). The exact mechanisms in the regulation of chlorovirus transcription are currently unknown, but viral promoter motifs are likely to have an influence, at least, for early gene expression (41, 42). These promoter motifs seem to be conserved among the different groups of chloroviruses, a characteristic that may have existed prior to the diversification of this group of viruses.

For the functional classification of COGs, we assigned only one major function to each COG, but we know that some chlorovirus genes have multidomains with specific functions. For example, the PBCV-1 A064R protein, initially described as a glycosyltransferase, has three functional domains: domain 1 is a β-l-rhamnosyltransferase, domain 2 is an α-l-rhamnosyltransferase, and domain 3 is a methyltransferase that methylates the C-2 hydroxyl group of an l-rhamnose unit (43). In these cases, we classified the proteins based on the initial functional annotation. These proteins attest to the great coding potential of chloroviruses, being a good example of genetic economy among giant viruses.

Regarding the COGs that have defined functions, we did not observe a large proportional difference in quantity when considering the pan-genome or just the core genome, with a few exceptions. Of these, it is interesting to note a greater presence of conserved genes related to transcription and RNA processing, suggesting an important ability to manipulate the process once it is triggered in the host cell. Interestingly, no chlorovirus has an RNA polymerase homologue, which makes it completely dependent on the cellular enzyme to produce its transcripts (40). It is interesting to note, however, that all chloroviruses encode two enzymes involved in mRNA capping. One enzyme is responsible for capping the 5′ end of the primary transcripts (44, 45), and the other is an RNA triphosphatase that is part of the capping apparatus (45, 46). The third component of the capping apparatus, methylation of the GpppN cap by S-adenosylmethionine:RNA (guanine-N7) methyltransferase, is presumed to be provided by the algal host. Furthermore, four transcription factors are present in all chloroviruses and are possibly used to initiate the gene transcription process, recruiting the host RNA polymerase.

Like transcription-related COGs, those belonging to other functional groups were present in viruses that infect all three host species. A slightly higher number of COGs was exclusively associated with NC64A viruses (368 COGs). It is worth noting that one of the viruses included in the analysis was the isolate named Osy-NE5, a virus isolated from the Chlorella variabilis Syngen 2-3 strain (21). Osy-NE5 can infect C. variabilis NC64A but does not complete replication in the alga (susceptible but not competent). A total of 348 COGs from NC64A viruses were shared with OSy-NE5, of which 51 were unique to it. It is possible that with the isolation and characterization of new specific viruses for this strain of C. variabilis, we will have new information that will allow a more detailed comparative genomic analysis between the groups. This isolate as well as those closest to it are considered Osy viruses, a new group (possibly a new subgenus) within Chlorovirus, and further studies should be carried out to better understand the genomics and biology of Osy viruses.

Conclusions.

Our analyses revealed an open pan-genome for chloroviruses, suggesting a great potential for genetic innovation for this group of viruses, once new isolates are obtained from unexplored sites on the planet. Similar to other organisms, the pan-genome evolution of chloroviruses may have been shaped by HGT events, but other forces are likely to have an influence, considering the conserved structures of gene gangs in these viruses (13). Despite the great genetic diversity, the chlorovirus core genome appears to be well delimited, with approximately 1/3 of the genome of these viruses being well conserved, regardless of the subgroup to which it belongs (related to the host). Furthermore, our analyses reinforce the monophyletism of chloroviruses but call attention to the need for a restructuring at the taxonomic level of this important group of aquatic viruses, ubiquitous on the planet. The study of chloroviruses has revealed unique characteristics of these viruses, including an emerging potential for relevant biotechnological applications (47–52). Thus, exploring new sites to isolate and characterize chloroviruses could reveal great surprises, further expanding the pan-genome of these viruses, being a valuable source of genetic diversity and biotechnological innovation.

MATERIALS AND METHODS

Data set.

Fasta files containing the complete genomes as well as the amino acid sequences (CDSs) of different chlorovirus isolates were downloaded from GenBank, constituting the raw data set. The search was done using the words “chlorovirus” and “chlorella virus” as bait. Furthermore, some genomes have been identified based on previously described deposits (9). The search was carried out in February 2021, and all sequences available at the time were obtained (42 isolated viruses). Of the 42 isolates, 14 were isolated from Chlorella variabilis NC64A, 1 was from Chlorella variabilis Syngen 2-3, 14 were from Micractinium conductrix Pbi, and 13 were from Chlorella heliozoae SAG 3.83. Of the 42 viruses, 6 had a number of CDSs much higher than the average number of CDSs (average of 350 genes) and were excluded from the pan-genome construction in order to avoid distortions in the analysis. Thus, the final data set for pan-genome analysis was composed of CDSs from 36 virus isolates, with 12 representatives of each host species in which these viral genomes have been annotated.

Phylogenetic analysis.

The phylogenetic reconstruction was based on the DNA polymerase B gene, as it is a well-conserved gene in all members of the Phycodnaviridae family (28). Amino acid sequences from all 42 chloroviruses were included, along with sequences from other representatives of the family, distributed among the other five genera. Sequence alignment was performed using Muscle software with default parameters (53). Noninformative positions were removed using TrimAl with a gap threshold of 0.2 (54). The tree was built using the maximum likelihood method with IQ-TREE, with the evolutionary model being selected by the program (55). Bootstrap analysis was performed with 1,000 replicates, and the tree was visualized with iTOL (56).

Pan-genome construction.

To estimate the size of the pan-genome, the chlorovirus gene sequences were clustered using the Proteinortho tool (57), based on the reciprocal best-hit strategy, using an amino acid sequence identity of 30% and a sequence coverage of 60% as thresholds and an E value of <10⁻⁵. The genes were clustered based on homology, constituting clusters of orthologous genes (COGs). For analysis of the evolution of the pan-genome and core genome, we performed a stepwise inclusion of each virus annotation in the pairwise comparisons of the gene contents of the 36 chlorovirus genome sequences included in the final data set.

Functional analysis and network construction.

Once the COGs were established, a representative gene for each of them was used for functional annotation and classification. This was made possible considering the homology of the genes included in each COG; thus, the functions of the genes included in the same cluster should be similar. The gene annotation was performed using BLASTP with an E value of <10⁻⁵ against the NCBI nr database (58). Conserved domains in the genes were searched using InterProScan (59). After function annotation, the COGs were classified into distinct functional groups based on NCVOG (Nucleo-Cytoplasmic Virus Orthologous Groups) classification and additional categories previously established for chloroviruses (39, 60). To obtain insights on the putative origin of the COGs, a BLASTP search was also performed excluding the Chlorovirus (taxid 181083) and unclassified Chlorovirus (taxid 346674) taxa and categorized into ORFans (genes with no homologues outside Chlorovirus) and those having the best hits with Bacteria, Eukarya, Archaea, and viruses. This analysis was performed using the DIAMOND tool in very-sensitive mode (61). To obtain a general picture of COG sharing among chloroviruses, a bipartite network was built using Gephi (62). The layout was generated using a force-based algorithm, followed by manual rearrangement of the nodes to better visualize the connections and unique groups of genes shared among each virus group based on the host with which they are associated.