Introduction

In the first decade after the discovery of amoeba-infecting giant viruses in 2003, two taxonomic families were created due to their structural and genomic characteristics, which differed significantly from those of other viruses known then. These families, Mimiviridae and Marseilleviridae, were created to classify representatives of the only two taxa that had been established at the time. However, with the increased interest and intensification of research on this topic in the last 10 years, several other groups of giant amoeba viruses have been discovered [1,2,3,4,5,6,7,8,9,10]. Despite recent progress in taxonomy of giant viruses, many of them have not yet been formally assigned to any taxon recognized by the ICTV [11,12,13,14].

Despite of their great diversity, these viruses are hypothetically monophyletic and are grouped within the phylum Nucleocytoviricota. Within this phylum, the giant amoeba viruses recognized to date have been classified in the class Megaviricetes. Within this class, the order Pimascovirales previously included three recognized families: Ascoviridae and Iridoviridae, which have invertebrates and some groups of vertebrates as hosts, and Marseilleviridae, which comprises amoeba-infecting giant viruses [15]. However, three related groups of ovoid amoeba-infecting giant viruses had not yet received formal taxonomic recognition. These groups are the pithoviruses, cedratviruses, and orpheoviruses.

Despite having similar shapes and sizes, these three groups of viruses have many morphological and genomic differences that strongly distinguish them from other families within this order and from each other [16]. Therefore, in order to start the process of official classification of these isolates within the ICTV taxonomy framework, we proposed the formal creation of seven species: two for the pithoviruses, four for the cedratviruses, and one for the orpheoviruses, to be classified in the newly established genera “Alphapithovirus”, “Alphacedratvirus”, and “Alphaorpheovirus”, belonging to the new families “Pithoviridae”, “Cedratviridae”, and “Orpheoviridae”, respectively.

Etymology of taxon nomenclature

The family Pithoviridae and its representatives

We formally proposed to create the family “Pithoviridae”, composed of a new genus, “Alphapithovirus”, and two new species, “Alphapithovirus sibericum” and “Alphapithovirus massiliense”. “Pitho-” is derived from the Greek pithos, a kind of large amphora. The names of the species refer to the names of the founding isolates already used in the literature, with “alpha” as a prefix to denote the genus in which the original isolate is classified. The epithets were derived from the informal names initially proposed by the original authors who discovered the viruses and subsequently modified to Latinized singular forms, as exemplified by the standardized classification of the amoebal virus species Yaravirus brasiliense [2, 13, 17, 18].

The family Cedratviridae and its representatives

We also proposed to create the family “Cedratviridae”, composed of a new genus, “Alphacedratvirus”, and four species within this genus: “Alphacedratvirus aljazairense”, comprising the isolate cedratvirus A11, “Alphacedratvirus franciense”, comprising cedratvirus lausannensis and cedratvirus zaza, “Alphacedratvirus brasiliense”, comprising Brazilian cedratvirus, and “Alphacedratvirus rossiense” comprising cedratvirus kamchatka. Once more, the particle morphology served as an inspiration for the naming: “Cedrat-” from French cédrat, citron [3, 19,20,21].

The first cedratvirus, named cedratvirus A11, was isolated in 2016 from water samples collected in different regions of Algeria. To standardize the nomenclature, we propose naming the species, founded by the isolate cedratvirus A11, “Alphacedratvirus aljazairense”. “Aljazair-” is the Romanized form of the name Algeria in Arabic, al-Jazāʾir. The other proposed species names also follow the same pattern, derived from the geographic locations where isolates were originally collected. In the following years, other cedratviruses were isolated from water samples from France (cedratvirus lausannensis and cedratvirus zaza) and Brazil (Brazilian cedratvirus) and a soil sample collected in Russia (cedratvirus kamchatka) [3, 19,20,21].

The family Orpheoviridae and its representative

Lastly, we proposed the creation of the species “Alphaorpheovirus massiliense”, the genus “Alphaorpheovirus”, and the family “Orpheoviridae” to classify one viral isolate named orpheovirus IHUMI-LCC2. “Orpheo-” was inspired by the Greek legend of Orpheus [1].

Morphological properties

Instead of exhibiting icosahedral capsid symmetry, pithoviruses, cedratviruses, and orpheovirus have unconventional virion morphologies that are similar to each other. They feature elongated, ovoid-shaped particles measuring approximately 1 µm in length, but these viruses can be distinguished morphologically by the arrangement and number of delivery portals present in their particles. Whilst cedratviruses contain two corks, one on each apex of the particle, pithoviruses more often contain only one cork, and orpheovirus has one ostiole, also located on the apex of the particle. The delivery portals can often be observed by transmission electron microscopy [1, 3, 17, 22,23,24].

Occurrence and host range

The representatives of this clade have been isolated on at least four different continents. Over 10 cedratvirus isolates and four pithovirus isolates have been discovered using Acanthamoeba sp. as a cellular support host, but so far, only one orpheovirus isolate has been discovered, using Vermamoeba vermiformis [1,2,3, 17, 19,20,21, 23, 25,26,27,28,29,30,31]. These viruses have been found in a diverse array of samples, obtained from rivers, rat feces, permafrost, sewage water, and muddy soil. Despite the increased prospecting efforts in recent years, metagenomics work indicates that this clade exhibits significant diversity, with many different hierarchical taxa within the order Pimascovirales that remain largely unexplored [32,33,34,35].

Comparative genomic features

Viruses of this clade exhibit a large amount of variation in the size of their genomes, GC content, and coding density, but when compared at the family level, the characteristics become fairly consistent among the isolates. The main characteristics of the circular dsDNA genomes of these viruses are listed in Table 1.

Table 1 Main genomic characteristics of the isolated viruses of the three proposed viral families

The same pattern can be observed in the synteny analysis. A whole-genome alignment made using Mauve software showed strong synteny between the pithoviruses. Similar genomic arrangements could also be observed within each lineage of cedratvirus, with cedratvirus A11, cedratvirus zaza, and cedratvirus lausannensis belonging to lineage A and cedratvirus kamchatka and Brazilian cedratvirus to lineage B. However, despite the clear synteny between pithoviruses and cedratviruses when orpheovirus was added to the analysis, major differences were observed in terms of genome size, composition, and genetic organization (Fig. 1).

Fig. 1
figure 1

Genomic synteny analysis. A comparison of the genome organization of members of the three new families “Cedratviridae”, “Orpheoviridae”, and “Pithoviridae” is shown. The schematic diagram was produced using the Mauve software package [36]

The pairwise average nucleotide identity (ANI) and average amino acid identity (AAI) for pithoviruses were > 84% for the two isolates, and their gene sharing level was 78%. The pairwise ANI and AAI values for all of the cedratviruses was > 70%, and their gene sharing level was > 67%. An intergroup comparison between the proposed families showed a pairwise ANI < 30%. The average pairwise AAI and gene sharing level values between the pithoviruses and cedratviruses was 42.1% and 24.8% respectively, and between the orpheovirus and the other viruses, the average was 32.1% and 8.2%, respectively (Fig. 2).

Fig. 2
figure 2

(A and B) Pairwise average nucleotide identity (ANI) for the whole genome and for the ORFs. ANI values < 70% were set to 0 [37]. (C) Pairwise average amino acid identity (AAI) [37]. (D) Normalized gene sharing level [12]

Phylogenomics

A concatenated phylogenetic approach based on seven core genes of giant viruses, named DNA polymerase family B, NCLDV major capsid protein, DEAD/SNF2-like helicase, DNA topoisomerase II, DNA-directed RNA polymerase beta subunit, transcription initiation factor IIB, and poxvirus late transcription factor VLTF3, was performed using representatives of the family Marseilleviridae as an outgroup. The analysis showed that cedratviruses are a sister group of pithoviruses with an equivalent phylogenetic breadth, and the orpheovirus forms a divergent branch with a wider breadth, supporting the creation of the three new families. The analysis was performed using only the isolates for which the whole genome sequence had been deposited in public databases at the time of the study, but complete high-quality viral genome sequences from metagenomic studies can provide additional evolutionary context that could result in revisions of the taxonomy proposed here.

Fig. 3
figure 3

Maximum-likelihood-based phylogenetic tree of seven marker genes, constructed using IQ-TREE multicore version 2.2.0 [38]. The best-fit model was LG + F + I + G4, chosen according to the Bayesian information criterion (BIC) [39]. Branch support was computed using 1000 ultrafast bootstrap replicates. The phylogenetic tree was visualized and edited using iTOL [40]. The members of the three proposed new families are indicated by different colors

Demarcation criteria

We propose that isolates with pairwise ANI > 95% should be considered members of the same species. Therefore, in our proposal, the species “Alphacedratvirus franciense” includes two viral isolates: cedratvirus lausannensis and cedratvirus zaza. All other isolates included here should be classified in separate new species.

In consensus with the newly recognized family Mamonoviridae, we propose that representatives of the same genus within the three families proposed here should have a pairwise ANI > 70% and similar morphological traits. Thus, according to these criteria, each of the three families is monogenic [12].

Families should be monophyletic clades with a high level of bootstrap support. Inclusion in the families should be validated based on comparative genomic analysis similar to that presented in this study, as well as phylogenetic analysis based on the seven genes (DNA PolB, MCP, DEAD/SNF2-like, DNA topoII, RNAPII, TIFIIB, VLTF3) that have been identified as the best markers for phylogenetic reconstructions of giant viruses (Fig. 3) [41]. Future genomes can be included in the phylogenomic analysis even if they differ and lack a subset of these genes.

We acknowledge that the proposed demarcation criteria are adaptable and subject to future reassessment as new viruses are isolated and their complete genomes are sequenced.

Conclusion

The number of characterized giant viruses within the order Pimascovirales has increased over the last decade, indicating a vast diversity yet to be discovered [1,2,3, 17, 19,20,21, 23, 25,26,27,28,29,30,31,32,33,34,35]. However, a group of phylogenetically related and well-characterized giant viruses that were putative members of this order still lacked formal classification. This group comprises three well-established clades of isolated viruses, commonly referred to as “pithoviruses”, “cedratviruses”, and “orpheoviruses”. Despite their common features, these ovoid-shaped giant viruses exhibit significant differences in gene content, genomic architecture, and the phylogenetic history of several core genes [16]. Additionally, despite the morphological similarity of their virions (ovoid shape), these viruses have distinct delivery portal characteristics. Cedratviruses feature two corks at each end of the viral particle, pithoviruses more often have only one cork, and orpheovirus has an ostiole at the apex of its particle. Moreover, orpheovirus also infects a different host, Vermamoeba vermiformis, whereas cedratviruses and pithoviruses infect amoebae of the genus Acanthamoeba [1,2,3]. Considering the dissimilarities within this group, and in line with the demarcation criteria of recently approved new taxa of giant viruses, our phylogenomic data suggest the creation of three family-level taxa in the order Pimascovirales. Accordingly, we proposed the creation of the families “Orpheoviridae”, “Cedratviridae”, and “Pithoviridae” [11, 12], each composed of a single genus. ("Alphaorpheovirus”, “Alphacedratvirus”, and “Alphapithovirus”, respectively), to formally accommodate eight isolates classified into seven new species.

The official taxonomy of giant viruses is currently in a very early stage. Therefore, the taxonomic framework proposed here represents an initial attempt to classify some of these viruses and is surely destined for future changes as new representatives are described and their genomes sequenced. The formal taxonomic proposal for classification of cedrat- and pithoviruses as well as an orpheovirus was submitted to the ICTV in June 2023 and was approved by a ratification vote in April 2024.