A Large Open Pangenome and a Small Core Genome for Giant Pandoraviruses

Abstract

Giant viruses of amoebae are distinct from classical viruses by the giant size of their virions and genomes. Pandoraviruses are the record holders in size of genomes and number of predicted genes. Three strains, P. salinus, P. dulcis, and P. inopinatum, have been described to date. We isolated three new ones, namely P. massiliensis, P. braziliensis, and P. pampulha, from environmental samples collected in Brazil. We describe here their genomes, the transcriptome and proteome of P. massiliensis, and the pangenome of the group encompassing the six pandoravirus isolates. Genome sequencing was performed with an Illumina MiSeq instrument. Genome annotation was performed using GeneMarkS and Prodigal softwares and comparative genomic analyses. The core genome and pangenome were determined using notably ProteinOrtho and CD-HIT programs. Transcriptomics was performed for P. massiliensis with the Illumina MiSeq instrument; proteomics was also performed for this virus using 1D/2D gel electrophoresis and mass spectrometry on a Synapt G2Si Q-TOF traveling wave mobility spectrometer. The genomes of the three new pandoraviruses are comprised between 1.6 and 1.8 Mbp. The genomes of P. massiliensis, P. pampulha, and P. braziliensis were predicted to harbor 1,414, 2,368, and 2,696 genes, respectively. These genes comprise up to 67% of ORFans. Phylogenomic analyses showed that P. massiliensis and P. braziliensis were more closely related to each other than to the other pandoraviruses. The core genome of pandoraviruses comprises 352 clusters of genes, and the ratio core genome/pangenome is less than 0.05. The extinction curve shows clearly that the pangenome is still open. A quarter of the gene content of P. massiliensis was detected by transcriptomics. In addition, a product for a total of 162 open reading frames were found by proteomic analysis of P. massiliensis virions, including notably the products of 28 ORFans, 99 hypothetical proteins, and 90 core genes. Further analyses should allow to gain a better knowledge and understanding of the evolution and origin of these giant pandoraviruses, and of their relationships with viruses and cellular microorganisms.

Keywords: core genome; giant virus; megavirales; pandoravirus; pangenome.

FIGURE 1

Electron microscopy pictures of pandoravirus isolates by negative staining (a,c) or after inclusion (b). (a) Pandoravirus massiliensis; (b) Pandoravirus pampulha; (c) Pandoravirus braziliensis.

FIGURE 2

Evolution of the size of the pangenome and core genome of pandoraviruses with the description of each of the six pandoravirus isolates.

FIGURE 3

Venn diagram of genes shared and not shared between the gene contents of the six pandoravirus isolates. Venn diagram was built using the following online tool: http://bioinformatics.psb.ugent.be/webtools/Venn/.

FIGURE 4

Distribution of the number of pandoravirus ORFs included in clusters of orthologous groups of proteins according to the number of pandoravirus genomes for which genes were involved in these clusters (A) and proportion for each pandoravirus genome of the genes involved in clusters including genes from one to six pandoraviruses (B).

FIGURE 5

Alignment by muscle of the 13 full-length miniature inverted repeat transposable elements (MITEs) described in the genome of Pandoravirus salinus (Sun et al., 2015) and those detected in the genomes of P. massiliensis and P. inopinatum. Alignment is limited to regions including the flanking target site duplication (TSD) sequences and the start and end of the MITEs.

FIGURE 6

Phylogenetic reconstruction based on amino acid sequences of the DNA-dependent RNA polymerase subunit 1 from representatives of megavirales. Phylogenetic tree was drawn using the maximum likelihood model with the FastTree program (Price et al., 2010).

FIGURE 7

Hierarchical clustering based on the presence/absence patterns of clusters of orthologous groups of proteins of megavirales members in the genomes of pandoraviruses.

FIGURE 8

Average amino acid identity between ORFs predicted for pandoravirus genomes. For each comparison, estimates were obtained using both best hits and reciprocal best hits between two sets of proteins from a pandoravirus newly described here and Pandoravirus salinus. (A) Pandoravirus braziliensis versus Pandoravirus salinus, (B) Pandoravirus massiliensis versus Pandoravirus salinus, (C) Pandoravirus pampulha versus Pandoravirus salinus.

FIGURE 9

Whole genome alignment of pandoravirus genomes by the MAUVE program (Darling et al., 2010).

FIGURE 10

Distribution of core genes and paralogous genes along the pandoravirus genomes. (A) Pandoravirus braziliensis; (B) Pandoravirus pampulha; (C) Pandoravirus massiliensis. Core genes are indicated by red dots; paralogous genes are indicated by blue diamonds.