Identification of nucleotide-level changes impacting gene content and genome evolution in orthopoxviruses

Abstract

Poxviruses are composed of large double-stranded DNA (dsDNA) genomes coding for several hundred genes whose variation has supported virus adaptation to a wide variety of hosts over their long evolutionary history. Comparative genomics has suggested that the Orthopoxvirus genus in particular has undergone reductive evolution, with the most recent common ancestor likely possessing a gene complement consisting of all genes present in any existing modern-day orthopoxvirus species, similar to the current Cowpox virus species. As orthopoxviruses adapt to new environments, the selection pressure on individual genes may be altered, driving sequence divergence and possible loss of function. This is evidenced by accumulation of mutations and loss of protein-coding open reading frames (ORFs) that progress from individual missense mutations to gene truncation through the introduction of early stop mutations (ESMs), gene fragmentation, and in some cases, a total loss of the ORF. In this study, we have constructed a whole-genome alignment for representative isolates from each Orthopoxvirus species and used it to identify the nucleotide-level changes that have led to gene content variation. By identifying the changes that have led to ESMs, we were able to determine that short indels were the major cause of gene truncations and that the genome length is inversely proportional to the number of ESMs present. We also identified the number and types of protein functional motifs still present in truncated genes to assess their functional significance.

Importance: This work contributes to our understanding of reductive evolution in poxviruses by identifying genomic remnants such as single nucleotide polymorphisms (SNPs) and indels left behind by evolutionary processes. Our comprehensive analysis of the genomic changes leading to gene truncation and fragmentation was able to detect some of the remnants of these evolutionary processes still present in orthopoxvirus genomes and suggests that these viruses are under continual adaptation due to changes in their environment. These results further our understanding of the evolutionary mechanisms that drive virus variation, allowing orthopoxviruses to adapt to particular environmental niches. Understanding the evolutionary history of these virus pathogens may help predict their future evolutionary potential.

FIG 1

Genome alignment. A multiple-sequence alignment (MSA) of representative orthopoxviruses is presented. A histogram of sequence conservation (nucleotide identity) is presented above the sequences. The locations of the central and end regions are shown below the sequences. Thick black lines indicate the presence of nucleotide sequence, intermediate black lines represent nucleotides interspersed with short gap insertions, and thin black lines indicate runs of gaps. Green arrows below the sequence lines mark intact genes, gray arrows mark truncated genes, and yellow arrows mark fragmented ORFs. The purple bars overlaying the sequence lines illustrate the positions of the ITRs. The SINE sequence present in TATV, which is the only nonancestral sequence detected, is denoted by a red bar.

FIG 2

Large genomic duplications. The duplications characterized in Table 2 are displayed as ribbons within the graph, with black circles indicating the proposed origin, providing the direction of the duplication. The tick marks refer to the base position (in thousands of bases) within the whole-genome MSA (Fig. 1). The positions of the ITRs for each genome are displayed outside the graph. Isolate abbreviations are provided in the center of the graph and are the same color as the duplication ribbons and the ITRs. MPXV-ZAI and MPXV-WR share the darker purple duplication.

FIG 3

Phylogenies resulting from DNA sequences of the whole genome (A), the 5′ region (B), the center (C), and the 3′ region (D) of the selected orthopoxvirus genomes. Phylogenies were inferred using Bayesian methods under the GTR+I+G nucleotide substitution model. Indel positions were included in the analysis as separate characters. Numbers at internal nodes provide clade credibility values for each node. Table 4 shows the positions for the 5′, center, and 3′ regions.

FIG 4

HSPV sequences absent in RPXV and VACV. Each panel displays a segment of the whole-genome alignment, with the original location within the MSA indicated along the top of the panel. Thick black lines indicate the presence of nucleotides, and thin black lines represent gaps in the alignment. Intact and truncated ORFs are shown as green and gray arrows, respectively. Sequence that is present in HSPV but has been deleted from either or both RPXV or VACV-WR is highlighted with red lines. Early stop mutations (ESMs) are shown as vertical red bars. In the top panel, a total of 2,898 bases are missing in VACV-WR. RPXV and VACV-WR each exhibit deletions of 10,639 bases in the middle panel. In the bottom panel, RPXV is missing a total of 8,925 bases and VACV-WR is missing a total of 11,778 bases.

FIG 5

Sequence variation in intact, truncated, and fragmented genes. Gene variants categorized as SNPs or indels are quantitated according to their presence in intact, truncated, or fragmented genes. Variants in the truncated region of an ORF are those that occur prior to the stop codon that interrupts a truncated gene, and variants in the degraded region of a truncated ORF occur after the stop codon that interrupts a truncated gene.

FIG 6

Frequency of orthopoxvirus deletions. (A) Each point on this graph shows the percentage of positions that consist of gaps within a sliding window of 100 bp for each isolate. This graph does not include the extreme ends of alignments, and the 0% and 100% y axis values are expanded vertically in order to allow separation of data points that would otherwise overlap. Each isolate is shown individually at each position. (B) Distribution of gap size for each isolate in the MSA after removal of the telomeres and large duplications. Each gap length bin (x axis) is colored according to the number of gaps present in each isolate.

FIG 7

Example of early stop mutations. Excerpts from the whole-genome alignment, with the nucleotide alignment shown as black bars and breaks in the bars representing gaps. Intact and truncated ORFs are shown as green and gray arrows, respectively. ESMs consisting of indels that lead to frame shifts and truncation of the ORF are shown as red squares, and ESMs due to nonsense SNPs are shown as red triangles. Large deletions are marked as transparent orange lines overlaying the nucleotide track. Orange hatch marks represent deletions within a truncated gene that maintain the reading frame following the deletion. The numbers above the alignments refer to the positions in the whole-genome alignment. (A) CP77 gene. The alignment has been reversed for this figure so that all genes run from left to right. (B) IL-1β receptor homolog gene.

FIG 8

Length of deletions that result in ESMs. Each bin in this histogram indicates the number of deletions of the indicated length present across all analyzed virus isolates. The x axis is divided into 100-bp-length bins; however, the graph labels are shown at only 500-bp intervals. The indicated length does not include gaps inserted due to insertions in other sequences. The y axis, on a log scale, displays the number of occurrences of each deletion of the indicated size. Bins for deletions of up to 50 bp have been magnified in the inset graph; bins from 1 to 9 are at intervals of 1 bp, bins from 10 to 50 are at intervals of 10 bp, and the y axis is on a linear scale.

FIG 9

Frequency of ESMs in orthopoxvirus genomes. (A) Relationship between genome length and ESM frequency. Related isolates that are located close to each other on the graph are circled. (B) Distribution of ESMs for each isolate relative to gene state.

FIG 10

Protein functional motifs by gene state (A) and position of the gene (B). Functional protein motifs were detected using InterProScan. (A) Number of protein motifs for each gene state in each isolate. (B) Number of protein motifs across the genomes, with the relative genomic position of the genes shown on the x axis (see Table 5 for more information on the genes shown). The number of total domains possible is the product of the number of domains present in the intact gene and the number of virus isolates analyzed (i.e., 15). The number of motifs present is the total of all of the motifs in intact genes and motifs found in the 5′ ORF of truncated genes. Ankyrin genes, which contain multiple domains, are marked by stars. The center and variable ends of genomes are indicated below the graph.