Population genomics of picophytoplankton unveils novel chromosome hypervariability

Abstract

Tiny photosynthetic microorganisms that form the picoplankton (between 0.3 and 3 μm in diameter) are at the base of the food web in many marine ecosystems, and their adaptability to environmental change hinges on standing genetic variation. Although the genomic and phenotypic diversity of the bacterial component of the oceans has been intensively studied, little is known about the genomic and phenotypic diversity within each of the diverse eukaryotic species present. We report the level of genomic diversity in a natural population of Ostreococcus tauri (Chlorophyta, Mamiellophyceae), the smallest photosynthetic eukaryote. Contrary to the expectations of clonal evolution or cryptic species, the spectrum of genomic polymorphism observed suggests a large panmictic population (an effective population size of 1.2 × 10⁷) with pervasive evidence of sexual reproduction. De novo assemblies of low-coverage chromosomes reveal two large candidate mating-type loci with suppressed recombination, whose origin may pre-date the speciation events in the class Mamiellophyceae. This high genetic diversity is associated with large phenotypic differences between strains. Strikingly, resistance of isolates to large double-stranded DNA viruses, which abound in their natural environment, is positively correlated with the size of a single hypervariable chromosome, which contains 44 to 156 kb of strain-specific sequences. Our findings highlight the role of viruses in shaping genome diversity in marine picoeukaryotes.

Keywords: GC content evolution; chromothripsis; evolutionary genomics; linkage disequilbrium; mating type locus; multiple nucleotide mutation events; picophytoplankton; population genomics; prasinovirus; sex evolution.

Fig. 1. Polymorphism features on 18 standard chromosomes.

(A) Site frequency spectrum of total SNPs. (B) Number of observed and expected mutation events involved in MNPs. (C) Distribution of fitness effects on nonsynonymous sites. (D) LD between pairs of SNPs, measured as r², with distance on chromosome 10. Gray line, expected LD for unlinked sites estimated from site frequency spectrum on intergenic sites.

Fig. 2. Variation of the population recombination rate (ρ), nucleotide diversity (π), and GC content (GC) along chromosomes of O. tauri.

(A) Chromosome 1, a standard chromosome. (B) Chromosome 2. ρ, population-scaled recombination rate (per kilobase) inferred from SNP using the program interval of LDhat package; π, nucleotide diversity per site averaged across 2.5-kb windows; GC, averaged GC percent across 10-kb windows; NCR, noncallable regions of at least 1 kb, where polymorphisms could not be called because of insufficient coverage (<10×) or because of too high coverage suggestive of regions in multiple copies, which are not represented in the reference sequence. ρ and π are computed using SNPs segregating among the 13 strains for chromosome 1 and 12 strains (RCC4221 is excluded) for chromosome 2. For chromosome 2, the reference strain RCC4221 was excluded because of its high divergence with the other strains on the candidate MAT locus region; the sequence of chromosome 2 of strain RCC1115 was used instead.

Fig. 3. Genetic structure of the two candidate MAT loci on chromosome 2 in O. tauri.

(A) Genes are indicated by vertical lines. Blue lines, orthologous genes between plus and minus type strains; green lines, plus specific genes; orange lines, minus specific genes; gray lines, orthologous relationship between genes on different strains. “*” indicates the position of the five orthologous genes used to build the phylogeny from the different strains. (B) Phylogeny of five orthologous housekeeping genes trapped inside the MAT locus region in seven species of Mamiellophyceae (O. tauri, Ostreococcus mediterraneus, Ostreococcus RCC809, Ostreococcus lucimarinus, Bathycoccus prasinos RCC1105, Micromonas pusilla RCC299, and M. pusilla CCMP1545). Bootstrap maximum likelihood (ML) percentages are indicated on nodes.

Fig. 4. Chromosome 19 sequence conservation in six O. tauri strains.

(A) Large synteny blocks along chromosome 19 are represented as rectangles (fig. S6), and red lines indicate location of strain-specific DNA. Pairwise local alignments between strains are represented in gray (sense) or blue (antisense) (blastn with scores between 50 and 500 and percentage alignment identities >95%). (B) Proportion of DNA in chromosome in syntenic regions, in rearranged regions, and in strain-specific regions. (C) Taxonomic affiliation of strain-specific sequences using sequence homology against GenBank (tblastx) (69).

Fig. 5. Length of chromosome 19 versus susceptibility to prasinoviruses.

Susceptibility of each strain is estimated by the percent of prasinoviruses lysing a strain and varies from 0% (strain is resistant to all 32 viruses) to 100% (strain is lysed by all 32 viruses). Pearson correlation coefficient ρ = −0.74, P < 0.005.