Fig 1.
Nearly-strictly vertically transmitted chemosynthetic endosymbionts exhibit genome erosion despite ongoing horizontal transmission events in their populations.
A) Transmission mode spectrum from strict horizontal transmission to strict vertical transmission, with a diversity of mixed modes, incorporating both strategies, in between. B) Genome sizes from this and previous studies [11–13,92] reveal consistent patterns of moderate genome erosion among the vesicomyid symbionts, but not in the other groups with higher rates of horizontal transmission. C) Mitochondrial and symbiont whole genome genealogies are discordant for all groups, indicating that sufficient amounts of horizontal transmission occur in vertically transmitted vesicomyid populations to erode the association between these cytoplasmic genomes. Maximum likelihood cladograms are midpoint rooted, and nodes below 50% bootstrap support are collapsed. Species are color coded by their symbiont transmission mode as in A).
Fig 2.
Chemosynthetic bacterial symbionts and their bivalve hosts exhibit ancient divergence times.
A) Maximum likelihood phylogeny inferred from 108 orthologous protein coding genes and the 16S and 23S rRNA genes (outgroup = Alphaproteobacteria; branch labels = bootstrap support fraction) with RelTime divergence date estimates (node bars = 95% confidence intervals). Host-associated bacteria are listed as symbionts of their host species. Bacterial genome sizes are written to the right of the taxon names in the tip labels to highlight trends in genome size across clades. B) Whole mitochondrial Bayesian phylogeny for bivalves (outgroup = Gastropoda; branch labels = posterior probabilities) with divergence dates co-inferred in Beast2 (node bars = 95% highest posterior densities). In both phylogenies, members of vesicomyid, solemyid, and bathymodiolin (both thioautotrophic and methanotrophic) associations are colored yellow, green, and blue, respectively.
Fig 3.
Horizontal transmission and recombination introduce novel alleles into symbiont populations.
A) Model of endosymbiont genotype (pink vs. white) distributions under well-mixed high horizontal transmission rates and differentiated, low horizontal transmission rates. B) Horizontally transmitted mytilid (blue) and mixed mode transmitted solemyid (green) symbionts are well mixed among hosts, whereas the nearly strictly vertically transmitted vesicomyid symbionts (yellow) are highly differentiated among hosts. Error bars = 95% confidence intervals from non-parametric bootstrapping. C) Intrahost population folded allele frequency spectra (AFS) are shaped by access to gene flow, which is enabled by horizontal transmission and recombination. D) Recombination rates are significantly higher in the mytilid (blue) and solemyid (green) symbiont genomes compared to the vesicomyid symbiont genomes (yellow). Error bars = 95% confidence intervals.
Fig 4.
Consequences of access to gene flow via horizontal transmission and recombination on the distribution of symbiont genetic diversity between hosts.
A) Diagram showing how beneficial alleles (pink) are linked to deleterious alleles (grey) in populations experiencing strong selection on linked sites versus free recombination, and how these processes are reflected in the underlying population genealogies and allele frequency spectra. B-G) Symbiont genealogies and between host allele frequency spectra (AFS) for each host/symbiont species.
Fig 5.
Genome structure is shaped by horizontal transmission and recombination.
A) Models of recombination-based structural mutation mechanisms. B-D) Whole genome alignments for B) sulfur-oxidizing mytilid, C) solemyid, and D) vesicomyid symbiont genome assemblies with >1 Mb scaffolds.
Fig 6.
A conceptual model of the prevention of endosymbiont genome degradation through horizontal transmission and recombination.
Sufficient levels of genetic diversity, which can be introduced via horizontal transmission of symbiont genotypes between hosts and recombination between genotypes in mixed infections, prevents or delays genome degradation by restoring functional versions of mutated or deleted regions. Prevention can continue until recombination capabilities (RecA-dependent and independent) are completely lost, at which point, genetic rescue is no longer possible without wholesale symbiont replacement.