Gene gain and loss push prokaryotes beyond the homologous recombination barrier and accelerate genome sequence divergence

Abstract

Bacterial and archaeal evolution involve extensive gene gain and loss. Thus, phylogenetic trees of prokaryotes can be constructed both by traditional sequence-based methods (gene trees) and by comparison of gene compositions (genome trees). Comparing the branch lengths in gene and genome trees with identical topologies for 34 clusters of closely related bacterial and archaeal genomes, we show here that terminal branches of gene trees are systematically compressed compared to those of genome trees. Thus, sequence evolution is delayed compared to genome evolution by gene gain and loss. The extent of this delay differs widely among bacteria and archaea. Mathematical modeling shows that the divergence delay can result from sequence homogenization by homologous recombination. The model explains how homologous recombination maintains the cohesiveness of the core genome of a species while allowing extensive gene gain and loss within the accessory genome. Once evolving genomes become isolated by barriers impeding homologous recombination, gene and genome evolution processes settle into parallel trajectories, and genomes diverge, resulting in speciation.

Horizontal gene transfer barrier shapes the evolution of prokaryotic pangenomes

“…Within this modeling framework, the intersection of genomes, , decays exponentially with the total evolutionary distance (see Methods for details). The exponential decay constant, , is given by the per-gene loss rate ~−⁄ , where is the number of genes [19]. The exponential decay of genome intersections is obtained under two assumptions.…”

“…The evolution of the genome size can be modelled as a random process, where genes are gained and lost stochastically, with rates + and − , respectively. Accordingly, the dynamics of the number of genes is given by the equation [24] d d = + − − [2] For IGP-CGS model, genomes intersection of genomes is given by [19] ( ) = e −λ [3] where the decay constant is given by the per-gene loss rate = − / , and is the total evolutionary distance spanned by those genomes. Specifically, is given by the sum of all branch lengths in the phylogenetic tree that describes the evolutionary relations of the genomes.…”

“…To allow reasonable computational times, ATGCs with more than 20 genomes were sampled such that each ATGC contains at most 20 representative genomes. Throughout the analysis, phylogenetic trees were rescaled to compensate for the systematic underestimation of branch lengths that is apparently due to homologous recombination [19]. Figure 1.…”

“…Here, we analyze the divergence of prokaryotic genomes within the theoretical framework we developed previously [19,24] and extract gene commonality from genomes intersection using the inclusion-exclusion principle [25]. This allows us to obtain the observed U-shape distribution without assuming any specific functional form for the gain and loss rates.…”

Slightly beneficial genes are retained by evolving Horizontal Gene Transfer despite selfish elements

“…Horizontal Gene Transfer (HGT), the transmission of genetic material between unrelated individuals, is a major factor driving prokaryotic evolution (Ochman et al, 2000;Doolittle and Zhaxybayeva, 2009;Vogan and Higgs, 2011). Recent estimates of the rate of HGT in closely related bacteria are staggeringly high (Iranzo et al, 2019;Sakoparnig et al, 2019), with HGT possibly even outpacing gradual sequence evolution (Hao and Golding, 2006;Puigbò et al, 2014;Vos et al, 2015). Combining this with the fact that prokaryotes adapt mostly through rapid gene loss (Kuo and Ochman, 2009;Morris et al, 2012), bacterial adaptation appears to be mainly driven by changes in gene content (Snel et al, 2002;Treangen and Rocha, 2011;Nowell et al, 2014).…”