| INVESTIGATION Recombination-Driven Genome Evolution and Stability of Bacterial Species Purushottam D. Dixit,* Tin Yau Pang,† and Sergei Maslov‡,§,1 *Department of Systems Biology, Columbia University, New York, New York 10032, †Institute for Bioinformatics, Heinrich-Heine- Universität Düsseldorf, 40221, Germany, and ‡Department of Bioengineering and §Carl R. Woese Institute for Genomic Biology, University of Illinois at Urbana-Champaign, Illinois 61801 ABSTRACT While bacteria divide clonally, horizontal gene transfer followed by homologous recombination is now recognized as an important contributor to their evolution. However, the details of how the competition between clonality and recombination shapes genome diversity remains poorly understood. Using a computational model, we find two principal regimes in bacterial evolution and identify two composite parameters that dictate the evolutionary fate of bacterial species. In the divergent regime, characterized by either a low recombination frequency or strict barriers to recombination, cohesion due to recombination is not sufficient to overcome the mutational drift. As a consequence, the divergence between pairs of genomes in the population steadily increases in the course of their evolution. The species lacks genetic coherence with sexually isolated clonal subpopulations continuously formed and dissolved. In contrast, in the metastable regime, characterized by a high recombination frequency combined with low barriers to recombination, genomes continuously recombine with the rest of the population. The population remains genetically cohesive and temporally stable. Notably, the transition between these two regimes can be affected by relatively small changes in evolutionary parameters. Using the Multi Locus Sequence Typing (MLST) data, we classify a number of bacterial species to be either the divergent or the metastable type. Generalizations of our framework to include selection, ecologically structured populations, and horizontal gene transfer of non- homologous regions are discussed as well. KEYWORDS bacterial evolution; recombination; population genetics acterial genomes are extremely variable, comprising both average number of genetic differences between pairs of indi- Ba consensus “core” genome, which is present in the ma- viduals in a population, often denoted by u. jority of strains in a population, and an “auxiliary” genome, During the last two decades, exchange of genetic fragments comprising genes that are shared by some but not all strains between closely related organisms has also been recognized (Medini et al. 2005; Tettelin et al. 2005; Hogg et al. 2007; as a significant factor in bacterial evolution (Guttman and Lapierre and Gogarten 2009; Touchon et al. 2009; Dixit et al. Dykhuizen 1994; Milkman 1997; Falush et al. 2001; Thomas 2015; Marttinen et al. 2015). and Nielsen 2005; Studier et al. 2009; Touchon et al. 2009; Vos Multiple factors shape the diversification of the core ge- and Didelot 2009; Dixit et al. 2015). Transferred fragments are nome. For example, point mutations generate single-nucleo- integrated into the recipient chromosome via homologous re- tide polymorphisms (SNPs) within the population that are combination. Notably, recombination between pairs of strains passed on from mother to daughter. At the same time, sto- is limited by the divergence in transferred regions. The prob- 2d=dTE chastic elimination of lineages leads to fixation of polymor- ability psuccess e of successful recombination of foreign phisms, which effectively reduces population diversity. The DNA into a recipient genome decays exponentially with d, balance between point mutations and fixation determines the the local divergence between the donor DNA fragment and the corresponding DNA on the recipient chromosome (Vuli´c Copyright © 2017 by the Genetics Society of America et al. 1997; Majewski 2001; Thomas and Nielsen 2005; Fraser doi: https://doi.org/10.1534/genetics.117.300061 d Manuscript received November 9, 2016; accepted for publication July 18, 2017; et al. 2007; Polz et al. 2013). Segments with divergence published Early Online July 27, 2017. greater than divergence dTE have negligible probability of suc- 1Corresponding author: 3406, Carl R. Woese Institute for Genomic Biology, University of Illinois at Urbana-Champaign, 1206 W Gregory Dr., MC-195, Urbana, IL 61801. cessful recombination. In this work, we refer to the divergence E-mail: [email protected] dTE as the transfer efficiency. dTE is shaped at least in part by Genetics, Vol. 207, 281–295 September 2017 281 the restriction modification (RM), the mismatch repair over mutations, r=m (defined as the ratio of the number of (MMR) systems, and the biophysical mechanisms of homol- SNPs brought by recombinations and those generated by ogous recombination (Vuli´c et al. 1997; Majewski 2001). point mutations in a pair of closely related strains), and our The transfer efficiency dTE imposes an effective limit on the second composite parameter u=dTE: Based on our analysis divergence among subpopulations that can successfully ex- of the existing Multi Locus Sequence Typing (MLST) data, change genetic material with each other (Vuli´c et al. 1997; we find that different real-life bacterial species belong to Majewski 2001). either divergent or metastable regimes. We discuss possi- In this work, we develop an evolutionary theoretical frame- ble molecular mechanisms and evolutionary forces that work that allows us to study in broad detail the nature of decide the role of recombination in a species’ evolutionary competition between recombinations and point mutations fate. We also discuss possible extensions of our analysis across a range of evolutionary parameters. We identify two to include adaptive evolution, effects of ecological niches, composite parameters that govern how genomes diverge and genome modifications such as insertions, deletions, and from each other over time. Each of the two parameters inversions. corresponds to a competition between vertical inheritance of polymorphisms and their horizontal exchange via homol- Computational Models ogous recombination. First is the competition between the recombination rate r We consider a population of Ne coevolving bacterial strains. and the mutation rate m. Within a coevolving population, The population evolves with nonoverlapping generations and consider a pair of strains diverging from each other. The av- in each new generation each of the strains randomly chooses erage time between consecutive recombination events affect- its parent (Gillespie 2010). As a result, the population remains 6 ing any given small genomic region is 1=ð2rltrÞ where ltr is the constant over time. Strain genomes have length lG ¼ 5 3 10 : average length of transferred regions. The total divergence Individual base pairs acquire point mutations at a constant accumulated in this region due to mutations in either of the rate m and recombination events are attempted at a constant two genomes is dmut 2m=2rltr: If dmut dTE; the pair of rate r (see Figure 1A). The mutations and recombination genomes is likely to become sexually isolated from each other events are assumed to have no fitness effects (later, we discuss in this region within the time that separates two successive how this assumption can be relaxed). The probability of a recombination events. In contrast, if dmut , dTE; frequent re- successful integration of a donor gene decays exponentially, 2d=dTE combination events would delay sexual isolation resulting in psuccess e ; with the local divergence d between the a more homogeneous population. Second is the competition donor and the recipient. Table 1 lists all important param- between the population diversity u and dTE: If dTE u; one eters in our model. expects spontaneous fragmentation of the entire population Unlike point mutations that occur anywhere on the ge- into several transient sexually isolated subpopulations that nome, genomic segments involved in recombination events rarely exchange genetic material between each other. In con- have a well-defined starting point and length. To understand trast, if dTE u; unhindered exchange of genetic fragments the effect of these two factors, below we introduce three may result in a single cohesive population. variants of a model of recombination with increasing com- Using computational models, we show that the two com- plexity illustrated in Figure 1B. In the first and the only math- posite parameters identified above, u=dTE and dmut=dTE; deter- ematically tractable model, we fix both length and start/end mine qualitative evolutionary dynamics of bacterial species. points of recombined segments. In the second model, recom- Furthermore, we identify two principal regimes of this dynam- bined segments have a fixed length but variable starting/ ics. In the divergent regime, characterized by a high dmut=dTE; ending positions. Finally, in the most realistic third model, local genomic regions acquire multiple mutations between suc- recombined segments have variable lengths [drawn from an cessive recombination events and rapidly isolate themselves exponential distribution with an average of 5000 bp (Dixit from the rest of the population. The population remains mostly et al. 2015)] and variable starting/ending positions. Prima clonal where transient sexually isolated subpopulations are con- facie, these three models appear quite distinct from each tinuously formed and dissolved. In contrast, in the metastable other, potentially leading to divergent conclusions about