How clonal are bacteria over time?
Introduction
Here, I revisit the question posed in the title of a classic paper by John Maynard Smith and colleagues [1]: How clonal are bacteria, and more specifically how does clonality vary among different microbial populations and over time? First, what do we mean by clonality? Perfectly clonal bacteria replicate by cell division (vertical descent) and evolve by random mutations that occur during DNA replication. In this theoretical population, there is negligible horizontal transfer of DNA by recombination across the resulting tree of vertical descent. Very few (if any) natural bacterial populations fit this idealized, theoretical definition of clonality. Or, as discussed below, they might only fit it for a short amount of time. However, knowing where a bacterial population of interest happens to fall along a spectrum of clonality can help us understand its biology, and even make predictions about its evolution.
The opposite of clonality is panmixis — a situation in which the rate of horizontal transfer is higher than the rate of vertical cell division, resulting in random association (linkage equilibrium) among loci in the genome [1, 2]. However, rates of horizontal transfer (recombination) vary widely across the genome, such that a population can be mostly clonal, except for a few loci in the genome [3]. These loci came to be termed genomic islands — a metaphor I will build upon below. Some of the first islands identified were called pathogenicity islands because they contained virulence factors [4]. However, non-pathogenic environmental bacteria also contain islands, conferring adaptation to different ecological niches. For example, genes in Prochlorococcus genomic islands confer adaptation to light and nutrient conditions [5, 6]. But islands need not confer niche adaptation to their host genome; they can be neutral to host fitness or even detrimental, selfish parasites. Here, I define genomic islands broadly as any piece of DNA that is transferred horizontally (by either homologous or nonhomologous recombination) from cell to cell and therefore evolves independently (i.e. is unlinked) from the rest of the genome.
I will begin by extending the use of island analogies to include continents, peninsulas and archipelagos (Table 1). I will then use these analogies to discuss to what extent microbial populations are clonal or panmictic, and how often they transition between the two regimes.
Section snippets
Are some islands really peninsulas?
In the classic analogy, an island is totally disconnected from the mainland, meaning that genes in the island evolve independently of the genome (Table 1). Examples of islands that fit this strict independence might include integrated phages and other ‘selfish’ elements, or genes that reside in a particular niche but not in a particular genome (e.g. a gene ecology model [7•]). Peninsulas provide an analogy that might better describe how islands are related to microbial genomes. A peninsula (or
Are some genomes archipelagos?
The very concept of one or a few islands implies a contrast with the large, clonal genomic mainland or continent. But some microbial genomes may contain so many islands that there is no mainland, only a vast archipelago (Table 1). A striking recent example is a population of hotspring cyanobacteria in which virtually every gene in the genome evolved independently due to frequent recombination [11•], leading the authors to call the population ‘quasi-sexual’ (in other words, panmictic). Frequent
Clonal expansions from panmictic pools
I propose that archipelagos are not necessarily static over time, and that archipelagos can sometimes coalesce into continents. Given the right ecological opportunity, a genome from a panmictic gene pool can escape the ‘gravitational pull’ of recombination and take off into a clonal expansion. An example mentioned earlier is V. cholerae, a genetically diverse group of coastal marine bacteria, some of which cause cholera. Virulence is mainly determined by two loci in the genome: the cholera
The balance between recombination and selection
Let us consider the evolutionary forces that determine clonality: natural selection and recombination. The effect of recombination on clonality is straightforward: more recombination means less clonality. The effect of natural selection is more complex, but is defined here simply as a force which favors clonal expansions of adaptive mutants within an ecological niche. Selection, as defined here, therefore includes ecological effects. When driven by ecological selection, clonal expansions are
Modeling the recombination-selection balance
When rates of recombination are relatively low compared to rates of natural selection on adaptive genes within niches, entire genomes will sweep to fixation before they can be shuffled by recombination. Following previous modeling work, s is defined here as the selective coefficient of a niche-adaptive allele, and r is the recombination rate, per locus per generation [7•, 32, 33]. The s ≫ r regime is well described in the Stable Ecotype Model [17], which predicts that most of the genome will
Genome-wide and gene-specific sweeps in nature
To date, empirical evidence for gene-specific and genome-wide sweeps has come mostly from cross-sectional studies of a single population of genomes at a single point in time, with recombination and selection inferred backward in time [11•, 13•, 14, 35]. Sequencing microbial genomes or metagenomes sampled over time — already a typical practice in genomic epidemiology (e.g. [28, 36]) — promises to elucidate the rates of gene-specific and genome-wide sweeps in nature (Figure 1).
In a pioneering study,
Is clonality a stable trait?
As described in the V. cholerae example, some pathogenic bacterial populations can switch between panmictic and clonal lifestyles [11•, 13•, 14, 19•, 25, 34, 35]. Therefore clonality can vary over time, but how much and how often? To quantify the stability of clonality over time, not just in pathogens but in free-living environmental bacteria, I re-analyzed the lake time-course of Bendall et al. [37••]. Because estimates of selection and recombination rates were not readily available for this
History repeats itself
It appears that pathogens are more likely than free-living bacteria to undergo clonal expansions, due in part to their ecology and transmission dynamics [41]. Free-living aquatic bacteria, on the other hand, seem to be more likely to live in large, panmictic populations and behave like archipelagos [11•, 13•, 14, 37••]. If clonality is indeed a stable trait, this implies that history will repeat itself, and that the future behavior of microbial populations can be predicted with some confidence
References and recommended reading
Papers of particular interest, published within the period of review, have been highlighted as:
• of special interest
•• of outstanding interest
Acknowledgements
I am grateful to Rex Malmstrom, Yan Boucher, and Salvador Almagro-Moreno for their thoughtful comments which improved the manuscript. I was supported by a Canada Research Chair.
References (48)
- et al.
Ordering microbial diversity into ecologically and genetically cohesive units
Trends Microbiol
(2014) - et al.
Lysogenic conversion by a filamentous phage encoding cholera toxin
Science
(1996) - et al.
Genome-wide selective sweeps and gene-specific sweeps in natural bacterial populations
ISME J
(2016) - et al.
Microbial diversity and the genetic nature of microbial species
Nat Rev Microbiol
(2008) - et al.
Dense genomic sampling identifies highways of pneumococcal recombination
Nat Genet
(2014) - et al.
How clonal are bacteria?
Proc Natl Acad Sci U S A
(1993) - et al.
Clonal divergence in Escherichia coli as a result of recombination, not mutation
Science
(1994) - et al.
The landscape of realized homologous recombination in pathogenic bacteria
Mol Biol Evol
(2016) - et al.
Pathogenicity islands of virulent bacteria: structure, function and impact on microbial evolution
Mol Microbiol
(1997) - et al.
Genomic islands and the ecology and evolution of Prochlorococcus
Science
(2006)