they can also bear features that exploit hosts (Frank, 1996a,b; Sachs and Wilcox, 2006; Simms et al., 2006; Weeks et al., 2007; Oliver et al., 2008; Heath et al., 2010; Sachs et al., 2010a, 2011). As we detail later, each of these variables (degree of reliance on hosts, type of host habitat, and type of benefit provided to host) can modulate evolutionary transitions in bacterial symbiosis and can explain how and why transitions occur.
Here, we investigate evolutionary transitions that have occurred in the history of bacterial mutualism. We focus on (i) the origins of host association in bacteria (transitions in which environmental bacteria evolve to form intimate and persistent associations with hosts irrespective of effects on host fitness), (ii) the origins of bacterial mutualism from other types of bacterial lifestyles, (iii) shifts to the stable maintenance of bacterial mutualism, (iv) the capture of bacterial mutualists (via the evolution of strict vertical transmission within host lineages), and (v) the evolutionary breakdown of bacterial mutualism. Each of these events has occurred multiple times in the evolution of bacteria. Only symbiont capture possibly constitutes a “major evolutionary transition,” defined as an integrating event in which partners lose the ability to replicate independently (Szathmáry and Smith, 1995). However, loss of independence often only occurs for the symbiont.
To study broad patterns and genetic drivers of transitions, we investigate phylogenomic data that span the bacterial domain (Williams et al., 2007, 2010; Merhej et al., 2009; Wu et al., 2009; Philippot et al., 2010; Toft and Andersson, 2010) (Fig. 2.1), and to study fine-scale patterns, we also analyze a focal set of bacterial mutualists (Table 2.1). Our domain-level data sources include a phylogeny with 350 bacterial taxa sampled from 20 phyla (Wu et al., 2009), coupled with phenotypic host-association data (Boussau et al., 2004; Merhej et al., 2009; Bright and Bulgheresi, 2010; Philippot et al., 2010; Toft and Andersson, 2010). The focal systems include beneficial symbionts chosen to represent host and bacterial diversity, breadth in symbiotic services, and variety in transmission modes. Our analysis of historical and selective scenarios that characterize transitions in bacterial symbiosis complements other work that has focused on genomic changes (Merhej et al., 2009; Carvalho et al., 2010; Medina and Sachs, 2010; Toft and Andersson, 2010). The phylogeny of Wu and colleagues (2009) and the review by Toft and Andersson (2010) are particularly germane to this study as they provide the domain-level dataset that we use to test hypotheses.
There are caveats to consider when inferring the evolutionary history of bacterial symbiosis at broad phylogenetic scales. First is the challenge of assigning host-association traits to bacterial species. Recent work suggests that fitness benefits provided by bacteria to hosts can be context dependent (Heath and Tiffin, 2007; Oliver et al., 2008; Heath et