Evolution of mutualism from parasitism
in experimental virus populations
Jason W. Shapiro
and Paul E. Turner
Department of Ecology and Evolutionary Biology, Yale University, New Haven, Connecticut 06520
Current Address: Department of Biology, Loyola University Chicago, 1032 W Sheridan Rd., Chicago, Illinois 60660
Program in Microbiology, Yale School of Medicine, New Haven, Connecticut 06510
Received April 18, 2017
Accepted January 16, 2018
While theory suggests conditions under which mutualism may evolve from parasitism, few studies have observed this transition
empirically. Previously, we evolved Escherichia coli and the ﬁlamentous bacteriophage M13 in 96-well microplates, an environment
in which the ancestral phage increased the growth rate and yield of the ancestral bacteria. In the majority of populations,
mutualism was maintained or even enhanced between phages and coevolving bacteria; however, these same phages evolved
traits that harmed the ancestral E. coli genotype. Here, we set out to determine if mutualism could evolve from this new parasitic
interaction. To do so, we chose six evolved phage populations from the original experiment and used them to establish new
infections of the ancestral bacteria. After 20 passages, mutualism evolved in almost all replicates, with the remainder growing
commensally. Many phage populations also evolved to beneﬁt both their local, evolving bacteria and the ancestral bacteria,
though these phages were less beneﬁcial to their co-occurring hosts than phages that harm the ancestral bacteria. These results
demonstrate the rapid recovery of mutualism from parasitism, and we discuss how our ﬁndings relate to the evolution of phages
that enhance the virulence of bacterial pathogens.
Bacteriophage, coevolution, symbiosis, tradeoff.
Microbial symbioses are often thought to exist along a parasitism-
mutualism continuum. How readily an interaction can evolve from
mutualism to parasitism or vice versa depends on many factors,
including the abiotic environment, the ecological, and evolution-
ary history of the interaction, the specificity of the host-symbiont
pair, and the mechanisms behind the benefits received by both
partners (Ewald 1987; Michalakis et al. 1992; Edwards 2009;
Oliver et al. 2009; Sachs et al. 2011). In the case of parasitic
interactions, the relative availability of vertical and horizontal
symbiont transmission is often thought to correspond to the de-
gree of parasite virulence (e.g., Turner et al. 1998). Past work has
shown that mutualists can evolve from parasites and vice versa
(Weeks et al. 2007; Sachs et al. 2010, 2011; Shapiro et al. 2016).
Relatively few studies, though, have explored the evolution of
mutualism breakdown experimentally, and we are aware of no
prior work that has explored a transition back to mutualism after
the evolution of parasitism.
The filamentous phage M13 provides an excellent model
system for studying the evolution of host-symbiont interactions.
Unlike other phage families, filamentous phages establish chronic
infections in their bacterial hosts and are typically secreted by in-
fected cells without causing host death (Marvin and Hohn 1969).
Instead, phage-encoded proteins embedded in the host membrane,
as well as the metabolic costs of phage replication and secretion,
typically results in reduced host growth rates (Mai-Prochnow et al.
2015). These phages can also be vertically transmitted for several
generations (Lerner and Model 1981). Together, these features
make it possible to study transitions between mutualism and par-
asitism using experimental evolution (Bull et al. 1991; Shapiro
et al. 2016).
In a previous study, we found that “wild-type” M13 in-
creases the growth rate and yield of Escherichia coli in 96-well
microplates despite decreasing bacterial growth in well-mixed Er-
lenmeyer flasks (Shapiro et al. 2016). This conditional mutualism
2018 The Author(s). Evolution
2018 The Society for the Study of Evolution.
Evolution 72-3: 707–712