SEARCH

SEARCH BY CITATION

Keywords:

  • antagonistic coevolution;
  • bacteria;
  • bacteriophage;
  • host;
  • parasite;
  • parasite-mediated coexistence

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results and discussion
  6. Acknowledgments
  7. References

Parasites can promote diversity by mediating coexistence between a poorer and superior competitor, if the superior competitor is more susceptible to parasitism. However, hosts and parasites frequently undergo antagonistic coevolution. This process may result in the accumulation of pleiotropic fitness costs associated with host resistance, and could breakdown coexistence. We experimentally investigated parasite-mediated coexistence of two genotypes of the bacterium Pseudomonas fluorescens, where one genotype underwent coevolution with a parasite (a virulent bacteriophage), whereas the other genotype was resistant to the evolving phages at all time points, but a poorer competitor. In the absence of phages, the resistant genotype was rapidly driven extinct in all populations. In the presence of the phages, the resistant genotype persisted in four of six populations and eventually reached higher frequencies than the sensitive genotype. The coevolving genotype showed a reduction in the growth rate, consistent with a cost of resistance, which may be responsible for a decline in its relative fitness. These results demonstrate that the stability of parasite-mediated coexistence of resistant and susceptible species or genotypes is likely to be affected if parasites and susceptible hosts coevolve.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results and discussion
  6. Acknowledgments
  7. References

Natural enemies have been shown both theoretically and experimentally to play a major role in structuring ecosystem diversity (Bohannan & Lenski, 2000; Chesson, 2000; Buckling & Rainey, 2002b; Chase et al., 2002; Brockhurst et al., 2004; Morgan & Buckling, 2004; Harcombe & Bull, 2005; Vamosi, 2005; Buckling & Hodgson, 2007; Gallet et al., 2007; Meyer & Kassen, 2007; Benmayor et al., 2008) sometimes promoting, and sometimes inhibiting the coexistence of other species or genotypes. One often cited mechanism by which parasites may promote species or genotype coexistence is if there is a trade-off between resistance (behavioural or physiological) to parasites and other fitness-related traits, e.g. growth rate, fecundity and longevity (Holt & Lawton, 1994; Leibold, 1996; Abrams, 1999; Bohannan & Lenski, 2000; Chesson, 2000; Chase et al., 2002). This idea holds for the evolution as well as the maintenance of host diversity (Doebeli & Dieckmann, 2000; Abrams & Chen, 2002; Brockhurst et al., 2004).

It is, however, unclear how antagonistic coevolution (the reciprocal evolution of host defence and parasite counter-defence) between hosts and parasites may affect the coexistence of the coevolving species or genotype with a poorer, but parasite-resistant competitor. Antagonistic coevolution is believed to be ubiquitous (Thompson, 1994) and, crucially, can occur over what are normally considered to be purely ecological time scales (Thompson, 1998). Coevolution could affect the coexistence between two species or genotypes by altering their relative fitnesses. For example, investment into host defence is often associated with pleiotropic fitness costs (Kraaijeveld & Godfray, 1997; Bohannan & Lenski, 2000; Brockhurst et al., 2004; Cotter et al., 2004; Buckling et al., 2006), which could reduce the growth rate advantage of the susceptible organism. If coevolution is characterized by an arms race, where resistance and infectivity increase through time, costs of resistance (and infectivity) may similarly increase (Sasaki, 2000). Coevolution may therefore alter the probability of coexistence through time. Coevolution could also affect the strength of selection imposed by the parasite (hence the relative benefit of host resistance), if virulence (damage caused by the parasite) or the parasite transmission rate altered. Alternatively, coevolution could result in parasites evolving to infect the resistant organism.

Here, we experimentally test the importance of host–parasite antagonistic coevolution on the coexistence of susceptible and resistant hosts. We used a lytic bacteriophage, SBW25Φ2 (Buckling & Rainey, 2002a) as a parasite; the bacterium Pseudomonas fluorescens SBW25 (Rainey & Bailey, 1996) as the susceptible host; and a mutant derived from P. fluorescens SBW25 that is resistant to SBW25Φ2 (including phages that have coevolved with the wild type). Phages invade bacterial cells, and then replicate, and release phage progeny through bacterial cell lysis. As such, there is strong selection for host resistance, which in turn imposes strong selection on phages to overcome this resistance. The interaction between the bacteria and phages results in a coevolutionary arms race, with bacteria becoming resistant to an increasingly wide range of phage genotypes, and phages becoming infective to an increasingly wide range of bacterial genotypes, during the initial stages of coevolution (Buckling & Rainey, 2002a; Brockhurst et al., 2003; Morgan et al., 2005). Resistance to phages is associated with a growth rate cost in the absence of phages in this system (Buckling et al., 2006; Lopez-Pascua & Buckling, 2008). We followed the dynamics over approximately 200 generations of mixtures (initially in a 1 : 1 ratio) of wild-type SBW25 and the resistant mutant in either the presence or absence of the phages. We hypothesized that (i) coexistence between the sensitive and resistant genotypes will be more likely in the presence than the absence of phages, because of a trade-off between growth rate and phage resistance, and (ii) the probability of coexistence will alter through time, as a result of cumulative costs of coevolution between the susceptible host and the phages.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results and discussion
  6. Acknowledgments
  7. References

Setting up coevolving populations

Twelve replicate populations were set up by inoculating microcosms [30 mL glass universals, containing 6 mL of King’s Media B (King et al., 1954) (KB)], with approximately 107 cells of wild-type P. fluorescens SBW25 and approximately 107 cells of a phage-resistant P. fluorescens SBW25 mutant. The mutant is resistant to SBW25Φ2 that has coevolved with wild-type SBW25 bacteria for different durations, and three other naturally isolated phages that can infect wild-type ancestral SBW25. The wild-type P. fluorescens was marked by an entire deletion of the PanB gene, turning it into a pantothenate auxotroph (PanB mutant). This marker has no effects on fitness when grown in an excess of pantothenic acid (Rainey & Travisano, 1998). Approximately 105 particles of SBW25Φ2 (Buckling & Rainey, 2002a) were added to the first six replicates, whereas the other six contained no phages. The microcosms were supplemented with an excess of pantothenic acid (0.0024%) and grown under static conditions, with loose lids at 28 °C, for 48 h. After 48 h, the microcosms were vortexed and 60 μL of culture was transferred to a fresh microcosm. This experiment was continued for a total of 26 experimental transfers. Cultures were frozen in 20% KB/glycerol every two experimental transfers and stored at −85 °C.

Enumeration of host densities

All replicates were serially diluted and plated from frozen onto casein amino acid agar and grown for 48 h at 28 °C. The agar was supplemented with a small amount of pantothenic acid (4.8 × 10−6%) to allow the pantothenic auxotroph (PanB mutants) to grow, but at a much-reduced rate compared to the unmarked phage-resistant mutant. This allowed the easy distinction of the two strains: the wild-type PanB formed small colonies and the resistant mutant formed the larger normal sized colonies (Rainey & Travisano, 1998; Buckling et al., 2000).

Measuring resistance of mutant

Every second transfer, we checked whether the resistant mutant bacteria were still resistant to the phages from their own population (Buckling & Rainey, 2002a). The phages were isolated by adding 100 μL of chloroform to 900 μL of culture, and centrifuged at 13 800 g for 2 min. This lysed and pelleted the bacteria leaving the phages in the supernatant. A 20-μL line of phages was dried onto a KB plate and 20 independent bacterial colonies were streaked across it. If there was any inhibition of the bacteria, it was classed as sensitive to phages, if not, resistant.

Measuring growth rates

At the end of the experiment, at transfer 26, we compared the growth rates in the absence of phages of both the wild type and mutant that evolved in the presence of phages, to their respective ancestors. Eight random colonies each of the coexisting susceptible bacteria and resistant mutant were isolated from each of the replicate populations (those where the mutant persisted). Eight clones were also isolated from cultures of both the ancestral susceptible bacteria, and the ancestral resistant mutant. Three replicate cultures were set up for all of the colonies and grown in KB with an excess of pantothenic acid (0.0024%) at 28 °C. Optical densities were used to measure the increase in cell number at 1-h intervals, and the slope of the log-phase growth rate was used as the measure of growth.

Results and discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results and discussion
  6. Acknowledgments
  7. References

Consistent with our first hypothesis, phages promoted the coexistence of the susceptible and resistant bacteria. Resistant mutants were driven to undetectable levels by the second transfer (14 generations) in all populations in the absence of phages, but coexisted with the susceptible bacteria for the duration of the experiment in four of six populations (Figs 1 and 2; Mann–Whitney U-test of persistence time of mutant in the presence vs. absence of phages, = 21.0, = 0.0031). This clearly demonstrates a fitness cost to this resistance phenotype in the absence of phages. Such costs are frequently associated with phage resistance (Bohannan & Lenski, 2000; Lythgoe & Chao, 2003), and have resulted in the coexistence of sensitive and resistant bacteria in the presence of phages in other studies (Chao et al., 1977; Lenski & Levin, 1985; Bohannan & Lenski, 2000; Lythgoe & Chao, 2003; Brockhurst et al., 2004, 2005, 2006). In the presence of phages, the susceptible populations had a greatly reduced density (approximately 10-fold) compared to populations where the phages were absent (compare Figs 1 and 2). This possibly allowed the initial coexistence of the resistant mutant in the early stages of the experiment when the susceptible bacteria were much fitter. Note that susceptibility of the resistant mutant to phages was never observed.

image

Figure 1.  Dynamics of individual populations without phages [replicates (a)–(f)]. Filled circles and solid lines are wild-type bacteria. Open circles and dashed lines are phage-resistant mutant bacteria.

Download figure to PowerPoint

image

Figure 2.  Dynamics of individual populations with phages [replicates (a)–(f)]. Filled circles and solid lines are wild-type bacteria. Open circles and dashed lines are phage-resistant mutant bacteria.

Download figure to PowerPoint

Despite coexistence in the presence of the phages, there is evidence that the fitness of the resistant mutant relative to the susceptible, coevolving bacteria increased through time. First, resistant mutants were driven to undetectable levels in all populations containing phages by transfer 2, like the populations containing no phages, but then returned to detectable frequencies in all populations by transfer 4. Resistant mutants were subsequently lost in two of six populations (at transfers 6 and 8 respectively; Fig. 2c,f). Second, the proportion of the resistant bacteria in the total population increased in the remaining four populations through time (Fig. 2a,b,d and e; t-test of slopes vs. zero: = 26.84, = 0.0001). These data suggest that coexistence may eventually breakdown if experiments were continued for longer, as a result of exclusion of the initially susceptible bacteria.

There are a number of possible explanations for the increase in the relative fitness of the resistant mutant. First, phage-imposed density reductions of the susceptible bacteria increased through time. This seems unlikely, given that the density of the susceptible bacteria did not decrease through time (Fig. 2; t-tests on slopes vs. zero, = 0.12, = 0.91. Note that data were analysed from transfer 2, by which time carrying capacity in the absence of phages had been reached; Fig. 1). Second, the resistant mutant may have adapted more rapidly to the abiotic elements of the selective environment. This also seems unlikely: despite the population size of the resistant mutant being greater towards the end of the experiment, the susceptible population size was greater at early stages of coevolution. Furthermore, a recent study has demonstrated that coevolution with phages may have resulted in elevated mutation rates of the susceptible bacteria (Pal et al., 2007), potentially increasing the rate of adaptation to the abiotic environment. Third, compensatory mutations to the large cost of resistance may have occurred in the mutant. Such compensatory mutations may also occur in the coevolving susceptible bacteria, but to a lesser degree, as compensatory mutations would have to evolve in response to each novel resistance mutation. Fourth, the continual evolution of resistance to coevolving phages of the ‘susceptible’ bacteria may have resulted in increasing pleiotropic fitness costs. Previous work has demonstrated a decline in the fitness of bacteria in the absence of phages, the more time they spend coevolving with phages (Buckling et al., 2006).

The relative importance of the latter two explanations (compensatory mutations and costs of resistance), which are not mutually exclusive, cannot be determined from the results reported so far. In an attempt to distinguish between these explanations, we measured the growth rate of clones of the susceptible and resistant bacteria at the beginning and end of the experiment. The growth rates of the susceptible coevolving bacteria at the end of the experiment showed that indeed they were significantly lower than the ancestral susceptible bacteria (= 3.38, = 0.043), whereas there was no difference between the evolved and ancestral resistant bacteria (= 2.31, = 0.10). These data suggest that increasing costs of resistance were the primary explanation for the decreasing relative fitness of the susceptible bacteria. Growth rate, however, is not the only component of bacteria fitness.

We are currently unaware of the mechanism of phage resistance, and hence are also unaware of the mechanism responsible for the resultant reduction in bacterial growth rate. We observed that the initially resistant mutant produced a visibly greater amount of cellulose-like polymer compared with the wild type. This may confer resistance by forming a physical barrier between the phages and the receptors on the cell surface. This material may be costly to produce, and so be responsible for the mutant’s poor competitive ability in the absence of phages.

We have shown, like several other studies, that parasites can play a key ecological role in promoting diversity by initially mediating coexistence between competitors. However, we have shown that the population dynamics of competitors can be dramatically altered if the parasitized host coevolves with the parasites. In the long-term this may reduce the diversity that the parasite initially promoted. Given the potential of microbes for very rapid evolution, coevolutionary dynamics with phages should be considered when trying to understand how microbial communities are structured. Furthermore, the high potential for rapid coevolutionary change, in general, suggests coevolution may play a crucial role in structuring communities of all organisms.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results and discussion
  6. Acknowledgments
  7. References

We would like to thank Mike Brockhurst, Pauline Manhes and Laura Lopez Pascua for helpful comments on the manuscript. This work was funded by the Leverhulme Trust and the Royal Society, and ADM was supported by a studentship from the Department of Zoology, University of Oxford.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results and discussion
  6. Acknowledgments
  7. References