Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Microbial Ecology and Functional Diversity of Natural Habitats

Using experimental evolution to explore natural patterns between bacterial motility and resistance to bacteriophages

Abstract

Resistance of bacteria to phages may be gained by alteration of surface proteins to which phages bind, a mechanism that is likely to be costly as these molecules typically have critical functions such as movement or nutrient uptake. To address this potential trade-off, we combine a systematic study of natural bacteria and phage populations with an experimental evolution approach. We compare motility, growth rate and susceptibility to local phages for 80 bacteria isolated from horse chestnut leaves and, contrary to expectation, find no negative association between resistance to phages and bacterial motility or growth rate. However, because correlational patterns (and their absence) are open to numerous interpretations, we test for any causal association between resistance to phages and bacterial motility using experimental evolution of a subset of bacteria in both the presence and absence of naturally associated phages. Again, we find no clear link between the acquisition of resistance and bacterial motility, suggesting that for these natural bacterial populations, phage-mediated selection is unlikely to shape bacterial motility, a key fitness trait for many bacteria in the phyllosphere. The agreement between the observed natural pattern and the experimental evolution results presented here demonstrates the power of this combined approach for testing evolutionary trade-offs.

Introduction

Resistance to parasites is typically thought to be associated with fitness costs, such as reduced growth rate or competitive ability. Such costs can have a key role in maintaining polymorphism in host resistance, and hence parasite persistence within populations (Antonovics and Thrall, 1994; Burdon and Thrall, 2003; Boots and Bowers, 2004; Morgan et al., 2005, 2009), and may also impact on interactions with other species both within and across trophic levels (Clancy and Price, 1986; Omacini et al., 2001; Lennon and Martiny, 2008; Hall et al., 2009). For example, resistance of bacteria to bacteriophages has been associated with substantial fitness costs (Lenski, 1988a; Bohannan et al., 1999), including an increased cost of deleterious mutations (Buckling et al., 2006), and decreased competitive ability (Brockhurst et al., 2005; Lennon et al., 2007; Quance and Travisano, 2009). Given the ubiquity of bacteria–phage interactions, and their key role in all ecosystems, these costs are likely to have important ecological consequences (Bohannan and Lenski, 2000a; Fuhrman and Schwalbach, 2003).

Here, we investigate fitness costs associated with resistance to lytic phages in the bacterial pathogen Pseudomonas syringae using two complementary approaches. First, we carry out a correlational study between resistance to phages and other bacterial fitness traits (growth rate and motility) in natural plant-associated isolates. This approach allows for the direct characterization of phenotypic diversity in natural bacterial and phage populations, but can be difficult to interpret because isolates will inevitably have numerous genetic differences between them in addition to resistance to phages. Next, we test for a causal link between resistance to phages and other fitness traits using a subset of these isolates by experimentally evolving bacteria in the presence and absence of phages. This experimental evolution approach is known to be a powerful tool for studying evolutionary trade-offs among fitness traits (Ebert, 1998; Kassen, 2002; Buckling et al., 2009), but may be less helpful in elucidating the importance of these trade-offs in explaining natural diversity. Combining these two approaches allows for a more robust assessment for the role of trade-offs in maintaining phenotypic diversity in natural populations.

Lytic phages have the potential to impose strong selection on host populations, as they are obligate killers. They replicate by injecting their viral DNA into a host bacterium, hijacking the host replication machinery to propagate and then bursting the host cell to release their viral progeny (Lenski, 1988b). Infection begins with the binding of molecules on the phage tail fiber to a bacterial cell surface receptor (Lindberg, 1973), and resistance to phages can be gained by loss or change of these receptors, which is likely to impact on other bacterial functions (Whitchurch and Mattick, 1994; Brockhurst et al., 2005). For example, phages commonly exploit bacterial surface motility appendages (flagella and pili); flagellatropic phages are known to reversibly bind to helical grooves on the bacterial flagellum and use the rotation of the flagellum to spiral toward the cell surface (Samuel et al., 1999), and similarly, pilus-specific phages will attach to pili and fuse their membrane with that of the bacterial cell during pili retraction (Romantschuk and Bamford, 1985; Mattick, 2002). A first step toward bacterial resistance may therefore be the loss or alteration of these structures. For example, phage-resistant mutants often show defective flagella that are unable to rotate (Icho and Iino, 1978), and abnormal unpiliated or hyperpiliated bacteria may arise to prevent phage attachment (Bradley, 1980; Mattick, 2002; Brockhurst et al., 2005).

Reduced motility function is likely to have important implications for bacterial fitness in both pathogenic and non-pathogenic bacteria (Drake and Montie, 1988; Korber et al., 1994; O'Toole and Kolter, 1998). For example, immotile mutants of the opportunistic animal pathogen P. aeruginosa show reduced infectivity on human hosts and impaired biofilm formation (Drake and Montie, 1988; O'Toole and Kolter, 1998). Similarly, epiphytic, non-motile strains of the plant pathogen P. syringae were found to have reduced fitness and competitive ability compared with more motile strains, especially under conditions of environmental stress (Haefele and Lindow, 1987). Indeed, motility is a key component of fitness for bacteria in the plant phyllosphere and is necessary for successful pathogenicity, as bacteria colonizing leaf surfaces are better able to invade the leaf interior through the stomata if they are motile (Panopoulos and Schroth, 1974; Beattie and Lindow, 1999; Melotto et al., 2006).

Despite the predicted link between phage resistance and motility, and some correlative work suggesting a trade-off between the two (Joys, 1965; Whitchurch and Mattick, 1994), the association has never been systematically investigated. Here, we combine an examination of natural bacteria and phage isolates (from the leaves of horse chestnut trees) with experimental evolution to address this relationship. Contrary to our expectations, although natural populations of Pseudomonads show a positive relationship between bacterial motility and resistance to phages, subsequent experimental work suggests that this relationship is not clearly causal.

Materials and methods

Relationship between resistance and motility in natural populations

To investigate the natural variation in resistance to phage and motility, we examined 80 natural isolates, from either the surface or the interior of horse chestnut leaves collected around Oxfordshire, UK, that were part of a larger sampling design from a previous experiment (Koskella et al., 2011). The study included a reciprocal cross-inoculation of culturable bacteria and communities of phages isolated from each of 32 leaves. Bacteria were isolated from either the leaf surface, using buffer from leaf washes, or leaf interior, using homogenates from surface-sterilized leaves. Washes/homogenates were plated on 1.2% King's medium B (KB) broth (10 g l−1 glycerol, 20 g l−1 proteose peptone (no. 3; Becton Dickinson UK Ltd, Oxford, UK), 12 g l−1 agar, 1.5 g l−1 K2HPO4·3H20, 1.5 g l−1 MgSO4·7H2O) and, after 48 h of growth, colonies were picked at random, based on proximity to a randomly chosen spot on the plate. Phages from the interior and surface of each leaf were separated from bacteria by chloroform treatment of the buffer solutions. This allowed us to generate an inoculum that was representative of the natural phage community, as it did not require passaging through a bacterial host. Using a cross-inoculation design, 7 μl of each phage inoculum was spotted in a grid formation onto a lawn, grown in soft KB agar (0.6% wt/vol), of each bacterial isolate. Phage plaque formation within the spot was compared with bacterial growth across the lawn, allowing us to define each bacterial isolate as either susceptible or resistant to the local, sympatric phage population (that is, to quantify susceptibility to phages from the same leaf). We then randomly chose 40 bacterial isolates from each category (susceptible or resistant) to further characterize and measure motility and growth rate. Of the 80 isolates examined, only 6 pairs showed strong phenotypic and genotypic similarity, and each of these pairs was combined in our statistical analyses.

Bacteria typically exhibit three types of motility: swimming and swarming, which are primarily flagella dependent, and twitching, which is dependent on type IV pili (Mattick, 2002; Harshey, 2003). To examine these motility mechanisms, we measured dispersal capability of each of the bacterial isolates under different agar environments. For each motility assay, bacterial isolates were first grown overnight from freezer stocks in KB broth at 28 °C. A small amount of each culture was then used to inoculate the center of 90 cm2 Petri dishes containing 25 ml of KB medium with the appropriate concentration of agar, dried briefly before use. Twitching motility was assessed on KB medium solidified with 1.2% (wt/vol) agar and estimated using the bacterial movement between the interface of the Petri dish and agar surface, while swarming and swimming motility assays were performed on KB medium containing 0.6% and 0.3% (wt/vol) agar, respectively (Rashid and Kornberg, 2000), and estimated by area dispersed through the agar. The area of dispersal was measured after either 24 h (for swimming) or 48 h (for swarming and twitching) of incubation at 24 °C by demarcating the area covered, photographing the plate with a measurement standard and analyzing the area digitally using ImageJ 1.41o (Abramoff et al., 2004). All area data were square root transformed, and three replicate assays were run for each bacterial isolate.

In vitro growth rate and density assays

We measured the growth rate and final bacterial density for each of the 80 natural bacterial isolates. For growth rate assays, KB cultures were grown overnight at 28 °C and diluted by a factor of 1:100. Then, 10 μl of each dilution was added to a 96-well microplate containing 90 μl of KB per well, and optical density at 600 nm was measured every 45 min at an incubation temperature of 24 °C, with 5 s shaking before reading for 24 h using a microplate spectrophotometer (BioTek Powerwave XS, Northstar Scientific Ltd, Bedfordshire, UK). The period of exponential growth occurred between 4 and 12 h, during which time Vmax (measured as milli-optical density units per minute (mOD min−1)), the maximal rate of change in optical density during log growth, was calculated. For density assays, cultures incubated for 24 h were diluted by a factor of 1:1 (to bring them within a range of measurements for which the readings were most accurate) and measured for optical density at 600 nm.

Characterization of isolates

The identity of each bacterial isolate, to the genus level, was determined by sequencing an 800-bp region of the 16S rRNA subunit using the forward primer 27F (Lane, 1991) and reverse primer 907R (Muyzer et al., 1998). The reaction contained 1 U Taq DNA polymerase (Invitrogen, Paisley, UK), 1X Taq Buffer, 3 mM MgCl2, 0.2 μM deoxyribonucleotide triphosphates, 0.2 pm of each primer and 0.5 μl of a 1:10 dilution of an overnight KB culture. PCR amplification was performed at 95 °C for 4 min, 29 cycles of 95 °C for 45 s, 52 °C for 1 min and 72 °C for 2 min, with a final elongation at 72 °C for 10 min. The product was then sequenced by Geneservice (Oxford, UK) using the reverse primer. These sequence data have been submitted to the GenBank database under accession numbers HQ529384–HQ529465. Each bacterial isolate was assigned to the genus level using the NCBI database, based on highest sequence similarity; all but five of the isolates had over a 97% similarity to a previously characterized isolate (the other five with a similarity between 93 and 95%), and all isolates had an e-value of 0. Isolates were not assigned to the species level because of the highly conserved nature of the sequenced 16S rRNA region.

Selection experiment

To specifically examine how the acquisition of phage resistance might alter motility function, we performed an experimental evolution study using a random subset of the natural isolates. Because of both the observed positive correlation between motility and resistance to phages and the epidemiological significance of the species (Hirano and Upper, 2000; Webber et al., 2008; Green et al., 2009), we chose to focus exclusively on 10 bacterial isolates from the leaf interior that had >99% sequence similarity to known isolates of P. syringae. Importantly, these isolates were all sampled from separate leaves to decrease the probability of pseudoreplication. To experimentally examine the relationship between swarming motility and phages in the environment, we performed 10 serial transfers (approximately 10–16 bacterial generations per transfer) of each bacterial isolate into fresh soft agar (0.6% wt/vol) that contained either high concentrations, low concentrations or no phages (Figure 1). To initiate the experiment, overnight cultures from a single bacterial colony were grown in KB broth, which was then pipetted directly onto the center of a 144-cm2 square Petri-dish (Fisherbrand, Leicestershire, UK) containing 40 ml soft KB agar. Plates were left in a humid incubator at 24 °C for 24 h, after which samples were taken for the next transfer under one of two selection regimes: positive (termed ‘dispersal’ treatment) or neutral (termed ‘random’ treatment) selection for dispersal (Figure 1). For the dispersal selection lines, we took six samples, equally spaced apart, from the outer edge of the colony range. For the random selection bacterial lines, we took six samples from throughout the colony, as chosen by a random number grid. Each sample was taken by stabbing a sterile 1-ml pipette (Finnpipette, Northumberland, UK) through the agar to the bottom of the plate and then transferring the agar stab to 1 ml of M9 solution (1 mM thiamine hydrochloride, 0.4% glycerol, 0.2% casamino acids, 2 mM MgSO4 and 0.1 mM CaCl2). The pipette tip was washed thoroughly in the solution and after vortexing, 2.5 μl of the solution was used to inoculate fresh agar plates, as described above. This was repeated every 24 h for 10 transfers. Plates were poured with fresh media at each transfer so as not to confound age of agar plate with dispersal ability.

Figure 1
figure1

Ten natural isolates were experimentally evolved over ten transfers in one of three phage environments: high phage concentration (40 μl phage inoculum/40 ml 0.6% agar), low phage concentration (4 μl phage inoculum/40 ml 0.6% agar) or no phage treatment (where no phage was added to 0.6% agar). In addition, each line was evolved under one of two selection regimes: positive (termed ‘dispersal’ treatment) or neutral (termed ‘random’ treatment) selection for dispersal. In total, we had 60 selection lines evolved over approximately 140 generations of selection.

The phage inoculum was generated by isolating 16 individual phage types (that is, independent plaques) from 16 different horse chestnut leaves used in the previous cross-inoculation (Koskella et al., 2011). Each phage isolate was passaged once through one of two previously characterized strains of P. syringae pv. aesculi (P. syringae. pv. aesculi 6617 and 6623; Green et al., 2010), to amplify numbers of phage particles, and separated from the bacteria using chloroform. The 16 phage isolates were mixed to produce a stock inoculum, representing a subset of the naturally occurring phage community in the leaf environment, and stored at 4 °C. This design allowed us to hold the phage environment relatively constant, while the bacteria evolved in response. To create a homogeneous selective environment, phage inoculum was vortexed into the soft agar before solidifying, when the agar reached about 40 °C. After 10 transfers, each line was assayed for growth rate and motility in a phage-free common garden (that is, all treatments were grown under the same laboratory environment) and under three agar concentrations. Bacterial densities (optical density at 600 nm) of overnight cultures of each evolved and ancestral strain were also measured in both the presence and absence of phages.

Statistical analyses comparing bacterial dispersal and susceptibility to phages

Analyses and figures were produced on PASW Statistics 18 (SPSS; part of IBM UK Ltd, Middlesex, UK). We first used a two-way analysis of variance to compare the area dispersed (square root transformed) across bacterial isolates of different genera and susceptibility to phages. In addition, independent samples t-tests were run to compare susceptibility to phages and dispersal within the Pseudomonas and Erwinia isolates from both the leaf surface and leaf interior. For the experimental evolution results, we examined the initially susceptible and initially resistant isolates separately, due to the dramatic differences in means and variance between them, and we included bacterial strain as a random factor in each model. At the end of the experiment, we examined the area dispersed within a common garden (in the absence of phages) at the end of the experiment using a separate two-way analysis of variance for (1) the 30 susceptible lines and (2) the 30 resistant lines and compared population growth parameters with dispersal using Pearson's correlation coefficients. In all cases, the area dispersed was square root transformed.

Results

Relationship between resistance, growth and motility in natural populations

Sequencing of the chosen bacterial isolates revealed that the surface of the leaf was primarily dominated by Erwinia-like species (>85%), whereas the leaf interior community was comprised of both Erwinia-like (50%) and Pseudomonas-like species (47%). We excluded bacterial isolates from other genera, including Rhanella and Pantoea, from subsequent analyses due to low replication. Overall, 44% (N/Ntot=24/54) of the Erwinia-like isolates and 61% (N/Ntot=14/23) of Pseudomonas-like isolates were susceptible to sympatric phages, that is, those collected directly from the same leaf as the bacteria being tested. We chose to focus specifically on susceptibility to sympatric phages, using the leaf homogenate as an inoculum, because this measure more accurately reflects local selection pressures and did not require amplification through a bacterial host.

The relationship between swarming motility and resistance to sympatric phages significantly differed across bacterial genera (Figure 2a; interaction effect for genus × phage susceptibility: F1,38=5.99 and P=0.020). Specifically, there was no difference in dispersal capability between resistant and susceptible strains of the Erwinia-like isolates from the leaf surface or interior (Figure 2a). However, for the Pseudomonas-like isolates from the leaf interior (surface isolates were excluded because of low sample size), resistant isolates showed a higher swarming motility range than the susceptible isolates (Figure 2a; t17=2.241 and P=0.039) but did not show a difference for swimming or twitching motilities (swimming: t17=1.143 and P=0.271; twitching: t17=1.219 and P=0.099).

Figure 2
figure2

Relationship between dispersal (a) and growth rate (b) for the 40 Pseudomonas-like and Erwinia-like isolates collected from the leaf interior. Resistance (dark grey) or susceptibility (light grey) to local phage is compared to determine whether there exists a cost to resistance with regard to bacterial motility or growth. Dispersal was measured as area covered (cm2) over 48 h on soft agar (0.6%). Values were square root transformed to correct for non-normality. Error bars represent ±1 s.e.m.

There was no difference in growth rate, measured as Vmax (mOD min−1) between resistant and susceptible Erwinia-like isolates from the surface or the interior (P>0.05), but for Pseudomonas-like isolates from the leaf interior, resistant isolates had a higher growth rate than susceptible isolates (Figure 2b; t17=3.234 and P=0.005). Importantly, the dispersal area during swimming and swarming motility was significantly correlated with growth rate for both the Pseudomonas-like isolates (Pearson's correlation, swimming: r=0.517 and P=0.012; swarming: r=0.669 and P<0.001) and the Erwinia-like isolates (swimming; r=0.317 and P=0.019; swarming: r=0.359 and P=0.008). However, twitching dispersal was not correlated with growth rate for either Pseudomonas-like isolates (r=0.310 and P=0.150) or Erwinia-like isolates (r=−0.100 and P=0.473).

Selection for bacterial resistance to phages and/or motility

Of the 10 Pseudomonad isolates chosen for experimental evolution, 5 were initially susceptible and 5 were initially resistant to the phage inoculum. Of the initially susceptible strains, all were susceptible to at least half of the 16 phage isolates used in the inoculum (mean susceptibility of 70.0% ±19.7 s.d.). Of the five initially resistant bacterial isolates, resistance was complete across all 16 phage isolates (susceptibility of 0%). At the start of the experiment, the presence of phage had a significant negative effect on motility for the initially susceptible strains (F1,24=7.223 and P=0.013), but had no effect on the initially resistant strains (F1,24=1.624 and P=0.215). After 10 transfers of experimental evolution in the presence of phages (at both high and low concentration and across random and dispersal selection regimes), each of the five bacterial isolates that were initially susceptible to phages had evolved resistance to most or all of the 16 phage isolates used in the inoculum (Figure 3). For these isolates, there was no longer an effect of phages in the environment on motility (F1,24=0.400 and P=0.533). These strains did not evolve increased resistance in the absence of phages (general linear model (GLM) for mean proportion of infective phages, effect of time: F1,18=16.56 and P=0.002; phage treatment: F2,18=4.173 and P=0.032; time × phage treatment interaction: F2,18=4.173 and P=0.032). The bacterial isolates that were initially resistant remained resistant throughout the duration of the experiment. Although we did not directly allow for coevolution between bacteria and phages (Bohannan and Lenski, 2000a; Buckling et al., 2009), some degree of coevolution may have occurred between phages that were passively collected along with bacteria at each transfer; creating an additional benefit to dispersal from the inoculation site, as coevolved phages are likely to have increased infectivity to the evolving bacteria (Bohannan and Lenski, 2000a; Brockhurst et al., 2005). However, all resistance assays were run using ancestral phages, as these represented the selection pressure throughout the experimental environment.

Figure 3
figure3

Evolution of resistance for the five initially susceptible isolates over 10 serial transfers in either the presence (dashed lines) or the absence (solid line) of phage. Proportion of infective phage represents susceptibility to each of the 16 phage clones used in the experimental evolution inoculum. Error bars represent ±1 s.e.m.

To confirm our measures of phage resistance, we also measured growth rates of ancestral and evolved lines in the presence of phages. For those strains that were initially resistant, we found no difference in population density in either the presence or absence of phages (GLM with arcsin-sqrt-transformed density, interaction effect of time × phage presence: F1,42=0.206 and P=0.652). However, for those strains that were initially susceptible and evolved resistance to phages over the course of the experiment, the ancestral bacterial lines had a significantly lower density than evolved bacterial lines in the presence of phages, but not in the absence of phages (time × phage presence: F1,42=5.977 and P=0.019), indicating that phages were initially reducing population size of susceptible bacteria but that population size was not affected by phages once resistance had evolved.

We measured area dispersed and growth rate of all isolates at the start of the experiment in the absence of phages and found that the initially resistant bacterial isolates had a higher mean swarming dispersal (mean area dispersed 56.45 cm2±7.05 s.d. in 0.6% agar) than the susceptible isolates (3.14 cm2±3.36s.d.) and that initially resistant isolates had a higher growth rate (mean Vmax 3.49±0.36s.d.) than susceptible isolates (1.55±0.79 s.d.). These results were consistent with the findings from the full sample of natural bacterial isolates.

After the 10 serial transfers of experimental evolution, we again assayed motility in a phage-free environment to examine the effect of both phage and dispersal selection regime on the evolution of dispersal, without the confounding ecological effects of phages (Figure 4). The initially susceptible strains did not show a response to selection for increased dispersal (F1,20=0.333 and P=0.571), and there was no evidence for a direct effect of phage treatment on dispersal (F2,20=0.337 and P=0.718), nor for any interaction between phage and selection (F2,20=0.080 and P=0.923), suggesting a lack of a causal link between phage resistance and motility (Figure 4a). However, the initially resistant strains were able to respond to selection for increased dispersal (F1,20=10.505 and P=0.004), regardless of phage treatment (main effect of phage: F2,20=1.995 and P=0.162; interaction between phage and selection: F2,20=0.514 and P=0.606, Figure 4b). Finally, there were few correlated changes in swimming or twitching motility resulting from either the dispersal or phage selection regimes: the only significant effect, after controlling for multiple tests, was increased swimming motility under positive selection for dispersal (compared with random) in the initially resistant lines (Figure 4b; main effect of selection: F1,20=8.268 and P=0.009).

Figure 4
figure4

Results from the common garden experiment run at the end of the selection experiment (that is, transfer 10) for initially resistant (a) and initially susceptible (b) isolates. Measured on 0.3% agar (swimming; left panel), 0.6% agar (swarming; middle panel) and 1.2% agar (twitching; right panel). These experiments were run in the absence of phage in the environment to examine differences between the lines that are not resulting from interactions with or ecological feedback from phage. ‘Treatment’ therefore represents the phage environment of each line during the course of experimental evolution and not within the common garden experiment. Area dispersed is square root transformed and error bars represent ±1 s.e.m.

We also investigated how the selection regimes affected growth rate of each population, with the specific focus on whether there were costs associated with the acquisition of resistance. We found no evidence that dispersal selection regime or phage-imposed selection affected population growth rate for either initially resistant or initially susceptible bacteria (P>0.10 for all treatments). Importantly, although we found a correlation between growth rate and dispersal for the initially resistant strains (Pearson's r=0.390 and P=0.033), there was no correlation between growth rate and dispersal for the initially susceptible lines (r=0.000 and P=0.999), demonstrating that dispersal estimates were not simply a function of population growth. This result was qualitatively the same when comparing population density with dispersal.

Discussion

We combined an examination of natural phenotypic variation with experimental evolution to investigate the potential link between resistance to phages and bacterial fitness (growth rate and motility). We found that, contrary to expectation, natural bacterial isolates that were resistant to local phages had neither reduced motility nor reduced growth rates relative to those that were susceptible. Generally, there was no relationship between bacterial motility or growth rate and resistance to phages from the local environment (as present in the leaf homogenate) for Erwinia-like isolates and a positive correlation for Pseudomonas-like isolates under favorable laboratory conditions (Figure 2). This result is in line with previous work from marine Cyanobacteria showing that costs of resistance to phage are not ubiquitous and, instead, are highly dependent on the virus and bacteria strain being examined, and on whether strains are in direct competition (Lennon et al., 2007). Importantly, we could not rule out the possibility that more motile and resistant strains represented different species or pathovars than less motile and susceptible strains. Therefore, to investigate any causal links between resistance to phages and bacterial motility, we experimentally evolved 10 Pseudomonas-like bacterial lines in either the presence or absence of phages, and under either positive or random selection for dispersal.

After the 10 serial transfers of experimental evolution, we found that bacteria that were initially susceptible to phages had evolved resistance in the presence, but not the absence, of phages (Figure 3). Moreover, at the start of the experiment, we saw decreased motility in the presence of phages for those strains that were susceptible to infection, but not for those that were initially resistant. This result adds to a growing body of evidence that parasites can have a direct impact on host demography (Cameron et al., 1993; Bradley and Altizer, 2005; Fellous et al., 2010). In addition, we found that bacterial motility decreased over the course of the experiment for most bacterial isolates (regardless of phage treatment and even under selection for increased dispersal), suggesting a cost to motility such that flagella and/or pili function is reduced under favorable laboratory conditions. Importantly, the fact that the initially resistant bacteria showed reduced motility but no loss of resistance to phages in the control (no-phage) treatments suggests that there is no negative correlation between resistance and either growth rate or motility in the laboratory. This overall reduction in motility is initially surprising, given that half the lines were selected for increased dispersal, and suggests that the imposed selection regime was relatively weak, as only the initially resistant, and not the initially susceptible, lines showed increased motility under selection for high dispersal (Figure 4). This result warrants further study as we cannot provide a clear mechanistic explanation, and understanding whether phages directly hinder the response of susceptible bacteria to selection for increased motility is of key interest. In addition, under the no-phage, common garden conditions at the end of the experiment, we found no evidence that phage-imposed selection for resistance in the initially susceptible lines was associated with a change in either the motility trait under selection (swarming) or twitching and swimming motility (Figure 4). Finally, there was no evidence of growth costs associated with the acquisition of resistance (although the strains were never put in direct competition). These data suggest no causal relationship, either positive or negative, between the acquisition of phage resistance and dispersal ability.

A number of other studies, like ours, have found no clear cost in terms of population growth associated with resistance to phages (Lenski, 1988a; Lythgoe and Chao, 2003; Meyer et al., 2010). However, the lack of an observed pattern is somewhat surprising, given the predicted association of many phages with flagella or pili number and function (Icho and Iino, 1978; Bradley, 1980; Mattick, 2002). There are a number of possible explanations. First, phages that use motility organelles (that is, pili and flagella) as attachment sites may be relatively uncommon in the natural leaf environment and hence do not impose very strong selection against bacterial motility. This interpretation may have been influenced by our phage isolation method; chloroform treatment is known to destroy primarily lipid-based phages (Leers, 1969), and if a correlation exists between phage-targeted bacterial receptors and phage coat composition, we could have missed an effect of phage-mediated selection. Second, observed dispersal behaviors may be influenced by traits other than motility organelles, such as cell size, chemotaxis or quorum sensing (Harshey, 2003), which are not altered by the acquisition of phage resistance, a possibility that would indicate bacterial motility can evolve independently of phage-mediated selection. Third, it is plausible that motility-associated costs may only be visible under stressful environmental conditions. For example, previous experimental work has shown that costs of resistance to phages in both Escherichia coli B (Bohannan and Lenski, 2000b) and P. fluorescens (Lopez-Pascua and Buckling, 2008) are increased in nutrient-poor environments. Fourth, it is possible that any costs were rapidly compensated by second site mutations, as is commonly observed with antibiotic resistance (MacLean et al., 2010). However, such rapid compensation would suggest that costs of phage resistance are relatively unimportant in natural populations.

Overall, our results do not support a causal link between the evolution of resistance to phages and bacterial motility in natural populations, although we did find a surprising positive association between phage resistance and bacterial motility for Pseudomonads in the phyllosphere. This of course does not rule out an important role of phage-imposed selection on the evolution of motility, as costs associated with resistance are likely to be contingent upon genetic background, local environment and the precise measures of fitness, but does suggest that the effect of phage-imposed selection is unpredictable and is likely to depend on the natural phage community and bacterial environment. This is an important consideration for phage therapy of pathogenic bacteria, where virulence is often associated with motility traits and any association between phage resistance and bacterial motility could have important consequences over coevolutionary time (Josenhans and Suerbaum, 2002). Understanding these potential costs will be key as phage therapy becomes a more common method for controlling pathogenic bacterial populations (Goodridge, 2004; Levin and Bull, 2004). More generally, these results add to a growing body of work across a range of taxa that points to the wide range of host fitness traits correlated with pathogen resistance (Boots and Begon, 1993; Ferdig et al., 1993; Fellowes et al., 1998; Langand et al., 1998; Williams et al., 1999; Yourth et al., 2002; Lythgoe and Chao, 2003; Sanders et al., 2005; Zhong et al., 2005; Buckling et al., 2006; Morgan et al., 2009). This method of combining natural observations with laboratory selection experiments shows promise in furthering our understanding of the importance of phages, and other selective pressures, as drivers of bacterial evolution and diversity in natural environments.

Accession codes

Accessions

GenBank/EMBL/DDBJ

References

  1. Abramoff MD, Magelhaes PJ, Ram SJ . (2004). Image processing with ImageJ. Biophot Int 11: 36–42.

    Google Scholar 

  2. Antonovics J, Thrall PH . (1994). The cost of resistance and the maintenance of genetic polymorphism in host-pathogen systems. Proc Biol Sci 257: 105–110.

    Article  Google Scholar 

  3. Beattie GA, Lindow SE . (1999). Bacterial colonization of leaves: a spectrum of strategies. Phytopathol 89: 353–359.

    CAS  Article  Google Scholar 

  4. Bohannan BJM, Lenski RE . (2000a). Linking genetic change to community evolution: insights from studies of bacteria and bacteriophage. Ecol Lett 3: 362–377.

    Article  Google Scholar 

  5. Bohannan BJM, Lenski RE . (2000b). The relative importance of competition and predation varies with productivity in a model community. Am Nat 156: 329–340.

    Article  Google Scholar 

  6. Bohannan BJM, Travisano M, Lenski RE . (1999). Epistatic interactions can lower the cost of resistance to multiple consumers. Evolution 53: 292–295.

    Article  Google Scholar 

  7. Boots M, Begon M . (1993). Trade-offs with resistance to a granulosis virus in the Indian meal moth, examined by a laboratory evolution experiment. Funct Ecol 7: 528–534.

    Article  Google Scholar 

  8. Boots M, Bowers RG . (2004). The evolution of resistance through costly acquired immunity. Proc R Soc B Biol Sci 271: 715–723.

    Article  Google Scholar 

  9. Bradley CA, Altizer S . (2005). Parasites hinder monarch butterfly flight: implications for disease spread in migratory hosts. Ecol Lett 8: 290–300.

    Article  Google Scholar 

  10. Bradley DE . (1980). A function of Pseudomonas aeruginosa PAO polar pili: twitching motility. Can J Microbiol 26: 146–154.

    CAS  Article  Google Scholar 

  11. Brockhurst MA, Buckling A, Rainey PB . (2005). The effect of a bacteriophage on diversification of the opportunistic bacterial pathogen, Pseudomonas aeruginosa. Proc R Soc B Biol Sci 272: 1385–1391.

    Article  Google Scholar 

  12. Buckling A, Maclean RC, Brockhurst MA, Colegrave N . (2009). The Beagle in a bottle. Nature 457: 824–829.

    CAS  Article  Google Scholar 

  13. Buckling A, Wei Y, Massey RC, Brockhurst MA, Hochberg ME . (2006). Antagonistic coevolution with parasites increases the cost of host deleterious mutations. Proc R Soc B Biol Sci 273: 45–49.

    Article  Google Scholar 

  14. Burdon J, Thrall P . (2003). The fitness costs to plants of resistance to pathogens. Genome Biol 4: 227.

    Article  Google Scholar 

  15. Cameron PG, Semlitsch RD, Bernasconi MV . (1993). Effects of body size and parasite infection on the locomotory performance of juvenile toads, Bufo bufo. Oikos 66: 129–136.

    Article  Google Scholar 

  16. Clancy KM, Price PW . (1986). Temporal variation in three-trophic-level interactions among willows, sawflies, and parasites. Ecology 67: 1601–1607.

    Article  Google Scholar 

  17. Drake D, Montie TC . (1988). Flagella, motility and invasive virulence of Pseudomonas aeruginosa. J Gen Microbiol 134: 43–52.

    CAS  PubMed  Google Scholar 

  18. Ebert D . (1998). Experimental evolution of parasites. Science 282: 1432–1436.

    CAS  Article  Google Scholar 

  19. Fellous S, Quillery E, Duncan AB, Kaltz O . (2010). Parasitic infection reduces dispersal of ciliate host. Biol Lett; e-pub ahead of print.

  20. Fellowes D, Kraaijeveld AR, Godfray HC . (1998). Trade-off associated with selection for increased ability to resist parasitoid attack in Drosophila melanogaster. Proc R Soc B Biol Sci 265: 1553–1558.

    CAS  Article  Google Scholar 

  21. Ferdig MT, Beerntsen BT, Spray FJ, Li J, Christensen BM . (1993). Reproductive costs associated with resistance in a mosquito-filarial worm system. Am J Trop Med Hyg 49: 756–762.

    CAS  Article  Google Scholar 

  22. Fuhrman JA, Schwalbach M . (2003). Viral influence on aquatic bacterial communities. Biol Bull 204: 192–195.

    CAS  Article  Google Scholar 

  23. Goodridge LD . (2004). Bacteriophage biocontrol of plant pathogens: fact or fiction? Trends Biotechnol 22: 384–385.

    CAS  Article  Google Scholar 

  24. Green S, Laue B, Fossdal CG, A'Hara SW, Cottrell JE . (2009). Infection of horse chestnut (Aesculus hippocastanum) by Pseudomonas syringae pv. aesculi and its detection by quantitative real-time PCR. Plant Pathol 58: 731–744.

    CAS  Article  Google Scholar 

  25. Green S, Studholme DJ, Laue BE, Dorati F, Lovell H, Arnold D et al. (2010). Comparative genome analysis provides insights into the evolution and adaptation of Pseudomonas syringae pv. aesculi on Aesculus hippocastanum. PLoS ONE 5: e10224.

    Article  Google Scholar 

  26. Haefele DM, Lindow SE . (1987). Flagellar motility confers epiphytic fitness advantages upon Pseudomonas syringae. Appl Environ Microbiol 53: 2528–2533.

    CAS  PubMed  PubMed Central  Google Scholar 

  27. Hall SR, Simonis JL, Nisbet RM, Tessier AJ, Caceres CE . (2009). Resource Ecology of Virulence in a Planktonic Host-Parasite System: An Explanation Using Dynamic Energy Budgets. University of Chicago Press: Chicago, IL.

    Google Scholar 

  28. Harshey RM . (2003). Bacterial motility on a surface: many ways to a common goal. Annu Rev Microbiol 57: 249–273.

    CAS  Article  Google Scholar 

  29. Hirano SS, Upper CD . (2000). Bacteria in the leaf ecosystem with emphasis on Pseudomonas syringae—a pathogen, ice nucleus, and epiphyte. Microbiol Mol Biol Rev 64: 624–653.

    CAS  Article  Google Scholar 

  30. Icho T, Iino T . (1978). Isolation and characterization of motile Escherichia coli mutants resistant to bacteriophage chi. J Bacteriol 134: 854–860.

    CAS  PubMed  PubMed Central  Google Scholar 

  31. Josenhans C, Suerbaum S . (2002). The role of motility as a virulence factor in bacteria. Int J Med Microbiol 291: 605–614.

    CAS  Article  Google Scholar 

  32. Joys TM . (1965). Correlation between susceptibility to bacteriophage PBS1 and motility in Bacillus subtilis. J Bacteriol 90: 1575–1577.

    CAS  PubMed  PubMed Central  Google Scholar 

  33. Kassen R . (2002). The experimental evolution of specialists, generalists, and the maintenance of diversity. J Evol Biol 15: 173–190.

    Article  Google Scholar 

  34. Korber DR, Lawrence JR, Caldwell DE . (1994). Effect of motility on surface colonization and reproductive success of Pseudomonas fluorescens in dual-dilution continuous culture and batch culture systems. Appl Environ Microbiol 60: 1421–1429.

    CAS  PubMed  PubMed Central  Google Scholar 

  35. Koskella B, Thompson JN, Preston GM, Buckling A . (2011). Local biotic environment shapes the spatial scale of bacteriophage adaptation to bacteria. Am Nat 177: 440–451.

    Article  Google Scholar 

  36. Lane DJ . (1991). 16S/23S rRNA sequencing. In: Stackebrandt E, Goodfellow M (eds). Nucleic Acid Techniques in Bacterial Systematics. Wiley Chichester: NY, pp 115–175.

    Google Scholar 

  37. Langand J, Jourdane J, Coustau C, Delay B, Morand S . (1998). Cost of resistance, expressed as a delayed maturity, detected in the host-parasite system Biomphalaria glabrata-Echinostoma caproni. Heredity 80: 320–325.

    Article  Google Scholar 

  38. Leers WD . (1969). Action of chloroform on the hemagglutinin of ECHO virus types 7 and 11. Arch Virol 28: 116–121.

    CAS  Google Scholar 

  39. Lennon J, Khatana S, Marston M, Martiny J . (2007). Is there a cost of virus resistance in marine cyanobacteria? ISME 1: 300–312.

    Article  Google Scholar 

  40. Lennon JT, Martiny JBH . (2008). Rapid evolution buffers ecosystem impacts of viruses in a microbial food web. Ecol Lett 11: 1178–1188.

    Article  Google Scholar 

  41. Lenski RE . (1988a). Experimental studies of pleiotropy and epistasis in Escherichia coli. I. Variation in competitive fitness among mutants resistant to virus T4. Evolution 42: 425–432.

    PubMed  PubMed Central  Google Scholar 

  42. Lenski RE . (1988b). Dynamics of interactions between bacteria and virulent bacteriophage. Adv Microb Ecol 10: 1–44.

    CAS  Article  Google Scholar 

  43. Levin BR, Bull JJ . (2004). Population and evolutionary dynamics of phage therapy. Nat Rev Micro 2: 166–173.

    CAS  Article  Google Scholar 

  44. Lindberg AA . (1973). Bacteriophage receptors. Annu Rev Microbiol 27: 205–241.

    CAS  Article  Google Scholar 

  45. Lopez-Pascua LC, Buckling A . (2008). Increasing productivity accelerates host-parasite coevolution. J Evol Biol 21: 853–860.

    CAS  Article  Google Scholar 

  46. Lythgoe KA, Chao L . (2003). Mechanisms of coexistence of a bacteria and a bacteriophage in a spatially homogeneous environment. Ecol Lett 6: 326–334.

    Article  Google Scholar 

  47. MacLean RC, Hall AR, Perron GG, Buckling A . (2010). The population genetics of antibiotic resistance: integrating molecular mechanisms and treatment contexts. Nat Rev Genet 11: 405–414.

    CAS  Article  Google Scholar 

  48. Mattick JS . (2002). Type IV pili and twitching motility. Annu Rev Microbiol 56: 289–314.

    CAS  Article  Google Scholar 

  49. Melotto M, Underwood W, Koczan J, Nomura K, He SY . (2006). Plant stomata function in innate immunity against bacterial invasion. Cell 126: 969–980.

    CAS  Article  Google Scholar 

  50. Meyer JR, Agrawal AA, Quick RT, Dobias DT, Schneider D, Lenski RE . (2010). Parallel changes in host resistance to viral infection during 45,000 generations of relaxed selection. Evolution 64: 3024–3034.

    CAS  PubMed  Google Scholar 

  51. Morgan AD, Gandon S, Buckling A . (2005). The effect of migration on local adaptation in a coevolving host-parasite system. Nature 437: 253–256.

    CAS  Article  Google Scholar 

  52. Morgan AD, Maclean RC, Buckling A . (2009). Effects of antagonistic coevolution on parasite-mediated host coexistence. J Evol Biol 22: 287–292.

    CAS  Article  Google Scholar 

  53. Muyzer G, Brinkhoff T, Nübel U, Santegoeds C, Schäfer H, Wawer C . (1998). Denaturing gradient gel electrophoresis (DGGE) in microbial ecology. In: Akkermans ADL, van Elsas JD, de Bruijn FJ (eds). Molecular Microbial Ecology Manual. Kluwer: The Netherlands, pp 1–23.

    Google Scholar 

  54. O'Toole GA, Kolter R . (1998). Flagellar and twitching motility are necessary for Pseudomonas aeruginosa biofilm development. Mol Microbiol 30: 295–304.

    CAS  Article  Google Scholar 

  55. Omacini M, Chaneton EJ, Ghersa CM, Ller CB . (2001). Symbiotic fungal endophytes control insect host-parasite interaction webs. Nature 409: 78–81.

    CAS  Article  Google Scholar 

  56. Panopoulos NJ, Schroth MN . (1974). Role of flagellar motility in the invasion of bean leaves by Pseudomonas phaseolicola. Phytopathology 64: 1389–1397.

    Article  Google Scholar 

  57. Quance MA, Travisano M . (2009). Effects of temperature on the fitness cost of resistance to bacteriophage T4 in Escherichia coli. Evolution 63: 1406–1416.

    Article  Google Scholar 

  58. Rashid MH, Kornberg A . (2000). Inorganic polyphosphate is needed for swimming, swarming, and twitching motilities of Pseudomonas aeruginosa. Proc Natl Acad Sci USA 97: 4885–4890.

    CAS  Article  Google Scholar 

  59. Romantschuk M, Bamford DH . (1985). Function of pili in bacteriophage Φ6 penetration. J Gen Virol 66: 2461–2469.

    CAS  Article  Google Scholar 

  60. Samuel AD, Pitta TP, Ryu WS, Danese PN, Leung EC, Berg HC . (1999). Flagellar determinants of bacterial sensitivity to chi-phage. Proc Natl Acad Sci USA 96: 9863–9866.

    CAS  Article  Google Scholar 

  61. Sanders AE, Scarborough C, Layen SJ, Kraaijeveld AR, Godfray HCJ . (2005). Evolutionary change in parasitoid resistance under crowded conditions in Drosophila melanogaster. Evolution 59: 1292–1299.

    Article  Google Scholar 

  62. Webber J, Parkinson NM, Rose J, Stanford H, Cook RTA, Elphinstone JG . (2008). Isolation and identification of Pseudomonas syringae pv. aesculi causing bleeding canker of horse chestnut in the UK. Plant Pathol 57: 368.

    Article  Google Scholar 

  63. Whitchurch CB, Mattick JS . (1994). Characterization of a gene, pilU, required for twitching motility but not phage sensitivity in Pseudomonas aeruginosa. Mol Microbiol 13: 1079–1091.

    CAS  Article  Google Scholar 

  64. Williams TD, Christians JK, Aiken JJ, Evanson M . (1999). Enhanced immune function does not depress reproductive output. Proc R Soc B Biol Sci 266: 753–757.

    Article  Google Scholar 

  65. Yourth CP, Forbes MR, Smith BP . (2002). Immune expression in a damselfly is related to time of season, not to fluctuating asymmetry or host size. Ecol Entomol 27: 123–128.

    Article  Google Scholar 

  66. Zhong D, Pai A, Yan G . (2005). Costly resistance to parasitism: Evidence from simultaneous quantitative trait loci mapping for resistance and fitness in Tribolium castaneum. Genetics 169: 2127–2135.

    CAS  Article  Google Scholar 

Download references

Acknowledgements

We thank G Preston and M Boots for helpful discussion. This work was supported by funding granted by the European Research Council (AB) and the National Science Foundation (BK, DEB-0754399).

Author information

Affiliations

Authors

Corresponding author

Correspondence to Tiffany B Taylor.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Koskella, B., Taylor, T., Bates, J. et al. Using experimental evolution to explore natural patterns between bacterial motility and resistance to bacteriophages. ISME J 5, 1809–1817 (2011). https://doi.org/10.1038/ismej.2011.47

Download citation

Keywords

  • coevolution
  • dispersal
  • host–parasite
  • trade-off
  • cost of resistance

Further reading

Search

Quick links