ANTIBIOTIC RESISTANCE AND STRESS IN THE LIGHT OF FISHER'S MODEL

The role of mutations in evolution depends upon the distribution of their effects on fitness. This distribution is likely to depend on the environment. Indeed genotype‐by‐environment interactions are key for the process of local adaptation and ecological specialization. An important trait in bacterial evolution is antibiotic resistance, which presents a clear case of change in the direction of selection between environments with and without antibiotics. Here, we study the distribution of fitness effects of mutations, conferring antibiotic resistance to Escherichia coli, in benign and stressful environments without drugs. We interpret the distributions in the light of a fitness landscape model that assumes a single fitness peak. We find that mutation effects (s) are well described by a shifted gamma distribution, with a shift parameter that reflects the distance to the fitness peak and varies across environments. Consistent with the theoretical predictions of Fisher's geometrical model, with a Gaussian relationship between phenotype and fitness, we find that the main effect of stress is to increase the variance in s. Our findings are in agreement with the results of a recent meta‐analysis, which suggest that a simple fitness landscape model may capture the variation of mutation effects across species and environments.     

Genetic variance in fitness and its cross-sex covariance predict adaptation during experimental evolution

The additive genetic variation (V_A) of fitness in a population is of particular importance to quantify its adaptive potential and predict its response to rapid environmental change. Recent statistical advances in quantitative genetics and the use of new molecular tools have fostered great interest in estimating fitness V_A in wild populations. However, the value of V_A for fitness in predicting evolutionary changes over several generations remains mostly unknown. In our study, we addressed this question by combining classical quantitative genetics with experimental evolution in the model organism Tribolium castaneum (red flour beetle) in three new environmental conditions (Dry, Hot, Hot-Dry). We tested for potential constraints that might limit adaptation, including environmental and sex genetic antagonisms captured by negative genetic covariance between environments and female and male fitness, respectively. Observed fitness changes after 20 generations mainly matched our predictions. Given that body size is commonly used as a proxy for fitness, we also tested how this trait and its genetic variance (including nonadditive genetic variance) were impacted by environmental stress. In both traits, genetic variances were sex and condition dependent, but they differed in their variance composition, cross-sex and cross-environment genetic covariances, as well as in the environmental impact on V_A.     

THE CONTRIBUTION OF NEW MUTATIONS TO GENOTYPE-ENVIRONMENT INTERACTION FOR FITNESS IN DROSOPHILA MELANOGASTER

Many studies have documented the existence of genotype-environment interaction (GEI) for traits closely related to fitness in natural populations. A type of GEI that is commonly observed is changes in the fitness ranking of genetic groups (families, clones, or inbred lines) in different environments. We refer to such changes in ranking as crossing of reaction norms for fitness. A common interpretation of crossing of reaction norms for fitness is that selection favors different alleles in the different environments (i.e., that “trade-offs” exist). If this is the case, selection could maintain genetic variation, and even lead to reproductive isolation between subpopulations using different environments. Even if the same alleles are favored in every environment, however, deleterious mutations that vary in the magnitude of their effect depending on environment could cause reaction norms for fitness to cross. If deleterious mutations with environment-dependent effects are responsible for maintaining much of the variation leading to crossing of reaction norms for fitness in natural populations, it should be possible to observe crossing of reaction norms for fitness among otherwise genetically identical lines bearing newly arisen spontaneous mutations. We examined the contribution of new mutations to GEI for fitness in Drosophila melanogaster. Eighteen lines were derived from a common, highly inbred base stock, and maintained at a population size of 10 pairs for over 200 generations, to allow them to accumulate spontaneous mutations. Because of the small population size of the lines, selection against mildly deleterious mutations should have been relatively ineffective. The lines were tested for productivity (number of surviving adult progeny from a standard number of parents) in five different environmental treatments, comprising different food media, temperatures, and levels of competition. The lines showed highly significant GEI for productivity, owing largely to considerable changes in ranking in the different environments. We conclude that mutations that are deleterious on average, but whose quantitative effects depend on environment, could be responsible for maintaining much of the variation leading to crossing of reaction norms for fitness that has been observed in samples of D. melanogaster from the wild.     

A POPULATION GENETIC THEORY OF CANALIZATION

Canalization is the suppression of phenotypic variation. Depending on the causes of phenotypic variation, one speaks either of genetic or environmental canalization. Genetic canalization describes insensitivity of a character to mutations, and the insensitivity to environmental factors is called environmental canalization. Genetic canalization is of interest because it influences the availability of heritable phenotypic variation to natural selection, and is thus potentially important in determining the pattern of phenotypic evolution. In this paper a number of population genetic models are considered of a quantitative character under stabilizing selection. The main purpose of this study is to define the population genetic conditions and constraints for the evolution of canalization. Environmental canalization is modeled as genotype specific environmental variance. It is shown that stabilizing selection favors genes that decrease environmental variance of quantitative characters. However, the theoretical limit of zero environmental variance has never been observed. Of the many ways to explain this fact, two are addressed by our model. It is shown that a “canalization limit” is reached if canalizing effects of mutations are correlated with direct effects on the same character. This canalization limit is predicted to be independent of the strength of stabilizing selection, which is inconsistent with recent experimental data (Sterns et al. 1995). The second model assumes that the canalizing genes have deleterious pleiotropic effects. If these deleterious effects are of the same magnitude as all the other mutations affecting fitness very strong stabilizing selection is required to allow the evolution of environmental canalization. Genetic canalization is modeled as an influence on the average effect of mutations at a locus of other genes. It is found that the selection for genetic canalization critically depends on the amount of genetic variation present in the population. The more genetic variation, the stronger the selection for canalizing effects. All factors that increase genetic variation favor the evolution of genetic canalization (large population size, high mutation rate, large number of genes). If genetic variation is maintained by mutation-selection balance, strong stabilizing selection can inhibit the evolution of genetic canalization. Strong stabilizing selection eliminates genetic variation to a level where selection for canalization does not work anymore. It is predicted that the most important characters (in terms of fitness) are not necessarily the most canalized ones, if they are under very strong stabilizing selection (k > 0.2V_e). The rate of decrease of mutational variance V_m is found to be less than 10% of the initial V_m. From this result it is concluded that characters with typical mutational variances of about 10^–3 V_e are in a metastable state where further evolution of genetic canalization is too slow to be of importance at a microevolutionary time scale. The implications for the explanation of macroevolutionary patterns are discussed.     

Quantitative genetic study of the adaptive process

The additive genetic variance with respect to absolute fitness, V_A(W), divided by mean absolute fitness, , sets the rate of ongoing adaptation. Fisher''s key insight yielding this quantitative prediction of adaptive evolution, known as the Fundamental Theorem of Natural Selection, is well appreciated by evolutionists. Nevertheless, extremely scant information about V_A(W) is available for natural populations. Consequently, the capacity for fitness increase via natural selection is unknown. Particularly in the current context of rapid environmental change, which is likely to reduce fitness directly and, consequently, the size and persistence of populations, the urgency of advancing understanding of immediate adaptive capacity is extreme. We here explore reasons for the dearth of empirical information about V_A(W), despite its theoretical renown and critical evolutionary role. Of these reasons, we suggest that expectations that V_A(W) is negligible, in general, together with severe statistical challenges of estimating it, may largely account for the limited empirical emphasis on it. To develop insight into the dynamics of V_A(W) in a changing environment, we have conducted individual-based genetically explicit simulations. We show that, as optimizing selection on a trait changes steadily over generations, V_A(W) can grow considerably, supporting more rapid adaptation than would the V_A(W) of the base population. We call for direct evaluation of V_A(W) and in support of prediction of rates adaptive evolution, and we advocate for the use of aster modeling as a rigorous basis for achieving this goal.     

THE DISTRIBUTION OF MUTATIONAL FITNESS EFFECTS OF PHAGE φX174 ON DIFFERENT HOSTS

Adaptation depends greatly on the distribution of mutation fitness effects (DMFE), but the phenotypic expression of mutations is often environment dependent. The environments faced by multihost pathogens are mostly governed by their hosts and therefore measuring the DMFE on multiple hosts can inform on the likelihood of short‐term establishment and longer term adaptation of emerging pathogens. We explored this by measuring the growth rate of 36 mutants of the lytic bacteriophage φX174 on two host backgrounds, Escherichia coli (EcC) and Salmonella typhimurium (StGal). The DMFE showed higher mean and variance on EcC than on StGal. Most mutations were either deleterious or neutral on both hosts, but a greater proportion of mutations were deleterious on StGal. We identified two mutations with beneficial fitness effects on EcC that were neutral on StGal. Host‐specific differences in fitness were associated with particular functional classes of genes involved in the initial stages of infection in accordance with previous studies of host specificity. Overall, there was a positive correlation between the effects of mutations on each host, suggesting that most new mutations will have general, rather than host‐specific fitness effects. We consider these results in light of simple fitness landscape models of adaptation and discuss the relevance of context‐dependent DMFE for multihost pathogens.     

The Environment Affects Epistatic Interactions to Alter the Topology of an Empirical Fitness Landscape

The fitness effect of mutations can be influenced by their interactions with the environment, other mutations, or both. Previously, we constructed 32 ( = 2⁵) genotypes that comprise all possible combinations of the first five beneficial mutations to fix in a laboratory-evolved population of Escherichia coli. We found that (i) all five mutations were beneficial for the background on which they occurred; (ii) interactions between mutations drove a diminishing returns type epistasis, whereby epistasis became increasingly antagonistic as the expected fitness of a genotype increased; and (iii) the adaptive landscape revealed by the mutation combinations was smooth, having a single global fitness peak. Here we examine how the environment influences epistasis by determining the interactions between the same mutations in two alternative environments, selected from among 1,920 screened environments, that produced the largest increase or decrease in fitness of the most derived genotype. Some general features of the interactions were consistent: mutations tended to remain beneficial and the overall pattern of epistasis was of diminishing returns. Other features depended on the environment; in particular, several mutations were deleterious when added to specific genotypes, indicating the presence of antagonistic interactions that were absent in the original selection environment. Antagonism was not caused by consistent pleiotropic effects of individual mutations but rather by changing interactions between mutations. Our results demonstrate that understanding adaptation in changing environments will require consideration of the combined effect of epistasis and pleiotropy across environments.     

Quantitative genetic variance in experimental fly populations evolving with or without environmental heterogeneity

Heterogeneous environments are typically expected to maintain more genetic variation in fitness within populations than homogeneous environments. However, the accuracy of this claim depends on the form of heterogeneity as well as the genetic basis of fitness traits and how similar the assay environment is to the environment of past selection. Here, we measure quantitative genetic (QG) variance for three traits important for fitness using replicated experimental populations of Drosophila melanogaster evolving under four selective regimes: constant salt‐enriched medium (Salt), constant cadmium‐enriched medium (Cad), and two heterogeneous regimes that vary either temporally (Temp) or spatially (Spatial). As theory predicts, we found that Spatial populations tend to harbor more genetic variation than Temp populations or those maintained in a constant environment that is the same as the assay environment. Contrary to expectation, Salt populations tend to have more genetic variation than Cad populations in both assay environments. We discuss the patterns for QG variances across regimes in relation to previously reported data on genome‐wide sequence diversity. For some traits, the QG patterns are similar to the diversity patterns of ecological selected SNPs, whereas the QG patterns for some other traits resembled that of neutral SNPs.     

Environmental variation alters the fitness effects of rifampicin resistance mutations in Pseudomonas aeruginosa

The fitness effects of antibiotic resistance mutations in antibiotic‐free conditions play a key role in determining the long‐term maintenance of resistance. Although resistance is usually associated with a cost, the impact of environmental variation on the cost of resistance is poorly understood. Here, we test the impact of heterogeneity in temperature and resource availability on the fitness effects of antibiotic resistance using strains of the pathogenic bacterium Pseudomonas aeruginosa carrying clinically important rifampicin resistance mutations. Although the rank order of fitness was generally maintained across environments, fitness effects relative to the wild type differed significantly. Changes in temperature had a profound impact on the fitness effects of resistance, whereas changes in carbon substrate had only a weak impact. This suggests that environmental heterogeneity may influence whether the costs of resistance are likely to be ameliorated by second‐site compensatory mutations or by reversion to wild‐type rpoB. Our results highlight the need to consider environmental heterogeneity and genotype‐by‐environment interactions for fitness in models of resistance evolution.     

Diverse phenotypic and genetic responses to short‐term selection in evolving Escherichia coli populations

Beneficial mutations fuel adaptation by altering phenotypes that enhance the fit of organisms to their environment. However, the phenotypic effects of mutations often depend on ecological context, making the distribution of effects across multiple environments essential to understanding the true nature of beneficial mutations. Studies that address both the genetic basis and ecological consequences of adaptive mutations remain rare. Here, we characterize the direct and pleiotropic fitness effects of a collection of 21 first‐step beneficial mutants derived from naïve and adapted genotypes used in a long‐term experimental evolution of Escherichia coli. Whole‐genome sequencing was able to identify the majority of beneficial mutations. In contrast to previous studies, we find diverse fitness effects of mutations selected in a simple environment and few cases of genetic parallelism. The pleiotropic effects of these mutations were predominantly positive but some mutants were highly antagonistic in alternative environments. Further, the fitness effects of mutations derived from the adapted genotypes were dramatically reduced in nearly all environments. These findings suggest that many beneficial variants are accessible from a single point on the fitness landscape, and the fixation of alternative beneficial mutations may have dramatic consequences for niche breadth reduction via metabolic erosion.     

Estimating the capacity of Chamaecrista fasciculata for adaptation to change in precipitation

Adaptation through natural selection may be the only means by which small and fragmented plant populations will persist through present day environmental change. A population's additive genetic variance for fitness (V_A(W)) represents its immediate capacity to adapt to the environment in which it exists. We evaluated this property for a population of the annual legume Chamaecrista fasciculata through a quantitative genetic experiment in the tallgrass prairie region of the Midwestern United States, where changing climate is predicted to include more variability in rainfall. To reduce incident rainfall, relative to controls receiving ambient rain, we deployed rain exclusion shelters. We found significant V_A(W) in both treatments. We also detected a significant genotype‐by‐treatment interaction for fitness, which suggests that the genetic basis of the response to natural selection will differ depending on precipitation. For the trait‐specific leaf area, we detected maladaptive phenotypic plasticity and an interaction between genotype and environment. Selection for thicker leaves was detected with increased precipitation. These results indicate capacity of this population of C. fasciculata to adapt in situ to environmental change.     

EXPRESSION OF GENETIC VARIATION IN BODY SIZE OF THE COLLARED FLYCATCHER UNDER DIFFERENT ENVIRONMENTAL CONDITIONS

Heritability of body size in two experimentally created environments, representing good and poor feeding conditions, respectively, was estimated using cross-fostered collared flycatcher Ficedula albicollis nestlings. Young raised under poor feeding conditions attained smaller body size (tarsus length) than their full-sibs raised under good feeding conditions. Parent-offspring regressions revealed lower heritability (h²) of body size under poor than under good feeding conditions. Hence, as the same set of parents were used in the estimation of h² in both environments, this suggests environment-dependent change in additive genetic component of variance (V_A), or that the genetic correlation between parental and poor offspring environment was less than that between parental and good offspring environment. However, full-sib analyses failed to find evidence for genotype-environment interactions, although the power of these tests might have been low. Full-sib heritabilities in both environments tended to be higher than estimates from parent-offspring regressions, indicating that prehatching or early posthatching common environment/maternal effects might have inflated full-sib estimates of V_A. The effect of sibling competition on estimates of V_A was probably small as the nestling size-hierarchy at day 2 posthatch was not generally correlated with size-hierarchy at fledging. Furthermore, there was no correlation between maternal body condition during the incubation and final size of offspring, indicating that direct maternal effects related to nutritional status were small. A review of earlier quantitative genetic studies of body size variation in birds revealed that in eight of nine cases, heritability of body size was lower in poor than in good environmental conditions. The main implication of this relationship will be a decreased evolutionary response to selection under poor environmental conditions. On the other hand, this will retard the loss of genetic variation by reducing the accuracy of selection and might help explain the moderate to high heritabilities of body-size traits under good environmental conditions.     

Effect of migration and environmental heterogeneity on the maintenance of quantitative genetic variation: a simulation study

The paradox of high genetic variation observed in traits under stabilizing selection is a long‐standing problem in evolutionary theory, as mutation rates appear too low to explain observed levels of standing genetic variation under classic models of mutation–selection balance. Spatially or temporally heterogeneous environments can maintain more standing genetic variation within populations than homogeneous environments, but it is unclear whether such conditions can resolve the above discrepancy between theory and observation. Here, we use individual‐based simulations to explore the effect of various types of environmental heterogeneity on the maintenance of genetic variation (V_A) for a quantitative trait under stabilizing selection. We find that V_A is maximized at intermediate migration rates in spatially heterogeneous environments and that the observed patterns are robust to changes in population size. Spatial environmental heterogeneity increased variation by as much as 10‐fold over mutation–selection balance alone, whereas pure temporal environmental heterogeneity increased variance by only 45% at max. Our results show that some combinations of spatial heterogeneity and migration can maintain considerably more variation than mutation–selection balance, potentially reconciling the discrepancy between theoretical predictions and empirical observations. However, given the narrow regions of parameter space required for this effect, this is unlikely to provide a general explanation for the maintenance of variation. Nonetheless, our results suggest that habitat fragmentation may affect the maintenance of V_A and thereby reduce the adaptive capacity of populations.     

Field measurements of genotype by environment interaction for fitness caused by spontaneous mutations in Arabidopsis thaliana

As the ultimate source of genetic diversity, spontaneous mutation is critical to the evolutionary process. The fitness effects of spontaneous mutations are almost always studied under controlled laboratory conditions rather than under the evolutionarily relevant conditions of the field. Of particular interest is the conditionality of new mutations—that is, is a new mutation harmful regardless of the environment in which it is found? In other words, what is the extent of genotype–environment interaction for spontaneous mutations? We studied the fitness effects of 25 generations of accumulated spontaneous mutations in Arabidopsis thaliana in two geographically widely separated field environments, in Michigan and Virginia. At both sites, mean total fitness of mutation accumulation lines exceeded that of the ancestors, contrary to the expected decrease in the mean due to new mutations but in accord with prior work on these MA lines. We observed genotype–environment interactions in the fitness effects of new mutations, such that the effects of mutations in Michigan were a poor predictor of their effects in Virginia and vice versa. In particular, mutational variance for fitness was much larger in Virginia compared to Michigan. This strong genotype–environment interaction would increase the amount of genetic variation maintained by mutation‐selection balance.     

Environment changes epistasis to alter trade‐offs along alternative evolutionary paths

The fitness effect of a mutation can depend on both its genetic background, known as epistasis, and the prevailing external environment. Many examples of these dependencies are known, but few studies consider both aspects in combination, especially as they affect mutations that have been selected together. We examine interactions between five coevolved mutations in eight diverse environments. We find that mutations are, on average, beneficial across environments, but that there is high variation in their fitness effects, including many examples of mutations conferring a cost in some, but not other, genetic background‐environment combinations. Indeed, even when global interaction trends are accounted for, specific local mutation interactions are common and differed across environments. One consequence of this dependence is that the range of trade‐offs in genotype fitness across selected and alternative environments are contingent on the particular evolutionary path followed over the mutation landscape. Finally, although specific interactions were common, there was a consistent pattern of diminishing returns epistasis whereby mutation effects were less beneficial when added to genotypes of higher fitness. Our results underline that specific mutation effects are highly dependent on the combination of genetic and external environments, and support a general relationship between a genotype's current fitness and its potential to increase in fitness.     

Adaptive evolution of the lactose utilization network in experimentally evolved populations of Escherichia coli

Adaptation to novel environments is often associated with changes in gene regulation. Nevertheless, few studies have been able both to identify the genetic basis of changes in regulation and to demonstrate why these changes are beneficial. To this end, we have focused on understanding both how and why the lactose utilization network has evolved in replicate populations of Escherichia coli. We found that lac operon regulation became strikingly variable, including changes in the mode of environmental response (bimodal, graded, and constitutive), sensitivity to inducer concentration, and maximum expression level. In addition, some classes of regulatory change were enriched in specific selective environments. Sequencing of evolved clones, combined with reconstruction of individual mutations in the ancestral background, identified mutations within the lac operon that recapitulate many of the evolved regulatory changes. These mutations conferred fitness benefits in environments containing lactose, indicating that the regulatory changes are adaptive. The same mutations conferred different fitness effects when present in an evolved clone, indicating that interactions between the lac operon and other evolved mutations also contribute to fitness. Similarly, changes in lac regulation not explained by lac operon mutations also point to important interactions with other evolved mutations. Together these results underline how dynamic regulatory interactions can be, in this case evolving through mutations both within and external to the canonical lactose utilization network.     

Effects of inbreeding on fitness‐related traits in a small isolated moose population

Inbreeding can affect fitness‐related traits at different life history stages and may interact with environmental variation to induce even larger effects. We used genetic parentage assignment based on 22 microsatellite loci to determine a 25 year long pedigree for a newly established island population of moose with 20–40 reproducing individuals annually. We used the pedigree to calculate individual inbreeding coefficients and examined for effects of individual inbreeding (f) and heterozygosity on fitness‐related traits. We found negative effects of f on birth date, calf body mass and twinning rate. The relationship between f and calf body mass and twinning rate were found to be separate but weaker after accounting for birth date. We found no support for an inbreeding effect on the age‐specific lifetime reproductive success of females. The influence of f on birth date was related to climatic conditions during the spring prior to birth, indicating that calves with a low f were born earlier after a cold spring than calves with high f. In years with a warm spring, calf f did not affect birth date. The results suggest that severe inbreeding in moose has both indirect effects on fitness through delayed birth and lower juvenile body mass, as well as separate direct effects, as there still was a significant relationship between f and twinning rate after accounting for birth date and body mass as calf. Consequently, severe inbreeding as found in the study population may have consequences for population growth and extinction risk.     

The effect of spontaneous mutations on competitive ability

Understanding the impact of spontaneous mutations on fitness has many theoretical and practical applications in biology. Although mutational effects on individual morphological or life‐history characters have been measured in several classic genetic model systems, there are few estimates of the rate of decline due to mutation for complex fitness traits. Here, we estimate the effects of mutation on competitive ability, an important complex fitness trait, in a model system for ecological and evolutionary genomics, Daphnia. Competition assays were performed to compare fitness between mutation‐accumulation (MA) lines and control lines from eight different genotypes from two populations of Daphnia pulicaria after 30 and 65 generations of mutation accumulation. Our results show a fitness decline among MA lines relative to controls as expected, but highlight the influence of genomic background on this effect. In addition, in some assays, MA lines outperform controls providing insight into the frequency of beneficial mutations.     

FREQUENCY-DEPENDENT SELECTION IN DROSOPHILA: ESTIMATION OF NET FITNESS IN PSEUDOHAPLOID POPULATIONS

There is much evidence that the viability of Drosophila larvae depends on the genotypes of other larvae with which they develop. There is, however, little evidence concerning frequency dependence of net fitness. This report documents variation in net fitness among male D. melanogaster from 17 lines studied in 49 discrete-generation populations. The experimental design, which uses attached-X females, eliminates effects of dominance, heterosis, recombination, genetic background, and selection in females. Selection operates only on patroclinously inherited X-chromosomes in males, causing frequency dynamics to mimic those of haploid populations. Analysis of one-generation transitions shows a general pattern of frequency dependence of net fitnesses. Analysis of multigeneration trajectories suggests frequency dependence in 12 populations, but trajectories tend to be heterogeneous among replicates. For a deterministic model in which fitness w(p) is a function of allelic frequency (p), oscillatory dynamics can occur where |dw(p)/dp| > w(p?) (Curtsinger, 1984a). The derivative measures the strength of frequency-dependent selection, and divides models into two categories: dynamically “well-behaved” and “complex.” For the pooled data, this criterion is violated in about half of the frequency space, though several factors probably cause strength to be overestimated. The strength criterion is a useful predictor of evolutionary dynamics: observed oscillations tend to occur most often where frequency dependence is strong, and less often where it is weak.     

Genotype‐by‐environment interactions due to antibiotic resistance and adaptation in Escherichia coli

Mutations that are beneficial in one environment can have different fitness effects in other environments. In the context of antibiotic resistance, the resulting genotype‐by‐environment interactions potentially make selection on resistance unpredictable in heterogeneous environments. Furthermore, resistant bacteria frequently fix additional mutations during evolution in the absence of antibiotics. How do these two types of mutations interact to determine the bacterial phenotype across different environments? To address this, I used Escherichia coli as a model system, measuring the effects of nine different rifampicin resistance mutations on bacterial growth in 31 antibiotic‐free environments. I did this both before and after approximately 200 generations of experimental evolution in antibiotic‐free conditions (LB medium), and did the same for the antibiotic‐sensitive wild type after adaptation to the same environment. The following results were observed: (i) bacteria with and without costly resistance mutations adapted to experimental conditions and reached similar levels of competitive fitness; (ii) rifampicin resistance mutations and adaptation to LB both indirectly altered growth in other environments; and (iii) resistant‐evolved genotypes were more phenotypically different from the ancestor and from each other than resistant‐nonevolved and sensitive‐evolved genotypes. This suggests genotype‐by‐environment interactions generated by antibiotic resistance mutations, observed previously in short‐term experiments, are more pronounced after adaptation to other types of environmental variation, making it difficult to predict long‐term selection on resistance mutations from fitness effects in a single environment.