Epistasis and the Evolution of Antimicrobial Resistance

Limited Evolutionary Conservation of the Phenotypic Effects of Antibiotic Resistance Mutations

Multidrug-resistant clinical isolates are common in certain pathogens, but rare in others. This pattern may be due to the fact that mutations shaping resistance have species-specific effects. To investigate this issue, we transferred a range of resistance-conferring mutations and a full resistance gene into Escherichia coli and closely related bacteria. We found that resistance mutations in one bacterial species frequently provide no resistance, in fact even yielding drug hypersensitivity in close relatives. In depth analysis of a key gene involved in aminoglycoside resistance (trkH) indicated that preexisting mutations in other genes-intergenic epistasis-underlie such extreme differences in mutational effects between species. Finally, reconstruction of adaptive landscapes under multiple antibiotic stresses revealed that mutations frequently provide multidrug resistance or elevated drug susceptibility (i.e., collateral sensitivity) only with certain combinations of other resistance mutations. We conclude that resistance and collateral sensitivity are contingent upon the genetic makeup of the bacterial population, and such contingency could shape the long-term fate of resistant bacteria. These results underlie the importance of species-specific treatment strategies.

Pan-Genomic and Polymorphic Driven Prediction of Antibiotic Resistance in Elizabethkingia

The Elizabethkingia are a genetically diverse genus of emerging pathogens that exhibit multidrug resistance to a range of common antibiotics. Two representative species, Elizabethkingia bruuniana and E. meningoseptica, were phenotypically tested to determine minimum inhibitory concentrations (MICs) for five antibiotics. Ultra-long read sequencing with Oxford Nanopore Technologies (ONT) and subsequent de novo assembly produced complete, gapless circular genomes for each strain. Alignment based annotation with Prokka identified 5,480 features in E. bruuniana and 5,203 features in E. meningoseptica, where none of these identified genes or gene combinations corresponded to observed phenotypic resistance values. Pan-genomic analysis, performed with an additional 19 Elizabethkingia strains, identified a core-genome size of 2,658,537 bp, 32 uniquely identifiable intrinsic chromosomal antibiotic resistance core-genes and 77 antibiotic resistance pan-genes. Using core-SNPs and pan-genes in combination with six machine learning (ML) algorithms, binary classification of clindamycin and vancomycin resistance achieved f1 scores of 0.94 and 0.84, respectively. Performance on the more challenging multiclass problem for fusidic acid, rifampin and ciprofloxacin resulted in f1 scores of 0.70, 0.75, and 0.54, respectively. By producing two sets of quality biological predictors, pan-genome genes and core-genome SNPs, from long-read sequence data and applying an ensemble of ML techniques, our results demonstrated that accurate phenotypic inference, at multiple AMR resolutions, can be achieved.

On the evolutionary ecology of multidrug resistance in bacteria

Resistance against different antibiotics appears on the same bacterial strains more often than expected by chance, leading to high frequencies of multidrug resistance. There are multiple explanations for this observation, but these tend to be specific to subsets of antibiotics and/or bacterial species, whereas the trend is pervasive. Here, we consider the question in terms of strain ecology: explaining why resistance to different antibiotics is often seen on the same strain requires an understanding of the competition between strains with different resistance profiles. This work builds on models originally proposed to explain another aspect of strain competition: the stable coexistence of antibiotic sensitivity and resistance observed in a number of bacterial species. We first identify a partial structural similarity in these models: either strain or host population structure stratifies the pathogen population into evolutionarily independent sub-populations and introduces variation in the fitness effect of resistance between these sub-populations, thus creating niches for sensitivity and resistance. We then generalise this unified underlying model to multidrug resistance and show that models with this structure predict high levels of association between resistance to different drugs and high multidrug resistance frequencies. We test predictions from this model in six bacterial datasets and find them to be qualitatively consistent with observed trends. The higher than expected frequencies of multidrug resistance are often interpreted as evidence that these strains are out-competing strains with lower resistance multiplicity. Our work provides an alternative explanation that is compatible with long-term stability in resistance frequencies.

Effects of a previously selected antibiotic resistance on mutations acquired during development of a second resistance in Escherichia coli

Background

The effect of mutations conferring antibiotic resistance can depend on the genetic background. To determine if a previously de novo acquired antibiotic resistance influences the adaptation to a second antibiotic, antibiotic resistance was selected for by exposure to stepwise increasing sublethal levels of amoxicillin, enrofloxacin, kanamycin, or tetracycline. E. coli populations adapted to either a single or two antibiotics sequentially were characterized using whole genome population sequencing and MIC measurements.

Results

In a wild-type background, adaptation to any of the antibiotics resulted in the appearance of well-known mutations, as well as a number of mutated genes not known to be associated with antibiotic resistance. Development of a second resistance in a strain with an earlier acquired resistance to a different antibiotic did not always result in the appearance of all mutations associated with resistance in a wild-type background. In general, a more varied set of mutations was acquired during secondary adaptation. The ability of E. coli to maintain the first resistance during this process depended on the combination of antibiotics used. The maintenance of mutations associated with resistance to the first antibiotic did not always predict the residual MIC for that compound.

Conclusions

In general, the data presented here indicate that adaptation to each antibiotic is unique and independent. The mutational trajectories available in already resistant cells appear more varied than in wild-type cells, indicating that the genetic background of E. coli influences resistance development. The observed mutations cannot always fully explain the resistance pattern observed, indicating a crucial role for adaptation on the gene expression level in de novo acquisition of antibiotic resistance.

Mycoplasmas under experimental antimicrobial selection: The unpredicted contribution of horizontal chromosomal transfer

Horizontal Gene Transfer was long thought to be marginal in Mycoplasma a large group of wall-less bacteria often portrayed as minimal cells because of their reduced genomes (ca. 0.5 to 2.0 Mb) and their limited metabolic pathways. This view was recently challenged by the discovery of conjugative exchanges of large chromosomal fragments that equally affected all parts of the chromosome via an unconventional mechanism, so that the whole mycoplasma genome is potentially mobile. By combining next generation sequencing to classical mating and evolutionary experiments, the current study further explored the contribution and impact of this phenomenon on mycoplasma evolution and adaptation using the fluoroquinolone enrofloxacin (Enro), for selective pressure and the ruminant pathogen Mycoplasma agalactiae, as a model organism. For this purpose, we generated isogenic lineages that displayed different combination of spontaneous mutations in Enro target genes (gyrA, gyrB, parC and parE) in association to gradual level of resistance to Enro. We then tested whether these mutations can be acquired by a susceptible population via conjugative chromosomal transfer knowing that, in our model organism, the 4 target genes are scattered in three distinct and distant loci. Our data show that under antibiotic selective pressure, the time scale of the mutational pathway leading to high-level of Enro resistance can be readily compressed into a single conjugative step, in which several EnroR alleles were transferred from resistant to susceptible mycoplasma cells. In addition to acting as an accelerator for antimicrobial dissemination, mycoplasma chromosomal transfer reshuffled genomes beyond expectations and created a mosaic of resistant sub-populations with unpredicted and unrelated features. Our findings provide insights into the process that may drive evolution and adaptability of several pathogenic Mycoplasma spp. via an unconventional conjugative mechanism.

Insights into the processes that drive the evolution of drug resistance in Mycobacterium tuberculosis

At present, the successful transmission of drug-resistant Mycobacterium tuberculosis, including multidrug-resistant (MDR) and extensively drug-resistant (XDR) strains, in human populations, threatens tuberculosis control worldwide. Differently from many other bacteria, M. tuberculosis drug resistance is acquired mainly through mutations in specific drug resistance-associated genes. The panel of mutations is highly diverse, but depends on the affected gene and M. tuberculosis genetic background. The variety of genetic profiles observed in drug-resistant clinical isolates underlines different evolutionary trajectories towards multiple drug resistance, although some mutation patterns are prominent. This review discusses the intrinsic processes that may influence drug resistance evolution in M. tuberculosis, such as mutation rate, drug resistance-associated mutations, fitness cost, compensatory mutations and epistasis. This knowledge should help to better predict the risk of emergence of highly resistant M. tuberculosis strains and to develop new tools and strategies to limit the development and spread of MDR and XDR strains.

Genetics and roadblocks of drug resistant tuberculosis

Considering the extensive evolutionary history of Mycobacterium tuberculosis, anti-Tuberculosis (TB) drug therapy exerts a recent selective pressure. However, in a microorganism devoid of horizontal gene transfer and with a strictly clonal populational structure such as M. tuberculosis the usual, but not sole, path to overcome drug susceptibility is through de novo mutations on a relatively strict set of genes. The possible allelic diversity that can be associated with drug resistance through several mechanisms such as target alteration or target overexpression, will dictate how these genes can become associated with drug resistance. The success demonstrated by this pathogenic microbe in this latter process and its ability to spread is currently one of the major obstacles to an effective TB elimination. This article reviews the action mechanism of the more important anti-TB drugs, including bedaquiline and delamanid, along with new findings on specific resistance mechanisms. With the development, validation and endorsement of new in vitro molecular tests for drug resistance, knowledge on these resistance mechanisms and microevolutionary dynamics leading to the emergence and fixation of drug resistance mutations within the host is highly important. Additionally, the fitness toll imposed by resistance development is also herein discussed together with known compensatory mechanisms. By elucidating the possible mechanisms that enable one strain to reacquire the original fitness levels, it will be theoretically possible to make more informed decisions and develop novel strategies that can force M. tuberculosis microevolutionary trajectory down through a path of decreasing fitness levels.

Epistatic control of intrinsic resistance by virulence genes in Listeria

Elucidating the relationships between antimicrobial resistance and virulence is key to understanding the evolution and population dynamics of resistant pathogens. Here, we show that the susceptibility of the gram-positive bacterium Listeria monocytogenes to the antibiotic fosfomycin is a complex trait involving interactions between resistance and virulence genes and the environment. We found that a FosX enzyme encoded in the listerial core genome confers intrinsic fosfomycin resistance to both pathogenic and non-pathogenic Listeria spp. However, in the genomic context of the pathogenic L. monocytogenes, FosX-mediated resistance is epistatically suppressed by two members of the PrfA virulence regulon, hpt and prfA, which upon activation by host signals induce increased fosfomycin influx into the bacterial cell. Consequently, in infection conditions, most L. monocytogenes isolates become susceptible to fosfomycin despite possessing a gene that confers high-level resistance to the drug. Our study establishes the molecular basis of an epistatic interaction between virulence and resistance genes controlling bacterial susceptibility to an antibiotic. The reported findings provide the rationale for the introduction of fosfomycin in the treatment of Listeria infections even though these bacteria are intrinsically resistant to the antibiotic in vitro.

Community interactions and spatial structure shape selection on antibiotic resistant lineages

Polymicrobial interactions play an important role in shaping the outcome of antibiotic treatment, yet how multispecies communities respond to antibiotic assault is still little understood. Here we use an individual-based simulation model of microbial biofilms to investigate how competitive and mutualistic interactions between an antibiotic-resistant and a susceptible strain (or species) influence the two-lineage community response to antibiotic exposure. Our model predicts that while increasing competition and antibiotics leads to increasing competitive release of the antibiotic-resistant strain, hitting a mutualistic community of cross-feeding species with antibiotics leads to a mutualistic suppression effect where both susceptible and resistant species are harmed. We next show that the impact of antibiotics is further governed by emergent spatial feedbacks within communities. Mutualistic cross-feeding communities can rescue susceptible members by subsidizing their growth inside the biofilm despite lack of access to the nutrient-rich and high-antibiotic growing front. Moreover, we show that antibiotic detoxification by resistant cells can protect nearby susceptible cells, but such cross-protection is more effective in mutualistic communities because mutualism drives mixing of resistant and susceptible cells. In contrast, competition leads to segregation, which ultimately prevents susceptible cells to profit from detoxification. Understanding how the interplay between microbial metabolic interactions and community spatial structuring shapes the outcome of antibiotic treatment can be key to effectively leverage the power of antibiotics and promote microbiome health.

Pathogens -microorganisms that make us sick- often live within dynamic and complex multispecies communities, where they may not only compete for limiting resources but also exchange beneficial resources or services with other resident species. While antibiotics are commonly used to get rid of such harmful microbes, the community-wide effects of antibiotic treatment and its consequences for antibiotic resistance are still not well understood. How do competitive or mutually beneficial interactions between antibiotic resistant and susceptible species influence community resistance to antibiotics? Here we investigate this question using a computational model. We find that antibiotic exposure favours the resistant lineage when resistant and susceptible strains are competitors but harms both types when they are mutualists. With antibiotic-detoxifying resistant cells, cross-protection of susceptible cells is more effective in mutualistic communities because mutualism drives mixing of susceptible and resistant cells. In contrast, competition leads to their segregation, precluding susceptible cells to profit from their competitor’s local detoxification. Our findings highlight that knowing not only what species are present but also how they interact with each other and arrange themselves in space is central to understanding antibiotic resistance and to informing the development of strategies that promote microbiome health.

Predictable Phenotypes of Antibiotic Resistance Mutations

Antibiotic-resistant bacteria represent a major threat to our ability to treat bacterial infections. Two factors that determine the evolutionary success of antibiotic resistance mutations are their impact on resistance level and the fitness cost. Recent studies suggest that resistance mutations commonly show epistatic interactions, which would complicate predictions of their stability in bacterial populations. We analyzed 13 different chromosomal resistance mutations and 10 host strains of Salmonella enterica and Escherichia coli to address two main questions. (i) Are there epistatic interactions between different chromosomal resistance mutations? (ii) How does the strain background and genetic distance influence the effect of chromosomal resistance mutations on resistance and fitness? Our results show that the effects of combined resistance mutations on resistance and fitness are largely predictable and that epistasis remains rare even when up to four mutations were combined. Furthermore, a majority of the mutations, especially target alteration mutations, demonstrate strain-independent phenotypes across different species. This study extends our understanding of epistasis among resistance mutations and shows that interactions between different resistance mutations are often predictable from the characteristics of the individual mutations.

The spread of antibiotic-resistant bacteria imposes an urgent threat to public health. The ability to forecast the evolutionary success of resistant mutants would help to combat dissemination of antibiotic resistance. Previous studies have shown that the phenotypic effects (fitness and resistance level) of resistance mutations can vary substantially depending on the genetic context in which they occur. We conducted a broad screen using many different resistance mutations and host strains to identify potential epistatic interactions between various types of resistance mutations and to determine the effect of strain background on resistance phenotypes. Combinations of several different mutations showed a large amount of phenotypic predictability, and the majority of the mutations displayed strain-independent phenotypes. However, we also identified a few outliers from these patterns, illustrating that the choice of host organism can be critically important when studying antibiotic resistance mutations.

Plasmid persistence: costs, benefits, and the plasmid paradox

Plasmids are extrachromosomal DNA elements that can be found throughout bacteria, as well as in other domains of life. Nonetheless, the evolutionary processes underlying the persistence of plasmids are incompletely understood. Bacterial plasmids may encode genes for traits that are sometimes beneficial to their hosts, such as antimicrobial resistance, virulence, heavy metal tolerance, and the catabolism of unique nutrient sources. In the absence of selection for these traits, however, plasmids generally impose a fitness cost on their hosts. As such, plasmid persistence presents a conundrum: models predict that costly plasmids will be lost over time or that beneficial plasmid genes will be integrated into the host genome. However, laboratory and comparative studies have shown that plasmids can persist for long periods, even in the absence of positive selection. Several hypotheses have been proposed to explain plasmid persistence, including host-plasmid co-adaptation, plasmid hitchhiking, cross-ecotype transfer, and high plasmid transfer rates, but there is no clear evidence that any one model adequately resolves the plasmid paradox.

Evolutionary Rescue over a Fitness Landscape

Evolutionary rescue describes a situation where adaptive evolution prevents the extinction of a population facing a stressing environment. Models of evolutionary rescue could in principle be used to predict the level of stress beyond which extinction becomes likely for species of conservation concern, or, conversely, the treatment levels most likely to limit the emergence of resistant pests or pathogens. Stress levels are known to affect both the rate of population decline (demographic effect) and the speed of adaptation (evolutionary effect), but the latter aspect has received less attention. Here, we address this issue using Fisher's geometric model of adaptation. In this model, the fitness effects of mutations depend both on the genotype and the environment in which they arise. In particular, the model introduces a dependence between the level of stress, the proportion of rescue mutants, and their costs before the onset of stress. We obtain analytic results under a strong-selection-weak-mutation regime, which we compare to simulations. We show that the effect of the environment on evolutionary rescue can be summarized into a single composite parameter quantifying the effective stress level, which is amenable to empirical measurement. We describe a narrow characteristic stress window over which the rescue probability drops from very likely to very unlikely as the level of stress increases. This drop is sharper than in previous models, as a result of the decreasing proportion of stress-resistant mutations as stress increases. We discuss how to test these predictions with rescue experiments across gradients of stress.

Fitness Tradeoffs of Antibiotic Resistance in Extraintestinal Pathogenic Escherichia coli

Evolutionary trade-offs occur when selection on one trait has detrimental effects on other traits. In pathogenic microbes, it has been hypothesized that antibiotic resistance trades off with fitness in the absence of antibiotic. Although studies of single resistance mutations support this hypothesis, it is unclear whether trade-offs are maintained over time, due to compensatory evolution and broader effects of genetic background. Here, we leverage natural variation in 39 extraintestinal clinical isolates of Escherichia coli to assess trade-offs between growth rates and resistance to fluoroquinolone and cephalosporin antibiotics. Whole-genome sequencing identifies a broad range of clinically relevant resistance determinants in these strains. We find evidence for a negative correlation between growth rate and antibiotic resistance, consistent with a persistent trade-off between resistance and growth. However, this relationship is sometimes weak and depends on the environment in which growth rates are measured. Using in vitro selection experiments, we find that compensatory evolution in one environment does not guarantee compensation in other environments. Thus, even in the face of compensatory evolution and other genetic background effects, resistance may be broadly costly, supporting the use of drug restriction protocols to limit the spread of resistance. Furthermore, our study demonstrates the power of using natural variation to study evolutionary trade-offs in microbes.

Clinical Implications of Digenic Inheritance and Epistasis in Primary Immunodeficiency Disorders

The existence of epistasis in humans was first predicted by Bateson in 1909. Epistasis describes the non-linear, synergistic interaction of two or more genetic loci, which can substantially modify disease severity or result in entirely new phenotypes. The concept has remained controversial in human genetics because of the lack of well-characterized examples. In humans, it is only possible to demonstrate epistasis if two or more genes are mutated. In most cases of epistasis, the mutated gene products are likely to be constituents of the same physiological pathway leading to severe disruption of a cellular function such as antibody production. We have recently described a digenic family, who carry mutations of TNFRSF13B/TACI as well as TCF3 genes. Both genes lie in tandem along the immunoglobulin isotype switching and secretion pathway. We have shown they interact in an epistatic way causing severe immunodeficiency and autoimmunity in the digenic proband. With the advent of next generation sequencing, it is likely other families with digenic inheritance will be identified. Since digenic inheritance does not always cause epistasis, we propose an epistasis index which may help quantify the effects of the two mutations. We also discuss the clinical implications of digenic inheritance and epistasis in humans with primary immunodeficiency disorders.

Plasticity of the MFS1 Promoter Leads to Multidrug Resistance in the Wheat Pathogen Zymoseptoria tritici

The ascomycete Zymoseptoria tritici is the causal agent of Septoria leaf blotch on wheat. Disease control relies mainly on resistant wheat cultivars and on fungicide applications. The fungus displays a high potential to circumvent both methods. Resistance against all unisite fungicides has been observed over decades. A different type of resistance has emerged among wild populations with multidrug-resistant (MDR) strains. Active fungicide efflux through overexpression of the major facilitator gene MFS1 explains this emerging resistance mechanism. Applying a bulk-progeny sequencing approach, we identified in this study a 519-bp long terminal repeat (LTR) insert in the MFS1 promoter, a relic of a retrotransposon cosegregating with the MDR phenotype. Through gene replacement, we show the insert as a mutation responsible for MFS1 overexpression and the MDR phenotype. Besides this type I insert, we found two different types of promoter inserts in more recent MDR strains. Type I and type II inserts harbor potential transcription factor binding sites, but not the type III insert. Interestingly, all three inserts correspond to repeated elements present at different genomic locations in either IPO323 or other Z. tritici strains. These results underline the plasticity of repeated elements leading to fungicide resistance in Z. tritici and which contribute to its adaptive potential. IMPORTANCE Disease control through fungicides remains an important means to protect crops from fungal diseases and to secure the harvest. Plant-pathogenic fungi, especially Zymoseptoria tritici, have developed resistance against most currently used active ingredients, reducing or abolishing their efficacy. While target site modification is the most common resistance mechanism against single modes of action, active efflux of multiple drugs is an emerging phenomenon in fungal populations reducing additionally fungicides' efficacy in multidrug-resistant strains. We have investigated the mutations responsible for increased drug efflux in Z. tritici field strains. Our study reveals that three different insertions of repeated elements in the same promoter lead to multidrug resistance in Z. tritici. The target gene encodes the membrane transporter MFS1 responsible for drug efflux, with the promoter inserts inducing its overexpression. These results underline the plasticity of repeated elements leading to fungicide resistance in Z. tritici.

Diversity and evolution of drug resistance mechanisms in Mycobacterium tuberculosis

Despite the efficacy of antibiotics to protect humankind against many deadly pathogens, such as Mycobacterium tuberculosis, nothing can prevent the emergence of drug-resistant strains. Several mechanisms facilitate drug resistance in M. tuberculosis including compensatory evolution, epistasis, clonal interference, cell wall integrity, efflux pumps, and target mimicry. In this study, we present recent findings relevant to these mechanisms, which can enable the discovery of new drug targets and subsequent development of novel drugs for treatment of drug-resistant M. tuberculosis.

Widespread epistasis regulates glucose homeostasis and gene expression

The relative contributions of additive versus non-additive interactions in the regulation of complex traits remains controversial. This may be in part because large-scale epistasis has traditionally been difficult to detect in complex, multi-cellular organisms. We hypothesized that it would be easier to detect interactions using mouse chromosome substitution strains that simultaneously incorporate allelic variation in many genes on a controlled genetic background. Analyzing metabolic traits and gene expression levels in the offspring of a series of crosses between mouse chromosome substitution strains demonstrated that inter-chromosomal epistasis was a dominant feature of these complex traits. Epistasis typically accounted for a larger proportion of the heritable effects than those due solely to additive effects. These epistatic interactions typically resulted in trait values returning to the levels of the parental CSS host strain. Due to the large epistatic effects, analyses that did not account for interactions consistently underestimated the true effect sizes due to allelic variation or failed to detect the loci controlling trait variation. These studies demonstrate that epistatic interactions are a common feature of complex traits and thus identifying these interactions is key to understanding their genetic regulation.

Most complex traits and diseases are regulated by the combined influence of multiple genetic variants. However, it remains controversial whether these genetic variants independently influence complex traits, and therefore the impact of each variant could be simply added together (additivity), or whether the variants work together to influence trait variation, in which case the combined impact of multiple variants would differ from the summed impact of each individual variant (epistasis). In this study in mice, we discovered that the genetic regulation of blood sugar levels and gene expression in the liver were predominantly controlled by non-additive interactions, whereas body weight was predominantly controlled by additive interactions. Remarkably, the expression level of nearly 25% of all genes in the liver was controlled by non-additive interactions. The non-additive interactions typically acted to return trait values to the levels detected in control mice, thus contributing to a reduction in trait variation. We also demonstrated that not accounting for non-additive interactions significantly underestimated the phenotypic effect of a genetic variant on a particular genetic background, suggesting that many previously identified risk loci may have significantly larger effects on disease susceptibility in a subset of individuals. These studies highlight the importance of understanding interactions between genetic variants to better understand disease risk and personalize clinical care.

Multidrug-resistant bacteria compensate for the epistasis between resistances

Mutations conferring resistance to antibiotics are typically costly in the absence of the drug, but bacteria can reduce this cost by acquiring compensatory mutations. Thus, the rate of acquisition of compensatory mutations and their effects are key for the maintenance and dissemination of antibiotic resistances. While compensation for single resistances has been extensively studied, compensatory evolution of multiresistant bacteria remains unexplored. Importantly, since resistance mutations often interact epistatically, compensation of multiresistant bacteria may significantly differ from that of single-resistant strains. We used experimental evolution, next-generation sequencing, in silico simulations, and genome editing to compare the compensatory process of a streptomycin and rifampicin double-resistant Escherichia coli with those of single-resistant clones. We demonstrate that low-fitness double-resistant bacteria compensate faster than single-resistant strains due to the acquisition of compensatory mutations with larger effects. Strikingly, we identified mutations that only compensate for double resistance, being neutral or deleterious in sensitive or single-resistant backgrounds. Moreover, we show that their beneficial effects strongly decrease or disappear in conditions where the epistatic interaction between resistance alleles is absent, demonstrating that these mutations compensate for the epistasis. In summary, our data indicate that epistatic interactions between antibiotic resistances, leading to large fitness costs, possibly open alternative paths for rapid compensatory evolution, thereby potentially stabilizing costly multiple resistances in bacterial populations.

Antibiotics target essential cellular functions, such as translation or cell wall biogenesis, and bacteria can become resistant to antibiotics by acquiring mutations in genes encoding those functions. This causes most drug-resistance mutations to be detrimental in the absence of the drug. However, bacteria can reduce this handicap by acquiring additional mutations, known as compensatory mutations. Compensatory evolution is crucial for the maintenance and dissemination of antibiotic resistances in bacterial populations. While compensation for single resistances has been extensively studied, compensatory evolution of multidrug-resistant bacteria remains unexplored. Importantly, interactions between resistance mutations are frequent, and this may cause compensation of multidrug-resistant bacteria to differ significantly from that of single-resistant strains. By comparing compensation of single- and double-drug–resistant E. coli, we found that double-drug–resistant bacteria compensate faster than single-drug–resistant strains. This is due to the acquisition of compensatory mutations with larger effects and possibly driven by the large fitness cost of double-drug resistance. Strikingly, we identified mutations that compensate specifically for the interaction between drug resistances, since they are beneficial only for double-drug–resistant bacteria and in conditions in which the interaction between resistances occurs. In summary, our data indicate that certain interactions between antibiotic-resistance mutations can open alternative paths for rapid compensatory evolution, thereby potentially stabilizing multiple drug resistances in bacterial populations.