Regulatory fine-tuning of *mcr-1* increases bacterial fitness and stabilises antibiotic resistance in agricultural settings

Antimicrobial resistance (AMR) poses a significant threat to global health. Acquiring antibiotic resistance, whether through mutation or horizontal gene transfer, typically incurs fitness costs, reducing bacterial competitiveness and virulence. Reducing antibiotic consumption should, in theory, select against resistance, leading to its decline. However, interventions aimed at reducing antibiotic use have often yielded only marginal decreases in resistance prevalence. A key challenge is understanding how resistance persists in the absence of continuous antibiotic exposure. Experimental evolution studies show that resistance often persists due to compensatory mutations that offset the fitness costs. However, evidence of this in clinical pathogen populations is scarce, with notable exceptions in *Mycobacterium tuberculosis*. Alternative mechanisms, such as co-selection (selection for linked genes) and high rates of horizontal gene transfer via mobile genetic elements (e.g., conjugative plasmids), can also maintain resistance. This study focuses on colistin, an agricultural antibiotic increasingly used as a last resort for treating multi-drug resistant infections. The widespread use of colistin as a growth promoter fueled the spread of *mcr-1* (mobile colistin resistance) in *E. coli*. A 2017 ban on colistin use in animal feed in China resulted in a 90% reduction in consumption and a decline in *mcr-1* prevalence. However, the decline rate was slower than expected given reported colistin resistance costs. This research investigates the hypothesis that compensatory adaptation stabilized colistin resistance in *E. coli* by combining functional assays to measure the fitness effects of *mcr-1* polymorphisms with analysis of large-scale genomic and epidemiological data.

The literature extensively documents the fitness costs associated with antibiotic resistance and the various mechanisms by which bacteria compensate for these costs. Studies have shown that compensatory mutations can alleviate fitness costs without compromising resistance, and this idea is central to evolutionary models of AMR. However, empirical evidence demonstrating compensatory adaptation in clinical settings remains limited. The role of co-selection, where resistance genes are linked to genes providing other advantages, has also been explored. Furthermore, the importance of horizontal gene transfer via conjugative plasmids in maintaining resistance, even when plasmids impose fitness costs, is well-established. The literature on *mcr-1*, its spread, and its associated fitness costs is also reviewed, highlighting its prevalence and the impact of colistin use in agriculture on its dissemination. Existing studies on the decline of *mcr-1* following colistin bans provide a backdrop for the investigation of compensatory mechanisms maintaining resistance.

The study employed a multi-faceted approach combining experimental evolution and genomic epidemiology. First, *mcr-1* regulatory variants, identified from a previously published genomic analysis, were cloned into a mini-RK2-derived expression vector (pSEVA121). The fitness effects of these variants were assessed by measuring their growth rates in colistin-free medium. MCR-1 protein activity was measured via a FITC-PLL binding assay, assessing cell surface charge. *mcr-1* transcription levels were determined by RT-qPCR. Colistin resistance was measured using MIC assays and IC50 calculations. Competitive fitness assays were conducted by co-culturing *E. coli* with regulatory variants and wild-type strains across a gradient of colistin concentrations. Linear and multiple regression models were used to analyze the relationships between fitness, resistance, and *mcr-1* expression and activity. To further investigate the role of *mcr-1* expression, the gene was integrated into the *E. coli* chromosome, decreasing its expression from a multi-copy plasmid to a single-copy chromosomal location. For the genomic epidemiology component, a previously published dataset of *mcr-1*-positive *E. coli* isolates from various sources in China was analyzed to identify associations between regulatory variants and plasmid replicons. Phylogenetic analysis was performed on IncX4 plasmids to study the evolutionary origins of regulatory variants. Finally, changes in the prevalence of regulatory polymorphisms before and after the colistin ban were calculated using data from pig farms, examining *mcr-1* stability over time.

The study found that seven of eight regulatory variants increased growth rates compared to wild-type *mcr-1*, demonstrating that these polymorphisms reduce the fitness cost of colistin resistance. In silico analysis indicated that these polymorphisms reduce *mcr-1* transcription and/or translation. Experimental measurements confirmed that these variants reduced MCR-1 protein activity and *mcr-1* expression. Despite reduced expression, colistin resistance was not compromised; in fact, six of the eight regulatory variants showed increased or equal colistin resistance compared to the wild-type. Competitive fitness assays showed that regulatory variants exhibited increased fitness in both the presence and absence of colistin, highlighting a lack of fitness trade-off. A strong correlation was found between fitness and colistin resistance in colistin-free media. Chromosomally integrated *mcr-1* displayed increased growth rate and colistin MIC compared to plasmid-borne *mcr-1*, further supporting the link between reduced expression and increased fitness. Analysis of the genomic dataset revealed strong associations between regulatory variants and plasmid replicons, notably IncX4 and PV3. Phylogenetic analysis of IncX4 plasmids suggested that regulatory fine-tuning occurred independently multiple times. The most prevalent regulatory variant, PV3, showed increased stability in pig farms following the colistin ban. The increase in prevalence of tested regulatory variants compared to the wild type following the ban was statistically significant (p=0.037). This suggests that the regulatory mutations improved fitness and stability of *mcr-1* under decreased colistin selection.

The findings demonstrate the remarkable ability of bacteria to adapt and fine-tune gene expression to minimize the fitness costs of antibiotic resistance while maintaining high resistance levels. The interplay between regulatory evolution and plasmid transfer is key to stabilizing *mcr-1* despite efforts to reduce antibiotic use. The high copy number of plasmids carrying *mcr-1* likely accelerated regulatory evolution, highlighting the role of plasmid copy number in shaping resistance evolution. The observed correlation between fitness and colistin resistance suggests that intermediate *mcr-1* expression levels optimize resistance while minimizing deleterious effects on membrane structure. The study contradicts the classical compensatory adaptation model, where mutations overcome resistance costs without affecting susceptibility. Instead, *mcr-1* regulatory mutations demonstrate a mechanism to fine-tune expression, enhancing both fitness and resistance. The dissemination of low-cost/high-resistance *mcr-1* alleles via conjugative plasmids represents a worst-case scenario for resistance management. However, the overall decline of *mcr-1* after the ban highlights the importance of sustained efforts in reducing antibiotic consumption.

This research provides clear evidence that regulatory evolution and plasmid transfer can effectively stabilize antibiotic resistance in the face of reduced antibiotic use. The fine-tuning of *mcr-1* expression via regulatory mutations allows for both high resistance and increased bacterial fitness. The dissemination of these advantageous alleles via conjugative plasmids poses a challenge to resistance management. Future research should explore the interactions between *mcr-1* and chromosomal genes involved in LPS biosynthesis and the impact of horizontal gene transfer on resistance persistence in different environments.

The study primarily focused on *mcr-1* in *E. coli*. Generalizing these findings to other resistance genes and bacterial species requires further research. The genomic analysis relied on a previously published dataset, and the sample size and sampling strategy could influence results. The study did not completely rule out the contribution of other mutations outside the *mcr-1* regulatory region to the observed fitness increases.