Escherichia coli is a common pathogen responsible for multiple infections, including those in the urinary tract, bloodstream, and wounds, in humans and animals worldwide (Nang et al., 2019). It can easily acquire multi-drug resistance, owing to its ability to accumulate multiple resistance genes, primarily through horizontal gene transfer of mobile elements. Currently, its acquisition of genes encoding extended-spectrum β-lactamases (ESBLs), carbapenem-hydrolyzing ß-lactamase (CHßls), 16S rRNA methylases, plasmid-mediated quinolone resistance (PMQR), and plasmid-mediated glutathione S-transferase (PMGST) is of great clinical concern (Simoni et al., 2021; Yin et al., 2021; Guo et al., 2022), accelerating the evolution of drug resistance, making colistin the last-resort antibiotic for treating infections caused by multi-drug resistant bacteria. The use of colistin in livestock as a growth promoter has been banned in China (Wang et al., 2020) due to increasing polymyxin resistance. However, the use of colistin in clinical settings is increasing because of the emergence of increased multi-drug resistance (MDR), particularly carbapenem-resistant E. coli (CRE) (Liu et al., 2020); therefore, the emergence and spread of colistin-resistant E. coli are inevitable.

Colistins primarily bind to the outer membrane lipopolysaccharide through the interaction of their cationic residues with the phosphate groups of lipid A, changing the permeability of the cell envelope, leading to leakage of cell contents and bacterial death (Rodríguez-Santiago et al., 2021). However, mobile colistin resistance (mcr) genes have been reported to be the primary mechanism conferring resistance to colistin, although two-component systems (TCSs), such as PmrAB and PhoPQ; mutation(s) in the MgrB regulator; LPS modification (Bialvaei and Samadi Kafil, 2015); LPS overall charge changes, and the reduced affinity of colistin for the outer membrane, also contribute to colistin resistance based on chromosomal-encoded mechanisms (Zhang et al., 2021; Phuadraksa et al., 2022). The mcr genes encode transferable phosphoethanolamine transferases that modify the lipopolysaccharide of the bacterial outer membrane to weaken its binding to colistin (Yin et al., 2021). To date, more than 10 mcr variants (mcr-1 to mcr-10) have been identified, carried on different conjugative and non-conjugative plasmid backbones (Mmatli et al., 2022). These variants have been widely identified in many bacteria, including E. coli, Klebsiella, Enterobacter, Citrobacter, Proteus, Providencia, Salmonella, Pseudomonas, Acinetobacter, Aeromonas, Kluyvera, and Raoultella, as well as in diverse ecosystems, including soil, botanicals, wildlife, animal environments, and public places (Nang et al., 2019; Anyanwu et al., 2020). However, despite its status as the most common host for mcr genes, data on mcr distribution in E. coli are quite limited, although global surveillance showed an approximately 1.26% prevalence of mcr-positive E. coli among humans, animals, and environments (Dadashi et al., 2022; Mmatli et al., 2022), with the mcr-1, mcr-3, mcr-5, mcr-9, and mcr-10 variants most prevalent (Dadashi et al., 2022). These investigations all relied on PCR detection methods, with few data on the prevalence of mcr from global genomic databases.

Most importantly, mcr genes are carried by mobile genetic elements, especially conjugative plasmids, leading to their broad dissemination and establishment worldwide under the selective pressure of antimicrobial agents. Although isolates with mcr as well as other antimicrobial resistance genes (ARG), such as bla_CHßLs, bla_ESBLs, PMQR, and fosA, have been frequently reported, posing clinical challenges (Haenni et al., 2016; Zhang W. et al., 2022), the prevalence of clinically endemic ARGs among mcr-positive strains remains unclear. Furthermore, a high diversity of plasmid reservoirs has been shown to be associated with mcr genes (Nang et al., 2019); however, the distribution of plasmid replicons among global mcr-positive E. coli and the associations between mcr and plasmid replicons remain unknown. Furthermore, multiple virulence factors (VFs) involved in bacterial adherence, invasion, immune modulation, effector delivery systems, and nutritional/metabolic factors have been identified in E. coli (Zhao et al., 2020; Kubelová et al., 2021). However, their distribution among mcr-positive E. coli and the association between VFs and ARGs require further investigation. Finally, the prevalence of VFs among internationally popular sequence types (STs) of mcr-positive E. coli is of vital importance. All of these will contribute to optimizing prevention and control measures to prevent further spread and outbreaks of mcr-positive infections.

In this study, we first investigated the distribution of mcr among global E. coli based on the whole genome sequencing (WGS) data from GenBank. We analyzed STs and the prevalence of ARGs, VFs, and plasmid replicons. Additionally, the distribution consistency of VFs and ARGs was tested, providing epidemiological data for the implementation of infection and prevention.

As the last-line defense against MDR E. coli infections, colistin has been widely used to treat infections caused by extensive-drug resistant (XDR) bacteria, such as CRE, albeit it has been banned for use in livestock as a growth promoter. The clinical utility of colistin is currently threatened by the global dissemination of mcr and its co-occurrence with other endemic ARGs, such as ESBL and CHßl, through different plasmids. Thus, understanding the prevalent characteristics of global mcr-positive E. coli isolates, including the prevalence of STs, endemic ARGs, VFs, and plasmid replicons, will be useful to guide clinical treatment for serious infections, as well as the implementation of preventative measures.

In this study, we observed a rapid increase in the genomes of global mcr-positive E. coli from 2014 to 2016. Since 2016, many countries have prohibited colistin as a growth promoter for livestock, including Brazil (November 2016), Thailand (February 2017), China (April 2017), Japan (July 2018), Malaysia (January 2019), Argentina (February 2019), and India (July 2019) (Center for Veterinary Drug Development, 2018; Shen et al., 2018; Wang et al., 2020). Although the number of mcr-positive E. coli genomes declined after 2016, given that some of the sequence data from 2019 to 2021 had not yet been released when we downloaded the data and conducted this study in 2021, the actual prevalence of mcr-positive strains worldwide should be more than the number we analyzed, so current global prevalence would likely be significantly higher. Additionally, as the poultry and swine industries accounted for 96% of total colistin sulfate livestock use (Wang et al., 2020), the presence of mcr in them provides evidence that colistin treatment has promoted the transmission of mcr, with livestock as the primary reservoir. Colistin-resistant E. coli can then spread by contaminating animal-derived food or contaminating crops by excrement to threaten public safety (Elbediwi et al., 2019).

The high incidence of mcr in our study is in accordance with a previous report (Dadashi et al., 2022), consistent with its identification as the globally predominant colistin-resistance gene. At present, mcr-1 could confer colistin resistance with broad minimal inhibitory concentrations (MICs) of 2–16 mg/L (Nang et al., 2019), for instance, 74% of mcr-1 positive E. coli isolates had MICs at 2 mg/L (Nguyen et al., 2022), while 52% of mcr-1 positive E. coli isolates in another study exhibited MICs of 4∼8 mg/L (Ćwiek et al., 2021). For other mcr genes, mcr-3 has been globally disseminated since it was first reported in Yin et al. (2017), and this gene inside the linker has been found to appear as a facilitator of colistin resistance (Xu et al., 2021). It also has been reported to facilitate evasion of host phagocytosis, but generally mediates low-level (≤8 mg/L) colistin resistance among Enterobacteriaceae (Yin et al., 2021). The mcr-9 variant was generally susceptible to colistin with MICs of ≤1 mg/L, because it did not confer colistin resistance itself, its colistin resistance replied to gene regulation (Nakamura et al., 2021), and mcr-5 was demonstrated to generate colistin resistance with MICs of 4–8 mg/L in S. enterica (Nang et al., 2019). Notably, more than 20% mcr-positive strains co-carried CHßLs, mainly bla_NDM followed by bla_KPC, indicating that almost one-fifth of the strains in our study may belong to CRE. In addition, more than 50% of the strains co-carried bla_CTX–M, such a high prevalence of bla_CTX–M showed that most of these strains belonged to MDR strains. Meanwhile, the co-carriage of 16s RNA methylase, PMQRs, and fosfomycin genes suggested that the mcr-positive strains were also reservoirs of numerous resistance genes (Guo et al., 2022).

The VF analysis showed that almost all mcr-positive strains also carried terC and gad genes. To the best of our knowledge, terC represents one of the key proteins in tellurite and colicine resistance, implicated in phage inhibition, resistance, and pathogenicity, and is widespread among bacterial species, particularly in pathogens (Peng et al., 2021; Rodríguez-Santiago et al., 2021). In contrast, gad has been reported to enzymatically decrease intracellular protons linked to E. coli flagellar motility (Yamanaka et al., 2022) and is involved in the colonization of the gastrointestinal tract. In addition, more than half of the strains contained iss, traT, and ompT, which participate in resistance to serum complement, outer membrane protease, and long polar fimbriae, respectively; these affect resistance to host innate immunity and facilitate colonization and adhesion. Moreover, iroN was detected in approximately 25% of all isolates and has been associated with increased 30-day mortality in patients with bacteremia (Hung et al., 2019).

We identified many STs, demonstrating the diversity of mcr-positive strains. The high prevalence of ST10 in our study was similar to that in a previous report (Dadashi et al., 2022), however, it was different from a report that ST131 was the most common ST for extraintestinal pathogenic E. coli (ExPEC) strains (Manges et al., 2019). To date, E. coli ST10 has been shown to be a high-risk mcr-1-positive isolate from cattle farm environments (Ali et al., 2021); its clonal spread with ∼50% fosfomycin resistance has been found among diarrheal calves in Xinjiang province, China (He et al., 2021). It also has been isolated from poultry in Poland (Ćwiek et al., 2021). The biological costs imposed by plasmid-mediated resistance and virulence affect its survival and spread (Yang et al., 2020), thus, the relatively rare ARGs and VFs within ST10 clones may explain its rapid spread.

Curiously, STs differed from continent to continent, and as frequent carriers for mcr-1 and bla_NDM (Lin et al., 2020; Zhang X. et al., 2022), more ARGs and VFs were observed within the ST156 strain prevailing in Asia and Europe, indicating that it represents a major risk to public health, highlighting the need for increased surveillance. For example, E. coli ST156, carrying bla_NDM–5, bla_CTX–M–65, bla_OXA–10, bla_TEM–1, and mcr-1, has been found in clinical isolates in China (Lin et al., 2020; Zhang X. et al., 2022). Notably, ST648 has been reported as a high-risk, MDR, ESBL-producing strain in public aquatic environments (Furlan et al., 2020), but ST648 was not frequently detected in our study, with almost all of them being in North America. The differences in the distributions of ST clones may be related to antibiotic use, diet, and environmental factors. As the predominant ST among ExPEC isolates worldwide (Alangari et al., 2022), ST131 was shown to have higher virulence than other important ExPEC clones (Alqasim et al., 2020), whereas only 11 mcr-positive E. coli ST131 isolates were identified in this study, mainly from China, suggesting that ST10, rather than ST131 and ST648, is the main host for mcr, which may result from the fitness between the clones and mcr genes. Moreover, although ST216, with fewer VFs and ARGs, was dominant in Oceania, their association with IncHI2 plasmids may be of great concern (Tarabai et al., 2021).

The great diversity of plasmid replicons within mcr-positive strains indicates the great dissemination potential of ARGs and VFs. The high incidence of IncFIB is consistent with previous reports showing that IncFIB is common in Enterobacteriaceae (Khine et al., 2020). A previous study found that IncI2 and IncX conferred fitness advantages to host bacteria and outcompeted other plasmids (Bahador et al., 2018); the approximately 40% incidence of IncX and IncFII among mcr-positive strains suggests that most of the strains in this study have relatively high adaptability. Notably, IncI2 and IncX were common plasmids in mcr-1-positive strains, indicating that they may be dominant vectors driving mcr-1 transmission, but this finding of correlation analysis should be further confirmed. Considering the distribution deviation, we hypothesized that mcr-3 and mcr-9 were harbored by IncFII and IncHI2 plasmids, respectively, and mcr-5 may be borne on IncFII and IncI1-I plasmids. These findings were in accordance with previous reports that mcr-1 was carried by a range of plasmids such as IncHI2, IncI2, IncX4, IncFIA, and IncP (Haenni et al., 2016; Elbediwi et al., 2019; Liu et al., 2020; Sonnevend et al., 2022), and mcr-9 was always carried by IncHI2 (Simoni et al., 2021). Recently, bla_NDM was reported to be disseminated via the IncP plasmid with a broad host range (Choudhury et al., 2019); moreover, most VFs that were associated with enteroinvasive E. coli, such as sit, iron, iss, iutA, ompT, and iroN, were reported to be located mainly on IncFIA/FIB and Col plasmids (Touzain et al., 2018; Hayashi et al., 2019). Both mcr-1 and bla_CTX–M–1 were colocalized on IncHI2 and IncI2 plasmids in two studies (Haenni et al., 2016; Yin et al., 2021), which support the consistent distribution of plasmids and ARGs/VFs analyzed in our study, suggesting the dissemination potential of the related ARGs and VFs.

To the best of our knowledge, ours is the first study to evaluate the prevalence of global mcr-harboring E. coli isolates using WGS data. The 778 mcr-positive isolates, widely distributed among the six continents during 1905–2019, could be representative of the current global mcr-carriers. However, this study had some limitations. First, the actual resistance phenotypes of these strains were not available for assessment of differences between genotypes and phenotypes. Second, whether the resistance and mcr genes were carried on a single plasmid was unclear, so we cannot clearly illustrate the specific relationships between VFs and ARGs among plasmids. Third, our study depended on the submission of genomes to the database, so it could not represent areas where no data had been uploaded.

Collectively, our data found that mcr-1 is the most widespread gene that confers resistance to colistin. More than 50% of the global mcr-positive E. coli isolates had numerous ARGs, and the fifth strain belonged to CRE. Diverse STs have been identified, further attention should be focused on the ST10 and ST156 clones, considering the rapid transmission of ST10 and the convergence of ARGs and VFs within ST156.

The original contributions presented in this study are included in this article/Supplementary material, further inquiries can be directed to the corresponding authors.

XC and HS contributed to the experimental design of the study. JS, HZ, CL, HX, and CCL performed data acquisition and statistical analysis. JS, HZ, XC, and HS performed bioinformatics analysis and writing. All authors contributed to the article and approved the submitted version.