Silent Threat Evolution: Critically Important Carbapenem and Colistin Resistance Genes in the Natural Aquatic Environment

Abstract

The rise in antimicrobial resistance (AMR) among the most clinically significant bacteria presents a global threat. The coexistence of resistance mechanisms to both carbapenems and colistin is particularly concerning, as these are last-line treatments, specifically reserved for the most challenging infections caused by clinically multidrug-resistant Enterobacterales. Natural aquatic environments have become environmental reservoirs for the transmission of AMR, particularly concerning mechanisms against these two types of critically important drugs. The crucial role of environmental settings as a driving force for the spread and evolution of AMR associated with these drugs is underestimated, and scientific knowledge on this topic is limited. This review aims to fill an important gap in the scientific literature and comprehensively consolidate the available data on carbapenem- and colistin-associated AMR in the aquatic environment. This study provides a comprehensive synthesis of the current knowledge by integrating bibliographic data with a detailed genomic analysis of 278 bacterial genomes sourced from natural waters. It explores the distribution of carbapenemase and mobile colistin resistance (mcr) genes, identifying their hosts, geographical spread, and complex gene–plasmid–host associations. This review distinguishes two critical host groups for genes that provide resistance to last-resort drugs, Enterobacterales and autochthonous aquatic microbiota, highlighting both confirmed and potential interactions between them. Crucially, genomic analysis highlights the alarming co-occurrence of carbapenem and colistin resistance in single cells and on single plasmids, contributing to the spread of multidrug resistance phenotypes. These findings clearly indicate that aquatic environments are not merely passive recipients but active, evolving hubs for high-risk AMR determinants. Future research should focus on the interplay between allochthonous vectors and autochthonous microbiota to better understand the long-term stabilization of carbapenemase and mcr genes. Such efforts, combined with advanced sequencing technologies, are essential to ensure that carbapenems and colistin remain viable treatment options in clinical settings.

1. Introduction

Antimicrobial resistance (AMR) is a significant global health issue resulting from the unavoidable consequence of the misuse of antibiotics in animal and human health, as well as in agriculture [1]. The increasing use of antimicrobial drugs contributes to the spread of antibiotic-resistant bacteria (ARB) and antibiotic resistance genes (ARGs), creating a “silent pandemic” [2]. The widespread prevalence of ARB and ARGs has severely undermined the significant advancements made in antibacterial therapy, causing common infections to be untreatable [3]. At present, estimates suggest that deaths from infections caused by resistant organisms could reach 10 million by 2050 [4]. Currently, AMR is recognized as one of the most pressing health challenges worldwide, driving us toward a post-antibiotic era [1,5,6]. The most significant consequence of AMR is the reduced effectiveness of many antimicrobials, leaving only a limited number of last-resort options for treating infections. The high rates of AMR observed among the most clinically important bacteria that cause common infections pose a major problem [4,7].

On 17 May 2024, the World Health Organization (WHO) released an updated list of ARB that pose the most significant threat to human health [8]. This update aims to guide the development of new treatments and strategies to prevent the spread of AMR [9]. The list of priority pathogens includes members of the Enterobacterales order that are resistant to carbapenems and produce extended-spectrum β-lactamases (ESBLs), which are of particular concern in hospital epidemiology [10]. Carbapenems, belonging to β-lactam antibiotics, present a broad spectrum of antibacterial activity. These antimicrobial substances are highly effective against both Gram-positive and Gram-negative bacteria [11,12]. Carbapenems work primarily by inhibiting the synthesis of peptidoglycan, a crucial component of bacterial cell walls. The unique molecular structure of antimicrobial substances is due to the presence of a carbapenem together with the β-lactam ring, which confers exceptional stability against most β-lactamases, including ESBLs [12]. These antibiotics are frequently used in human medicine to treat infections caused by various bacteria, including Escherichia coli, Salmonella sp., Klebsiella pneumoniae, Pseudomonas aeruginosa, Mycobacterium tuberculosis, and methicillin-resistant Staphylococcus aureus (MRSA). Their safety profile has contributed to their widespread use across all continents, which has unfortunately led to an increase in carbapenem resistance among microorganisms [11].

Bacteria develop carbapenem resistance by either producing carbapenemase, an enzyme that degrades the carbapenem, or acquiring structural mutations that induce the production of other β-lactamases [13,14]. The existence of carbapenemases has been recognized for over 30 years. The IMP-1 was the first carbapenemase detected in Japan in 1991 [15]. A few years later, in Italy, VIM-1 was reported [16]. The genes responsible for producing these enzymes, namely bla_IMP and bla_VIM, have successfully spread worldwide among microorganisms. Currently, the most common carbapenemases recognized among bacteria include Klebsiella pneumoniae carbapenemase (KPC), metallo-β-lactamases (MBLs), New Delhi metallo-β-lactamase (NDM), and oxacillinases (OXA) [14]. These enzymes allow bacteria to become resistant not only to carbapenems but also to many other β-lactam antibiotics, such as cephalosporins, monobactams, and penicillins [17]. In contrast to the sporadic reports of GIM, SIM, AIM, DIM, FIM, and POM carbapenemases, NDM has shown a dramatic global spread since its detection in India in 2008 [18]. In turn, among the class A carbapenem-hydrolyzing enzymes, KPCs have exhibited the most rapid and widespread dissemination [12]. The genes encoding these enzymes, bla_NDM and bla_KPC, respectively, are widely distributed both in clinical and non-clinical settings [19].

The broad use of carbapenems in clinical practice and the potential to transfer their ARGs through mobile genetic elements (MGEs) have promoted the emergence of carbapenem-resistant bacteria, which include species of particular clinical importance. Especially the pathogenic species belonging to Enterobacteriaceae are commonly associated with hospital infections such as pneumonia, bacteriemia, or urinary tract infections [20,21]. Carbapenems, which display a broad spectrum of activity, are considered a last-line drug against infections caused by multidrug-resistant (MDR) bacteria belonging to this family [22]. However, due to the spread of carbapenem resistance among Enterobacteriaceae, there is a pressing need to find alternative treatment methods. Polymyxin E, commonly known as colistin, has emerged as a critical last-resort agent for treating infections caused by carbapenem-resistant, MDR strains, untreated by other antimicrobials, especially carbapenems [23,24,25,26,27,28]. Colistin, discovered over fifty years ago and commonly known as colistin, is a cationic lipopeptide antibiotic. Unlike carbapenems used in human medicine, colistin is primarily used in animal treatment. In the mid-20th century, colistin gained global, across-the-board use in veterinary medicine. Two commercially available forms of polymyxin E are commonly used: colistin sulphate for oral and topical use (the only approved product in some countries used in pig production to control infections caused by Enterobacteriaceae) and colistin methanesulphonate sodium for parenteral use. In addition to its therapeutic purposes, colistin is also commonly used to promote animal growth [29,30]. Currently, the greatest consumption of colistin is associated with poultry treatment practices. The long-lasting misuse and overuse of colistin have been reflected in the dissemination of colistin resistance among microorganisms [24].

Until 2015, all known polymyxin resistance mechanisms were chromosomally mediated [31,32]. The first case of polymyxin resistance, distributed via horizontal gene transfer (HGT), was identified 10 years ago during studies on antimicrobial resistance among bacteria belonging to Enterobacteriaceae. The first mobile determinant of polymyxin resistance, the mcr-1 gene, was reported in China [31]. The emergence of the mobile colistin resistance (mcr) gene was a breakthrough in polymyxin resistance mediated by plasmids. Over the last 10 years, various studies have documented new unknown variants of mcr genes ranging from mcr-1 to mcr-10 [30]. The rapid increase in plasmid-mediated mcr gene variants has raised major public health concerns. Occasionally, some mcr are noticed on bacterial genomes, indicating a relic of recombination/integration between chromosomes and plasmids [33]. The presence of mcr genes can transfer horizontally among microorganisms and further distribute among humans, animals, and the environment [24].

It is particularly dangerous that carbapenem resistance is frequently connected with accompanying resistance to other antimicrobials, including mobile resistance to colistin, representing a pandrug-resistant phenotype [20,34,35]. Coexistence of carbapenem- and colistin resistance determinants is especially worrisome and results in dire public health consequences. Several studies have already reported overlapping carbapenem and colistin resistance in pathogenic strains within the Enterobacteriaceae family, such as Escherichia coli and Klebsiella pneumoniae, from various matrices, especially of clinical origin [36,37,38,39,40]. However, data on the emergence and dissemination of pandrug-resistant bacteria of high clinical significance in the natural environment are severely limited [22].

Anthropogenic activities and the associated release of antimicrobials, ARB, and ARGs into the environment have significantly impacted AMR evolution. Aquatic environments, including groundwater and surface water (mainly rivers and lakes), have become a breeding ground for AMR transmission [20,41,42,43,44]. It should be especially emphasized that AMR in the natural environment can provide a reflection of the regional and global clinical resistance situation. Moreover, the differences and similarities in AMR profiles among microorganisms isolated from clinical and natural environments reveal a complex interplay that requires a complex study to fully understand the risk to public health [41,43]. Given the interconnectedness of the natural environment, animals, and humans, the AMR issue requires a One Health approach to evaluate its dissemination and evolution [20] (Figure 1).

Taking the above into account, we focused on the aquatic environment as a particularly important, often overlooked reservoir of resistance to carbapenems and colistin, a major cause of critical infections. The role of the natural water sources as a medium for the spread and evolution of resistance to these critical last-line drugs is crucial for public health protection and has not yet been summarized in the scientific literature. To grasp the core of the problem and facilitate accurate tracking of the spread of AMR to these last-line drugs, our study integrates data from both the scientific literature and whole-genome sequencing. This approach allows for precise monitoring of AMR’s dissemination to carbapenems and colistin, which are reserved for only the most difficult-to-treat diseases. Our review aims to thoroughly examine the role of aquatic environments as reservoirs and hotspots for the spread of resistance to critically important pharmaceuticals, whose retransmission poses a direct threat to the effectiveness of medical treatment. By simultaneously combining bibliographic and genomic data analysis, our study provides a foundation for studying the evolution of AMR in the context of last-line drugs in the natural water environment and tracking its transmission pathways.

2. Materials and Methods

2.1. Data Sources

In compliance with the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guidelines [45], the articles were selected according to the four criteria: (i) identification, (ii) screening studies, (iii) eligibility, and (iv) inclusion. In this work, two data sources were used: bibliographic and genomic. In the first case, the PubMed and Google Scholar scientific literature databases were surveyed to find reviewed papers published from 1 January to 30 November 2025. In the case of a genomic data source, specific sequences were explored in the Pathogen Detection tool provided by the National Center for Biotechnology Information (NCBI).

2.2. Bibliographic Data Search Strategy

The bibliographic data search strategy employed is illustrated in Figure S1 in the Supplementary Materials. A preliminary search of the available scientific literature related to the subject of this review was conducted to identify the keywords to be employed in the advanced search. The keywords used in the search strategy were the following: “((mcr[Title]) OR (colistin[Title]) OR (bla[Title]) OR (carbapenem[Title])) AND ((water[Title]) OR (river[Title]) OR (aquatic[Title]))” in the PubMed database and “TITLE (mcr OR colistin OR polymyxin) AND (bla OR carbapenem) AND (water OR river OR natural OR aquatic)” in the Scopus database. These were tailored to each database. The keywords of all resulting articles are presented in Figure S2. In addition, a reference list of articles was checked manually so as to find adequate scientific publications for this review of data from the literature. After filtering the resulting bibliographic data, 150 scientific publications were selected for this review article and their findings were analyzed.

2.3. Genomic Data Search Strategy

Genomes of bacteria isolated from surface waters that possess last-resort antibiotic resistance genes were searched for using the Pathogen Detection tool provided by the National Center for Biotechnology Information (NCBI). To search for genomes with mobile colistin resistance mechanisms, the “AMR genotypes” filter was applied, selecting all variants of the mcr genes. This was further refined by overlaying the “Isolation source” filter using the following terms: “river,” “lake,” “ocean,” “sea,” “marine,” “aquatic,” and “water.” Records that did not fit the paper scope, such as those indicating “wastewater,” were manually excluded. To search for genomes carrying carbapenemase genes, we selected carbapenemase genes from “AMR genotypes” filter. Only genes of mechanisms classified as carbapenemases according to the BLDB database were selected [46]. The same “Isolation source” search parameters as above were maintained. For both carbapenemase and colistin resistance genes, only genomes with a complete sequence length were included in the analysis (as indicated by the ‘COMPLETE’ status within the AMR genotype filter). To ensure high data quality, genomes containing only partial ARG sequences or those truncated at the end of a contig were excluded from further investigation. Detailed information regarding the search queries performed in the Pathogen Detection database is provided in Table S1. The complete list of 278 genomes ultimately included in the analysis is available in File S1.

2.4. Genome-Based Analyses

The genomic sequences were retrieved from the NCBI database and visualized/analyzed using the Proksee (5-12-2025 version) [47] and CARD Resistance Gene Identifier (RGI 6.0.3, 1.3.1 Tool Version) [48]. Within Proksee, contigs containing the mcr and carbapenemase genes were identified. Sequences of the selected contigs were exported from the bacterial genomes using the Geneious Prime software (2025.2.1) and were subsequently analyzed with PlasmidFinder 2.1 [49] to detect plasmid replicon sequences (thresholds of 95% for minimum identity and 60% for minimum % coverage). Contigs harboring successfully identified ARGs and plasmid replicons were further investigated using the COPLA tool [50] to predict their potential host range. Visual comparison of plasmid similarity was performed using Proksee’s built-in BLAST tool (BLAST+ 2.16.0, 1.5.1 Tool Version). Bacterial assignment to specific sequence types (STs) was determined using the MLST 2.0 tool [51]. The similarity of plasmid and whole-genome sequences (Average Nucleotide Identity (ANI) values) was calculated using the FastANI tool (v1.34) [52].

3. Results

3.1. Carbapenem Resistance in the Natural Aquatic Environment

Antimicrobial resistance to carbapenems is a significant global health issue. More than 80% of ARGs against carbapenems are located on plasmids and can be easily transferred among bacteria via HGT [53]. Carbapenemase-producing microorganisms are frequently isolated not only in hospitals and healthcare facilities but also in the natural environment, such as natural water sources. This raises concerns, as it appears that after the clinical environment, a significant reservoir for carbapenem resistance may exist in the natural aquatic environment. Microorganisms are crucial components of natural aquatic environments, which include groundwater, inland surface waters, seas, and oceans. Freshwater ecosystems include rivers, lakes, streams, ponds, and wetlands, as well as drinking water sources like wells. Marine ecosystems encompass vast oceans, coral reefs, and mangroves [54,55]. All these types of water sources were considered during the data collection process for this review.

A review of bibliographic data on the presence of carbapenem resistance genes in natural aquatic environments is summarized in

Table 1. Summary of a review of bibliographic data on the occurrence of carbapenem resistance genes and their hosts in samples from natural aquatic environments.

Carbapenem Resistance Gene	Bacterial Host	Environmental Water Type	Country	Reference
-	Aeromonas veronii, Pseudomonas aeruginosa, Stenotrophomonas spp.	natural (forest) and bore hole water sources	Algeria	[59]
blaOXA-48-like (blaOXA-538)	Shewanella xiamenensis	river water		[60]
blaGES-5	Comamonas spp.	lake water, wetland	Australia	[61]
-	Klebsiella pneumoniae	river water	Bangladesh	[62]
blaGES, blaKPC, blaNDM, blaOXA-51	Enterobacter cloacae, Aeromonas caviae, Acinetobacter venetianus	river water	Brazil	[63]
blaKPC-2	Enterobacter kobei	coastal water		[64]
blaKPC-2	Enterobacter cloacae	coastal water		[65]
blaNDM-1	Klebsiella pneumoniae	coastal water		[66]
blaNDM-1	Raoultella ornithinolytica	well water	China	[67]
blaNDM-1, blaKPC-2	Citrobacter spp.	river water, pond water		[68]
blaKPC-2	Raoultella ornithinolytica	well water		[69]
blaNDM-1	-	drinking water supply system		[70]
blaOXA-58	Pseudomonas sp., Rheinheimera sp., Stenotrophomonas sp., Shewanella sp., Raoultella sp., Vibrio sp., Pseudoalteromonas sp., Algoriphagus sp., Bowmanella sp., Thalassospira sp.	seawater		[19]
blaOXA-69	Acinetobacter baumannii	river water		[71]
blaNDM-5	Escherichia coli	river water	France	[72]
-	Escherichia coli	river water		[73]
blaVCC-1	Vibrio cholerae	coastal waters	Germany	[74]
blaPOM-1	Pseudomonas otitidis	river water	Ghana	[75]
blaOXA-181, blaNDM-5, blaOXA-48	Escherichia coli, Klebsiella pneumoniae, Enterobacter cloacae	river water, pond water		[76]
blaNDM-1, blaNDM-2, blaNDM-5	Serratia marcescens, Escherichia coli, Klebsiella pneumoniae	lake water, pond water	India	[77]
blaNDM, blaOXA-48	Klebsiella pneumoniae	river water		[78]
blaNDM, blaOXA-48-like, blaKPC	Klebsiella pneumoniae	river water		[79]
blaOXA-48, blaKPC, blaNDM-5	Klebsiella pneumoniae, Escherichia coli	seawater	Ireland	[80]
blaIMP	Escherichia coli	river water	Japan	[81]
blaOXA-731	Shewanella sp.	drinking water storage system	Myanmar	[82]
blaIMP, blaVIM, blaOXA-51, blaOXA-58, blaIMP-1, blaVIM-2	Acinetobacter spp.	river water	Poland	[83]
-	Acinetobacter spp.	river water		[84]
blaKPC-3	Klebsiella pneumoniae	river sediment	Portugal	[85]
blaGES-5, blaKPC-3, blaNDM-1	Klebsiella pneumoniae, Enterobacter spp., Citrobacter spp.	river water		[86]
blaCphA, blaL1, blaOXA-48-like, blaVIM-2	Aeromonas spp., Stenotrophomonas maltophilia, Stenotrophomonas xiamenensis, Pseudomonas spp.	river water		[57]
blaNDM-5	Escherichia coli ST410	reservoir water	Singapore	[87]
-	Escherichia coli, Klebsiella pneumoniae	river water	Tanzania	[88]
blaOXA-204, blaNDM-1	Shewanella spp.	river water	Tunisia	[89]
blaIMI-2, blaL1	Enterobacter asburiae, Stenotrophomonas maltophilia	lake water, pond water	USA	[58]
blaOXA-252, blaOXA-547	Shewanella xiamenensis	water canals	Vietnam	[90]

. Escherichia coli (n = 8, 13.5%) and Klebsiella pneumoniae (n = 10, 16.9%) were identified as the most common carbapenemase-producing isolates. Moreover, carbapenem resistance genes have also been confirmed in several clinically significant bacteria belonging to the Acinetobacter, Enterobacter, and Pseudomonas genera. However, these genes were also found in typical environmental, non-clinical isolates of the genera Aeromonas, Comamonas, Shewanella, Stenotrophomonas, and Vibrio, indicating that the natural environment is becoming an active reservoir of resistance to last-line drugs. In addition, it should be emphasized that members of the genus Stenotrophomonas have inherent resistance to a wide range of β-lactam antibiotics due to two β-lactamases encoded chromosomally (by bla_L1/bla_L2 genes) [56]. Various samples from natural water sources contained such isolates [57,58]. Interestingly, some studies have reported Stenotrophomonas maltophilia isolated from seawater that carried the bla_OXA-58 gene [19]. This discovery illustrates that environmental bacteria can acquire and integrate clinically relevant resistance genes, potentially leading to their transmission and contributing to the spread of AMR.

Regardless of taxonomy, various bacteria have been reported to harbor several genetic determinants that enable the production of different carbapenemases. Most of the bibliographic data reports originated from China (17%) and Brazil (11%), while river water samples were the most commonly examined (44%) (Table 1).

Based on available genomic data from bacteria isolated from surface waters, we compiled information on carbapenemase-carrying isolates originating from 27 countries worldwide. The largest numbers of genomes were derived from isolates collected in Brazil (n = 31; 14.15%), Ghana (n = 31; 14.15%), and China (n = 22; 10.04%) (Figure 2A). Across all datasets, the predominant hosts of carbapenemase genes were members of the genera Escherichia/Shigella, Klebsiella, and Enterobacter (Figure 2B). The most frequently detected carbapenemase families included NDM (n = 76, 33.33%), KPC (n = 45, 19.73%), and OXA (n = 37, 16.22%), and several genomes carried more than one carbapenemase gene (n = 9, 4.11%), indicating co-occurrence of multiple resistance determinants within single strains (Figure 2C). Most genomes were classified as Escherichia coli (n = 60, 27.40%) and Klebsiella pneumoniae (n = 55, 25.11%), which not only carried the widest diversity of carbapenemase genes but also most commonly accumulated multiple carbapenemase determinants within the same genome (Figure 2D).

Environmental bacteria constitute an important reservoir of natural carbapenem resistance in aquatic settings, mediated by genes such as bla_CphA (in Aeromonas spp.), bla_VCC (in Vibrio cholerae), and bla_L (in Stenotrophomonas maltophilia) (Figure 2D). Clinically relevant carbapenemase genes—including bla_NDM, bla_KPC, and bla_OXA—were also identified in genomes of typically environmental bacteria, such as those belonging to the genera Vibrio, Shewanella, and Aeromonas, underscoring the increasing role of natural aquatic habitats as reservoirs and potential exchange hubs for high-priority resistance genes.

Among the analyzed bacterial genomes, 73 carbapenem resistance genes (32.02%) were successfully linked to specific plasmid replicons based on their co-localization on the same genomic contig (Figure 3). Conversely, for 155 ARGs (67.98%), a definitive plasmid association could not be established. Mobilization was confirmed for six carbapenemase family genes (NDM, OXA, KPC, VIM, GES, and FRI), which were associated with 21 unique plasmid replicon types across six bacterial genera. Klebsiella species exhibited the highest diversity in terms of plasmid-mediated resistance, carrying carbapenemase genes on 13 distinct plasmid types. The COPLA tool successfully assigned the analyzed plasmid contigs harboring carbapenemase genes to specific Plasmid Taxonomic Units (PTUs) in 61 out of 73 cases (83.56%), thereby predicting their theoretical host range. Based on the COPLA analysis, the identified carbapenemase-carrying plasmids exhibited a diverse range of potential hosts, spanning from the narrowest at the species level (IncR), through the family level (ColKP3; IncX3; IncFIB(AP001918)_IncFIA_IncFII(pRSB107); IncFII_IncFIA; IncFII(K)_IncFIB(pB171); IncFII(Yp)_ IncFIB(pB171); pKPC-CAV1321; IncR_IncFII(K); IncFII(K)), to the order level (IncN; IncL; IncM1), and reaching the class level (IncC and IncQ2) (Figure 3, File S1). Notably, the IncC plasmid was confirmed as a broad-host-range vector, facilitating the presence of bla_NDM in both Enterobacterales and Vibrio isolates (Figure 3).

As the sole non-Enterobacterales representatives within the observed ARG–plasmid–host associations, three Vibrio genomes were identified as carriers of the bla_NDM-1 gene on IncC-type plasmids (Figure 3). In-depth analysis of these specific genomes revealed that the plasmids were reconstructed using long-read sequencing technology, which ensured high assembly continuity and provided complete, circularized sequences. Specifically, our comparative plasmid analysis (Figure 4) demonstrated that these Vibrio vulnificus isolates, recovered from a coastal beach in Nigeria, carried a plasmid identical/nearly identical to those found in several clinical Enterobacterales isolates, while a related plasmid was also reported in Providencia vermicola from the Democratic Republic of Congo. This geographic proximity of environmental and clinical reservoirs, combined with the high mobility of plasmids carrying carbapenemase genes, underscores the potential for inter-species and inter-ecosystem dissemination. Together, these observations illustrate how aquatic environments can act as contact zones where environmental bacteria and clinically derived resistance plasmids co-occur, enabling genetic exchange and contributing to the broader spread of AMR.

3.2. Mobile Colistin Resistance in the Natural Aquatic Environment

The global distribution of mcr genes, responsible for coding mobile resistance to polymyxin E, is influenced by HGT through plasmids, as well as transposons and integrons [91]. The mcr genes are generally harbored by various conjugative and non-conjugative plasmids, but in some isolates, they can also be integrated into the chromosome [92]. Interestingly, it is assumed that because of the variability in the sequence similarity, mcr genes have a different genetic origin. The sequence comparison reveals specific variations; the mcr-1 gene shares varying degrees of amino acid sequence identity with other mcr variants: mcr-2 (81%), mcr-3 (34%), mcr-4 (33%), mcr-5 (31%), mcr-6 (82%), mcr-7 (29%), and mcr-8 (31%) (Mondal 2024) [30]. Moreover, various variants of mcr genes are possessed by different species of Gram-negative bacteria. In any environment, whether clinical or natural, nearly 90% of all mcr-positive isolates are identified as Escherichia coli [30,93]. In general, the mcr-1 and mcr-9 genes are most frequently reported among Enterobacterales, mcr-3 in Aeromonas sp., and mcr-10 in Enterobacter sp. [30,94]. The findings from the review of bibliographic data regarding the presence of mcr genes in natural aquatic environments are consistent with established knowledge and are summarized in

Table 2. Summary of a review of bibliographic data on the presence of mcr genes and their hosts in samples from natural aquatic environments.

Mobile Colistin Resistance Gene	Bacterial Host	Environmental Water Type	Country	Reference
mcr-5.1	Cupriavidus gilardii	well water	Algeria	[99]
mcr-1.1	Escherichia coli	drinking water	Armenia	[100]
mcr-1	Escherichia coli	surface water	Bangladesh	[101]
mcr-3	Escherichia coli	pond water		[102]
mcr-1, mcr-1.26	-	river water	Brazil	[103]
mcr-3, mcr-7.1	-	water from a recreation club		[104]
mcr-1, mcr-2, mcr-3, mcr-4, mcr-5, mcr-6, mcr-7, mcr-8, mcr-9	Acinetobacter spp., Enterobacter spp.	urban recreational estuary water		[105]
mcr-1	Escherichia coli ST683/CC155	touristic coastal water		[106]
mcr-3, mcr-3.6	Aeromonas veronii TR112	river water		[107]
mcr-1	-	river water	China	[108]
mcr-1	Escherichia coli	river water		[109]
mcr-1	Escherichia coli	watershed		[110]
mcr-1, mcr-3	Escherichia coli, Enterobacter cloacae, Aeromonas veronii, Aeromonas hydrophila	river water		[111]
mcr-1	-	drinking water supply system		[70]
mcr-3	Aeromonas spp.	river water		[112]
mcr-1	Escherichia coli	well water		[69]
mcr-3	Stenotrophomonas maltophilia	lake water		[113]
mcr-1	Escherichia coli	river water		[114]
mcr-3	Aeromonas spp.	river water, well water, mud water, surface water	Ghana	[115]
mcr-1, mcr-2, mcr-3	Salmonella spp.	river water, lake water	India	[116]
mcr-1	Escherichia coli	irrigation water	Lebanon	[117]
mcr-3, mcr-5	-	river water storm water	South Africa (Western Cape)	[92]
mcr-10	Enterobacter cloacae	river water	Switzerland	[118]
mcr-1	Escherichia coli	river water		[119]
mcr-1, mcr-8, mcr-9	Escherichia coli	surface water	Thailand	[120]
mcr-3, mcr-4, mcr-5, mcr-8	Aeromonas veronii Aeromonas media	springs, ponds, drinking water	Turkey	[121]
mcr-3	Aeromonas jandaei	surface water	USA	[122]
mcr-1.1	Escherichia coli	irrigation water		[123]
mcr-1	-	lake water, river water, groundwater	Vietnam	[124]

. Escherichia coli was identified as the most common mcr-positive isolate (n = 13; 46%). However, the presence of mcr genes has been confirmed in several clinically important bacterial species, including Escherichia coli, Klebsiella pneumoniae, and bacteria belonging to Salmonella sp. and Enterobacter sp. Similarly to the hosts of carbapenem resistance genes (see Table 1), mcr genes were also found in bacteria unrelated to clinical settings, originating from typical natural sources such as the genera Aeromonas, Stenotrophomonas, and Cupriavidus. This underscores the significant role of the environment, both as a medium and as a source of typical environmental bacteria that carry resistance to critical drugs. Notably, bacteria from the Aeromonas genus have been identified to possess the mcr-3 gene in their genomes (n = 6). The scientific literature suggests that the origin of mcr-3 can be traced back to this Gram-negative bacterium, which is commonly found in soil and water environments. Some studies have shown that mcr-3 originated and evolved from Aeromonas species, eventually leading to its presence in significant members of the Enterobacteriaceae family, such as Escherichia coli and Klebsiella pneumoniae [95,96,97]. Although we have not found any reports of colistin-resistant representatives of the water environmental genus Shewanella, it should be emphasized that the scientific literature indicates this bacterium, found in marine environments and aquatic ecosystems, possesses nonmobile colistin resistance determinants (nmcr), which are probably progenitors of the mcr-4 gene [33,98].

A review of bibliographic data shows that various microorganisms can simultaneously carry numerous types of mcr genes. River water was the most frequently studied sample among natural water reservoirs (n = 13). Notably, the majority of these reports originated from China (31%) and Brazil (17%) (Table 2), as in the case of reports regarding carbapenem resistance bacteria isolated from natural water environment (Table 1).

Based on available genomic data from bacteria isolated from surface waters, we compiled information on mcr gene-carrying isolates originating from 11 countries worldwide. The largest numbers of genomes were derived from isolates collected in Brazil (n = 21, 30.43%), Lebanon (n = 19, 27.54%), and China (n = 8, 11.59%) (Figure 5A). The vast majority of reported genomes belonged to members of the genera Escherichia/Shigella (n = 49, 71.01%) (Figure 5B). In total, five unique mcr gene variants were detected among bacterial genomes originating from surface waters, with mcr-1 being the most prevalent (n = 47, 64.38%) (Figure 5C). All mcr variants were detected in representatives of Enterobacterales (Figure 5D), with the exception of mcr-3, which is strongly associated with the genus Aeromonas—a potential progenitor of this resistance mechanism [115].

The mobilization analysis of mobile colistin resistance genes revealed that 47 mcr genes (64.38%) were successfully linked to plasmid replicons, while 26 genes (35.62%) remained non-typeable (Figure 6). The mcr genes were identified across seven unique plasmid types within the genomes of three distinct bacterial genera belonging to the order Enterobacterales. A high degree of replicon specificity was observed, particularly for mcr-1, which was predominantly associated with the IncX4 and IncI2 plasmids in Escherichia/Shigella. Furthermore, the IncHI2 and IncHI2A plasmids exhibited the greatest diversity, both in terms of the mcr variants they carried (mcr-1 and mcr-9) and their bacterial host range, which included Escherichia/Shigella, Enterobacter, and Klebsiella. The COPLA tool successfully assigned the analyzed plasmid contigs harboring mcr genes to specific Plasmid Taxonomic Units in 43 out of 46 cases (93.48%), thereby predicting their theoretical host range (File S1). According to the COPLA analysis, these identified mcr-carrying plasmids exhibited a diverse range of potential hosts, spanning from the family level (IncX4 and IncI2) to the order level (IncHI2_IncHI2A and IncN_IncHI2_IncHI2A) (Figure 6, File S1).

3.3. Co-Resistance to Carbapenems and Colistin Among Bacteria from Environmental Water Sources

Several studies have identified colistin- and carbapenem-resistance genes simultaneously occurring in bacteria isolated from various natural water sources (

Table 3. Summary of information on isolates simultaneously carrying genetic determinants for carbapenem resistance (CRG column) and mobile colistin resistance (MCR column). The data is derived from both a comprehensive literature search and the analysis of publicly available genomes in the NCBI database. Isolates for which full genomic data is available (indicated by the Assembly ID) are further characterized in detail in Figure 7.

Bacterial Host	Isolation Source	Year	Location	CRGs	MCR	No. of Isolates	Detection Method	Assembly ID	Reference
Klebsiella pneumoniae	Surface water	2021	Brazil: Restinga	blaKPC-2	mcr-10.1	1	WGS	GCA_052220985.1	-
Enterobacter hormaechei	Surface water	2015	Canada: Ontario	blaVIM-1	mcr-9.1	3	WGS	GCA_015910325.1 GCA_015910345.1 GCA_022023895.1	[126]
Klebsiella pneumoniae	Natural water sources	2022	Thailand: Ratchaburi	blaNDM-1	mcr-1.1	2	WGS	GCA_029542165.1 GCA_029542185.1	[127]
Enterobacter hormaechei	Eutrophic lake	2021	India: Puducherry	blaNDM-1	mcr-9.1	1	WGS	GCA_025290855.1	[128]
Enterobacter kobei	Coastal Water	2014	Brazil: Rio de Janeiro	blaKPC-2	mcr-9.1 mcr-5.1 mcr-10.1	1	WGS	GCA_024623675.1	[129]
		2013			mcr-9.1	1		GCA_024623685.1
Escherichia coli	River	2021	China: Shandong Province	blaNDM	mcr-1	10	WGS	-	[130]
Escherichia coli	River	2021	Thailand: Nongkhai Province	blaOXA-48	mcr-8	2	PCR	-	[120]
					mcr-9.1	2
					mcr-8 mcr-9	6
Citrobacter freundii	River	2022	Germany: Lower Saxony	blaKPC-2 blaVIM-1	mcr-9	1	WGS	-	[131]

CRGs—carbapenem resistance genes; MCR—mobile colistin resistance.

). These reports are limited to representatives of Enterobacterales—Enterobacter (n = 3), Klebsiella (n = 2), Escherichia (n = 2), and Citrobacter (n = 1)—recovered from diverse types of waters across the world, which is consistent with global clinical trends [125]. The authors of these studies highlight direct links between the environmental isolates listed in Table 3 and related clinical isolates [126,127] as well as isolates recovered from discharge water originating from a nearby hospital and a fish farm [120]. Of particular concern is the fact that the waters from which these isolates were obtained included, among others, a lake serving as a source of irrigation water for surrounding villages [128] and coastal waters suitable for primary contact recreation [129].

Genomic data derived from available long-read sequencing assemblies highlight the high mobility of carbapenem and colistin resistance determinants within bacterial cells, carried on plasmids together with additional antimicrobial resistance genes (Figure 7). Plasmids encoding resistance to last-resort antibiotics also exhibited substantial variability, despite occurring in bacterial isolates that were otherwise highly similar. Although carbapenemase and mcr genes were typically located on separate plasmids, co-localization on a single plasmid was also observed, for instance, in Enterobacter kobei.

Although co-resistant bacteria pose a significant threat, they currently appear to be mostly sporadic; nevertheless, it remains crucial to implement appropriate measures to limit their dissemination [125]. The reported findings also indicate that the detection of multiple isolates (2–10) within a single sample is unlikely to be accidental, suggesting either in situ proliferation or a strong point-source contamination event.

4. Discussion

The findings of this review highlight an important aspect of the ‘One Health’ approach, which encompasses human, animal, and environmental health sectors—the last of which, unfortunately, is often overlooked in global surveillance. Our work is intended to initiate intensified efforts to ensure that carbapenems and colistin remain effective treatment options in clinical settings. However, the presented findings extend beyond expected patterns, showing that the aquatic environment serves as a dynamic evolutionary interface. The results of this review are consistent with global monitoring trends, where the high prevalence of ARGs in Enterobacterales is a logical consequence of their WHO priority status and focused surveillance. However, the path to a comprehensive global strategy for monitoring and combating AMR faces certain challenges. Although this review synthesizes data from both bibliographic sources and genomic analyses—covering 41 countries across several continents—significant geographical gaps remain. There is a notable disproportion in the number of available reports, with a dominance of data originating from China and Brazil. Consequently, many regions remain underrepresented, highlighting the urgent need for more geographically diverse surveillance to achieve a truly global understanding of the mechanisms driving the dissemination of carbapenem and colistin resistance.

The data presented in the preceding sections show that genetic determinants of carbapenem and colistin resistance are widespread among bacteria inhabiting aquatic environments—both within the native environmental microbiota and among members of Enterobacterales. These two distinct microbial groups coexist and interact within water systems, creating opportunities for evolutionary adaptation, HGT, and ecological expansion. Such continuous circulation of microorganisms and their resistance determinants across natural environments, animals, and humans reflects the interconnected dynamics central to the One Health framework (Figure 8).

An intriguing observation emerges regarding the role of Escherichia coli in aquatic resistomes. Although E. coli was identified as the predominant host for mcr and carbapenemase genes in surface waters, its associated plasmids—typically IncX3, ColKP3, and IncF for carbapenemases and IncX4 or IncI2 for mcr—appear to exhibit a more restricted host range, potentially limiting their dissemination to indigenous aquatic genera such as Vibrio, Aeromonas, and Pseudomonas [132,133]. In contrast, our findings suggest that other Enterobacterales species, particularly Klebsiella and Enterobacter, may play a more pivotal role as high-risk reservoirs. Specifically, Klebsiella is recognized as a critical hub for the accumulation and transmission of resistance determinants [42,134], while together with Enterobacter, it serves as a primary driver for the evolution of AMR across diverse environments [135]. These taxa were associated with broad-host-range scaffolds, including IncC and IncQ2, which facilitate cross-taxonomic resistance transfer. While the current scarcity of environmental genomic data warrants caution, these patterns suggest that the ecological risk of AMR dissemination in water bodies is driven by the plasticity of the plasmid scaffold and the fitness of the intermediary host.

Figure 8. Schematic summary of data regarding the prevalence of bacteria carrying carbapenem resistance genes and mobile colistin resistance genes in natural surface waters. Based on the data collected in this study, the figure highlights the fraction of last-resort antibiotic resistance genes detected in waterborne Enterobacterales, as well as in environmental bacteria identified in the review. The specific host taxa for these resistance mechanisms are listed in the bottom right corner. Potential transmission routes of the aforementioned bacterial taxa are presented, considering the water environment-to-human pathway based on [136,137,138], the environment-to-animal pathway [136,139,140,141,142,143,144], and the animal-to-human pathway [136,140], within the context of the One Health concept. Green arrows indicate potential transmission routes for bacteria classified as ‘environmental bacteria’, while red arrows represent those classified as ‘Enterobacterales’. Gray arrows denote the potential influx of bacteria into the ‘Environment’ component. Dashed lines indicate transmission occurring independently of the ‘Environment’ component.

Figure 8. Schematic summary of data regarding the prevalence of bacteria carrying carbapenem resistance genes and mobile colistin resistance genes in natural surface waters. Based on the data collected in this study, the figure highlights the fraction of last-resort antibiotic resistance genes detected in waterborne Enterobacterales, as well as in environmental bacteria identified in the review. The specific host taxa for these resistance mechanisms are listed in the bottom right corner. Potential transmission routes of the aforementioned bacterial taxa are presented, considering the water environment-to-human pathway based on [136,137,138], the environment-to-animal pathway [136,139,140,141,142,143,144], and the animal-to-human pathway [136,140], within the context of the One Health concept. Green arrows indicate potential transmission routes for bacteria classified as ‘environmental bacteria’, while red arrows represent those classified as ‘Enterobacterales’. Gray arrows denote the potential influx of bacteria into the ‘Environment’ component. Dashed lines indicate transmission occurring independently of the ‘Environment’ component.

Wastewater discharges, runoff, and other human-associated inputs represent a major pathway through which clinically associated bacteria, including healthcare-linked pathogens, are released into surface waters [145]. In this way, aquatic ecosystems become recipients of genetic material previously confined to clinical settings, including last-resort resistance determinants located on mobile genetic elements. Our review indicates that plasmids such as IncQ2 and IncC serve as particularly critical vectors for transmission beyond the Enterobacterales order due to their broad predicted host range. By enriching the environmental resistome with such clinically relevant components, these inputs create conditions that promote extensive interactions between hospital-derived strains and the complex native microbiota. Within this shared habitat, the boundaries between environmental and clinical microbial communities begin to blur. Close coexistence facilitates HGT, enabling the transfer of antimicrobial resistance genes across taxonomic and ecological barriers. Consequently, microorganisms that are already well adapted to natural conditions may acquire highly efficient resistance systems characteristic of clinical pathogens, potentially increasing their ecological competitiveness.

Importantly, this exchange may be bidirectional. Numerous studies indicate that several clinically relevant AMR mechanisms, including resistance to colistin and carbapenems, have environmental origins [89,97,146,147]. Thus, it is plausible to consider the aquatic environment as a cryptic reservoir where yet-undiscovered resistance mechanisms may reside, diversify, and potentially become mobilized within natural habitats [148]. Currently, environmental bacteria are well known to possess intrinsic, chromosomally encoded carbapenem resistance genes—including bla_L (Stenotrophomonas maltophilia), bla_CphA (Aeromonas spp.), bla_POM (Pseudomonas otitidis), and bla_VCC (Vibrio cholerae). While our analysis failed to demonstrate the mobility of these determinants (specifically their plasmid-borne localization), the emergence of their mobile variants in clinical settings has already been documented [149]. Such novel or currently overlooked variants may ultimately enter clinical settings, where continuous and intense selective pressures facilitate their emergence as significant threats [147].

Numerous reports describe infections caused by environmental bacteria which, as shown in our review, frequently harbor genes conferring resistance to last-resort antibiotics. These infections include open-wound and burn-related injuries following contact with water and have been attributed to diverse taxa such as Vibrio spp., Shewanella spp., Aeromonas spp., Escherichia coli, Enterobacter spp., Klebsiella pneumoniae, Serratia marcescens, and Citrobacter spp. [137,138,140,150]. In several cases, infections resulted in severe clinical outcomes, including limb amputation or death. Transmission may occur during routine occupational activities [140,150], through recreational exposure [140], or through the use of river water as a drinking source [136]. Animals also constitute an important vector: fish, oysters, and shrimp can both carry and become infected with opportunistic environmental bacteria [136,139,140,141], as well as Enterobacterales resistant to carbapenems or/and colistin [141,142,143].

Hospitalization due to waterborne infections commonly requires antimicrobial treatment; however, increasing resistance to first-line antibiotics significantly constrains therapeutic options [138,151]. As a result, last-resort antibiotics are increasingly used, including in severe or rapidly progressing infections among immunocompetent individuals [152]. Continued acquisition of last-resort antibiotic resistance among these taxa could therefore further complicate the management of water-associated infections.

The evidence presented above indicates that such processes are already underway and are likely to intensify in the future, partly as a consequence of climate change [153]. We may increasingly observe both the emergence of environmental bacteria in clinical settings and the re-entry of hospital-associated pathogens into the clinic via waterborne infections. This scenario provides environmental bacteria with expanding opportunities to enter clinical ecosystems, where frequent contact with clinically relevant pathogens and strong selective pressures facilitate further acquisition of antimicrobial resistance mechanisms. Ultimately, these bacteria may again be released from healthcare facilities into surface waters through wastewater effluents, thereby closing the cycle of bacterial circulation across the interconnected components of the One Health framework.

Taken together, these findings underscore that aquatic environments function not merely as passive recipients, but as dynamic, evolutionarily active hubs that mediate the continuous exchange, adaptation, and dissemination of high-risk antimicrobial resistance determinants.

When investigating carbapenem and colistin resistance in aquatic environments, it is essential to recognize that many environmental bacteria possess intrinsic, taxon-specific resistance mechanisms that may resemble acquired resistance phenotypes. This is particularly relevant when screening for clinically significant determinants such as the “Big Five” carbapenemases (KPC, NDM, VIM, IMP, and OXA-48-like), whose detection in environmental isolates can be confounded by naturally occurring β-lactamase activities. Therefore, robust risk assessment requires confirmation of phenotypic findings using molecular approaches capable of distinguishing intrinsic resistance from horizontally acquired ARGs. Whole-genome sequencing, especially long-read technologies, provides a powerful framework for this purpose, as it enables reconstruction of complete plasmid sequences and resolves the genomic context of resistance genes with high precision. Such data are invaluable for tracking the mobility, dissemination pathways, and evolutionary dynamics of clinically relevant ARGs in natural settings. Taxonomic identification remains a critical first step, as knowledge of species-specific intrinsic resistomes, ecological niches, and pathogenic potential allows researchers to interpret resistance profiles more reliably and to differentiate environmental background signals from emerging public health threats. However, investigating native aquatic bacteria remains paramount, regardless of their intrinsic resistance. Unlike transient clinical strains, these indigenous species are highly fitness-adapted to natural waters, allowing them to effectively stabilize and maintain acquired ARGs like bla_NDM or bla_KPC. By providing a stable genomic backbone for clinical determinants, these well-adapted populations turn into permanent environmental reservoirs, ensuring the long-term persistence of last-resort resistance in the ecosystem.

5. Conclusions and Future Directions

5.1. Public Health and One Health Implications

Our review identifies two distinct facets of carbapenem and colistin resistance within natural aquatic environments. The first involves allochthonous Enterobacterales, which act as the primary vectors for clinically significant ARGs, frequently localized on mobile plasmids. The second facet comprises the autochthonous aquatic microbiota, which typically harbor intrinsic, non-mobile resistance mechanisms. Natural water bodies serve as a dynamic interface for interaction between these two bacterial fractions.

A particularly critical finding is the simultaneous presence of both carbapenem and colistin resistance determinants within single bacterial cells, even co-located on the same plasmid. This phenomenon of co-resistance on shared MGEs highlights the extreme risk of multidrug resistance dissemination in the aquatic environment.

The findings of this review demonstrate that natural aquatic environments have transitioned from passive recipients of pollution to active, stable reservoirs for carbapenem and colistin resistance. This evolution poses a direct threat to public health by creating environmental routes for high-risk pathogens that can bypass traditional clinical surveillance. We conclude that the presence of genes that provide resistance to last-resort options in both Enterobacterales and autochthonous microbiota increases the risk of community-acquired infections that are inherently difficult to treat. Ultimately, gaining a deeper understanding of these environmental reservoirs is essential to safeguard the clinical efficacy of carbapenems and colistin, ensuring they remain viable treatment options for life-threatening infections.

To effectively address the global AMR crisis, the environmental sector should be recognized as an important driver of resistance. Aquatic systems, in particular, act as a primary reservoir where plasmids emerge as significant vectors for carbapenemase and mcr gene dispersal. From a One Health perspective, these water bodies represent the missing link; ignoring them compromises the efficacy of interventions in other sectors by allowing resistance to circulate freely between human, animal, and environmental compartments.

5.2. Limitations and Future Research

The data synthesized in this review highlight three primary strategic directions for future research concerning the prevalence and dynamics of critical resistance genes in aquatic environments:

• Monitoring allochthonous vectors: Research should continue to focus on the detection of Enterobacterales carrying ARGs against last-resort antibiotics. As the primary “introducers” of clinically significant, mobile ARGs into natural waters, these organisms represent the most direct threat to public health. While this area is currently the most extensively studied, continued surveillance is essential to track emerging high-risk clones.
• Investigating native recipients: There is an urgent need to shift focus toward identifying clinically relevant ARGs within the natural (autochthonous) microbiota. This will allow for an assessment of which environmental organisms serve as the most effective “recipients” of critical ARGs and which MGEs play the most significant roles in the stabilization and long-term environmental persistence of AMR. Based on our findings, particular attention should be directed toward the genera Aeromonas, Pseudomonas, Vibrio, Shewanella, and Comamonas. This area remains significantly under-researched, with genomic data for these groups still being relatively scarce.
• Evolution of intrinsic resistomes: It is crucial to monitor known intrinsic resistance mechanisms and discover novel environmental determinants. Tracking the specific genes—both in the environment and in clinical settings—will enable researchers to observe the potential transition of innate ARGs toward mobility, providing an early warning system for future clinical threats.

Furthermore, a significant limitation of currently available data is the high degree of genomic fragmentation, which obscures critical ARG–plasmid associations and limits comparative analyses. The adoption of long-read sequencing technologies is essential to reconstruct complete plasmid sequences. Such high-resolution data will enable precise tracking of the dissemination and structural evolution of last-resort resistance across all One Health sectors.