Preplanned Studies: High Prevalence and Genomic Characterization of Extended-Spectrum β-Lactamase-Producing Escherichia coli in the Yellow River and Source Water from A One Health Perspective — Henan Province, China, 2023–2024

Introduction: Understanding the prevalence and dissemination pathways of the clinically relevant extended-spectrum β-lactamase (ESBL) genes and ESBL-producing Escherichia coli (E. coli) within the Yellow River is essential from a One Health perspective to control antibiotic resistance dissemination from environmental reservoirs to human populations.

Methods: Water samples were collected from the Yellow River and two of its major tributaries in Henan Province during 2023 and 2024. TaqMan quantitative polymerase chain reaction (qPCR) was used to quantify the abundance of the ESBL gene bla_CTX-M-G9. Twenty-three E. coli isolates underwent whole-genome sequencing (WGS), and 167 publicly available E. coli genomes from diverse sources were incorporated into the comparative phylogenetic analysis.

Results: The bla_CTX-M-G9 gene was ubiquitous across all sampling sites, exhibiting significantly higher relative abundance during the dry season compared to the flat season. The multidrug-resistant E. coli sequence type (ST) 6802 carrying bla_CTX-M-14 emerged as the predominant clone. Strong positive correlations were observed between bla_CTX-M-G9 abundance and plasmids carried by E. coli ST6802 in the Yellow River, providing evidence for clonal expansion during the dry season. Furthermore, comparative phylogenetic analysis integrating human, animal, and environmental isolates demonstrated that ST6802 strains from this study were closely related to those previously identified in anaerobic digestion systems treating pig manure in China, suggesting an animal-to-environment transmission pathway.

Conclusion: These findings emphasize the urgent need to implement targeted interventions that prevent the transmission of antibiotic resistance from animal sources into aquatic environments, thereby protecting public health and preserving the integrity of critical water resources.

The emergence and spread of extended-spectrum β-lactamase (ESBL)-producing Escherichia coli (E. coli) poses a substantial threat to public health worldwide. In 2019, antimicrobial resistance accounted for an estimated 145,000 deaths in China, with ESBL-producing E. coli ranking as the third leading cause (8,032 deaths) (1). Recognizing this growing concern, the World Health Organization launched the “ESBL E. coli tricycle AMR surveillance project” in 2017, which encompasses monitoring across human populations, the food chain, and environmental reservoirs. Environmental surveillance has historically been the weakest component of this integrated approach. However, recent years have witnessed increasing attention to ESBL-producing E. coli surveillance in environmental settings (2). Among the diverse ESBLs, the cefotaxime from Munich (CTX-M) has emerged as the most prevalent ESBL type globally. In China, E. coli strains carrying CTX-M group 9 variants, particularly the CTX-M-14 subtype, predominate, despite CTX-M group 1 subtype CTX-M-15 being the most widespread variant internationally (3). Surveillance data from Chinese hospitals revealed a high prevalence (62.8%) of CTX-M group genes in E. coli isolated from outpatients. The most common genotype was the CTX-M group 9 subtype bla_CTX-M-14 (57.7%), followed by bla_CTX-M-55 (23.4%) and bla_CTX-M-15 (15.4%) (4). Notably, E. coli isolates from community-acquired urinary tract infections across various regions demonstrated a high prevalence of bla_CTX-M-14 genotype, with an incidence rate of 31.8% (5). Our previous research identified bla_CTX-M-14 as the dominant genotype in E. coli from anaerobic digestion systems treating pig manure (6).

The Yellow River, China's second longest river at 5,464 km, plays a critical role in water supply and agricultural irrigation throughout northern China. Flowing through densely populated urban areas and extensive agricultural lands, the river basin receives environmental discharges exceeding 3,000 tons of antibiotics annually (7), thereby functioning as both a reservoir and transmission medium for antibiotic resistance genes (ARGs). Understanding the prevalence, transmission dynamics, and potential sources of CTX-M group 9 genes and CTX-M-producing E. coli in the Yellow River and its associated source waters is essential from a One Health perspective to effectively control ARG dissemination from environmental reservoirs to human populations. Despite this critical need, the abundance of CTX-M group 9 genes and the dissemination routes of CTX-M-producing E. coli in the Yellow River remain poorly characterized.

Henan Province is located in the east-central region of China, encompassing the middle and lower reaches of the Yellow River. Although urban areas have primarily relied on the South-to-North Water Transfer Project since 2014, several regions in Henan continue to utilize water from the Yellow River as their primary or backup source for drinking water, irrigation, and landscaping. In this study, we collected water samples from the Yellow River at six cities in Henan Province (Lingbao, Luoyang, Jiaozuo, Xinxiang, Zhengzhou, and Puyang), an associated water source in Zhengzhou City, and two major tributaries (the Qin River and the Yiluo River) upstream of the water source. Sampling was conducted during three distinct hydrological periods: the wet season (August 2023), the dry season (December 2023), and the flat season (April 2024).

We employed TaqMan quantitative polymerase chain reaction (qPCR) to quantify the abundance of the bla_CTX-M-G9 gene and its associated mobile genetic elements (MGEs). The qPCR primers are provided in Supplementary Table S1. During the dry season, we filtered 100 mL water samples from each sampling point through 0.45 µm microporous membranes for bacterial isolation. The filtered membranes were incubated on fuchsin sodium sulfite medium supplemented with 4 mg/L cefotaxime at 37 °C for 24 h. All colonies were isolated and purified. A total of 75 cefotaxime-resistant E. coli strains were recovered from Lingbao (n=10), Luoyang (n=7), Yiluo River (n=10), Jiaozuo (n=1), Qin River (n=4), Xinxiang (n=11), the water source in Zhengzhou City (n=17), and Puyang (n=15). All strains underwent polymerase chain reaction (PCR) detection of CTX-M group 9 genes and antibiotic susceptibility testing by disk diffusion methods as previously reported (6, 8-9). We selected 23 E. coli isolates for whole-genome sequencing (WGS) based on their diverse antibiotic resistance profiles (Supplementary Table S2), ensuring representation of the range of resistance patterns observed. Additionally, we downloaded 167 E. coli genomes from the National Center for Biotechnology Information (NCBI) database (CTX-M-14-producing E. coli, n=162) and the EnteroBase database (E. coli ST6802, n=5) for comparative phylogenetic analysis.

WGS of 23 E. coli strains was performed using the Illumina NovaSeq PE150 platform (Sinobiocore Biotechnology Co., Ltd., Beijing, China). The genome sequences were deposited in the NCBI database under accession number PRJNA1306565. The assembled contig files were uploaded to the Centre for Genomic Epidemiology (CGE) platform (http://www.genomicepidemiology.org/) for screening of ARGs, virulence factors (VFs), and plasmid replicon types, as well as multilocus sequence typing (MLST) analysis. Single nucleotide polymorphisms (SNPs) were determined on the CGE platform, and the SNP-based phylogenetic tree was visualized in iTOL online. The complete genome multilocus sequence typing (cgMLST) analysis was conducted using PubMLST (https://pubmlst.org/) and R (Version 4.4.1, R Foundation for Statistical Computing, Vienna, Austria). The cgMLST-based minimum spanning tree was generated using GrapeTree (Version 1.5.0, University of Melbourne, Melbourne, Australia).

The bla_CTX-M-G9 gene was ubiquitously detected across all sampling sites in the Yellow River, its tributaries, and the associated water source (Figure 1A). Relative abundance of bla_CTX-M-G9 was significantly higher during the dry season (10^-5-10^-3 copies/16S rDNA) compared to the flat season (10^-6-10^-4 copies/16S rDNA) (Figure 1B), with peak levels observed in the water source during the dry season. Antibiotic susceptibility testing results for all 75 cefotaxime-resistant E. coli isolates are presented in Supplementary Table S2. Among these isolates, 98.7% (n=74) harbored the bla_CTX-M-14 genotype and demonstrated multidrug resistance phenotypes, with 78.7% (n=59) exhibiting resistance to seven or more of the tested antibiotics. Whole-genome sequencing revealed that 21 of 23 E. coli strains belonged to sequence type (ST) 6802, all of which carried bla_CTX-M-14 (Figure 1C). These ST6802 strains co-harbored 8–13 additional ARGs, 19 VFs, and two to four plasmid replicons (IncFIB, IncHI2, IncHI2A, and IncX1). SNP-based phylogenetic analysis demonstrated zero to four SNP differences among all ST6802 isolates, confirming clonal dissemination of ST6802 throughout the Yellow River. Moreover, during the dry season, bla_CTX-M-G9 abundance exhibited strong correlations with the plasmid replicons IncHI2 (R²=0.86), IncHI2A (R²=0.75), IncX1 (R²=0.83), and IncFIB (R²=0.67) in the Yellow River, providing additional evidence for clonal ST6802 spread (Figure 2).

Figure 1. 

Prevalence of the bla_CTX-M-G9 gene and CTX-M-producing E. coli in the Yellow River and associated source water. Relative abundance of bla_CTX-M-G9 across different (A) cities and (B) seasons. (C) SNP-based phylogenetic tree of 23 CTX-M-producing E. coli isolates, displaying associated antibiotic resistance genes, virulence factors, and plasmid replicon types.

Note: For (A) and (B), one-way analysis of variance (ANOVA) with Tukey's post-hoc tests was employed to assess significant differences among groups.

Abbreviation: LB=Lingbao City; LY=Luoyang City; JZ=Jiaozuo City; XX=Xinxiang City; PY=Puyang City; WS=water source in Zhengzhou City; Q=Qin River; YL=Yiluo River; ARGs=antibiotic resistance genes; VFs=virulence factors; MLST=multilocus sequence typing; SNP=single nucleotide polymorphism.

Genome analysis identified 47 distinct STs among the 185 CTX-M-14-producing E. coli isolates examined (23 isolates from this study and 162 publicly available CTX-M-14-producing E. coli genomes), which were broadly distributed across human, animal, and environmental sources (Figure 3A and 3B). E. coli ST6802 has been previously detected in livestock and poultry feces, wild animals, animal-derived food products, anaerobic digestion systems, wastewater, and river water; however, no human isolates have been reported to date. cgMLST and SNP-based phylogenetic analyses demonstrated that the ST6802 isolates from this study were closely related to those recovered from anaerobic digestion systems treating pig manure, exhibiting 0–35 SNP differences (Figure 3C and 3D). Notably, ST6802 isolates from the drinking water source displayed zero SNP differences from the E. coli strain EFF60 (SAMN22853575) isolated from an anaerobic digestion system in China. These findings indicate environmental dissemination of CTX-M-14-producing E. coli ST6802 from a One Health perspective.

Figure 2. 

Relative abundances of plasmids and IS26 carried by E. coli ST6802 in (A) the Yellow River and their correlations with bla_CTX-M-G9 during the (B) wet, (C) dry, and (D) flat seasons.

Figure 3. 

Genome-based phylogenetic analysis of CTX-M-14-producing E. coli and E. coli ST6802 from Yellow River water in this study, NCBI, and EnteroBase databases. (A) cgMLST-based minimum spanning tree of 190 E. coli isolates from diverse sources, comprising 185 CTX-M-14-producing E. coli and 5 E. coli ST6802 isolates. (B) Distribution of multilocus sequence types (MLSTs) among the 190 E. coli strains across different sources. (C) cgMLST-based minimum spanning tree of 39 ST6802 isolates from various sources. (D) SNP-based phylogenetic tree of 39 ST6802 isolates.

Understanding the occurrence and genetic characteristics of ESBL-producing E. coli in the Yellow River is essential for tracking the origin and dissemination of clinically relevant antimicrobial resistance in surface waters. This study investigated the abundance of the CTX-M group 9-type ESBL gene bla_CTX-M-G9 in the Yellow River across three seasons at nine sampling sites. We further characterized the genomic features of the dominant CTX-M-14-producing E. coli during the dry season and identified potential contamination sources through comparative genomic analysis. Our findings reveal critical insights into the environmental transmission pathways of ESBL-producing E. coli from a One Health perspective, demonstrating both the necessity of continuous ESBL surveillance in the Yellow River and the urgent need to prevent environmental dissemination of ESBL-producing E. coli from animal sources.

Despite significant progress in China over recent decades in regulating antibiotic use in livestock and in reducing environmental ARG dissemination (10-11), bla_CTX-M-G9 was detected in all water samples, with significantly a higher relative abundance during the dry season compared to the flat season. This seasonal pattern aligns with a previous study that reported peak ARG abundance in the Yellow River during winter (12). Three factors likely contribute to this observation. First, reduced water volume during the dry season concentrates ARGs. Second, increased antibiotic prescriptions for seasonal epidemic diseases during winter months elevate ARG abundance in aquatic environments through wastewater discharge (13). Third, low winter temperatures may slow microbial community turnover, favoring the survival and proliferation of specific ESBL-producing E. coli clones. Consistent with a previous report (14), Puyang City exhibited the highest bla_CTX-M-G9 abundance during both dry and flat seasons, likely reflecting its intensive livestock and poultry farming operations.

The number of cefotaxime-resistant E. coli isolates varied substantially across sampling sites. The water source (n=17) and the Puyang section (n=15) yielded the highest numbers of isolates, consistent with the elevated spatial abundance of the bla_CTX-M-G9 gene observed during the dry season (Figure 1A). These findings reflect the spatial distribution patterns of cefotaxime-resistant E. coli across the study region. Notably, only a single E. coli strain was isolated from the Jiaozuo section, likely due to either the selectivity constraints of the culture medium employed or the presence of alternative bacterial hosts harboring bla_CTX-M-G9.

The predominant genotype among cefotaxime-resistant E. coli from the Yellow River and its water source was bla_CTX-M. The emergence and dissemination of specific E. coli clones represent a key mechanism driving the spread of bla_CTX-M genes. For instance, Falgenhauer et al. (15) documented the emergence of an E. coli ST949 clone harboring chromosomal bla_CTX-M-15 in German surface waters. In contrast, bla_CTX-M-14 typically resides on diverse plasmids, and CTX-M-14-producing E. coli in Chinese surface waters have demonstrated considerable genetic diversity, with no single clonal group predominating (16). This diversity is evident in Figure 3A, where the 167 CTX-M-14-producing E. coli strains retrieved from public databases belonged to 47 distinct ST types, underscoring the broad host range of bla_CTX-M-14. However, our study revealed a notable exception: the E. coli ST6802 clone carrying bla_CTX-M-14 was detected at all sampling sites across multiple cities, suggesting widespread clonal dissemination.

The ST6802 isolates exhibited fewer than four SNP differences from one another, with highly similar ARG profiles, plasmid replicons, and VF patterns, confirming clonal dissemination. To validate this finding, we quantified the IncHI2, IncHI2A, IncX1, and IncFIB plasmids within ST6802, along with the IS26 element, which typically flanks bla_CTX-M-14, in Yellow River samples. As anticipated, bla_CTX-M-G9 demonstrated strong correlations with these plasmids during the dry season, most notably with the IncHI2 plasmid (R²=0.86). This correlation likely reflects the location of bla_CTX-M-14 on the IncHI2 plasmid in E. coli ST6802, as we previously reported (6). Remarkably, the ST6802 isolates harbored 19 distinct VF types, predominantly associated with adhesion and biofilm formation (fimH, csgA, fdeC, and yehA-D), iron acquisition (chuA and sitA), toxin production (hlyE), serum resistance (iss and kpsE), virulence gene expression (eilA), and stress response (gad, nlpI, and terC). The eilA gene has been identified in enteroaggregative E. coli strains that cause both acute and chronic diarrhea (17). The presence of VFs mediating bacterial adhesion and biofilm formation likely enhances the survival and environmental persistence of ST6802 isolates in the Yellow River ecosystem.

E. coli ST6802 has been documented in diverse sources globally, including chicken feces in China, a silver gull in Australia, wastewater in the United States, chicken and poultry products in Canada, and anaerobic digestion systems treating pig manure in China (Supplementary Table S3). Both cgMLST and SNP-based phylogenetic analyses demonstrated that the ST6802 isolates from this study were genetically identical to E. coli strain EFF60 (SAMN22853575) from an anaerobic digestion system and closely related to isolates from livestock and poultry feces. In contrast, these isolates showed substantial phylogenetic divergence from ST6802 strains recovered from urban wastewater. These findings strongly support an animal origin for E. coli ST6802 and its subsequent environmental dissemination. The zero-SNP difference between ST6802 isolates from river water and anaerobic digestion systems is particularly striking, as it underscores the public health risk posed by transmission of animal-origin-resistant bacteria into aquatic environments. This transmission pathway — from animal feces through aquatic environments to potential human exposure — identifies critical intervention points for preventing cross-environmental spread of antibiotic resistance.

Although ST6802 has not been detected in human clinical samples, the IncHI2 plasmid harboring bla_CTX-M-14 within E. coli ST6802 demonstrates broad host range, exceptional conjugative transferability, and remarkable genetic stability (6). These characteristics position ST6802 isolates as critical vectors for bla_CTX-M-14 dissemination across bacterial populations. Moreover, the co-occurrence of multiple virulence factors in E. coli ST6802 amplifies concerns regarding its environmental persistence and spread throughout the Yellow River system, presenting substantial public health risks. The potential for this multidrug-resistant, virulent clone to establish itself in drinking water sources and irrigation systems warrants immediate attention and enhanced monitoring strategies.

This study advances beyond previous investigations of ESBL-producing E. coli in animal manure waste treatment systems (6) by tracing the transmission pathways of animal-origin ST6802 isolates into aquatic environments through integrated whole-genome sequencing and phylogenomic analysis of 167 publicly available E. coli genomes from diverse animal, human, and environmental sources. These findings underscore the critical need for sustained surveillance programs and stringent source control measures to interrupt the transmission of ESBL-producing E. coli from agricultural settings to natural water systems. Furthermore, comprehensive public education initiatives regarding antibiotic resistance are essential to mobilize broader societal engagement and support for antimicrobial stewardship efforts.