Escherichia coli ST2797 Is Abundant in Wastewater and Might Be a Novel Emerging Extended-Spectrum Beta-Lactamase E. coli

Abstract

The increasing prevalence of antibiotic-resistant bacteria is an emerging threat to global health. The analysis of antibiotic-resistant enterobacteria in wastewater can indicate the prevalence and spread of certain clonal groups of multiresistant bacteria. In a previous study of Escherichia coli that were isolated from a pump station in Norway over 15 months, we found a recurring E. coli clone that was resistant to trimethoprim, ampicillin, and tetracycline in 201 of 3,123 analyzed isolates (6.1%). 11 representative isolates were subjected to whole-genome sequencing and were found to belong to the MLST ST2797 E. coli clone with plasmids carrying resistance genes, including bla_TEM-1B, sul2, dfrA7, and tetB. A phenotypic comparison of the ST2797 isolates with the uropathogenic ST131 and ST648 that were repeatedly identified in the same wastewater samples revealed that the ST2797 isolates exhibited a comparable capacity for temporal survival in wastewater, greater biofilm formation, and similar potential for the colonization of mammalian epithelial cells. ST2797 has been isolated from humans and has been found to carry extended spectrum β-lactamase (ESBL) genes in other studies, suggesting that this clonal type is an emerging ESBL E. coli. Collectively, these findings show that ST2797 was more ubiquitous in the studied wastewater than were the infamous ST131 and ST648 and that ST2797 may have similar abilities to survive in the environment and cause infections in humans. IMPORTANCE The incidence of drug-resistant bacteria found in the environment is increasing together with the levels of antibiotic-resistant bacteria that cause infections. The COVID-19 pandemic has shed new light on the importance of monitoring emerging threats and finding early warning systems. Therefore, to mitigate the antimicrobial resistance burden, the monitoring and early identification of antibiotic-resistant bacteria in hot spots, such as wastewater treatment plants, are required to combat the occurrence and spread of antibiotic-resistant bacteria. Here, we applied a PhenePlate system as a phenotypic screening method for genomic surveillance and discovered a dominant and persistent E. coli clone ST2797 with a multidrug resistance pattern and equivalent phenotypic characteristics to those of the major pandemic lineages, namely, ST131 and ST648, which frequently carry ESBL genes. This study highlights the continuous surveillance and report of multidrug resistant bacteria with the potential to spread in One Health settings.

Keywords: Escherichia coli; PhenePlate; ST2797; bacterial survival; biochemical fingerprinting; biofilm; genomics; multidrug resistant; multidrug resistant bacteria; persistent clones; urban wastewater.

FIG 1

Comparison of the PhP-AREB profiles of the dominant clones found in urban wastewater. (A) Dendrogram inferred from the unweighted pair-group sequential clustering method using arithmetic averages (UPGMA) of the bacterial fingerprint of representative, whole-genome sequenced isolates of ST131 (PhP₁), ST648 (PhP₂), and ST2797 (PhP₃). The absorbance profiles of sugar fermentation by bacterial isolates and the dendrogram were generated using the PhenePlate software package (PhPlate AB). The absorbance values ranged from positive or 0 (bright yellow) to negative or 25 (dark blue). The AREB plates were read as 0 (susceptible, growth ≤ 10% of the control well), 1 (intermediate, requiring visual inspection, growth = 10 to 25% of the control well), and 2 (resistant, growth ≥ 25% of the control well). The pairwise similarities among the isolates that were obtained from the data generated via the PhP method were expressed as correlation coefficients. LACTS, lactose; RHMNS, rhamnose; DXRIB, deoxyribose; SUCRS, sucrose; SORBS, sorbose; TAGTS, tagatose; d-ARL, d-arabitol; RAFF, raffinose; GL-LAC, gal-lactone; ORNTN, ornithine. The final concentrations (in mg/L) of antibiotics were: ampicillin (32), cefotaxime (2), chloramphenicol (32), ciprofloxacin (4), gentamicin (16), nalidixic acid (32), cefpodoxime (3), tetracycline (16) and trimethoprim (16). The PhP-AREB images of each plate were produced using a desktop scanner (HP G4050). (B) The SNP-based maximum likelihood dendrogram from 19 ST2797 isolates, of which 11 were obtained in this study and 8 were downloaded from EnteroBase (8) from multiple sources. The core genome alignment was generated by using the Panaroo pipeline (24) in the strict mode, a core threshold of 95%, and a maximum length difference between genes in a cluster of 75%, and it was composed of 4,054 genes. The tree was calculated using PhyML (25) with the standard parameters and SH-like branch supports. (C) The top five plasmid genomes with the highest percentage of identity (>70%) and query coverage were included in the comparison using BRIG36 (v0.95, http://brig.sourceforge.net/). Contigs that were longer than 10 kb from the ST2797 genomes and contained at least one antibiotic resistance gene were used as the reference sequences (outer rings). The inner rings represent GC Skew− (purple), GC Skew+ (green), E. coli strain EC 528 plasmid pME5 (MT868879.1), S. enterica subsp. enterica serovar Typhimurium strain B71 plasmid pB71 (KP899806.1), S. enterica subsp. enterica serovar Typhimurium strain F8475 plasmid pF8475 (KP899804.1), and S. enterica strain SRC27 plasmid pSRC27-H (CP058810.1). The outermost ring indicates the plasmid pB71-191kp with annotated genes.

FIG 2

Bacterial survival, biofilm formation, and bacterial adherence to epithelial cells. (A) The survival ratio (%) and (B) the bacterial count (CFU/mL) of representative ST131 (n = 3), ST68 (n = 3), and ST2797 (n = 3 during incubation at 20°C in lake water microcosms [LWM]). Bacterial growth (>100%) and cell death (<50%) were colored yellow and red, respectively. The statistical analysis of panel B was performed using a repeated measures (RM) two-way analysis of variance (ANOVA) with Tukey’s multiple-comparison test. Comparisons that produced significant changes were coded as “a”, “b”, and “c” in panel B, and they correspond to ST2797 versus ST648, ST2797 versus ST131, and ST648 versus ST131, respectively. The error bars show the standard error of the mean of 3 independent experiments. (C) The quantification and comparison of surface-adhered biofilms in ST131 (red), ST648 (green), and ST2797 (black) isolates after 1, 2, 3, 7, 14, and 28 days of incubation in LWM at 20°C. Six independent CV experiments were performed, and the average values are plotted along with the standard deviations (error bars). The P values were determined via a one-way analysis of variance (ANOVA) with Tukey’s multiple-comparison test. (D) The quantification and comparison of the surface-adhered biofilms in ST131 (n = 3), ST648 (n = 3), and ST2797 (n = 11) after 48 h of incubation in rich medium (LB) at 20°C or 37°C. (E) The biofilm production of each isolate was measured to characterize the within-strain variability in the production of biofilm in response to temperature. Three independent CV experiments were performed, and the average values are plotted along with the standard deviations (error bars). The P values (*, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001) were determined via t tests. (F) Adherence assays were performed for 5 h in DMEM with 10% FBS. The numbers of bacterial cells (CFU/mL) adhered to HT-29/Caco-2 cell monolayers for the ST131, ST648, and ST2797 strains. The assays were performed in triplicate, and the graphs show the mean and standard deviation (error bars). The statistical analysis was performed using a one-way analysis of variance (ANOVA), but no significant differences were found.