Biocide-Resistant Escherichia coli ST540 Co-Harboring ESBL, dfrA14 Confers QnrS-Dependent Plasmid-Mediated Quinolone Resistance

Abstract

Emerging sequence types of pathogenic bacteria have a dual ability to acquire resistance islands/determinants, and remain renitent towards disinfection practices; therefore, they are considered “critical risk factors” that contribute significantly to the global problem of antimicrobial resistance. Multidrug-resistant Escherichia coli was isolated, its genome sequenced, and its susceptibilities characterized, in order to understand the genetic basis of its antimicrobial resistance.The draft genome sequencing of E. coli ECU32, was performed with Illumina NextSeq 500, and annotated using a RAST server. The antibiotic resistome, genomic island, insertion sequences, and prophages were analyzed using bioinformatics tools. Subsequently, analyses including antibiotic susceptibility testing, E-test, bacterial growth, survival, and efflux inhibition assays were performed.The draft genome of E. coli ECU32 was 4.7 Mb in size, the contigs were 107, and the G+C content was 50.8%. The genome comprised 4658 genes, 4543 CDS, 4384 coding genes, 115 RNA genes, 88 tRNAs, and 3 CRISPR arrays. The resistome characterization of ST540 E. coli ECU32 revealed the presence of ESBL, APH(6)-Id, APH(3′)-IIa, dfrA14, and QnrS1, with broad-spectrum multidrug and biocide resistance. Comparative genome sequence analysis revealed the presence of transporter and several virulence genes. Efflux activity and growth inhibition assays, which were performed with efflux substrates in the presence of inhibitor PAβN, exhibited significant reduced growth relative to its control.This study discusses the genotypic and phenotypic characterization of the biocide-tolerant multidrug-resistant E. coli O9:H30 strain, highlighting the contributory role of qnrS-dependent plasmid-mediated quinolone resistance, in addition to innate enzymatic modes of multidrug resistance mechanisms.

1. Introduction

Among the different members of genus Escherichia, serotypes from bacterial species E. coli that belong to Enterobacteriaceae are highly pathogenic for humans, birds, and animals. This Gram-negative bacterium is a rod-shaped, facultative anaerobe found in soil, food, environment, and the intestines of animals and humans, with an unmatched capacity to survive in diverse, stressful conditions. E. coli produces an arsenal of virulence factors, such as fimbrial adhesins, different iron acquisition systems, heat-labile/stable toxins and hemolysin, capsules, a type III secretion system, and colonization invasion factors. The toxin-producing E. coli strains are responsible for causing mild sicknesses, such as diarrhea and vomiting, to severe illnesses such as meningitis, respiratory diseases, pneumonia, and urinary tract infections [1,2]. As per the 2017–2018 GLASS report, E. coli was the most frequently reported pathogen with high-level resistance to ciprofloxacin and imipenem. Subsequently, in its 2019–2020 report, focusing on AMR data from 65 different countries, E. coli isolates tested from urine samples (from 64% countries) were resistant to ceftriaxone, cefotaxime, ampicillin, and ciprofloxacin, including carbapenems. In 2017, the WHO published its priority list of clinically significant human pathogens, and MDR E. coli was classified under the critical threat category, in order to ensure enhanced research activities to control AMR spread, and infection control [3,4].

The major pathotypes of E. coli include intestinal (IPEC) and extraintestinal (ExPEC) pathogenic strains. Of the ExPEC E. coli, UPEC (uropathogenic E. coli) evades the host’s innate immunity and colonizes the urinary bladder, kidneys, and is well recognized as a causative agent for urinary tract infections (UTIs), which account for ~80% of urinary infections [5,6].

Recently, India published its first pathogen priority list with a similar frame of objectives, carbapenem and tigecycline resistant E. coli topped the critical threat category. According to recent ICMR-AMR data, out of the total 107,387 isolates studied during the year 2019, the relative distribution of E. coli remained the highest at 28% (n =30822) in Indian hospitals from different locations. E. coli was the most predominant isolate from urine (56%), with strains exhibiting reduced susceptibility to cephalosporins and fluoroquinolones, including imipenem (from 86% in 2016 to 63% in 2019). While ST648, ST2659, and ST540 have been reported to be NDM-5-producing E. coli globally, the frequently reported ExPEC sequence types that are linked with UTI cases include ST131, ST73, ST95, and ST69 [7].

Primary susceptibility testing of E. coli urinary isolate ECU32 illustrated its antimicrobial resistance behavior, with an innate capability to co-produce multiple resistance determinants. The phylogenetic analysis classified this isolate to the ST540 serotype, one which is usually reported from birds and animals.Therefore, to reach a deeper perspective on its genetic content, genome organization, and phenotypic behaviors, asystematic study was initiated.This study reports the draft genome sequence, analysis, and resistome characterization of ST540 E. coli urinary isolate ECU32 from India, co-producing ESBL, APH(6)-Id, APH(3′)-IIa, dfrA14, and QnrS1, with broad multidrug and biocide resistance.

2. Results and Discussion

2.1. Genomic Features and Phylogenetic Analysis of E. coli

The draft genome reads of E. coli ECU32 were assembled to a single chromosome of size 4.7 Mb, an N50 spanning 94, 946 bp, L50 being 15, number of contigs being 107, and a G+C content of 50.8% (Assembly: GCA_002872235.1; GenBank: LZGD01000000). Analysis revealed 4658 genes (total), 4543 CDS (total), 4384 coding genes, 115 RNA genes, 8, 3, 2, 5S, 16S, 23S rRNA genes, 88 tRNAs, and 159 pseudogenes (

Table 1. Genomic features of E. coli ECU32.

Type	Assembly Statistics
Genome	Escherichia coli ECU32
Size	4,734,193
GC content	50.8
Number of coding sequences	4601
Number of RNAs	105
Number of subsystems	595
Contigs generated	107
Maximum contig length	239,491
Minimum contig length	502
Average contig length	44,244.8 ± 55,998.5
Median contig length	8912
Total contigs length	4,734,193
Total number of non-ATGC characters	663
Percentage of non-ATGC characters	0.014
Contigs ≥ 500 bp	107
Contigs ≥ 1 kbp	97
Contigs ≥ 10 kbp	67
Contigs ≥ 1 Mbp	0
N50 value	95,844
L50	15
Genome coverage	93.0X

). Furthermore, PADLOC analysis revealed three CRISPR arrays, including cas_type_I-E array system (

Table 2. Probable CRISPR systems in E.coli ECU32 using PADLOC analysis.

CRISPR System	Protein	Target	Sequence Id	Start	End	Strand
CRISPR_array	CRISPR_array	CRISPR001	LZGD01000013.1	92,385	92,901	-
cas_type_I-E	Cas2e	A8A11_20845	LZGD01000013.1	93,006	93,291	-
cas_type_I-E	Cas1e	A8A11_20850	LZGD01000013.1	93,292	94,210	-
cas_type_I-E	Cas6e	A8A11_20855	LZGD01000013.1	94,225	94,825	-
cas_type_I-E	Cas5e	A8A11_20860	LZGD01000013.1	94,811	95,486	-
cas_type_I-E	Cas7e	A8A11_20865	LZGD01000013.1	95,488	96,580	-
cas_type_I-E	Cas11e	A8A11_20870	LZGD01000013.1	96,592	97,075	-
cas_type_I-E	Cas8e	A8A11_20875	LZGD01000013.1	97,067	98,576	-
CRISPR_array	CRISPR_array	CRISPR002	LZGD01000043.1	123,708	123,910	-
retron_I-C	RT-Toprim_I-C	A8A11_16640	LZGD01000043.1	170,971	172,729	-
retron_I-C	msr-msd	NA	LZGD01000043.1	172,765	172,895	-
RM_type_II	MTase_II	A8A11_21460	LZGD01000051.1	19,587	21,006	+
RM_type_II	REase_II	A8A11_21465	LZGD01000051.1	20,986	21,457	+
DMS_other	BrxD	A8A11_09690	LZGD01000074.1	2659	3976	+
DMS_other	BrxHI	A8A11_09695	LZGD01000074.1	3972	6168	+
CRISPR_array	CRISPR_array	CRISPR003	LZGD01000087.1	154,485	155,734	+

) [8].

The genomic features of selected ST540 E. coli strains were investigated with Indian E. coli strains from different available sources in the BacWGSTdb server, while phylogenetic relationshipswere analyzed using the cgMLST approach with the ST540 reference strain. The analysis revealed two major clusters based on the isolation source; the ECU32 strain was found in the second cluster, along with strains predominantly isolated from humans (Figure 1A). Further analysis with only E. coli ST540 Indian isolate revealed its clustering with strains of human gut origin (Figure 1B). Overall, the cgMLST analysis revealed that the E. coli ECU32 strain was closely related to the human and gut origin multidrug isolates.

Additionally, in the RAST server, functional annotation of the genome revealed the presence of 595 subsystems (39% unassigned). The gene ontology data showed the distribution of genes for different components and functions. The biological processes includedgenes involved in transcription 5.37%, regulation of transcription 2.62%, transmembrane transport 1.93%, carbohydrate metabolic process 1.77%, cell adhesion 0.85%, cell wall organization 0.83%, and integral component of membrane 20.25%. The cellular components included cytoplasm with 10.22%, and the molecular function included genes with oxidoreductase activity 1.93%, zinc ion binding 2.27%, magnesium ion binding 2.36%, plasma membrane 6.43%, DNA binding 7.99%, ATP binding 8.49%, and transporter activity 2.53% (Figure 1C,D).

The clinical strain belonged to ST540 as the alleles adk_6, fumC_7, gyrB_57, icd_1, mdh_8, purA_8, and recA_2 exhibited 100% identity to their respective locus, as per MLST typing. Based on Clermon typing, the strain belonged to phylogroup A, a group largely dominated with commensal origins of strains. The E. coli ECU32 belongs to serotype H30-H with 100% identity for fliC in scaffold8|size160413, fimH54 (scaffold11|size140663), and O9-O antigen (>99% identity to wzm and wzt in scaffold34|size42957). These gene clusters were found between hisI and gnd genes, and belong to group 1 K antigens.

2.2. Antimicrobial-Resistant Genes

Upon analyzing the sequence of E. coli strain ECU32 for genetic clusters, a 9852-base-pair genomic island was identified that carried the CRISPR-associated proteins, and the 5540-base-pair genomic island (scaffold94xsize3049) showed the presence of ESBL, the class-A blaTEM beta-lactamase (A8A11_02620).

The 32,081-base-pair genomic island in the strain (scaffold7xsize174057) highlighted the presence of members from the EscJ/YscJ/HrcJ family type III secretion system (from A8A11_10065 to A8A11_10125).

The largest genomic island present in strain ECU32, 131,243bp in size (consisting of scaffold78xsize4056 and scaffold83xsize2691), indicated the presence of aminoglycoside-modifying enzymes APH(3′) (A8A11_04060), APH(6)-Id (A8A11_04065),and class 1 integrase, followed by gene cassette dfrA1 (A8A11_21925). The class-C ampC detected in this study exhibited 97% identity to a homolog found in E. coli O157:H7; 95% to Shigella sonnei; 78% to Enterobacteriaceae bacterium; 73% to Shigella flexneri; 71% to Enterobacter, and 70% to Yersinia ruckeri.

The quinolone resistance gene qnrS1 was found in scaffold57xsize16952 (A8A11_12295) with an adjacent ISKra4-like element (A8A11_12305), while the sulfonamide-resistant dihydropteroate synthase sul2 was found in scaffold78xsize4056 (A8A11_04055). The quinolone resistance determinant exhibited identities to homologs found in Vibrio mytili CAIM 528 (95%); Photobacterium ganghwense strain DSM 22954 (92%); Vibrionales bacterium SWAT-3 (83%); and Photobacterium halotolerans strain MELD1 (66%). The sequence analysis of QnrS1 indicated the conserved pentapeptide sequences, and homology modeling highlighted the differences in loops A and B regions, as shown in Figure S1.

Homologs of well characterized efflux pumps, AcrAB, EmrAB, MdtABC, EmrE, AcrD, and TetA (scaffold82xsize4187; A8A11_05610), were present in the genome; moreover, the marR gene harbored sense mutations at the Y137H and G103S residues. The tetracycline efflux pump exhibited varied identities to homologs found in Aeromonas simiae CIP 107798 (98%), Pseudomonas fluorescens HK44 (97%), E. coli O83:H1 str. NRG 857C (97%), and Clostridium nexile DSM 1787 (78%).

Using the BacWGSTdb server, the resistance genes were assessed amongst ST540 strains from different countries, hosts, and isolation sources, and the ECU32 strain exhibited the presence of diverse resistance genes (Figure 2A).

2.3. Virulence and Its Associated Genes

Comparative genome sequence analysis revealed that E. coli strain ECU32, in scaffold21xsize74584, carried ORFs for both the ferric enterobactin ABC transporter system (fepA, fepB, fepC, fepD, fepG, fes), and enterobactin (siderophore) biosynthesis exporter system (entB, entC, entD, entE, entF, entS). Important virulent factors that were identified included the type 1 fimbriae adhesin fimH (scaffold11xsize140663-1, A8A11_14300), hemolysin E hlyE (scaffold22xsize79984, A8A11_04330), invasin of brain endothelial cells, ibeB (scaffold60xsize15487, A8A11_08140), gspC, gspD, gspE, (scaffold35xsize41361), and intimin-like adhesion fdeC (scaffold1xsize239491 with 93% to EC958_0448).

Virulence genes were analyzed and compared with different ST540 strains and pathotypes. The analysis demonstrated the prevalence of virulence genes that are involved in adherence, autotransporter, invasion, non-LEE-encoded TTSS effectors, secretion system, toxin, and others (Figure 2B,C).

2.4. Phenotypic Characterization of Multidrug-Resistant E. coli

The Kirby Bauer assay revealed the E. coli strain ECU32 to be multidrug resistant (AMP, AMK, CLI, CST, ERY, KAN, LZD, MET, OXA, PEN, RIF, STR). The minimum inhibitory concentrations of this Indian isolate for different antibiotics, as evaluated by the agar dilution method, were ceftazidime 0.5 µg/mL, ticarcillin >1024 µg/mL, neomycin 4 µg/mL, norfloxacin 16 µg/mL, doxycycline 64 µg/mL, tetracycline 64 µg/mL, nalidixic acid >1024 µg/mL, ampicillin >1024 µg/mL, kanamycin 16 µg/mL, erythromycin >1024 µg/mL, trimethoprim 64 µg/mL, and chloramphenicol 4 µg/mL.

E. coli, a common flora in the human gastrointestinal tract, has adaptive survival response strategies under different stress conditions. The growth profile of E. coli ECU32 was assessed in the presence of low to high pH concentrations. The strain retained the ability to grow in pH 5.0 to pH 8.0; meanwhile, in pH 10, the strain displayed less (~0.5-fold) growth compared to pH 5.0 (Figure 3A). Concentration-dependent growth of the strain in the presence of the antibiotic norfloxacin was observed. The strain was able to grow in doses upto 64 µg/mL, and exhibited nine-fold lower growth compared to its control (Figure 3B).

However, the E. coli ECU32 strain was able to survive in the presence of both oxidative and nitrostative stress-inducing agents, such as hydrogen peroxide, sodium nitroprusside, and sodium nitrite, respectively (Figure 3C–E).

The survival of E. coli ECU32 under osmotic stress conditions was determined in the presence of different sodium chloride concentrations. The strain exhibited more than 80% survival upto 0.75 M, and 40% in 1 M of sodium chloride (Figure 3F). Altogether, these observations emphasize the adaptability of E. coli under any intracellular stress conditions, and its survival inside the host to cause disease severity.

The E. coli ECU32 strain has an ability to cope with different intracellular stress responses. Furthermore, the ability of the strains to survive in the presence of a various range of antimicrobial agents was examined. Analyses revealed a more than 50% survival rate was observed for the strain at 1024 µg/mL for ampicillin, 4 µg/mL for neomycin, and 0.5 µg/mL for tetracycline. Furthermore, the strain was found to have retained the ability to survive in various antimicrobial compounds, such as acriflavine, acridine orange at 256 µg/mL, and saffranine and deoxycholate at 1024 µg/mL (Figure 3G).

In order to detect efflux activity in the E. coli strain, MICs for the following efflux-based substrates were initially analyzed: acridine orange 256 µg/mL, acriflavine 64 µg/mL rhodamine >1024 µg/mL, saffranine >1024 µg/mL, deoxycholate >1024 µg/mL, and SDS >1024 µg/mL. Hence, a growth inhibition assay was performed with the E. coli strain ECU32, using ampicillin 256 µg/mL, in the presence of known efflux pump inhibitor PAβN, which exhibited an approximatelyfive-fold reduced growth relative to its control. Overall, the assays demonstrated that the strain has the ability to survive under different assails, including antimicrobial compounds. Furthermore, the sequence analysis and phenotypic efflux assays indicated the role of active efflux as the primary mechanism used to confer antimicrobial resistance. In order to substantiate these observations, genomic analysis of the strain further confirmed the presence of well-characterized efflux families, depicting the importance and possible involvement of these pumps in antimicrobial resistance and multiple cellular functions.

The strain was examined for its resistance level towards hospital-based disinfectants, and it exhibited tolerance as follows: benzalkonium chloride 12.8 µg/mL, chlorhexidine <3.2 µg/mL, and triclosan >0.1 µg/mL (Figure 3H(i,ii)), indicating that the urinary strain was broad-spectrum antimicrobial, as well as biocide-resistant. Additionally, the E. coli strain ECU32 displayed an ability to form biofilms (ratio 570/600 nm = 0.252), as it harbored genes that are required for adhesion and virulence.

Upon transforming the plasmid preparation from the MDR strain into E. coli JM109, transformants were confirmed on ampicillin (>256 µg/mL) and ciprofloxacin (0.5 µg/mL) plates; subsequent PCR detection indicated the presence of quinolone resistance determinant in the transformant, and further sequencing confirmed the role of qnrS-dependent plasmid-mediated resistance in E. coli strain ECU32.

It is worthwhile to state here that ST540 has usually been reported in E. coli that was isolated from birds and chickens. Identifying the serotype ST540 from a biological sample strictly emphasizes the periodic monitoring of emerging E. coli strains at the One-Health Interface. The E. coli strains TOP2386 (accession no: AORB01), TOP2396_1 (accession no: AOQQ01), TOP2515 (accession no: AOQT01), TOP2522_1 (accession no: AOQU01), and TUM3433 (accession no: BGLY01), isolated from human samples in USA and Japan, belong to the same ST, as that of E. coli strain ECU32. As per the BacWGSTdb server, E. coli strain ECU32 (accession no: LZGD01) with ST540 is being reported for the first time in India.

Figure 3. Biological characterization of the E. coli ECU32 strain. The growth profile of the E. coli ECU32 strain was monitored in the presence of different pHs (A) and concentrations of norfloxacin (B). The growth of the ECU32 strain was monitored in the presence of varied concentrations of oxidative stress-inducing agent—hydrogen peroxide (C) and nitrostative stress-inducing agents—sodium nitrite and sodium nitroprusside (D,E). The osmotic stress response survival was determined in the presence of sodium chloride concentrations (F). The survival ability of the E. coli ECU32 strain was determined under different concentrations of antibiotics [ampillicin, neomycin, tetracycline], bile [deoxycholate], structurally unrelated compounds and dyes [acriflavine, acridine orange, saffranine] (G), and biocides [benzalkonium chloride, chlorhexidine, and triclosan] (H(i,ii)). The mean values of the different independent experiments were used for plotting graphs, using GraphPad Prism.

Figure 3. Biological characterization of the E. coli ECU32 strain. The growth profile of the E. coli ECU32 strain was monitored in the presence of different pHs (A) and concentrations of norfloxacin (B). The growth of the ECU32 strain was monitored in the presence of varied concentrations of oxidative stress-inducing agent—hydrogen peroxide (C) and nitrostative stress-inducing agents—sodium nitrite and sodium nitroprusside (D,E). The osmotic stress response survival was determined in the presence of sodium chloride concentrations (F). The survival ability of the E. coli ECU32 strain was determined under different concentrations of antibiotics [ampillicin, neomycin, tetracycline], bile [deoxycholate], structurally unrelated compounds and dyes [acriflavine, acridine orange, saffranine] (G), and biocides [benzalkonium chloride, chlorhexidine, and triclosan] (H(i,ii)). The mean values of the different independent experiments were used for plotting graphs, using GraphPad Prism.

3. Materials and Methods

3.1. The Genome Draft Sequence, Annotation and Analysis

The E. coli strain, ECU32, selected for this analysis, was obtained in 2013 from a urine sample during our longitudinal study. Draft genome sequencing was carried out with paired-end sequencing technology, Illumina NextSeq 500, and scaffold annotation was performed using a RAST server [9]. Basic local alignment search tool, (BLAST) (GO), was utilized for determining percent identities [10]. The E. coli ECU32 genomesequence was compared with available E. coli genomes (n =300) that were isolated from India, and core genome multilocus sequence typing (cgMLST) was performed with the selected reference genome (E. coli_ AZ147 GenBank CP018995 ST540) for phylogenetic analysis. The graph tree and minimum spanning tree were constructed and analyzed, based on cgMLST profiles of Indian E. coli isolates. The antibiotic resistome, genomic island, insertion sequences, and prophages were analyzed using bioinformatic tools such as Pathosystems Resource Integration Center tool, Resistance Gene Identifier, Island Viewer 4, Mobile Element Finder, PHAge Search Tool Enhanced Release, and NCBI-BLAST [9,10,11]. The CRISPR was determined using PADLOC server (https://padloc.otago.ac.nz/ accessed on 26 September 2022). Comprehensive Antibiotic Resistance Database (CARD; https://card.mcmaster.ca/analyze/rgi accessed on 10 May 2021) was used to predict AMR genes and the VR2 profile, in order to identify the associated mobile elements [8,12,13]. Virulence, transposons, and antibiotic resistance genes were determined using the BacWGSTdb server. The distribution of virulence factors in E. coli ECU32 with other E.coli strains was deciphered using virulence factor database analyzer [14].

3.2. Antibiotic Resistance Pattern and Growth Analysis

The antibiotic susceptibility testing, E-test, survival assays, stress response, growth curves, and efflux inhibition assays were performed, as described previously [9,11].

3.3. Data Deposition

The E. coli ECU32 whole genome sequence was submitted to GenBank NCBI, with accession number LZGD00000000.1.

4. Conclusions

Overall, this study highlighted the presence of diverse virulence factors and the antibiotic resistome in ST540 E. coli strain ECU32, emphasizing a pressing need to conduct large-scale epidemiological surveillance to monitor the dominant ST/strain that is circulating within the cohort, and also to track the changing patterns in antibiotic susceptibility and resistome amongst swiftly disseminating bacteria such as E. coli in India. Emerging STs remain renitent towards hospital sterilization protocols, and havean unprecedented ability to acquire additional resistance determinants; therefore, they are “critical risk factors” that contribute significantly to the global problem of antimicrobial resistance.