Introduction

Extended-spectrum β-lactamases (ESBLs) in Enterobacteriaceae are categorized by World Health Organization as the most critical cause of antimicrobial resistance (AMR) requiring discovery of new antibiotics1. In addition, most ESBL-producing Enterobacteriaceae are also multidrug-resistant (MDR), burdening the treatment2. Extraintestinal pathogenic Escherichia coli (ExPEC) is a major ESBL-producing organism which in addition to intestine, infects, urinary tract, bloodstream, meningitis, and wounds and causes sepsis. ESBL associated AMR in ExPEC is not only disseminated in health care settings but also in community-acquired infections3. The global increase of ESBL-ExPEC strains is causing clinical and economic losses similar in magnitude to that caused by pathogenic E. coli. Unlike for, intestinal pathogenic E. coli or commensal E. coli, defining the origin or primary reservoir of ExPEC is the major challenge in its treatment4. Moreover, the influence of AMR and virulence factors (VF) genes on ExPEC pathogenicity has become a serious global concern. Thus, researchers majorly rely on ExPEC genotyping to explore the association between AMR genes, VF, and their phylogenetic distribution.

The distribution of ESBL-producing E. coli (ESBL-E coli) in extraintestinal infections is diverse and varies among different geographical regions. The clonal spread of E. coli ST131 (associated with ExPEC infections, particularly urinary tract and blood stream infections) contributed to globally disseminated MDR clone5. Among the ESBL genes, blaCTX-M-15 is highly prevalent, followed by CTX-M, TEM, SHV, PER, VEB, GES, BES, TLA, and OXA genes6. The commensal E. coli in healthy cattle, pigs, and chickens serve as a reservoir of AMR genes7. The prevalence of CTX-M genes is higher in uropathogenic E. coli (UPEC) than in the commensal isolates from healthy volunteers8. Moreover, ESBLs producing Enterobacterales (ESBL-Enterobacterales) can colonize long-term (> 12 months) as intestinal microbiota9 enhancing the spread of ESBL AMR in a given health system, including humans, animals, and environments10. The widespread presence of ESBL genes in systemic infections also significantly impacts therapeutic and mortality outcomes. Clinical and Laboratory Standard Institute (CLSI) recommends phenotypic screening and confirmatory testing for ESBLs production as a part of regular clinical treatment of microbial infections11. However, the genotypic methods of ESBL screening are more advantageous for epidemiological management and help in overcoming challenges associated with phenotypic expression variance12.

According to Ambler’s molecular classification, ESBLs are serine β-lactamases comprising three significant class A genes (blaCTX-M, blaTEM, and blaSHV), and a class D blaOXA gene. Recently, various newly discovered isothermal nucleic acid amplification techniques, which do not require expensive thermal cycling machines, have been extensively applied for rapid and simple detection of AMR genes13. In 2021, a 10 min naked-eye rapid multiplex recombinase polymerase amplification (RPA) lateral flow assay was developed to detect the three common ESBLs genes (blaCTX-M, blaOXA, and blaSHV)14. However, the use of engineered recombinase enzyme from E. coli K12 or BL21 strains (carrying their own blaTEM gene) proved to be a major setback; the use of these RPA kits was plagued with cross reactivity and false positive results15. Helicase-dependent amplification technique, like RPA, utilizes a primer pair to amplify a specific DNA/RNA target at a set temperature. To unwind double-stranded DNA and generate a new amplicon, the RPA kit (TwistDx) uses recombinase and DNA polymerase with strand displacement activity (Sau polymerase), whereas the IsoAmp® II Universal tHDA kit (New England Biolabs) employs a thermostable helicase Tte-UvrD from Thermoanaerobacter tengcongenesis and Bacillus stearothermophilus DNA polymerase (Bst-DNA polymerase)16. The optimum amplification conditions for RPA and tHDA are 37–42 °C for 20–40 min and 60–65 °C for 90–120 min, respectively17. Both, RPA and tHDA, are portable nucleic acid detection assays ideally suited for use in point-of-care or resource-limited field settings13.

The ExPEC pathogenicity is influenced by several VF genes, with wide functions, ranging from bacterial colonization to virulence18. Moreover, most VF genes are encoded by plasmids, pathogenicity islands (PAI)18,19 or other mobile genetic elements18 which horizontally transfer to other pathogenic and non-pathogenic bacteria. E. coli is classified into eight phylogenetic groups (A, B1, B2, C, D, E, F, and Clade I) using the easier quadruplex PCR (which is less complex and more rapid than multilocus sequence typing or ribotyping methods)20. The association of ESBLs, VF genes, and phylogroups has been investigated in both ExPEC and commensal E. coli. Milenkov et al., found that blaCTX-M-15 was the most prevalent ESBL gene in E. coli specimens isolated from healthy pregnant Madagascar woman. In addition, 90% of these isolates belonged to phylogenetic groups A, B1, and C (which are associated with commensalism and carry a few VF genes involved in adhesion and iron acquisition). While, 10% of these isolates belonged to the extraintestinal virulent phylogenetic groups B2, D, and F21. Persistent long-term carriage (average 3.5 months) of ESBL-Enterobacterales (which belonged to extraintestinal virulence associated phylogroup B2/D/F) was observed in travelers visiting tropical areas, 3 or more months after their return. Whereas, commensal associated phylogroups A/B1/E persisted for shorter carriage durations (approx. 0.5 months)21. Phylogroup B2 of uropathogenic E. coli (UPEC) exhibited maximum AMR and carried six VFs22. ESBL-ExPECs carrying VF genes, isolated from bloodstream infections, showed the following order of phylogroup predominance: B2 (45.8%) > B1 (18.8%) > E (14.6%). The ESBL blaCTX-M-15 and VF traT genes were the most predominant23. E. coli, the most frequent pathogen, second only to group B streptococci, causing neonatal meningitis in early-onset infections, belonged to extraintestinal phylogroup B2; > 70% of this pathogenic E. coli strain carry kpsII, K1, neuC, iucC, sitA, and vat genes. In contrast, E. coli obtained from healthy individuals belonged to groups A and D; they carry < 27% of VF genes24,25.

We aimed to develop isothermal amplification assays to characterize ESBL genes (blaCTX-M, blaOXA, blaSHV, and blaTEM) in ESBL-ExPEC strains. The sensitivity and specificity of these assays were validated using nucleotide sequencing. These simple isothermal platforms can be established in low resources settings for monitoring ESBL in all clinical samples. Moreover, we characterized and phylogenetically analyzed VF genes (traT, ompT, iutA, fimH, iha, hlyA, papC, and cvaC) found in the ESBL-ExPEC clinical isolates. Identifying associations between ESBL genes, VF genes, and ESBL-ExPEC phylogroups is important for improved global AMR surveillance, targeted antibiotic treatment, and infection control.

Results

Identification of ESBLs, their variants, and antimicrobial susceptibility pattern

Of the 144 ExPEC isolated from various extraintestinal specimens including blood, urine, pus, body fluid, and sputum, 95 ESBL- and 49 non-ESBL-ExPEC strains were identified by combination disk method given in CLSI guideline11. All 95 ESBL-producing E. coli were subjected to genotyping. We found blaCTX-M genes (n = 95) in single or in combination with blaTEM (n = 40), blaOXA (n = 29), and blaSHV (n = 2) genes (Supplementary Table S1). In addition, 21 out of 49 phenotypically screened non-ESBL-producing E. coli harbored blaTEM-1 gene, while none of blaTEM, blaOXA, blaSHV, and blaTEM genes were observed in the rest. The PCR-sequencing results of CTX-M variants revealed the following order of prevalence: blaCTX-M-15 (50.5%) > blaCTX-M-55 (17.9%) > blaCTX-M-27 (16.8%) > blaCTX-M-14 (14.7%). The blaSHV was found only in 2 isolates. The antibiotic susceptibility pattern of most common ESBL–ExPEC variants (CTX-M-15, CTX-M-27, CTX-M-14, and CTX-M-55 types) showed 100% resistance to cefotaxime and cefdinir as described in Table 1. The following order (high–low) was observed in MDR toward ceftriaxone, cefepime, ciprofloxacin, ceftazidime, levofloxacin, and gentamycin: blaCTX-M-15 isolates > blaCTX-M-55 > blaCTX-M-27 isolates.

Table 1 The antimicrobial susceptibility pattern of blaCTX-M variants in 95 ESBLs producing E. coli.
Full size table

Development of multiplex RPA and tHDA assays

The optimum concentration of blaCTX-M, blaOXA, and blaSHV primers required in multiplex RPA reaction were 0.2 µM CTX-M, 0.1 µM OXA and 0.1 µM SHV (Fig. 1A), respectively. All products were amplified at 37–39 °C (Fig. 1B) at 25–30 min (Fig. 1C). The DNA template concentration ranging at 12.5–25 ng showed the apparent bands of all 3 amplicons (Fig. 1D). The blaSHV amplicon was absent at 41℃ and incubation duration < 25 min. Finally, the following multiplex RPA reaction condition was chosen as optimum: 37 °C for 25 min using 25 ng template DNA.

Figure 1
figure 1

Optimization of the RPA assay. Agarose gel electrophoresis shows the blaCTX-M, blaOXA, and blaSHV amplicons (A) when 0.2 µM CTX-M, 0.1 µM OXA, 0.1 µM SHV and 0.2 µM CTX-M, 0.05 µM OXA and 0.05 µM SHV were used; (B) at 37–41 °C; (C) at 10–30 min incubation; and (D) at different DNA template concentrations (1.5–25 ng) [M, 100 bp Marker; C+, positive DNA control; C−, no template control].

Full size image

The tHDA was employed instead of RPA to amplify 111 bp of blaTEM. The blaTEM amplicon was observed at 65–67 °C (Fig. 2A), with 30–90 min incubation (Fig. 2B). The optimum DNA template and primers concentrations were 50–100 ng (Fig. 2C) and 0.025–0.05 µM, respectively (Fig. 2D). The following tHDA condition was selected as optimum: 65 °C for 30 min using 50 ng template DNA and 0.025 µM primers.

Figure 2
figure 2

Optimization of the tHDA assay. Agarose gel electrophoresis showed the blaTEM amplicons (A) at 59–67 °C; (B) at 15–90 min; (C) at different DNA template concentrations (0.05–100 ng); and (D) at different primers concentrations (0.025–0.075 µM) [M, 100 bp Marker; C+, positive DNA control; C−, no template control].

Full size image

Limit of detection (LOD) and specificity of multiplex RPA and tHDA assays

The DNA templates were isolated from E. coli clinical strain ESBL120 carrying blaCTX-M, strain KP125 carrying blaOXA, strain EC137 carrying blaTEM, and K. pneumoniae ATCC 700,603 carrying blaSHV. The LOD of multiplex RPA assays for blaCTX-M, blaOXA, blaSHV genes (Fig. 3A–C) was 5 ng, 0.5 ng, and 0.5 ng, respectively. The sensitivity was 10–100 times higher than that of tHDA assay for blaTEM (LOD of 50 ng) (Fig. 3D).

Figure 3
figure 3

(A–C) LODs for blaCTX-M, blaOXA, and blaSHV detection by RPA assay; (D) LODs for blaTEM detection by tHDA; (E) Specificity of the RPA assay for blaCTX-M, blaOXA, and blaSHV genes; and (F) Specificity of the tHDA assay for blaTEM gene [M, 100 bp Marker; C+, positive DNA control; C−, no template control].

Full size image

Amplicons were absent (Fig. 3E and F) in RPA and tHDA assays of the DNA of the following organisms: E. coli ATCC 25,922, Proteus mirabilis ATCC 25,933, Pseudomonas aeruginosa ATCC 27,853, Acinetobacter baumannii ATCC 19,606, Staphylococcus aureus ATCC 25,923 and Enterococcus faecalis ATCC 29,212. A 214 bp amplicon was produced from DNA of K. pneumoniae ATCC 700,603, which naturally contains blaSHV.

Validation of multiplex RPA and tHDA assays

A total of 95 ESBL-ExPEC and 49 non-ESBL-ExPEC were subjected to multiplex RPA and tHDA assays. The results of blaCTX-M, blaOXA, and blaSHV genes from multiplex RPA assays were in concordance with PCR-sequencing results exhibiting 100% sensitivity (95% CI = 96.19–100%, 88.06–100%, 15.81–100%, respectively), 100% specificity, positive predictive value (PPV), and negative predictive value (NPV) (95% CI: 92.75–100%, 96.84–100%, and 97.44–100%, respectively) (Table 2). While tHDA assay showed eight false-negative samples of blaTEM gene yielding 86.89% (95% CI = 93.28–100), 100% (95% CI = 83.41–96.13%), 100% (95% CI = 28.93–45.42%), and 91.21% (95% CI = 77.37–92.78%) sensitivity, specificity, PPV, and NPV, respectively. The identification accuracy for blaCTX-M, blaOXA, and blaSHV genes were 100%, and blaTEM gene was 94.44%.

Table 2 Sensitivity, specificity, and predictive value of multiplex RPA and tHDA assays.
Full size table

Distribution of VF genes, phylogroups among ESBL variants

In the 95 ESBL-ExPEC clinical isolates, the VF genes exhibited the following order in frequency: traT (67.4%) > ompT (52.6%) > iutA (50.5%) > fimH (47.4%) > iha (33.7%) > hlyA (26.3%) > papC (12.6%) > cvaC (3.2%), respectively (Table 3). All VF genes were distributed among CTX-M variants, excepting papC and cvaC that were not found in the CTX-M-27 variant. The traT gene was found frequently in CTX-M-15 (75%), CTX-M-14 (64.3%), and CTX-M-55 (76.5%). While iha gene was predominate in CTX-M-27 (70.6%). Most common phylogroups among ESBL-ExPEC strains included: A (28.4%), B2 (28.4%), F (22.1%). The CTX-M-14, 15, and 55 (35.7%, 29.2%, and 47.1%) were predominant in phylogroup A, while most of CTX-M-27 belonged to phylogroup B2 (70.6%). Only CTX-M-15, and CTX-M-55 variants were found in the rare phylogroups B1, and E, respectively.

Table 3 Distribution of VF genes and phylogenetic groups in 95 CTX-M isolates.
Full size table

Relationships of VF genes, phylogenetic groups, and antibiotic resistance

ESBL-ExPEC and non-ESBL-ExPEC isolates predominantly carried traT and ompT, respectively (Table 4). Surprisingly, cvcC was absent in our non-ESBL isolates. The association of traT, ompT, and hlyA were significantly different (P < 0.05) between the ESBL-ExPEC and non-ESBL-ExPEC isolates. Phylogroup predominance was as follows: phylogroup B2 (35.4%) > A (23.6%) > F (19.4%). ESBL-ExPEC were predominated by phylogroups A and B2, while non-ESBL-ExPEC were predominated by phylogroup B2. Only phylogroup B2 was significantly different between ESBL and non-ESBL groups (P < 0.05). Phylogroup Clades I was absent in all clinical isolates. Analysis of variance by Friedman’s test revealed a significant difference in VF gene distribution (P = 0.000). Three VF genes (hlyA, iha, and ompT) were distributed differently across phylogroups (Table 5). Pairwise analysis of phylogroup showed that hlyA was associated with phylogroup A and iha was associated with phylogroups F, A, and B2 (P < 0.05). Whereas, ompT was associated with phylogroups B2 and F (P < 0.05).

Table 4 Association of VF genes and phylogroups between ESBLs and non-ESBL clinical isolates.
Full size table
Table 5 Relationship of VF genes and phylogroups in E. coli clinical isolates.
Full size table

Discussion

ExPEC is a major cause of morbidity and mortality in both hospital and community-acquired infections. Apart from epidemiological factors, acquisition of VFs and AMR, is likely to contribute to the global pandemic of ExPEC lineages4,26. Moreover, plasmid-mediated horizontal transfer of ESBL genes occur easily among species, causing widespread infection6. Accumulating evidence suggested that VFs helps gastrointestinal pathogens outcompete the commensal microbiota and impair host immunity via inducing colonization, resistance, and invasion27.

ESBL-Enterobacteriaceae, especially E. coli was one of the most frequent isolates in blood stream infection samples collected from a pediatric oncology center28. The MDR E. coli from febrile neutropenic cancer patients showed high resistance to ampicillin, cefepime, ceftriaxone, and cephradine29. Here, all ESBL-ExPEC isolated from cancer patients carried blaCTX-M, showing the following predominance: blaCTX-M1 group (68.4%; 50.5% blaCTX-M-15 and 17.9% blaCTX-M-55) > blaCTX-M9 group (31.5%; 16.8% blaCTX-M-27 and 14.7% blaCTX-M-14). MDR was observed in all CTX-M variants. They were resistant to ceftriaxone, cefepime, ciprofloxacin, ceftazidime, levofloxacin, and gentamycin. Similarly, ESBL-ExPECs (from tertiary hospitals in Thailand) predominantly carried blaCTX-M1 (71.23%) and blaCTX-M9 (38.95%)30. The global pathogenic E. coli ST 131 strain harbors blaCTX-M-15 (67.6%), blaCTX-M-27 (20.6%), and blaCTX-M-14 (11.8%)31. Globally, the blaCTX-M-15 is frequently reported ESBL gene, especially in the bloodstream and urinary tract infections23,32,33,34. The blaCTX-M-55 is present in most E. coli isolated from pork and fecal samples14,35.

Several assays for genotyping ESBL genes, exist: (1) Kanokudom et al.,14 developed a naked-eye rapid multiplex RPA assay for detecting blaCTX-M, blaOXA, and blaSHV in pork E. coli isolates. Amplicons were visualized on a single-stranded tag hybridization chromatographic printed-array strip (STH-PAS, a commercial lateral flow assay strip); (2) Higgins et al.,36 set up a portable loop-primer endonuclease cleavage-loop-mediated isothermal amplification (loop-primer endonuclease cleavage-LAMP) assay for detecting blaCTX-M-1 and blaCTX-M-15 in porcine fecal E. coli isolates; and (3) Wang et al. developed a probe-based real-time PCR assay for detecting blaCTX-M, blaTEM, and blaSHV in broiler chicken E. coli isolates37.

Here, we developed the following isothermal assays to detect common ESBL genes: (1) a multiplex RPA assay to detect blaCTX-M, blaOXA, and blaSHV genes; and (2) a tHDA assay to detect blaTEM gene. The amplicons were visualized by agarose gel electrophoresis method, which is a cheap and widely used technique in general molecular laboratory. These simple isothermal platforms for nucleic acid amplification utilize only one pair of primers (like PCR) and the commonly available heating instruments. Moreover, compared to other isothermal platforms, the RPA kit (TwistAmp Basic kit) is a highly stable lyophilized reagent with long shelf life38. The lyophilized pellets in RPA kit are stable for at least one year when stored at temperatures below − 15 °C or at 2–8 °C, and up to 6 months at room temperature (22–28 °C)39. However, the RPA kit is slightly more expensive (~ $4.3 USD) than other amplification assays, such as PCR and LAMP (expired patents)38. Further, RPA patent is set to expire in 202340, making way for the development of a new cost-effective in-house RPA formula which will allow for its large scale applications.

The nucleotide compositions and length of primers affect the optimum annealing temperature and incubation time in multiplex RPA reactions. Here high temperatures (leading to poor primers binding) or short incubation times did not yield blaSHV amplicons, probably because the blaSHV primers have lower GC contents and shorter lengths than blaCTX-M and blaOXA primers. High primer concentrations increases the chance of primer-dimer and non-specific amplicon formation in multiplex RPA reactions, while low concentrations may lead to low yields. Therefore, good primer designing and extensive optimization are critical for effective multiplexed isothermal amplification41. The LOD of our multiplex RPA-gel electrophoresis assay in detecting blaCTX-M, blaOXA, and blaSHV genes was slightly lower than that of the previous multiplex RPA-Lateral Flow Assay (LFA)14. RPA amplicons must be purified before loading onto the agarose gel to remove protein contaminants42. This RPA post-amplification purification step causes amplicon loss. Moreover, the detection sensitivity using agarose gel electrophoresis depends on efficacy of various steps (pre-loading, pre-casting, and post-staining)43. Thus, the RPA post-amplification purification is a major drawback for gel electrophoresis detection44. Several methods for RPA products purification exist, including heat denaturation (65 °C or 95 °C for 10 min), sodium dodecyl sulfate treatment, proteinase K digestion, protein sedimentation via high-speed centrifugation and purification using commercial DNA purification kits44. Although agarose gel electrophoresis is a common method for visualization of amplification products, it is time consuming due to gel preparation, electrophoresis, gel staining, and imaging steps. These limitations can be circumvented in the future—by using various in-house detection methods (which employ bridge flocculation, SYBR green I or lateral flow assays)—to develop a cost-effective, simple, equipment-free, rapid, and naked-eye assay for detecting the desired genes39,44.

Assessing LOD of tHDA assay for blaTEM required 10x–100 × the template concentration used in RPA assays for blaCTX-M, blaOXA, blaSHV genes. The high-temperature tHDA operating condition (60–65 °C) may affect the primer binding efficiency; however specificity could be higher than mesophilic isothermal amplification38. Moreover, compared to RPA and PCR, we used lower primer concentrations (0.1–1 μM vs. 75–100 nM) for tHDA assay. The tHDA amplicon is generally < 150 nucleotides long, due to helicase processivity which limits the multiplexing capacity41,45. Our multiplex RPA assay showed 100% sensitivity and specificity for blaCTX-M, blaOXA, and blaSHV genes. However, tHDA had 86.89% sensitivity and 100% specificity for blaTEM. The higher LOD and lower primer concentrations used may have caused tHDA assay to have lower sensitivity, as compared to the RPA assay. The false-negative results may occur for samples with low quantity and quality of DNA (old/degraded DNA).

ExPEC strains, compared to commensal E. coli strains, have complex phylogenetic structure and diverse VFs4. ExPEC strains, such as uropathogenic E. coli (UPEC), neonatal meningitis E. coli (NMEC), sepsis-associated E. coli (SEPEC), and avian pathogenic E. coli (APEC), have several VFs. ExPEC VFs—such as adhesins, toxins and iron acquisition, lipopolysaccharides, capsules, and invasins—facilitate their colonization and systemic infection. UPEC prevalently causes urinary tract infection and secondary bacteremia18. Here, three VF genes (traT, ompT, and hlyA) were significantly associated with ESBL-ExPEC strains. traT was the most prevalent one, present in CTX-M-14, 15, and 27 variants. Other prevalent VF genes were aerobactin acquisition (iutA) and adhesins (fimH and iha). While a few cvaC existed only in our ESBL-ExPEC clinical isolates. The serum resistance gene traT, which inhibits the classical pathway of complement activity, is present in all pathotypes with more prevalence in UPEC46,47,48. The outer membrane protein (ompT) associated with UPEC enables intracellular survival and evasion from the host defense. Hemolysin A (hlyA) is a membrane lysis toxin present in UPEC strains. The afimbrial adhesin (afa) was found associated with bacteremia mortality23.

Only phylogroup B2 was significantly different between ESBL-ExPEC and non-ESBL-ExPEC. Similarly, the extended-spectrum cephalosporin resistant E. coli isolates belonging to phylogroup B2 carried ESBL and/or plasmid-mediated AmpC genes; blaCTX-M-15, blaCTX-M-14, blaCTX-M-55 were the most prevalent49. The phylogroup A and F are commensal and avian associated strains50, respectively. They are the second and third abundant phylogroups in our E. coli isolates. In addition, only three VF genes (ompT, iha, and hlyA) were distributed differently across phylogroups, especially in B2, F, and A. The phylogroup B2 was the most prevalent in clinical isolates (from many systemic infections) and carried a greater number of VFs than other phylogroups4. Although phylogroup F was less virulent than B2, its clinical significance in mediating AMR in ExPEC is established49,51,52.

In conclusion, most ESBL-ExPECs possessed a MDR pattern belonging to phylogroup B2 and carried more VF genes. The traT, ompT, and hlyA were found associated with ESBL-ExPEC strains. The relationship of VF genes was first demonstrated across phylogroups with diverse ompT, iha, and hlyA. The E. coli pathotypes should be characterized in the future. The identification and characterization of AMR, VF genes, and their pathotypes and phylogenetic analysis may contribute to developing novel strategies for treating E. coli-mediated systemic infection. Moreover, our multiplex RPA and tHDA assays, presented here, are simple, rapid, and reliable in detecting the most common ESBLs genes. These assays are beneficial for strategizing targeted therapy in low resource settings (using common heating equipments) and epidemiological control.

Materials and methods

Bacterial isolates

A total of 144 E. coli clinical isolates from various specimens including blood, urine, pus, body fluid, and sputum were collected from a tertiary cancer institute in Bangkok (Thailand) between February 2017–September 2018. All isolates were preserved in skimmed milk at − 80 °C until use.

The sample size used here was calculated using the Buderer method,53 which helps evaluate sensitivity and specificity of diagnostic tests at a 95% confidence interval. The study was approved by the Ethical Committee of the National Cancer Institute, Thailand (certificate number 020/2562).

Antimicrobial susceptibility testing and ESBLs detection

Ninety-five ESBL-ExPEC and forty-nine non-ESBL-ExPEC isolates were selected consecutively and non-duplicate from clinical samples by phenotypic ESBLs screening. Their antimicrobial susceptibility to various antibiotics—including, amikacin (30 µg), gentamicin (10 µg), ampicillin (10 µg), amoxicillin/clavulanic acid (20/10 µg), ceftazidime (30 µg), cefotaxime (30 µg), cefdinir (5 µg), cefepime (30 µg), ciprofloxacin (5 µg), ceftriaxone (30 µg), doripenem (10 µg), ertapenem (10 µg), imipenem (10 µg), meropenem (10 µg), levofloxacin (5 µg), and piperacillin/tazobactam (1.25/23.75 µg)—was determined using disk diffusion method. The ESBLs confirmatory testing was carried out using the combination disk method according to the CLSI guidelines11. ESBLs was considered to be produced when size of either inhibition zones of cefotaxime/clavulanic acid (30/10 µg) or ceftazidime/clavulanic acid (30/10 µg) were ≥ 5 mm compared to that of cefotaxime (30 µg), or ceftazidime (30 µg).

DNA extraction

DNA was isolated from E. coli pure colonies cultured on MacConkey agar by boiling method. Briefly, bacterial isolates were suspended in 400 µL of Tris–EDTA buffer (TE buffer), vortex mixed, and boiled at 95°C for 10 min. The supernatant was separated by centrifugation at 12,000 rpm for 5 min. DNA was precipitated by adding 1/10 volume of 3 M NaOAc, pH 5.2, and 2–3 volumes of chilled absolute ethanol and centrifuged at 12,000 rpm for 15 min. The bacterial DNA was washed with 70% ethanol, dried, and dissolved in 50 µL of TE buffer. DNA was stored at − 20 °C for subsequent amplification.

PCR amplification and DNA Sequencing

All E. coli clinical isolates were investigated for the presence of ESBL genes (blaCTX-M, blaOXA, blaSHV, and blaTEM). Further, the genes were sequenced to identify the variants. The primer sequences and amplicon sizes of each ESBL gene are described in Supplementary Table S2. The PCR reaction was carried out in a total volume of 25 µL comprising 50 ng of DNA, 1.5 mM MgCl2, 0.2 mM dNTPs, 0.4 µM of each primer, 1 × Standard Taq reaction buffer, and 1.25 U Taq polymerase (New England Biolabs, UK). The amplification of each ESBL gene were performed as previously described54,55,56,57. Positive control of blaTEM, blaCTX-M, blaOXA, and blaSHV were derived from E. coli EC137 and K.pneumoniae KP125, respectively (kindly provided by Prof. Visanu Thamlikitkul, Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok, Thailand). All PCR products were examined using 2% agarose gel electrophoresis. The amplified fragments were sequenced (Bioneer Corporation, South Korea) and aligned with the GenBank database using the BLASTn program (https://blast.ncbi.nlm.nih.gov/) and Clustal Omega (https://www.ebi.ac.uk/Tools/msa/ ustalo/), respectively.

Multiplex RPA reaction for bla CTX-M, bla OXA, and bla SHV genes

The RPA primers for blaCTX-M, blaOXA, and blaSHV, genes described previously were used (Supplementary Table S3). The multiplex RPA reaction was performed using the TwistAmp Basic reaction kit (TwistDx, UK). The RPA master mix contained 0.2 µM of blaCTX-M primers, 0.1 µM each of blaOXA and blaSHV primers, 29.5 µL rehydration buffer, 50 ng of DNA template, and sterile distilled water added to obtain a final volume of 47.5 µL. The reaction was vortex mixed, and then transferred to a freeze-dried tube, and finally mixed with 2.5 µL of 280 mM MgOAc. The reaction was incubated at 37 °C for 25 min. To explore the optimum conditions of the multiplex RPA reactions for blaCTX-M, blaOXA, and blaSHV, the following conditions were tried: (a) two sets of CTX-M, OXA, and SHV primer concentrations—0.2, 0.1, 0.1 and 0.2, 0.05, 0.05 µM; (b) three reaction temperatures—37 °C, 39 °C, and 41 °C; and (c) five incubation durations—10, 15, 20, 25, and 30 min. The RPA amplicons were purified using a GeneJet PCR purification kit (Thermo Fisher Scientific Inc., USA) and detected using 2% agarose gel electrophoresis.

tHDA reaction for bla TEM gene

The blaTEM gene of E. coli (accession no. MG515250.1) was used for primer design using Primer3plus program (http://www.bioinformatics.nl/cgi-bin/primer3plus/primer3plus.cgi). The sequences were as follows: forward primer, 5′-TGAGTGATAACACTGCGGCCAACTTAC-3′; reverse primer, 5′-CCCTACGATCAAGGCGAGTTACATGAT-3′. The tHDA reaction was performed in a total volume of 50 µL using IsoAmp® II Universal tHDA Kit (New England Biolabs, Inc., USA). The reaction mixer contained 1X Annealing buffer II, 4 mM MgSO4, 40 mM NaCl, 3.5 µL IsoAmp® dNTP Solution, 0.075 µM of each primer, and 3.5 µL IsoAmp® Enzyme Mix. The reaction was overlaid with mineral oil and incubated at 67 °C for 75 min. The tHDA condition was optimized by varying primer concentration (0.025, 0.050, and 0.075 µM), temperature (63 °C, 65 °C, 67 °C, 69 °C, 71 °C) and incubation time (30, 45, 60, 75, and 90 min). The tHDA amplicon of 111 bp was detected using 2% agarose gel electrophoresis.

Analytical LOD and specificity of multiplex RPA and tHDA assays

The template DNA from E. coli EC120, K. pneumoniae KP125, K. pneumoniae ATCC 700,603, E. coli EC137 harboring blaCTX-M, blaOXA, blaSHV, and blaTEM genes, respectively, were diluted to 100, 50, 5, 0.5, 0.05 ng/µL concentrations. The LOD was assessed as the lowest DNA concentration required for amplicon production using the optimized (in this study) multiplex RPA and tHDA assays. The specificity was also evaluated using DNA isolated from E. coli ATCC 25,922, K. pneumoniae ATCC 700,603, Proteus mirabilis ATCC 25,933, Pseudomonas aeruginosa ATCC 27,853, Acinetobacter baumannii ATCC 19,606, Staphylococcus aureus ATCC 25,923, and Enterococcus faecalis ATCC 29,212.

Validation of multiplex RPA and tHDA assays

All 144 clinical isolates were examined for the presence of blaCTX-M, blaOXA, and blaSHV genes (multiplex RPA assay) and blaTEM gene (tHDA assay). The sensitivity, specificity, positive and negative predictive values were calculated by comparing with nucleotide sequencing results obtained using MEDCALC® easy-to-use statistical software (https://www.medcalc.org/calc/diagnostic_test.php).

Detection of virulence genes

Eight virulence genes, viz. traT, ompT, iutA, fimH, hlyA, iha, papC, and cvaC, which are commonly associated with ExPEC were characterized by multiplex PCR methods. Three positive control genes (fimH, hlyA, and iutA) were synthesized and commercially cloned in GeneArt vectors (Invitrogen, USA). The recombinant plasmids were then transformed into a DH5α competent cell using Subcloning Efficiency™ DH5α Competent Cell (Invitrogen, USA). The primers for these three virulence genes (designed here) and five primers sets for traT, ompT, cvaC, iha, and papC genes (described previously) are tabulated in Supplementary Table S4. The first pool consists of fimH, hlyA, and iutA, while the second pool comprises cvaC, iha, traT, ompT, and papC. A total 25 µL reaction comprising 1 × Standard Taq reaction buffer and 1.5 U Taq polymerase (New England Biolabs, UK), 0.2 µM of each set of primers (forward and reverse), 0.2 mM dNTPs, ~ 100 ng DNA template and ultra-pure water were used. The optimized PCR condition for virulence genes was: initial denaturation at 94 ℃ for 15 min followed by 29 cycles at 94 ℃ for 1 min, annealing at 56 ℃ for 1 min, extension at 72 ℃ for 1 min, and post extension at 72 ℃ for 10 min. All amplicons were analyzed using 2% agarose gel electrophoresis.

Phylogroup analysis

Quadruplex PCR was performed by targeting arpA, chuA, yjaA, and TspE4.C2 (Clermont et al.). An additional PCR was carried out as described previously20 using specific primers for groups C, E, and cryptic Clades. Previously published primers and the amplicon size are illustrated (Supplementary Table S5). The phylogroups B1, B2, and F were assigned based on quadruplex results, while specific primers were used to differentiate phylogroups C, E, and other cryptic Clades from A, D, and Clade I20. A total reaction of 25 µL consisting of 2.5 µL 10 × buffer, 0.2 mM of each dNTPs, 1 µL DNA template (100 ng), 1.5 U Taq polymerase (New England Biolabs, UK), 0.2 µM primers except for internal control (trpBA = 0.12 µM) and Milli Q water was used. We followed the same PCR conditions as described by Clermont et al. 201320. PCR products were analyzed using 2% agarose gel electrophoresis.

Statistical analysis

The association of VF genes and phylogroups was compared between ESBL and non-ESBL-ExPEC isolates using Mann–Whitney U test. The distribution of VF genes across phylogroups was tested using Friedman’s and Kruskal–Wallis tests. P value < 0.05 was considered statistically significant.

Ethical approval

The study protocol was approved by Research Committee of National Cancer Institute (certificate number 020/2562).