Detection of Extended-spectrum β-lactamase-producing Escherichia coli isolates by isothermal amplification and association of their virulence genes and phylogroups with extraintestinal infection

Extraintestinal pathogenic Escherichia coli (ExPEC) producing extended-spectrum β-lactamases (ESBL) cause serious human infections due to their virulence and multidrug resistance (MDR) profiles. We characterized 144 ExPEC strains (collected from a tertiary cancer institute) in terms of antimicrobial susceptibility spectrum, ESBL variants, virulence factors (VF) patterns, and Clermont’s phylogroup classification. The developed multiplex recombinase polymerase amplification and thermophilic helicase-dependent amplification (tHDA) assays for blaCTX-M, blaOXA, blaSHV, and blaTEM detection, respectively, were validated using PCR-sequencing results. All ESBL-ExPEC isolates carried blaCTX-M genes with following prevalence frequency of variants: blaCTX-M-15 (50.5%) > blaCTX-M-55 (17.9%) > blaCTX-M-27 (16.8%) > blaCTX-M-14 (14.7%). The multiplex recombinase polymerase amplification assay had 100% sensitivity, and specificity for blaCTX-M, blaOXA, blaSHV, while tHDA had 86.89% sensitivity, and 100% specificity for blaTEM. The VF genes showed the following prevalence 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%), in ESBL-ExPEC isolates which belonged to phylogroups A (28.4%), B2 (28.4%), and F (22.1%). The distribution of traT, ompT, and hlyA and phylogroup B2 were significantly different (P < 0.05) between ESBL-ExPEC and non-ESBL-ExPEC isolates. Thus, these equipment-free isothermal resistance gene amplification assays contribute to effective treatment and control of virulent ExPEC, especially antimicrobial resistance strains.

Limit of detection (LOD) and specificity of multiplex RPA and tHDA assays. The DNA templates were isolated from E. coli clinical strain ESBL120 carrying bla CTX-M , strain KP125 carrying bla OXA , strain EC137 carrying bla TEM , and K. pneumoniae ATCC 700,603 carrying bla SHV . The LOD of multiplex RPA assays for bla CTX-M , bla OXA , bla SHV 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 bla TEM (LOD of 50 ng) (Fig. 3D).

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 lineages 4,26 . Moreover, plasmid-mediated horizontal transfer of ESBL genes occur easily among species, causing widespread infection 6 . Accumulating evidence suggested that VFs helps gastrointestinal pathogens www.nature.com/scientificreports/ outcompete the commensal microbiota and impair host immunity via inducing colonization, resistance, and invasion 27 . ESBL-Enterobacteriaceae, especially E. coli was one of the most frequent isolates in blood stream infection samples collected from a pediatric oncology center 28 . The MDR E. coli from febrile neutropenic cancer patients showed high resistance to ampicillin, cefepime, ceftriaxone, and cephradine 29  Here, we developed the following isothermal assays to detect common ESBL genes: (1) a multiplex RPA assay to detect bla CTX-M , bla OXA , and bla SHV genes; and (2) a tHDA assay to detect bla TEM 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 life 38 . 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,    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 2023 40 , 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 bla SHV amplicons, probably because the bla SHV primers have lower GC contents and shorter lengths than bla CTX-M and bla OXA 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 amplification 41 . The LOD of our multiplex RPA-gel electrophoresis assay in detecting bla CTX-M , bla OXA , and bla SHV 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 contaminants 42 . 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 detection 44 . 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 kits 44 . 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 genes 39,44 .
Assessing LOD of tHDA assay for bla TEM required 10x-100 × the template concentration used in RPA assays for bla CTX-M , bla OXA , bla SHV genes. The high-temperature tHDA operating condition (60-65 °C) may affect the primer binding efficiency; however specificity could be higher than mesophilic isothermal amplification 38 . 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 capacity 41,45 . Our multiplex RPA assay showed 100% sensitivity and specificity for bla CTX-M , bla OXA , and bla SHV genes. However, tHDA had 86.89% sensitivity and 100% specificity for bla TEM . 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 VFs 4 . 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 bacteremia 18 . 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 UPEC [46][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 mortality 23 .
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; bla CTX-M-15 , bla CTX-M-14 , bla CTX-M-55 were the most prevalent 49 . The phylogroup A and F are commensal and avian associated strains 50 , 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 www.nature.com/scientificreports/ isolates (from many systemic infections) and carried a greater number of VFs than other phylogroups 4 . Although phylogroup F was less virulent than B2, its clinical significance in mediating AMR in ExPEC is established 49,51,52 .
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).
PCR amplification and DNA Sequencing. All E. coli clinical isolates were investigated for the presence of ESBL genes (bla CTX-M , bla OXA , bla SHV , and bla TEM ). 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 MgCl 2 , 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 described [54][55][56][57] . Positive control of bla TEM , bla CTX-M, bla OXA , and bla SHV 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 bla CTX-M , bla OXA , and bla SHV , 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 bla CTX-M primers, 0.1 µM each of bla OXA and bla SHV 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.  ). An additional PCR was carried out as described previously 20 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 I 20 . 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. 2013 20 . 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).

Data availability
The datasets generated during the current study are available in the genbank NCBI repository, with the accession number including OP999005-OP999011. These datasets were derived from the following public domain resources