Whole genome analysis of cephalosporin-resistant Escherichia coli from bloodstream infections in Australia, New Zealand and Singapore: high prevalence of CMY-2 producers and ST131 carrying blaCTX-M-15 and blaCTX-M-27

Whole genome analysis of cephalosporin-resistant Escherichia coli from bloodstream infections in... Abstract Objectives To characterize MDR Escherichia coli from bloodstream infections (BSIs) in Australia, New Zealand and Singapore. Methods We collected third-generation cephalosporin-resistant (3GC-R) E. coli from blood cultures in patients enrolled in a randomized controlled trial from February 2014 to August 2015. WGS was used to characterize antibiotic resistance genes, MLST, plasmids and phylogenetic relationships. Antibiotic susceptibility was determined using disc diffusion and Etest. Results A total of 70 3GC-R E. coli were included, of which the majority were ST131 (61.4%). BSI was most frequently from a urinary source (69.6%), community associated (62.9%) and in older patients (median age 71 years). The median Pitt score was 1 and ICU admission was infrequent (3.1%). ST131 possessed more acquired resistance genes than non-ST131 (P = 0.003). Clade C1/C2 ST131 predominated (30.2% and 53.5% of ST131, respectively) and these were all ciprofloxacin resistant. All clade A ST131 (n = 6) were community associated. The predominant ESBL types were blaCTX-M (80.0%) and were strongly associated with ST131 (95% carried blaCTX-M), with the majority blaCTX-M-15. Clade C1 was associated with blaCTX-M-14 and blaCTX-M-27, whereas blaCTX-M-15 predominated in clade C2. Plasmid-mediated AmpC genes (mainly blaCMY-2) were frequent (17.1%) but were more common in non-ST131 (P < 0.001) isolates from Singapore and Brisbane. Two strains carried both blaCMY-2 and blaCTX-M. The majority of plasmid replicon types were IncF. Conclusions In a prospective collection of 3GC-R E. coli causing BSI, community-associated Clade C1/C2 ST131 predominate in association with blaCTX-M ESBLs, although a significant proportion of non-ST131 strains carried blaCMY-2. Introduction In recent decades, resistance to β-lactam antibiotics in Enterobacteriaceae has become increasingly common. Of particular concern has been the rising prevalence of resistance to third-generation cephalosporins (3GCs), a class that includes key agents such as ceftriaxone, cefotaxime or ceftazidime.1–3 This phenomenon has largely arisen from the dissemination of genes encoding ESBL or, less frequently, plasmid-mediated AmpC (p-AmpC) enzymes.4–6 These resistance genes are often acquired by plasmid transfer and may be associated with other antibiotic resistance determinants, rendering organisms MDR.7 The global emergence of infections caused by ESBL-producing Escherichia coli, in both the community and hospital setting, has been driven by the acquisition of CTX-M-type ESBL genes (especially blaCTX-M-15) in the successful pandemic clone of E. coli, sequence type 131 (ST131).8–11,E. coli ST131 lineages belong to the B2 phylogenetic subgroup I, and are mostly serotype O25b:H4.12 Within ST131, further sublineages have been delineated according to fimH (type 1 fimbrial adhesin, FimH) alleles, phylogenetic clades (A, B, C1 and C2) and associated resistance genes.8,13 A globally dominant fluoroquinolone-resistant fimH30 subclonal lineage, defined as H30-R (to differentiate from the ancestral fluoroquinolone susceptible H30 strains) or clade C, has been described.8,14,H30-R/clade C strains contain fluoroquinolone resistance mutations in the chromosomal gyrA and parC genes15 and have been associated with poor clinical outcomes.16 Within this sublineage, a pathogenic ST131 subclone containing blaCTX-M-15 has been referred to as H30-Rx14 or clade C2.8 Specific Incompatibility (Inc) F-type plasmids have also been described in association with fluoroquinolone-resistant ST131-H30 clades, with IncF-type F1:A2:B20 plasmids associated with the H30-R/C1 clade and IncF-type F2:A1:B- plasmids associated with the H30-Rx/C2 clade.17–19 Resistance to β-lactams mediated by ESBLs drives the use of broader-spectrum antibiotics such as carbapenems (e.g. meropenem),4 providing selection pressure for carbapenem resistance in Gram-negative bacteria. Of great concern has been the emergence of transmissible carbapenemases in common Enterobacteriaceae.20 As part of an international randomized controlled trial of piperacillin/tazobactam (a potential ‘carbapenem-sparing’ agent) versus meropenem for the treatment of 3GC-resistant (3GC-R) E. coli causing bloodstream infection (BSI),21 we aimed to further characterise these isolates using WGS. Methods Bacterial isolates and clinical data 3GC-R E. coli were collected prospectively during an international multicentre randomized trial comparing treatment options for BSI caused by ceftriaxone non-susceptible E. coli or Klebsiella spp. [the ‘MERINO’ trial: Australian New Zealand Clinical Trials Register (ANZCTR), ref. no. ACTRN12613000532707 and the US National Institute of Health ClinicalTrials.gov register, ref. no. NCT02176122].21 All initial 3GC-R E. coli blood culture isolates (i.e. the first positive culture prompting study inclusion) were included from 70 patients enrolled in the trial for the first 18 months of recruitment (from February 2014 until August 2015) from eight hospital sites in three countries [Australia (n = 21), New Zealand (n = 5) and Singapore (n = 44)]. To be eligible for the trial, patients had to have at least one monomicrobial blood culture growing E. coli, with non-susceptibility to ceftriaxone, ceftazidime or cefotaxime and susceptibility to piperacillin/tazobactam and meropenem, determined by methods used in the local laboratories. All blood culture isolates from patients meeting inclusion criteria who consented to participate in the study were stored at the recruiting site laboratory at −80 °C in cryovials containing glycerol and nutrient broth and later shipped to the coordinating laboratory in Queensland, Australia. For the purposes of this study, only the first 3GC-R E. coli isolated from blood cultures for each enrolled patient were included in the genomic analysis. Relevant clinical data were collected and managed using the REDCap22 electronic data capture tool hosted at the University of Queensland. Healthcare-associated infection (HAI) was defined as a positive blood culture collected in a patient within 48 h of hospital admission, if the patient had received specialized nursing care at home (e.g. intravenous therapy, wound care), attended a hospital, haemodialysis unit, or received intravenous chemotherapy in the 30 days prior to the BSI event, was a resident of a nursing home or long-term care facility or had been admitted to an acute care facility for >48 h in the preceding 90 days.23 HAI also included nosocomial bacteraemia occurring >48 h following hospital admission. BSI was defined as community associated if it occurred within 48 h of admission and the patient did not meet any criteria for HAI.23 Ethics Approval for the study was provided by the Royal Brisbane and Women’s Hospital (ref. HREC/12/QRBW/440), the National Healthcare Group (NHG) Domain Specific Review Board (DSRB) in Singapore (NHG DSRB ref. 2013/00453) and the New Zealand Health and Disability Ethics Committee (ref. 14/NTB/52). Phenotypic susceptibility testing All isolates were tested at the coordinating laboratory against a standard panel of antibiotics used to treat Gram-negative infections by disc diffusion according to EUCAST standards.24 Agents tested included ceftriaxone, ceftazidime, cefepime, cefoxitin, aztreonam, ertapenem, gentamicin, amikacin, ciprofloxacin, co-trimoxazole and amoxicillin/clavulanate. In addition, MICs of piperacillin/tazobactam, meropenem (the two comparator drugs used in the trial) and ceftriaxone were determined by Etest (bioMérieux). ESBL production was confirmed using combination disc testing with ceftriaxone and ceftazidime with and without clavulanate; an increase in zone diameter ≥5 mm with the addition of clavulanate confirmed ESBL production.25 DNA extraction and library preparation After subculture onto LB agar to check for viability and purity, genomic DNA was extracted using the MoBio Ultrapure kit and quantified by spectrophotometry (NanoDrop; ThermoFisher) and fluorometry (Qubit; ThermoFisher). Paired-end DNA libraries were prepared using the Nextera kit (Illumina; Australia) in accordance with the manufacturer’s instructions. WGS WGS was performed in two batches using Illumina HiSeq (100 bp paired end) and MiSeq (300 bp paired end) at the Australian Genome Research Facility (AGRF), University of Queensland, St Lucia. MiSeq raw reads were trimmed conservatively to 150 bp and filtered using Nesoni (version 0.130) to remove Illumina adaptor sequences, reads <80 bp and bases below Phred quality 5 (https://github.com/Victorian-Bioinformatics-Consortium/nesoni). Strains were checked for contamination using Kraken (0.10.5-beta) as implemented through Nullarbor (default settings).26 Resistance gene detection, MLST and plasmid typing Antibiotic resistance genes were detected by using Abricate (version 0.2) with the ResFinder database against SPAdes assemblies (version 3.6.2) as implemented through the pipeline analysis tool Nullarbor (default settings).26 MLST was undertaken using the mlst tool as implemented through Nullarbor. Plasmid replicon typing and plasmid multilocus typing were performed using PlasmidFinder and pMLST.27 Fluoroquinolone resistance SNP detection Filtered reads were mapped to the complete ST131 E. coli reference strain EC958 (GenBank: HG941718.1) using Bowtie as implemented through Nesoni. Non-synonymous mutations were identified using Nesoni nway and manually compared with known mutations in gyrA and parC associated with quinolone resistance.13,28 Phylogenetic analysis Reads for all isolates (n = 70) were mapped to the complete ST131 reference EC958 (GenBank: HG941718.1)29 using Nesoni under default settings. SNPs identified between isolates and the reference EC958 were used to create pseudogenomes for each isolate by substituting the relevant SNPs into the EC958 chromosome using an in-house script. Multiple sequence alignment of the pseudogenomes was used as input for Gubbins (version 1.3.4)30 using the (GTR)GAMMA substitution model to parse recombinant regions. The remaining 211 920 SNPs were used to generate a phylogenetic tree using RAxML (version 8.2.9)31 under the (GTR)GAMMA substitution model with Lewis ascertainment bias correction and a random seed of 456 (100 bootstraps). An ST131-only tree (n = 43) was also created in the same manner using 2248 recombination-free SNPs and 1000 bootstraps. In silico phylo-grouping for all STs was undertaken according to the updated Clermont scheme.32 Phylogenetic trees and associated metadata were visualized using Evolview.33 Sequence data Raw sequence reads and associated metadata have been uploaded to NCBI (Bioproject Accession no. PRJNA398288). A full description of clinical metadata, MLST, acquired resistance genes and plasmid typing is available online at Harvard Dataverse (doi: 10.7910/DVN/YC9WDV). Statistical tests Data describing patient demographics, phenotypic susceptibility, clinical variables and genotypic data for all cases were tabulated, with proportions expressed as percentages and median, mean or IQR calculated as appropriate for scale variables. Categorical variables were compared using Pearson’s χ2 test.Comparisons of mean values in normally distributed data were compared using the t-test. The Mann–Whitney U-test was used for non-parametric data. Statistical analysis was performed using Stata version 13.1 (StataCorp; TX, USA) and graphical images prepared using Prism version 7.0 (GraphPad Software; CA, USA). A P value <0.05 was considered significant. Results Baseline clinical data A total of 70 3GC-R E. coli bloodstream isolates were included. The background clinical and demographic details of enrolled patients are summarized in Table 1. The source of BSI was most frequently the urinary tract (48/70, 69.6%) and infections were mostly community associated (44/70, 62.9%). There was a predominance of patients reporting Chinese ethnicity (38/70, 54.3% of all cases), reflecting the demographics of the largest recruiting sites in Singapore. There was also a greater proportion of male patients (60%), but this was not statistically significant (P = 0.12). Patients tended to be older (median age 71, IQR 64–81 years, range 20–94 years), although only a small proportion (5.8%) were admitted from nursing homes. The majority of patients had less severe acute illness (median Pitt score 1, IQR 0–2, range 0–3; where a score ≥4 reflects the presence of critical illness with high mortality34) and relatively low comorbidity scores (Charlson score median 2, IQR 1–4, range 0–11) and were infrequently admitted to the ICU (3.1%). Table 1. Baseline clinical and demographic variables Factor/level  No. (%) (N = 70)  Region     Singapore  44 (63)   Brisbane (AUS)  13 (19)   Melbourne (AUS)  8 (11)   Auckland (NZ)  5 (7)  Age, years, mean (SD)  69.9 (15.2)  Gender     female  28 (40)   male  42 (60)  Source of bacteraemia     urinary tract infection  48 (70)   intra-abdominal infection  8 (12)   pneumonia  1 (1)   other  9 (13)   unknown  3 (4)  Acquisition     community associated  44 (63)   healthcare associated  26 (37)  Pitt score, median (IQR)  1 (0–2)  Charlson score, median (IQR)  2 (1–4)  Any CTX-M ESBL  56 (80)  AmpC β-lactamase  12 (17)  Surgery within 14 days  5 (7)  Central venous catheter  6 (9)  Immune suppression  10 (14)  ICU admission  2 (3)  Nursing home resident  4 (6)  Factor/level  No. (%) (N = 70)  Region     Singapore  44 (63)   Brisbane (AUS)  13 (19)   Melbourne (AUS)  8 (11)   Auckland (NZ)  5 (7)  Age, years, mean (SD)  69.9 (15.2)  Gender     female  28 (40)   male  42 (60)  Source of bacteraemia     urinary tract infection  48 (70)   intra-abdominal infection  8 (12)   pneumonia  1 (1)   other  9 (13)   unknown  3 (4)  Acquisition     community associated  44 (63)   healthcare associated  26 (37)  Pitt score, median (IQR)  1 (0–2)  Charlson score, median (IQR)  2 (1–4)  Any CTX-M ESBL  56 (80)  AmpC β-lactamase  12 (17)  Surgery within 14 days  5 (7)  Central venous catheter  6 (9)  Immune suppression  10 (14)  ICU admission  2 (3)  Nursing home resident  4 (6)  Antibiotic resistance phenotypes By MIC testing, 97.1% (68/70) were susceptible to piperacillin/tazobactam (median 2 mg/L, range 1–24 mg/L, IQR 1.5–4; EUCAST breakpoint for susceptibility ≤8 mg/L24) (see Figure S1, available as Supplementary data at JAC Online). All strains were susceptible to meropenem (MIC90 0.047 mg/L; median 0.023 mg/L, range 0.012–0.19 mg/L; EUCAST breakpoint for susceptibility ≤2 mg/L24), although one strain (MER-86) was non-susceptible to ertapenem. The majority (90.1%) demonstrated ceftriaxone MICs ≥32 mg/L (range 0.064 to ≥32 mg/L; median ≥32 mg/L, MIC90 and MIC50 ≥32 mg/L). Two strains, which were susceptible to ceftriaxone by MIC, were resistant to ceftazidime [one (MER-15) contained blaDHA-1, the other (MER-100) had changes to the ampC promotor region, discussed in detail later]. Phenotypic resistance to 3GCs could not be detected in one strain (MER-34) when retested in the coordinating laboratory, and was found to only possess TEM-176 (a narrow-spectrum β-lactamase) by sequencing. Resistance to gentamicin was common (25/70, 35.7%), although all strains remained susceptible to amikacin. Strains demonstrated a variable antibiogram according to ST131 clade (see Table 2), but were frequently resistant to trimethoprim/sulfamethoxazole (46/70, 65.7%) or fluoroquinolones (52/70, 74.3%). There was universal resistance to ciprofloxacin in clade C1/C2 ST131 (n = 36), compared with only 50% (3/6) and 48.2% (13/27) in clade A and non-ST131 strains, respectively (P < 0.001). Table 2. Antibiotic resistance profile of E. coli strains according to ST131 clade     Non-susceptible, n (%)   ST131 clade    CTX  CAZ  FEP  FOX  AMC  TZP  MEM  ETP  ATM  SXT  GEN  AMK  CIP  A  6  5 (83)  5 (83)  5 (83)  1 (17)  4 (67)  0 (0)  0 (0)  0 (0)  5 (83)  4 (67)  3 (50)  0 (0)  3 (50)  B  1  1 (100)  1 (100)  1 (100)  0 (0)  1 (100)  0 (0)  0 (0)  0 (0)  1 (100)  1 (100)  1 (100)  0 (0)  0 (0)  C1  13  13 (100)  4 (31)  12 (92)  1 (8)  5 (38)  0 (0)  0 (0)  0 (0)  12 (92)  8 (62)  4 (31)  0 (0)  13 (100)  C2  23  23 (100)  21 (91)  22 (96)  1 (4)  21 (91)  2 (9)  0 (0)  0 (0)  23 (100)  19 (83)  13 (57)  0 (0)  23 (100)  Non-ST131  27  25 (93)  22 (81)  14 (52)  10 (37)  20 (74)  0 (0)  0 (0)  1 (4)  21 (78)  14 (52)  4 (15)  0 (0)  13 (48)  All  70  67 (96)  53 (76)  54 (77)  13 (19)  51 (73)  2 (3)  0 (0)  1 (1)  62 (89)  46 (66)  25 (36)  0 (0)  52 (74)      Non-susceptible, n (%)   ST131 clade    CTX  CAZ  FEP  FOX  AMC  TZP  MEM  ETP  ATM  SXT  GEN  AMK  CIP  A  6  5 (83)  5 (83)  5 (83)  1 (17)  4 (67)  0 (0)  0 (0)  0 (0)  5 (83)  4 (67)  3 (50)  0 (0)  3 (50)  B  1  1 (100)  1 (100)  1 (100)  0 (0)  1 (100)  0 (0)  0 (0)  0 (0)  1 (100)  1 (100)  1 (100)  0 (0)  0 (0)  C1  13  13 (100)  4 (31)  12 (92)  1 (8)  5 (38)  0 (0)  0 (0)  0 (0)  12 (92)  8 (62)  4 (31)  0 (0)  13 (100)  C2  23  23 (100)  21 (91)  22 (96)  1 (4)  21 (91)  2 (9)  0 (0)  0 (0)  23 (100)  19 (83)  13 (57)  0 (0)  23 (100)  Non-ST131  27  25 (93)  22 (81)  14 (52)  10 (37)  20 (74)  0 (0)  0 (0)  1 (4)  21 (78)  14 (52)  4 (15)  0 (0)  13 (48)  All  70  67 (96)  53 (76)  54 (77)  13 (19)  51 (73)  2 (3)  0 (0)  1 (1)  62 (89)  46 (66)  25 (36)  0 (0)  52 (74)  CTX, ceftriaxone; CAZ, ceftazidime; FEP, cefepime; FOX, cefoxitin; AMC, amoxicillin/clavulanate; TZP, piperacillin/tazobactam; MEM, meropenem; ETP, ertapenem; ATM, aztreonam; SXT, trimethoprim/sulphamethoxazole; GEN, gentamicin; AMK, amikacin; CIP, ciprofloxacin. MLST and ST131 phylogeny A clear majority of strains were ST131 by in silico MLST (43/70, 61.4%), with other strains broadly distributed across a number of other STs (Figure 1). ST131 was prevalent in all city locations (Singapore, Melbourne, Brisbane and Auckland). The only STs with more than a single representative strain were ST1193, ST410, ST38, ST69 and ST963. The majority of ST131 strains belonged to clades C1 (30.2%) or C2 (53.5%), with strains from clades B (2.3%) and A (14.0%) seen less frequently. Figure 1. View largeDownload slide In silico MLST of ESBL- or AmpC-producing E. coli isolated from blood, by region. Figure 1. View largeDownload slide In silico MLST of ESBL- or AmpC-producing E. coli isolated from blood, by region. A phylogenetic tree of all 3GC-R E. coli strains can be found in Figure S2. All clade A ST131 were community associated, with a mixture of community- and healthcare-associated infections observed for strains in clades C1 and C2. There was evidence of clustering of closely related paired strains (with <10 core genome SNPs different) within certain hospitals (e.g. MER-27/25 in Hospital E; MER-8/10 and MER-37/39 in Hospital A; Figure 2), indicating a direct relationship or very recent transmission, although these represented both community- and healthcare-associated infections. Figure 2. View largeDownload slide Phylogenetic tree of ST131 E. coli based on core genome SNPs. Clade, antibiotic resistance, ESBL/p-AmpC-type and IncF plasmid type are shown. EC958 genome (NZ HG941718.1) used as reference.29 Figure 2. View largeDownload slide Phylogenetic tree of ST131 E. coli based on core genome SNPs. Clade, antibiotic resistance, ESBL/p-AmpC-type and IncF plasmid type are shown. EC958 genome (NZ HG941718.1) used as reference.29 Two of the Australian strains (MER-2 and MER-4) were both ST963, with chromosomally integrated blaCMY-2, and occurred in patients whose admissions overlapped in the same hospital. However, they were separated by >350 SNPs, suggesting a more distant common ancestor. Antibiotic resistance genes The median number of acquired antibiotic resistance genes detected for each isolate was 9. One strain (MER-90) possessed a total of 17 acquired resistance genes, including β-lactamases (blaCMY-2, blaTEM-1B), aminoglycoside resistance genes [aac(3)-IId-like, aadA1-like, aadA2, aph(3′)-Ic-like, strA, strB], resistance genes related to folate metabolism (dfrA12, dfrA14-like), fluoroquinolones (qnrS1), sulphonamides (sul1, sul2, sul3), tetracyclines [tet(A)], phenicols (floR-like) and macrolides [mph(A)]. The number of acquired antibiotic resistance genes was significantly greater in ST131 than non-ST131 strains (P = 0.003) (Figure S3A). However, the number of resistance genes did not vary significantly across ST131 clades (Figure S3C). No carbapenemase genes were identified. In one strain (MER-100), resistance to ceftazidime (but not ceftriaxone) was not clearly explained by resistance gene profiling, with no β-lactamase genes identified, although an altered −35 box (TTGACA) was found in its promoter, which has been associated with overexpression of the ampC promoter.35 A single strain (MER-86) demonstrated resistance to ertapenem and was found to have disruption in ompF, which has been associated with reduced susceptibility to ertapenem when associated with broad-spectrum β-lactamases.36 β-Lactamases The predominant ESBL genes identified were blaCTX-M, seen in 80.0% (56/70) of isolates. The presence of blaCTX-M was strongly associated with ST131; 95% of ST131 possessed blaCTX-M ESBLs, compared with only 56% of non-ST131 (P < 0.001) (Table S1). These were either from CTX-M group 9 [blaCTX-M-14 (n = 8),blaCTX-M-27 (n = 14)] or CTX-M group 1 [blaCTX-M-15 (n = 31),blaCTX-M-55 (n = 3)], with the majority blaCTX-M-15 (44.3%). Clade C1 ST131 (n = 13) were associated with blaCTX-M-14 (n = 4) and blaCTX-M-27 (n = 9), whereas blaCTX-M-15 predominated in clade C2 (95.7%; 22/23) (Figure S3B). ESBL genes found in non-ST131 strains included blaCTX-M-15 (n = 7; ST12, ST95, ST648, ST973 and ST1193), blaCTX-M-14 (n = 3; ST10, ST69 and ST4702), and blaCTX-M-27 (n = 1; ST10). The second most common group of β-lactamases with the ability to hydrolyse 3GCs were acquired AmpC β-lactamase genes, found in 17.1% (12/70). Acquired AmpC-producers were most frequently seen in strains from Brisbane (5/13, 38%), and Singapore (7/44, 15.9%), but none was detected in a smaller sample of patients from Melbourne or Auckland. The presence of acquired AmpC was not clearly associated with specific STs, but was more common in non-ST131 strains (37.0% versus 4.7%; P < 0.001). These were predominantly blaCMY-2, although a single strain carried blaDHA-1. A single clade C2 ST131 strain from Singapore possessed blaCMY-2 (in association with blaCTX-M-15) (Figure 2). The genetic context of blaCMY-2 in five strains from Singapore showed homology with a plasmid (GenBank: LC019731) identified in Klebsiella pneumoniae from Taiwan (Figure S4). blaCMY-2 located within an IncI1 plasmid backbone from two strains isolated in Australia was similar to a plasmid (pS10584; GenBank: KX058576.1) identified in Salmonella enterica from a food source in China (Figure S5). ISEcp1 was identified in all but two of the E. coli strains carrying blaCMY-2, with these associated with IS1294 and a truncated ISEcp1 (Figure S5). Although blaCMY-2 are usually acquired on plasmids, in four strains {three from Australia [MER-2 (ST963), MER-4 (ST963), MER-43 (ST38)] and one from Singapore [MER-99 (ST93)]} there was evidence to suggest chromosomal integration based on flanking genes (Figure S6). However, owing to repetitive regions surrounding the blaCMY-2 region, the complete context could not be ascertained in all isolates using short-read sequencing alone. A single strain (MER-89) possessed blaLAP-2, and two (MER-86 and MER-110) carried both blaCTX-M and blaCMY. Further details of the genetic context of blaCMY-2 and blaDHA-1 can be found in Figures S4–S7. Fluoroquinolone resistance genes Acquired quinolone resistance genes (i.e. those not mediated by SNPs in regions associated with quinolone resistance) were seen in 11% (8/70) of strains. These genes included qnrS1, qnrB4, qnrB66-like, oqxA or aac(6′)Ib-cr. The presence of these genes was more commonly seen in non-ST131 than ST131 (22% versus 5%; P = 0.025). In some strains (e.g. MER-34, MER-26) phenotypic ciprofloxacin resistance was not evident despite the presence of acquired resistance genes, although these genes usually only confer low-level fluoroquinolone resistance in the absence of additional mechanisms. All clade C strains [and a single clade A strain (MER-42)] were identical to the EC958 reference strain with respect to mutations in parC and gyrA (Tables S2 and S3). Phenotypic ciprofloxacin resistance was largely congruent with the presence of SNPs in parC and gyrA known to be associated with fluoroquinolone resistance, or the presence of acquired quinolone-resistance determinants.37 Certain gyrA SNPs (e.g. 83L) were not by themselves associated with phenotypic ciprofloxacin resistance unless accompanied by additional SNPs (e.g. 87N or 87Y) (Table S3). Sulphonamide and folate pathway resistance genes Sulphonamide resistance genes (sul1, sul2 or sul3) were common, and present in 69% (48/70) of strains, as were folate synthesis pathway (e.g. trimethoprim) resistance genes (54%, 38/70), such as dfrA1, dfrA7, dfrA12, dfr14 and dfrA17. The presence of sulphonamide resistance and trimethoprim resistance genes was more common in ST131 (81% versus 48%, P = 0.004, and 74% versus 22%, P < 0.001, respectively). Other resistance genes The presence of aminoglycoside-modifying enzymes (AMEs) was common (seen in 76% of strains) and was encountered more frequently in ST131 (86% versus 59%; P = 0.011). Genes mediating resistance to tetracyclines [specifically tet(A) and tet(B)] were seen in 56% (39/70) of strains, but were equally distributed between ST131 and non-ST131. Other frequently identified resistance genes included those mediating resistance to chloramphenicol (e.g. catA, florR) or macrolides [e.g. mph(A)]. Plasmids The majority of plasmid replicon types were identified as IncF. IncF-type plasmids were closely associated with blaCTX-M. Almost all (53/55, 96.4%) E. coli with blaCTX-M also carried IncF-type plasmids, whereas only 53.3% of strains without blaCTX-M carried plasmids typed as IncF. According to the PubMLST scheme (www.pubmlst.org/plasmid), plasmids seen in clade C1 ST131 were mainly IncF plasmid type F1:A2:B20 (76.9%), with the remainder IncF-type F1:A2:B- or IncI1 types (ST79 or unknown ST). Amongst clade C2 ST131, IncF-types F31:A4:B1 or F36:A4:B1 were most common (22.7% and 27.3%, respectively), with IncF-type F2:A1:B- plasmids only seen in 18.2% (Figure 2). Only three clade C2 strains contained IncI1 or IncN plasmids. The presence of blaCMY-2 was associated with IncI1-type plasmids (unless chromosomally integrated). In a single strain, blaDHA-1 was located within an IncF-type plasmid (Figure S7). However, given the limitations of reconstructing plasmids from short-read data alone, the location of β-lactamase genes within specific plasmids remains uncertain. A full summary of plasmid replicon types can be found in the Supplementary data. Discussion This prospective collection of ESBL- and p-AmpC-producing E. coli bloodstream isolates provides insight into the current clinical and molecular epidemiology of these infections within Australia, New Zealand and Singapore. The clear predominance of ST131 carrying CTX-M-type ESBLs is striking and reflects how this pandemic clone has emerged as a highly successful human pathogen. As has been described elsewhere, CTX-M-type ESBLs have now displaced TEM- or SHV-type ESBLs in many parts of the world,10,38 and the latter were not seen in this contemporary collection of E. coli bloodstream isolates. It is also notable that the majority of cases were community associated, with their origin from the urinary tract and in patients over the age of 65 years. An increasing proportion of infections caused by ESBL-producing E. coli are reported as being community acquired.39 Our study would suggest a community reservoir for 3GC-R E. coli, including CTX-M-producing ST131 (65% of which were associated with BSI in patients without any recent healthcare exposure). This contrasts to previous decades, following the first description of ESBLs, where nosocomial acquisition was common and TEM- and SHV-type EBSLs predominated.4 Different β-lactamase genes were associated with certain E. coli lineages; blaCTX-M-15 was largely restricted to the C2 clade amongst ST131, whereas blaCTX-M-27 and blaCTX-M-14 were found in clade C1, which is consistent with previous studies.8,40 Within ST131, clades C1/H30-R and C2/H30-Rx are globally disseminated, and have not demonstrated clear geographical clustering,8 although within some regions limited clonal expansion has occurred.40 The global prevalence of blaCTX-M-15 has increased in most countries over recent years, and dominates in most regions, although group 9 variants (including blaCTX-M-14) are increasingly common in areas such as China, South-East Asia, Korea, Japan and Spain.41 A second notable finding is the emerging prevalence of 3GC-R E. coli with acquired AmpC as a cause of BSIs; these were the second most commonly encountered broad-spectrum β-lactamase after CTX-M-type ESBLs in this cohort. Having been previously under-appreciated, p-AmpC enzymes (particularly blaCMY-2) are increasingly recognized as a prominent mediator of resistance in E. coli, and have been described in many parts of the world in both animal- and human-derived strains.42–46 This is of clinical relevance because, despite isolates often testing susceptible in vitro to piperacillin/tazobactam, AmpC (Ambler class C) enzymes may be less efficiently inhibited than class A β-lactamases.5,47 Most clinical laboratories are unable to reliably identify p-AmpC enzymes in 3GC-R Enterobacteriaceae, which may have consequences for the reliability of piperacillin/tazobactam. In this cohort, p-AmpC (mainly blaCMY-2) were not associated with ST131 or any other ST. Previous studies have demonstrated the predominant p-AmpC enzyme amongst E. coli to be CMY-2,42,45,48 with evidence that blaCMY-2 has been mobilized from the Citrobacter freundii chromosome in association with ISEcp1.49 IncF-type plasmids have a host range that is limited to Enterobacteriaceae and contribute to bacterial fitness via antibiotic resistance and virulence determinants.50 These plasmids have been associated with the rapid emergence and global spread of blaCTX-M-15, as well as genes encoding resistance to aminoglycosides and fluoroquinolones [e.g. aac(6′)-Ib-cr, qnr, armA, rmtB].50,51 Similar patterns were also seen in this study, where the majority of plasmids were of IncF type. There was an association between blaCTX-M-15, blaOXA-1, as well as the AMEs aac(3)-IIa and aac(6′)-Ib-cr in clade C2 ST131 carrying IncF plasmids, the majority of which came from patients in Singapore. Previous work, mainly including isolates from North America, suggested that the H30-R/C1 clade of ST131 most commonly carries IncF-type F1:A2:B20 plasmids and the H30-Rx/C2 clade is associated with IncF-type F2:A1:B- plasmids.17–19 In this cohort, plasmid types were associated with different sublineages of ST131. For instance, IncF-type F31:A4:B1 or F36:A4:B1 plasmids were most frequently seen in clade C2, whereas IncF-type F2:A1:B- were only seen in a single clade C1 strain. These variations may reflect sampling from different geographical locations, rather than associations with specific E. coli lineages. This study has some limitations. Enrolment into the clinical trial required susceptibility to piperacillin/tazobactam at the local testing laboratory, and therefore bias may exist in the selection of strains. In addition, as enrolment of patients into a clinical trial may preclude those with severe comorbidities or early mortality (prior to randomization), it is possible that the E. coli were obtained from patients with less severe disease, which may be associated with less virulent strains. The limitations of short-read sequencing data to reconstruct plasmid structures52 are also acknowledged. Conclusions In this analysis of 3GC-R E. coli causing BSI in patients from Singapore, Australia and New Zealand, a clear predominance of fluoroquinolone-resistant clade C1/C2 ST131 was observed. The majority of 3GC-R E. coli carried CTX-M-type ESBLs in association with IncF-type plasmids. The most common clinical source of BSI was community-associated urinary tract infections, consistent with a potential reservoir of these strains outside the healthcare environment. We also observed a significant proportion of non-ST131 strains carrying blaCMY-2, which present additional challenges for laboratory detection and treatment. Acknowledgements We thank all the members of the MERINO study teams from the recruiting sites. Additional members of the MERINO study team Royal Brisbane & Women’s Hospital: Tiffany Harris-Brown, Penelope Lorenc, John McNamara. Princess Alexandria Hospital: Neil Underwood, Jared Eisenmann, James Stewart, Andrew Henderson. National University Hospital: Jaminah Ali, Donald Chiang. Tan Tock Seng Hospital: Soh Siew Hwa, Yvonne Kang, Ong Siew Pei, Ding Ying. North Shore Hospital: Umit Holland. Monash Health: Tony Korman. Funding This project was supported by funding from the Pathology Queensland Study, Education and Research Committee (SERC), the National University Hospital Singapore (NUHS) Clinician Researcher Grant, the Australian Society of Antimicrobials (ASA), the International Society for Chemotherapy (ISC) and the National Health and Medical Research Council (NHMRC) of Australia (GNT1067455). P. N. A. H. is supported by the Royal College of Pathologists of Australasia (RCPA) Foundation Postgraduate Research Fellowship and an Australian Postgraduate Award (APA) from the University of Queensland. S. A. B. is supported by an NHMRC Career Development Fellowship (GNT1090456). M. A. S. is supported by an NHMRC Senior Research Fellowship (GNT1106930). Transparency declarations P. N. A. H. and S. A. B. have spoken at an educational event sponsored by Pfizer. B. R. has consulted for Mayne Pharma and received honoraria for advisory board participation from Merck. P. A. T. has received research support from GSK, Shionogi, Sanofi-Pasteur and Janssen in the last 12 months. D. L. P. has received honoraria for advisory board participation and speaking at events sponsored by Achaogen, Merck and GlaxoSmithKline. All other authors have none to declare. Author contributions P. N. A. H. wrote the first and final drafts and undertook the laboratory work. A. M. W. and H. M. Z. helped with the WGS and N. L. B. Z., L. W. R. and S. A. B. undertook the genomic data analysis. All other authors are site investigators for the trial and helped to recruit patients and collect bacterial isolates. D. L. P. is the chief investigator for the MERINO study and conceived the concept for the paper with P. N. A. H., N. L. B. Z., L. W. R. and S. A. B. All authors contributed to the writing of the paper and have approved the final version. Supplementary data Tables S1–S3 and Figures S1–S7 appear as Supplementary data at JAC Online. References 1 World Health Organisation. Antimicrobial Resistance Global Report on Surveillance . Geneva: WHO, 2014. 2 European Centre for Disease Prevention and Control. Antimicrobial Resistance Surveillance in Europe 2014. Annual Report of the European Antimicrobial Resistance Surveillance Network (EARS-Net) . Stockholm: ECDC, 2015. 3 Centers for Disease Control and Prevention. Antibiotic Resistance Threats in the United States . US Department of Health and Human Services. Atlanta: CDC, 2013. PubMed PubMed  4 Paterson DL, Bonomo RA. Extended-spectrum β-lactamases: a clinical update. Clin Microbiol Rev  2005; 18: 657– 86. Google Scholar CrossRef Search ADS PubMed  5 Jacoby GA. AmpC β-lactamases. Clin Microbiol Rev  2009; 22: 161– 82. Google Scholar CrossRef Search ADS PubMed  6 Jacoby GA, Munoz-Price LS. The new β-lactamases. N Engl J Med  2005; 352: 380– 91. Google Scholar CrossRef Search ADS PubMed  7 Magiorakos AP, Srinivasan A, Carey RB et al.   Multidrug-resistant, extensively drug-resistant and pandrug-resistant bacteria: an international expert proposal for interim standard definitions for acquired resistance. Clin Microbiol Infect  2012; 18: 268– 81. Google Scholar CrossRef Search ADS PubMed  8 Petty NK, Ben Zakour NL, Stanton-Cook M et al.   Global dissemination of a multidrug resistant Escherichia coli clone. Proc Natl Acad Sci U S A  2014; 111: 5694– 9. Google Scholar CrossRef Search ADS PubMed  9 Johnson JR, Johnston B, Clabots C et al.   Escherichia coli sequence type ST131 as the major cause of serious multidrug-resistant E. coli infections in the United States. Clin Infect Dis  2010; 51: 286– 94. Google Scholar CrossRef Search ADS PubMed  10 Rogers BA, Sidjabat HE, Paterson DL. Escherichia coli O25b-ST131: a pandemic, multiresistant, community-associated strain. J Antimicrob Chemother  2011; 66: 1– 14. Google Scholar CrossRef Search ADS PubMed  11 Nicolas-Chanoine MH, Blanco J, Leflon-Guibout V et al.   Intercontinental emergence of Escherichia coli clone O25: H4-ST131 producing CTX-M-15. J Antimicrob Chemother  2008; 61: 273– 81. Google Scholar CrossRef Search ADS PubMed  12 Nicolas-Chanoine MH, Bertrand X, Madec JY. Escherichia coli ST131, an intriguing clonal group. Clin Microbiol Rev  2014; 27: 543– 74. Google Scholar CrossRef Search ADS PubMed  13 Johnson JR, Tchesnokova V, Johnston B et al.   Abrupt emergence of a single dominant multidrug-resistant strain of Escherichia coli. J Infect Dis  2013; 207: 919– 28. Google Scholar CrossRef Search ADS PubMed  14 Price LB, Johnson JR, Aziz M et al.   The epidemic of extended-spectrum-β-lactamase-producing Escherichia coli ST131 is driven by a single highly pathogenic subclone. H30-Rx. MBio  2013; 4: e00377– 13. Google Scholar PubMed  15 Johnson JR, Johnston B, Kuskowski MA et al.   Intensity and mechanisms of fluoroquinolone resistance within the H30 and H30Rx subclones of Escherichia coli sequence type 131 compared with other fluoroquinolone-resistant E. coli. Antimicrob Agents Chemother  2015; 59: 4471– 80. Google Scholar CrossRef Search ADS PubMed  16 Johnson JR, Thuras P, Johnston BD et al.   The pandemic H30 subclone of Escherichia coli sequence type 131 is associated with persistent infections and adverse outcomes independent from its multidrug resistance and associations with compromised hosts. Clin Infect Dis  2016; 62: 1529– 36. Google Scholar CrossRef Search ADS PubMed  17 Johnson TJ, Danzeisen JL, Youmans B et al.   Separate F-type plasmids have shaped the evolution of the H30 subclone of Escherichia coli sequence type 131. mSphere  2016; 1: pii=e00121-16. 18 Phan MD, Forde BM, Peters KM et al.   Molecular characterization of a multidrug resistance IncF plasmid from the globally disseminated Escherichia coli ST131 clone. PLoS One  2015; 10: e0122369. Google Scholar CrossRef Search ADS PubMed  19 Kanamori H, Parobek CM, Juliano JJ et al.   Genomic analysis of multidrug-resistant Escherichia coli from North Carolina community hospitals: ongoing circulation of CTX-M-producing ST131-H30Rx and ST131-H30R1 strains. Antimicrob Agents Chemother  2017; 61: pii=e00912-17. 20 Nordmann P, Dortet L, Poirel L. Carbapenem resistance in Enterobacteriaceae: here is the storm! Trends Mol Med  2012; 18: 263– 72. Google Scholar CrossRef Search ADS PubMed  21 Harris PN, Peleg AY, Iredell J et al.   Meropenem versus piperacillin-tazobactam for definitive treatment of bloodstream infections due to ceftriaxone non-susceptible Escherichia coli and Klebsiella spp (the MERINO trial): study protocol for a randomised controlled trial. Trials  2015; 16: 24. Google Scholar CrossRef Search ADS PubMed  22 Harris PA, Taylor R, Thielke R et al.   Research electronic data capture (REDCap)—a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform  2009; 42: 377– 81. Google Scholar CrossRef Search ADS PubMed  23 Friedman ND, Kaye KS, Stout JE et al.   Health care-associated bloodstream infections in adults: a reason to change the accepted definition of community-acquired infections. Ann Intern Med  2002; 137: 791. Google Scholar CrossRef Search ADS PubMed  24 European Committee on Antimicrobial Susceptibility Testing. Breakpoint Tables for Interpretation of MICs and Zone Diameters, Version 5.0. http://www.eucast.org/fileadmin/src/media/PDFs/EUCAST_files/Breakpoint_tables/v_5.0_Breakpoint_Table_01.pdf. 25 European Committee on Antimicrobial Susceptibility Testing. EUCAST Guidelines for Detection of Resistance Mechanisms and Specific Resistances of Clinical and/or Epidemiological Importance. http://www.eucast.org/fileadmin/src/media/PDFs/EUCAST_files/Resistance_mechanisms/EUCAST_detection_of_resistance_mechanisms_170711.pdf. 26 Seemann T, Bulach D, Kwong J. Nullarbor. http://github.com/tseemann/nullarbor. 27 Carattoli A, Zankari E, Garcia-Fernandez A et al.   In silico detection and typing of plasmids using PlasmidFinder and plasmid multilocus sequence typing. Antimicrob Agents Chemother  2014; 58: 3895– 903. Google Scholar CrossRef Search ADS PubMed  28 Hopkins KL, Davies RH, Threlfall EJ. Mechanisms of quinolone resistance in Escherichia coli and Salmonella: recent developments. Int J Antimicrob Agents  2005; 25: 358– 73. Google Scholar CrossRef Search ADS PubMed  29 Forde BM, Ben Zakour NL, Stanton-Cook M et al.   The complete genome sequence of Escherichia coli EC958: a high quality reference sequence for the globally disseminated multidrug resistant E. coli O25b: H4-ST131 clone. PLoS One  2014; 9: e104400. Google Scholar CrossRef Search ADS PubMed  30 Croucher NJ, Page AJ, Connor TR et al.   Rapid phylogenetic analysis of large samples of recombinant bacterial whole genome sequences using Gubbins. Nucleic Acids Res  2015; 43: e15. Google Scholar CrossRef Search ADS PubMed  31 Stamatakis A. RAxML version 8: a tool for phylogenetic analysis and post-analysis of large phylogenies. Bioinformatics  2014; 30: 1312– 3. Google Scholar CrossRef Search ADS PubMed  32 Clermont O, Christenson JK, Denamur E et al.   The Clermont Escherichia coli phylo-typing method revisited: improvement of specificity and detection of new phylo-groups. Environ Microbiol Rep  2013; 5: 58– 65. Google Scholar CrossRef Search ADS PubMed  33 He Z, Zhang H, Gao S et al.   Evolview v2: an online visualization and management tool for customized and annotated phylogenetic trees. Nucleic Acids Res  2016; 44: W236– 41. Google Scholar CrossRef Search ADS PubMed  34 Chow JW, Yu VL. Combination antibiotic therapy versus monotherapy for Gram-negative bacteraemia: a commentary. Int J Antimicrob Agents  1999; 11: 7– 12. Google Scholar CrossRef Search ADS PubMed  35 Tracz DM, Boyd DA, Hizon R et al.   ampC gene expression in promoter mutants of cefoxitin-resistant Escherichia coli clinical isolates. FEMS Microbiol Lett  2007; 270: 265– 71. Google Scholar CrossRef Search ADS PubMed  36 Guillon H, Tande D, Mammeri H. Emergence of ertapenem resistance in an Escherichia coli clinical isolate producing extended-spectrum β-lactamase AmpC. Antimicrob Agents Chemother  2011; 55: 4443– 6. Google Scholar CrossRef Search ADS PubMed  37 Hooper DC, Jacoby GA. Mechanisms of drug resistance: quinolone resistance. Ann N Y Acad Sci  2015; 1354: 12– 31. Google Scholar CrossRef Search ADS PubMed  38 Canton R, Gonzalez-Alba JM, Galan JC. CTX-M enzymes: origin and diffusion. Front Microbiol  2012; 3: 110. Google Scholar CrossRef Search ADS PubMed  39 Doi Y, Park YS, Rivera JI et al.   Community-associated extended-spectrum β-lactamase-producing Escherichia coli infection in the United States. Clin Infect Dis  2013; 56: 641– 8. Google Scholar CrossRef Search ADS PubMed  40 Stoesser N, Sheppard AE, Pankhurst L et al.   Evolutionary history of the global emergence of the Escherichia coli epidemic clone ST131. MBio  2016; 7: e02162. Google Scholar CrossRef Search ADS PubMed  41 Bevan ER, Jones AM, Hawkey PM. Global epidemiology of CTX-M β-lactamases: temporal and geographical shifts in genotype. J Antimicrob Chemother  2017; 72: 2145– 55. Google Scholar CrossRef Search ADS PubMed  42 Tagg KA, Ginn AN, Jiang X et al.   Distribution of acquired AmpC β-lactamase genes in Sydney, Australia. Diagn Microbiol Infect Dis  2015; 83: 56– 8. Google Scholar CrossRef Search ADS PubMed  43 Pascual V, Alonso N, Simo M et al.   Bloodstream infections caused by Escherichia coli producing AmpC β-lactamases: epidemiology and clinical features. Eur J Clin Microbiol Infect Dis  2016; 35: 1997– 2003. Google Scholar CrossRef Search ADS PubMed  44 Harris PN. Clinical management of infections caused by Enterobacteriaceae that express extended-spectrum β-lactamase and AmpC enzymes. Semin Respir Crit Care Med  2015; 36: 56– 73. Google Scholar CrossRef Search ADS PubMed  45 Drinkovic D, Morris AJ, Dyet K et al.   Plasmid-mediated AmpC β-lactamase-producing Escherichia coli causing urinary tract infection in the Auckland community likely to be resistant to commonly prescribed antimicrobials. N Z Med J  2015; 128: 50– 9. Google Scholar PubMed  46 Belmahdi M, Bakour S, Al Bayssari C et al.   Molecular characterisation of extended-spectrum β-lactamase- and plasmid AmpC-producing Escherichia coli strains isolated from broilers in Bejaia, Algeria. J Glob Antimicrob Resist  2016; 6: 108– 12. Google Scholar CrossRef Search ADS PubMed  47 Harris PN, Ferguson JK. Antibiotic therapy for inducible AmpC β-lactamase-producing Gram-negative bacilli: what are the alternatives to carbapenems, quinolones and aminoglycosides? Int J Antimicrob Agents  2012; 40: 297– 305. Google Scholar CrossRef Search ADS PubMed  48 Sidjabat HE, Seah KY, Coleman L et al.   Expansive spread of IncI1 plasmids carrying blaCMY-2 amongst Escherichia coli. Int J Antimicrob Agents  2014; 44: 203– 8. Google Scholar CrossRef Search ADS PubMed  49 Verdet C, Gautier V, Chachaty E et al.   Genetic context of plasmid-carried blaCMY-2-like genes in Enterobacteriaceae. Antimicrob Agents Chemother  2009; 53: 4002– 6. Google Scholar CrossRef Search ADS PubMed  50 Villa L, Garcia-Fernandez A, Fortini D et al.   Replicon sequence typing of IncF plasmids carrying virulence and resistance determinants. J Antimicrob Chemother  2010; 65: 2518– 29. Google Scholar CrossRef Search ADS PubMed  51 Carattoli A. Resistance plasmid families in Enterobacteriaceae. Antimicrob Agents Chemother  2009; 53: 2227– 38. Google Scholar CrossRef Search ADS PubMed  52 Arredondo-Alonso S, Willems RJ, van Schaik W et al.   On the (im)possibility of reconstructing plasmids from whole-genome short-read sequencing data. Microbial Genomics  2017; Epub online Aug 2017. http://mgen.microbiologyresearch.org/content/journal/mgen/10.1099/mgen.0.000128.v1. © The Author 2017. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved. For Permissions, please email: journals.permissions@oup.com. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Antimicrobial Chemotherapy Oxford University Press

Whole genome analysis of cephalosporin-resistant Escherichia coli from bloodstream infections in Australia, New Zealand and Singapore: high prevalence of CMY-2 producers and ST131 carrying blaCTX-M-15 and blaCTX-M-27

Loading next page...
 
/lp/ou_press/whole-genome-analysis-of-cephalosporin-resistant-escherichia-coli-from-BaEhtcnec4
Publisher
Oxford University Press
Copyright
© The Author 2017. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved. For Permissions, please email: journals.permissions@oup.com.
ISSN
0305-7453
eISSN
1460-2091
D.O.I.
10.1093/jac/dkx466
Publisher site
See Article on Publisher Site

Abstract

Abstract Objectives To characterize MDR Escherichia coli from bloodstream infections (BSIs) in Australia, New Zealand and Singapore. Methods We collected third-generation cephalosporin-resistant (3GC-R) E. coli from blood cultures in patients enrolled in a randomized controlled trial from February 2014 to August 2015. WGS was used to characterize antibiotic resistance genes, MLST, plasmids and phylogenetic relationships. Antibiotic susceptibility was determined using disc diffusion and Etest. Results A total of 70 3GC-R E. coli were included, of which the majority were ST131 (61.4%). BSI was most frequently from a urinary source (69.6%), community associated (62.9%) and in older patients (median age 71 years). The median Pitt score was 1 and ICU admission was infrequent (3.1%). ST131 possessed more acquired resistance genes than non-ST131 (P = 0.003). Clade C1/C2 ST131 predominated (30.2% and 53.5% of ST131, respectively) and these were all ciprofloxacin resistant. All clade A ST131 (n = 6) were community associated. The predominant ESBL types were blaCTX-M (80.0%) and were strongly associated with ST131 (95% carried blaCTX-M), with the majority blaCTX-M-15. Clade C1 was associated with blaCTX-M-14 and blaCTX-M-27, whereas blaCTX-M-15 predominated in clade C2. Plasmid-mediated AmpC genes (mainly blaCMY-2) were frequent (17.1%) but were more common in non-ST131 (P < 0.001) isolates from Singapore and Brisbane. Two strains carried both blaCMY-2 and blaCTX-M. The majority of plasmid replicon types were IncF. Conclusions In a prospective collection of 3GC-R E. coli causing BSI, community-associated Clade C1/C2 ST131 predominate in association with blaCTX-M ESBLs, although a significant proportion of non-ST131 strains carried blaCMY-2. Introduction In recent decades, resistance to β-lactam antibiotics in Enterobacteriaceae has become increasingly common. Of particular concern has been the rising prevalence of resistance to third-generation cephalosporins (3GCs), a class that includes key agents such as ceftriaxone, cefotaxime or ceftazidime.1–3 This phenomenon has largely arisen from the dissemination of genes encoding ESBL or, less frequently, plasmid-mediated AmpC (p-AmpC) enzymes.4–6 These resistance genes are often acquired by plasmid transfer and may be associated with other antibiotic resistance determinants, rendering organisms MDR.7 The global emergence of infections caused by ESBL-producing Escherichia coli, in both the community and hospital setting, has been driven by the acquisition of CTX-M-type ESBL genes (especially blaCTX-M-15) in the successful pandemic clone of E. coli, sequence type 131 (ST131).8–11,E. coli ST131 lineages belong to the B2 phylogenetic subgroup I, and are mostly serotype O25b:H4.12 Within ST131, further sublineages have been delineated according to fimH (type 1 fimbrial adhesin, FimH) alleles, phylogenetic clades (A, B, C1 and C2) and associated resistance genes.8,13 A globally dominant fluoroquinolone-resistant fimH30 subclonal lineage, defined as H30-R (to differentiate from the ancestral fluoroquinolone susceptible H30 strains) or clade C, has been described.8,14,H30-R/clade C strains contain fluoroquinolone resistance mutations in the chromosomal gyrA and parC genes15 and have been associated with poor clinical outcomes.16 Within this sublineage, a pathogenic ST131 subclone containing blaCTX-M-15 has been referred to as H30-Rx14 or clade C2.8 Specific Incompatibility (Inc) F-type plasmids have also been described in association with fluoroquinolone-resistant ST131-H30 clades, with IncF-type F1:A2:B20 plasmids associated with the H30-R/C1 clade and IncF-type F2:A1:B- plasmids associated with the H30-Rx/C2 clade.17–19 Resistance to β-lactams mediated by ESBLs drives the use of broader-spectrum antibiotics such as carbapenems (e.g. meropenem),4 providing selection pressure for carbapenem resistance in Gram-negative bacteria. Of great concern has been the emergence of transmissible carbapenemases in common Enterobacteriaceae.20 As part of an international randomized controlled trial of piperacillin/tazobactam (a potential ‘carbapenem-sparing’ agent) versus meropenem for the treatment of 3GC-resistant (3GC-R) E. coli causing bloodstream infection (BSI),21 we aimed to further characterise these isolates using WGS. Methods Bacterial isolates and clinical data 3GC-R E. coli were collected prospectively during an international multicentre randomized trial comparing treatment options for BSI caused by ceftriaxone non-susceptible E. coli or Klebsiella spp. [the ‘MERINO’ trial: Australian New Zealand Clinical Trials Register (ANZCTR), ref. no. ACTRN12613000532707 and the US National Institute of Health ClinicalTrials.gov register, ref. no. NCT02176122].21 All initial 3GC-R E. coli blood culture isolates (i.e. the first positive culture prompting study inclusion) were included from 70 patients enrolled in the trial for the first 18 months of recruitment (from February 2014 until August 2015) from eight hospital sites in three countries [Australia (n = 21), New Zealand (n = 5) and Singapore (n = 44)]. To be eligible for the trial, patients had to have at least one monomicrobial blood culture growing E. coli, with non-susceptibility to ceftriaxone, ceftazidime or cefotaxime and susceptibility to piperacillin/tazobactam and meropenem, determined by methods used in the local laboratories. All blood culture isolates from patients meeting inclusion criteria who consented to participate in the study were stored at the recruiting site laboratory at −80 °C in cryovials containing glycerol and nutrient broth and later shipped to the coordinating laboratory in Queensland, Australia. For the purposes of this study, only the first 3GC-R E. coli isolated from blood cultures for each enrolled patient were included in the genomic analysis. Relevant clinical data were collected and managed using the REDCap22 electronic data capture tool hosted at the University of Queensland. Healthcare-associated infection (HAI) was defined as a positive blood culture collected in a patient within 48 h of hospital admission, if the patient had received specialized nursing care at home (e.g. intravenous therapy, wound care), attended a hospital, haemodialysis unit, or received intravenous chemotherapy in the 30 days prior to the BSI event, was a resident of a nursing home or long-term care facility or had been admitted to an acute care facility for >48 h in the preceding 90 days.23 HAI also included nosocomial bacteraemia occurring >48 h following hospital admission. BSI was defined as community associated if it occurred within 48 h of admission and the patient did not meet any criteria for HAI.23 Ethics Approval for the study was provided by the Royal Brisbane and Women’s Hospital (ref. HREC/12/QRBW/440), the National Healthcare Group (NHG) Domain Specific Review Board (DSRB) in Singapore (NHG DSRB ref. 2013/00453) and the New Zealand Health and Disability Ethics Committee (ref. 14/NTB/52). Phenotypic susceptibility testing All isolates were tested at the coordinating laboratory against a standard panel of antibiotics used to treat Gram-negative infections by disc diffusion according to EUCAST standards.24 Agents tested included ceftriaxone, ceftazidime, cefepime, cefoxitin, aztreonam, ertapenem, gentamicin, amikacin, ciprofloxacin, co-trimoxazole and amoxicillin/clavulanate. In addition, MICs of piperacillin/tazobactam, meropenem (the two comparator drugs used in the trial) and ceftriaxone were determined by Etest (bioMérieux). ESBL production was confirmed using combination disc testing with ceftriaxone and ceftazidime with and without clavulanate; an increase in zone diameter ≥5 mm with the addition of clavulanate confirmed ESBL production.25 DNA extraction and library preparation After subculture onto LB agar to check for viability and purity, genomic DNA was extracted using the MoBio Ultrapure kit and quantified by spectrophotometry (NanoDrop; ThermoFisher) and fluorometry (Qubit; ThermoFisher). Paired-end DNA libraries were prepared using the Nextera kit (Illumina; Australia) in accordance with the manufacturer’s instructions. WGS WGS was performed in two batches using Illumina HiSeq (100 bp paired end) and MiSeq (300 bp paired end) at the Australian Genome Research Facility (AGRF), University of Queensland, St Lucia. MiSeq raw reads were trimmed conservatively to 150 bp and filtered using Nesoni (version 0.130) to remove Illumina adaptor sequences, reads <80 bp and bases below Phred quality 5 (https://github.com/Victorian-Bioinformatics-Consortium/nesoni). Strains were checked for contamination using Kraken (0.10.5-beta) as implemented through Nullarbor (default settings).26 Resistance gene detection, MLST and plasmid typing Antibiotic resistance genes were detected by using Abricate (version 0.2) with the ResFinder database against SPAdes assemblies (version 3.6.2) as implemented through the pipeline analysis tool Nullarbor (default settings).26 MLST was undertaken using the mlst tool as implemented through Nullarbor. Plasmid replicon typing and plasmid multilocus typing were performed using PlasmidFinder and pMLST.27 Fluoroquinolone resistance SNP detection Filtered reads were mapped to the complete ST131 E. coli reference strain EC958 (GenBank: HG941718.1) using Bowtie as implemented through Nesoni. Non-synonymous mutations were identified using Nesoni nway and manually compared with known mutations in gyrA and parC associated with quinolone resistance.13,28 Phylogenetic analysis Reads for all isolates (n = 70) were mapped to the complete ST131 reference EC958 (GenBank: HG941718.1)29 using Nesoni under default settings. SNPs identified between isolates and the reference EC958 were used to create pseudogenomes for each isolate by substituting the relevant SNPs into the EC958 chromosome using an in-house script. Multiple sequence alignment of the pseudogenomes was used as input for Gubbins (version 1.3.4)30 using the (GTR)GAMMA substitution model to parse recombinant regions. The remaining 211 920 SNPs were used to generate a phylogenetic tree using RAxML (version 8.2.9)31 under the (GTR)GAMMA substitution model with Lewis ascertainment bias correction and a random seed of 456 (100 bootstraps). An ST131-only tree (n = 43) was also created in the same manner using 2248 recombination-free SNPs and 1000 bootstraps. In silico phylo-grouping for all STs was undertaken according to the updated Clermont scheme.32 Phylogenetic trees and associated metadata were visualized using Evolview.33 Sequence data Raw sequence reads and associated metadata have been uploaded to NCBI (Bioproject Accession no. PRJNA398288). A full description of clinical metadata, MLST, acquired resistance genes and plasmid typing is available online at Harvard Dataverse (doi: 10.7910/DVN/YC9WDV). Statistical tests Data describing patient demographics, phenotypic susceptibility, clinical variables and genotypic data for all cases were tabulated, with proportions expressed as percentages and median, mean or IQR calculated as appropriate for scale variables. Categorical variables were compared using Pearson’s χ2 test.Comparisons of mean values in normally distributed data were compared using the t-test. The Mann–Whitney U-test was used for non-parametric data. Statistical analysis was performed using Stata version 13.1 (StataCorp; TX, USA) and graphical images prepared using Prism version 7.0 (GraphPad Software; CA, USA). A P value <0.05 was considered significant. Results Baseline clinical data A total of 70 3GC-R E. coli bloodstream isolates were included. The background clinical and demographic details of enrolled patients are summarized in Table 1. The source of BSI was most frequently the urinary tract (48/70, 69.6%) and infections were mostly community associated (44/70, 62.9%). There was a predominance of patients reporting Chinese ethnicity (38/70, 54.3% of all cases), reflecting the demographics of the largest recruiting sites in Singapore. There was also a greater proportion of male patients (60%), but this was not statistically significant (P = 0.12). Patients tended to be older (median age 71, IQR 64–81 years, range 20–94 years), although only a small proportion (5.8%) were admitted from nursing homes. The majority of patients had less severe acute illness (median Pitt score 1, IQR 0–2, range 0–3; where a score ≥4 reflects the presence of critical illness with high mortality34) and relatively low comorbidity scores (Charlson score median 2, IQR 1–4, range 0–11) and were infrequently admitted to the ICU (3.1%). Table 1. Baseline clinical and demographic variables Factor/level  No. (%) (N = 70)  Region     Singapore  44 (63)   Brisbane (AUS)  13 (19)   Melbourne (AUS)  8 (11)   Auckland (NZ)  5 (7)  Age, years, mean (SD)  69.9 (15.2)  Gender     female  28 (40)   male  42 (60)  Source of bacteraemia     urinary tract infection  48 (70)   intra-abdominal infection  8 (12)   pneumonia  1 (1)   other  9 (13)   unknown  3 (4)  Acquisition     community associated  44 (63)   healthcare associated  26 (37)  Pitt score, median (IQR)  1 (0–2)  Charlson score, median (IQR)  2 (1–4)  Any CTX-M ESBL  56 (80)  AmpC β-lactamase  12 (17)  Surgery within 14 days  5 (7)  Central venous catheter  6 (9)  Immune suppression  10 (14)  ICU admission  2 (3)  Nursing home resident  4 (6)  Factor/level  No. (%) (N = 70)  Region     Singapore  44 (63)   Brisbane (AUS)  13 (19)   Melbourne (AUS)  8 (11)   Auckland (NZ)  5 (7)  Age, years, mean (SD)  69.9 (15.2)  Gender     female  28 (40)   male  42 (60)  Source of bacteraemia     urinary tract infection  48 (70)   intra-abdominal infection  8 (12)   pneumonia  1 (1)   other  9 (13)   unknown  3 (4)  Acquisition     community associated  44 (63)   healthcare associated  26 (37)  Pitt score, median (IQR)  1 (0–2)  Charlson score, median (IQR)  2 (1–4)  Any CTX-M ESBL  56 (80)  AmpC β-lactamase  12 (17)  Surgery within 14 days  5 (7)  Central venous catheter  6 (9)  Immune suppression  10 (14)  ICU admission  2 (3)  Nursing home resident  4 (6)  Antibiotic resistance phenotypes By MIC testing, 97.1% (68/70) were susceptible to piperacillin/tazobactam (median 2 mg/L, range 1–24 mg/L, IQR 1.5–4; EUCAST breakpoint for susceptibility ≤8 mg/L24) (see Figure S1, available as Supplementary data at JAC Online). All strains were susceptible to meropenem (MIC90 0.047 mg/L; median 0.023 mg/L, range 0.012–0.19 mg/L; EUCAST breakpoint for susceptibility ≤2 mg/L24), although one strain (MER-86) was non-susceptible to ertapenem. The majority (90.1%) demonstrated ceftriaxone MICs ≥32 mg/L (range 0.064 to ≥32 mg/L; median ≥32 mg/L, MIC90 and MIC50 ≥32 mg/L). Two strains, which were susceptible to ceftriaxone by MIC, were resistant to ceftazidime [one (MER-15) contained blaDHA-1, the other (MER-100) had changes to the ampC promotor region, discussed in detail later]. Phenotypic resistance to 3GCs could not be detected in one strain (MER-34) when retested in the coordinating laboratory, and was found to only possess TEM-176 (a narrow-spectrum β-lactamase) by sequencing. Resistance to gentamicin was common (25/70, 35.7%), although all strains remained susceptible to amikacin. Strains demonstrated a variable antibiogram according to ST131 clade (see Table 2), but were frequently resistant to trimethoprim/sulfamethoxazole (46/70, 65.7%) or fluoroquinolones (52/70, 74.3%). There was universal resistance to ciprofloxacin in clade C1/C2 ST131 (n = 36), compared with only 50% (3/6) and 48.2% (13/27) in clade A and non-ST131 strains, respectively (P < 0.001). Table 2. Antibiotic resistance profile of E. coli strains according to ST131 clade     Non-susceptible, n (%)   ST131 clade    CTX  CAZ  FEP  FOX  AMC  TZP  MEM  ETP  ATM  SXT  GEN  AMK  CIP  A  6  5 (83)  5 (83)  5 (83)  1 (17)  4 (67)  0 (0)  0 (0)  0 (0)  5 (83)  4 (67)  3 (50)  0 (0)  3 (50)  B  1  1 (100)  1 (100)  1 (100)  0 (0)  1 (100)  0 (0)  0 (0)  0 (0)  1 (100)  1 (100)  1 (100)  0 (0)  0 (0)  C1  13  13 (100)  4 (31)  12 (92)  1 (8)  5 (38)  0 (0)  0 (0)  0 (0)  12 (92)  8 (62)  4 (31)  0 (0)  13 (100)  C2  23  23 (100)  21 (91)  22 (96)  1 (4)  21 (91)  2 (9)  0 (0)  0 (0)  23 (100)  19 (83)  13 (57)  0 (0)  23 (100)  Non-ST131  27  25 (93)  22 (81)  14 (52)  10 (37)  20 (74)  0 (0)  0 (0)  1 (4)  21 (78)  14 (52)  4 (15)  0 (0)  13 (48)  All  70  67 (96)  53 (76)  54 (77)  13 (19)  51 (73)  2 (3)  0 (0)  1 (1)  62 (89)  46 (66)  25 (36)  0 (0)  52 (74)      Non-susceptible, n (%)   ST131 clade    CTX  CAZ  FEP  FOX  AMC  TZP  MEM  ETP  ATM  SXT  GEN  AMK  CIP  A  6  5 (83)  5 (83)  5 (83)  1 (17)  4 (67)  0 (0)  0 (0)  0 (0)  5 (83)  4 (67)  3 (50)  0 (0)  3 (50)  B  1  1 (100)  1 (100)  1 (100)  0 (0)  1 (100)  0 (0)  0 (0)  0 (0)  1 (100)  1 (100)  1 (100)  0 (0)  0 (0)  C1  13  13 (100)  4 (31)  12 (92)  1 (8)  5 (38)  0 (0)  0 (0)  0 (0)  12 (92)  8 (62)  4 (31)  0 (0)  13 (100)  C2  23  23 (100)  21 (91)  22 (96)  1 (4)  21 (91)  2 (9)  0 (0)  0 (0)  23 (100)  19 (83)  13 (57)  0 (0)  23 (100)  Non-ST131  27  25 (93)  22 (81)  14 (52)  10 (37)  20 (74)  0 (0)  0 (0)  1 (4)  21 (78)  14 (52)  4 (15)  0 (0)  13 (48)  All  70  67 (96)  53 (76)  54 (77)  13 (19)  51 (73)  2 (3)  0 (0)  1 (1)  62 (89)  46 (66)  25 (36)  0 (0)  52 (74)  CTX, ceftriaxone; CAZ, ceftazidime; FEP, cefepime; FOX, cefoxitin; AMC, amoxicillin/clavulanate; TZP, piperacillin/tazobactam; MEM, meropenem; ETP, ertapenem; ATM, aztreonam; SXT, trimethoprim/sulphamethoxazole; GEN, gentamicin; AMK, amikacin; CIP, ciprofloxacin. MLST and ST131 phylogeny A clear majority of strains were ST131 by in silico MLST (43/70, 61.4%), with other strains broadly distributed across a number of other STs (Figure 1). ST131 was prevalent in all city locations (Singapore, Melbourne, Brisbane and Auckland). The only STs with more than a single representative strain were ST1193, ST410, ST38, ST69 and ST963. The majority of ST131 strains belonged to clades C1 (30.2%) or C2 (53.5%), with strains from clades B (2.3%) and A (14.0%) seen less frequently. Figure 1. View largeDownload slide In silico MLST of ESBL- or AmpC-producing E. coli isolated from blood, by region. Figure 1. View largeDownload slide In silico MLST of ESBL- or AmpC-producing E. coli isolated from blood, by region. A phylogenetic tree of all 3GC-R E. coli strains can be found in Figure S2. All clade A ST131 were community associated, with a mixture of community- and healthcare-associated infections observed for strains in clades C1 and C2. There was evidence of clustering of closely related paired strains (with <10 core genome SNPs different) within certain hospitals (e.g. MER-27/25 in Hospital E; MER-8/10 and MER-37/39 in Hospital A; Figure 2), indicating a direct relationship or very recent transmission, although these represented both community- and healthcare-associated infections. Figure 2. View largeDownload slide Phylogenetic tree of ST131 E. coli based on core genome SNPs. Clade, antibiotic resistance, ESBL/p-AmpC-type and IncF plasmid type are shown. EC958 genome (NZ HG941718.1) used as reference.29 Figure 2. View largeDownload slide Phylogenetic tree of ST131 E. coli based on core genome SNPs. Clade, antibiotic resistance, ESBL/p-AmpC-type and IncF plasmid type are shown. EC958 genome (NZ HG941718.1) used as reference.29 Two of the Australian strains (MER-2 and MER-4) were both ST963, with chromosomally integrated blaCMY-2, and occurred in patients whose admissions overlapped in the same hospital. However, they were separated by >350 SNPs, suggesting a more distant common ancestor. Antibiotic resistance genes The median number of acquired antibiotic resistance genes detected for each isolate was 9. One strain (MER-90) possessed a total of 17 acquired resistance genes, including β-lactamases (blaCMY-2, blaTEM-1B), aminoglycoside resistance genes [aac(3)-IId-like, aadA1-like, aadA2, aph(3′)-Ic-like, strA, strB], resistance genes related to folate metabolism (dfrA12, dfrA14-like), fluoroquinolones (qnrS1), sulphonamides (sul1, sul2, sul3), tetracyclines [tet(A)], phenicols (floR-like) and macrolides [mph(A)]. The number of acquired antibiotic resistance genes was significantly greater in ST131 than non-ST131 strains (P = 0.003) (Figure S3A). However, the number of resistance genes did not vary significantly across ST131 clades (Figure S3C). No carbapenemase genes were identified. In one strain (MER-100), resistance to ceftazidime (but not ceftriaxone) was not clearly explained by resistance gene profiling, with no β-lactamase genes identified, although an altered −35 box (TTGACA) was found in its promoter, which has been associated with overexpression of the ampC promoter.35 A single strain (MER-86) demonstrated resistance to ertapenem and was found to have disruption in ompF, which has been associated with reduced susceptibility to ertapenem when associated with broad-spectrum β-lactamases.36 β-Lactamases The predominant ESBL genes identified were blaCTX-M, seen in 80.0% (56/70) of isolates. The presence of blaCTX-M was strongly associated with ST131; 95% of ST131 possessed blaCTX-M ESBLs, compared with only 56% of non-ST131 (P < 0.001) (Table S1). These were either from CTX-M group 9 [blaCTX-M-14 (n = 8),blaCTX-M-27 (n = 14)] or CTX-M group 1 [blaCTX-M-15 (n = 31),blaCTX-M-55 (n = 3)], with the majority blaCTX-M-15 (44.3%). Clade C1 ST131 (n = 13) were associated with blaCTX-M-14 (n = 4) and blaCTX-M-27 (n = 9), whereas blaCTX-M-15 predominated in clade C2 (95.7%; 22/23) (Figure S3B). ESBL genes found in non-ST131 strains included blaCTX-M-15 (n = 7; ST12, ST95, ST648, ST973 and ST1193), blaCTX-M-14 (n = 3; ST10, ST69 and ST4702), and blaCTX-M-27 (n = 1; ST10). The second most common group of β-lactamases with the ability to hydrolyse 3GCs were acquired AmpC β-lactamase genes, found in 17.1% (12/70). Acquired AmpC-producers were most frequently seen in strains from Brisbane (5/13, 38%), and Singapore (7/44, 15.9%), but none was detected in a smaller sample of patients from Melbourne or Auckland. The presence of acquired AmpC was not clearly associated with specific STs, but was more common in non-ST131 strains (37.0% versus 4.7%; P < 0.001). These were predominantly blaCMY-2, although a single strain carried blaDHA-1. A single clade C2 ST131 strain from Singapore possessed blaCMY-2 (in association with blaCTX-M-15) (Figure 2). The genetic context of blaCMY-2 in five strains from Singapore showed homology with a plasmid (GenBank: LC019731) identified in Klebsiella pneumoniae from Taiwan (Figure S4). blaCMY-2 located within an IncI1 plasmid backbone from two strains isolated in Australia was similar to a plasmid (pS10584; GenBank: KX058576.1) identified in Salmonella enterica from a food source in China (Figure S5). ISEcp1 was identified in all but two of the E. coli strains carrying blaCMY-2, with these associated with IS1294 and a truncated ISEcp1 (Figure S5). Although blaCMY-2 are usually acquired on plasmids, in four strains {three from Australia [MER-2 (ST963), MER-4 (ST963), MER-43 (ST38)] and one from Singapore [MER-99 (ST93)]} there was evidence to suggest chromosomal integration based on flanking genes (Figure S6). However, owing to repetitive regions surrounding the blaCMY-2 region, the complete context could not be ascertained in all isolates using short-read sequencing alone. A single strain (MER-89) possessed blaLAP-2, and two (MER-86 and MER-110) carried both blaCTX-M and blaCMY. Further details of the genetic context of blaCMY-2 and blaDHA-1 can be found in Figures S4–S7. Fluoroquinolone resistance genes Acquired quinolone resistance genes (i.e. those not mediated by SNPs in regions associated with quinolone resistance) were seen in 11% (8/70) of strains. These genes included qnrS1, qnrB4, qnrB66-like, oqxA or aac(6′)Ib-cr. The presence of these genes was more commonly seen in non-ST131 than ST131 (22% versus 5%; P = 0.025). In some strains (e.g. MER-34, MER-26) phenotypic ciprofloxacin resistance was not evident despite the presence of acquired resistance genes, although these genes usually only confer low-level fluoroquinolone resistance in the absence of additional mechanisms. All clade C strains [and a single clade A strain (MER-42)] were identical to the EC958 reference strain with respect to mutations in parC and gyrA (Tables S2 and S3). Phenotypic ciprofloxacin resistance was largely congruent with the presence of SNPs in parC and gyrA known to be associated with fluoroquinolone resistance, or the presence of acquired quinolone-resistance determinants.37 Certain gyrA SNPs (e.g. 83L) were not by themselves associated with phenotypic ciprofloxacin resistance unless accompanied by additional SNPs (e.g. 87N or 87Y) (Table S3). Sulphonamide and folate pathway resistance genes Sulphonamide resistance genes (sul1, sul2 or sul3) were common, and present in 69% (48/70) of strains, as were folate synthesis pathway (e.g. trimethoprim) resistance genes (54%, 38/70), such as dfrA1, dfrA7, dfrA12, dfr14 and dfrA17. The presence of sulphonamide resistance and trimethoprim resistance genes was more common in ST131 (81% versus 48%, P = 0.004, and 74% versus 22%, P < 0.001, respectively). Other resistance genes The presence of aminoglycoside-modifying enzymes (AMEs) was common (seen in 76% of strains) and was encountered more frequently in ST131 (86% versus 59%; P = 0.011). Genes mediating resistance to tetracyclines [specifically tet(A) and tet(B)] were seen in 56% (39/70) of strains, but were equally distributed between ST131 and non-ST131. Other frequently identified resistance genes included those mediating resistance to chloramphenicol (e.g. catA, florR) or macrolides [e.g. mph(A)]. Plasmids The majority of plasmid replicon types were identified as IncF. IncF-type plasmids were closely associated with blaCTX-M. Almost all (53/55, 96.4%) E. coli with blaCTX-M also carried IncF-type plasmids, whereas only 53.3% of strains without blaCTX-M carried plasmids typed as IncF. According to the PubMLST scheme (www.pubmlst.org/plasmid), plasmids seen in clade C1 ST131 were mainly IncF plasmid type F1:A2:B20 (76.9%), with the remainder IncF-type F1:A2:B- or IncI1 types (ST79 or unknown ST). Amongst clade C2 ST131, IncF-types F31:A4:B1 or F36:A4:B1 were most common (22.7% and 27.3%, respectively), with IncF-type F2:A1:B- plasmids only seen in 18.2% (Figure 2). Only three clade C2 strains contained IncI1 or IncN plasmids. The presence of blaCMY-2 was associated with IncI1-type plasmids (unless chromosomally integrated). In a single strain, blaDHA-1 was located within an IncF-type plasmid (Figure S7). However, given the limitations of reconstructing plasmids from short-read data alone, the location of β-lactamase genes within specific plasmids remains uncertain. A full summary of plasmid replicon types can be found in the Supplementary data. Discussion This prospective collection of ESBL- and p-AmpC-producing E. coli bloodstream isolates provides insight into the current clinical and molecular epidemiology of these infections within Australia, New Zealand and Singapore. The clear predominance of ST131 carrying CTX-M-type ESBLs is striking and reflects how this pandemic clone has emerged as a highly successful human pathogen. As has been described elsewhere, CTX-M-type ESBLs have now displaced TEM- or SHV-type ESBLs in many parts of the world,10,38 and the latter were not seen in this contemporary collection of E. coli bloodstream isolates. It is also notable that the majority of cases were community associated, with their origin from the urinary tract and in patients over the age of 65 years. An increasing proportion of infections caused by ESBL-producing E. coli are reported as being community acquired.39 Our study would suggest a community reservoir for 3GC-R E. coli, including CTX-M-producing ST131 (65% of which were associated with BSI in patients without any recent healthcare exposure). This contrasts to previous decades, following the first description of ESBLs, where nosocomial acquisition was common and TEM- and SHV-type EBSLs predominated.4 Different β-lactamase genes were associated with certain E. coli lineages; blaCTX-M-15 was largely restricted to the C2 clade amongst ST131, whereas blaCTX-M-27 and blaCTX-M-14 were found in clade C1, which is consistent with previous studies.8,40 Within ST131, clades C1/H30-R and C2/H30-Rx are globally disseminated, and have not demonstrated clear geographical clustering,8 although within some regions limited clonal expansion has occurred.40 The global prevalence of blaCTX-M-15 has increased in most countries over recent years, and dominates in most regions, although group 9 variants (including blaCTX-M-14) are increasingly common in areas such as China, South-East Asia, Korea, Japan and Spain.41 A second notable finding is the emerging prevalence of 3GC-R E. coli with acquired AmpC as a cause of BSIs; these were the second most commonly encountered broad-spectrum β-lactamase after CTX-M-type ESBLs in this cohort. Having been previously under-appreciated, p-AmpC enzymes (particularly blaCMY-2) are increasingly recognized as a prominent mediator of resistance in E. coli, and have been described in many parts of the world in both animal- and human-derived strains.42–46 This is of clinical relevance because, despite isolates often testing susceptible in vitro to piperacillin/tazobactam, AmpC (Ambler class C) enzymes may be less efficiently inhibited than class A β-lactamases.5,47 Most clinical laboratories are unable to reliably identify p-AmpC enzymes in 3GC-R Enterobacteriaceae, which may have consequences for the reliability of piperacillin/tazobactam. In this cohort, p-AmpC (mainly blaCMY-2) were not associated with ST131 or any other ST. Previous studies have demonstrated the predominant p-AmpC enzyme amongst E. coli to be CMY-2,42,45,48 with evidence that blaCMY-2 has been mobilized from the Citrobacter freundii chromosome in association with ISEcp1.49 IncF-type plasmids have a host range that is limited to Enterobacteriaceae and contribute to bacterial fitness via antibiotic resistance and virulence determinants.50 These plasmids have been associated with the rapid emergence and global spread of blaCTX-M-15, as well as genes encoding resistance to aminoglycosides and fluoroquinolones [e.g. aac(6′)-Ib-cr, qnr, armA, rmtB].50,51 Similar patterns were also seen in this study, where the majority of plasmids were of IncF type. There was an association between blaCTX-M-15, blaOXA-1, as well as the AMEs aac(3)-IIa and aac(6′)-Ib-cr in clade C2 ST131 carrying IncF plasmids, the majority of which came from patients in Singapore. Previous work, mainly including isolates from North America, suggested that the H30-R/C1 clade of ST131 most commonly carries IncF-type F1:A2:B20 plasmids and the H30-Rx/C2 clade is associated with IncF-type F2:A1:B- plasmids.17–19 In this cohort, plasmid types were associated with different sublineages of ST131. For instance, IncF-type F31:A4:B1 or F36:A4:B1 plasmids were most frequently seen in clade C2, whereas IncF-type F2:A1:B- were only seen in a single clade C1 strain. These variations may reflect sampling from different geographical locations, rather than associations with specific E. coli lineages. This study has some limitations. Enrolment into the clinical trial required susceptibility to piperacillin/tazobactam at the local testing laboratory, and therefore bias may exist in the selection of strains. In addition, as enrolment of patients into a clinical trial may preclude those with severe comorbidities or early mortality (prior to randomization), it is possible that the E. coli were obtained from patients with less severe disease, which may be associated with less virulent strains. The limitations of short-read sequencing data to reconstruct plasmid structures52 are also acknowledged. Conclusions In this analysis of 3GC-R E. coli causing BSI in patients from Singapore, Australia and New Zealand, a clear predominance of fluoroquinolone-resistant clade C1/C2 ST131 was observed. The majority of 3GC-R E. coli carried CTX-M-type ESBLs in association with IncF-type plasmids. The most common clinical source of BSI was community-associated urinary tract infections, consistent with a potential reservoir of these strains outside the healthcare environment. We also observed a significant proportion of non-ST131 strains carrying blaCMY-2, which present additional challenges for laboratory detection and treatment. Acknowledgements We thank all the members of the MERINO study teams from the recruiting sites. Additional members of the MERINO study team Royal Brisbane & Women’s Hospital: Tiffany Harris-Brown, Penelope Lorenc, John McNamara. Princess Alexandria Hospital: Neil Underwood, Jared Eisenmann, James Stewart, Andrew Henderson. National University Hospital: Jaminah Ali, Donald Chiang. Tan Tock Seng Hospital: Soh Siew Hwa, Yvonne Kang, Ong Siew Pei, Ding Ying. North Shore Hospital: Umit Holland. Monash Health: Tony Korman. Funding This project was supported by funding from the Pathology Queensland Study, Education and Research Committee (SERC), the National University Hospital Singapore (NUHS) Clinician Researcher Grant, the Australian Society of Antimicrobials (ASA), the International Society for Chemotherapy (ISC) and the National Health and Medical Research Council (NHMRC) of Australia (GNT1067455). P. N. A. H. is supported by the Royal College of Pathologists of Australasia (RCPA) Foundation Postgraduate Research Fellowship and an Australian Postgraduate Award (APA) from the University of Queensland. S. A. B. is supported by an NHMRC Career Development Fellowship (GNT1090456). M. A. S. is supported by an NHMRC Senior Research Fellowship (GNT1106930). Transparency declarations P. N. A. H. and S. A. B. have spoken at an educational event sponsored by Pfizer. B. R. has consulted for Mayne Pharma and received honoraria for advisory board participation from Merck. P. A. T. has received research support from GSK, Shionogi, Sanofi-Pasteur and Janssen in the last 12 months. D. L. P. has received honoraria for advisory board participation and speaking at events sponsored by Achaogen, Merck and GlaxoSmithKline. All other authors have none to declare. Author contributions P. N. A. H. wrote the first and final drafts and undertook the laboratory work. A. M. W. and H. M. Z. helped with the WGS and N. L. B. Z., L. W. R. and S. A. B. undertook the genomic data analysis. All other authors are site investigators for the trial and helped to recruit patients and collect bacterial isolates. D. L. P. is the chief investigator for the MERINO study and conceived the concept for the paper with P. N. A. H., N. L. B. Z., L. W. R. and S. A. B. All authors contributed to the writing of the paper and have approved the final version. Supplementary data Tables S1–S3 and Figures S1–S7 appear as Supplementary data at JAC Online. References 1 World Health Organisation. Antimicrobial Resistance Global Report on Surveillance . Geneva: WHO, 2014. 2 European Centre for Disease Prevention and Control. Antimicrobial Resistance Surveillance in Europe 2014. Annual Report of the European Antimicrobial Resistance Surveillance Network (EARS-Net) . Stockholm: ECDC, 2015. 3 Centers for Disease Control and Prevention. Antibiotic Resistance Threats in the United States . US Department of Health and Human Services. Atlanta: CDC, 2013. PubMed PubMed  4 Paterson DL, Bonomo RA. Extended-spectrum β-lactamases: a clinical update. Clin Microbiol Rev  2005; 18: 657– 86. Google Scholar CrossRef Search ADS PubMed  5 Jacoby GA. AmpC β-lactamases. Clin Microbiol Rev  2009; 22: 161– 82. Google Scholar CrossRef Search ADS PubMed  6 Jacoby GA, Munoz-Price LS. The new β-lactamases. N Engl J Med  2005; 352: 380– 91. Google Scholar CrossRef Search ADS PubMed  7 Magiorakos AP, Srinivasan A, Carey RB et al.   Multidrug-resistant, extensively drug-resistant and pandrug-resistant bacteria: an international expert proposal for interim standard definitions for acquired resistance. Clin Microbiol Infect  2012; 18: 268– 81. Google Scholar CrossRef Search ADS PubMed  8 Petty NK, Ben Zakour NL, Stanton-Cook M et al.   Global dissemination of a multidrug resistant Escherichia coli clone. Proc Natl Acad Sci U S A  2014; 111: 5694– 9. Google Scholar CrossRef Search ADS PubMed  9 Johnson JR, Johnston B, Clabots C et al.   Escherichia coli sequence type ST131 as the major cause of serious multidrug-resistant E. coli infections in the United States. Clin Infect Dis  2010; 51: 286– 94. Google Scholar CrossRef Search ADS PubMed  10 Rogers BA, Sidjabat HE, Paterson DL. Escherichia coli O25b-ST131: a pandemic, multiresistant, community-associated strain. J Antimicrob Chemother  2011; 66: 1– 14. Google Scholar CrossRef Search ADS PubMed  11 Nicolas-Chanoine MH, Blanco J, Leflon-Guibout V et al.   Intercontinental emergence of Escherichia coli clone O25: H4-ST131 producing CTX-M-15. J Antimicrob Chemother  2008; 61: 273– 81. Google Scholar CrossRef Search ADS PubMed  12 Nicolas-Chanoine MH, Bertrand X, Madec JY. Escherichia coli ST131, an intriguing clonal group. Clin Microbiol Rev  2014; 27: 543– 74. Google Scholar CrossRef Search ADS PubMed  13 Johnson JR, Tchesnokova V, Johnston B et al.   Abrupt emergence of a single dominant multidrug-resistant strain of Escherichia coli. J Infect Dis  2013; 207: 919– 28. Google Scholar CrossRef Search ADS PubMed  14 Price LB, Johnson JR, Aziz M et al.   The epidemic of extended-spectrum-β-lactamase-producing Escherichia coli ST131 is driven by a single highly pathogenic subclone. H30-Rx. MBio  2013; 4: e00377– 13. Google Scholar PubMed  15 Johnson JR, Johnston B, Kuskowski MA et al.   Intensity and mechanisms of fluoroquinolone resistance within the H30 and H30Rx subclones of Escherichia coli sequence type 131 compared with other fluoroquinolone-resistant E. coli. Antimicrob Agents Chemother  2015; 59: 4471– 80. Google Scholar CrossRef Search ADS PubMed  16 Johnson JR, Thuras P, Johnston BD et al.   The pandemic H30 subclone of Escherichia coli sequence type 131 is associated with persistent infections and adverse outcomes independent from its multidrug resistance and associations with compromised hosts. Clin Infect Dis  2016; 62: 1529– 36. Google Scholar CrossRef Search ADS PubMed  17 Johnson TJ, Danzeisen JL, Youmans B et al.   Separate F-type plasmids have shaped the evolution of the H30 subclone of Escherichia coli sequence type 131. mSphere  2016; 1: pii=e00121-16. 18 Phan MD, Forde BM, Peters KM et al.   Molecular characterization of a multidrug resistance IncF plasmid from the globally disseminated Escherichia coli ST131 clone. PLoS One  2015; 10: e0122369. Google Scholar CrossRef Search ADS PubMed  19 Kanamori H, Parobek CM, Juliano JJ et al.   Genomic analysis of multidrug-resistant Escherichia coli from North Carolina community hospitals: ongoing circulation of CTX-M-producing ST131-H30Rx and ST131-H30R1 strains. Antimicrob Agents Chemother  2017; 61: pii=e00912-17. 20 Nordmann P, Dortet L, Poirel L. Carbapenem resistance in Enterobacteriaceae: here is the storm! Trends Mol Med  2012; 18: 263– 72. Google Scholar CrossRef Search ADS PubMed  21 Harris PN, Peleg AY, Iredell J et al.   Meropenem versus piperacillin-tazobactam for definitive treatment of bloodstream infections due to ceftriaxone non-susceptible Escherichia coli and Klebsiella spp (the MERINO trial): study protocol for a randomised controlled trial. Trials  2015; 16: 24. Google Scholar CrossRef Search ADS PubMed  22 Harris PA, Taylor R, Thielke R et al.   Research electronic data capture (REDCap)—a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform  2009; 42: 377– 81. Google Scholar CrossRef Search ADS PubMed  23 Friedman ND, Kaye KS, Stout JE et al.   Health care-associated bloodstream infections in adults: a reason to change the accepted definition of community-acquired infections. Ann Intern Med  2002; 137: 791. Google Scholar CrossRef Search ADS PubMed  24 European Committee on Antimicrobial Susceptibility Testing. Breakpoint Tables for Interpretation of MICs and Zone Diameters, Version 5.0. http://www.eucast.org/fileadmin/src/media/PDFs/EUCAST_files/Breakpoint_tables/v_5.0_Breakpoint_Table_01.pdf. 25 European Committee on Antimicrobial Susceptibility Testing. EUCAST Guidelines for Detection of Resistance Mechanisms and Specific Resistances of Clinical and/or Epidemiological Importance. http://www.eucast.org/fileadmin/src/media/PDFs/EUCAST_files/Resistance_mechanisms/EUCAST_detection_of_resistance_mechanisms_170711.pdf. 26 Seemann T, Bulach D, Kwong J. Nullarbor. http://github.com/tseemann/nullarbor. 27 Carattoli A, Zankari E, Garcia-Fernandez A et al.   In silico detection and typing of plasmids using PlasmidFinder and plasmid multilocus sequence typing. Antimicrob Agents Chemother  2014; 58: 3895– 903. Google Scholar CrossRef Search ADS PubMed  28 Hopkins KL, Davies RH, Threlfall EJ. Mechanisms of quinolone resistance in Escherichia coli and Salmonella: recent developments. Int J Antimicrob Agents  2005; 25: 358– 73. Google Scholar CrossRef Search ADS PubMed  29 Forde BM, Ben Zakour NL, Stanton-Cook M et al.   The complete genome sequence of Escherichia coli EC958: a high quality reference sequence for the globally disseminated multidrug resistant E. coli O25b: H4-ST131 clone. PLoS One  2014; 9: e104400. Google Scholar CrossRef Search ADS PubMed  30 Croucher NJ, Page AJ, Connor TR et al.   Rapid phylogenetic analysis of large samples of recombinant bacterial whole genome sequences using Gubbins. Nucleic Acids Res  2015; 43: e15. Google Scholar CrossRef Search ADS PubMed  31 Stamatakis A. RAxML version 8: a tool for phylogenetic analysis and post-analysis of large phylogenies. Bioinformatics  2014; 30: 1312– 3. Google Scholar CrossRef Search ADS PubMed  32 Clermont O, Christenson JK, Denamur E et al.   The Clermont Escherichia coli phylo-typing method revisited: improvement of specificity and detection of new phylo-groups. Environ Microbiol Rep  2013; 5: 58– 65. Google Scholar CrossRef Search ADS PubMed  33 He Z, Zhang H, Gao S et al.   Evolview v2: an online visualization and management tool for customized and annotated phylogenetic trees. Nucleic Acids Res  2016; 44: W236– 41. Google Scholar CrossRef Search ADS PubMed  34 Chow JW, Yu VL. Combination antibiotic therapy versus monotherapy for Gram-negative bacteraemia: a commentary. Int J Antimicrob Agents  1999; 11: 7– 12. Google Scholar CrossRef Search ADS PubMed  35 Tracz DM, Boyd DA, Hizon R et al.   ampC gene expression in promoter mutants of cefoxitin-resistant Escherichia coli clinical isolates. FEMS Microbiol Lett  2007; 270: 265– 71. Google Scholar CrossRef Search ADS PubMed  36 Guillon H, Tande D, Mammeri H. Emergence of ertapenem resistance in an Escherichia coli clinical isolate producing extended-spectrum β-lactamase AmpC. Antimicrob Agents Chemother  2011; 55: 4443– 6. Google Scholar CrossRef Search ADS PubMed  37 Hooper DC, Jacoby GA. Mechanisms of drug resistance: quinolone resistance. Ann N Y Acad Sci  2015; 1354: 12– 31. Google Scholar CrossRef Search ADS PubMed  38 Canton R, Gonzalez-Alba JM, Galan JC. CTX-M enzymes: origin and diffusion. Front Microbiol  2012; 3: 110. Google Scholar CrossRef Search ADS PubMed  39 Doi Y, Park YS, Rivera JI et al.   Community-associated extended-spectrum β-lactamase-producing Escherichia coli infection in the United States. Clin Infect Dis  2013; 56: 641– 8. Google Scholar CrossRef Search ADS PubMed  40 Stoesser N, Sheppard AE, Pankhurst L et al.   Evolutionary history of the global emergence of the Escherichia coli epidemic clone ST131. MBio  2016; 7: e02162. Google Scholar CrossRef Search ADS PubMed  41 Bevan ER, Jones AM, Hawkey PM. Global epidemiology of CTX-M β-lactamases: temporal and geographical shifts in genotype. J Antimicrob Chemother  2017; 72: 2145– 55. Google Scholar CrossRef Search ADS PubMed  42 Tagg KA, Ginn AN, Jiang X et al.   Distribution of acquired AmpC β-lactamase genes in Sydney, Australia. Diagn Microbiol Infect Dis  2015; 83: 56– 8. Google Scholar CrossRef Search ADS PubMed  43 Pascual V, Alonso N, Simo M et al.   Bloodstream infections caused by Escherichia coli producing AmpC β-lactamases: epidemiology and clinical features. Eur J Clin Microbiol Infect Dis  2016; 35: 1997– 2003. Google Scholar CrossRef Search ADS PubMed  44 Harris PN. Clinical management of infections caused by Enterobacteriaceae that express extended-spectrum β-lactamase and AmpC enzymes. Semin Respir Crit Care Med  2015; 36: 56– 73. Google Scholar CrossRef Search ADS PubMed  45 Drinkovic D, Morris AJ, Dyet K et al.   Plasmid-mediated AmpC β-lactamase-producing Escherichia coli causing urinary tract infection in the Auckland community likely to be resistant to commonly prescribed antimicrobials. N Z Med J  2015; 128: 50– 9. Google Scholar PubMed  46 Belmahdi M, Bakour S, Al Bayssari C et al.   Molecular characterisation of extended-spectrum β-lactamase- and plasmid AmpC-producing Escherichia coli strains isolated from broilers in Bejaia, Algeria. J Glob Antimicrob Resist  2016; 6: 108– 12. Google Scholar CrossRef Search ADS PubMed  47 Harris PN, Ferguson JK. Antibiotic therapy for inducible AmpC β-lactamase-producing Gram-negative bacilli: what are the alternatives to carbapenems, quinolones and aminoglycosides? Int J Antimicrob Agents  2012; 40: 297– 305. Google Scholar CrossRef Search ADS PubMed  48 Sidjabat HE, Seah KY, Coleman L et al.   Expansive spread of IncI1 plasmids carrying blaCMY-2 amongst Escherichia coli. Int J Antimicrob Agents  2014; 44: 203– 8. Google Scholar CrossRef Search ADS PubMed  49 Verdet C, Gautier V, Chachaty E et al.   Genetic context of plasmid-carried blaCMY-2-like genes in Enterobacteriaceae. Antimicrob Agents Chemother  2009; 53: 4002– 6. Google Scholar CrossRef Search ADS PubMed  50 Villa L, Garcia-Fernandez A, Fortini D et al.   Replicon sequence typing of IncF plasmids carrying virulence and resistance determinants. J Antimicrob Chemother  2010; 65: 2518– 29. Google Scholar CrossRef Search ADS PubMed  51 Carattoli A. Resistance plasmid families in Enterobacteriaceae. Antimicrob Agents Chemother  2009; 53: 2227– 38. Google Scholar CrossRef Search ADS PubMed  52 Arredondo-Alonso S, Willems RJ, van Schaik W et al.   On the (im)possibility of reconstructing plasmids from whole-genome short-read sequencing data. Microbial Genomics  2017; Epub online Aug 2017. http://mgen.microbiologyresearch.org/content/journal/mgen/10.1099/mgen.0.000128.v1. © The Author 2017. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved. For Permissions, please email: journals.permissions@oup.com.

Journal

Journal of Antimicrobial ChemotherapyOxford University Press

Published: Mar 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off