J Antimicrob Chemother 2017; 72: 700–704 doi:10.1093/jac/dkw511 Advance Access publication 30 December 2016 Prediction of antibiotic resistance from antibiotic resistance genes detected in antibiotic-resistant commensal Escherichia coli using PCR or WGS Downloaded from https://academic.oup.com/jac/article-abstract/72/3/700/2762720 by Biomedical Library user on 04 April 2019 Robert A. Moran1, Sashindran Anantham1, Kathryn E. Holt2,3 and Ruth M. Hall1* 1School of Life and Environmental Sciences, The University of Sydney, NSW 2006, Australia; 2Centre for Systems Genomics, University of Melbourne, Parkville, Victoria 3010, Australia; 3Department of Biochemistry and Molecular Biology, Bio21 Molecular Science and Biotechnology Institute, University of Melbourne, Parkville, Victoria 3010, Australia *Corresponding author. School of Life and Environmental Sciences, Molecular Bioscience Building G08, The University of Sydney, NSW 2006, Australia. Tel: þ61-2-9351-3465; Fax: þ61-2-9351-5858; E-mail: [email protected] Received 22 August 2016; accepted 27 October 2016 Objectives: To assess the effectiveness of bioinformatic detection of resistance genes in whole-genome se- quences in correctly predicting resistance phenotypes. Methods: Genomes of a collection of well-characterized commensal Escherichia coli were sequenced using Illumina HiSeq technology and assembled with SPAdes. Antibiotic resistance genes identified by PCR, SRST2 ana- lysis of reads and ResFinder analysis of SPAdes assemblies were compared with known resistance phenotypes. Results: Generally, the antibiotic resistance genes detected using bioinformatic methods were concordant, but only ARG-ANNOT included sat2. However, the presence or absence of genes was not always predictive of the phenotype. In one strain, trimethoprim resistance was due to a known mutation in the chromosomal folA gene. In cases where the copy number was low, the aadA5 gene downstream of dfrA17 did not confer streptomycin or spectinomycin resistance. Resistance genes were found in the genomes that were not detected previously by PCRs targeting a limited gene set and gene cassettes in class 1 or class 2 integrons. In one isolate, the aadA1 gene cassette in the estX-aadA1 cassettes pair was outside an integron context and was not expressed. The qnrS1 gene, conferring reduced susceptibility to fluoroquinolones, and the blaCMY-2 gene, encoding an ESBL, were each detected in a single isolate and mphA (macrolide resistance) was present in six isolates surrounded by IS26 and IS6100. Conclusions: WGS analysis detected more genes than PCR. Some were not expressed, causing inconsistencies with the experimentally determined phenotype. An unpredicted chromosomal folA mutation causing trimetho- prim resistance was found. Introduction volunteers using a number of measures of strain diversity, includ- Bacteria resistant to therapeutic antibiotics represent a significant ing the antibiotic resistance phenotype of each isolate, phylogen- global health challenge as infections caused by multiply, exten- etic group and random amplified polymorphic DNA (RAPD) 3–6 sively or pan antibiotic-resistant Gram-negative and Gram-positive profiling. A single representative of each strain type detected bacteria continue to increase. As WGS becomes more affordable was retained. For the resistant strains, PCR was used to detect class and searchable databases of acquired antibiotic resistance genes 1 and class 2 integrons and the gene cassettes they harbour, as have been made available,1,2 predicting the antibiotic resistance well as a limited set of other resistance genes. The plasmid content 6 profile by identifying antibiotic resistance genes in WGS data has was determined recently and a few plasmids that carry resistance 5,7 become feasible. However, studies that compare experimentally genes have been studied or completely sequenced. determined resistance profiles with resistance gene content are Here, the genomes of the antibiotic-resistant isolates in the needed in order to assess the reliability of WGS-based approaches. collection were sequenced via Illumina and both reads-based We have established a non-redundant collection of commensal analysis with SRST2 using the ARG-ANNOT database and Escherichia coli recovered from healthy Australian adults by exam- assembly-based analysis with ResFinder were used to determine ining the population structure of E. coli from the colons of the resistance gene content. The outputs of this analysis were VC The Author 2016. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved. For Permissions, please email: [email protected]. 700 Resistance genes in commensal Escherichia coli JAC Table 1. Antibiotic-resistant commensal E. coli collection Resistance Additional genes a b Strain phenotype Phylogroup blaTEM strAB aadA sul dfrA tetA Other WGS 14.3-R4 TET A0 ÀÀÀÀÀA— — 24.16-R4 AMP STR SMX TMP A0 11225 2 —— 24.20-R5 TET A0 22222A— — Downloaded from https://academic.oup.com/jac/article-abstract/72/3/700/2762720 by Biomedical Library user on 04 April 2019 1-R1 AMP STR SMX TET A1 þþÀ1, 27 A— — TMP 2.2-R2 NAL TET A1 ÀÀÀÀÀA— — 3.6-R4 AMP STR SMX A1 þþÀ2 ÀÀ—— 3.6-R5 AMP CHL GEN NAL A1 þþ51, 217 A catA1, aacC2d mphA STR SMX TET TOB TMP 21.1-R1 AMP STR SMX TET A1 þþÀ2 À A— — 1.4-R4 SMX TET TMP A1 ÀÀÀ17 A— — 15.1-R1 AMP STR SMX TMP A1 þþÀ25 À —— 1.2-R2 TET A1 ÀÀÀÀÀA— — 1.2-R3 SMX TET TMP A1 þÀ214A— — 1.9-R7 AMP TET A1 þÀÀÀA — qnrS1 14.2-R3 SMX TET TMP A1 ÀÀÀ2 À A— aadA5, dfrA17 24-R3 AMP STR SMX TMP A1 11225 2 —— 3.5-R3 STR SPT TMP B1c ÀÀ1 À 1 À sat2 — 19.1-R1a STR SPT TMP B1 ÀÀ1 À 1 À sat2 — 19.1-R1 TMP B1 ÀÀÀÀ1 À sat2 — 1.10-R8 SMX TET TMP B2 22215 B— — 2.1-R1 AMP B2 þÀÀÀÀ—— 3-S1R AMP B2 þÀÀÀÀ—— 3.3-R2 AMP B2 þÀÀÀÀ—— 13.1-R2 AMP SMX B2 þÀ2 ÀÀ—— 13.1-R2a AMP SMX TET B2 þÀ2 À B— — 14.2-R2 AMP B2 þÀÀÀÀ—— 22.1-R1 AMP CHL STR SPT B2 þ1, 23 12B cmlA1 mef(B) SMX TET TMP 10.1-R1 AMP STR SMX TET B2 þþÀ21 B sat2 — TMP 11.1-R1 AMP TMP B2 þÀÀÀÀ—— 11.3-R3 AMP TET B2 þÀÀÀA— — 19.1-R3 AMP STR SPT SMX B2 þþ11, 21 B— — TET TMP 1.9-R6 AMP CIP GEN NAL B2 þþ51, 217 A aacC2d mphA STR SMX TET TOB TMPd 5.1-R1 AMP STR SMX B2 þþÀ2 ÀÀ—— 2.3-R3 STR SMX TMP D þÀ1, 25 À —— 2.3-R4 SMX TET TMP D ÀÀÀ15 C— — 2.3-R5 STR SMX TET TMP D þÀ1, 25 B— — 3-R1 AMP TET D þÀÀÀB— — 4-R1 AMP TET D þÀÀÀB— — 4.2-R3 TET D ÀÀÀÀÀB— — 5.2-R2 AMP CHL STR SMX D þþÀ25e D catA1 — TET TMP 11.4-R4 SMX TMP D 22215 2 —— 18.1-R1 AMP TET D þÀÀÀB— — 4-R2 AMP STR SPT SMX D þþ51, 217 A — mphA TET TMP Continued 701 Moran et al. Table 1. Continued Resistance Additional genes a b Strain phenotype Phylogroup blaTEM strAB aadA sul dfrA tetA Other WGS 4.3-R2a AMP D þÀÀÀÀ—— 4.4-R2b AMP STR SMX TET D þþÀ2 À A — mphA 11.2-R2 AMP STR SPT SMX D þþ51, 217 A — mphA TET TMP Downloaded from https://academic.oup.com/jac/article-abstract/72/3/700/2762720 by Biomedical Library user on 04 April 2019 14.1-R1 AMP STR SPT SMX D þþ51, 217 A— mphA TET TMP 6.2-R1 SMX TMP D þÀ214À —— 9.1-R1 SMX D ÀÀÀ2 ÀÀ— aadA1 13.1-R1 TET D ÀÀÀÀÀB— — 24.1-R1 AMP CAZ CTX TET D 22222B— blaCMY-2 24.1-R2 AMP CHL STR SPT D 111, 23 12A cmlA1 — SMX TET TMP aNew strains and ciprofloxacin and nalidixic acid resistance are bold. bAMP, ampicillin; CAZ, ceftazidime; CTX, cefotaxime; CHL, chloramphenicol; CIP, ciprofloxacin; GEN, gentamicin; NAL, nalidixic acid; STR, streptomycin; SPT, spectinomycin; SMX, sulfamethoxazole; TET, tetracycline; TOB, tobramycin; TMP, trimethoprim. c 6 Previously reported as phylogroup A0 in Moran et al. (2015). dPreviously reported as AMP (CHL) STR SMX TET TMP in Anantham and Hall5 (2012). eIncorrectly reported as dfrA1 in Anantham and Hall5 (2012). reconciled with the resistance phenotype and resistance gene JAC Online). Resistance genes in assembled contigs were detected using determination using PCR-based methods. ResFinder (https://cge.cbs.dtu.dk//services/ResFinder/)1 and raw reads were used to query ARG-ANNOT (http://en.mediterranee-infection.com/article. php?laref¼283%26titre¼arg-annot)2 using SRST2 with default settings.11 Materials and methods The coverage of resistance genes relative to the average coverage for genes E. coli isolates used for MLST was used to assess copy number. Assembled sequences were compared with those found in the GenBank non-redundant DNA The strains used were either derived from a published collection of com- database using the BLAST alignment facility (http://blast.ncbi.nlm.nih.gov). mensal E. coli strains recovered from the faeces of 22 healthy human sub- Gene Construction Kit version 2.5 (Textco, West Lebanon, NH, USA) was jects between 2008 and 20108 or were isolated from further samples used to draw figures to scale. collected from the same subjects and from an additional subject over the time frame 2008–14.4–6 Sample collection followed protocols approved by the University of Sydney Human Research Ethics Committee (HREC) (04- GenBank accession numbers 2008/10778) with informed consent from subjects. The protocols for isolation Sequences of fragments containing blaCMY-2 and the estX-aadA1 cassettes and analysis are described elsewhere.4 The 51 unique isolates that are resist- have been deposited in GenBank with the accession numbers KX462017 ant to at least 1 of the 12 antibiotics (ampicillin, ceftazidime, cefotaxime, and KX462014, respectively.