A coliform-targeted metagenomic method facilitating human exposure estimates to Escherichia coli-borne antibiotic resistance genes

A coliform-targeted metagenomic method facilitating human exposure estimates to Escherichia... Abstract Antimicrobial resistance and the spread of antibiotic resistance genes (ARGs) pose a threat to human health. Community-acquired infections resistant to treatment with first-line antibiotics are increasing, and there are few studies investigating environmental exposures and transmission. Our objective is to develop a novel targeted metagenomic method to quantify the abundance and diversity of ARGs in a faecal indicator bacterium, and to estimate human exposure to resistant bacteria in a natural environment. Sequence data from Escherichia coli metagenomes from 13 bathing waters in England were analysed using the ARGs Online Analysis Pipeline to estimate the abundance and diversity of resistance determinants borne by this indicator bacterium. These data were averaged over the 13 sites and used along with data on the levels of E. coli in English bathing waters in 2016 and estimates of the volume of water that water users typically ingest in an average session of their chosen activityto quantify the numbers of ARGs that water users ingest. Escherichia coli in coastal bathing waters were found to harbour on average 1.24 ARGs per cell. Approximately 2.5 million water sports sessions occurred in England in 2016 that resulted in water users ingesting at least 100 E. coli-borne ARGs. antibiotic resistance, Escherichia coli, bathing waters, recreation, high-throughput sequencing INTRODUCTION Antimicrobial resistance is one of the greatest public health threats faced today. Infections caused by resistant organisms are on the rise, with resulting treatment failure and increases in morbidity, mortality and costs associated with increased demands on healthcare services and loss in productivity. According to a recent review of antimicrobial resistance, if current trends continue, 10 million deaths are predicted to occur annually by the year 2050, with a global economic cost of $100 billion USD by this time (Review on antimicrobial resistance 2014). Much research concerning the transmission and acquisition of resistant infections has focused on the role of hospitals. However, the frequency of resistance to antibiotics among community-acquired pathogens is increasing, and with up to 80% of antibiotics prescribed in primary care settings, research into non-nosocomial transmission routes is of increasing importance (Costelloe et al. 2014). One way in which people might be exposed to and infected by resistant bacteria is through exposure to natural waters contaminated by faecal pollution. Natural aquatic environments, such as rivers and oceans, are regularly contaminated by faecal pollution from anthropogenic and zoonotic sources, which introduces large numbers of microorganisms to these aquatic environments, some of which can be pathogenic and/or harbour clinically important antibiotic resistance genes (ARGs) (Pruden et al. 2006; Review on Antimicrobial Resistance 2015; Blackburn et al. 2017). Approximately 50% of the world's population lives within 100 km of the coast (Shuval 2003), and in the UK, millions of people visit the coast each year (Kantar TNS 2017), and engage in activities involving head immersion in seawater and water ingestion (Leonard et al. 2015). Very few studies have investigated these natural environments as sites where people could be exposed to and infected by resistant bacteria. The few that have investigated transmission of resistant bacteria via aquatic environments used culture-based methods to isolate bacteria phenotypically resistant to antibiotics, and found evidence of an association between exposure to natural waters and faecal colonisation by resistant bacteria (Leonard 2018) or community-acquired urinary tract infections caused by resistant bacteria (Soraas et al. 2013). High-throughput sequencing methods offer opportunities to explore the diversity and abundance of ARGs in environmental samples and reduce biases introduced by PCR-based methods and cultivation of phenotypically resistant bacteria (Li et al. 2015). These techniques are increasingly used to study environmental microbiology and the ecology of resistant bacteria in various matrices and are able to produce data on unculturable bacteria (Logares et al. 2012; Ercolini 2013). However, a problem often identified with culture-independent sequencing of bacteria is a lack of information on the bacterial hosts of identified genes in samples, many of which may be harmless to humans. The primary aim of this study was to develop a novel approach using targeted metagenomics to investigate the relative abundance and diversity of ARGs in the faecal indicator bacterium, Escherichia coli, in coastal bathing waters. Escherichia coli were selected as indicator organisms because of their presence in the gut microbiomes of animals (including humans), and their ease of culture leading to widespread use to monitor water quality (including drinking water, bathing waters and shellfish waters) (International Organization for Standardization 2005; Monaghan and Hutchinson 2010; Odonkor and Ampofo 2013). In addition E. coli also readily undergo horizontal gene transfer with other members of the microbial community (Levy, FitzGerald and Macone 1976; Handel et al. 2015) and clinically important AGRs can easily spread from non-pathogenic bacteria to ones capable of causing disease in humans (Ashbolt et al. 2013). Escherichia coli, for example E. coli sequence type 131 (Johnson et al. 2010) and E. coli O157:H7 (Ihekweazu et al. 2006), are increasingly important clinical pathogens causing a wide variety of intestinal and extra-intestinal infections (Russo and Johnson 2003). Escherichia coli are the leading cause of bloodstream infections, responsible for causing 40 580 cases of bacteraemia in England between 2016 and 2017, more than three times greater than the number of bacteraemias caused by methicillin-resistant and methicillin-susceptible Staphylococcus aureus combined (Public Health England 2017). Escherichia coli are therefore useful indicator organisms for monitoring antibiotic resistance in aquatic environments, such as bathing waters, to which humans have high exposure levels. A secondary aim was to use data generated by this study to estimate bather exposure to ARGs borne by E. coli in coastal bathing waters in England in 2016. METHODS Isolation of E. coli from coastal bathing waters The Environment Agency, as part of their routine monitoring of designated bathing waters (European Parliament Council of the European Union 2006), collect water samples from 415 bating waters in England on a weekly basis every year between mid-May and the end of September (Environment Agency 2017). The levels of E. coli in water samples are quantified using standard methods for membrane filtration (performed by Environment Agency staff), in which 0.1, 1, and 10 mL water samples are passed through a filter (0.45 µm pore size) and these filters are incubated on Tryptone Bile X-Glucuronide agar for 4 ± 1 h at 30°C ± 1°C, followed by 14 ± 2 h at 44°C ± 0.5°C (Environment Agency 2009; International Organization for Standardization 2014a; International Organization for Standaridzation 2014b). After incubation, all blue colonies on the filter are counted as confirmed E. coli. Validation confirmed that 99% of phenotypic E. coli colonies are genotypic E. coli (unpublished data, Jonathan Porter, Environment Agency, pers. comm), meaning very few non-E. coli species will be included in the E. coli composite samples. For this study, the Environment Agency provided filters with cultured E. coli from water samples taken at 13 different designated coastal bathing waters. All water samples were collected between 21 September 2016 and 22 September 2016 and were transported on ice from the Environment Agency Laboratory (Starcross, Exeter) to the University of Exeter Medical School laboratory for processing (Penryn, Cornwall). Sample preparation and DNA extraction All E. coli colonies from each site were counted, and every E. coli colony that was growing separately (i.e. not touching other colonies) was picked using a sterile 1000 µL pipette tip and combined in an Eppendorf with 0.85% sterile NaCl solution to form a composite E. coli metagenome sample for each site. In addition, all the non-E. coli coliform (NEC) colonies were picked into a composite NEC sample across all 13 sites. DNA was extracted from the 14 composite samples using the FastDNA Spin kit (MP Biomedicals, Santa Ana, California) according to the manufacturer's instructions, followed by RNAse digestion and clean up (Agencourt AMPure XP System Beckman Coulter, Brea, California). DNA concentrations were determined using the Qubit dsDNA high-sensitivity assay kit (Invitrogen, Carlsbad, California). The DNA was sent for sequencing (University of Exeter, Exeter Sequencing Service) that provided Nextera XT library preparation and Illumina high-throughput sequencing on MiSeq, generating 2 × 250 bp paired end reads. The amount of biomass contributed to the composite samples by each E. coli colony was not standardised by cell count. Therefore, larger colonies could contribute more genetic material to a sample than smaller colonies, and genes in larger colonies be over-represented in the absolute abundance of genes in each sample. To assess whether the methods used to pick cells from filters resulted in similar numbers of cells contributing biomass to the composite samples, the number of E. coli picked using a 1000 µL pipette tip was quantified for 18 E. coli colonies (see supplementary material for methods). Bioinformatic and statistical analyses Sequence data underwent analysis using the validated AGRs Online Analysis Pipeline (OAP) (Yang et al. 2016). Sample sequences were aligned against sequences in the structured antibiotic resistance gene (SARG) database that contains 4049 ARG sequences. Sample sequences were identified as ARGs based on identity being more than 85% and the alignment length ratio being more than 75%. Results were categorised into resistance ‘type’ (24) and ‘subtype’ (1209) whereby ‘type’ is the class of antibiotic to which the ARG confers resistance (e.g. beta-lactams), and ‘subtype’ is the identity of the resistance gene (e.g. CTX-M-15 of beta-lactams). The abundance of ARGs was quantified along with the relative abundance of genes per E. coli cell, with E. coli abundance being calculated as previously described (Yang et al. 2016), and the Simpson Index of diversity of ARGs at each site were calculated using the R package, phyloseq. Water user exposure to ARGs borne by E. coli in coastal bathing waters The mean number of resistance genes per E. coli over all 13 sites was calculated for each type of antibiotic class and are presented in Table S2 (Supporting Information). However, for the main analysis, data on multi-drug resistance (MDR) genes (82 subtypes) and unclassified resistance gene (17 subtypes) were excluded because although they contribute to the antibiotic resistance potential, they may be present in both antibiotic-susceptible and antibiotic-sensitive bacteria (Li et al. 2015). The mean number of genes for each antibiotic per E. coli were then summed to give the average number of resistance genes per E. coli. Similar methods to those described by Leonard et al. (2015) were used to estimate the number of E. coli-borne ARGs ingested by coastal water users in England in 2016. Briefly, data on the density of E. coli in English designated bathing waters were obtained from the Environment Agency website for the 2016 bathing season (Environment Agency 2017). Escherichia coli data from bathing water sample were categorised into three groups based upon E. coli density data: (i) E. coli density ≤250 E. coli colony-forming units (cfu) per 100 mL; (ii) E. coli density between 251 and 500 cfu 100 mL−1; (iii) E. coli density >500 cfu 100mL−1. Escherichia coli densities were converted from cfu 100 mL−1 to cfu mL−1 and multiplied by the number of resistance genes carried by the average E. coli. These were then applied to the average volume of water that various water users ingest during a typical session of their chosen activity to give X, the number of resistance genes borne by E. coli ingested by water users:   \begin{equation*} X = E.R.V \end{equation*} where E is the number of E. coli cfu mL−1, R is the number of resistance genes carried by the average E. coli, and V is the volume of water ingested (mL). The mean of X was then estimated for each water sport by water quality category. The number of water sports sessions in which E. coli-borne ARGs were ingested by water users in England in 2016 was then estimated (Leonard et al. 2015). The Office for National Statistics estimated population of England in mid-2016 to be 55 268 200, with 13 107 000 aged 19 years or under, and 42 161 200 aged 20 years and above (Office for National Statistics 2016). A variety of estimates were produced to understand the scale of exposure if different thresholds of resistance gene carriage were considered: exposure to at least one ARG borne by E. coli, at least 10 E. coli-borne ARGs, 100 E. coli-borne ARGs and 1000 E. coli-borne ARGs. RESULTS Isolation of E. coli from coastal bathing waters Filters with E. coli and other coliform colonies from 13 designated coastal bathing sites were acquired: six from beaches in the south of England and seven from beaches in the northeast (Table 1). A total of 315 E. coli colonies were picked across all thirteen sites (mean number of E. coli per site = 24, standard deviation 21). In addition, all 105 non-E. coli coliform (NEC) colonies from across the 13 sites were picked into a single composite NEC sample. Table 1. The number of E. coli colonies and non-E. coli colonies picked at each site, and the region of England in which each site was located. Site number  Region  Number of E. coli colonies sampled  Number of non-E. coli coliform colonies sampled  1  South  15  20  2  South  20  16  3  South  20  13  4  South  10  5  5  North east  40  3  6  South  12  2  7  North east  20  1  8  North east  18  1  9  North east  27  2  10  North east  90  5  11  South  21  5  12  North east  9  14  13  North east  13  18  Total     315  105  Site number  Region  Number of E. coli colonies sampled  Number of non-E. coli coliform colonies sampled  1  South  15  20  2  South  20  16  3  South  20  13  4  South  10  5  5  North east  40  3  6  South  12  2  7  North east  20  1  8  North east  18  1  9  North east  27  2  10  North east  90  5  11  South  21  5  12  North east  9  14  13  North east  13  18  Total     315  105  View Large The numbers of E. coli picked by a 1000 µL pipette tip was in the same order of magnitude, and the majority of replicates were within one standard deviation of the mean (2.41 × 108) (Fig. S1, Supporting Information). Using a pipette tip was a fast and simple method of picking consistent numbers of cells. Furthermore, the number of E. coli colonies picked at each site appears to have no association with the abundance of ARGs identified at each site, although given the small sample size, it is difficult to see relationships (Fig. S2, Supporting Information). Results of sequence analyses The average library size was 356 MB and assuming a typical E. coli genome of 5MB (Lukjancenko, Wassenaar and Ussery 2010), this gave a coverage equivalent to >70 genomes per composite sample. The sequencing depth for each sample is reported in Table S1 (Supporting Information). Among all 13 sites, resistance genes to 12 different classes of antibiotic were identified in E. coli. All 13 sites had E. coli with resistance genes to at least seven different antibiotic classes, and one site had resistance genes to 11 classes of antibiotic (Fig. 1). Resistance genes against 10 types of antibiotics were not detected. The mean number of resistance genes per E. coli across all sites was 1.24 (Table S2, Supporting Information). Figure 1. View largeDownload slide The abundance of resistance genes to types of antibiotic per E. coli at each sampled site. High relative abundance is indicated by pink, low relative abundance is coloured in green. Grey cells indicate no detection. MLS is the antibiotic type macrolide-lincosamide-streptogramin. Figure 1. View largeDownload slide The abundance of resistance genes to types of antibiotic per E. coli at each sampled site. High relative abundance is indicated by pink, low relative abundance is coloured in green. Grey cells indicate no detection. MLS is the antibiotic type macrolide-lincosamide-streptogramin. The relative abundance of resistance determinants in NEC from all sites was also quantified. While there was a lower abundance of genes compared to E. coli (0.19 resistance genes per 16S rRNA in NEC compared to 0.45 per 16S rRNA in E. coli) (Table S2, Supporting Information), genes for resistance to 13 different classes of antibiotic were detected in the NEC, some of which were not found in E. coli (e.g. chloramphenicol and quinolone). Genes conferring sulphonamide resistance were detected in E. coli but not in NEC (Fig. S3, Supporting Inoformation). Resistance genes against the beta-lactam class of antibiotics was the most common type found among E. coli, accounting for 20% of the resistance genes across all sites (range 14% to 24%). Similarly, ARGs against beta-lactams were the most prevalent type of resistance among NEC, making up 31% of the resistance genes detected. The percentage of the total numbers of genes that confer resistance to each class of antibiotic among E. coli and NEC are displayed in Fig. 2. Figure 2. View largeDownload slide The percentage of genes conferring resistance to each type of antibiotic in the total E. coli and total non-E. coli coliforms populations. MLS is the antibiotic type macrolide-lincosamide-streptogramin. Figure 2. View largeDownload slide The percentage of genes conferring resistance to each type of antibiotic in the total E. coli and total non-E. coli coliforms populations. MLS is the antibiotic type macrolide-lincosamide-streptogramin. The Simpson index of diversity at each site was quantified using data on the gene subtype count among E. coli at each site. E. coli from site number 2 had the highest Simpsons Index of diversity compared to the E. coli from other sampled sites (Fig. 3), although samples from sites 1 and 3 also demonstrated high levels of ARG diversity compared to the 10 remaining samples. The diversity of ARGs among the composite NEC sample was also high. Other indices of diversity are displayed in Fig. S4 (Supporting Information). Figure 3. View largeDownload slide Simpson index of diversity of E. coli-borne antibiotic resistance genes at the 13 sample sites and among non-E. coli coliforms at all sites (NEC). Figure 3. View largeDownload slide Simpson index of diversity of E. coli-borne antibiotic resistance genes at the 13 sample sites and among non-E. coli coliforms at all sites (NEC). Bather exposure The average E. coli in English bathing waters in 2016 was estimated to harbour 1.24 ARGs. Table 2 presents the mean number of E. coli-borne ARGs ingested during a typical session of various high- and low-contact water sports. Table 2. Mean number of E. coli-borne antibiotic resistance genes ingested by different water users in bathing waters in different categories of water quality: (i) E. coli density ≤250 cfu 100 mL−1. (ii) E. coli density 251–500 cfu 100 mL−1. (iii) E. coli density >500 cfu 100 mL−1.   Mean number of E. coli-borne antibiotic resistance genes ingested by different water users in bathing waters in different categories of water quality    E. coli density ≤250 cfu 100 mL−1  E. coli density 251–500 cfu 100 mL−1  E. coli density >500 cfu 100 mL−1  High-contact water sportsa  Swimming (adults)  5.70  69.3  280  Swimming non-adults  13.2  160  648  Surfing  60.7  739  2987  Diving  3.52  42.9  173  Low-contact water sportsa  Boating  1.32  16.0  64.8  Canoeing  1.39  16.9  68.3  Fishing  1.28  15.6  63.0  Rowing  1.25  15.2  61.3  Kayaking  1.35  16.5  66.5  Wading  1.32  16.0  64.8    Mean number of E. coli-borne antibiotic resistance genes ingested by different water users in bathing waters in different categories of water quality    E. coli density ≤250 cfu 100 mL−1  E. coli density 251–500 cfu 100 mL−1  E. coli density >500 cfu 100 mL−1  High-contact water sportsa  Swimming (adults)  5.70  69.3  280  Swimming non-adults  13.2  160  648  Surfing  60.7  739  2987  Diving  3.52  42.9  173  Low-contact water sportsa  Boating  1.32  16.0  64.8  Canoeing  1.39  16.9  68.3  Fishing  1.28  15.6  63.0  Rowing  1.25  15.2  61.3  Kayaking  1.35  16.5  66.5  Wading  1.32  16.0  64.8  a High-contact water sports include those with a high chance of head immersion in the water, low-contact water sports include those activities with a low chance of head immersion. View Large These data were used to calculate a population-level estimate of exposure to ARGs borne by E. coli in coastal bathing waters. Of the 8218 water samples collected by the Environment Agency between 15 May 2016 and 30 September 2016, 7874 (95.8%) had E. coli densities of ≤250 cfu 100 mL−1. 205 (2.49%) had E. coli densities between 250 and 500 cfu 100 mL−1, and 139 (1.69%) had E. coli densities >500 cfu 100 mL−1 (Table S3, Supporting Information). Over 123 million water sports sessions occurred in England in 2016 that resulted in the ingestion of at least one ARG borne by E. coli in coastal bathing waters (Table 3), representing 100% of water sports sessions involving ingestion of resistant E. coli. Nearly 2.5 million sessions occurred if the threshold of exposure is raised to the ingestion of E. coli harbouring at least 100 ARGs. Table 3. Number of water sports sessions occurring in England in 2016 that resulted in at least 1, 10, 100 or 1000 E. coli-borne ARGs being ingested. Number of E. coli-borne ARGs ingested  Number of exposures in bathing waters in three different categories of water quality  Total number of exposure events occurring in England in 2016    E. coli density ≤250 cfu 100 mL−1  E. coli density 251–500 cfu 100 mL−1  E. coli density >500 cfu 100 mL−1    1  118 028 329  3 072 874  2 083 558  123 184 760  10  36 251 904  3 072 874  2 083 558  41 408 336  100  0  943 820  1 540 037  2 483 857  1000  0  0  373 925  373 925  Number of E. coli-borne ARGs ingested  Number of exposures in bathing waters in three different categories of water quality  Total number of exposure events occurring in England in 2016    E. coli density ≤250 cfu 100 mL−1  E. coli density 251–500 cfu 100 mL−1  E. coli density >500 cfu 100 mL−1    1  118 028 329  3 072 874  2 083 558  123 184 760  10  36 251 904  3 072 874  2 083 558  41 408 336  100  0  943 820  1 540 037  2 483 857  1000  0  0  373 925  373 925  View Large DISCUSSION For the first time, a targeted metagenomic method to estimate the abundance and diversity of antibiotic resistance determinants in an indicator bacterium, E. coli, is described. Data generated on the relative abundance and diversity of antibiotic resistance genes (ARGs) were used to demonstrate that this approach can facilitate estimations of human exposure to E. coli-borne ARGs in coastal bathing waters. Based on the abundance of ARGs per E. coli isolated from 13 designated bathing waters in England, it was estimated that the average E. coli in this environmental compartment harboured 1.24 ARGs, and ingestion of E. coli-borne resistance genes is a certainty when taking part in water sports (100% of water sports sessions in 2016 involved the ingestion of E. coli harbouring ARGs). If the threshold for exposure to resistance is raised to ingestion of E. coli harbouring 100 or more ARGs, 2.5 million exposure events were estimated to have occurred. This demonstrates that bathing in coastal waters could be a significant means by which a large number of people are exposed to high numbers of genes conferring a wide variety of resistance to antibiotics, including genes borne on mobile genetic elements, harboured by bacteria capable of colonising the human gut (Leonard et al. 2015). This is likely to be an underestimate of exposure, because genes thought to confer multidrug resistance were excluded from the analyses and mutation based resistance was not considered. These estimates of exposure are greater than those previously estimated by Leonard et al. (2015), when exposure to E. coli phenotypically resistant to one group of antibiotics (third-generation cephalosporins) was estimated among water users. The SARG database used in the OAP (ARGs-OAP) enabled the classification of ARGs into the types of antibiotic to which identified genes conferred resistance. Genes conferring resistance to beta-lactam antibiotics were most commonly found among both E. coli and non-E. coli coliform metagenomes, accounting for 20% and 31%, respectively, of the resistance genes identified. It was also possible to quantify the diversity of resistance gene subtypes in each sample. The highest diversity of ARGs was found among E. coli from sample number 2, and diversity was also relatively high in samples 1 and 3. Diversity was otherwise quite consistent among E. coli across all sites. A possible explanation for variations in ARG diversity across sites could be differences in the sources of pollution affecting each site. For example, sites heavily impacted by human activities and those with multiple pollution sources might have a higher abundance and diversity of ARGs compared to more pristine sites (Li et al. 2015). The approach described has the potential to address important research questions regarding the ecology of antimicrobial resistance determinants in the environment, such as understanding the landscape-scale factors driving high levels of resistance in environments affected by faecal pollution, and to monitor the impact of antibiotic stewardship programmes in the community on ARGs in wastewater and receiving waters. Furthermore, the method provides data on ARGs borne by viable cells and need not be limited to investigating resistance genes in E. coli: any culturable bacterium of interest could be used to assess the abundance and diversity of ARGs they harbour. A limitation to this method is that it cannot be used for cells that cannot be cultured in the laboratory, or for samples with very low cell densities, since a minimum amount of genetic material is required for the construction of sequence libraries and for the detection of rare sequences (Tringe and Rubin 2005; Wang et al. 2013). On average, a sequencing depth equivalent to 71 genomes was achieved for the samples (Table S1, Supporting Information), suggesting the coverage was high enough to sequence a representative from each colony in the majority of samples. In composite samples with large numbers of colonies some colonies may be underrepresented, however, a similar coverage of E. coli genomes is achieved in each water sample independent of E. coli density in the original water samples. Interestingly there was no relationship between E. coli number in each pool and the number of annotated resistance genes, suggesting variation in resistance gene relative abundance and diversity is not a function of sampling effort. Compared to conventional metagenomics, where a small proportion of thousands of species and strains are sequenced giving very low coverage of total genetic diversity, this targeted approach not only links ARG diversity with a specific host species, but also gives relatively high coverage of E. coli diversity with the majority of isolates being sequenced. Therefore, it offers a pragmatic, cost effective approach to surveying resistance genes in environmental E. coli. Escherichia coli is a highly useful indicator organism for the purposes of studying ARGs in a variety of matrices because of its widespread use as a faecal indicator bacterium for monitoring the bacteriological quality of aquatic environments, food and drinking water (Odonkor and Ampofo 2013). In addition, data on the abundance and types of ARGs present in viable bacterial cells that are capable of colonising the human gut are provided. Therefore, this approach can be used to understand the abundance and diversity of ARGs harboured by E. coli isolated in a range of contexts. CONCLUSIONS A metagenomic method targeting E. coli and other coliforms was used to investigate the abundance and diversity of ARGs in a commonly measured indicator bacterium, E. coli, and facilitated an estimation of human exposure to ARGs borne by E. coli in coastal bathing waters. On average, E. coli were found to harbour 1.24 ARGs per cell. In 2016 all water sports sessions in England were estimated to result in ingestion of one or more E. coli-borne ARGs and 2.5 million sessions involved the ingestion of E. coli harbouring 100 or more ARGs. This targeted metagenomic approach has a variety of possible applications to facilitate the monitoring of resistance genes, as well as investigations into of the ecology of ARGs in various environments and matrices. SUPPLEMENTARY DATA Supplementary data are available at FEMSEC online. ACKNOWLEDGEMENTS Thanks to Dr Jonathan Porter and the staff at the Environment Agency for supplying us with E. coli specimens for this work. Thanks also to the Exeter Sequence Service facility. The authors have no conflicts of interest to declare. FUNDING This work was supported by the University of Exeter and the Chinese University of Hong Kong. The Exeter Sequencing Service is supported by the Wellcome Trust Institutional Strategic Support Fund (WT097835MF), the Wellcome Trust Multi User Equipment Award (WT101650MA), the Medical Research Council Clinical Infrastructure Funding (MR/M008924/1) and the Biotechnology and Biological Sciences Research Council Longer and Larger grants award (BB/K003240/1). Conflict of interest. None declared. REFERENCES Ashbolt NJ, Amezquita A, Backhaus T et al.   Human Health Risk Assessment (HHRA) for environmental development and transfer of antibiotic resistance. Environ Health Perspect . 2013; 121: 993– 1001. Google Scholar PubMed  Blackburn H, O'Neill R, Rangeley-Wilson C et al.   Flushed away: how sewage is still polluting the rivers of England and Wales: World Wildlife Fund, 2017; https://www.wwf.org.uk/updates/flushed-away-0 (05 December 2017, date last accessed). Costelloe C, Williams OM, Montgomery AA et al.   Antibiotic prescribing in primary care and antimicrobial resistance in patients admitted to hospital with urinary tract infection: a controlled observational pilot study. Antibiotics (Basel) . 2014; 3: 29– 38. Google Scholar CrossRef Search ADS PubMed  Environment Agency. Find a bathing water. 2017; https://environment.data.gov.uk/bwq/profiles/ (19 October 2017, date last accessed). Environment Agency. The Microbiology of Drinking Water (2009)—Part 4—methods for the isolation and enumeration of coliform bacteria and Escherichia coli (including O157:H7). Standing Committee of Analysts Blue Books: Methods for the Examination of Waters and Associated Materials . 2009; https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/316786/MoDWPart4-223MAYh.pdf (22 Januay 2018, date last accessed). Ercolini D High-throughput sequencing and metagenomics: moving forward in the culture-independent analysis of food microbial ecology. Appl Environ Microbiol . 2013; 79: 3148– 55. Google Scholar CrossRef Search ADS PubMed  European Parliament Council of the European Union. Directive 2006/7/EC of the European Parliament and of the Council of 15 February 2006 concerning the management of bathing water quality and repealing Directive 76/160/EEC. In: Union EPatCotE  (ed). 2006. Handel N, Otte S, Jonker M et al.   Factors that affect transfer of the IncI1 beta-Lactam resistance plasmid pESBL-283 between E-coli strains. PLoS One . 2015; 10, e0123039, doi: https://doi.org/10.1371/journal.pone.0123039. Google Scholar CrossRef Search ADS PubMed  Ihekweazu C, Barlow M, Roberts S et al.   Outbreak of E. coli O157 infection in the south west of the UK: risks from streams crossing seaside beaches. Euro Surveill . 2006; 11: 128– 30. Google Scholar CrossRef Search ADS PubMed  International Organization for Standardization. ISO/TS 16649-3:2005 Microbiology of food and animal feeding stufffs—horizontal method for the enumeration of beta-glucuronidase-positive Escherichia coli Part 3: Most probable number technique using 5-bromo-4-chloro-3-indolyl-beta-D-glucuronide, 2005. International Organization for Standardization. Water quality—enumeration of Escherichia coli and coliform bacteria—Part 1: Membrane filtration method for waters with low bacterial background flora Water quality, Microbiological Methods, 2014a. International Organization for Standardization. Water quality—requirements for the comparison of the relative recovery of microorganisms by two quantitative methods Water quality, Microbiology, 2014b. 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  Kantar TNS The Great Britain Day Visitor 2016 Annual Report, 2017; https://www.visitbritain.org/gb-day-visits-survey-latest-results (01 November 2017, date last accessed). Leonard, A, Zhang, L, Balfour, A et al. Exposure to and colonisation by antibiotic-resistant E. coli in UK coastal water users: Environmental surveillance, exposure assessment, and epidemiological study (Beach Bum Survey). Environ Int . 2018; doi: https://doi.org/10.1016/j.envint.2017.11.003. Leonard A, Zhang L, Balfour A et al.   Human recreational exposure to antibiotic resistant bacteria in coastal bathing waters. Environ Int . 2015; 82: 92– 100. Google Scholar CrossRef Search ADS PubMed  Levy SB, FitzGerald GB, Macone AB Spread of antibiotic-resistant plasmids from chicken to chicken and from chicken to man. Nature . 1976; 260: 40– 2. Google Scholar CrossRef Search ADS PubMed  Li B, Yang Y, Ma L et al.   Metagenomic and network analysis reveal wide distribution and co-occurrence of environmental antibiotic resistance genes. ISME J . 2015; 9: 2490– 502. Google Scholar CrossRef Search ADS PubMed  Logares R, Haverkamp TH, Kumar S et al.   Environmental microbiology through the lens of high-throughput DNA sequencing: synopsis of current platforms and bioinformatics approaches. J Microbiol Methods . 2012; 91: 106– 13. Google Scholar CrossRef Search ADS PubMed  Lukjancenko O, Wassenaar TM, Ussery DW Comparison of 61 sequenced Escherichia coli genomes. Microb Ecol . 2010; 60: 708– 20. Google Scholar CrossRef Search ADS PubMed  Monaghan J, Hutchinson M Monitoring microbial food safety of fresh produce. In: Food Standards Agency, Horitcultural Development Company . Horiticultural Development Company, Kenilworth, Warwickshire, 2010. Odonkor ST, Ampofo JK Escherichia coli as an indicator of bacteriological quality of water: an overview. Microbiol Res . 2013; 4: 5– 11. Google Scholar CrossRef Search ADS   Office for National Statistics. Population estimates for UK, England and Wales, Scotland and Northern Ireland: mid-2016. 2016; https://www.ons.gov.uk/peoplepopulationandcommunity/populationandmigration/populationestimates/datasets/populationestimatesforukenglandandwalesscotlandandnorthernireland (20 October 2017, date last accessed). Pruden A, Pei R, Storteboom H et al.   Antibiotic resistance genes as emerging contaminants: studies in northern Colorado. Environ Sci Technol . 2006; 40: 7445– 50. Google Scholar CrossRef Search ADS PubMed  Public Health England. Accompanying data for annual epidemiological commentary: MRSA, MSSA and E. coli bacteraemia and C. difificile infection data, up to and including financial year 2016 to 2017. 2017; https://www.gov.uk/government/statistics/mrsa-mssa-and-e-coli-bacteraemia-and-c-difficile-infection-annual-epidemiological-commentary (22 November 2017, date last accessed). Review on Antimicrobial Resistance. Antimicrobial Resistance: Tackling a crisis for the health and wealth of nations, 2014; https://amr-review.org/sites/default/files/AMR%20Review%20Paper%20-%20Tackling%20a%20crisis%20for%20the%20health%20and%20wealth%20of%20nations_1.pdf (02 October 2017, date last accessed). Review on Antimicrobial Resistance. Antimicrobials in agriculture and the environment: reducing unnecessary use and waste. 2015; https://amr-review.org/sites/default/files/Antimicrobials%20in%20agriculture%20and%20the%20environment%20-%20Reducing%20unnecessary%20use%20and%20waste.pdf (03 October 2017, date last accessed). Russo TA, Johnson JR Medical and economic impact of extraintestinal infections due to Escherichia coli: focus on an increasingly important endemic problem. Microbes Infect . 2003; 5: 449– 56. Google Scholar CrossRef Search ADS PubMed  Shuval H Estimating the global burden of thalassogenic diseases: human infectious diseases caused by wastewater pollution of the marine environment. J Water Health . 2003; 1: 53– 64. Google Scholar PubMed  Soraas A, Sundsfjord A, Sandven I et al.   Risk factors for community-acquired urinary tract infections caused by ESBL-producing enterobacteriaceae—a case-control study in a low prevalence country. PLoS One . 2013; 8: e69581. Google Scholar CrossRef Search ADS PubMed  Tringe SG, Rubin EM Metagenomics: DNA sequencing of environmental samples. Nat Rev Genet . 2005; 6: 805– 14. Google Scholar CrossRef Search ADS PubMed  Wang J, McLenachan PA, Biggs PJ et al.   Environmental bio-monitoring with high-throughput sequencing. Brief Bioinform . 2013; 14: 575– 88. Google Scholar CrossRef Search ADS PubMed  Yang Y, Jiang X, Chai B et al.   ARGs-OAP: online analysis pipeline for antibiotic resistance genes detection from metagenomic data using an integrated structured ARG-database. Bioinformatics . 2016; 32: 2346– 51. Google Scholar CrossRef Search ADS PubMed  © FEMS 2018. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png FEMS Microbiology Ecology Oxford University Press

A coliform-targeted metagenomic method facilitating human exposure estimates to Escherichia coli-borne antibiotic resistance genes

Loading next page...
 
/lp/ou_press/a-coliform-targeted-metagenomic-method-facilitating-human-exposure-Lrn7Ogf6wl
Publisher
Oxford University Press
Copyright
© FEMS 2018. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
ISSN
0168-6496
eISSN
1574-6941
D.O.I.
10.1093/femsec/fiy024
Publisher site
See Article on Publisher Site

Abstract

Abstract Antimicrobial resistance and the spread of antibiotic resistance genes (ARGs) pose a threat to human health. Community-acquired infections resistant to treatment with first-line antibiotics are increasing, and there are few studies investigating environmental exposures and transmission. Our objective is to develop a novel targeted metagenomic method to quantify the abundance and diversity of ARGs in a faecal indicator bacterium, and to estimate human exposure to resistant bacteria in a natural environment. Sequence data from Escherichia coli metagenomes from 13 bathing waters in England were analysed using the ARGs Online Analysis Pipeline to estimate the abundance and diversity of resistance determinants borne by this indicator bacterium. These data were averaged over the 13 sites and used along with data on the levels of E. coli in English bathing waters in 2016 and estimates of the volume of water that water users typically ingest in an average session of their chosen activityto quantify the numbers of ARGs that water users ingest. Escherichia coli in coastal bathing waters were found to harbour on average 1.24 ARGs per cell. Approximately 2.5 million water sports sessions occurred in England in 2016 that resulted in water users ingesting at least 100 E. coli-borne ARGs. antibiotic resistance, Escherichia coli, bathing waters, recreation, high-throughput sequencing INTRODUCTION Antimicrobial resistance is one of the greatest public health threats faced today. Infections caused by resistant organisms are on the rise, with resulting treatment failure and increases in morbidity, mortality and costs associated with increased demands on healthcare services and loss in productivity. According to a recent review of antimicrobial resistance, if current trends continue, 10 million deaths are predicted to occur annually by the year 2050, with a global economic cost of $100 billion USD by this time (Review on antimicrobial resistance 2014). Much research concerning the transmission and acquisition of resistant infections has focused on the role of hospitals. However, the frequency of resistance to antibiotics among community-acquired pathogens is increasing, and with up to 80% of antibiotics prescribed in primary care settings, research into non-nosocomial transmission routes is of increasing importance (Costelloe et al. 2014). One way in which people might be exposed to and infected by resistant bacteria is through exposure to natural waters contaminated by faecal pollution. Natural aquatic environments, such as rivers and oceans, are regularly contaminated by faecal pollution from anthropogenic and zoonotic sources, which introduces large numbers of microorganisms to these aquatic environments, some of which can be pathogenic and/or harbour clinically important antibiotic resistance genes (ARGs) (Pruden et al. 2006; Review on Antimicrobial Resistance 2015; Blackburn et al. 2017). Approximately 50% of the world's population lives within 100 km of the coast (Shuval 2003), and in the UK, millions of people visit the coast each year (Kantar TNS 2017), and engage in activities involving head immersion in seawater and water ingestion (Leonard et al. 2015). Very few studies have investigated these natural environments as sites where people could be exposed to and infected by resistant bacteria. The few that have investigated transmission of resistant bacteria via aquatic environments used culture-based methods to isolate bacteria phenotypically resistant to antibiotics, and found evidence of an association between exposure to natural waters and faecal colonisation by resistant bacteria (Leonard 2018) or community-acquired urinary tract infections caused by resistant bacteria (Soraas et al. 2013). High-throughput sequencing methods offer opportunities to explore the diversity and abundance of ARGs in environmental samples and reduce biases introduced by PCR-based methods and cultivation of phenotypically resistant bacteria (Li et al. 2015). These techniques are increasingly used to study environmental microbiology and the ecology of resistant bacteria in various matrices and are able to produce data on unculturable bacteria (Logares et al. 2012; Ercolini 2013). However, a problem often identified with culture-independent sequencing of bacteria is a lack of information on the bacterial hosts of identified genes in samples, many of which may be harmless to humans. The primary aim of this study was to develop a novel approach using targeted metagenomics to investigate the relative abundance and diversity of ARGs in the faecal indicator bacterium, Escherichia coli, in coastal bathing waters. Escherichia coli were selected as indicator organisms because of their presence in the gut microbiomes of animals (including humans), and their ease of culture leading to widespread use to monitor water quality (including drinking water, bathing waters and shellfish waters) (International Organization for Standardization 2005; Monaghan and Hutchinson 2010; Odonkor and Ampofo 2013). In addition E. coli also readily undergo horizontal gene transfer with other members of the microbial community (Levy, FitzGerald and Macone 1976; Handel et al. 2015) and clinically important AGRs can easily spread from non-pathogenic bacteria to ones capable of causing disease in humans (Ashbolt et al. 2013). Escherichia coli, for example E. coli sequence type 131 (Johnson et al. 2010) and E. coli O157:H7 (Ihekweazu et al. 2006), are increasingly important clinical pathogens causing a wide variety of intestinal and extra-intestinal infections (Russo and Johnson 2003). Escherichia coli are the leading cause of bloodstream infections, responsible for causing 40 580 cases of bacteraemia in England between 2016 and 2017, more than three times greater than the number of bacteraemias caused by methicillin-resistant and methicillin-susceptible Staphylococcus aureus combined (Public Health England 2017). Escherichia coli are therefore useful indicator organisms for monitoring antibiotic resistance in aquatic environments, such as bathing waters, to which humans have high exposure levels. A secondary aim was to use data generated by this study to estimate bather exposure to ARGs borne by E. coli in coastal bathing waters in England in 2016. METHODS Isolation of E. coli from coastal bathing waters The Environment Agency, as part of their routine monitoring of designated bathing waters (European Parliament Council of the European Union 2006), collect water samples from 415 bating waters in England on a weekly basis every year between mid-May and the end of September (Environment Agency 2017). The levels of E. coli in water samples are quantified using standard methods for membrane filtration (performed by Environment Agency staff), in which 0.1, 1, and 10 mL water samples are passed through a filter (0.45 µm pore size) and these filters are incubated on Tryptone Bile X-Glucuronide agar for 4 ± 1 h at 30°C ± 1°C, followed by 14 ± 2 h at 44°C ± 0.5°C (Environment Agency 2009; International Organization for Standardization 2014a; International Organization for Standaridzation 2014b). After incubation, all blue colonies on the filter are counted as confirmed E. coli. Validation confirmed that 99% of phenotypic E. coli colonies are genotypic E. coli (unpublished data, Jonathan Porter, Environment Agency, pers. comm), meaning very few non-E. coli species will be included in the E. coli composite samples. For this study, the Environment Agency provided filters with cultured E. coli from water samples taken at 13 different designated coastal bathing waters. All water samples were collected between 21 September 2016 and 22 September 2016 and were transported on ice from the Environment Agency Laboratory (Starcross, Exeter) to the University of Exeter Medical School laboratory for processing (Penryn, Cornwall). Sample preparation and DNA extraction All E. coli colonies from each site were counted, and every E. coli colony that was growing separately (i.e. not touching other colonies) was picked using a sterile 1000 µL pipette tip and combined in an Eppendorf with 0.85% sterile NaCl solution to form a composite E. coli metagenome sample for each site. In addition, all the non-E. coli coliform (NEC) colonies were picked into a composite NEC sample across all 13 sites. DNA was extracted from the 14 composite samples using the FastDNA Spin kit (MP Biomedicals, Santa Ana, California) according to the manufacturer's instructions, followed by RNAse digestion and clean up (Agencourt AMPure XP System Beckman Coulter, Brea, California). DNA concentrations were determined using the Qubit dsDNA high-sensitivity assay kit (Invitrogen, Carlsbad, California). The DNA was sent for sequencing (University of Exeter, Exeter Sequencing Service) that provided Nextera XT library preparation and Illumina high-throughput sequencing on MiSeq, generating 2 × 250 bp paired end reads. The amount of biomass contributed to the composite samples by each E. coli colony was not standardised by cell count. Therefore, larger colonies could contribute more genetic material to a sample than smaller colonies, and genes in larger colonies be over-represented in the absolute abundance of genes in each sample. To assess whether the methods used to pick cells from filters resulted in similar numbers of cells contributing biomass to the composite samples, the number of E. coli picked using a 1000 µL pipette tip was quantified for 18 E. coli colonies (see supplementary material for methods). Bioinformatic and statistical analyses Sequence data underwent analysis using the validated AGRs Online Analysis Pipeline (OAP) (Yang et al. 2016). Sample sequences were aligned against sequences in the structured antibiotic resistance gene (SARG) database that contains 4049 ARG sequences. Sample sequences were identified as ARGs based on identity being more than 85% and the alignment length ratio being more than 75%. Results were categorised into resistance ‘type’ (24) and ‘subtype’ (1209) whereby ‘type’ is the class of antibiotic to which the ARG confers resistance (e.g. beta-lactams), and ‘subtype’ is the identity of the resistance gene (e.g. CTX-M-15 of beta-lactams). The abundance of ARGs was quantified along with the relative abundance of genes per E. coli cell, with E. coli abundance being calculated as previously described (Yang et al. 2016), and the Simpson Index of diversity of ARGs at each site were calculated using the R package, phyloseq. Water user exposure to ARGs borne by E. coli in coastal bathing waters The mean number of resistance genes per E. coli over all 13 sites was calculated for each type of antibiotic class and are presented in Table S2 (Supporting Information). However, for the main analysis, data on multi-drug resistance (MDR) genes (82 subtypes) and unclassified resistance gene (17 subtypes) were excluded because although they contribute to the antibiotic resistance potential, they may be present in both antibiotic-susceptible and antibiotic-sensitive bacteria (Li et al. 2015). The mean number of genes for each antibiotic per E. coli were then summed to give the average number of resistance genes per E. coli. Similar methods to those described by Leonard et al. (2015) were used to estimate the number of E. coli-borne ARGs ingested by coastal water users in England in 2016. Briefly, data on the density of E. coli in English designated bathing waters were obtained from the Environment Agency website for the 2016 bathing season (Environment Agency 2017). Escherichia coli data from bathing water sample were categorised into three groups based upon E. coli density data: (i) E. coli density ≤250 E. coli colony-forming units (cfu) per 100 mL; (ii) E. coli density between 251 and 500 cfu 100 mL−1; (iii) E. coli density >500 cfu 100mL−1. Escherichia coli densities were converted from cfu 100 mL−1 to cfu mL−1 and multiplied by the number of resistance genes carried by the average E. coli. These were then applied to the average volume of water that various water users ingest during a typical session of their chosen activity to give X, the number of resistance genes borne by E. coli ingested by water users:   \begin{equation*} X = E.R.V \end{equation*} where E is the number of E. coli cfu mL−1, R is the number of resistance genes carried by the average E. coli, and V is the volume of water ingested (mL). The mean of X was then estimated for each water sport by water quality category. The number of water sports sessions in which E. coli-borne ARGs were ingested by water users in England in 2016 was then estimated (Leonard et al. 2015). The Office for National Statistics estimated population of England in mid-2016 to be 55 268 200, with 13 107 000 aged 19 years or under, and 42 161 200 aged 20 years and above (Office for National Statistics 2016). A variety of estimates were produced to understand the scale of exposure if different thresholds of resistance gene carriage were considered: exposure to at least one ARG borne by E. coli, at least 10 E. coli-borne ARGs, 100 E. coli-borne ARGs and 1000 E. coli-borne ARGs. RESULTS Isolation of E. coli from coastal bathing waters Filters with E. coli and other coliform colonies from 13 designated coastal bathing sites were acquired: six from beaches in the south of England and seven from beaches in the northeast (Table 1). A total of 315 E. coli colonies were picked across all thirteen sites (mean number of E. coli per site = 24, standard deviation 21). In addition, all 105 non-E. coli coliform (NEC) colonies from across the 13 sites were picked into a single composite NEC sample. Table 1. The number of E. coli colonies and non-E. coli colonies picked at each site, and the region of England in which each site was located. Site number  Region  Number of E. coli colonies sampled  Number of non-E. coli coliform colonies sampled  1  South  15  20  2  South  20  16  3  South  20  13  4  South  10  5  5  North east  40  3  6  South  12  2  7  North east  20  1  8  North east  18  1  9  North east  27  2  10  North east  90  5  11  South  21  5  12  North east  9  14  13  North east  13  18  Total     315  105  Site number  Region  Number of E. coli colonies sampled  Number of non-E. coli coliform colonies sampled  1  South  15  20  2  South  20  16  3  South  20  13  4  South  10  5  5  North east  40  3  6  South  12  2  7  North east  20  1  8  North east  18  1  9  North east  27  2  10  North east  90  5  11  South  21  5  12  North east  9  14  13  North east  13  18  Total     315  105  View Large The numbers of E. coli picked by a 1000 µL pipette tip was in the same order of magnitude, and the majority of replicates were within one standard deviation of the mean (2.41 × 108) (Fig. S1, Supporting Information). Using a pipette tip was a fast and simple method of picking consistent numbers of cells. Furthermore, the number of E. coli colonies picked at each site appears to have no association with the abundance of ARGs identified at each site, although given the small sample size, it is difficult to see relationships (Fig. S2, Supporting Information). Results of sequence analyses The average library size was 356 MB and assuming a typical E. coli genome of 5MB (Lukjancenko, Wassenaar and Ussery 2010), this gave a coverage equivalent to >70 genomes per composite sample. The sequencing depth for each sample is reported in Table S1 (Supporting Information). Among all 13 sites, resistance genes to 12 different classes of antibiotic were identified in E. coli. All 13 sites had E. coli with resistance genes to at least seven different antibiotic classes, and one site had resistance genes to 11 classes of antibiotic (Fig. 1). Resistance genes against 10 types of antibiotics were not detected. The mean number of resistance genes per E. coli across all sites was 1.24 (Table S2, Supporting Information). Figure 1. View largeDownload slide The abundance of resistance genes to types of antibiotic per E. coli at each sampled site. High relative abundance is indicated by pink, low relative abundance is coloured in green. Grey cells indicate no detection. MLS is the antibiotic type macrolide-lincosamide-streptogramin. Figure 1. View largeDownload slide The abundance of resistance genes to types of antibiotic per E. coli at each sampled site. High relative abundance is indicated by pink, low relative abundance is coloured in green. Grey cells indicate no detection. MLS is the antibiotic type macrolide-lincosamide-streptogramin. The relative abundance of resistance determinants in NEC from all sites was also quantified. While there was a lower abundance of genes compared to E. coli (0.19 resistance genes per 16S rRNA in NEC compared to 0.45 per 16S rRNA in E. coli) (Table S2, Supporting Information), genes for resistance to 13 different classes of antibiotic were detected in the NEC, some of which were not found in E. coli (e.g. chloramphenicol and quinolone). Genes conferring sulphonamide resistance were detected in E. coli but not in NEC (Fig. S3, Supporting Inoformation). Resistance genes against the beta-lactam class of antibiotics was the most common type found among E. coli, accounting for 20% of the resistance genes across all sites (range 14% to 24%). Similarly, ARGs against beta-lactams were the most prevalent type of resistance among NEC, making up 31% of the resistance genes detected. The percentage of the total numbers of genes that confer resistance to each class of antibiotic among E. coli and NEC are displayed in Fig. 2. Figure 2. View largeDownload slide The percentage of genes conferring resistance to each type of antibiotic in the total E. coli and total non-E. coli coliforms populations. MLS is the antibiotic type macrolide-lincosamide-streptogramin. Figure 2. View largeDownload slide The percentage of genes conferring resistance to each type of antibiotic in the total E. coli and total non-E. coli coliforms populations. MLS is the antibiotic type macrolide-lincosamide-streptogramin. The Simpson index of diversity at each site was quantified using data on the gene subtype count among E. coli at each site. E. coli from site number 2 had the highest Simpsons Index of diversity compared to the E. coli from other sampled sites (Fig. 3), although samples from sites 1 and 3 also demonstrated high levels of ARG diversity compared to the 10 remaining samples. The diversity of ARGs among the composite NEC sample was also high. Other indices of diversity are displayed in Fig. S4 (Supporting Information). Figure 3. View largeDownload slide Simpson index of diversity of E. coli-borne antibiotic resistance genes at the 13 sample sites and among non-E. coli coliforms at all sites (NEC). Figure 3. View largeDownload slide Simpson index of diversity of E. coli-borne antibiotic resistance genes at the 13 sample sites and among non-E. coli coliforms at all sites (NEC). Bather exposure The average E. coli in English bathing waters in 2016 was estimated to harbour 1.24 ARGs. Table 2 presents the mean number of E. coli-borne ARGs ingested during a typical session of various high- and low-contact water sports. Table 2. Mean number of E. coli-borne antibiotic resistance genes ingested by different water users in bathing waters in different categories of water quality: (i) E. coli density ≤250 cfu 100 mL−1. (ii) E. coli density 251–500 cfu 100 mL−1. (iii) E. coli density >500 cfu 100 mL−1.   Mean number of E. coli-borne antibiotic resistance genes ingested by different water users in bathing waters in different categories of water quality    E. coli density ≤250 cfu 100 mL−1  E. coli density 251–500 cfu 100 mL−1  E. coli density >500 cfu 100 mL−1  High-contact water sportsa  Swimming (adults)  5.70  69.3  280  Swimming non-adults  13.2  160  648  Surfing  60.7  739  2987  Diving  3.52  42.9  173  Low-contact water sportsa  Boating  1.32  16.0  64.8  Canoeing  1.39  16.9  68.3  Fishing  1.28  15.6  63.0  Rowing  1.25  15.2  61.3  Kayaking  1.35  16.5  66.5  Wading  1.32  16.0  64.8    Mean number of E. coli-borne antibiotic resistance genes ingested by different water users in bathing waters in different categories of water quality    E. coli density ≤250 cfu 100 mL−1  E. coli density 251–500 cfu 100 mL−1  E. coli density >500 cfu 100 mL−1  High-contact water sportsa  Swimming (adults)  5.70  69.3  280  Swimming non-adults  13.2  160  648  Surfing  60.7  739  2987  Diving  3.52  42.9  173  Low-contact water sportsa  Boating  1.32  16.0  64.8  Canoeing  1.39  16.9  68.3  Fishing  1.28  15.6  63.0  Rowing  1.25  15.2  61.3  Kayaking  1.35  16.5  66.5  Wading  1.32  16.0  64.8  a High-contact water sports include those with a high chance of head immersion in the water, low-contact water sports include those activities with a low chance of head immersion. View Large These data were used to calculate a population-level estimate of exposure to ARGs borne by E. coli in coastal bathing waters. Of the 8218 water samples collected by the Environment Agency between 15 May 2016 and 30 September 2016, 7874 (95.8%) had E. coli densities of ≤250 cfu 100 mL−1. 205 (2.49%) had E. coli densities between 250 and 500 cfu 100 mL−1, and 139 (1.69%) had E. coli densities >500 cfu 100 mL−1 (Table S3, Supporting Information). Over 123 million water sports sessions occurred in England in 2016 that resulted in the ingestion of at least one ARG borne by E. coli in coastal bathing waters (Table 3), representing 100% of water sports sessions involving ingestion of resistant E. coli. Nearly 2.5 million sessions occurred if the threshold of exposure is raised to the ingestion of E. coli harbouring at least 100 ARGs. Table 3. Number of water sports sessions occurring in England in 2016 that resulted in at least 1, 10, 100 or 1000 E. coli-borne ARGs being ingested. Number of E. coli-borne ARGs ingested  Number of exposures in bathing waters in three different categories of water quality  Total number of exposure events occurring in England in 2016    E. coli density ≤250 cfu 100 mL−1  E. coli density 251–500 cfu 100 mL−1  E. coli density >500 cfu 100 mL−1    1  118 028 329  3 072 874  2 083 558  123 184 760  10  36 251 904  3 072 874  2 083 558  41 408 336  100  0  943 820  1 540 037  2 483 857  1000  0  0  373 925  373 925  Number of E. coli-borne ARGs ingested  Number of exposures in bathing waters in three different categories of water quality  Total number of exposure events occurring in England in 2016    E. coli density ≤250 cfu 100 mL−1  E. coli density 251–500 cfu 100 mL−1  E. coli density >500 cfu 100 mL−1    1  118 028 329  3 072 874  2 083 558  123 184 760  10  36 251 904  3 072 874  2 083 558  41 408 336  100  0  943 820  1 540 037  2 483 857  1000  0  0  373 925  373 925  View Large DISCUSSION For the first time, a targeted metagenomic method to estimate the abundance and diversity of antibiotic resistance determinants in an indicator bacterium, E. coli, is described. Data generated on the relative abundance and diversity of antibiotic resistance genes (ARGs) were used to demonstrate that this approach can facilitate estimations of human exposure to E. coli-borne ARGs in coastal bathing waters. Based on the abundance of ARGs per E. coli isolated from 13 designated bathing waters in England, it was estimated that the average E. coli in this environmental compartment harboured 1.24 ARGs, and ingestion of E. coli-borne resistance genes is a certainty when taking part in water sports (100% of water sports sessions in 2016 involved the ingestion of E. coli harbouring ARGs). If the threshold for exposure to resistance is raised to ingestion of E. coli harbouring 100 or more ARGs, 2.5 million exposure events were estimated to have occurred. This demonstrates that bathing in coastal waters could be a significant means by which a large number of people are exposed to high numbers of genes conferring a wide variety of resistance to antibiotics, including genes borne on mobile genetic elements, harboured by bacteria capable of colonising the human gut (Leonard et al. 2015). This is likely to be an underestimate of exposure, because genes thought to confer multidrug resistance were excluded from the analyses and mutation based resistance was not considered. These estimates of exposure are greater than those previously estimated by Leonard et al. (2015), when exposure to E. coli phenotypically resistant to one group of antibiotics (third-generation cephalosporins) was estimated among water users. The SARG database used in the OAP (ARGs-OAP) enabled the classification of ARGs into the types of antibiotic to which identified genes conferred resistance. Genes conferring resistance to beta-lactam antibiotics were most commonly found among both E. coli and non-E. coli coliform metagenomes, accounting for 20% and 31%, respectively, of the resistance genes identified. It was also possible to quantify the diversity of resistance gene subtypes in each sample. The highest diversity of ARGs was found among E. coli from sample number 2, and diversity was also relatively high in samples 1 and 3. Diversity was otherwise quite consistent among E. coli across all sites. A possible explanation for variations in ARG diversity across sites could be differences in the sources of pollution affecting each site. For example, sites heavily impacted by human activities and those with multiple pollution sources might have a higher abundance and diversity of ARGs compared to more pristine sites (Li et al. 2015). The approach described has the potential to address important research questions regarding the ecology of antimicrobial resistance determinants in the environment, such as understanding the landscape-scale factors driving high levels of resistance in environments affected by faecal pollution, and to monitor the impact of antibiotic stewardship programmes in the community on ARGs in wastewater and receiving waters. Furthermore, the method provides data on ARGs borne by viable cells and need not be limited to investigating resistance genes in E. coli: any culturable bacterium of interest could be used to assess the abundance and diversity of ARGs they harbour. A limitation to this method is that it cannot be used for cells that cannot be cultured in the laboratory, or for samples with very low cell densities, since a minimum amount of genetic material is required for the construction of sequence libraries and for the detection of rare sequences (Tringe and Rubin 2005; Wang et al. 2013). On average, a sequencing depth equivalent to 71 genomes was achieved for the samples (Table S1, Supporting Information), suggesting the coverage was high enough to sequence a representative from each colony in the majority of samples. In composite samples with large numbers of colonies some colonies may be underrepresented, however, a similar coverage of E. coli genomes is achieved in each water sample independent of E. coli density in the original water samples. Interestingly there was no relationship between E. coli number in each pool and the number of annotated resistance genes, suggesting variation in resistance gene relative abundance and diversity is not a function of sampling effort. Compared to conventional metagenomics, where a small proportion of thousands of species and strains are sequenced giving very low coverage of total genetic diversity, this targeted approach not only links ARG diversity with a specific host species, but also gives relatively high coverage of E. coli diversity with the majority of isolates being sequenced. Therefore, it offers a pragmatic, cost effective approach to surveying resistance genes in environmental E. coli. Escherichia coli is a highly useful indicator organism for the purposes of studying ARGs in a variety of matrices because of its widespread use as a faecal indicator bacterium for monitoring the bacteriological quality of aquatic environments, food and drinking water (Odonkor and Ampofo 2013). In addition, data on the abundance and types of ARGs present in viable bacterial cells that are capable of colonising the human gut are provided. Therefore, this approach can be used to understand the abundance and diversity of ARGs harboured by E. coli isolated in a range of contexts. CONCLUSIONS A metagenomic method targeting E. coli and other coliforms was used to investigate the abundance and diversity of ARGs in a commonly measured indicator bacterium, E. coli, and facilitated an estimation of human exposure to ARGs borne by E. coli in coastal bathing waters. On average, E. coli were found to harbour 1.24 ARGs per cell. In 2016 all water sports sessions in England were estimated to result in ingestion of one or more E. coli-borne ARGs and 2.5 million sessions involved the ingestion of E. coli harbouring 100 or more ARGs. This targeted metagenomic approach has a variety of possible applications to facilitate the monitoring of resistance genes, as well as investigations into of the ecology of ARGs in various environments and matrices. SUPPLEMENTARY DATA Supplementary data are available at FEMSEC online. ACKNOWLEDGEMENTS Thanks to Dr Jonathan Porter and the staff at the Environment Agency for supplying us with E. coli specimens for this work. Thanks also to the Exeter Sequence Service facility. The authors have no conflicts of interest to declare. FUNDING This work was supported by the University of Exeter and the Chinese University of Hong Kong. The Exeter Sequencing Service is supported by the Wellcome Trust Institutional Strategic Support Fund (WT097835MF), the Wellcome Trust Multi User Equipment Award (WT101650MA), the Medical Research Council Clinical Infrastructure Funding (MR/M008924/1) and the Biotechnology and Biological Sciences Research Council Longer and Larger grants award (BB/K003240/1). Conflict of interest. None declared. REFERENCES Ashbolt NJ, Amezquita A, Backhaus T et al.   Human Health Risk Assessment (HHRA) for environmental development and transfer of antibiotic resistance. Environ Health Perspect . 2013; 121: 993– 1001. Google Scholar PubMed  Blackburn H, O'Neill R, Rangeley-Wilson C et al.   Flushed away: how sewage is still polluting the rivers of England and Wales: World Wildlife Fund, 2017; https://www.wwf.org.uk/updates/flushed-away-0 (05 December 2017, date last accessed). Costelloe C, Williams OM, Montgomery AA et al.   Antibiotic prescribing in primary care and antimicrobial resistance in patients admitted to hospital with urinary tract infection: a controlled observational pilot study. Antibiotics (Basel) . 2014; 3: 29– 38. Google Scholar CrossRef Search ADS PubMed  Environment Agency. Find a bathing water. 2017; https://environment.data.gov.uk/bwq/profiles/ (19 October 2017, date last accessed). Environment Agency. The Microbiology of Drinking Water (2009)—Part 4—methods for the isolation and enumeration of coliform bacteria and Escherichia coli (including O157:H7). Standing Committee of Analysts Blue Books: Methods for the Examination of Waters and Associated Materials . 2009; https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/316786/MoDWPart4-223MAYh.pdf (22 Januay 2018, date last accessed). Ercolini D High-throughput sequencing and metagenomics: moving forward in the culture-independent analysis of food microbial ecology. Appl Environ Microbiol . 2013; 79: 3148– 55. Google Scholar CrossRef Search ADS PubMed  European Parliament Council of the European Union. Directive 2006/7/EC of the European Parliament and of the Council of 15 February 2006 concerning the management of bathing water quality and repealing Directive 76/160/EEC. In: Union EPatCotE  (ed). 2006. Handel N, Otte S, Jonker M et al.   Factors that affect transfer of the IncI1 beta-Lactam resistance plasmid pESBL-283 between E-coli strains. PLoS One . 2015; 10, e0123039, doi: https://doi.org/10.1371/journal.pone.0123039. Google Scholar CrossRef Search ADS PubMed  Ihekweazu C, Barlow M, Roberts S et al.   Outbreak of E. coli O157 infection in the south west of the UK: risks from streams crossing seaside beaches. Euro Surveill . 2006; 11: 128– 30. Google Scholar CrossRef Search ADS PubMed  International Organization for Standardization. ISO/TS 16649-3:2005 Microbiology of food and animal feeding stufffs—horizontal method for the enumeration of beta-glucuronidase-positive Escherichia coli Part 3: Most probable number technique using 5-bromo-4-chloro-3-indolyl-beta-D-glucuronide, 2005. International Organization for Standardization. Water quality—enumeration of Escherichia coli and coliform bacteria—Part 1: Membrane filtration method for waters with low bacterial background flora Water quality, Microbiological Methods, 2014a. International Organization for Standardization. Water quality—requirements for the comparison of the relative recovery of microorganisms by two quantitative methods Water quality, Microbiology, 2014b. 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  Kantar TNS The Great Britain Day Visitor 2016 Annual Report, 2017; https://www.visitbritain.org/gb-day-visits-survey-latest-results (01 November 2017, date last accessed). Leonard, A, Zhang, L, Balfour, A et al. Exposure to and colonisation by antibiotic-resistant E. coli in UK coastal water users: Environmental surveillance, exposure assessment, and epidemiological study (Beach Bum Survey). Environ Int . 2018; doi: https://doi.org/10.1016/j.envint.2017.11.003. Leonard A, Zhang L, Balfour A et al.   Human recreational exposure to antibiotic resistant bacteria in coastal bathing waters. Environ Int . 2015; 82: 92– 100. Google Scholar CrossRef Search ADS PubMed  Levy SB, FitzGerald GB, Macone AB Spread of antibiotic-resistant plasmids from chicken to chicken and from chicken to man. Nature . 1976; 260: 40– 2. Google Scholar CrossRef Search ADS PubMed  Li B, Yang Y, Ma L et al.   Metagenomic and network analysis reveal wide distribution and co-occurrence of environmental antibiotic resistance genes. ISME J . 2015; 9: 2490– 502. Google Scholar CrossRef Search ADS PubMed  Logares R, Haverkamp TH, Kumar S et al.   Environmental microbiology through the lens of high-throughput DNA sequencing: synopsis of current platforms and bioinformatics approaches. J Microbiol Methods . 2012; 91: 106– 13. Google Scholar CrossRef Search ADS PubMed  Lukjancenko O, Wassenaar TM, Ussery DW Comparison of 61 sequenced Escherichia coli genomes. Microb Ecol . 2010; 60: 708– 20. Google Scholar CrossRef Search ADS PubMed  Monaghan J, Hutchinson M Monitoring microbial food safety of fresh produce. In: Food Standards Agency, Horitcultural Development Company . Horiticultural Development Company, Kenilworth, Warwickshire, 2010. Odonkor ST, Ampofo JK Escherichia coli as an indicator of bacteriological quality of water: an overview. Microbiol Res . 2013; 4: 5– 11. Google Scholar CrossRef Search ADS   Office for National Statistics. Population estimates for UK, England and Wales, Scotland and Northern Ireland: mid-2016. 2016; https://www.ons.gov.uk/peoplepopulationandcommunity/populationandmigration/populationestimates/datasets/populationestimatesforukenglandandwalesscotlandandnorthernireland (20 October 2017, date last accessed). Pruden A, Pei R, Storteboom H et al.   Antibiotic resistance genes as emerging contaminants: studies in northern Colorado. Environ Sci Technol . 2006; 40: 7445– 50. Google Scholar CrossRef Search ADS PubMed  Public Health England. Accompanying data for annual epidemiological commentary: MRSA, MSSA and E. coli bacteraemia and C. difificile infection data, up to and including financial year 2016 to 2017. 2017; https://www.gov.uk/government/statistics/mrsa-mssa-and-e-coli-bacteraemia-and-c-difficile-infection-annual-epidemiological-commentary (22 November 2017, date last accessed). Review on Antimicrobial Resistance. Antimicrobial Resistance: Tackling a crisis for the health and wealth of nations, 2014; https://amr-review.org/sites/default/files/AMR%20Review%20Paper%20-%20Tackling%20a%20crisis%20for%20the%20health%20and%20wealth%20of%20nations_1.pdf (02 October 2017, date last accessed). Review on Antimicrobial Resistance. Antimicrobials in agriculture and the environment: reducing unnecessary use and waste. 2015; https://amr-review.org/sites/default/files/Antimicrobials%20in%20agriculture%20and%20the%20environment%20-%20Reducing%20unnecessary%20use%20and%20waste.pdf (03 October 2017, date last accessed). Russo TA, Johnson JR Medical and economic impact of extraintestinal infections due to Escherichia coli: focus on an increasingly important endemic problem. Microbes Infect . 2003; 5: 449– 56. Google Scholar CrossRef Search ADS PubMed  Shuval H Estimating the global burden of thalassogenic diseases: human infectious diseases caused by wastewater pollution of the marine environment. J Water Health . 2003; 1: 53– 64. Google Scholar PubMed  Soraas A, Sundsfjord A, Sandven I et al.   Risk factors for community-acquired urinary tract infections caused by ESBL-producing enterobacteriaceae—a case-control study in a low prevalence country. PLoS One . 2013; 8: e69581. Google Scholar CrossRef Search ADS PubMed  Tringe SG, Rubin EM Metagenomics: DNA sequencing of environmental samples. Nat Rev Genet . 2005; 6: 805– 14. Google Scholar CrossRef Search ADS PubMed  Wang J, McLenachan PA, Biggs PJ et al.   Environmental bio-monitoring with high-throughput sequencing. Brief Bioinform . 2013; 14: 575– 88. Google Scholar CrossRef Search ADS PubMed  Yang Y, Jiang X, Chai B et al.   ARGs-OAP: online analysis pipeline for antibiotic resistance genes detection from metagenomic data using an integrated structured ARG-database. Bioinformatics . 2016; 32: 2346– 51. Google Scholar CrossRef Search ADS PubMed  © FEMS 2018. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com

Journal

FEMS Microbiology EcologyOxford 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