Date Published: January 24, 2019
Publisher: Public Library of Science
Author(s): Andrew Cameron, Ruth Barbieri, Ron Read, Deirdre Church, Emelia H. Adator, Rahat Zaheer, Tim A. McAllister, Asad U. Khan.
The biocide triclosan is in many consumer products and is a frequent contaminant of wastewater (WW) such that there is concern that triclosan promotes resistance to important antibiotics. This study identified functional mechanisms of triclosan resistance (TCSR) in WW metagenomes, and assessed the frequency of TCSR in WW-derived and clinical isolates of Escherichia coli and Enterococcus spp. Metagenomic DNA extracted from WW was used to profile the microbiome and construct large-insert cosmid libraries, which were screened for TCSR. Resistant cosmids were sequenced and the TCSR determinant identified by transposon mutagenesis. Wastewater Enterococcus spp. (N = 94) and E. coli (N = 99) and clinical Enterococcus spp. (N = 146) and vancomycin-resistant E. faecium (VRE; N = 149) were collected and tested for resistance to triclosan and a comprehensive drug panel. Functional metagenomic screening revealed diverse FabV homologs as major WW TCSR determinants. Resistant clones harboured sequences likely originating from Aeromonas spp., a common WW microbe. The triclosan MIC90s for E. coli, E. faecalis, and E. faecium isolates were 0.125, 32, and 32 mg/L, respectively. For E. coli, there was no correlation between the triclosan MIC and any drug tested. Negative correlations were detected between the triclosan MIC and levofloxacin resistance for E. faecalis, and between triclosan and vancomycin, teicoplanin, and ampicillin resistance for E. faecium. Thus, FabV homologs were the major contributor to the WW triclosan resistome and high-level TCSR was not observed in WW or clinical isolates. Elevated triclosan MICs were not positively correlated with antimicrobial resistance to any drug tested.
Triclosan is a synthetic chlorinated bisphenol antimicrobial drug, commonly referred to as a biocide, and is effective against a wide variety of microorganisms, including bacteria, fungi, and apicomplexan parasites [1, 2]. Many clinical and personal hygiene products, such as soaps, sanitizers, and toothpaste include triclosan at concentrations up to 0.3% (300 g/L or ~1 M). Triclosan is also found in human urine , milk , wastewater (WW) and aquatic systems . The widespread inclusion of triclosan in consumer products is controversial , as it is structurally similar to T3 and T4 thyroid hormones [6, 7] and has the potential to promote antimicrobial resistance (AMR) [5, 8].
This study characterized the TCSR gene reservoir in an environment likely impacted by triclosan. Our results highlighted the prevalence of triclosan-refractory ENRs potentially linked to Aeromonas spp. These are widespread microbes that are frequently found in both WW and treated drinking water, and are associated with fish, shellfish, and waterfowl . Several species, including A. hydrophila, A. caviae, A. media, A. schubertii, A. veronii biovar sobria, and A. veronii biovar veronii are emerging human pathogens and the causative agents of gastroenteritis, wound and soft tissue infections, necrotizing fasciitis, urinary tract infections, cystic fibrosis, and septicemia [38, 39]. The fish pathogen, A. salmonicida subsp. salmonicida harbors a TCSR FabV ENR homolog with high identity with ENRs in other fish pathogens, including Edwardsiella spp., Vibrio spp., and Flavobacterium spp. . Our functional work suggests that TCSR FabV homologs may be ubiquitous in Aeromonas spp. Intrinsic TCSR in the related γ-Proteobacteria Pseudomonas aeruginosa, is also due to FabV [10, 17]. Although Pseudomonadales were more prevalent than Aeromonadales in our WW samples, no TCSR determinants were predicted to have originated from Pseudomonadales. This may be because the predominant WW Pseudomonadales were Acinetobacter, which typically encode triclosan-sensitive fabI in lieu of fabV. Similarly, FabI is the terminal ENR in Acidovorax, the most abundant Burkholderiales in this WW metagenome. Several fabV had highly divergent sequences from previously characterized FabV in V. cholerae, P. aeruginosa, and A. salmonicida. With the exception of the fabV in cosmid Tri-4, which was identical to that found in some Yersinia spp., all but one shared the most identity with putative fabV in Aeromonas spp. and Serratia spp. Although rare, Y. intermedia and Y. enterocolitica were present in the WW microbiome (0.12 ± 0.032% of total reads) and only 17 reads (out of ~140M) mapped to fabV. Thus, the cosmid library screen functionally selected a rare insert, validating comprehensive functional genomics approaches as powerful methods of identifying new genes or resistance determinants. Despite this, inherent drawbacks of using functional screening of metagenomic libraries for studying TCSR include the fact that they may be unable to reveal the true depth and variety of TCSR (and other AMR) mechanisms, particularly from uncultured organisms whose genes might be incompatible with the library host strain genetics. For example, if TCSR determinants were not expressed or translated, toxic, encoded by a large number of non-contiguous genes, or otherwise irrelevant in the E. coli host they would not be identified. Additionally, metagenomic libraries frequently exhibit cloning bias compared to the original sample, and are time-consuming and challenging to construct .