Co-occurance of antibiotic concentrations and resistance gene abundances in the global environment

DOI
  1. Data and methods2.1. Search strategyA literature review was performed by searching the Web of Science platform in May 2018. The titles, abstracts, and keywords were screened using the following search string “antibiotic AND ARG$ AND water”. The symbol “$” represents zero or one character, while “*” represents any group of characters, including no character. The publication year was coerced to equal or <2017 as to encompass complete years. The search returned 428 publications.2.2. Selection criteriaThe suitability of the publications was first assessed by scanning the titles and abstracts. Publications were selected for data extraction only when antibiotic concentrations and resistance genes abundance were measured simultaneously in the samples. Different techniques are currently employed to detect environmental DNA but only publications using quantitative polymerase chain reaction (qPCR) were considered since it has been widely applied and allows gene quantification. This study focused on three main environmental matrices, i.e. surface water, wastewater and sediment. Non-original research publications (e.g. reviews) were not considered but used as a source for cross-references. This process resulted in the selection of 42 publications for data extraction.2.3. Data extractionThe following data were extracted and compiled: antibiotic concentrations, antibiotic-resistance gene copy numbers, environmental matrix type, sampling year, country. The data were collected from tables and texts. Data expressed in figures were extracted by use of WebPlotDigitizer 3.12 (Rohatgi, 2017). When not possible, the authors were contacted to request the numerical data. The mean or median values of replicates from the same samples were collected. Aggregated samples over time or space were excluded. If both descriptors were available, the mean value was selected over the median, given its extensive use. Reported concentrations below the limits of detection or quantification were not considered for analysis. Data from the same samples partitioned into separate publications were also recovered. A total of 256 environmental samples were identified containing 87 antibiotics, 63 ARGs and 3 mobile genetic elements.2.4. Data structure2.4.1. ARG abundanceFor each sample, if not reported in the study, the total 16S rRNA copy number was used to calculate the relative abundance of individual ARGs (Eq. (1)), as well as the total ARG abundance (TARG; Eq. (2)).rARG_x,j=(ARG_x / 16S rRNA_j) (1)TARG_y,j = ∑ x ∈ y (rARG_x,j ) (2)where rARGx, j is the relative abundance of antibiotic-resistance gene x in sample j, ARG_x is the number of copies of gene x, 16S rRNA_j is the number of copies of 16S ribosomal RNA gene in sample j, and TARG_y, j is the total relative abundance of genes x in sample j which confer resistance against antibiotics belonging to therapeutic class y (x ∈ y).2.4.2. Resistance mapping and antibiotic classificationIndividual resistance genes were linked to the individual antibiotics which they confer resistance against, according to the Comprehensive Antibiotic Resistance Database (Jia et al., 2017). Then, these antibiotics were grouped following the Anatomical Therapeutic Chemical (ATC) classification system (Table 1). Certain genes allow phenotypic resistance to more than one specific antibiotic, like extended-spectrum β-lactamase genes such as blaCTX. In such cases, these genes were assumed to be associated with all individual antibiotics belonging to a class (Table 1). Antibiotic transformation products suspected of antibacterial activity were included in the analysis (e.g. dehydrated erythromycin). Besides individual rARGs, the relative abundance of genetic elements intI1, intI2 and tnpA was also considered because of their important role as facilitators of gene mobilization and spread of antibiotic resistance (Boerlin and Reid-Smith, 2008).2.4.3. Antibiotic selective pressureAll antibiotics were standardized to concentrations of ng/l for surface water and wastewater, and ng/kg dw for sediment. Concentrations of individual antibiotics were used to determine the resistance selection pressure potential by applying representative PNEC values, according to Bengtsson-Palme and Larsson (2016) (Eq. (3)). Sediment PNEC values were calculated using organic carbon-normalized sorption coefficients estimates from the software KOCWIN v2.01 (EPA, 2015) at an assumed 5.8% organic carbon content (RIVM, 2015). To allow a coherent comparison across samples, a measure of total selection pressure potential was calculated (Eq. (4)).ASP_i,j = (MEC_i,j / PNEC_i) (3)TASP_y,j = ∑ i ∈ y (ASP_i,j) (4)where ASP_i, j is the selection pressure potential of antibiotic i in sample j, MEC_i, j is the measured environmental concentration of antibiotic i in sample j, PNEC_i is the predicted no effect concentration for selection of resistance by antibiotic i, and TASP_y, j is the total selection pressure potential in sample j of antibiotics i belonging to therapeutic class y (i ∈ y).2.4.4. Environmental matricesSamples of WWTP influents, hospital wastewater, urban sewage and industrial wastewater origin were classified as ‘wastewater’. Water samples collected from rivers, estuaries, water reservoirs, bays, lakes and creeks were classified as ‘surface water’. The environmental matrix ‘sediments’, includes sediment samples from rivers, estuaries, lakes, water reservoirs, bays and coast. Wastewater was included in this study for comparability since it is a heavily antibiotic and ARG loaded matrix of anthropogenic origin.2.4.5. DatabaseA final database was created comprising 342 unique entries for each antibiotic class nested by sample and study. These represent 26 studies (Study), 11 countries (Country), 3 environmental matrices (Matrix), 197 samples (Sample), 10 sampling years (Year) and 11 antibiotic classes (Class).
Identifier
DOI https://doi.org/10.17026/DANS-ZG8-M5CC
Metadata Access https://lifesciences.datastations.nl/oai?verb=GetRecord&metadataPrefix=oai_datacite&identifier=doi:10.17026/DANS-ZG8-M5CC
Provenance
Creator D.J. Da Silva Tavares Duarte; R. Oldenkamp; A.M.J. Ragas
Publisher DANS Data Station Life Sciences
Contributor RU Radboud University
Publication Year 2022
Rights CC-BY-4.0; info:eu-repo/semantics/openAccess; http://creativecommons.org/licenses/by/4.0
OpenAccess true
Contact RU Radboud University
Representation
Resource Type Dataset
Format application/pdf; application/vnd.oasis.opendocument.spreadsheet; application/zip
Size 836722; 8142; 27198; 176574; 197598
Version 2.0
Discipline Earth and Environmental Science; Environmental Research; Geosciences; Life Sciences; Medicine; Natural Sciences