Antibiotic Resistance and Sewage-Associated Marker Genes in Untreated Sewage and a River Characterized During Baseflow and Stormflow

Abstract

Since sewage is a hotspot for antibiotic resistance genes (ARGs), the identification of ARGs in environmental waters impacted by sewage, and their correlation to fecal indicators, is necessary to implement management strategies. In this study, sewage treatment plant (STP) influent samples were collected and analyzed using quantitative polymerase chain reaction (qPCR) to investigate the abundance and correlations between sewage-associated markers (i.e., Bacteroides HF183, Lachnospiraceae Lachno3, crAssphage) and ARGs indicating resistance to nine antibiotics (belonging to aminoglycosides, beta-lactams, sulfonamides, macrolides, and tetracyclines). All ARGs, except bla _VIM, and sewage-associated marker genes were always detected in untreated sewage, and ermF and sul1 were detected in the greatest abundances. intl1 was also highly abundant in untreated sewage samples. Significant correlations were identified between sewage-associated marker genes, ARGs and the intl1 in untreated sewage (τ = 0.488, p = 0.0125). Of the three sewage-associated marker genes, the BIO-ENV procedure identified that HF183 alone best maximized correlations to ARGs and intl1 (τ = 0.590). Additionally, grab samples were collected from peri-urban and urban sites along the Brisbane River system during base and stormflow conditions, and analyzed for Escherichia coli, ARGs, the intl1, and sewage-associated marker genes using quantitative polymerase chain reaction (qPCR). Significant correlations were identified between E. coli, ARGs, and intl1 (τ = 0.0893, p = 0.0032), as well as with sewage-associated marker genes in water samples from the Brisbane River system (τ = 0.3229, p = 0.0001). Of the sewage-associated marker genes and E. coli, the BIO-ENV procedure identified that crAssphage alone maximized correlations with ARGs and intl1 in river samples (τ = 0.4148). Significant differences in E. coli, ARGs, intl1, and sewage-associated marker genes, and by flow condition (i.e., base vs. storm), and site types (peri-urban vs. urban) combined were identified (R = 0.3668, p = 0.0001), where percent dissimilarities between the multi-factorial groups ranged between 20.8 and 11.2%. Results from this study suggest increased levels of certain ARGs and sewage-associated marker genes in stormflow river water samples compared to base flow conditions. E. coli, HF183 and crAssphage may serve as potential indicators of sewage-derived ARGs under stormflow conditions, and this merits further investigation. Data presented in this study will be valuable to water quality managers to understand the links between sewage pollution and ARGs in urban environments.

Keywords: antibiotic resistance; human health risks; microbial source tracking; sewage pollution; stormwater.

FIGURE 1

Map of the study sites along the Brisbane River system and its tributaries Oxley Creek, and Boggy Creek, located in Brisbane, Australia. Site OX1 is located downstream of a WWTP, site BR8 is located in proximity to a storm water drain, and sites BR9, BR10, BR11, and BR12 are located downstream of hospitals.

FIGURE 2

Abundance (log₁₀ copies/L) of ARGs (aacA, bla_ctx–m–32, bla_KPC, ermF, sul1, sul2, tet(M), and vanA), intl1, and sewage-associated marker genes (HF183, crAssphage CPQ_056, and Lachno3) in untreated sewage samples, which were always detected in quantifiable abundances. ARG bla_VIM was positively detected, but not in quantifiable abundances (between 2.42- and 2.59-log₁₀ copies/L), in only two of the six untreated sewage samples. +denotes mean while the outer box lines represent 25th and 75th percentiles and the whiskers extended to the range, and lines inside the boxes represent median values.

FIGURE 3

Median abundance (log₁₀ copies/L) with maximum and minimum abundances (error bars) reported for E. coli (PLOD = 2.00 log₁₀ copies/L), aacA (PLOD = 2.30 log₁₀ copies/L), bla_ctx–m–32 (PLOD = 2.60 log₁₀ copies/L), bla_KPC (PLOD = 2.30 log₁₀ copies/L), bla_VIM (PLOD = 2.42 log₁₀ copies/L), ermF (PLOD = 2.90 log₁₀ copies/L), intl1 (PLOD = 2.30 log₁₀ copies/L), sul1 (PLOD = 2.00 log₁₀ copies/L), sul2 (PLOD = 3.00 log₁₀ copies/L), tet(M) (PLOD = 2.60 log₁₀ copies/L), and vanA (PLOD = 2.90 log₁₀ copies/L), and sewage-associated marker genes (HF183, crAssphage CPQ_056, and Lachno3; all PLOD = 2.00 log₁₀ copies/L) in water samples collected from peri-urban and urban sites from the Brisbane River system during the base and stormflow (solid-filled shape). If a particular microbial target had left-censored values, the uncertainty of the minimum abundance was depicted with a straight, vertical line. If it was not possible to calculate a median abundance (>80% left-censored), then the maximum abundance measured was depicted with an “X” inside the site/flow shape. When a microbial target was only detected below the limit of quantification, then a half-filled shape depicted the limit of quantification. Finally, an un-filled shape represents 100% left-censored values, and the process limit of detection is depicted.

FIGURE 4

Non-metric multidimensional scaling (NMDS) plot of E. coli, ARGs, and sewage-associated markers, measured in water samples (n = 56) collected from peri-urban sites during baseflow (black circle) and stormflow (red triangle), as well as and urban sites during baseflow (green square) and stormflow (blue diamond) along the Brisbane River system, Brisbane, Australia.