Demonstrating an Integrated Antibiotic Resistance Gene Surveillance Approach in Puerto Rican Watersheds Post-Hurricane Maria

Comprehensive surveillance approaches
are needed to assess sources,
clinical relevance, and mobility of antibiotic resistance genes (ARGs)
in watersheds. Here, we examined metrics derived from shotgun metagenomic
sequencing and relationship to human fecal markers (HFMs; crAssphage,
enterococci) and anthropogenic antibiotic resistance markers (AARMs; intI1, sul1) in three distinct Puerto Rican
watersheds as a function of adjacent land use and wastewater treatment
plant (WWTP) input 6 months after Hurricane Maria, a category V storm.
Relative abundance and diversity of total ARGs increased markedly
downstream of WWTP inputs, with ARGs unique to WWTP and WWTP-impacted
river samples predominantly belonging to the aminoglycoside and β-lactam
resistance classes. WWTP and other anthropogenic inputs were similarly
associated with elevated resistome risk scores and mobility incidence
(M%). Contig analysis indicated a wide variety of mobile β-lactam
ARGs associated with pathogens downstream of WWTP discharge that were
consistent with regional clinical concern, e.g., Klebsiella
pneumoniae contigs containing KPC-2 within an ISKpn6-like transposase. HFMs and AARMs correlated strongly
with the absolute abundance of total ARGs, but AARMs better predicted
the majority of ARGs in general (85.4 versus <2%) and β-lactam
ARGs in particular. This study reveals sensitive, quantitative, mobile,
clinically relevant, and comprehensive targets for antibiotic resistance
surveillance in watersheds.