Radmila Novakovic1*, Milos Jovicevic2, Dusan Kekic3, Svetlana Grujić4, Ljiljana Tolic Stojadinovic5, Nemanja Mirkovic6, Olja Sovljanski7, Aleksandra Vitkovac1, Petra Stojanovic1, Milos Veselinovic1, Ivan Vicic8, Milica Mirkovic9, Nedeljko Karabasil8, Natasa Opavski2 and Ina Gajic2
1Institute of Molecular Genetics and Genetic Engineering, University of Belgrade, Serbia
2Medical Faculty, University of Belgrade, Belgrade, Serbia
3Institute of MicrobioMedical Faculty, University of Belgrade, Belgrade, Serbia
4Faculty of Technology and Metallurgy, University of Belgrade, Belgrade, Serbia
5Innovation Centre of the Faculty of Technology and Metallurgy, Belgrade, Serbia
6Faculty of Agriculture University of Belgrade, Belgrade, Serbia
7Faculty of Technology University of Novi Sad, Serbia
8Faculty of Veterinary Medicine, University of Belgrade, Belgrade, Serbia
9Faculty of Agriculture, University of Belgrade, Belgrade, Serbia
radmila.novakovic [at] imgge.bg.ac.rs
Abstract
Urban wastewater is a convergence point for antibiotic residues and antibiotic resistance genes (ARGs) drained from human, clinical and environmental sources, and is increasingly used as an integrative matrix for antimicrobial resistance surveillance. Chemical quantification with shotgun metagenomics on the same sampling point strengthens the interpretation of the local resistome and offers a more biologically meaningful readout than either approach in isolation. In this work we combined HPLC–MS/MS quantification of nine antibiotics with shotgun metagenomic characterisation of a single grab sample collected from a major urban collector Dorcol in Belgrade, Serbia. Reads were taxonomically profiled with Kraken2 and the resistome was screened in parallel with two complementary tools, CARD–RGI and ResFinder. HPLC–MS/MS detected six out of nine target antibiotics in the sample (ng/L), dominated by ciprofloxacin (1940.5), followed by doxycycline (77.4), trimethoprim (70.0), sulfamethoxazole (47.7), amoxicillin (20.1) and azithromycin (13.9). Gentamicin, colistin and ceftriaxone were below the limit of detection. Metagenomic sequencing yielded ~6.07 million reads (90.07% bacterial), with Bacillota (27.96%), Pseudomonadota (26.43%) and Bacteroidota (20.98%) dominating, and the top genera (Acinetobacter 8.94%, Prevotella 4.65%, Bacteroides 4.32%, Aeromonas 3.51%) matching taxonomic signatures consistently reported for raw municipal influents. CARD–RGI identified 1,402 ARO terms (159 at ≥80% gene coverage) and ResFinder confirmed 86 high-confidence acquired ARGs across 14 drug classes. The resistome was internally consistent with the chemical profile: the classes detected at the highest concentrations corresponded to the most abundant and most diverse resistance determinants, while colistin, undetected chemically, yielded no resistance genes. Beyond this expected baseline, the sample carried clinically significant last-resort determinants, including extended-spectrum and carbapenem-hydrolysing beta-lactamases (blaVEB-1, blaGES-5, blaOXA-427), the tigecycline-inactivating tet(X), and cfr(C), conferring cross-resistance to linezolid, phenicols and pleuromutilins. The integrated chemical–metagenomic profile of this single collector point falls within the range expected for untreated urban wastewater, while the recovery of last-resort resistance markers confirms that clinically critical determinants are already circulating in the local sewerage system. Although based on a single grab sample, the study establishes a methodological and gene-prioritizing method for subsequent longitudinal, multi-site surveillance of urban wastewater in Serbia.
Keywords: OneHealth, Wastewater, Metagenomics, Resistome, ResFinder