Pangenome Analysis of Proteus mirabilis Reveals Lineage-Specific Antimicrobial Resistance Profiles and Discordant Genotype-Phenotype Correlations
Discuss this preprint
Start a discussion What are Sciety discussions?Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
Urinary tract infections (UTIs) impose a large healthcare burden, with escalating antimicrobial resistance (AMR) and treatment failure. Proteus mirabilis is an under-characterized and challenging UTI pathogen due to intrinsic resistances and biofilm formation. To understand P. mirabilis population genomics, we combined pangenome analysis, in silico AMR predication, and phenotypic antimicrobial susceptibility testing (AST) across 1,027 P. mirabilis genomes derived from human urine specimens. This revealed a mosaic pangenome driven by extensive accessory genome plasticity. Multilocus sequence typing (MLST) identified 213 MLSTs with only 7% having ≥10 genomes, highlighting strain diversity. AMR gene profiles were largely lineage specific, with 25% of genomes harboring resistances for >6 antimicrobial subclasses. ST135 was identified as a highly MDR lineage, with 95% of genomes carrying ≥16 resistance genes. Mobile genetic element (MGE) analysis of 22 clinical isolates with complete, reference level genomes revealed that Tn7 transposons, IS26-mediated genomic islands, and class 1 integrons act as vehicles for high AMR gene dissemination, including IS26-mediated gene stacking within a Proteus mirabilis Genomic Resistance Island 1 ( PmGRI1) in ST135 isolates. While presence of genes like aph(3’)-la reliably predicted kanamycin resistance, discordance for antibiotics such as trimethoprim-sulfamethoxazole and chloramphenicol revealed that AMR gene stacking, regulatory context, and intrinsic mechanisms, like efflux pumps, modulate phenotypic outcomes. In summary, our study provides a comprehensive and phenotypic resolution of P. mirabilis AMR, establishing that resistance architecture is lineage structured, MGE-driven, and phenotypically non-deterministic. We emphasize the need to shift towards standardized, genome-informed surveillance framework to translate into diagnostic and therapeutic strategies.
