Spatial transcriptomics identifies novelPseudomonas aeruginosavirulence factors

Abstract

The interactions of a pathogen with the host take place within the confines of infected tissues. However, current technologies do not allow for contextual characterization of infection. We report a unified spatial transcriptomic analysis detecting pathogen- and host-transcripts simultaneously. We showcase our approach using aP. aeruginosa-induced ocular infection, where we observed differential enrichment of host and bacterial transcripts at specific anatomical sites. The ridge regression model, trained using a minimum of 150 host gene features, accurately predicted bacterial burden in tissue regions. The tissue-in-depth gene expression enrichment analysis identified bacterial transcript PA2590, encoding a currently uncharacterized gene as deeply penetrant. Infection experiments with the PA2590 transposon deletion mutant strain caused less disease than the WT strain. Comparative structural analysis identified the PA2590 gene product as an iron-scavenging and cobalamin transporter. Cumulatively, our data highlight coordinated spatial interplay between the host and the pathogen representing an approach to identify novel virulence traits.