Pseudomonas aeruginosa is a prominent opportunistic pathogen whose virulence is closely linked to its Type III Secretion System (T3SS), a specialized apparatus that injects effector proteins into host cells. T3SS chaperones are essential for stabilizing, delivering, and regulating T3SS expression. Multiple T3SS effectors and regulators have been identified and characterized in the P. aeruginosa reference strains such as PAO1 and PA14. However, the full repertoire of T3SS chaperones and their interacting partners in the pan-genomes of thousands of P. aeruginosa isolates remains to be explored. Here, I systematically screened over 15,000 high-quality P. aeruginosa genomes to identify T3SS chaperones using structure-assisted homology searches. Subsequently, I applied AlphaFold2 and AlphaFold3 to predict protein-protein interactions between chaperones and other proteins in the pan-proteome. A benchmark analysis suggests that our approach can effectively distinguish true interacting partners of T3SS chaperones on a proteome-wide scale. Our analysis identified several high-confidence candidate T3SS effectors and regulators, including putative lipoproteins, calcium-binding proteins, and transcriptional regulators, that are potentially involved in host adaptation, T3SS regulation, and virulence modulation. I also found T3SS chaperone homologs that may serve as non-cognate antitoxins or virulence regulators. These findings demonstrate the power of combining genomics with deep learning-based structure and interaction prediction to uncover hidden components of bacterial virulence pathways and provide new targets for antimicrobial development.
J D Zhang (Sun,) studied this question.