Detecting transcriptional responses and comparing the virulence of Pseudomonas aeruginosa cystic fibrosis isolates in a mung bean model

A mung bean infection model has previously been shown to differentiate between non-virulent and virulent Pseudomonas aeruginosa bacteria. However, it remains unclear how plant and bacteria adjust their gene expression during infection and whether the mung bean model can be used to compare the virulence of clinical cystic fibrosis (CF) P. aeruginosa lung isolates. Here, we first explored temporal transcriptomics of P. aeruginosa PAO1 and mung bean during an infection. We found that bacterial gene expression followed temporal changes, with an increase in the expression of O-antigen biosynthetic genes, chemotaxis, phosphate intake and phenazine production. Mung bean responded by upregulating genes associated with defence mechanisms and downregulating genes involved in the plant development. From the PAO1 perspective, the core-transcriptomic responses in the mung bean were similar to its responses previously observed in wound and excision and in in vitro media and sputum models, while differed from those observed in the bronchial cell model. Furthermore, we used the mung bean to assess the virulence of 119 clinical P. aeruginosa CF strains originating from the Copenhagen CF clinic. By quantifying bacterial virulence as a reduction in root and shoot growth and weight of the seeds, we found that CF strains isolated at later compared to early stages of lung infections showed higher virulence. This difference corresponded with the higher number of immune modulation-associated virulence genes and lower number of motility and effector genes, present in the genomes of late compared to early isolated CF strains. IMPORTANCEOur results demonstrate that based on PAO1 transcriptional profile, the mung bean model is similar to in vitro and wound infection models but differs from cell and bronchial models. Moreover, the mung bean model can detect virulence differences between clinical P. aeruginosa CF strains, making it a potentially useful high-throughput in vivo model for bacterial virulence screening.