Carbon source competition within a wound can significantly influence infection progression.

It is becoming increasingly apparent that commensal skin bacteria have an important role in wound healing and infection progression. However, the precise mechanisms underpinning many of these probiotic interactions remain to be fully uncovered. In this work, we demonstrate that the common skin commensal Cutibacterium acnes can limit the pathogenicity of the prevalent wound pathogen Pseudomonas aeruginosa in vivo. We show that this impact on pathogenicity is independent of any effect on growth, but occurs through a significant down regulation of the Type Three Secretion System (T3SS), the primary toxin secretion system utilized by P. aeruginosa in eukaryotic infection. We also show a down regulation in glucose acquisition systems, a known regulator of the T3SS, suggesting that glucose availability in a wound can influence infection progression. This suggests that introducing carbon source competition within the wound microenvironment may be an effective way to prevent or limit wound infection. P. aeruginosa PA14 was grown in the overnight tube at 37°C at 180 rpm. Cultures were then adjusted to OD=0.1 with LB and added in proportion 1:1 to TSB or the supernatant of C. acnes CCUG 38584 to be grown to OD600=0.6-0.7 at 37°C at 180 rpm. Upon reaching the desired optical density, 1 ml of the cultures was aliquoted and spun down at 5000 rpm for 10 mins, the supernatant was discarded, and the pellets were resuspended in the RNAlater buffer (ThermoFischer). The resuspended cells were stored in the buffer at 4°C overnight. RNA extraction procedure was performed with RNeasy Mini Kit (Qiagen). Extracted RNA was quantified using Nanodrop. The quality of the extraction was assessed with Bioanalyser. Samples were stored at -20°C until shipment to the sequencing facility. Illumina sequencing was performed by Microbial Genome Sequencing Centre (MiGS). Raw fastq files were obtained from the sequencing facility. Adapter and quality trimming was performed by MiGS using bcl2fastq. Reads and their quality was assessed with fastqc and visualised through multiqc (Andrews S, 2010, Ewels et al., 2016). Read mapping was performed with hisat2 with ‘--very-sensitive’ parameter (Kim et al., 2019). Read quantification was performed with featureCounts available within the Subread package (Liao et al., 2014). Read normalisation was performed using the edgeR package in R with the Trimmed Mean of M values algorithm (Robinson et al., 2010). Differential expression analysis was performed using edgeR’s Quasi-Linear F-Test (qlfTest) functionality against treatment groups. All quantified genes were subset by log-fold change (logFC) > |1| and p-value < 0.05 to create a list of differentially expressed genes. All quantified genes were visualised in a volcano plot in ggplot2 R package.