ABSTRACT: Bacterial pathogens use various strategies to interfere with host cell functions. Among these strategies, bacteria induce transcriptional changes, in order to modify the set of proteins synthetized by the host cell, or target pre-existing proteins by modulating their post-translational modifications or by triggering their degradation. Protein levels variations in host cells during infection integrates both transcriptional and post-transcriptional regulations induced by pathogens. Here, we focused on host proteome alterations induced by the bacterial pathogen Listeria monocytogenes, and more specifically the Listeria toxin Listeriolysin O (LLO). In order to characterize host proteome alterations induced by LLO, we performed a shotgun proteomics analysis on HeLa cells treated or not with the LLO toxin. To this end, in a first analysis we used SILAC labelling (stable isotope labelling by amino acids in cell culture) to compare in a quantitative manner the protein content from two differentially treated cell populations: one control population and one population incubated with a sublytic dose of LLO (3 nM). We decided to expose cells to LLO during only 20 min to limit protein level changes resulting from transcriptional changes. We performed two independent experiments (experiment #1 and experiment #2) with swapped SILAC labelling to rule out putative labelling-dependent effects. A total of 2,009 and 2,577 proteins was quantified in each independent experiment, respectively, and 1,834 proteins were quantified in both experiments. Among these 1834 proteins, we identified a total of 151 proteins for which protein levels were consistently decreased in cells treated with LLO (i.e. with a log2 control/LLO ratio >0.5 in both experiments). In addition to this preliminary study, we carried out a second experiment using label-free quantitative shotgun proteomics to compare protein abundance in cells treated or not with LLO. Four independent biological replicates were included in this second screen in order to perform a robust statistical analyses of downregulated proteins. Among the 2,973 proteins that were quantified in all independent replicates, we identified a total of 149 proteins (5.0 %) for which protein levels were significantly decreased in cells treated with LLO. In contrast to these decreases, only 16 proteins (0.5 %) showed significant increased levels in LLO-treated cells. This result confirmed that LLO remodels the cell proteome mainly by decreasing the level of host targets. To assess whether the host proteasome was involved in the decrease of these 149 proteins, we repeated our label-free proteomics analysis on HeLa cells pre-treated with proteasome inhibitors before exposure to LLO. Interestingly, we observed that the vast majority (i.e. 83%) of host proteins identified as strongly downregulated in response to LLO also displayed significant decreased levels in the presence of proteasome inhibitors, indicating that the majority of observed decreases in host protein levels induced by LLO are not due to proteasome-mediated degradation. In follow-up experiments, we validated that LLO decreases in particular the protein level of two E2 ubiquitin enzymes, UBE2K and UBE2N, leading to major changes in the host ubiquitylome. This toxin-induced proteome remodeling involves post-transcriptional regulations, as no modification in the transcription levels of the corresponding genes was observed. These decreases in host protein levels were observed in different epithelial cell lines but not in macrophages. In addition, we could show that Perfringolysin O, another bacterial pore-forming toxin similar to LLO, also induces host proteome changes. Taken together, our data show that different bacterial pore-forming toxins induce deep proteome remodeling, that may impair host epithelial cell functions.
Impens Francis 
