Legionella pneumophila (LP) is the major etiologic bacterium of Legionnaires' disease. This lung infection causes up to 10% of pneumonia with a mortality rate of 15%. The purpose of this study is to perform comparative proteomic analyses to answer biological questions such as showing genes differential expression between strains, study effectors secretion or study the clinically observed evolution of the infectious strain over time. In some cases the treatment is not efficient to stop the infection. A early sample was taken at the time of infection diagnosis and a late one a month after. A third one was produced by growing the early strain in human serum to simulate the late strain.
Biological triplicates of 15 bacterial strains in exponential growth phase were analyzed with data-independent acquisition LC-MS/MS method optimized for maximum protein identification. Raw data were processed with DIA-NN. The result matrix was processed with Python to select only proteotypic precursors, filter the proteins with a protein q-value less than 1% and calculate the maxLFQ quantities mean on the triplicates for each protein and for each strain.
A method of 31 min gradient, 15 ms accumulation time for MS2 spectra and 65 variable windows covering 400 to 900 MS1 m/z range, identified an average of 63% of the proteome. Analysis of triplicates gave results on the identification of few proteins similarly expressed in late and simulated late strains, indicating a biological evolution of the strain during infection. About 20% of effector proteins and a majority of the Dot/Icm secretion system proteins are detected in bacteria in exponential growth phase. However, none of them were found in the culture supernatant. This would imply that no effectors are secreted during the replication phase of LP infection cycle. The next step would be to analyze a post-exponential growth phase sample to track the evolution of effectors secretion.