In vivo versus in vitro protein abundance analysis of Shigella dysenteriae type 1 reveals changes in the expression of proteins involved in virulence, stress and energy metabolism

Abstract

Background: Shigella dysenteriae serotype 1 (SD1) causes the most severe form of epidemic bacillary dysentery. Quantitative proteome profiling of Shigella dysenteriae serotype 1 (SD1) in vitro (derived from LB cell cultures) and in vivo (derived from gnotobiotic piglets) was performed by 2D-LC-MS/MS and APEX, a label-free computationally modified spectral counting methodology.

Results: Overall, 1761 proteins were quantitated at a 5% FDR (false discovery rate), including 1480 and 1505 from in vitro and in vivo samples, respectively. Identification of 350 cytoplasmic membrane and outer membrane (OM) proteins (38% of in silico predicted SD1 membrane proteome) contributed to the most extensive survey of the Shigella membrane proteome reported so far. Differential protein abundance analysis using statistical tests revealed that SD1 cells switched to an anaerobic energy metabolism under in vivo conditions, resulting in an increase in fermentative, propanoate, butanoate and nitrate metabolism. Abundance increases of transcription activators FNR and Nar supported the notion of a switch from aerobic to anaerobic respiration in the host gut environment. High in vivo abundances of proteins involved in acid resistance (GadB, AdiA) and mixed acid fermentation (PflA/PflB) indicated bacterial survival responses to acid stress, while increased abundance of oxidative stress proteins (YfiD/YfiF/SodB) implied that defense mechanisms against oxygen radicals were mobilized. Proteins involved in peptidoglycan turnover (MurB) were increased, while β-barrel OM proteins (OmpA), OM lipoproteins (NlpD), chaperones involved in OM protein folding pathways (YraP, NlpB) and lipopolysaccharide biosynthesis (Imp) were decreased, suggesting unexpected modulations of the outer membrane/peptidoglycan layers in vivo. Several virulence proteins of the Mxi-Spa type III secretion system and invasion plasmid antigens (Ipa proteins) required for invasion of colonic epithelial cells, and release of bacteria into the host cell cytosol were increased in vivo.

Conclusions: Global proteomic profiling of SD1 comparing in vivo vs. in vitro proteomes revealed differential expression of proteins geared towards survival of the pathogen in the host gut environment, including increased abundance of proteins involved in anaerobic energy respiration, acid resistance and virulence. The immunogenic OspC2, OspC3 and IpgA virulence proteins were detected solely under in vivo conditions, lending credence to their candidacy as potential vaccine targets.

Figure 1

Euler/Venn diagram representations of S. dysenteriae serotype 1 (SD1) proteins. Of the 4502 proteins predicted for the SD1 genome, 1761 proteins were identified at a 5% false discovery rate (FDR), with 1480 proteins identified from the in vitro analysis, and 1505 proteins from the in vivo analysis.

Figure 2

Sub-cellular localization (SCL) of SD1 proteins. SCL of 4502 proteins encoded by the SD1 genome was predicted using the bioinformatic algorithms PSORTb, SignalP, TatP, TMHMM, BOMP, LipoP and KEGG. 350 outer and inner membrane proteins corresponding to ca. 38% of the SD1 membrane proteome, and 1410 cytoplasmic and periplasmic proteins representing ca. 39% of SD1 soluble proteins were identified.

Figure 3

SD1 differential protein expression analysis using the two-tailed Z-test. Approximately 300 proteins were found to be differentially expressed at 99% confidence, including 151 in vivo and 142 in vitro SD1 proteins using the two-tailed Z-test utility in the APEX tool application.

Figure 4

Hierarchial clustering (HCL) analysis of differentially expressed SD1 proteins based on APEX abundance values using MeV. Protein abundance values from the in vitro sample are represented on the left, with in vivo protein abundances on the right. Abundance magnitude is depicted as a color gradient, with red indicating an increase in protein abundance, green indicating a corresponding decrease in abundance, and black for the median level of abundance. Based on biological interests, example clusters are enlarged to depict differentially expressed proteins.

Figure 5

Representation of functional role categories of SD1 proteins. Proteins identified from 2D-LC-MS/MS experiments of S. dysenteriae cells were analyzed based on protein functional assignments in the CMR database for the genome of SD1 strain Sd197. Distribution of role categories of SD1 proteins cultured from stationary phase cells (in vitro) are shown in the panel on the left (5A) and cells isolated from gut environment of infected piglets (in vivo) are depicted on the right (5B).