Proteomic profiling of Mycobacterium tuberculosis identifies nutrient-starvation-responsive toxin-antitoxin systems

Abstract

In order to successfully enter the latent stage, Mycobacterium tuberculosis must adapt to conditions such as nutrient limitation and hypoxia. In vitro models that mimic latent infection are valuable tools for describing the changes in metabolism that occur when the bacterium exists in a non-growing form. We used two complementary proteomic approaches, label-free LC-MS/MS analysis and two-dimensional difference gel electrophoresis, to determine the proteome profile of extracellular proteins from M. tuberculosis cultured under nutrient starvation. Through the label-free LC-MS/MS analysis of fractionated samples, 1176 proteins were identified from culture filtrates of log phase and nutrient-starved cultures, and the protein levels of 230 proteins were increased in nutrient-starved culture filtrates, whereas those of 208 proteins were decreased. By means of Gene Ontology clustering analysis, significant differences in the overall metabolism during nutrient starvation were detected. Notably, members of the toxin-antitoxin systems were present in larger quantities in nutrient-starved cultures, supporting a role for these global modules as M. tuberculosis switches its metabolism into dormancy. Decreased abundance of proteins involved in amino acid and protein synthesis was apparent, as well as changes in the lipid metabolism. Further analysis of the dataset identified increased abundance of lipoproteins and decreased abundance of ESAT-6 family proteins. Results from the two-dimensional difference gel electrophoresis proteomics demonstrated overall agreement with the LC-MS/MS data and added complementary insights about protein degradation and modification.

Fig. 1.

Functional protein groups were made according to the classification of the TubercuList server, and for each functional group the identified proteins are presented as increased (>1.5-fold, p < 0.01, red), decreased (>1.5-fold, p < 0.01, green) or unchanged under starvation conditions (yellow).

Fig. 2.

Presence of Esx proteins in triplicate CF from log phase and 6-week-starved cultures. Ten micrograms of each CF sample was tested via Western blot analysis with anti-EsxH/EsxR mouse monoclonal antibody, anti-EsxA mouse monoclonal antibody, and anti-EsxB rabbit polyclonal serum.

Fig. 3.

Representative two-dimensional DIGE image of culture filtrate proteins from log phase and starvation conditions. The total amount of protein used for CyDye labeling was 50 μg. The numbers on the left indicate molecular mass markers, and the numbers along the bottom of the image indicate the pI range. The numbered protein spots were excised from two-dimensional DIGE gels post-stained with silver and subjected to MS identification; for the underlined spot numbers, a successful identification was obtained, whereas no significant match was obtained in the Mascot search engine for spots not underlined. The spot numbering refers to the number of the spot assigned by the Imagemaster software. Protein spots that were more abundant in log phase appear as green, and spots more abundant under starvation conditions appear as red.

Fig. 4.

Comparison of fold change observed with the two-dimensional DIGE method and the label-free LC-MS/MS approach. Each bar shows the observed fold change in abundance (starved/log phase CF) as detected via the two-dimensional DIGE method (light gray bars representing the mean fold change of individual spots) and via the LC-MS/MS method (dark gray bars) for the individual proteins.