Proteomics from compartment-specific APEX2 labeling in Mycobacterium tuberculosis reveals Type VII secretion substrates in the cell wall

Abstract

The cell wall of mycobacteria plays a key role in interactions with the environment. Its ability to act as a selective filter is crucial to bacterial survival. Proteins in the cell wall enable this function by mediating the import and export of diverse metabolites, from ions to lipids to proteins. Identifying cell wall proteins is an important step in assigning function, especially as many mycobacterial proteins lack functionally characterized homologues. Current methods for protein localization have inherent limitations that reduce accuracy. Here we showed that although chemical labeling of live cells did not exclusively label surface proteins, protein tagging by the engineered peroxidase APEX2 within live Mycobacterium tuberculosis accurately identified the cytosolic and cell wall proteomes. Our data indicate that substrates of the virulence-associated Type VII ESX secretion system are exposed to the periplasm, providing insight into the currently unknown mechanism by which these proteins cross the mycobacterial cell envelope.

Keywords: APEX2; ESX; Type VII secretion; cell wall; mycobacteria; periplasm; proximity labeling; tuberculosis.

Figure 1.. Biotinylation of proteins in Mtb by Cyt-APEX2, Sec-APEX2 or N-hydroxysuccinimide reagents.

(A) Schematic of compartment-specific biotinylation of proteins in Mtb expressing Cyt-APEX2 or Sec-APEX2. Figure created with BioRender.com. (B) Mtb encoding inducible Cyt-APEX2 or Sec-APEX2 was cultured without (−) or with (+) theophylline and subjected to the labeling protocol with biotin-phenol. Image on the right is the same blot contrasted to highlight Sec-APEX2-dependent biotinylation. (C) Mtb whole cells were treated with sulfo-NHS-biotin (Sulfo) or NHS-PEG-biotin (PEG). For all blots, biotinylation in crude lysates were detected with streptavidin. Immunoblots are (B) representative of >3 independent experiments (see Figure S1 for additional replicates) and (C) from one experiment for sulfo-NHS-biotin and representative of 3 independent experiments for NHS-PEG-biotin.

Figure 2.. Validation of compartment-specific labeling by Cyt-APEX2, Sec-APEX2 or N-hydroxysuccinimide reagents.

Biotinylated proteins were enriched from crude lysates obtained from Mtb treated as follows: (A) Mtb encoding inducible Cyt-APEX2 or Sec-APEX2 were cultured without (−) or with (+) theophylline and subjected to the labeling protocol with biotin-phenol. (B) Mtb whole cells were treated with sulfo-NHS-biotin (Sulfo) or NHS-PEG-biotin (PEG). Immunoblots are (A) representative of 2 independent experiments (see Figure S2 for additional replicate) and (B) from one experiment for sulfo-NHS-biotin and representative of 2 independent experiments for NHS-PEG-biotin. The fold increase in biotinylation upon induction of APEX2 or treatment with NHS (“Ratio”) was calculated by taking the ratio of the + / − theophylline output intensities after normalizing to the corresponding inputs.

Figure 3.. Identification and analysis of Cyt, Sec, and NHS proteomes.

One-side volcano plots showing proteins detected by label-free affinity purification-mass spectrometry as a function of SAINT score and fold change from cultures (A, B) +theophylline vs. –theophylline of Mtb encoding inducible Cyt-APEX2 (Cyt) or Sec-APEX2 (Sec) or (C) of Mtb treated with NHS-PEG-biotin (NHS) vs. DMSO vehicle control. Proteins with SAINT score > 0.9 (orange circles) were considered as labeled (A,B) by APEX2 or (C) by NHS-PEG-biotin. ESX secretion-related proteins Esx (blue) and Esp (black) proteins are explicitly labeled. Multiple Esx proteins are shown for a single data point when redundant peptides precluded unique assignments. Data shown were derived from an aggregate analysis by SAINT of three independent experiments for each experimental condition. (D) Venn diagram of the overlap between the Cyt, Sec, and NHS proteomes. (E) Protein sequences from the Cyt (486), Sec (254), and NHS (416) proteomes were analyzed using TMHMM and SignalP to detect predicted transmembrane helices and Sec, Tat, and lipobox sequences. Data labels indicate the percent of total proteins in each category.