Chemical Proteomics Strategies for Analyzing Protein Lipidation Reveal the Bacterial O-Mycoloylome

Abstract

Protein lipidation dynamically controls protein localization and function within cellular membranes. A unique form of protein O-fatty acylation in Corynebacterium, termed protein O-mycoloylation, involves the attachment of mycolic acids─unusually large and hydrophobic fatty acids─to serine residues of proteins in these organisms' outer mycomembrane. However, as with other forms of protein lipidation, the scope and functional consequences of protein O-mycoloylation are challenging to investigate due to the inherent difficulties of enriching and analyzing lipidated peptides. To facilitate the analysis of protein lipidation and enable the comprehensive profiling and site mapping of protein O-mycoloylation, we developed a chemical proteomics strategy integrating metabolic labeling, click chemistry, cleavable linkers, and a novel liquid chromatography-tandem mass spectrometry (LC-MS/MS) method employing LC separation and complementary fragmentation methods tailored to the analysis of lipophilic, MS-labile O-acylated peptides. Using these tools in the model organism Corynebacterium glutamicum, we identified approximately 30 candidate O-mycoloylated proteins, including porins, mycoloyltransferases, secreted hydrolases, and other proteins with cell envelope-related functions─consistent with a role for O-mycoloylation in targeting proteins to the mycomembrane. Site mapping revealed that many of the proteins contained multiple spatially proximal modification sites, which occurred predominantly at serine residues surrounded by conformationally flexible peptide motifs. Overall, this study (i) discloses the putative protein O-mycoloylome for the first time, (ii) yields new insights into the undercharacterized proteome of the mycomembrane, which is a hallmark of important pathogens (e.g., Corynebacterium diphtheriae, Mycobacterium tuberculosis), and (iii) provides generally applicable chemical strategies for the proteomic analysis of protein lipidation.

Figure 1

Mycomembrane biosynthesis pathway and exploitation to label and analyze O-mycoloylated proteins. (A) Model for the biosynthesis of the major components of the Cg mycomembrane. Cytoplasmic TMM is translocated to the periplasm and used as the universal mycoloyl donor. Subsequently, the TMM donor is processed by mycoloyltransferase enzymes to construct the major components of the mycomembrane, including the glycolipids AGM and TDM (left path) and O-mycoloylated proteins, the latter of which are synthesized by the protein-selective mycoloyltransferase Cmt1 (right path). Note: the precise sequence and location of Cmt1-mediated protein O-mycoloylation events are not known. (B) Exploitation of mycomembrane biosynthesis to label and analyze O-mycoloylated proteins. The synthetic probe O-AlkTMM mimics the mycoloyl donor function of TMM and thus undergoes mycoloyltransferase-mediated alkyne-labeling of the mycomembrane components. In this work, Cmt1-mediated alkyne-labeling of O-mycoloylated proteins enabled click chemistry-mediated visualization, identification, and site mapping of this PTM on the whole-proteome scale. AG, arabinogalactan; AGM, arabinogalactan mycolate; MM, mycomembrane; PG, peptidoglycan; PM, plasma membrane; TDM, trehalose dimycolate; and TMM, trehalose monomycolate.

Figure 2

Specific labeling and identification of putative O-mycoloylated proteins. (A) Chemical proteomics strategies developed in this study to enable the visualization, identification, and site mapping of O-mycoloylated proteins. (B) Cg wild type (WT), cmt1 mutant (Δcmt1), or complement (Δcmt1::cmt1) were treated with O-AlkTMM or left untreated, then cell lysates were collected and subjected to CuAAC with azido-488 and analyzed by SDS-PAGE with visualization by Coomassie Brilliant Blue (CBB) staining and in-gel fluorescence scanning (Fluor). (C) Cg WT was treated with varying concentrations of O-AlkTMM, 6-heptynoic acid, or left untreated and then processed and analyzed as in (B). (D) As depicted in (Ai), lysates from O-AlkTMM-treated Cg WT were subjected to CuAAC with Az-T-B, and then the fluorescently labeled, biotinylated proteins (“input”) were enriched on streptavidin-coated beads and eluted (“output”). Input and output samples were analyzed by SDS-PAGE with visualization by CBB staining or silver staining, respectively, and in-gel fluorescence scanning. (E) Four replicate output samples prepared as in (D) were digested, and the resulting peptides were analyzed by label-free quantitative LC-MS/MS. The volcano plot depicts proteins in red that were identified as significantly enriched (p < 0.05) by a fold-change (FC) of at least 3 in the O-AlkTMM-treated samples versus the untreated controls. Proteins of interest are annotated and discussed in the main text. Raw and curated data from this study (LC-MS/MS study 1 (Az-T-B)) are found in Supporting Tables S4–S6. (F) Cg strains expressing either His-tagged wild-type PorB, His-tagged double mutant PorB-S₇A/S₉₈A, or empty vector control were treated with O-AlkTMM or left untreated; then, cell lysates were collected and subjected to CuAAC with azido-488. His-tagged proteins were enriched on Ni²⁺-NTA resin, eluted, and analyzed by SDS-PAGE with visualization by CBB staining and in-gel fluorescence scanning.

Figure 3

Identification of modified peptides using cleavable linker and LC-MS/MS strategies. (A) As depicted in Figure 2Aii, lysates from O-AlkTMM-treated Cg WT were subjected to CuAAC with Az-DADPS-B, and then the biotinylated proteins were captured on streptavidin beads. Beads were treated with 5% HCO₂H to cleave the DADPS linker and release intact proteins, which were analyzed by SDS-PAGE with silver staining. (B–F) Data from LC-MS/MS study 2 (DADPS) and protein-level enrichment of O-AlkTMM-modified proteins. Proteins were enriched with O-AlkTMM and Az-DADPS-B as described in (A) and digested to generate modified peptides, which were then analyzed by a custom LC-MS/MS method. (B) Cumulative percent of the number of tryptic modified (gray) and unmodified (green) peptides detected at various retention times over 120 min. (C) Venn diagram of unique modified peptides detected by AI-ETD and HCD analyses. (D) Charge state distributions of unique modified (gray) and unmodified (green) peptides. (E, F) Examples of annotated (E) HCD and (F) AI-ETD spectra showing thorough sequence coverage and the presence of characteristic ions at m/z of 252.17 and 270.18, corresponding to the protonated modification with and without an adducted water molecule. Lists of ions for these spectra are given in Supporting Tables S12 and S13. Additional annotated spectra are shown in Supporting Figure S4.

Figure 4

Characteristics of modified proteins determined by LC-MS/MS identification and site mapping studies. (A) Venn diagram comparison of proteins identified in LC-MS/MS studies 1–3. (B) Venn diagram comparison of the number of modification sites identified in LC-MS/MS studies 2 and 3. (C) Distribution of the number of modification sites observed per peptide and per protein in LC-MS/MS study 3. (D) Results of motif enrichment analysis conducted using WebLogo using data from LC-MS/MS study 2 (top), study 3 (middle), and the combined data from studies 2 and 3 (bottom). All data for LC-MS/MS studies 2 and 3 in panels (A–D) are from the curated lists (Tables 1, S8 and S10).

Figure 5

Modification sites are spatially clustered on multiply modified proteins. (A, left) Crystal structure of Cmt1 (PDB ID: 4H18). All putative O-mycoloylation sites on Cmt1 were identified within the highlighted disordered loop consisting of residues 281–304. Catalytic site, blue. (A, right) AlphaFold structure of Cmt1. Disordered loop in crystal structure, gray; modification sites, red; catalytic site, blue. (B, left) Crystal structure of PorB (PDB ID: 2VQL). O-Mycoloylation sites on PorB were identified within the disordered regions at the termini. (B, right) AlphaFold structure of PorB. Modification sites, red. Note: the signal sequences for both AlphaFold-predicted structures were removed. See Supporting Figure S5 for the AlphaFold-predicted structures of other candidate O-mycoloylated proteins.