3.4.2. LC-MS/MS Analysis

Peptides were separated with high-performance liquid chromatography (HPLC, Ultimate 3000 Nano LC System, Thermo Fisher Scientific,) in a 15-cm long C18 column with a diameter of 75 μm (Acclaim ®PepMap ™ RSLC, Thermo Fisher Scientific). The peptides were eluted with a gradient from 5% to 35% of bu ffer B (80% acetonitrile, 0.1% formic acid) over 115 min at a flow rate of 0.3 μl/min. Total run time including 15 min to reach 99% bu ffer B, flushing 10 min with 99% bu ffer B and 15 min re-equilibration to bu ffer A (0.1% formic acid) amounted to 65 min. Further analysis was performed with a Q Exactive HF mass spectrometer (Q ExactiveTM HF Hybrid Quadrupole-OrbitrapTM Mass spectrometer, Thermo Fisher Scientific). Mass spectra were acquired at a resolution of 60,000 (MS) and 15,000 (MS/MS) in a range of 400-1500 m/z (MS) and 200-2000 m/z (MS/MS). An isolation threshold of 67,000 was determined for precursor selection, and up to the top 25 precursors were chosen for fragmentation with high-energy collisional dissociation (HCD) at 25 V and 100 ms activation time. Precursors with a charged state of +1 were rejected and all measured precursors were excluded from measurement for 20 s. From 2 to 4 technical runs were analyzed for each sample.

### 3.4.3. Protein Identification and Quantitation

Raw data was captured from the mass spectrometer and converted to MGF (Mascot Generic Format) files using ProteoWizard with the following parameters: peakPicking true 2, msLevel 2, zeroSamples removeExtra [52]. For thorough protein identification, the generated peak lists were searched with the MASCOT (v 2.5.1, Matrix Science Ltd, UK) and X! Tandem (VENGEANCE, 2015.12.15, The Global Proteome Machine Organization) search engines. Database-searching parameters were as follows: tryptic hydrolysis, no more than one missed site, the precursor and fragment mass tolerance were set at 20 ppm and 50 ppm, respectively. Oxidation of methionine was set as a possible modification, carbamidomethylation of cysteine as a fixed. For X! Tandem, parameters that allowed a quick check for protein N-terminal residue acetylation, peptide N-terminal glutamine ammonia loss or peptide N-terminal glutamic acid water loss were selected. Resulting files were submitted to the Sca ffold 4 software (v 4.2.1, Proteome Software, Inc, USA) for validation and further analysis. For protein identification, the proteomic databases for the RUS\_B0 (RefSeq: CP030093.1) and reference strain H37Rv (RefSeq: NC\_000962.3) genomes were used. Additionally, in proteome searching database peptides containing strain-specific single amino acid polymorphisms were added according to the approach described earlier [51]. The local false discovery rate scoring algorithm with standard experiment-wide protein grouping was used. A 1% FDR threshold was applied to search results from individual datasets. For all detected proteins, functional categories (TubercuList v 2.6 (http://tuberculist.epfl.ch/)) and subcellular localizations (PSORTdb v 3.0 (http://db.psort.org/)) were established.

For label-free quantitation, raw MS data files (.wi ff files) were imported and processed in Progenesis LC-MS software v.4.1 (Nonlinear Dynamics, Newcastle, UK). The results of peptide quantitation were normalized using an iterative median-based normalization as implemented in the Progenesis software. Di fferences in the abundance of a protein between the three biological replicates of all strains were evaluated using a two-sided unpaired Student's T-test. *P*-values < 0.05 were considered

statistically significant. Adjusted *p*-values for multiple tests (*q*-values) were generated using the Benjamini–Hochberg method [53].
