Life science research faces an increasing demand for absolute quantification of biomolecules to determine the molecular composition of a cell and to support mathematical modelling of biochemical processes. Here, we implemented a universally applicable, label-free strategy to estimate absolute cellular protein concentrations on a proteome-wide scale based on SWATH mass spectrometry. We applied this strategy to study proteomic reorganisation in the human pathogen Mycobacterium tuberculosis during exponential growth, hypoxia-induced dormancy and resuscitation. The resulting data set covering >2000 proteins reveals how protein biomass is distributed among cellular functions and dynamically remodelled in response to hypoxic stress. We found that the DosR regulon contributes 20% to the entire cellular protein content during dormancy, whereas the fraction of ribosomal proteins remains largely unchanged at 5-7%. Knowledge of protein copies/cell furthermore allowed us to translate effects of protein regulation into changes in maximal enzymatic reaction velocities, enhancing our understanding of metabolic adaptations.

For absolute label-free abundance estimations of all proteins identified by SWATH MS, 30 anchor proteins were selected covering a wide abundance range of the Mtb proteome (Schubert et al., 2013). For each anchor protein one or two synthetic isotope-labelled reference peptides in defined concentrations as determined by amino acid analysis were spiked into the samples for accurate absolute quantification of these anchor proteins (see sample preparation paragraph). To ensure accurate absolute quantitation results, the linear dynamic quantification range and the lower limit of quantification for each reference peptide was determined by performing a dilution series experiment. In the end, 51 peptides showed good linear response (slope close to 1, axis intercept close to 0, R2 > 0.94) and suitable lower limit of quantification with respect to endogenous protein levels. All results of this experiment are summarised in Table S2 and can be visualised in Panorama: http://tinyurl.com/ka9elv5. Absolute concentrations of all anchor proteins over all samples were analysed manually using Skyline (MacLean et al., 2010) and corresponding raw chromatographic data are available in Table S2 and in Panorama. Integrated peak areas of the reference and endogenous peptides were summed and from the obtained ratios the endogenous peptide concentration was determined in fmol/μg. Of note, in the here described absolute quantification procedure, which is based on monoisotopic peak areas measured with high-resolution mass spectrometry, we did not account for differences in the theoretical isotopic distributions between the light and heavy peptide forms. The optimal model to combine SWATH MS intensities to a single protein MS signal was determined by Monte Carlo cross-validation (Ludwig et al., 2012) using the aLFQ R package (version 1.3.1) (Rosenberger et al., 2014). Summarising the five most intense transitions of the three most intense peptides per protein was the resulting model with the highest accuracy and was used to estimate proteome-wide concentrations from the SWATH MS signal intensities. For proteins detected with only two peptides absolute abundance estimation was based on those two peptides, while proteins with only a single quantified peptide were excluded from further analysis.

Wayne experiments in Mtb H37Rv and M. bovis BCG were performed as described earlier (Wayne and Hayes, 1996). Samples were taken before the start of the experiment in aerobic cultures (day 0) and at 5, 10 and 20 days of the hypoxic time course (“day 5”, “day 10” and “day 20”). After sampling at day 20, remaining glass vials were opened and cultures were transferred into new vessels for aerated incubation without addition of new culture medium. Further time points were sampled from cultures re-aerated for 6 or 48 hours (“day 20+6h” and “day 20+48h”). Bacterial pellets were resuspended in lysis buffer containing 8 M urea and 0.1% RapiGest subjected to sonication and bead beating to extract proteins. Protein concentration was determined using a BCA assay. Proteins were reduced and alkylated, followed by proteolysis with Lys-C and trypsin (Glatter et al., 2012). At this step isotopically labelled synthetic reference peptides (AQUA QuantPro, Thermo Fisher Scientific) for absolute quantification of the 30 anchor proteins were added to the samples in concentrations similar to the endogenous peptide. The peptide solution was desalted by C18 reverse-phase chromatography. iRT peptides (RT-kit WR, Biognosys) were added for quality control and to allow determination of system-independent retention times (Escher et al., 2012).