Human monocytic cells (THP-1) were cultured at 37°C in SILAC RPMI-1640 medium (Thermo Pierce) supplemented with 10% dialyzed fetal calf serum (FCS). Media was supplemented with either light (Arg 0, Lys 0, Fisher Scientific) or heavy (Arg 10, Lys 8, Fisher Scientific) amino acids at 50 mg/L and L-proline at 200 mg/L. The incorporation of heavy label was checked by analysis of a lysate of 3 × 10⁶ heavy labeled cells generated using 4% Sodium dodecyl sulfate (SDS)/Dithiothreitol (DTT)/Tris buffer and Filter Aided Sample Processing (FASP) (Wisniewski et al., 2009). Incorporation was >97% for both arginine and lysine-containing peptides. Bacille-Calmette–Guerin Pasteur 1173P2 strain (BCG) was grown in 7H9 medium (Difco) supplemented with oleic acid, albumin, dextrose, and catalase (OADC), glycerol and 0.025% Tween-20. BCG was grown to early log phase (OD₆₀₀ ∼0.4) prior to infection of macrophages. BCG expressing green fluorescent protein (BCG-GFP) was constructed by transforming BCG with the episomal plasmid pMV261-GFP (Stover et al., 1991) and selecting for kanamycin resistance. All experiments with BCG and M. abscessus (local clinical isolate, kind gift from Dr. Hairong Huang, Beijing Tuberculosis and Thoracic Tumor Institute) were carried out in Biosafety level-2 containment facilities according to local guidelines and all experiments with M. tuberculosis (strain H37Rv) in a BSL-3 facility according to local guidelines.

THP-1 cells were activated and differentiated into adherent macrophages by overnight incubation with phorbol-12-myristate-13-acetate (PMA) at a concentration of 5 ng/ml prior to infection. Cells were washed and adherent cells infected with BCG. BCG cultures in logarithmic growth phase were centrifuged, resuspended, sonicated, filtered through a 5 μm filter and diluted in serum free SILAC RPMI to achieve a multiplicity of infection of 5:1. Cells were incubated with BCG for 4 h at 37°C and then washed three times with phosphate buffered saline (PBS) to remove extracellular bacteria still in suspension. Adherent control cells underwent the same PMA activation and media changes with PBS washes, but did not undergo infection. For experiment A, light-labeled cells were infected with BCG and heavy-labeled cells were the control. For experiment B, heavy-labeled cells were infected, and light-labeled cells were the control as a ‘label-swap’.

Plasma membrane profiling was performed as described previously (Weekes et al., 2010, 2012), with minor modifications. Briefly, after 48 h, uninfected and infected adherent cells were scraped, resuspended, and mixed 1:1, using 5.6 × 10⁷ of each cell type. Cells were washed twice with ice-cold PBS. Sialic acid residues were oxidized with sodium meta-periodate (Thermo) then biotinylated with aminooxy-biotin (Biotium). The reaction was quenched, and the biotinylated cells resuspended in 1% Triton X-100 lysis buffer. Biotinylated glycoproteins were enriched with high affinity streptavidin agarose beads (Pierce) and washed extensively. Captured protein was denatured with DTT, alkylated with iodoacetamide (IAA, Sigma) and digested overnight on-bead with trypsin (Promega) in 50 mM ammonium bicarbonate pH at 37°C. For experiment B, ten percent of the resultant digest was desalted and concentrated by StageTip (Rappsilber et al., 2007) for immediate analysis. For experiment A, 90% of the tryptic peptide sample was fractionated by HpRP-HPLC (see below). For both experiments, beads were further washed, and incubated overnight with Peptide-N-glycosidase (PNGase) at 37°C in G7 buffer. Glycopeptides were collected, beads were washed once with G7 buffer, and eluates pooled and concentrated on a StageTip (Rappsilber et al., 2007). Peptides were enriched and desalted using StageTips.

High pH Reverse-phase High Pressure Liquid Chromatography (HpRp-HPLC) was performed as described previously (Weekes et al., 2012), using 90% of the tryptic peptide sample from experiment A with a Dionex Ultimate 3000 powered by an ICS-3000 SP pump with an Agilent ZORBAX Extend-C18 column (4.6 mm × 250 mm, 5 μm particle size). Mobile phases (H20, 0.1% NH4OH or MeCN, 0.1% NH4OH) were adjusted to pH 10.5 with the addition of formic acid and peptides were resolved using a linear 40 min 0.1%-40% MeCN gradient at pH 10.5. Eluting peptides were collected in 15 s fractions. Fractions were dried down using an Eppendorf Concentrator and resuspended in 8 μL MS solvent (3% MeCN, 0.1% TFA). Fractions 24 to 136 inclusive were analyzed and in each case 3 μL was injected and subjected to LC-MS/MS using a NanoAcquity uPLC (Waters, MA, USA) coupled to an LTQ OrbiTrapXL (Thermo, FL, UA). Peptides were eluted using a gradient rising from 10 to 25% MeCN by 25 min, 40% MeCN by 28 min and 85% MeCN by 29 min. MS data was acquired between 300 and 1600 m/z at 60,000 FWHM with lockmass enabled (445.120025 m/z). CID spectra were acquired in the LTQ with MSMS switching operating in a top 6 DDA fashion triggered at 500 counts. For both experiments, the glycopeptide samples were eluted using a gradient rising from 3 to 25% MeCN by 90 min, 40% MeCN by 100 min and 85% MeCN by 102 min. Spectra were acquired using the same MS parameters.

Raw MS files were processed using MaxQuant version 1.3.0.5 (Cox et al., 2009, 2011). Raw files from the glycopeptide samples were searched in parallel with tryptic samples. An experimental design was specified such that all raw files from Experiments A and B were analyzed together, with data output separated within the same spreadsheet. Data were searched against concatenated Uniprot human and BCG databases, and common contaminants (Cox et al., 2011). Fragment ion tolerance was set to 0.5 Da with a maximum of 2 missed tryptic cleavage sites. Carbamidomethyl cysteine was defined as a fixed modification, oxidized methionine, N-terminal acetylation and deamidation (NQ) were selected as variable modifications. Reversed decoy databases were used and the false discovery rate for both peptides and proteins were set at 0.01. Protein quantitation utilized razor and unique peptides and required a minimum of 2 ratio counts and normalized protein ratios reported. Peptide re-quantify was enabled in all analyses apart from where indicated. Significance B values were calculated and Gene Ontology Cellular Compartment (GOCC) terms added using Perseus version 1.2.0.16¹.

To assess the number of PM proteins identified, we summed the proteins with GOCC terms PM, cell surface but not PM (CS) and extracellular but not PM/CS (XC) as previously described (Weekes et al., 2010, 2012). We previously identified a subset of proteins annotated by GO as integral to the membrane, but with no subcellular assignment and a short GOCC (ShG) term (a 4-part term containing the terms ‘integral to membrane’ ‘intrinsic to membrane’, ‘membrane part’, ‘cell part’ or a 5-part term containing ‘membrane’ in addition to these terms) (Weekes et al., 2010, 2012). Where a majority of proteins identified from a given sample are annotated PM/CS/XC, it is likely that a substantial proportion of the proteins with a short GOCC term are also integral PM proteins – including these proteins provides a useful upper estimate of the number of identified PM proteins. To generate a list of proteins quantified with high confidence, we extracted all proteins quantified in both Experiments A and B. All such proteins were similarly modulated in both independent biological repeats.

THP-1 monocytes were differentiated into macrophages as described above, and infected with BCG-GFP at an MOI of 5:1 for 4 h. Non-phagocytosed bacteria were washed away as described above. Cells were harvested at 48 h post-infection and labeled with antibodies prior to analysis by flow cytometry. The degree of up- or down-regulation of a protein was calculated by comparing the mean fluorescent intensity of infected (FL-1^HI macrophages that had phagocytosed green BCG) macrophages with uninfected macrophages.

Antibodies used for surface staining for flow cytometry were: anti-ABCA1 (clone AB.H10), anti-CD14-APC (clone 61D3) and anti-SLAMF7-PE (clone 162) (from: eBioscience, Inc, CA, USA); and goat anti-mouse-APC (Invitrogen, NY, USA).

Cells were washed away from plates and stained in PBS with 5% FBS for 30 min at 4C using the monoclonal antibodies listed above. Cells which stained by unlabeled antibody were then washed by PBS for 2 times and stained in PBS with 5% FBS for 30 min at 4C using Alexa Fluor488, Alexa Fluor405 or APC labeled secondary antibody (from Invitrogen). Data were acquired on an Accuri^TMC6 (BD) or a MACSQuant^® VYB (Miltenyi Biotec) and analyzed using FlowJo software.

THP-1 cells were activated and differentiated into adherent macrophages by 16 μM phorbol-12-myristate-13-acetate (PMA) 48 h prior to infection. Bacterial cultures in logarithmic growth phase were centrifuged, washed, resuspended with RPMI-1640 culture medium with 10% FBS, and filtered through a 5 μm filter to get single cell bacteria. Macrophages were infected with BCG (MOI of 5:1) for 7 h or M. abscessus (MOI of 5:1) for 4 h or Mtb (MOI of 10:1) for 4 h at 37°C with 5% CO2. Then, cells were washed twice with PBS and cultured in the presence of Amikacin (200 μg/ml) for 1 h to kill extracellular mycobacteria. Cells were washed three times using warm PBS to remove Amikacin and then fresh RPMI-1640 with 10% FBS and antibiotics (penicillin and streptomycin) were added to cells. Simvastatin – 10uM – (Sigma) was added to activated THP-1 cells 16 h prior to infection and washed away prior to infection and fresh simvastatin and/or mevalonate (100 μM, Sigma) added in fresh medium following infection.

Virulent Mtb (H37Rv) or M. abscessus was grown to logarithmic growth phase in Middlebrook 7H9 broth (BD) with BBL Middlebrook OADC Enrichment (BD) and 0.05% Tween 80 (Sigma). After 4 h infection, infected macrophages were lysed at 0 h, 20 h, or 44 h incubation. Mycobacteria enumerated by plating serial dilutions of cell lysates on Middlebrook 7H10 agar plates. Briefly, infected macrophage were lysed with 1% Triton X-100 for 5 min and 10-fold dilutions were performed using 0.05% Tween 80 in PBS before plating on Middlebrook 7H10 (BD) supplemented with 10% oleic acid-albumin-dextrose-catalase enrichment (Difco). Colonies were counted after 21 days (Mtb) or 3 days (M. abscessus) of growth at 37°C. Survival rate was calculated as the percentage survival of bacteria (as enumerated by CFU) at time = X compared with lysis and plating immediately following infection.

Expression vector construction: the RNA interference sequence was designed as described (Fukuchi et al., 2004). DNA coding for an RNAi for human ABCA1 was prepared with the following oligonucleotides: 5-GATCCCCTGAGTTTAGGTATGGCGGCTTCAAGAGA-3 and 5-TCGAGAAAAATGAGTTTAGGTATGGCGGCTCTCTTGAA-3. These oligonucleotides were annealed and ligated into the pEN_H1 vector. Using the Gateway^® LR recombination reaction, the RNAi sequence was shuttled into a large expression vector pDSL_hpUGIP.

To package lentivirus: HEK293T cells were seeded 24 h before transfection. PDSL_hpUGIP or PDSL_hpUGIP contained RNAi sequence, pMD2.G and psPAX2 were transfected into HEK293T cells using Lipofectamine^® LTX with Plus^TM (from Invitrogen) following manufacturer’s instruction. Reagents without plasmid were added into HEK293T as a negative control. Forty-eight hours after transfection, the supernatants were collected and filtrated with 0.45 μm filter.

To infect THP-1 cells: fresh RPIM-1640 with 10% FBS was mix with the filtrated supernatant at 2 to 1 ratio and add to THP-1 cells. Twenty-four hours after infection, cells were collected and resuspended with fresh RPIM-1640 with 10% FBS. The expression of ABCA1 was detected by flow-cytometry 96 h after infection.

Statistical significance between groups was determined by unpaired two-tailed Student’s t-test. ^∗p < 0.05, ^∗∗p < 0.01.

To identify cell surface proteins modulated by BCG infection, we used ‘plasma membrane profiling’ to compare infected with control cells. We fractionated peptides using high pH reversed phase liquid chromatography to increase the number of proteins quantified. We quantified 873 proteins, of which 559 had a Gene Ontology annotation ‘plasma membrane’ (PM), ‘cell surface’ (CS), ‘extracellular’ (XC), or ‘short GO’ (ShG, see methods) (70% of all quantified proteins with a GO annotation; Experiment A). To ensure all light-labeled skin contaminants were fully excluded, a ‘label swap’ unfractionated sample was also analyzed which identified 110 proteins, of which 93% (101 proteins) were annotated PM/CS/XC/ShG (Experiment B, Figure 1). We considered the overlap between Experiments A and B ‘high confidence’ data, which is displayed in Figures 2A,B; Supplementary Table S1. Full data is shown in Supplementary Table S2.

We selected a subset of PM proteins whose expression was significantly affected by BCG infection and for which commercially available antibodies were available for flow cytometry (Figure 3). We confirmed upregulation of the most significantly upregulated high confidence proteins CD14 and ATP-binding cassette sub-family A member 1 (ABCA1) (Figures 2 and 3), in addition to SLAM Family Member 7 (SLAMF7) (Figure 3; Supplementary Table S2).

In order to identify pathways that were modulated upon BCG infection, we performed Ingenuity Pathways Analysis² (Ingenuity^® Systems, Mountain View, CA, USA) on all proteins up- or down-regulated >2.5-fold. In order to increase the ability of this analysis to detect regulated pathways, we used the full set of up- or down-regulated proteins derived from Experiment A. Multiple functional terms suggest an effect of BCG infection on host immunity, including ‘lipid metabolism’, ‘Immune Cell Trafficking’, ‘Inflammatory Response’, ‘Antigen Presentation’ and ‘Cell-mediated Immune Response’. Pathways significantly modulated by BCG infection include CD40 signaling, and leukocyte extravasation signaling as well as LXR/RXR activation. The LXR/RXR pathway is linked to lipid metabolism in macrophages (A-González and Castrillo, 2011).

It has been suggested that although mycobacteria are able to metabolize a wide variety of carbon sources in vitro, the key carbon source in vivo is likely to be cholesterol (Pandey and Sassetti, 2008; Griffin et al., 2011, 2012; Ouellet et al., 2011). We noted with interest, therefore, that several proteins highlighted by Ingenuity analysis of ‘lipid metabolism’ pathways were implicated in cholesterol metabolism and inflammation and were differentially expressed on the PM following BCG-infection. Upregulated proteins included Zinc alpha 2-glycoprotein (Ceperuelo-Mallafre et al., 2012), Low Density Lipoprotein (LDL) Receptor, CD82 (Delaguillaumie et al., 2004), CXCR7 (Li et al., 2014), and ATP-binding cassette subfamily A Member 1 (ABCA1) (Ye et al., 2011). Downregulated proteins included Plasminogen activator inhibitor 1 RNA-binding protein (Kudo et al., 2004), Sortilin 1 and the Sortilin-related receptor (Strong et al., 2014), Endothelial cell-selective adhesion molecule (Inoue et al., 2010), and Ephrin B1 with its receptor (Lee and Daar, 2009).

The cholesterol transporter ABCA1 pumps cholesterol from inside the cell to the outside (Jin et al., 2015) and may therefore alter the cellular content of cholesterol available to mycobacteria. We wished to investigate the functional role for upregulation of ABCA1 on mycobacterial infection. We noted a robust upregulation of ABCA1 following stimulation of activated THP-1 cells by both live and heat-killed BCG (Figure 4A). This was suggestive of a host response to mycobacterial infection. To verify that ABCA1 upregulation was not a generalized response to particle phagocytosis we fed activated THP-1 macrophages with fluorescently labeled beads. Phagocytosis of the beads was associated with a small down-regulation of surface ABCA1 expression (Supplementary Figure S1), suggesting that ABCA1 upregulation was a specific response to stimulation by mycobacterial components.

Our proteomic experiments were performed on BCG, a relatively non-pathogenic relative of pathogenic Mtb. We wished to extend the implications of our data to pathogenic mycobacteria such as M. tuberculosis and M. abscessus. Previous reports had noted upregulation of ABCA1 gene expression in murine macrophages infected by M. tuberculosis-H37Rv (Stavrum et al., 2012), but not examined ABCA1 protein expression. We verified that infection of activated THP-1 cells, a human monocytic cell-line by pathogenic mycobacteria such as M. tuberculosis-H37Rv (Figure 4B) and M. abscessus (Figure 4C) also led to increased plasma-membrane expression of ABCA1.

To determine whether ABCA1 upregulation benefits pathogen or host, we knocked-down ABCA1 expression by RNA interference. Lentiviral transduction of activated THP-1 cells with short-hairpin RNA (shRNA) targeting ABCA1 led to decreased cell-surface expression of ABCA1 (Figure 4D). This knockdown led to increased survival of intra-cellular M. tuberculosis (Figure 4E) and M. abscessus (Figure 4F) and this was not due to differences in infection rate in macrophages where ABCA1 was knocked down (Supplementary Figure S2).

To determine whether the phenotype associated with ABCA1 depletion was due to its effects on cellular cholesterol depletion, we used the pharmacological agents simvastatin and mevalonate (Parihar et al., 2013). Mevalonate is a water-soluble precursor of cholesterol and its synthesis by HMG-CoA reductase is inhibited by simvastatin. Simvastatin depletes cellular cholesterol stores, which would be rescued by the addition of mevalonate, and simvastatin treatment has recently been shown to potentiate TB treatment (Dutta et al., 2016). Treating THP-1 cells infected with M. abscessus with simvastatin following knock-down of ABCA1 or empty vector led to a substantial decrease in mycobacterial survival (Figure 5), with no difference in mock-treated or ABCA1 knock-down cells. This suggested that, with the cellular depletion of cholesterol by simvastatin, ABCA1 inhibition had no additive effect. Whereas addition of mevalonate rescued the simvastatin phenotype and mevalonate alone increased intra-cellular mycobacterial survival, regardless of ABCA1 knock-down (Figure 5). Taken together, these data are highly suggestive that ABCA1 upregulation following mycobacterial infection is a host response to restrict mycobacterial growth by intra-cellular cholesterol depletion.