Mass spectrometry (MS) for the detection of proteins is an indispensable tool for evaluating the biological processes of the proteome. Proteomics frequently requires proteolysis of proteins into peptide fragments. Proteins can be refractory to ideal proteolysis at the sequence level rendering them difficult to analyze by routine proteomics methods. EsxA (ESAT-6, Early Secreted Antigen, 6kDa) is a major virulence determinant of Mycobacterium tuberculosis, the cause of human tuberculosis. EsxA is routinely used to evaluate mycobacterial virulence in the laboratory and as a biomarker for tuberculosis in humans. The sequence of EsxA hinders deeper MS analysis beyond routine detection. Here we engineer the sequence of EsxA to add desirable tryptic properties aimed at improving complex MS analysis. We demonstrate that EsxA variants are amenable to MS analysis and remain functional in established in vitro and ex vivo assays of Esx-1-function. We provide the first demonstration of molecular engineering to specifically improve MS analysis of individual microbial proteins.
Proteomic detection using MS analysis is a standard approach to identify and quantify proteins. Protein mixtures are digested into peptide fragments using proteases. The most frequently used protease in MS analysis is trypsin, which cleaves following lysine (K) or arginine (R) with high fidelity. The size of the cleaved products, generally between eleven and 30 amino acids in length (Fig. 1), increases the likelihood of identification using MS and MS/MS methods. Due to the abundance of K and R in proteomes, most proteins when digested with trypsin produce peptides within this range that are sufficient for identification and for further quantitative analysis by MS (Fig. 1). However, MS analysis of some proteins is complicated by the amino acid composition; one is limited to the protease sites nature designed. In general, peptides ≤ 5 amino acids or greater than 30 amino acids perform poorly by chromatography or tandem MS and are not readily identified by common proteomics methods1,2. This subset of peptides, which comprises 28.1% of peptides generated by trypsin in silico for M. tuberculosis, are not commonly identified in experimental data and constitute a “dark/inaccessible proteotypic proteome” (Fig. 1). The dark-regions of the proteotypic proteome do not generally impact protein identification. Inconsistent recovery and poor reproducibility limit the quantitative analysis of specific peptides such as protein N- or C-termini, post-translational modifications, and domains with high or low distributions of K and R residues (Supplemental Fig. 1). Digestion with alternate proteases or using top/middle-down proteomics methods can be used to span dark regions1,3,4. While both of these techniques are promising, they can be less sensitive and more challenging than trypsin-based proteomics. The dark-regions of the proteome are also poor choices for MRM/SRM-based detection & quantification5. Some of the criteria for quantitative exclusion include the lack of a complete trypsin fragment, >20aa in length, and low digestion efficiency. Like trypsin, alternative proteases have inaccessible regions, necessitating larger sample size and multiple analyses to span a proteome. Finally, small proteins in particular are considerably constrained in the distribution of peptides due to a rarity of appropriate cleavage sites.
Although large MS peptides allow protein identification, stoichiometric quantification remains challenging. An underappreciated aspect of large peptides, which complicates MS analysis, is the relatively large number of sites capable of being modified in a non-physiological manner. Moreover, the number of sites that can result in a missed enzymatic cleavage is also large. Missed cleavages and non-physiological modification of peptides may only have a modest impact on peptide detection. However, they have a significant impact on quantification. Specifically, the large peptide is likely to exist in many proteoforms. No clean or consistent peptide form is observed across biological samples. Consequently, the signal for peptide quantification is diluted, creating severe challenges to peak integration. A typical MRM is a canonical cleavage peptide (high digestion efficiency and reproducibility), >7 and <20 amino acids in length with no M and typically avoiding C when feasible5.
Tuberculosis (TB) is an epidemic with a disease burden of ~9.6 million infections and 1.5 million deaths annually6. The Esx-1-exporter is a protein transport system which promotes the extracellular localization of mycobacterial virulence factors, including EsxA. EsxA is a major human T-cell antigen and a substrate of Esx-17,8,9. Loss of Esx-1 is the chief cause of attenuation in the TB vaccine strain, Mycobacterium bovis BCG10,11,12. EsxA is used in blood-based diagnostics for TB including QuantiFERON TB Gold (Cellestis, Carnegie, Australia) and T-SPOT.TB (Oxford Immunotec, Oxford, UK) tests13. EsxA is also a central candidate included in the rational design of novel TB vaccines14,15. In the laboratory, EsxA secretion and activity are regularly used as reporters of Esx-1-function and mycobacterial virulence.
EsxA and its binding partner EsxB are produced in the cytoplasm of mycobacterial cells and transported across the cytoplasmic membrane by the Esx-1-exporter8,9,10. During in vitro growth, the EsxB and EsxA proteins are displayed on the mycobacterial surface and released or secreted into the culture supernatant via unknown mechanisms14,15. In ex vivo cellular infection models the Esx-1-system perforates the phagosomal membrane promoting interaction between the mycobacterial cell and the cytosol of the macrophage and modulation of the host innate immune response. Mycobacterial strains deficient for Esx-1-export are retained in the phagosome and are attenuated in ex vivo and in vivo infection models16,17,18,19,20.
Generally Esx-1-mediated protein secretion into culture medium during bacteriological growth correlates with virulence in ex vivo and in vivo infection models10. EsxA and other Esx-associated proteins are N-terminally acetylated21,22,23,24,25,26,27. How modification of the protein N-termini of Esx-associated proteins impacts mycobacterial virulence remains unknown. We have developed several MS approaches to measure the levels of EsxA on the mycobacterial cell surface and the modification of its N-terminus by acetylation14,21,22,23. However, routine use of these assays is complicated by the sequence of EsxA, which results in a large N-terminal tryptic fragment of 33 amino acids (Fig. 2).
The N-terminal tryptic fragment of EsxA has most of the criteria of a weak proteotypic peptide, including poor chromatographic reproducibility, further reducing accurate quantification and sensitive detection of EsxA between strains5. The substantially poor qualities of the EsxA N-terminus are well documented in peptide-libraries that catalog protein and peptide identification. The Peptide Atlas contains 13 proteoforms of the EsxA N-terminal peptide alone (Supplemental Table 1)28,29. A single modification of the N-terminus and monitoring for the loss or acetylation of the initiator Met residue means no fewer than 52 N-terminal peptides might need to be simultaneously monitored just to measure changes in the canonical N-terminus. These properties are reflected in the observed pool of N-terminal peptide fragments and the PeptideAtlas entry for EsxA which has suitability and observability scores ~10-fold less than internal peptides. This means that under typical conditions, your likelihood of identifying EsxA is high however your reproducibility, recovery and quantification of the N-terminus itself will be poor (Supplemental Table 1). Other proteases such as GluC generate size appropriate N-terminal peptide fragments of EsxA (Data not shown), but we find these proteases more costly and less reliable (complete and reproducible digestion) than trypsin.
Genetic manipulation of organisms to facilitate scientific inquiry is commonplace. Virtually all laboratory microorganisms are modified to circumvent natural barriers to study. Rational design to improve proteome analysis at the genetic level is a logical next step. However, genetic manipulation of the proteome has largely not been described. We sought to modify the sequence of EsxA at the chromosomal level to improve tryptic digestion and detection while retaining protein function.
EsxA variants with improved MS performance are secreted from Mycobacterium
Prior studies have indicated that residues surrounding E12 are not required for EsxA secretion or functions that mediate virulence30. Moreover, changing residue E12 to alanine did not impact virulence of M. tuberculosis in a mouse model of infection31. We engineered the sequence of EsxA from M. tuberculosis to include an additional tryptic site following residue 12. We generated three variants which changed the E residue at position 12 to a K (E12K), an R (E12R) or a C (E12C) (Fig. 2). EsxA and EsxB are encoded by the esxBA operon and require each other for stability8,32. We introduced each mutation into an integrating plasmid containing the esxBA operon from M. tuberculosis (indicated with an “MT” subscript) and confirmed the mutagenesis by DNA sequencing analysis.
The Esx-1 exporter, as well as EsxA and EsxB, is highly conserved between M. tuberculosis and Mycobacterium marinum, a pathogenic mycobacterial species that is an established model for several aspects of M. tuberculosis pathogenicity and virulence33,34. One of the major benefits of using M. marinum as a model for M. tuberculosis is that this organism can be studied in a Biosafety Level 2 laboratory.
Deletion of the esxBA genes in M. marinum can be functionally complemented with the esxBA genes from M. tuberculosis. Therefore, we can study transport and function of the M. tuberculosis proteins in M. marinum. We introduced the plasmids expressing either wild-type or variant esxBA operons into M. marinum and M. tuberculosis. Because EsxA and EsxB are also required for Esx-1 function8,9, we expected that if the changes in the EsxAMT sequence negatively impacted function, the EsxAMT and EsxBMT substrates would not be secreted into the culture medium. We prepared whole cell lysate and secreted protein fractions from both M. tuberculosis Erdman and M. marinum M strains expressing each EsxAMT variant. We measured EsxBMT production and secretion into the culture supernatant by both mycobacterial species using immunoblot analysis (Fig. 3). We detected EsxBMT in the whole cell lysate and culture supernatant fractions from both wild-type M. marinum and wild-type M. tuberculosis strains but not in the fractions from the M. marinum ΔesxBA or M. tuberculosis ΔesxA strains (Fig. 3a). Expression of the wild-type or the three variant esxBAMT operons restored EsxBMT production and secretion from M. marinum and M. tuberculosis.
Because the changes to the EsxAMT sequence disrupted the epitope recognized by the EsxA antibody used for immunoblot analysis, we measured the production and secretion of the EsxAMT variants using proteomic analysis. We digested the protein fractions with trypsin and subjected the resulting peptides to LC/MS/MS analysis. Two novel tryptic peptides result from the introduction of a K or an R at the 12th position at the N-terminus of EsxA (Fig. 2). To circumvent potential problems with tolerance of K and R, we also generated an E12C variant, which when chemically aminoethylated at the Cys residue, generates a tryptic substrate and the novel peptide M.TEQQWNFAGI[AEC-MAEC].
We readily detected the acetylated and unacetylated forms of the novel tryptic peptides (M.TEQQWNFAGIK. and M.TEQQWNFAGIR., respectively) corresponding to the first 12 amino acids of EsxAMT in the whole cell lysate and culture supernatant fractions from M. marinum and M. tuberculosis strains expressing EsxAMT E12K and EsxAMT E12R (Fig. 3b,c and Supplemental Fig. 2). Following aminoethylation, we detected both the acetylated and unacetylated forms of the novel tryptic fragment corresponding to the EsxAMT E12C protein in the culture supernatant from M. tuberculosis (Supplemental Fig. 3). From these data we conclude that the engineered EsxAMT is facile for tryptic digestion and detection by LC/MS/MS analysis. Together, the immunoblot and the MS analyses demonstrate that the EsxAMT variants are expressed and actively exported from M. marinum and M. tuberculosis. Moreover, the three EsxA variants can functionally promote secretion of EsxBMT from both M. marinum and M. tuberculosis.
EsxA variants retain function in vitro and ex vivo
We sought to further characterize the functionality of the EsxAMT variants in in vitro and ex vivo Esx-1 assays. M. marinum lyses red blood cells in a contact-dependent, Esx-1-dependent manner35,36. As such, deleterious changes in EsxAMT structure or function due to the E12C, E12K or E12R substitutions would result in a loss of hemolysis. As shown in Fig. 4a, wild-type M. marinum lysed sRBCs as indicated by an increased OD405 as compared to the negative PBS control. The ΔesxBA strain was non-hemolytic. Hemolytic activity was restored in M. marinum strains expressing either the wild-type or variant EsxAMT proteins. These data demonstrate that the E12C, E12K and E12R variants of EsxAMT can functionally promote hemolysis.
Infection of macrophages with virulent M. marinum at high multiplicity of infection (MOI) results in cytolysis of the host cell. Cytolysis requires perforation of the phagosomal membrane by Esx-119,20. If the EsxA variants were nonfunctional, we expected decreased cytolysis of macrophages infected with strains expressing the EsxAMT variants. We infected RAW 264.7 cells at an MOI of 5 and detected cytolysis by staining the macrophage monolayers 24 hours post infection using Calcein-AM and Ethidium Homodimer (Eth-D1). Calcein-AM is a membrane permeable dye that emits a green fluorescence when cleaved by cytosolic esterases in live cells. Eth-D1 is a nucleic acid stain that cannot cross intact cellular membranes. As shown in Fig. 4b,c, the infection of RAW 264.7 cells with wild-type M. marinum resulted in decreased viability (decreased Calcein-AM signal) and increased cytolysis (increased EthD-1 signal) as compared to the uninfected monolayer (mock). Infection with the ΔesxBA strain, which cannot access the cytosol, resulted in increased viability (robust Calcein-AM signal) and minimal cytolysis (decreased EthD-1 signal) as compared to infection with the wild-type strain. Expression of the wild-type or E12C, E12K or E12R variants of esxBAMT in the ΔesxBA strain of M. marinum resulted in decreased viability and increased cytolysis as compared to the parent ΔesxBA strain. These data indicate that the EsxAMT variants can promote Esx-1-mediated cytosolic access of M. marinum in an ex vivo macrophage infection model.
Phagosomal membrane perforation by the Esx-1-exporter triggers the cytosolic surveillance pathway (CSP) and the induction of IFN-β transcription in macrophages17,19,20,30. Therefore, induction of IFN-β transcript in macrophages infected by M. marinum or M. tuberculosis indicate that the Esx-1-exporter is functional. If EsxAMT variants had altered activity, we expected a lack of the IFN-β transcript induction by strains expressing the EsxAMT variants. We infected RAW 264.7 cells with M. marinum or M. tuberculosis strains expressing the wild-type or variant EsxAMT proteins and measured the induction of IFN-β using qRT-PCR. Infection with either M. marinum or M. tuberculosis significantly induced transcription of IFN-β as compared to both the mock infected or ΔesxBA control strains (Fig. 4d, p = 0.0286). Expression of esxBAMT in the ΔesxBA (M. marinum) or ΔesxA (M. tuberculosis) strains restored IFN-β transcription to levels at or above those induced by the wild-type strains. Expression of EsxAMT E12C, E12K or E12R in the esxBA or esxA deficient strains also restored induction of IFNβ-transcription to levels at or above those induced by the wild-type strains. Together, these data demonstrate that the EsxAMT variants promoted Esx-1-exporter activity in an ex vivo cell-based infection model.
Here, we provide the first example of the rational engineering of a gene of a microorganism to permanently improve proteomic detection while retaining measurable protein function. Our results demonstrate that the engineered changes to EsxAMT that improved tryptic digestion and proteomic detection of the natural protein N-terminus resulted in functional EsxA protein in routine assays. Another engineering approach was used in E. coli, where trypsin sites were engineered into bacterial inner membrane proteins to facilitate the study of protein topology. However, these sites were not used to facilitate proteomic analysis37,38.
Other studies have suggested alternative protease combinations to address the lack of trypsin coverage2,4. While the use of protease combinations certainly has merit, our approach presented here is complementary. A strength of our approach is that altering the gene of interest prior to proteomic analysis will streamline the proteomic throughput, such that parallel digests are not required to span the protein of interest.
We envision several applications of our framework described here within Mycobacterium. The N-terminal acetylation of EsxA has been proposed to both impact the physical interaction between EsxA and EsxB and virulence24,25,39. The strains we have generated here could serve as parent strains for researchers interested in studying the biological significance of the N-terminus of EsxA and how its modification promotes mycobacterial virulence using proteomics approaches. Moreover, EsxA, and its binding partner, EsxB, have 20 paralogous proteins in M. tuberculosis35,36. In addition to the Esx-1-system, there are four additional type VII secretion systems, named Esx-2-540. EsxN is a substrate of Esx5, and is required for virulence21. Esx-3 is required for metal homeostasis, and secretes EsxG and EsxH41,42. In silico analysis of the Esx-paralogs revealed that several of these proteins have regions that are unamenable to MS methods. In particular, many peptides generated from this population are outside of the ideal ranges for reproducible detection and are notably limited in the protein N-termini (Supplemental Fig. 4). The N-terminal peptides from EsxG, EsxC and EsxD have low suitability scores of 0.15, 0.1 and 0.1 respectively for their N-terminal peptide. The N-terminal peptides from EsxJ, EsxK, EsxM, and EsxP are too short to be observed (4 aa). The suitability scores of the N-terminal peptides from EsxO, EsxE and EsxF range between 0 and 0.0128,29. EsxH has a predicted N-terminal tryptic fragment of 67 amino acids which would be refractory to analysis and would be ideal for the approach outlined here. When just considering the Esx-like proteins, 40% of the N-termini alone are inaccessible. This places limitations on researchers’ ability to characterize changes in the processing or modification of the protein N-termini, which may be essential for disease22,33. In the M. tuberculosis proteome overall, we observed that 7.3% and 13.8% of peptides generated were greater than 30 amino acids for the empirical and in silico analyses respectively indicating that refractory proteins exceed the Esx-systems (Supplemental Fig. 1). As we demonstrate here, C and K substitutions introduce chemical side-chains which are readily modified to alter protease sensitivity, and may allow the quantification of otherwise intractable modified peptides.
In cells, the natural distribution of K and R residues throughout the proteome limits complete coverage of the proteome or a specific set of proteins (Supplemental Fig. 1). Previous studies in Saccharomyces cerevisiae revealed that 56% of the theoretical peptides generated by trypsin are ≤6 amino acids, which are severely constrained for identification using proteomics1. In M. tuberculosis, at least 19.6% of the theoretical peptides generated by trypsin are ≤6 amino acids in length (Fig. 1 and Supplemental Fig. 1). Removal of existing K or R residues might be desired in cases where the ’natural’ polypeptide is too short for definitive MS assignment.
We envision that our approach would function with minor modifications for any target protein in any organism. Because this approach relies on altering single amino acids to introduce novel tryptic cleavage patterns, it is likely that the virtually all resulting protein variants can be designed to retain function. One consideration in applying this approach is that it has been reported that changing a single amino acid of a protein sequence could lead to the gain of new, undiscovered function (Reviewed recently by C.J. Jeffery43). It is formally possible that changing the 12th residue of EsxA to R, K or C could lead to new functional activity that we have not yet appreciated. However, here we demonstrate the canonical function of EsxA has not been lost.
Although we generated esxA variants using a PCR-based approach, a long-term implementation in higher organisms could be to employ other genome editing approaches including CRISPR to generate stable ‘proteome-ready’ lines-for research. Finally, we can envision an application by which several strains or cell lines, each bearing a unique C, K, or R substituted peptide in the same protein could serve as a proteomic ‘bar-coding’ enabling quantitative dissection of individual strains from mixed bacterial or complex cellular populations.
Overall, we provide a specific example of genetic manipulation of an M. tuberculosis virulence factor to improve proteomic analysis while retaining protein function. This approach is applicable to numerous targets across any systems to facilitate analysis of how alterations to proteins impact biological function.
Growth of mycobacterial strains
All mycobacterial strains used in this study are listed in Supplemental Table 2. All Mycobacterium marinum strains were derived from the M strain (ATCC BAA-535). M. marinum strains were grown at 30 °C and maintained in Middlebrook 7H9 defined liquid broth (Sigma-Aldrich, St. Louis, MO) with 0.5% glycerol and 0.1% Tween-80 (Fisher Scientific, Fair Lawn, NJ) or on Middlebrook 7H11 agar (Thermo Fisher, San Jose, CA) as previously described26. Kanamycin (20 μg/mL; IBI, Poesta, IA) or hygromycin (50 μg/mL; EMD Millipore, Billerica, MA) were added when appropriate. All M. marinum strains were grown and maintained in a BSL-2 laboratory under safety protocol (13–028), approved by the University of Notre Dame.
All M. tuberculosis strains were derived from M. tuberculosis Erdman (ATCC 35801), which was obtained from Jeffery S. Cox. M. tuberculosis strains were grown in 7H9 defined liquid broth with 10% Middlebrook OADC enrichment (Becton Dickinson and Company, Sparks, MD) and 0.05% Tween-80 at 37 °C. Kanamycin (20 μg/mL) or hygromycin (50 μg/mL) was added to the 7H9 broth when appropriate. All M. tuberculosis strains were grown and maintained in a BSL-3 laboratory under safety protocol (13-029) approved by the University of Notre Dame.
The pMH406 plasmid (KanR gift of David Sherman, HygR gift of Jeffery Cox) is an integrating plasmid bearing the esxBA genes from M. tuberculosis behind the mycobacterial optimal promoter (MOP)7,42. The pMH406 HygR plasmid was the parent for the constructs expressed in M. marinum. The pMH406 KanR plasmid was the parent for the constructs expressed in M. tuberculosis.
Site directed mutagenesis of the pMH406 plasmid was performed exactly as described previously44. The esxAMT E12C mutation was generated using OCR015 (5′-TTTCGCGGGTATCTGCGCCGCGGCAAG-3′) and the reverse complement, OCR016, oligonucleotide primers where bolded residues indicate the altered codon. The esxAMT E12K mutation was generated using the OCR011 (5′-TTTCGCGGGTATCAAGGCCGCGGCAAG-3′) and the reverse complement, OCR012, oligonucleotide primers. The esxAMT E12R was generated using the OCR013 (5′-TTTCGCGGGTATCAGGGCCGCGGCAAG-3′) and the reverse complement, OCR014, oligonucleotide primers. All primers were purchased from Integrated DNA Technologies (IDT, Coralville, IA). The resulting mutagenized plasmids were confirmed by DNA sequencing analysis performed at the Genomics and Bioinformatics Core Facility at the University of Notre Dame using the MOPs forward (5′- TGCTTCCGGCTCGTATAATG-3′) and reverse (5′-GTCTTTCGACTGAGCCTTTC-3′) primers.
Mycobacterial strain construction
Following confirmation by sequencing, the pMH406-based plasmids bearing either the wild-type or mutated alleles of esxAMT were introduced by electroporation into the ΔesxBA or ΔesxA strains of M. marinum and M. tuberculosis strains, respectively, as described previously45. The resulting strains bearing the plasmids were selected on 7H11 agar containing 50 μg/ml hygromycin or 20 μg/ml kanamycin as required. Colonies were streak purified onto 7H11 agar with antibiotics and grown in 7H9 broth. Genomic DNA from each strain was purified as described previously6,23. Integration of the plasmid in each mycobacterial strain was confirmed by using the MOPs sequencing primers described above.
Sheep red blood cell (sRBC; defibrinated sheep red blood cells; BD, Sparks, MD) lysis assays were performed as previously described44 with minor modifications. M. marinum cells were washed three times with phosphate buffered saline (PBS) and then resuspended in 300 μl of PBS with 100 μl of sRBCs, collected by centrifugation, and incubated at 30 °C for 2 hours. For each strain, at least three biological replicates were performed. Each sample was read in triplicate. For all hemolysis assays, incubation of the sRBCs with dH2O served as a positive control for maximal lysis. Incubation of the sRBCs with PBS served as a negative control for lysis. Error bars are the standard deviations for triplicate readings. Data shown is from a single, representative assay.
The RAW 264.7 murine macrophage cell line (ATCC TIB-71) was maintained in DMEM (Life Technologies, San Diego, CA) supplemented with 10% heat inactivated fetal bovine serum (FBS) (HyClone Laboratories, Logan, UT) and maintained at 37 °C with 5% CO2, as described previously14,22. Cells were plated at 5 × 105 cells per well in 24-well plates (Grenier Bio-one, Monroe, NC) for 24 hours and then infected with 2.5 × 106 cells of M. marinum. Cells were incubated for 2 hours at 37 °C with 5% CO2 and then treated with 100 μg/mL gentamycin (Research Products International [RPI], Mount Prospect, IL) for 2 hours at 37 °C to kill external mycobacterial cells. The cells were then washed three times using PBS and fresh media was added. Staining of macrophages was performed using Ethidium homodimer-1 (EthD-1; Life Technologies, Eugene, OR) and the Live/Dead viability/cytotoxicity kit for mammalian cells (Life Technologies) at 24 hours post infection. The cells were imaged using an AxioObserver A1 inverted microscope (Zeiss, Thornwood, NY) as in Kennedy et al.46. We performed three independent biological replicates (each with two wells per strain) of the macrophage infections. For each biological replicate, we counted the permeabilized cells over ten fields per sample and reported the average. All magenta images were adjusted in the exact same manner to increase visibility (brightness +40% and contrast −40%).
Macrophage infection and RNA extraction
The RAW 264.7 macrophage cell line was prepared and infected exactly as previously described39. RNA isolation was performed as described previously with the following minor changes: after the RNA binding step, the column was wash with 400 μL RNA wash buffer and the RNA was column digested by addition of 80 μL of the DNase I reaction mix (Zymo Research, Irvine, CA) directly to the column. After DNase I treatment, RNA was washed further according to the manufacturer’s instructions before being eluted. The RNA was quantified using a Nanodrop 2000 instrument (Thermo Fisher).
Quantitative real-time PCR (qRT-PCR) was performed exactly as previously described39. The reaction and thermal cycling parameters were performed exactly as described in Mba Medie et al.39. The mRNA level of beta interferon (IFN-β) was normalized to that of actin. The IFN-β, and the β-actin primers oligonucleotide primers used here are: IFNβ-F (5′-CTGGAGCAGCTGAATGGAAAG-3′), IFNβ-R (5′-CTTGAAGTCCGCCCTGTAGGT-3′), β-actin-F (5′-AGGTGTGATGGTGGGAATGG-3′) and β-actin-R (5′-GCCTCGTCACCCACATAGGA-3′)20,47. We performed four biological replicates in M. marinum and two biological replicates in M. tuberculosis. Each biological replicate included two technical replicates. Statistical analysis was performed with GraphPad Prism, version 6.07 (GraphPad Software, San Diego, CA) using a Mann-Whitney U nonparametric test48.
Esx-1 protein secretion assays
M. marinum protein secretion assays were performed exactly as described previously26,45. In short, M. marinum strains were grown in Middlebrook 7H9 broth and then diluted into Sauton’s liquid media with 0.01% Tween-80 at OD600 of 0.8 in 50 ml. Following 48 hours of growth, cell lysate and culture filtrate protein fractions were prepared and the protein concentration determined exactly as previously described26,45.
M. tuberculosis protein secretion assays were performed as previously described in Stanley et al. with minor changes8. Briefly, strains were cultured in Sauton’s liquid media supplemented with 0.05% Tween-80 and grown at 37 °C until OD600 1–1.5. The cultures were subsequently diluted into Sauton’s liquid media to an OD600 of 0.05 with 0.005% Tween-80 and grown for 5–7 days at 37 °C, with gentle shaking. Cultures were harvested and fractions prepared as previously described6,45.
Protein fractions were run on Mini-Protean TGX Tris-HCl precast polyacrylamide gels (Bio-Rad, Hercules, CA) and transferred to a nitrocellulose membrane. Proteins were visualized using antibodies against EsxB (CFP-10) (1:5,000 [NR-13801; BEI Resources]), MPT-32 (1:5,000 [NR-13807; BEI Resources]), and RNA polymerase subunit β (RNAP-β) (1:5,000 [ab12087; Abcam] Abcam, Cambridge, UK) as previously described39.
50 μg samples of cytosolic and secreted protein fractions were precipitated and digested as previously described22,23. Unless otherwise stated reagents were from Sigma-Aldrich (St Louis, MO). Briefly, acetone precipitated pellets were resuspended with vigorous vortexing in 50% 2,2,2 Trifluoroethanol, 50 mM ammonium bicarbonate, and Dithiotheritol (DTT) (Sigma, St Louis, MO) was added to a concentration of 10 mM. After incubation at 58 °C for 60 minutes, Iodoacetamide was added to a concentration of 20 mM and allowed to react in the dark for 15 minutes at room temperature. Additional DTT was added to quench the reaction. Aminoethylation was performed in lieu of alkylation using an identical procedure except the concentration of 2-Bromoethylamine HBr was increased to 50 mM × 2 1 hour reactions and ammonium bicarbonate was increased to 100 mM. The volume was brought to 200 μl in 50 mM ammonium bicarbonate and 500 ng of sequencing grade trypsin (Promega, Madison, WI) was added and incubated at 37 °C for 2 hours. An additional 500 ng of trypsin was added for overnight digestion. Digestion was quenched with addition of 5% formic acid (FA) (Thermo Fisher) until pH <4 and each sample was desalted with a C18 spin column using standard conditions (Protea, Morgantown, WV). Desalted-dried digests were resuspended to 500 ng/μl in 0.3% FA in water for LC/MS/MS analysis.
LC/MS/MS was performed essentially as described previously22,23. Briefly 2 μl samples (~1 μg digest) were injected onto a 100 μm × 100 mm C18BEH nano column with an Acquity nano UHPLC (Waters, Billerica, MA). A 90 minute gradient from 2–33% [A = H2O, 0.1%FA, B = Acetonitrile 0.1%FA (Burdick & Jackson, Morris Plains, NJ)] was run at a flow rate of 800 nl/min. MS & MS/MS were acquired on an Q-Exactive Orbitrap instrument (Thermo Fisher) running a TOP20 data dependent method47. MS and MS/MS spectra were converted to <mgf> peak lists using Proteome Discoverer 1.4 (Thermo Fisher) and exported for database searching. Database searching was performed using ProteinPilot/Paragon (ABSciex, Redwood City, CA) using a thorough-search parameter against a concatenated FASTA database of M. marinum M and M. tuberculosis Erdman strains respectively. This database contained a common contaminant file and the custom sequences of EsxA E12 C, K, and R substitutions to match the altered proteins. False discovery rates (FDR) were calculated using the method of Elias et al.49 and the instantaneous (Local) FDR from a quadratic fit was <10−6 % for EsxA peptides reported here. The 1% FDR population of proteins identified in culture filtrates and cytosolic fractions is typical for what we observe and is available for download through the MassIVE database http://massive.ucsd.edu with the accession # MSV000080101 or via ftp at ftp://MSV000080101@massive.ucsd.edu https://dl.dropboxusercontent.com/u/18562762/Reyna_SuppData_NB.zip.
Analysis of mycobacterial peptide distributions
For the data presented in Fig. 1, in silico digest of the M. tuberculosis Erdman strain proteome acquired from Uniprot (www.uniprot.org) was produced using the Pacific Northwest National Laboratories Protein Digestion Simulator (omics.pnl.gov) v2.2.5679; July 20, 2015. Digestion settings were set to full tryptic (KR not P) normal output with one missed cleavage allowed and a minimum fragment mass of 400 and maximum fragment mass of 6000 Da. The peptides generated were used to determine the distribution of peptide lengths. For the experimental data sets in-house digestion data of the M. tuberculosis Erdman strain was prepared using the filter–aided sample preparation method for LC/MS/MS analysis50. Briefly, 100 μg of cell pellet and culture filtrate fractions were denatured using 2.5% sodium dodecyl sulfate in 100 mM Tris, pH 8 with 100 mM DTT and reduced at 95 °C for 5 minutes. Samples were cooled and loaded onto 10 K microcon centrifugal filters (Millipore, Billerica, MA), volume added with 8 M Urea in 100 mM Tris (pH 8.0) spun and then alkylated with 50 mM iodoacetamide in 100 mM Tris and 8 M urea in the dark for 5 minutes. Four subsequent spin-washes were carried out using 8 M urea in 100 mM Tris pH 8.0. Two additional spin-washes in 100 mM ammonium bicarbonate were performed prior to digestion with trypsin at a 1:50 trypsin to substrate ratio at 37 °C for approximately 16 hours. Digests were spin-eluted and a second extraction with 500 mM NaCl was performed, followed by quenching with 5 μl FA. Desalting was performed using C18 spin columns (Protea, Morganstown, WV) and dried prior to LC/MS/MS analysis according to manufacturer’s instructions. Samples were re-suspended to 500 ng/μl and LC/MS/MS analysis was done essentially as described above. Identified peptides at a 1% FDR were used to determine peptide length distributions. Peptide lengths were determined by counting the number of amino acids. The percentage of peptides within each population was determined by taking the number of peptides within a bin and dividing by the total number of peptides at the 1% FDR. The percentage of peptides within each bin (≤5, 6–10, 11–15, etc.) were graphed for each individual data set (in silico and experimental). All calculations were done using Microsoft Excel. The missed cleavage rate for in silico digestion was chosen by analyzing the empirical data sets of the mycobacterial proteomes and determining the average number of missed cleavages per protein using the following calculation: (total number of missed cleavages/total number of peptides)*(average number of peptides per protein).
Visualization of the M. tuberculosis Peptide Map
For the analysis in Supplemental Fig. 1, in silico digestion of the M. tuberculosis Erdman proteome was done as described above. Settings were set to zero missed cleavages allowed and no mass cutoff for the purpose of determining the entirety of the theoretical peptides generated from a perfect trypsin digestion. Peptides generated are representative of digestion of individual proteins from the N-to C-terminus (left to right). Peptide lengths were determined and used to create a heat map using the open source programming software, R51. For analysis of the Esx-like proteins in Supplemental Fig. 4, protein sequences were obtained from Mycobrowser49 and were analyzed in silico for digestion by trypsin with PeptideCutter (ExPASy). Protein graphics were generated using Microsoft Excel.
How to cite this article: Reyna, C. et al. Rational engineering of a virulence gene from Mycobacterium tuberculosis facilitates proteomic analysis of a natural protein N-terminus. Sci. Rep. 6, 33265; doi: 10.1038/srep33265 (2016).
Swaney, D. L., Wenger, C. D. & Coon, J. J. Value of using multiple proteases for large-scale mass spectrometry-based proteomics. Journal of proteome research 9, 1323–1329, doi: 10.1021/pr900863u (2010).
Tsiatsiani, L. & Heck, A. J. Proteomics beyond trypsin. The FEBS journal 282, 2612–2626, doi: 10.1111/febs.13287 (2015).
Shortreed, M. R. et al. Elucidating Proteoform Families from Proteoform Intact-Mass and Lysine-Count Measurements. Journal of proteome research 15, 1213–1221, doi: 10.1021/acs.jproteome.5b01090 (2016).
Giansanti, P., Tsiatsiani, L., Low, T. Y. & Heck, A. J. Six alternative proteases for mass spectrometry-based proteomics beyond trypsin. Nature protocols 11, 993–1006, doi: 10.1038/nprot.2016.057 (2016).
Mohammed, Y. & Borchers, C. H. An extensive library of surrogate peptides for all human proteins. Journal of proteomics 129, 93–97, doi: 10.1016/j.jprot.2015.07.025 (2015).
Salinas, J. L. et al. Leveling of Tuberculosis Incidence — United States, 2013–2015. Morb Mortal Wkly Rep 65, 273–278; 10.15585/mmwr.mm6511a2 (2016)
Sorensen, A. L., Nagai, S., Houen, G., Andersen, P. & Andersen, A. B. Purification and characterization of a low-molecular-mass T-cell antigen secreted by Mycobacterium tuberculosis. Infection and immunity 63, 1710–1717 (1995).
Stanley, S. A., Raghavan, S., Hwang, W. W. & Cox, J. S. Acute infection and macrophage subversion by Mycobacterium tuberculosis require a specialized secretion system. Proceedings of the National Academy of Sciences of the United States of America 100, 13001–13006 (2003).
Guinn, K. M. et al. Individual RD1-region genes are required for export of ESAT-6/CFP-10 and for virulence of Mycobacterium tuberculosis. Molecular microbiology 51, 359–370 (2004).
Hsu, T. et al. The primary mechanism of attenuation of bacillus Calmette-Guerin is a loss of secreted lytic function required for invasion of lung interstitial tissue. Proceedings of the National Academy of Sciences of the United States of America 100, 12420–12425 (2003).
Mahairas, G. G., Sabo, P. J., Hickey, M. J., Singh, D. C. & Stover, C. K. Molecular analysis of genetic differences between Mycobacterium bovis BCG and virulent M. bovis. Journal of bacteriology 178, 1274–1282 (1996).
Pym, A. S., Brodin, P., Brosch, R., Huerre, M. & Cole, S. T. Loss of RD1 contributed to the attenuation of the live tuberculosis vaccines Mycobacterium bovis BCG and Mycobacterium microti. Molecular microbiology 46, 709–717 (2002).
Penn-Nicholson, A., Nemes, E., Hanekom, W. A., Hatherill, M. & Scriba, T. J. Mycobacterium tuberculosis-specific CD4 T cells are the principal source of IFN-gamma in QuantiFERON assays in healthy persons. Tuberculosis (Edinburgh, Scotland) 95, 350–351, doi: 10.1016/j.tube.2015.03.002 (2015).
Broset, E., Martin, C. & Gonzalo-Asensio, J. Evolutionary landscape of the Mycobacterium tuberculosis complex from the viewpoint of PhoPR: implications for virulence regulation and application to vaccine development. MBio 6, e01289–01215, doi: 10.1128/mBio.01289-15 (2015).
Samten, B., Wang, X. & Barnes, P. F. Immune regulatory activities of early secreted antigenic target of 6-kD protein of Mycobacterium tuberculosis and implications for tuberculosis vaccine design. Tuberculosis (Edinburgh, Scotland) 91 Suppl 1, S114–118, doi: 10.1016/j.tube.2011.10.020 (2011).
Stamm, L. M. et al. Mycobacterium marinum escapes from phagosomes and is propelled by actin-based motility. The Journal of experimental medicine 198, 1361–1368 (2003).
Houben, D. et al. ESX-1-mediated translocation to the cytosol controls virulence of mycobacteria. Cellular microbiology 14, 1287–1298 (2012).
van der Wel, N. et al. M. tuberculosis and M. leprae translocate from the phagolysosome to the cytosol in myeloid cells. Cell 129, 1287–1298 (2007).
Manzanillo, P. S., Shiloh, M. U., Portnoy, D. A. & Cox, J. S. Mycobacterium Tuberculosis Activates the DNA-Dependent Cytosolic Surveillance Pathway within Macrophages. Cell host & microbe 11, 469–480 (2012).
Simeone, R. et al. Phagosomal rupture by Mycobacterium tuberculosis results in toxicity and host cell death. PLoS pathogens 8, e1002507 (2012).
Bottai, D. et al. Disruption of the ESX-5 system of Mycobacterium tuberculosis causes loss of PPE protein secretion, reduction of cell wall integrity and strong attenuation. Molecular microbiology 83, 1195–1209 (2012).
Abdallah, A. M. et al. The ESX-5 secretion system of Mycobacterium marinum modulates the macrophage response. J Immunol 181, 7166–7175 (2008).
Abdallah, A. M. et al. A specific secretion system mediates PPE41 transport in pathogenic mycobacteria. Molecular microbiology 62, 667–679 (2006).
Okkels, L. M. et al. CFP10 discriminates between nonacetylated and acetylated ESAT-6 of Mycobacterium tuberculosis by differential interaction. Proteomics 4, 2954–2960 (2004).
Lange, S. et al. Analysis of protein species differentiation among mycobacterial low-Mr-secreted proteins by narrow pH range Immobiline gel 2-DE-MALDI-MS. Journal of proteomics 97, 235–244 (2014).
Champion, P. A., Champion, M. M., Manzanillo, P. & Cox, J. S. ESX-1 secreted virulence factors are recognized by multiple cytosolic AAA ATPases in pathogenic mycobacteria. Molecular microbiology 73, 950–962 (2009).
Li, Y. et al. Capillary Zone Electrophoresis-Electrospray Ionization-Tandem Mass Spectrometry as an Alternative Proteomics Platform to Ultraperformance Liquid Chromatography-Electrospray Ionization-Tandem Mass Spectrometry for Samples of Intermediate Complexity. Analytical chemistry 84, 1617–1622 (2012).
Desiere, F. et al. The PeptideAtlas project. Nucleic acids research 34, D655–658 (2006).
Deutsch, E. W. The PeptideAtlas Project. Methods in molecular biology (Clifton, N.J) 604, 285–296, doi: 10.1007/978-1-60761-444-9_19 (2010).
Brodin, P. et al. Functional analysis of ESAT-6, the dominant T-cell antigen of mycobacterium tuberculosis, reveals key residues involved in secretion, complex-formation, virulence and immunogenicity. J Biol Chem 40, 33953–9. (2005).
Gallegos, A. M. et al. A gamma interferon independent mechanism of CD4 T cell mediated control of M. tuberculosis infection in vivo. PLoS pathogens 7, e1002052, doi: 10.1371/journal.ppat.1002052 (2011).
Berthet, F. X., Rasmussen, P. B., Rosenkrands, I., Andersen, P. & Gicquel, B. A. Mycobacterium tuberculosis operon encoding ESAT-6 and a novel low-molecular-mass culture filtrate protein (CFP-10). Microbiology 144 (Pt 11), 3195–3203 (1998).
Tobin, D. M. & Ramakrishnan, L. Comparative pathogenesis of Mycobacterium marinum and Mycobacterium tuberculosis. Cellular microbiology 10, 1027–1039 (2008).
Stinear, T. P. et al. Insights from the complete genome sequence of Mycobacterium marinum on the evolution of Mycobacterium tuberculosis. Genome research 18, 729–741 (2008).
King, C. H., Mundayoor, S., Crawford, J. T. & Shinnick, T. M. Expression of contact-dependent cytolytic activity by Mycobacterium tuberculosis and isolation of the genomic locus that encodes the activity. Infection and immunity 61, 2708–2712 (1993).
Gao, L. Y. et al. A mycobacterial virulence gene cluster extending RD1 is required for cytolysis, bacterial spreading and ESAT-6 secretion. Molecular microbiology 53, 1677–1693 (2004).
Lee, M. H. & Manoil, C. Engineering trypsin-sensitive sites in a membrane transport protein. Protein engineering 10, 715–723 (1997).
Lee, M. H., Kosuk, N., Bailey, J., Traxler, B. & Manoil, C. Analysis of F factor TraD membrane topology by use of gene fusions and trypsin-sensitive insertions. Journal of bacteriology 181, 6108–6113 (1999).
Mba Medie, F., Champion, M. M., Williams, E. A. & Champion, P. A. Homeostasis of N-alpha-terminal acetylation of EsxA correlates with virulence in Mycobacterium marinum. Infection and immunity 82, 4572–4586, doi: 10.1128/IAI.02153-14 (2014).
Abdallah, A. M. et al. Type VII secretion--mycobacteria show the way. Nature reviews 5, 883–891 (2007).
Siegrist, M. S. et al. Mycobacterial Esx-3 is required for mycobactin-mediated iron acquisition. Proceedings of the National Academy of Sciences of the United States of America 106, 18792–18797 (2009).
Serafini, A., Boldrin, F., Palu, G. & Manganelli, R. Characterization of a Mycobacterium tuberculosis ESX-3 conditional mutant: essentiality and rescue by Iron and Zinc. Journal of bacteriology 191, 6340–4. (2009).
Jeffery, C. J. Protein species and moonlighting proteins: Very small changes in a protein’s covelant structure can changes its biochemical function. J Proteomics 134, 19–24. doi: 10.1016/j.jprot.2015.10.003 (2016).
Champion, P. A., Stanley, S. A., Champion, M. M., Brown, E. J. & Cox, J. S. C-terminal signal sequence promotes virulence factor secretion in Mycobacterium tuberculosis. Science 313, 1632–1636 (2006).
Champion, M. M., Williams, E. A., Kennedy, G. M. & Champion, P. A. Direct detection of bacterial protein secretion using whole colony proteomics. Mol Cell Proteomics 11, 596–604 (2012).
Kennedy, G. M., Hooley, G. C., Champion, M. M., Medie, F. M. & Champion, P. A. A novel ESX-1 locus reveals that surface associated ESX-1 substrates mediate virulence in Mycobacterium marinum. Journal of bacteriology 196, 1877–1888 (2014).
Stanley, S. A., Johndrow, J. E., Manzanillo, P. & Cox, J. S. The Type I IFN response to infection with Mycobacterium tuberculosis requires ESX-1-mediated secretion and contributes to pathogenesis. J Immunol 178, 3143–3152 (2007).
Sun, L. et al. Quantitative proteomics of Xenopus laevis embryos: expression kinetics of nearly 4000 proteins during early development. Scientific reports 4, 4365, doi: 10.1038/srep04365 (2014).
Elias, J. E. & Gygi, S. P. Target-decoy search strategy for increased confidence in large-scale protein identifications by mass spectrometry. Nature methods 4, 207–214 (2007).
Wisniewski, J. R., Zougman, A., Nagaraj, N. & Mann, M. Universal sample preparation method for proteome analysis. Nature methods 6, 359–362, doi: 10.1038/nmeth.1322 (2009).
R. Core Team R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL http://www.R-project.org/ (2013).
Research reported in this study was supported by the National Institute of Allergies and Infectious Diseases of the National Institutes of Health under award numbers R01AI106872 to P.A.C. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. We thank the Genomics and Bioinformatics Facility and the Mass Spectrometry and Proteomics Facility at the University of Notre Dame for their assistance in this research. We thank Dr. Stuart Jones for assistance with the use of R.
The authors declare no competing financial interests.
Electronic supplementary material
About this article
Cite this article
Reyna, C., Mba Medie, F., Champion, M. et al. Rational engineering of a virulence gene from Mycobacterium tuberculosis facilitates proteomic analysis of a natural protein N-terminus. Sci Rep 6, 33265 (2016). https://doi.org/10.1038/srep33265