Deep proteome coverage advances knowledge of Treponema pallidum protein expression profiles during infection

Propagation and isolation of T. pallidum

Treponema pallidum subsp. pallidum (Nichols strain) was propagated in male specific pathogen-free (SPF) New Zealand White rabbits as described elsewhere²⁶. The animal study was reviewed and approved by the local institutional review board at the University of Victoria under protocol 2020-024, and was conducted in strict accordance with standard accepted principles as set forth by the Canadian Council on Animal Care, National Institutes of Health and the United States Department of Agriculture in a facility accredited by the Canadian Council on Animal Care and the American Association for the Accreditation of Laboratory Animal Care. For extraction, treponemes were harvested from the testes of rabbits approximately 10–12 days after infection in sterile saline (0.9% w/v NaCl, pH 7.0) in the presence or absence (refer to method development section below for details) of 10% normal rabbit serum (NRS). Extractions were performed in an anaerobic chamber (Coy Laboratories, Grass Lake, MI, USA) at room temperature in an atmosphere of 1.5–3% O₂ and 5% CO₂, balanced with N₂. Rabbit cells and debris were separated and removed as described below. Treponemes in suspension were enumerated using a Nikon Eclipse E600 darkfield microscope (Nikon Canada, Mississauga, ON, Canada).

Treponema pallidum protein sample preparation; method optimization

Due to the inherent fragility of T. pallidum and the sample complexity arising from contaminating rabbit proteins/cellular debris from in vivo culturing of T. pallidum, we optimized a protocol for the preparation and mass spectrometry-based analyses of T. pallidum samples in order to maximize coverage of the T. pallidum proteome. The flow diagram shown in Fig. 1 outlines all major steps and parameters that were tested in the optimization steps, including the isolation of T. pallidum (as described above) and mass spectrometry sample preparation (as detailed below). Since some of the key methods used in our T. pallidum proteomics workflow had not been used in previous treponemal proteomics studies, there were limited literature reports that could inform workflow optimization. In our initial optimizations (samples 1–9), we used methods from previous proteomics analyses of other organisms. In our later optimization analysis (sample 10), we combined all the methods from our initial findings (samples 1–9) that had led to increased T. pallidum proteome coverage.

Optimization of a workflow for deep proteome coverage of in vivo-grown T. pallidum. Schematics showing (a) isolation of T. pallidum from rabbits, (b) the major steps comprising T. pallidum sample preparation, and (c) the main optimization step in mass spectrometry sample preparation, processing, and data acquisition. The workflow indicates each of the individual steps that were performed in the optimization of the protocol for global profiling of T. pallidum protein expression (left). The corresponding table shows the variable parameter conditions used in each of the biological replicate samples at each of the individual optimization steps (right). Samples 1–10 were used for optimizing the protocol during the method development stages; sample 10 conditions were found to be optimal (red text). Samples 11 and 12 (red text) correspond to two biological replicate samples that were processed using the optimized protocol used for sample 10 to obtain three biological replicate samples prepared using an identical protocol. The total number of treponemes used in the preparation of each of the 12 samples is indicated (bottom). Plus sign; a parameter condition has been included in a protocol step; minus sign, a parameter condition has been omitted in a protocol step. +/− NRS = the addition or omission of normal rabbit serum (NRS) during T. pallidum isolation. U/C = the use of either ultrafiltration (U) or high-speed centrifugation (C) during removal of contaminating rabbit components. +/− HI = the use or omission of heat inactivation during T. pallidum sample preparation. (d) The total number of T. pallidum proteins that were detected and identified in each of the 12 in-vivo grown T. pallidum biological replicate samples. For each bar, the number of T. pallidum proteins that were identified via the detection of one tryptic peptide (blue) or two tryptic peptides (orange) in each of the 12 individual samples is indicated.

Treponema pallidum was isolated from rabbits (treponeme cell number range: 1.7–4.1 × 10⁸). The in vivo samples were harvested in either the presence (six samples) or absence (four samples) of NRS (Fig. 1a). NRS helps to maintain treponeme viability, however, its addition also increases the complexity of the samples by increasing the amount of contaminating rabbit proteins, including highly abundant albumin.

The next parameters to be investigated pertained to sample preparation of harvested T. pallidum (Fig. 1b). The first step in sample preparation involved removing as much contaminating rabbit cellular debris and proteins as possible. Two methods, each comprised of two components, were investigated; (1) low-speed centrifugation followed by ultrafiltration (eight samples), and (2) low-speed centrifugation followed by high-speed centrifugation (two samples). In both methods, insoluble rabbit gross cellular debris was removed via two centrifugation steps at 220 × g (5 min each, room temperature) followed by two additional centrifugation steps at 400 × g (7 min each, room temperature). In method (1), the slow-speed centrifugation steps were followed by ultrafiltration in order to remove soluble contaminants in the treponemal supernatant and to wash the suspended treponemes. In this method, the T. pallidum-containing supernatants (2.5 mL) were centrifuged at 220 × g (room temperature) using Protein Ark Proteus-X-Spinner 2.5 ultrafiltration concentrators (300 kDa MWCO) (Canadian Life Science, Peterborough, ON, Canada) until 1 mL of the T. pallidum-containing supernatants (retentate) remained in the bottom of the ultrafiltration concentrators. The ultrafiltrates were removed and 1.5 mL of sterile saline (0.9% w/v NaCl, pH 7.0) was added to wash the T. pallidum-containing retentate. The ultrafiltration wash/concentration steps were repeated two more times. The final concentrated and washed 1 ml T. pallidum-containing saline suspensions were pooled for each of the individual rabbits and stored at – 80 °C for subsequent lyophilisation. In method (2), the slow-speed centrifugation steps were followed by high-speed centrifugation in order to remove soluble contaminants in the treponemal supernatant and to wash the suspended treponemes. In this method, the T. pallidum-containing supernatants were centrifuged at 17,000 × g for 5 min at room temperature. The supernatant was discarded, and the T. pallidum pellet was gently resuspended in sterile saline (0.9% w/v NaCl, pH 7.0) and centrifuged again at 17,000 × g for 5 min at room temperature, after which the supernatant was discarded.

As shown in Fig. 1b, the eight T. pallidum samples that were subjected to ultrafiltration were lyophilized to concentrate the samples for mass spectrometry analyses. Following removal from − 80 °C storage, the frozen samples were lyophilized overnight (> 16 h) using a VirTis Freezemobile freeze dryer/lyophiliser (model 12EL; SP Industries, Warminster, PA, USA).

Chemical lysis of all 10 T. pallidum samples was performed by resuspending and incubating the lyophilized and pelleted samples in lysis buffer (500 µL per sample; 50 mM ammonium bicarbonate pH 8.0, 0.9% sodium deoxycholate [Sigma-Aldrich Canada Co., Oakville, ON, Canada]) for 30 min on ice, with 30 s vortex mixing steps every five min (Fig. 1b). Physical lysis was then performed on all 10 samples by ultrasonication at 6 °C using a Covaris ME220 focused-ultrasonicator with the following parameters: 3 × 30 s at 6W, 20% duty factor, 200 cycles per burst, 30 s rest between the three cycles (Covaris, LLC., Woburn, MA, USA).

Following cell lysis, three of the 10 samples were heated at 95 °C for 5 min in order to inactivate proteases and prevent non-trypsin mediated proteolysis of T. pallidum proteins (Fig. 1b). However, this resulted in heavy precipitation in the three samples. These three samples were recovered by centrifugation at 17,000 × g for 10 min. The supernatants were separated from the pellets and stored at 4 °C. The three pellets were incubated with solubilisation buffer (500 µL per sample; 300 mM Tris pH 8.0, 8 M Urea) for 45 min at room temperature with 60 s vortex mixing steps every 15 min. The three samples were then centrifuged at 13,000 × g for 10 min and the three urea-extracted supernatants were stored until trypsin digestion, as described below. In order to ensure complete removal of precipitated, aggregated, or insoluble proteins and cellular debris, all 10 T. pallidum lysate samples were centrifuged at 17,000 × g for 10 min at 4 °C. The supernatants were removed and protein concentrations were determined by measuring the absorbance at 280 nm (Beckman Coulter DU 730 Life Science UV/Vis Spectrophotometer; Beckman Coulter Canada, Mississauga, ON, Canada) and by performing BCA assays using the Thermo Scientific Pierce BCA protein assay kit (Thermo Fisher Scientific, Ottawa, ON, Canada).

In solution trypsin digestion

Following the T. pallidum sample preparation optimization steps, all 10 lysate samples were digested with trypsin. For the three samples containing urea, 50 mM ammonium bicarbonate (pH 7.8) was added to 300 µg of protein (400 µL final volume). For samples containing no urea, 50 mM ammonium bicarbonate (pH 7.8) was added to obtain protein concentrations of ~ 1.7 mg/mL. Each sample was reduced at 37 °C for 30 min by the addition of 100 mM dithiothreitol (80 µL [urea-containing] or 120 µL [no urea samples]) and then alkylated at room temperature in the dark for 30 min by the addition of 240 mM iodoacetamide (80 µL [urea-containing] or 120 µL [no urea samples]). Urea-containing samples were diluted by the addition of 50 mM ammonium bicarbonate (pH 7.8) (2.6 mL). For the urea-containing samples, in-solution tryptic digestion was performed at 37 °C for 18 h by the addition of 30 µg of trypsin (Worthington Biochemical Corporation, Lakewood, NJ, USA). For the urea-free samples, 100 µg of trypsin was added per mg of protein. Protein digestion was stopped by the addition of formic acid (1% final concentration). Solid phase extraction cleanup was performed using Hydrophilic-Lipophilic-Balanced (HLB) columns (Waters Corporation, Milford, MA, USA). The digested protein samples were eluted with 60% acetonitrile/0.1% formic acid (300 µL per sample) and reduced to dryness in a SpeedVac vacuum concentrator. Samples were then re-suspended in 10 mM ammonium hydroxide, pH 10.0 (900 µL).

High-pH reversed phase fractionation

To determine the effect of sample complexity reduction on whole proteome coverage, eight of the 10 trypsin-digested T. pallidum samples were separated into 24 fractions based on hydrophobicity using high-pH reversed phase fractionation (Fig. 1c). Specifically, an Agilent 1290 HPLC system (Agilent, Santa Clara, CA, USA) was equipped with an XBridge BEH300 C18 peptide separation technology (PST) column (250 mm × 4.6 mm, 5 µm, 300 Å) (Waters Corporation, Milford, MA, USA). Buffer A consisted of 10 mM ammonium hydroxide (pH 10.0), and buffer B was comprised of 80% acetonitrile and 10 mM ammonium hydroxide (pH 10). The T. pallidum samples were diluted to 0.9 mL total volume with buffer A and injected onto the column with a constant flow rate set at 0.75 mL/min. The column was equilibrated for 5 min in buffer A before a gradient (5–45%) of buffer B was performed over 75 min. Fractions were collected every min for 96 min, reduced in volume by lyophilisation, rehydrated with 2% acetonitrile/0.1% formic acid (300 µL), and concatenated into 24 fractions by combining every 24th fraction (e.g. fractions 1, 25, 49, and 73 were combined).

LC–MS/MS analyses

The T. pallidum samples (fractionated and non-fractionated) were subjected to liquid chromatography-tandem mass spectrometry (LC–MS/MS) for global, high confidence protein identifications. An aliquot (5 µL) of each concatenated fraction or non-fractionated sample was separated by on-line reversed phase liquid chromatography using a Thermo Scientific EASY-nLC 1000 system with a reversed-phase pre-column packed with Magic C18-AQ resin (100 µm I.D., 2.5 cm length, 5 µm, 100 Å) and an in-house prepared reversed phase nano-analytical column packed with Magic C-18AQ resin (75 µm I.D., 15 cm length, 5 µm, 100 Å) (Michrom BioResources Inc., Auburn, CA, USA), at a flow rate of 300 nL/min. Solvent A was comprised of 2% acetonitrile and 0.1% formic acid while solvent B consisted of 90% acetonitrile and 0.1% formic acid. The T. pallidum samples were separated using a 120-min gradient comprised of the following steps: (1) 0–100 min, gradient change from 95% A/5% B to 58% A/42% B, (2) 100–115 min, gradient change from 58% A/42% B to 0% A/100% B, and (3) 115–120 min, gradient held at 0% A/100% B.

The chromatography system was coupled on-line with an Orbitrap Fusion Tribrid mass spectrometer (Thermo Fisher Scientific, San Jose, CA, USA) equipped with a Nanospray Flex NG source (Thermo Fisher Scientific). The Orbitrap Fusion Tribrid mass spectrometer instrument parameters (Fusion Tune 3.3 software) used were as follows: nano-electrospray ion source with spray voltage = 2.55 kV; capillary temperature = 275 ℃; survey MS1 scan = m/z range 350–1800, profile mode, resolution 120,000 FWHM@200 m/z, number of microscan = 1, automatic maximum inject time. Internal calibration was performed using the lock mass for siloxane (445.120024 m/z) as a reference. Data-dependent acquisition Orbitrap survey spectra were scheduled at least every 3 s, with the software determining “Automatic” number of MS/MS acquisitions during this period. The automatic gain control (AGC) target values for FTMS and MSn were 400,000 and 10,000 respectively. The most intense ions (m/z range 350–1800, charge state 2–5) exceeding 50,000 counts were selected for higher-energy collisional dissociation (HCD) ion trap MS/MS fragmentation with detection in centroid mode. Dynamic exclusion settings were: repeat count = 2; exclusion duration = 15 s with a 10 ppm mass window. The ddMS2 IT HCD scan used a quadrupole isolation window of 1.6 Da; rapid scan rate, auto mass range, centroid detection, 1 microscan, automatic maximum injection time, and stepped HCD collision energy 28, 30 and 32%.

Biological replicate sample preparation using the optimized proteomics workflow

The optimized proteomics workflow used for sample 10 was repeated two more times in order to prepare two additional optimized biological replicate samples (samples 11, 12; both samples from different rabbits). This method was performed as described above with the following key optimizations: (1) NRS was omitted during treponeme isolation; (2) high-speed centrifugation was used for the removal of contaminating proteins during sample preparation; (3) lyophilisation and heat inactivation steps were omitted; and (4) high-pH reversed phase fractionation (24 fractions) was used (Fig. 1). In total 12 biological replicate samples were prepared: samples 1–9 constituted method optimization samples while samples 10–12 constituted samples prepared using the optimized protein preparation method.

Mass spectrometry parameters and data analyses: protein identifications and validation

Raw files were created by XCalibur 4.3.73.11 (Thermo Scientific) software. Tandem mass spectra were extracted and charge state deconvoluted by Proteome Discoverer version 2.5 (Thermo Scientific). Deisotoping was not performed. All MS/MS samples were analyzed using Sequest (Thermo Fisher Scientific, San Jose, CA, USA; node in Proteome Discoverer 2.5.0.400) containing a customized T. pallidum database comprised of all unique protein entries from all National Centre for Biotechnology Information (NCBI) whole proteome annotation revisions of the T. pallidum Nichols strain, NCBI reference sequence NC_021490 (18 whole proteome annotation revisions, June 17th 2013–July 4th 2021; https://www.ncbi.nlm.nih.gov/nuccore/NC_021490.2?report=girevhist). This database, “Tpal_06_rabbit_review”, contained 1261 T. pallidum sequences, all the reviewed rabbit protein sequences from the UniProt Oryctolagus cuniculus proteome, UP000001811, and mass spectrometry common contaminants (https://www.thegpm.org/crap/) (Supplementary Table S1). Database search parameters were as follows: precursor tolerance 10 ppm; MS/MS tolerance 0.6 Da; enzyme specificity was set to trypsin, with a maximum of two missed cleavages allowed; ESI-TRAP instrument type; fixed modification: carbamidomethylation (C); variable modifications: acetyl of the N-terminus and oxidation (M). Scaffold (version Scaffold_5.1.2) (Proteome Software Inc., Portland, OR, USA) was used to validate MS/MS based peptide and protein identifications. Peptide identifications were accepted if they could be established at greater than 95.0% probability by the Percolator posterior error probability calculation²⁷. Similar to the previous T. pallidum whole proteome mass spectrometry-based study by Osbak et al.²⁵, protein identifications were accepted in the present study if they could be established at greater than 95.0% probability and contained at least one identified peptide. In the current study, protein probabilities were assigned by the Peptide and Protein Prophet algorithms²⁸. Percolator (as a node in Proteome Discoverer) was used to generate decoy sequences (randomized sequences from the customized T. pallidum database, as described above). Similar to the previous T. pallidum whole proteome mass spectrometry-based study by Osbak et al.²⁵, the false discovery rate (FDR) calculated by Scaffold for confident protein identification was set for less than 5% in the present study. Proteins that contained similar peptides and could not be differentiated based on MS/MS analysis alone were grouped to satisfy the principles of parsimony²⁹. Proteins sharing significant peptide evidence were grouped into clusters. Non-identical protein paralog identifications were confirmed via the detection of at least one tryptic peptide that is unique to a single paralog. For identical paralogs (full-length proteins that are identical at the amino acid sequence level e.g., TPANIC_0117 [TprC] and TPANIC_0131 [TprD]), our peptide identification pipeline could not distinguish these as separate protein identifications. The proteome of T. pallidum Nichols strain (NCBI reference sequence NC_021490, July 2021 annotation) was used to calculate proteome coverages (964 proteins from predicted protein-coding genes, 15 proteins potentially encoded by 15 genes annotated as “pseudo genes”, and three detected proteins from previous proteome annotations that are not annotated in the July 4th 2021 proteome).

Mass spectrometry parameters and data analyses: label free quantification

Relative protein abundances in the three T. pallidum samples that were prepared using our optimized method (samples 10, 11, 12) (described above and Fig. 1), were determined using label-free quantification (LFQ) based on peptide ion peak intensities as a relative quantitative measure. The LFQ analyses were performed within the FragPipe proteomics pipeline (version 17.1) using the MSFragger proteomic search engine (version 3.4) for protein database searching and peptide identification, Philosopher toolkit (version 4.1.1) for downstream post-processing of MSFragger search results (PeptideProphet and ProteinProphet), and IonQuant for LFQ with FDR-controlled match-between-run (MBR) functionality^30,31,32. The RAW spectral files for T. pallidum samples 10, 11, and 12 were converted to mzML format with ProteoWizard MS convert (http://proteowizard.sourceforge.net) and loaded into FragPipe and the workflow was configured for LFQ-MBR. The database used for searching and peptide identifications was “Tpal_06_rabbit_review” (described above). MSFragger database search parameters were as follows: precursor mass tolerance − 20 to 20 ppm; fragment mass (MS/MS) tolerance 0.6 Da; enzyme specificity was set to trypsin, with a maximum of two missed cleavages allowed; fixed modification: carbamidomethylation (C); variable modifications: acetylation of the peptide N-terminus and oxidation (M). Peptide and protein identifications were validated and filtered using PeptideProphet and ProteinProphet. Label-free quantification with FDR-controlled match-between-runs in MS1 Quant was performed with the following parameters: Ion quant selected; match between runs enabled; protein quant = MaxLFQ; min ions = 2. All other parameters were set by the LFQ-MBR workflow configuration. Identified proteins were filtered using protein probability (confidence score determined by ProteinProphet from combined evidence from the three samples) and top peptide probability (highest PeptideProphet confidence score from all peptides that map to the protein) thresholds equal to 95% or greater. The MaxLFQ intensity values from the “combined.protein” data output file for the three T. pallidum samples were ranked from highest to lowest intensities to determine the relative abundances of the treponemal proteins.

Functional classification of T. pallidum proteins

The genome wide functional annotation tool, eggNOG-mapper version 2.1.9 (http://eggnog-mapper.embl.de/)³³, was used to assign functional classification to T. pallidum proteins. All 1261 T. pallidum protein sequences from the “Tpal_06_rabbit_review” database, as described above, were submitted to eggNOG-mapper (default parameters; minimum hit e-value = 0.001, minimum hit bit-score = 60, percentage identity = 40, minimum % of query coverage = 20, minimum % of subject coverage = 20, search against database = eggNOG 5). Functional classifications were based on the “COG_category” (https://www.ncbi.nlm.nih.gov/research/cog) outputs generated for each of the submitted proteins.