Induction of NETs with enriched protease activity

NE, PR3 and CG were previously identified in NETs and together were estimated to make up ~9% of the total protein associated with the NETs [14]. To estimate proteolytic activity in PMA-induced NETs generated from healthy donor neutrophils we screened a set of internally quenched fluorescent peptides and identified a substrate that was readily cleaved by all three enzymes (Figure 1A, Figure S1). This substrate, K(Mca)-PLGKQVEY-K(Dnp), was previously used to assay a glutamic acid protease secreted from a fungus [18]. Using this probe, the proteolytic activity released from NETs derived from PMA and Micrococcal Nuclease (MNase) treated neutrophils ranged from 0.047 to 0.068 µmoles/min/mg. For individual donors this activity was approximately five-fold greater than control samples that lack PMA treatment, and twenty- to forty-fold greater than control samples that lacked MNase treatment (Figure 1B). These assays confirmed that activated neutrophils produce NETs that are enriched with proteases that can be released upon nuclease treatment.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Monitoring of proteolytic activity in neutrophil extracellular traps.

A. Identification of an internally quenched fluorescent substrate that is hydrolyzed by NE, CG and PR3. B. Extracellular proteolytic activity was analyzed from 3 donor neutrophils (Donor 1, dark grey; Donor 2, black; Donor 3, light grey) following treatment with PMA, MNase or a combination of both. Proteolytic activity was measured using (K-Amc) PLGKQVEY(K-Dnp).

https://doi.org/10.1371/journal.pone.0075141.g001

Comparison of substrate signatures of NET associated proteases from donor neutrophils

We employed MSP-MS to generate a substrate signature of proteases associated with NETs from PMA activated neutrophils and compared the substrate specificity for each donor sample (Figure 2A-C). Each donor sample contained proteases with a distinct preference for isoleucine, valine and threonine in the P1 position while arginine, glutamine and tryptophan were significantly enriched (p ≤ 0.05) at P4, P3 and P2′ positions, respectively. Furthermore, cleavage was rarely observed on the C-terminal side of amino acids with charged side chains with the exception of lysine, while proline is not well tolerated in either the S3, S1 or S1′ pockets of these proteases. When donors were directly compared, 40 shared cleavage sites were identified and each donor sample had between 7 and 11 unique cleaved bonds (Figure 2D). When the location of each cleavage site was analyzed, no hydrolysis was evident near the N-termini of the tetradecapeptides, indicating that these neutrophil derived enzymes lacked aminopeptidase specificity (Figure 2E). However, cleavage readily occurred near the C-termini, indicating the presence of carboxypeptidases or endoproteases that have little specificity at the non-prime side of the scissile bond.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 2. Determination of the proteolytic signatures in NETs.

A-C. IceLogos representing the P4 to P4′ sites for NETs isolated from three donor samples. Amino acids that are most frequently observed (above axis) and least frequently observed (below axis) are illustrated. The number of cleavage sites used to make each iceLogo are listed in the bottom right-hand corner. Residues that are highlighted in black text are significantly (p = 0.05) enriched relative to the frequency that these same amino acids are found in the peptide library (5.2 +/- 0.5%). The amino acid ‘n’ corresponds to norleucine. D. Determination of the number of cleavage sites that are common and unique to each donor sample. E. Positional frequency of all donor-derived cleavage sites within the tetradecapeptides (n=85).

https://doi.org/10.1371/journal.pone.0075141.g002

Identification of NET associated proteins

To identify the full complement of proteins embedded in the NETs, protein preparations from the same NETosis-induced neutrophils described above were subjected to proteomic analysis to evaluate the protein composition of neutrophils after PMA and MNase treatment (Table 1). Using mass spectrometry, 29 proteins were identified in PMA- and MNase-treated neutrophil samples, however only NE, alpha-enolase and Histones H2A and H3 were found to be specifically enriched relative to the control samples. Eleven proteins were reproducibly identified in all three PMA- and MNase-treated donor samples, which included NE and the inactive serine protease family member, azurocidin. Surprisingly, while our enzymatic studies indicated an enrichment of proteolytic activity in NETs from PMA- and MNase-treated neutrophils relative to the control samples, there was little or no enrichment of proteases in the same samples when analyzed by mass spectrometry-based proteomics. CG was found in one donor sample while PR3, NSP4 and MMP-8 were never observed. Two peptides corresponding to MMP-9 were observed in a MNase only treated sample.

MSP-MS profiling of NE and comparison to NETs

Based on our proteomic data, the majority of proteolytic activity in NETs was predicted to be derived from NE. To test this prediction, the substrate specificity of purified NE was profiled using both MSP-MS and the more established positional scanning synthetic combinatorial library (PS-SCL) assay [19]. This enzyme was isolated from human neutrophils and was found to be free of other neutrophil or serum proteases using mass spectrometry based proteomic analysis (data not shown). In the PS-SCL assay the P1 site of NE had a distinct preference for valine, alanine, threonine and isoleucine. In addition, proline was preferred in P2, glutamine, glutamic acid and methionine in P3 and norleucine in P4 (Figure 3A). Using the MSP-MS assay, NE had a preference for isoleucine, valine and threonine in the P1 position and had low tolerance for asparagine, lysine and most other amino acids with charged side chains. In addition arginine, glutamine, norleucine and tryptophan were significantly enriched (p ≤ 0.05) at the P4, P3, P1’ and P2′ positions, respectively. A Pearson correlation was performed on the NE profiles from the MSP-MS and PS-SCL methods which showed that each non-prime subsite had strong positive correlation with a score of ≥0.4 on a scale from -1.0 to 1.0 (Table 2). This validated the MSP-MS assay as a method for profiling the substrate specificity of NE and other serine neutrophil serine proteases.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 3. Substrate specificity profiling of human Neutrophil Elastase A.

Positional scanning of the P4 to P1 subsites of NE using the PS-SCL assay. B An iceLogo illustrating amino acids that are most frequently (above axis) and least frequently (below axis) observed in the P4 to P4′ sites of NE. Residues that are highlighted in black are significantly (p = 0.05) enriched or de-enriched in the subsites relative to the frequency that these same amino acids are found in the peptide library (5.2 +/- 0.5%). C. A representative “donor signature” consisting of 40 cleavage sites that are common to the three donors. D. A pie chart representing the 40 cleavage sites that are common to the donor samples. 33 of these sites are also hydrolyzed by NE.

https://doi.org/10.1371/journal.pone.0075141.g003

To generate a representative “donor signature”, the 40 cleavages observed in all three donor samples were aggregated into a single motif. An advantage of using the MSP-MS assay over the PS-SCL assay for profiling biological samples is that specific cleavage of peptide substrates can be directly linked to a protease. The cleavage sites that were common to all three donors were directly compared with the substrate profile of purified NE (Figure 3C) and correlated strongly between P4 and P3′ (Table 2). In addition we determined that 33 of the cleaved bonds that occur in donor samples could also be hydrolyzed by NE (Figure 3C) which provided evidence that the major proteolytic activity in NETs is derived from NE. However, seven of the cleavage sites could not be directly linked to NE activity, therefore additional proteases were active on NETs.

MSP-MS profiling of CG, PR3 and NSP4 and comparison to NETs

CG, PR3 and the newly identified neutrophil serine proteases NSP4 have been shown to be released from PMA activated neutrophils [14,15], therefore these enzymes were profiled using MSP-MS to identify the source of the unknown proteolytic activity in NETs. In the P1 positions, PR3 had approximately equal preference for alanine, valine, threonine and isoleucine, CG favored phenylalanine over tyrosine and lysine, while NSP4 had a strict preference for arginine (Figure 4A-C). Outside of the P1 subsite, PR3 showed selectivity for aspartic acid and asparagine at P2, and norleucine, leucine and glycine at P4, P3 and P2′, respectively while CG had a preference for norleucine at P2. In contrast to the other neutrophil serine proteases, NSP4 specificity was so selective for P1 arginine residues that only 19 bonds were cleaved in the library after 20 hours incubation. In order to further characterize non-prime side specificity, each enzyme was assayed using the PS-SCL method and the P4 to P1 site preferences were identified (Figure S2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 4. Determination of proteolytic signature of PR3, CG and NSP4.

A-C. IceLogos representing the P4 to P4′ sites for PR3, CG and NSP4. Amino acids that are most frequently observed (above axis) and least frequently observed (below axis) are illustrated. The number of cleavage sites used to make each iceLogo are listed in the bottom right-hand corner. Residues that are highlighted in black text are significantly (p ≤ 0.05) enriched relative to the frequency that these same amino acids are found in the peptide library (5.2 +/- 0.5%). The amino acid ‘n’ corresponds to norleucine. D. A 4-way venn diagram illustrating the number of unique and overlapping peptide bonds that are cleaved by each neutrophil serine protease. E. A pie chart representing the 40 cleavage sites that are common to the donor samples. ‘Shared’ corresponds to cleavage sites that are derived from more than one neutrophil serine protease.

https://doi.org/10.1371/journal.pone.0075141.g004

Using the MSP-MS datasets, a Venn diagram was generated to highlight the unique and shared cleavage sites for each of the neutrophil serine proteases. Interestingly, for a family of enzymes that share high sequence identity, they have each evolved to cleave a defined set of unique peptide bonds. When compared to the unknown cleavage sites in NETs, we determined that 5 of the 7 sites were matched to CG while one site was matched to PR3 activity. With the knowledge that CG and PR3 are active in NETs we were now only able to directly assign 18 of the cleavage sites to NE as many of the sites can be cleaved by more than one neutrophil serine protease (Figure 4E). A single unidentified cleavage site occurred between valine and isoleucine in a peptide containing the sequence PHWQRVIFFRLNTP. Although not observed in the individual enzyme assays, cleavage at this P1 valine site is likely to be the result of NE, CG or both.

Immunodepletion of NE from NETs

In order to confirm that NE was the major proteolytic activity in NETs, we selectively removed the enzyme by immunodepletion, and assayed the remaining proteases in the mixture. On this occasion, we were able to increase the total amount of NE-depleted donor protein in the assay by 15-fold, which resulted in only a 1.7 to 2.3-fold increase in the number of cleavage bonds identified in each donor sample (Figure S3). The NE-depleted donor samples shared 76 cleavage sites and 12 to 16 unique cleaved bonds were identified for each donor (Figure 5). The substrate signature of the shared cleavage sites showed a preference for phenylalanine, arginine and lysine in the P1 position, while arginine and norleucine were most often found in P3 and P2 sites, respectively. When compared to the individual serine protease datasets, we determined that 36 of the 76 cleavage sites were attributable to CG activity. In addition, PR3 and NSP4 accounted for 7 and 1 cleaved bonds, respectively, while the protease(s) responsible for 15 cleavage sites could not be determined. The successful depletion of NE was evident by the presence of a single NE cleaved bond, however, it is possible that cleavage at this site is due to an unidentified low abundance protease with overlapping specificity. In addition, hydrolysis of the PHWQRVIFFRLNTP peptide at the V-I site was not evident in the NE-depleted samples indicating that the cleavage site was the product of NE activity.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 5. Determination of the proteolytic signatures in NE-depleted NETs.

A. After immunodepletion of NE from donor NET samples, the MSP-MS assay was performed at 15-fold higher total protein concentration. A total of 153 cleaved bonds were identified in the 3 donor samples and 76 cleaved bonds were common to each donor. B. An iceLogo illustrating the proteolytic signature of cleavage sites that are common to each NE-depleted donor. C. A pie chart representing the 76 cleavage sites that are common to the donor samples. ‘Shared’ corresponds to cleavage sites that are derived from more than one neutrophil serine protease.

https://doi.org/10.1371/journal.pone.0075141.g005

Taken together, we have utilized a novel substrate based profiling method to determine that NE and to a lesser extent CG and PR3 are active in neutrophil derived NETs. In addition, by selectively removing NE we have been able to perform an in-depth characterization of less abundant proteolytic activity and determined that NSP4 and at least one other protease are also active in NETs.

Ethics statement

All data were analyzed anonymously and all aspects of this study were conducted according to the principles expressed in the Declaration of Helsinki. Human subject research approval for this study was received from the Western Institutional Review Board (study No: 1084207). All participants provided written informed consent prior to the start of the study.

Isolation of neutrophils from healthy donors

Human neutrophil cells were isolated by a two-step purification protocol using Red Blood Cell (RBC) sedimentation followed by removal of monocytes using Ficoll density gradient centrifugation. Briefly, 50 ml of fresh human whole blood was collected in a collection tube containing heparin. Blood was mixed with HetaSep (STEM cell, Cat^# 07906) at the ratio (5:1) to precipitate RBCs and platelets. Supernatants with enriched leukocytes and monocytes were layered on top of Ficoll-PAQUE PLUS (GE Healthcare, Cat^# 17-1440-03). After Ficoll gradient centrifugation, neutrophils were separated from monocytes in the supernatant and pelleted. Contaminating RBCs were further removed by repeated cell lysis using RBC lysis buffer (Miltenyi Biotec, Cat^# 130-094-183). By this method, 50-100 million neutrophils were isolated to greater than 98% purity, as confirmed by flow cytometry using CD66b antibodies (BD Pharmamingen, Cat# 555724).

NETosis induction and NET preparation

Purified neutrophils were washed five times to remove plasma proteins, then seeded at a density of 1.7 x 10⁶ cells/ml in RPMI 1640 media supplemented with glutamine in a 10 mm culture plate. NETosis was induced in vitro by stimulating neutrophils with 50 nM phorbol-12-myristate-13-acetate (PMA, Sigma P8139) at 37°C in a 5% CO₂ incubator. After induction for 3 hours, media was removed and plates were washed gently with warm media for three times. The induction time was selected for maximum protein release from NETs after PMA treatment. NETs were then digested by Micrococcal Nuclease (MNase) (20 U/ml) (Thermo Scientific, Cat^# 88216) for 10-40 minutes and released into media. The supernatant was subsequently centrifuged to remove cells and cellular debris. Supernatants from PMA-untreated neutrophils and PMA-treated neutrophils but without MNase digestion were prepared as negative controls. A fraction of each sample was treated with protease inhibitor cocktail to preserve the sample for proteomic analysis, whereas the remainder of the sample for protease activity screening was not treated with protease inhibitors. Progression of NETosis was monitored by measuring cell-free DNA using Sytox Orange (Life Technologies, S11368). DNA was quantified by relative fluorescence measurement with a SpectraMax M2 fluorometer (Molecular Devices) at a filter setting of 544 nm (ex)/ 590 nm (em) and calibrated by standard curve with DNA standard of known concentration. NETosis was also visually examined by confocal immunofluorescent microscopy. Neutrophils (5 x 10⁵ cells/ml) were seeded on poly-L-lysine coated cover slips and treated with or without 50 nM PMA. At different time points post-NETosis induction, cells were fixed with 4% paraformaldehyde, then permeablized and blocked with 10% FBS in phosphate buffered saline (PBS) with 0.05% Triton X-100. For histone staining, coverslips were incubated with a mouse anti-human core histone antibody (Millipore, Anti-histone Clone H11-4, MAB3422) followed a Tetramethyl Rhodamine Isothiocyanate (TRITC)-conjugated secondary antibody (Invitrogen Cat^# T2762). DNA was counterstained with Hoechst 33342 (AnaSpec Inc, Cat^# 83218). Coverslips were mounted onto glass slides using Prolong Gold mounting media (Invitrogen Cat^# P36930) before acquisition.

Determination of neutrophil serine protease activity

Human neutrophil serine proteases were purchased from Athens Research & Technology. A set of internally quenched fluorescent substrates (40 µM each) were assayed with 20 nM of NE (Cat^# 16-14-051200), CG (Cat^# 16-14-030107) and PR3 (Cat^# 16-14-161820) in Dulbecco’s-PBS containing 0.01% Tween-20 at room temperature. Triplicate assays were performed in 96-well plates in a Spectra Max Gemini EM (Molecular Devices) using a λ_ex 328nm and λ_em 393 nm. Initial velocities (relative fluorescent units/s) were calculated using SoftMax Pro software (Molecular Devices). NET samples were prepared as described above from three donors with the following combinations: +PMA/+MNase, +PMA/-MNase, and –PMA/+MNase treatment. 2 µg/ml of each NET sample was assayed in triplicate with 40 µM of (K-Mca) PLGKQVEY(K-Dnp). Initial velocities were converted to µmole/min/mg.

Peptide cleavage site identification by multiplex substrate profiling-mass spectrometry

Recombinant NSP4 was expressed and isolated from HEK 293 cells as described previously [15]. NE (1 nM), CG (5 nM), PR3 (2 nM) and NSP4 (25nM) were profiled using the MSP-MS assay in Dulbecco’s-PBS as described previously [17]. In addition, proteolytic activities in three PMA-induced and MNase-treated donor NET samples were determined using the MSP-MS assay. Control samples lacked PMA or MNase treatment and consisted of an equal mixture of total protein from each donor. All assays contained 0.4 µg/mL of donor protein and 500 nM of each peptide in a total reaction volume of 900 µl. Aliquots were removed after 15, 60, 240, and 1200 minutes and quenched with concentrated formic acid to a final pH of 2.5. Samples were desalted and analyzed by LC-MSMS peptide sequencing.

For LC-MS/MS, an LTQ-FT mass spectrometer (Thermo) equipped with a 10,000 psi system nanoACUITY (Waters) UPLC instrument was used for reversed phase chromatography with a C18 column (BEH130, 1.7 µm bead size, 100 µm x 100 mm). The LC was operated at 600 nL/min flow rate, and peptides were separated using a linear gradient over 42 min from 2% B to 30% B, with solvent A: 0.1% formic acid in water and solvent B: 0.1% formic acid in 70% acetonitrile. Survey scans were recorded over 350-1800 m/z range, and MS/MS was performed with CID fragmentation on the six most intense precursor ions. Mass spectrometry peak lists were generated using in-house software called PAVA, and data were searched using Protein Prospector software v. 5.10.0 [38]. Data was searched against a database containing the sequences of the 124 14-mer synthetic peptides, concatenated with 4 different copies of randomized sequences for the same 124 entries to create a final database of 620 sequences for estimation of false discovery rate [39]. For database searching, peptide sequences were matched with no enzyme specificity requirement, and variable modifications including oxidation of Trp, Pro and Phe, and N-terminal pyroGlu from Gln. Protein Prospector score thresholds were selected to be minimum protein score of 20, minimum peptide score of 15, and maximum expectation values of 0.1 for “protein” and 0.05 for peptide matches, and resulted in a peptide false discovery rate of 0.2%. Cleavage site data was extracted from Protein Prospector using an in house script called “MSP extractor” software. The earliest time interval that ≥2.5% (n=41) of all possible bonds in the library (n = 1612) were cleaved was chosen to compare enzymes specificity. NE, PR3 and CG reached this value after 240, 1200 and 60 minutes incubation, respectively while NSP4 cleaved only 1.2% (n=19) of peptide bonds after 1200 minutes. For comparison of substrate specificity, an iceLogo software was used to generate substrate specificity logos for amino acids at ±4 positions adjacent to the identified cleavage sites [40].

Protein identification of NETs by mass spectrometry

Protein identification in NET-induced samples was performed using peptide sequencing by mass spectrometry as previously reported [17]. The +PMA/+MNase treated samples were assayed individually for each donor, while the +PMA/-MNase, and -PMA/+MNase control samples were prepared as pooled mass-matched samples from the three donors. NET protein concentrations ranged from 25-40 µg/ml in PBS, therefore a slightly modified in solution trypsin digestion protocol was applied as follows. Samples were brought to a standardized concentration of 30 µg/ml with 100 mM ammonium bicarbonate buffer (~20 µg in a total volume of 700 µl) to which solid urea was added to 4M final. Sample was reduced with 10 mM DTT incubation for 10 min at 56 °C, then alkylated with 12 mM iodoacetamide (45 min, dark, 21 °C), and then quenched with 5 mM additional DTT. The final volume was adjusted to 1.4 ml with additional 100 mM ammonium bicarbonate, bringing urea concentration to 2M. Trypsin (sequencing grade, Promega) was added at 1:50 trypsin: total protein for digestion overnight at 37 °C. The sample was then acidified with formic acid to pH 2-3 and desalted using C18 OMIX tips (Varian). Each sample was assayed with two technical replicate LC-MS/MS analyses using an LTQ-Orbitrap (Thermo) mass spectrometer operated under identical separation and analysis conditions as the LTQ-FT system described above.

Database searches were performed against the H. sapiens UniProt database (downloaded March 21, 2012), containing 62,611 entries. For estimation of false discovery rate, this database was concatenated with a fully randomized set of sequence entries [36]. Data were searched with mass tolerances of 20 ppm for parent and 0.8 Da for fragment ions. Peptide sequences were matched as tryptic peptides with no missed cleavages, and carbamidomethylated cysteines as a fixed modification. Variable modifications included oxidation of Met, N-terminal pyroGlu from Gln, loss of Met and N-terminal acetylation. Protein Prospector score parameters were: minimum protein score of 22, minimum peptide score of 15, and maximum expectation values of 0.01 for protein and 0.001 for peptide matches, resulting in a protein false discovery rate of 1.1%. Protein identification results are reported with unique peptide count, peptide count as an approximation of protein abundance, percent sequence coverage and an expectation value for the probability of the protein identification [38,39]. Proteins were required to have been identified with at least two unique peptides in one of the three conditions tested (+PMA/+MNase, +PMA/-MNase, and -PMA/+MNase) to be reported.

Proteolytic activity using combinatorial fluorogenic substrate libraries

Non-prime side sequence specificity, as the N-terminal sequence relative to the scissile bond is termed in protease nomenclature, was assayed for proteases using combinatorial fluorogenic substrate libraries [19]. This fluorescent peptide library is amenable for detailed profiling of purified serine proteases and distinguishes between subsite preferences in closely related enzymes. Human NE (50 nM), CG (100 nM), PR3 (50 nM) and NSP4 (100 nM) were assayed with this fluorogenic library in Dulbecco’s-PBS containing 0.01% Tween-20. Amino acid preferences at each position can be determined by direct comparison of activity, in units of picomolar of fluorophore released per second.

Depletion of elastase from NET supernatants

MNase preparations of NET samples were secondarily digested with DNase (100U/ml) for 10 minutes at 37° to fully release NET-associated proteins. NET supernatants were then mixed with Pierce G/A magnetic beads (Thermo Scientific, Cat^# 88802) coated with elastase antibodies (Sigma, Cat^# PAI-74132) at 4 °C for 1 hour. After incubation, the supernatants were separated from beads on a DYNAL-magnet bead separation rack (Invitrogen Cat^# 123-21D).