Article pubs.acs.org/ac

Recombinant Nepenthesin II for Hydrogen/Deuterium Exchange Mass Spectrometry Menglin Yang,† Morgan Hoeppner,† Martial Rey,† Alan Kadek,‡,§ Petr Man,‡,§ and David C. Schriemer*,† †

Department of Biochemistry & Molecular Biology, University of Calgary, Calgary, Alberta, Canada T2N 1N4 Institute of Microbiology, Academy of Sciences of the Czech Republic, 117 20 Prague, Czech Republic § Department of Biochemistry, Faculty of Science, Charles University in Prague, 116 36 Prague, Czech Republic ‡

ABSTRACT: The pitcher secretions of the Nepenthes genus of carnivorous plants contain a proteolytic activity that is very useful for hydrogen/deuterium exchange mass spectrometry (HX-MS). Our efforts to reconstitute pitcher fluid activity using recombinant nepenthesin I (one of two known aspartic proteases in the fluid) revealed a partial cleavage profile and reduced enzymatic stability in certain HX-MS applications. We produced and characterized recombinant nepenthesin II to determine if it complemented nepenthesin I in HX-MS applications. Nepenthesin II shares many properties with nepenthesin I, such as fast digestion at reduced temperature and pH, and broad cleavage specificity, but in addition, it cleaves Cterminal to tryptophan. Neither enzyme reproduces the C-terminal proline cleavage we observed in the natural extract. Nepenthesin II is considerably more resistant to chemical denaturants and reducing agents than nepenthesin I, and it possesses a stability profile that is similar to that of pepsin. Higher stability combined with the slightly broader cleavage specificity makes nepenthesin II a useful alternative to pepsin and a more complete replacement for pitcher fluid in HX-MS applications.

H

involve plant aspartic proteases called nepenthesins.12,13 Nepenthesins belong to MEROPS subfamily A1 of pepsinlike enzymes, which contain a typical Asp-Ser/Thr-Asp catalytic triad in the active site.14 Two isozymes of nepenthesin have been discovered and purified from Nepenthes gracilis plants. Nepenthesin I (NepI) and Nepenthesin II (NepII) had been partially characterized, and they demonstrate at least some classic aspartic protease features.13,15,16 They are expressed as zymogens that autoactivate in acidic pH by cleavage of a propeptide. The enzymes appear sensitive to inhibition by pepstatin A and are quite stable. Because the fluid is an inconvenient source of enzymatic activity for HX-MS, we have begun to reconstitute fluid properties with recombinant enzymes. We previously expressed and characterized recombinant NepI (rNepI) from Escherichia coli.17,18 Our studies showed some differences between rNepI and the native form of NepI, as well as the fluid. While the selectivity profile approached that of the fluid, there were notable exceptions. We saw little evidence of enzymatic processing C-terminal to W and P. We also observed that the activity and stability of rNepI were lower than those of the native purified form, although the stability under reducing and denaturing conditions could be improved through immobilization.18 Native NepI is glycosylated, which likely explains its higher stability and perhaps activity, but the “missing” cleavage properties argue for other enzymes.

ydrogen/deuterium exchange mass spectrometry (HXMS) provides conformational and structural data about proteins and complexes that can complement those generated by conventional structural biology methods.1,2 The digestionbased, or “bottom-up”, approach offers a powerful way to conduct HX-MS studies. Compared to “top-down” methods, it requires low protein concentrations and low sample amounts and supports flexible buffer formulations.2 These attributes have led to new applications in the area of membrane protein analysis, complex multiprotein reconstructions and antibody characterization.3−6 However, efficient digestion under HX conditions has proven difficult to sustain as sample complexity increases.7 This limitation partly arises from the suboptimal kinetics of pepsin digestion for low-temperature and short digestion work, which has prompted a search for alternative proteases.8 In a previous study, we demonstrated an effective alternative in an extract of the Nepenthes genus of carnivorous plants, otherwise known as pitcher plants or monkey cups.9 We showed that the proteolytic properties of the concentrated pitcher fluid are compatible with HX-MS conditions (i.e., low pH and temperature) and that the fluid possesses several advantages over pepsin. Activity was higher by >1000-fold, and there was little evidence of protease autodigestion. Additional and surprising new cleavage sites were also observed (i.e., Cterminal to H, K, R, and P). Hooker observed in the 1870s that pitcher secretions contained proteolytic activity, but for a considerable period of time, the source of the activity was thought to be microbial in nature.10,11 Only recently has proteolytic activity been shown to © 2015 American Chemical Society

Received: March 2, 2015 Accepted: May 20, 2015 Published: May 20, 2015 6681

DOI: 10.1021/acs.analchem.5b00831 Anal. Chem. 2015, 87, 6681−6687

Article

Analytical Chemistry

and 300 mM β-mercaptoethanol (pH 11)] and incubated at 37 °C for 1 h (with shaking) followed by dialysis against 5 volumes of 50 mM Tris (pH 11) twice in two 1 h steps. The denatured protein was then dialyzed against 50 mM Tris-HCl (pH 7.4) overnight at 4 °C prior to dialysis against MOPS buffer [50 mM 3-(N-morpholino)propanesulfonic acid and 300 mM NaCl (pH 7)] for 24 h at 4 °C. Precipitated material was removed by centrifugation. The supernatant was adjusted to pH 2.5 with glycine-HCl (final concentration of 100 mM) prior to incubation overnight at 4 °C. After centrifugation, the supernatant was concentrated in a 10 kDa molecular mass cutoff filter (Millipore) and washed 1000 times with 100 mM glycine-HCl (pH 2.5) prior to purity and size analysis by sodium dodecyl sulfate−polyacrylamide gel electrophoresis (SDS−PAGE), matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF), and size-exclusion chromatography. The dried-droplet method was used for MALDI-TOF. Aliquots of recombinant enzymes were mixed in a 1:1 ratio with a saturated solution of sinapinic acid (Sigma-Aldrich) dissolved in 50% acetonitrile and 1% formic acid and deposited on a MALDI plate prior to analysis on an AB Sciex MALDI TOF/TOF 5800 mass spectrometer, operated in linear mode. Data were processed in mMass.19 Size-exclusion chromatography was performed with a Superdex S-75 column (Amersham Biosciences) on a BioLogic DuoFlow Chromatography system (Bio-Rad), to confirm the size and aggregation status of the enzyme. All samples were eluted with 20 mM glycine-HCl (pH 2.5) at a flow rate of 0.5 mL/min. Disulfide bond formation in the folded mature enzyme, and its role in enzyme activity, was assessed by incubation of rNepI or rNepII in 150 mM tris(2carboxyethyl)phosphine (TCEP) for 2 h at 37 °C and pH 2.5, prior to a 5 min digestion of 4 μg of bovine serum albumin (BSA) at pH 2.5 and 37 °C, and followed by SDS−PAGE. Activity and Stability Determinations. Proteolytic activity was measured using a modified version of the hemoglobin activity assay first described by Anson et al. in 1938.20 The assay applied enzyme (400 ng of porcine pepsin, 1 μg of rNepI, or 1 μg of rNepII) to 1.25 mg of equine hemoglobin in 100 mM glycine-HCl (pH 2.5) in a final volume of 100 μL. The digestion was allowed to proceed for 30 min at 37 °C and then stopped via the addition of 100 μL of 10% trichloroacetic acid (TCA). The precipitate was removed by centrifugation, and the supernatant was used to measure the absorbance of TCA-soluble peptides at 280 nm in a GE Nanovue plus spectrophotometer. Data points are the mean of three samples. To determine the effect of pH on activity, the assay was modified using glycine-HCl (100 mM) for pH 1.6−3, ammonium acetate (100 mM) for pH 4−5, citrate (100 mM) for pH 6−7, and Tris-HCl (100 mM) for pH 8. To determine the temperature profile, the amount of enzyme in the assay was decreased (100 ng of rNepI and rNepII and 50 ng of porcine pepsin) and the digestion time reduced to 15 min. To monitor enzyme stability across pH, a 1 mg/mL solution of each enzyme containing 100 mM buffer ranging from pH 1.6 to 8 was incubated at 37 °C. On days 7 and 30, the standard assay was used to quantify the remaining activity at pH 2.5. The standard assay was also used to determine the relative activity in the presence of 1−6 M denaturing agent, urea or guanidine hydrochloride (GndHCl). High levels of reducing agent (TCEP) interfered with the assay; therefore, to test the activity in the presence of a

Native NepII does not show evidence of glycosylation, and it may contribute overlapping properties. The degree of sequence homology between NepI and NepII is quite high (∼67%), but NepII may not be an exact isotype. In this study, we describe the production of recombinant NepII (rNepII) and compare its enzymatic properties to those of rNepI and porcine pepsin for use in bottom-up HX-MS. We show that rNepII provides substrate cleavage selectivity closer to that of the extract and is significantly more resistant to denaturants and reducing agents than rNepI.

■

EXPERIMENTAL SECTION Preparation of the NepII Expression Vector. The gene corresponding to pro-NepII (residues 25−438, UniProt entry Q766C3) from N. gracilis was amplified from a pUC57 vector encoding the full gene (Genscript) and cloned into a custom vector (e3884, kind gift from A. Schryvers, University of Calgary) with a forward primer (5′-ACT GAC GGATCC ACA TCG AGA GGC ACA TTA TTG CAT C-3′) containing a BamHI restriction site (underlined) and a reverse primer (5′CTG TCT AAGCTT TTA GCT CGC ACC GCA CTG-3′) containing a HindIII restriction site (underlined). The custom vector allows for gene insertion in frame and downstream of an N-terminal His tag and an N-terminal maltose binding protein (MBP) tag. The construct was verified by restriction digestion analysis and DNA sequencing. The gene corresponding to pro-NepI (residues 25−437, UniProt entry Q766C2), also from N. gracilis, was amplified from a pET21d vector encoding the full gene (kind gift from H. Inoue, Tokyo University of Pharmacy and Life Science, Tokyo, Japan) and cloned into e3884 with a forward primer (5′-ACT GAC AGATCT ACG TCA AGA ACA GCT CTC AAT C-3′) containing a BglII restriction site (underlined) and a reverse primer (5′-CTG TCT AAGCTT TTA CGA CGC ACC ACA TTG AG-3′) containing a HindIII restriction site (underlined). The construct was also verified by restriction digestion analysis and DNA sequencing. Protein Expression, Purification, and Analysis. The methods for protein expression and purification are similar to those described previously for rNepI.17 Briefly, E. coli BL21(DE3) cells were transformed with the pro-NepI or proNepII expression vector. A starter culture in 2XYT medium with ampicillin (100 μg/mL) was generated by incubation for 16 h at 37 °C. Larger scale cultures were grown from the starter until an OD600 of ∼0.8 was reached and then induced with 0.8 mM isopropyl β-D-thiogalactopyranoside (IPTG) followed by overnight incubation at 30 °C, with shaking. Cultures were harvested by centrifugation and lysed with a homogenizer after resuspension in buffer A [50 mM Tris and 0.5 M NaCl (pH 8)], supplemented with a cocktail of protease inhibitors (Roche). Inclusion bodies and debris were recovered by centrifugation and then resuspended in buffer A for centrifugation through a sucrose cushion [17400 rcf, 15 min, 4 °C, containing 25% (w/v) sucrose, 50 mM Tris-HCl, 1 mM ethylenediaminetetraacetic acid (EDTA) (pH 7.4)]. The pellet containing crude inclusion bodies was resuspended in buffer B [50 mM Tris-HCl, 10 mM NaCl, 1 mM β-mercaptoethanol, and 0.5% (v/v) Triton X-100 (pH 7.4)] and incubated for 30 min at 37 °C prior to centrifugation (11400 rcf, 15 min, 4 °C). The pelleted inclusion bodies were washed twice with Tritonfree buffer B. The inclusion bodies were resuspended in buffer C [50 mM Tris, 8 M urea, 1 mM EDTA, 1 mM glycine, 500 mM NaCl, 6682

DOI: 10.1021/acs.analchem.5b00831 Anal. Chem. 2015, 87, 6681−6687

Article

Analytical Chemistry

Figure 1. Recombinant production of rNepII. (A) SDS−PAGE of solubilized protein from purified E. coli inclusion bodies (lane 1), refolded rNepII after the final dialysis step, with the intact fusion protein indicated (MBP-pro-Nep II, lane 2), and mature enzyme after activation for 24 h at low pH (lane 3). (B) MALDI-TOF MS analysis of the mature enzyme, showing the [M + H]+ and [M + 2H]2+ peaks. (C) Size-exclusion chromatography profile of rNepII, eluted with 20 mM glycine-HCl (pH 2.5), suggesting monomeric mature enzyme. Sizing standards are dextran blue (2000 kDa, I), equine hemoglobin (64 kDa, II), and equine myoglobin (17 kDa, III). (D) SDS−PAGE of native and reduced protein, challenged with substrate. Disulfide bond formation upon refolding is shown by comparing 4 μg of rNepII under reducing conditions (lane 1; A = rNepII) with nonreducing conditions (lane 2; B = rNepII). rNepII is applied under nonreducing conditions with 4 μg of BSA showing complete digestion (lane 3; B = rNepII). Four micrograms of BSA is shown as a control (lane 4; A = BSA). Digestion was conducted for 5 min at 37 °C prior to loading.

loaded onto a C18 trap/column combination [C18 trap, 75 μm (inside diameter) × 2 cm, 3 μm particle diameter; C18 column, 75 μm (inside diameter) × 15 cm, 2 μm particle diameter] using mobile phase A and eluted using a 5 to 50% linear gradient of mobile phase B (mobile phase A consisted of 3% acetonitrile and 0.1% FA and mobile phase B 97% acetonitrile and 0.1% FA). The flow rate was 350 nL/min. Peptides were fragmented using collision-induced dissociation in a top-10 data-dependent acquisition mode. The MS/MS data were searched against a database of the five proteins supplemented with pepsin, rNepI and rNepII, using Mascot 2.3 with “no enzyme” specificity. To quantify cleavage specificity, the results were normalized according to the following equation:21

denaturing and reducing agent, digests were prepurified by HPLC and then analyzed by LC−MS/MS. Here, 1 μg of rNepI, 1 μg of rNepII, or 0.4 μg of porcine pepsin was used to digest 1 μg of human X-ray repair cross-complementing protein 4 (XRCC4) in 100 mM glycine-HCl (pH 2.5), alone or with 3 M GndHCl and 100 mM TCEP. The digestion was conducted for 3 min at 10 °C and quenched by increasing the pH to 8. The digest was injected into an Agilent 1100 series HPLC system (50 mm × 2.1 mm, Aeris Widepore 3.6 μm XB-C18 column, Phenomenex Inc.), and the peptides were recovered by step elution. The collected peptides were analyzed by datadependent LC−MS/MS (Agilent 1260 LC, and Infinity chip cube on an Agilent 6550 QTOF instrument) and searched on a custom database using Mascot version 2.3 (Matrix Science). All peptide sequences were verified manually and mapped against the protein sequence using MSTools.20 Digest Mapping and Enzyme Specificity. For determining cleavage preferences, 20 pmol of protein substrate was mixed with 100 ng of rNepII, rNepI, or Nepenthes fluid (estimated as previously described).9 Digestion was allowed to proceed for 5 min at 37 °C and quenched when the pH was increased to 8. The proteins used as substrates were XRCC4, XRCC4-like factor (XLF), polynucleotide kinase (PNK), BRCA1 C-terminal domain (BRCT), and myoglobin. Details regarding the recombinant production of the substrates can be found elsewhere.9 One picomole of digested substrate was injected into an Easy LC-1000 nanoLC column coupled to an Orbitrap Velos column (Thermo Scientific). Peptides were

Si =

Pi Pi′

where Pi represents the fraction of observed peptides with cleavage at amino acid i in the total pool of observed peptides and Pi′ represents the fraction of amino acid i in the total sequence set. In other words, in a sufficiently large set of substrates, an Si of >1 suggests a cleavage frequency greater than random. Protein sequence maps of APLF (Aprataxin and PNKP-Like Factor) were generated using similar amounts of immobilized rNepII and immobilized pepsin, and the digestion times were varied to produce the optimal sequence coverage using each enzyme. We define optimal coverage as the highest percent sequence coverage having the greatest degree of amino acid 6683

DOI: 10.1021/acs.analchem.5b00831 Anal. Chem. 2015, 87, 6681−6687

Article

Analytical Chemistry

Figure 2. Effect of (A) temperature and (B) pH on activity of rNepI (○), rNepII (●), and pepsin (△) and long-term stability of these enzymes after incubation for (C) 7 and (D) 30 days at the pH values noted, with activity measured at pH 2.5 and 37 °C after storage: black bars, rNepII; gray bars, rNepI; white bars, porcine pepsin. All data shown as the mean of three samples ± the standard deviation.

redundancy.22 We immobilized rNepII on POROS AL resin using standard Schiff base chemistry.23 Immobilized pepsin was commercially available. Immobilized rNepII (∼15 μg/0.6 μL of POROS resin) was used to digest APLF in 90 s. Immobilized pepsin (∼13 μg/5 μL of agarose gel) was used to digest APLF in 5 min. Chemicals. Water, acetonitrile (Burdick and Jackson HPLC grade), NaCl, and Tris were from VWR. Ampicillin was from Invitrogen, and both TCEP and immobilized pepsin were from Thermo Scientific. All other chemicals, reagents, and proteins not recombinantly produced were from Sigma-Aldrich.

■

RESULTS AND DISCUSSION

Expression and Purification of NepII. Proteolytic enzymes can be difficult to produce in soluble form from E. coli. To promote soluble production, we tried expressing the enzyme (pro-NepII, residues 25−438) using a construct containing the sequence of a large easily folded tag (MBP) on the N-terminus. Unfortunately, this was not sufficient to promote solubilization, and the enzyme went to inclusion bodies, which required a denaturing and refolding procedure. The presence of a soluble MBP tag can positively influence refolding, so the same construct was used. The mature enzyme (residues 80−438) was generated from the refolded protein after autoactivation at low pH and temperature, as shown by SDS−PAGE (Figure 1A). Tags were not needed for purification. Autoactivation of the enzyme removed the propeptide and the upstream tags, and digestion eliminated any contaminating protein present in the inclusion bodies (Figure 1A). The MALDI-measured mass was 37506 Da (Figure 1B), which agrees with the expected mass of 37497 Da for the mature enzyme, within error. (We assume all six cysteines form disulfide bonds.) In addition, gel-filtration chromatography indicates the enzyme is monomeric (Figure 1C). These data support a clean isolation of the mature

Figure 3. Effects of denaturing agents on the activity of enzymes: (A) urea and (B) guanidine hydrochloride, GndHCl. rNepI (○), rNepII (●), and pepsin (△). Activity measured in the standard assay, at 37 °C. Data shown as the mean of three samples ± the standard deviation.

enzyme, and as Figure 1D shows, chemical reduction of the mature enzyme with TCEP causes a strong shift in electrophoretic mobility, consistent with the presence of disulfide 6684

DOI: 10.1021/acs.analchem.5b00831 Anal. Chem. 2015, 87, 6681−6687

Article

Analytical Chemistry

respectively, at 37 °C, for equivalent amounts of enzyme. Our previous study estimated the activity of Nepenthes fluid to be ∼1400 times that of pepsin under HX-MS conditions (10 °C), where we used peptide ion chromatograms for quantitation.9 As the modified Anson assay was performed at 37 °C, we explored the effect of temperature on digestion efficiency more thoroughly. Temperature does appear to strongly influence the activity of both rNepI and rNepII (Figure 2A), but it parallels the trend observed for pepsin, at least between 4 and 50 °C. The lower temperatures used in the HX-MS experiments are clearly not sufficient to explain the difference in activity levels between the recombinant enzymes and the fluid. It could reflect a bias in the hemoglobin assay toward reporting partial digestion, but it can also indicate that the proteolytic activity of the fluid involves additional enzymes. Above 50 °C, we observed a large difference in activity between the enzymes: rNepII and porcine pepsin remained active up to ∼80 °C, but rNepI was active only up to ∼60 °C. (The thermal stability of rNepI produced as described is slightly lower than that from our earlier report on rNepI, but we note that the assays are not identical.17) The nepenthesins isolated from fluid show temperature−activity profiles different from those of the recombinant enzymes presented here. Glycosylated native NepI remains active up to ∼80 °C.13 Glycosylation is a well-known mechanism for maintaining stability,27 so the source-dependent activity for NepI probably reflects the lack of glycosylation in the rNepI from E. coli. Conversely, native NepII is active to 70 °C and is not glycosylated.13 Although our rNepII appears to be slightly more temperature stable, we used assay times shorter than those of Athauda et al.,13 so the difference could simply reflect enzyme denaturation kinetics. We next compared the activity profile of enzymes over a wide pH range (1.6−8) to determine if a classical aspartic protease pH profile applied to both nepenthesins (Figure 2B). The optimal activity for rNepI and rNepII was observed at pH ∼2.5, slightly higher than that of porcine pepsin (pH ∼2). However, when we compared rNepII to rNepI, we found that rNepII showed a steeper drop in activity with an increase in pH. As with the temperature profile, rNepII appears to be more pepsin-like. Above pH 6, all three enzymes were essentially inactive. These data for rNepI and rNepII are consistent with the pH profiles of the isolated native enzymes.13 Stability. We then conducted a longer-term stability study. The residual activity at pH 2.5 was measured after incubating the enzymes at 37 °C in buffers ranging from pH 1.6 to 8, for 7 and 30 days (Figure 2C,D). All enzymes retained significant levels of activity under acidic conditions, but long-term storage at pH