Multiplex Detection of Protein Toxins Using MALDI-TOF-TOF Tandem

Oct 19, 2012 - From a bioterrorism and warfare perspective, these agents are likely to cause maximum damage to a civilian or military population throu...
0 downloads 0 Views 422KB Size
Download PDF
Article pubs.acs.org/ac

Multiplex Detection of Protein Toxins Using MALDI-TOF-TOF Tandem Mass Spectrometry: Application in Unambiguous Toxin Detection from Bioaerosol Syed Imteyaz Alam,* Bhoj Kumar, and Dev Vrat Kamboj Biotechnology Division, Defence Research and Development Establishment, Gwalior-474002, India S Supporting Information *

ABSTRACT: Protein toxins, such as botulinum neurotoxins (BoNTs), Clostridium perf ringens epsilon toxin (ETX), staphylococcal enterotoxin B (SEB), shiga toxin (STX), and plant toxin ricin, are involved in a number of diseases and are considered as potential agents for bioterrorism and warfare. From a bioterrorism and warfare perspective, these agents are likely to cause maximum damage to a civilian or military population through an inhalational route of exposure and aerosol is considered the envisaged mode of delivery. Unambiguous detection of toxin from aerosol is of paramount importance, both for bringing mitigation protocols into operation and for implementation of effective medical countermeasures, in case a “biological cloud” is seen over a population. A multiplex, unambiguous, and qualitative detection of protein toxins is reported here using tandem mass spectrometry with MALDI-TOF-TOF. The methodology involving simple sample processing steps was demonstrated to identify toxins (ETX, Clostridium perf ringes phospholipase C, and SEB) from blind spiked samples. The novel directed search approach using a list of unique peptides was used to identify toxins from a complex protein mixture. The bioinformatic analysis of seven protein toxins for elucidation of unique peptides with conservation status across all known sequences provides a high confidence for detecting toxins originating from any geographical location and source organism. Use of tandem MS data with peptide sequence information increases the specificity of the method. A prototype for generation of aerosol using a nebulizer and collection using a cyclone collector was used to provide a proof of concept for unambiguous detection of toxin from aerosol using precursor directed tandem mass spectrometry combined with protein database searching. ETX prototoxin could be detected from aerosol at 0.2 ppb concentration in aerosol.

P

pathogenesis of toxinotypes B and D, enterotoxemia, or pulpy kidney disease in lambs, being a rapidly fatal disease.3 Staphylococcal enterotoxin B (SEB) is a single-chain polypeptide (28.4 kDa) consisting of 239 amino acid residues. It is a potent gastrointestinal toxin and a super antigen, activating CD4+ T cells and causing rapid and massive proliferation of cells and cytokine production.4 Protein toxin Clostridium perf ringens phospholipase C (alpha toxin; PLC) is the predominant lethal toxin and the main virulence factor associated with gas gangrene5 produced during vegetative growth of type A strains of C. perf ringens. Gas gangrene generally begins with infection of a deep wound and spreads rapidly with extensive myonecrosis of the infected tissue and often amputation becomes unavoidable if not treated early. Hundreds of thousands of soldiers died of gas gangrene as a result of battlefield injuries, and C. perf ringens was widely recognized as being the most important causal organism of the disease.6

rotein toxins, especially those of bacterial origin, are involved in a number of diseases and are considered as potential agents for bioterrorism and warfare. Among the criteria to prioritize these agents in terms of biothreat potential are the high toxicity, stability in the environment, the ease of production, and lack of prophylactic/therapeutic countermeasures. Protein toxins such as botulinum neurotoxins (BoNTs), Clostridium perf ringens epsilon toxin (ETX), staphylococcal enterotoxin B (SEB), shiga toxin (STX), and plant toxin ricin can act alone on the host system even in the absence of producing organism and are included in the list of biological and toxin warfare (BTW) agents (http://www.bt.cdc.gov/ agent/agentlist-category.asp). From bioterrorism and warfare perspectives, these agents are likely to cause maximum damage to a civilian or military population through an inhalational route of exposure and aerosol is considered the envisaged mode of delivery.1 Epsilon toxin (ETX) is produced by type B and type D strains of Clostridium perf ringens, is one of the most potent toxins known, and is classified as a category B biological agent.2 ETX is produced as a single-chain prototoxin (32.981 kDa) containing a signal peptide and is activated by proteases such as trypsin, α-chymotrypsin, and λ-protease. It is implicated in © XXXX American Chemical Society

Received: October 3, 2012 Accepted: October 19, 2012

A

dx.doi.org/10.1021/ac3028678 | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

precursor masses corresponding to unique peptides from these seven protein toxins was generated for directed search from tryptic digest on MALDI-TOF-TOF. The instrument was programmed to search for these selected masses (m/z) in a ± 1 Da window and to carry out subsequent MS/MS analysis if the precursors were detected in MS spectrum in reflector mode. Presence of at least three unique precursor peptides for a given toxin and MS/MS fragmentation data of at least one precursor peptide (with sequence information) showing significant database match was the criteria for scoring a toxin positive. The method involved minimal sample processing and was successfully demonstrated to unequivocally identify any of the three toxins (ETX, PLC, and SEB) tested; blind samples spiked with toxin, either alone or in combination and even in a background of complex protein mixture could be successfully annotated for the presence of toxin. Using bovine serum albumin (BSA) as simulant, aerosolization and capture was optimized using a nebulizer and cyclone collector, respectively. Unambiguous detection of epsilon prototoxin from aerosol was achieved using this directed search approach which can be extrapolated for accurate identification of any protein toxin from a suspected cloud.

Unambiguous detection of toxin from aerosol is of paramount importance, both for bringing mitigation protocols into operation and for implementation of effective medical countermeasures, in case a “biological cloud” is seen over a population. The detection technologies for toxins need to be protein based which is crippled by the lack of molecular amplification tools similar to polymerase chain reaction (PCR) that is generally used for infectious agents. Detection methods for toxins reported so far have been mainly focused toward quantitative assays from food samples, often employing an immunoassay such as enzyme linked immunosorbent assay (ELISA) or surface plasmon resonance (SPR) biosensor.7−11 Immunoassays such as immunochromatographic strips are of importance in providing rapid results in field conditions. However, their use in a biothreat scenario is of limited value as the method is liable to give false negative or false positive results due to cross-reactivity and/or interferences from complex environmental milieu. Ambiguity of result cannot be tolerated in a bioterrorism or warfare situation as the test result entails a series of mitigation and countermeasure actions involving several agencies. Moreover, in a biothreat scenario (especially for an aerosol route of delivery), our primary objective is to identify the agent from the list of potential agents, a multiplex platform for specific detection being more important than ambiguous, individual, sensitive methods. Hence, unambiguous and multiplex detection of select toxin agents from aerosol is an important and challenging task. Mass spectrometry-based analysis of toxins (such as those for ricin, SEB, BoNT, tetanus neurotoxin (TeNT), and ETX) has gained momentum, owing to the sensitivity and fidelity of the technology.12−21 These studies were either directed toward generation of reference spectra from purified toxins12,13 or used for detection and/or quantitation from food samples.14,15,18−21 The research groups often employed immunocapture of toxin from complex food matrixes as a sample processing step and worked in selected reaction monitoring (SRM) or multiple reaction monitoring (MRM) mode.21,22 Studies pertaining to MS-based detection methodologies using environmental samples have been scanty, and to the best of our knowledge, there is no report for application of this high throughput, unambiguous technique for toxin detection from aerosol. The only problem with the MS-based workflow in the biothreat scenario is a relatively longer time-to-result as the analysis of proteins in bottom-up proteomic approaches involves tryptic digestion. However, recent developments in sample processing promises to substantially reduce this time lapse and overcome the sample processing bottleneck.23 Notably, MS-based identification of bacterial protein toxins, reported so far, largely rely on indirect measurement of target peptides produced as a result of tryptic digestion of the toxin itself24 or direct determination of the molecular mass of the toxin after chromatographic or affinity separation.25 Tandem mass spectrometry using MALDI as an ion source obviates the need for a peptide separation step and provides more reliable data, as the results are based on sequence information rather than just the precursor masses. We report here detection of protein toxin from aerosol employing MALDI-TOF-TOF tandem mass spectrometry using a directed search approach. Unique precursor peptides were elucidated by in house MS/MS analysis (ETX, SEB, PLC). Bioinformatic analysis was carried out to ensure that the corresponding peptides were unique to the toxin and shared by all the known toxin sequences. A consolidated inclusion list of

■

EXPERIMENTAL SECTION Chemicals and Reagents. All the chemicals were of analytical grade from Sigma-Aldrich (Sigma Chemical Co., St Louis, MO, USA) unless specified otherwise. Sequencing grade modified trypsin was from Promega (Promega, Madison, WI, USA). Mass standards for MALDI-TOF-TOF analysis were obtained from Applied Biosystem, USA. Safety Considerations. Activated epsilon toxin is highly toxic while the other two toxins, SEB and PLC, are moderately toxic and must be handled in strict adherence to safety rules for handling toxic substances. Epsilon prototoxin before proteolytic cleavage of signal peptides exhibits very low toxicity as compared to activated toxin. For this work, a low amount of prototoxin was used for generation of aerosol under strict safety guidelines and the work was approved by institutional biosafety committee at DRDE, Gwalior. Handling of microgram quantities of active toxins and aerosol generation of epsilon prototoxin was carried out in a high containment facility (biosafety level 3). All contact with the substance should be avoided by taking appropriate protective measures. Toxin Preparation. The epsilon prototoxin with Nterminal 6X His was overexpressed in E. coli host, and the recombinant protein was purified to near homogeneity under nondenaturing conditions using Ni-NTA sepharose resin (Qiagen, Germany) as per manufacturer’s instructions. Purified proteins were dialyzed against phosphate-buffered saline (PBS), and the purity of the protein was analyzed by 12% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDSPAGE). Identity of the protein was confirmed by tandem mass spectrometry of excised band after tryptic digestion as described below. Recombinant SEB was similarly expressed as described above, and purified recombinant toxin was used for experiments. Partially purified PLC (type 1 and type XIV) was obtained from Sigma-Aldrich (Sigma Chemical Co., St Louis, MO, USA). The specific band corresponding to ∼43 kDa PLC was gel-excised for generation of reference spectra while partially purified fraction was used for other experimental work. Total protein concentration was determined using Quick Start Bradford Protein Assay kit (Bio-Rad, USA) as per manufacB

dx.doi.org/10.1021/ac3028678 | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

Table 1. MS and MS/MS Results of Peptides from Tryptic Digest of Three Protein Toxins (ETX, SEB, and PLC) Used for Generation of Inclusion List for Directed Search from Unknown Samples peptide

peptide sequence

epsilon toxin (ETX) E1 ASYDNVDTLIEK E2 YYPNAMAYFDK E3 VTINPQGNDFYINNPK E4 VELDGEPSMNYLEDVYVGK E5 NTDTVTATTTHTVGTSIQATAK E6 FTVPFNETGVSLTTSYSFANTNTNTNSK E7 EITHNVPSQDILVPANTTVEVIAYLK E8 LVGQVSGSEWGEIPSYLAFPR E9 FSLSDTVNKSDLNEDGTININGK E10 GNYSAVMGDELIVK E11 NLNTNNVQEYVIPVDK E12 NLNTNNVQEYVIPVDKK Staphylococcus enterotoxin B (SEB) S1 FTGLMENMK S2 VLYDDNHVSAINVK S3 SIDQFLYFDLIYSIK S4 SIDQFLYFDLIYSIKDTK S5 LGNYDNVR S6 DKYVDVFGANYYYQCYFSK S7 YVDVFGANYYYQCYFSK S8 TCMYGGVTEHNGNQLDKYR S9 VFEDGKNLLSFDVQTNK S10 NLLSFDVQTNK S11 KVTAQELDYLTR S12 VTAQELDYLTR S13 KLYEFNNSPYETGYIK S14 LYEFNNSPYETGYIK S15 FIENENSFWYDMMPAPGDK S16 FIENENSFWYDMMPAPGDKFDQSK S17 YLMMYNDNK alpha toxin (PLC) P1 DNSWYLAYSIPDTGESQIR P2 FETFAEER P3 SIYYSHASMSHSWDDWDYAAK P4 GTAGYIYR P5 IDDIQNMWIR a

position

m/zobsd (Da)a

Δ error (Da)

|Δ| (ppm)

E-value

MS/MS ion score

47−58 74−84 85−100 101−119 141−162 163−190 191−216 226−246 251−273 274−287 290−305 290−304

1367.716 1382.656 1833.998 2157.110 2219.220 3042.596 2864.606 2292.276 2481.229 1495.804 1860.033 1988.139

0.0467 0.0476 0.0849 0.1042 0.1052 0.0261 0.0255 0.1097 −0.1234 0.0577 0.0830 0.0938

34.1 34.4 46.3 48.3 47.4 8.5 8.9 47.8 49.7 38.5 44.6 47.1

0.0045 3.8 × 10−5 7.7 × 10−10 6.2 × 10−9 0.0300 3.00 × 10−5 1.40 × 10−15 2 × 10−12 1.2 × 10−6 3.3 × 10−8 3.2 × 10−9 2.5 × 10−9

67 88 135 126 59 215 178 161 107 119 129 130

16−24 25−38 39−53 39−56 57−64 78−96 80−96 111−129 135−151 141−151 153−164 154−164 172−187 173−187 188−206 188−211 212−220

1070.548 1586.910 1865.077 2209.282 950.508 2430.240 2187.095 2243.137 1954.118 1278.739 1436.859 1308.755 1966.082 1837.974 2291.133 2896.464 1191.575

0.0473 0.0935 0.1045 0.1398 0.0397 0.1655 0.1429 0.1406 0.1271 0.0703 0.0847 0.0756 0.1229 0.1101 0.1550 0.2056 0.0583

44.2 58.9 56.0 63.2 41.7 68.1 65.3 62.6 65.0 55.0 58.9 57.7 62.5 59.9 67.6 70.9 48.9

3.60 5.60 5.40 9.10 1.20 7.00 1.10 3.00 1.00 2.70 1.40 5.90 8.60 4.60 2.00 4.10 7.10

× × × × × × × × × × × × × × × × ×

10−2 10−10 10−7 10−5 10−3 10−11 10−9 10−2 10−9 10−3 10−5 10−7 10−10 10−9 10−3 10−9 10−2

58 137 107 84 73 145 133 59 134 70 92 106 134 127 71 127 56

109−127 179−186 230−250 259−266 346−355

2215.153 1028.513 2520.157 900.495 1303.712

0.1231 0.0445 0.1590 0.0380 0.0661

55.5 43.2 63.0 42.2 50.7

1.70 1.80 2.40 7.30 9.10

× × × × ×

10−7 10−2 100 10−1 10−4

111 61 40 45 75

Observed monoisotopic masses [M + H]+ are shown here.

turer’s instructions. The protein concentration was calculated using bovine serum albumin (BSA) as standard. Aerosol Generation and Collection. Aerosol was generated using a nebulizer (MRK Healthcare, India) meant for medicinal purposes and producing particles in the size range of 10−15 μm. The nebulizer was made up of an aerosol generation unit with a liquid container of 10 mL and an air compressor unit. It was shown to generate aerosol with a throughput of approximately 12 mL of liquid per hour. The assembly for aerosol generation with a chamber (0.30 × 0.45 × 0.60 m) has been shown in the schematic illustration (abstract graphic). The generated aerosol was collected using a wet cyclone collector (Bertin Technologies, France) capable of collecting large concentrations of aerosols ranging in size from 0.5 to 10 μm into a liquid sample for downstream analysis of biological material. The aerosol collection rate could be adjusted for 100, 200, and 300 L/min. Inactive prototoxin was tested for toxicity by i.v. injection in duplicate mice before aerosol generation to ensure that it carried no toxicity, and the

experiments were carried out in containment facility (biosafety level 3, BSL3). Sample Preparation. Reference spectra were generated after in-gel digestion of 1 μg of toxin separated on 12% SDSPAGE and excised with the help of a scalpel. In-gel digestion of protein was carried out using the method previously described26 with some modifications (details in Supporting Information). In-sol digestion was optimized using different solubilization buffers for protein samples as described in the results. Enrichment and cleanup was carried out for protein samples in 10 mL of Tris-HCl buffer ((10 mM, pH 7.5), either reflecting envisaged aerosol collection methodology or actual samples collected using a cyclone collector. Protein in the solution was precipitated with trichlo-acetic acid (TCA) at a final concentration of 10% (w/v) in the presence of β-mercapto ethanol (0.07%). After 2 h of incubation on ice, the protein pellet was collected by centrifugation (10 000g, 4 °C, 10 min) and washed twice with 500 μL of acetone after transferring the pellet to 1.5 mL microfuge tube (Eppendorf, USA). The pellet was air-dried for 5 min and resuspended in 10 μL of either C

dx.doi.org/10.1021/ac3028678 | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

(∼5−15 μM) was concentrated and collected in 10 mL of buffer using a cyclone collector. Sample preparation was optimized for this 10 mL aqueous sample as starting material, and a directed search using MALDI-TOF-TOF was demonstrated to successfully identify toxin in a qualitative manner with blind spiked samples. In view of biosafety concerns, we used only three toxins (ETX, SEB, and PLC) for preparing blind samples and ETX prototoxin was used for generation of aerosol. PLC, although not a BTW agent, was included in the study as the toxin is major virulence determinant for gas gangrene caused by Clostridium perf ringens which is a validated biological weapon.28 Previous mass spectrometry-based studies for toxins have been largely directed toward detection and/or quantitation from food samples.14,15,18−21 We provide here a proof of concept for confirmatory qualitative detection of protein toxins from aerosol that can be theoretically expanded to any number of protein toxins. Generation of Reference Spectra. Reference MS and MS/MS spectra for ETX, PLC, and SEB were generated after in-gel digestion of 1 μg of toxin as described in the methods. Three or four independent digestions were carried out, and peptides with significant MS/MS ion score were consolidated to generate a list of peptides (Table 1, Table S1, Supporting Information). Representative MS and MS spectra for the three toxins are shown in Figure S1, Supporting Information. The mass list obtained from the MS/MS fragmentation of precursor ions was searched against the MASCOT database, and the resulting “b” and “y” ion series has been summarized in Figure S1, Supporting Information. MS and MS/MS data from a representative analysis for a given precursor have been reflected in Table 1 and Figure S1, Supporting Information. Analysis of tryptic digest of epsilon prototoxin provided the most intense 12 monoisotopic [M + H]+ peptides (E1 to E12, signal-to-noise ratio of >20) with significant MS/MS ion score suitable for ETX detection. There is only one previous report pertaining to MS analysis of ETX21 wherein the authors have used LC coupled ESI-MS using both top-down and bottom-up approaches and reported 23 peptides for prototoxin upon tryptic digestion. Since we did not use any LC separation and our criteria for reporting a precursor peptide is based on meaningful fragmentation upon CID (with significant MS/MS ion score, signal-to-noise ratio of >10 for fragment ions), the number of peptides reported here is comparatively low. In total, we report here 89 distinct b-type and 145 y-type ions (Figure S1, Supporting Information) for prototoxin as against 71 b-type and 42 y-type ions reported by Seyer et al.21 Of the 12 peptides reported here, three peptides, E3 (m/z 1833.998, amino acid residues 85−100), E8 (m/z 2292.276, amino acid residues 226−246), and E12 (m/z 1988.139, amino acid residues 290− 304), were detected with high sensitivity and were amenable to fragmentation with CID. Notably, peptide E3 was one among the two peptides selected for quantitative MS-based detection of ETX from food samples using LC ESI-MS/MS.21 However, peptides E12 and E9, originating from one missed cleavage, were consistently observed in our study and not reported previously. Ours is the first report of tandem MS/MS analysis for the bacterial toxins SEB and PLC wherein we report 17 and 5 peptides, respectively, with significant fragmentation data. Peptides S2 (m/z 1586.910, amino acid residues 25−38), S5 (m/z 950.508, amino acid residues 57−64), S12 (m/z 1308.755, amino acid residues 154−164), and S14 (m/z 1837.974, amino acid residues 173−187) were consistently observed with high sensitivity and good MS/MS ion score in

NH4HCO3 (50 mM) or trifluoro-ethanol (TFE, 10%) before subjecting to in-sol tryptic digestion using the standard protocol as described in the Supporting Information. For digestion of a crude and complex protein mixture, a brief SDS-PAGE step (30 min) was added. In this case, the protein pellet obtained after TCA precipitation and washing was solubilized in SDS lysis buffer [12.5 mM Tris-HCl (pH 6.8), 4% glycerol (w/v), 0.4% SDS (w/v), and 1% β-mercaptoethanol (v/v)] and loaded onto a discontinuous denaturing SDSPAGE after heating at 95 °C for 5 min. The electrophoresis was carried out at a constant current of 20 mA, and just after the protein mixture crossed the stacking and resolving gel interface (as indicated by the tracking dye), the electrophoresis was terminated and the single band containing mixture of proteins was subjected to in-gel digestion as described above. MALDI-TOF-TOF Analysis. In a bottom-up approach, digested proteins were analyzed by Applied Biosystem 4800 plus MALDI TOF/TOF Analyzer (AB Sciex, USA) using conditions as detailed in methods, Supporting Information. The digested peptides were mixed with equal volume of the CHCA matrix solution (10 mg/mL) and spotted onto the target plate. MS mass spectra were recoded in the reflector positive mode using a laser with accelerated voltage of 2 kV. The MS/MS mass spectra were acquired by the data dependent acquisition method, and for routine identification and generation of reference spectra, the 20 strongest precursors were selected between 850 and 4000 Da and filtered with a signal-to-noise ratio greater than 20 from one MS scan. MS/MS fragment ions were generated by collision induced dissociation (CID) at 1 kV voltage. MS and MS/MS data were analyzed, and peak lists were generated using the 4000 Series Explorer Software v. 3.5 (Applied Biosystems). MS/MS peaks were selected on the basis of a signal-to-noise ratio greater than 10, and the data were analyzed using Protein Pilot version 4.0 (Applied Biosystem) employing the MASCOT 2.0 search engine (Matrix Science, London, UK). For directed search approach, precursor ions on a ± 1 Da window with a signal-to-noise ratio greater than 20 were selected for MS/MS analysis from the inclusion list (m/z for all unique toxin specific peptides) defined in the interpretation method of the instrument. Bioinformatic Analysis. Peptide sequences, their position on protein, and m/z values have been listed for seven protein toxins either from the present study (ETX, PLC, SEB) or from data reported earlier (Table S1, Supporting Information). Each of the peptide sequences were subjected to global protein blast (www.ncbi.nlm.nih.gov) with search parameters adjusted for short input sequence. All the reference sequences for a given toxin were retrieved from the database, and these sequences were further short listed using DAMBE software27 that recognizes identical sequences. The infile of all the unique representative sequences of a toxin type were used for global alignment using ClutalW (www.ebi.ac.uk). The alignment data was used to determine the conservation of each peptide across all the known sequences.

■

RESULTS AND DISCUSSION Several protein toxins, largely from bacterial origin, are potential biothreat agents from bioterrorism and warfare perspectives. Aerosol being the envisaged mode of delivery, a detection platform for unambiguous and multiplex identification of these toxins especially from aerosol is of paramount importance. A prototype simulating generation of a “biological cloud” of toxins was made, and aerosol of breathable size range D

dx.doi.org/10.1021/ac3028678 | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

5000 ng of ETX, each in 1 mL of Tris buffer, were prepared. Protein was precipitated and purified by the TCA−acetone method as described in the text. Protein pellets of one set were solubilized in 10 μL of 10% (v/v) trifloroethanol (TFE) while the pellets of the second set were solubilized in 10 μL of 50 mM NH4HCO3. Each sample was subjected to in-sol digestion in a 30 μL reaction volume [final concentration of TFE in the first set was 5% (v/v)] of which 0.4 μL was used for loading onto MALDI target plate after digestion. Comparable results were obtained with TCA−acetone enrichment and purification followed by solubilization in NH4HCO3 giving a marginal better sensitivity than the TFE method (Table S3, Supporting Information). However, with SEB, we observed a better peptide coverage using TFE for denaturation (data not shown) and used it for all subsequent digestions involving SEB and unknown samples. The reason for this could be the differences in intrinsic higher order structures of the proteins, as enteroroxins are adapted to protect themselves from the tryptic onslaught of digestive proteolytic enzymes in the stomach. Use of TFE as a denaturing agent has been reported to improve coverage of peptides and originates from the protocols for solvent-based enzymatic digestion of proteins.29,30 For digestion of crude and complex protein mixture from a 10 mL buffer, we used enrichment by TCA precipitation followed by a brief electrophoresis step (30 min) where the protein pellet was solubilized in SDS lysis buffer containing 0.4% SDS (w/v) and was subjected to discontinuous denaturing SDS-PAGE as described in methods. The presence of SDS, heating and gel-based cleanup, substantially improved digestion of the protein mixture. The polyacrylamide gel matrix further helped in maintaining the denatured state of protein and increasing the efficiency of tryptic digestion. The use of surfactant including SDS for protein denaturation has been shown to positively influence tryptic digestion efficiency.30,31 In case of a bioattack, it is imperative to ascertain rapidly and unambiguously whether the biological cloud (aerosol) contains any of the threat agents including protein toxins. Very high sensitivity is of a lesser importance as the amount of agent in such a scenario is likely to be high if it has to cause large scale ill effects. The present work describes a simple sample processing methodology for concentration and cleanup of samples collected through a cyclone collector. There is a large scope of improving the sensitivity of detection by refining the individual steps of sample processing which we shall be pursuing further, as with the current method we could achieve fairly good detection sensitivity (0.1 ppb) of simulant (BSA) from the aerosol. However, the time span of collection-to-result is still an area of concern and needs further improvement. Detection of ETX from Complex Protein Mixture Using Directed Search Approach. On the basis of the reference spectrum of ETX and subsequent bioinformatic analysis of resulting peptides to ensure unique identity and prevalence among known sequences, we generated an inclusion list of 12 peptides (E1 to E12, Table 1) and used it for directed search of epsilon toxin from complex mixtures of proteins. A list of precursor masses in a ± 1 Da window was added to the interpretation method so as to select only precursors near the listed m/z values for subsequent MS/MS analysis, if detected in MS spectrum of tryptic digest in a reflector mode. We first tested this directed search approach for ETX detection in samples with increasing complexity using one, two, or four proteins mixed with ETX. Five hundred micrograms of ETX alone or in a mixture of 1:1 (w/w) with one, two, or four

repeated experiments with SEB. For PLC, peptides P2 (m/z 1028.513, amino acid residues 179−186) and P5 (m/z 1303.712, amino acid residues 346−355) were detected consistently with high sensitivity and even with crude preparations of the toxin. Bioinformatic Analysis of Peptides. The peptide sequences, their position on protein, and m/z values for three protein toxins from the present study (ETX, PLC, SEB) have been reflected in Table S1 (Supporting Information). Each of the peptide sequences was subjected to global protein blast (www.ncbi.nlm.nih.gov) to rule out the possibility that the sequence is shared with any other known protein. Any peptide with 56 indicated identity or extensive homology (p < 0.05)] for at least one peptide. A 2-fold serial dilution of ETX (100 pg to 100 ng) was subjected to in-sol tryptic digestion in a 30 μL reaction volume, and 0.4 μL of the resulting peptide mixture was loaded onto the target plate after mixing with the matrix. With the aforesaid criteria, the sensitivity of detection for the ETX on MALDI target plate was observed to be 166 picogram (∼3.9 femtomole), corresponding to 833 ng·mL−1 of pure recombinant prototoxin. Our sensitivity of detection is substantially lower than that reported by Seyer et al.21 (5 ng·mL−1) as the authors used an immunocapture step for enrichment and ultraperformance LC for purification of peptides after tryptic digestion. However, the sensitivity is comparable to those reported for tetanus neurotoxin (1 μg·mL−1) using LC-MS analysis.12 We envisaged collection of aerosol for toxin detection using a cyclone collector for which 10 mL of Tris buffer (50 mM, pH 8.0) was used as collection buffer. Method for enriched and purification of toxin was optimized using two sets of 2-fold serial dilution of ETX (40 ng/μL to 5 μg/μL) in Tris buffer. Two sets of six samples, containing 38, 75, 150, 300, 1250, and E

dx.doi.org/10.1021/ac3028678 | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

Table 2. Detection of ETX Using Directed Search Approach from Complex Protein Mixture in the Form of Crude Lysate of Environmental Microbial Consortium (LEC) peptides (MS/MS ion score) sample (ng)

MOWSE score

sequence coverage

E1

E2

E3

E4

E5

E6

E7

E8

E9

E10

E11

E12

RMS error

1194 1105 680 559

63 55 40 33

75 82 38 27

67 65 54 73

123 118 122 132

117 154 89 30

43 32

118 82

14

179 194 134 43

98 70 19

94 69 38 45

75 79 60 84

136 122 113 126

54 58 63 61

815 926 818 556

70 63 55 44

34 48 39

46 51 52 59

119 121 105 118

82 77 95 51

31 29 30

72 110 54 29

26 18

179 179 164 94

28 54 38 17

50 44 46

23 57 83 48

126 124 104 101

59 70 71 69

ETX 1500 750 375 185 ETX/LEC 1500:6000 750:3000 375:1500 185:750

Table 3. Multiplex Detection of Protein Toxins from Blind Samples Using Directed Search Approacha sample

toxin detected

MOWSE score

sequence coverage

S1 S2

ND ETX

1108

67

S3

SEB ETX

270 1000

43 60

PLC ETX SEB ETX SEB PLC ND ND

72 338 114 694 513 112

4 19 21 44 41 9

S4 S5 S6 S7 S8 S9

peptides (MS/MS ion score) E1(24), E2(53), E3(108), E4(86), E5(16), E6(148) E7(147), E8(188), E9(69), E10(33), E11(44), E12(80) S2(56), S6(34), S8(34), S9(53), S11(20), S12(53), S14(21) E1(25), E2(51), E3(108), E4(129), E6(168), E7(103), E8(181), E9(31), E10(65), E11(46), E12(74) P2(42), P5(30) E2(59), E3(77), E8(160), E12(43) S2(46), S5(15), S12(40), S14(13) E2(57), E3(108), E4(96), E7(118), E8(169), E10(41), E12(104) S2(86), S3(64), S5(52), S8(66), S11(27), S12(80), S13(39), S14(65), S17(34) P1(22), P2(42), P5(48)

RMS error 6 11 9 4 10 32 5 12 10

a

Ten milliliters of Tris buffer alone or spiked with one, two, or three toxins (ETX/SEB/PLC) with or without other proteins of varying complexity were subjected to enrichment and cleanup and for detection of toxin using MS and MSMS analysis. S1 = 1 μg LEC; S2 = 1 μg ETX and 1 μg SEB; S3 = 1 μg ETX and 1 μg PLC; S4 = 1 μg ETX, 1 μg SEB, and 1 μg PLC; S5 = 1 μg ETX and 1 μg LEC; S6 = 1 μg SEB and 1 μg LEC; S7 = 1 μg PLC and 1 μg LEC; S8 = 1 μg LEC; S9 = distilled water. LEC = lysate of environmental microbial consortium.

wall components. The mixture of ETX and the crude lysate was subjected to TCA−acetone precipitation and washing, followed by a brief SDS-PAGE step as described in the methods. The mixture was allowed to just cross the stacking and resolving gel interface (20 mA, 20 min), and the stacked protein mixture in the form of a single band (the individual proteins were not allowed to resolve) was excised and subjected to in-gel digestion. This step improved the efficiency of tryptic digestion and resulted in removal of contaminating lipids, cell wall components, salts, and other small molecules and complexes. Notably, ETX could be detected even in the presence of this highly complex protein mixture using the directed search approach (Table 2). Multiplex Detection of Toxins (ETX/SEB/PLC) from Blind Samples. Having successfully detected ETX from complex milieu, we tested our methodology for ETX detection using ten blind samples using TCA−acetone cleanup and in-gel digestion. The details of the samples are provided in the footnote of Table S5 (Supporting Information). ETX was successfully identified from the positive samples with significant ion scores for several peptides. Finally, three toxins either alone or in combination of two, with or without a background of complex protein mixture, were used as blind spiked samples (Table 3) to validate unambiguous, multiplex detection of toxin using an inclusion

proteins (the total amount of other proteins in each case equal to ETX) was 2-fold serially diluted and subjected to in-sol digestion. Purified recombinant proteins ornithine carbamoyltransferase (cOTC), cystathionine beta lyase (CBL), hypothetical protein 2918, and ABC transporter from Clostridium perf ringens were used for making the mixture. ETX could be detected in a mixture of four proteins (1:1) with a slight decrease in sensitivity when compared with scores of toxin alone (Table S4, Supporting Information). An increase in complexity did not appear to significantly affect the MOWSE score in the MASCOT database search, and comparable results were obtained for ETX combined with a mixture of four proteins. We further tested the effect of interfering proteins on ETX detection by directed search approach using more complex protein in the form of Escherichia coli whole cell lysate. ETX and lysate were mixed in different proportions (1:1, 1:2, and 1:4) and subjected to in-sol digestion. The presence of a complex E. coli lysate even at 1:4 (ETX/lysate) ratio did not significantly alter the sensitivity of detection (data not shown) and was comparable to the results obtained with a mixture of four proteins (Table S4, Supporting Information). A crude lysate of an environmental microbial consortium (LEC) was used to further increase the complexity of the interfering proteins. The crude lysate, in addition to the protein, contained other impurities such as nucleic acid and cell F

dx.doi.org/10.1021/ac3028678 | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

Table 4. Unambiguous Detection of ETX from Aerosol by Directed Search Approach Using ETX Alone and with a Background of Complex Interfering Proteinsa ETX released (μg)

aerosol collected (L)

calculated ETX concentration (ppb)

MOWSE score

coverage (%)

100

100

1.00

1081

62

20 10

100 100

0.20 0.10

78 ND

11

100

100

1.00

1118

49

250 + 250

50

100

0.50

567

42

100 + 100 50 + 50

20 10

100 100

0.20 0.10

64 ND

11

ETX used for aerosol (μg/mL) ETX 500 100 50 ETX + LEC 500 + 500

a

peptides (MS/MS ion score) E2(81), E3(123), E4(101), E6(216), E7(185), E8(171), E9(18), E10(32), E11(37), E12(118) E3(46), E8(32)

E2(71), E3(123), E4(125), E6(206), E7(136), E8(178), E9(80), E11(32), E12(110) E2(49), E3(117), E4(28), E6(129), E7(44), E8(126), E12(74) E3(31), E8(33)

RMS error 8 36

41 33 28

Lysate of environmental microbial consortium, LEC.

facility (BSL3). Aerosol was collected for 1 min at a rate of 100 L/min in a 10 mL Tris buffer. After sample processing and tryptic digestion, the peptide mix was subjected to MS and MS/ MS analysis using the inclusion list and directed search approach. We could detect ETX prototoxin from aerosol at 0.2 ppb concentration even in the presence of an equivalent amount of interfering complex protein mixture (Table 4).

list of selected peptides (for all the seven toxins). Samples were processed as above, and the presence of toxin was successfully shown in all positive samples among the nine possibilities tested. Reproducibility of the qualitative detection was investigated by repeating the analysis three times using ten blind samples either without or with one or two toxins (alone or mixed with complex mixture) at a concentration of ∼100 ng·mL−1. The accuracy of assay was 100% as there was no false positive or false negative result with any of the samples. Detection of ETX from Aerosol. Aerosol was generated using a nebulizer, and the released aerosol was collected by a cyclone collector in 10 mL of Tris buffer (10 mM, pH 7.5), as reflected in the schematic illustration of the prototype (abstract graphic). The entire setup was kept inside the biosafety cabinet, and the initial optimizations with respect to rate of collection and limit of detection were carried out using bovine serum albumin (BSA) as simulant. Five different concentrations of BSA in the range of 0.01 to 0.1 mg/mL were used for generation of aerosol which was collected for 1 min at rates of 100 and 200 L/min using the cyclone collector. Theoretical concentration of BSA in aerosol was calculated from the initial and final weight of solution before and after the generation of aerosol and the volume of air collected using the cyclone collector. The collected sample in 10 mL of Tris buffer was subjected to enrichment and cleanup as described above and subjected to MS/MS analysis after tryptic digestion. On the basis of the densitometric analysis of coomassie stained SDSPAGE images, we calculated the efficiency of collection. Any loss of toxin could be due to a differential partitioning of protein molecules between the liquid and aerosol phase during the nebulization process, escape of particles from cyclone collector without getting into collection buffer, and the process of TCA precipitation and washing by acetone. The calculated efficiency of collection for BSA at 100 mL/min rate of cyclone collection was estimated to be approximately 1% (data not shown). The efficiency of collection was observed to decline with an increase in rate of collection, likely due to escape of protein molecules without being trapped in the collection buffer (data not shown). We could detect BSA with a significant MOWSE score at a concentration in aerosol as low as 0.1 ppb. Further, inactive prototoxin (ETX) was used to test the workflow for unambiguous detection of toxin from aerosol using the same setup as described above in a containment

■

CONCLUSION We describe here a multiplex, unambiguous, and qualitative detection of protein toxins by MALDI-TOF-TOF which can theoretically be extended to all protein toxins. Ours is the first report for MS-based multiplex detection of toxin from a simulated biological cloud and involves minimal sample processing. The novel directed search approach using tryptic digest of crude samples directly spotted onto MALDI target plate obviates the need for a LC-based separation and handles the interferences from complex samples. The bioinformatic analysis of seven protein toxins for elucidation of unique peptides with conservation status across all known sequences provides a high confidence for detecting toxins originating from any geographical location and source organism. Use of tandem MS data with peptide sequence information increases the specificity of the method. Together, we provide a proof of concept for unambiguous and multiplex detection of toxin from aerosol using precursor directed tandem mass spectrometry combined with protein database searching. The methodology can be substantially improved with respect to sample processing in order to reduce the time-to-result from the point of capture of the aerosol. Use of molecular size cutoff membranes, microfluidics for sample preparation, and/or purification of digested peptides on one or two dimensions can improve sensitivity and applicability of the methodology.

■

ASSOCIATED CONTENT

S Supporting Information *

Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*Phone: 91-751-2390276. Fax: 91-751-2341148. E-mail: [email protected]. G

dx.doi.org/10.1021/ac3028678 | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

Notes

(26) Speicher, K; Kolbas, O.; Harper, S.; Speicher, D. W. J. Biomol. Tech. 2000, 11, 74−86. (27) Xia, X.; Xie, Z. D. J. Hered. 2001, 92, 371−373. (28) Ecker, D. J.; Sampath, R.; Willett, P.; Wyatt, J. R.; Samant, V.; Massire, C.; Hall, T. A.; Hari, K.; McNeil, J. A.; Büchen-Osmond, C.; Budowle, B. BMC Microbiol. 2005, 5, 19. (29) Russell, W. K.; Park, Z.-Y.; Russell, D. H. Anal. Chem. 2001, 73, 2682−2685. (30) Proc, J. L.; Kuzyk, M. A.; Hardie, D. B.; Yang, J.; Smith, D. S.; Jackson, A. M.; Parker, C. E.; Borchers, C. H. J. Proteome Res. 2010, 9 (10), 5422−5437. (31) Norrgran, J.; Williams, T. L.; Woolfitt, A. R.; Solano, M. I.; Pirkle, J. L.; Barr, J. R. Anal. Biochem. 2009, 393, 48−55.

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank Prof. (Dr.) M. P. Kaushik, Director, DRDE, Gwalior, for providing all facilities and support required for this study. The work has been funded by Defence Research and Development Organization, Government of India.

■

REFERENCES

(1) Arnon, S. S.; Schechter, R.; Inglesby, T. V.; Henderson, D. A.; Bartlett, J. G.; Ascher, M. S.; Eitzen, E.; Fine, A. D.; Hauer, J.; Layton, M.; Lillibridge, S.; Osterholm, M. T.; O’Toole, T.; Parker, G.; Perl, T. M.; Russell, P. K.; Swerdlow, D. L.; Tonat, K. JAMA 2001, 285, 1059− 1070. (2) Brown, M. B.; Savva, C. G.; da Costa, S. P. F.; Naylor, C. E.; Basak, A. K.; Titball, R. W. FEBS J. 2011, 278, 4589−4601. (3) Popoff, M. R. FEBS J. 2011, 278, 4602−4615. (4) Franz, D. R.; Jaax, N. K. In Medical Aspects of Chemical and Biological Warfare; Borden Institute, Walter Reed Army Medical Center: Washington, DC, 1997; pp 631−642. (5) Awad, M. M.; Bryant, A. F.; Stevens, D. L.; Rood, J. I. Mol. Microbiol. 1995, 15, 191−202. (6) MacLennan, J. D. Lancet 1943, ii, 123−126. (7) Griffiths, G. D.; Newman, H.; Gee, D. J. J. Forensic Sci. Soc. 1986, 26, 349−358. (8) Shyu, H. F.; Chiao, D. J.; Liu, H.; Tang, S. S. Hybridoma Hybridomics 2002, 21, 69−73. (9) Sasaki, T.; Terano, Y.; Shibata, T.; Kawamoto, H.; Kuzuguchi, T.; Kohyama, E.; Watanabe, T.; Ohyama, T.; Gemba, M. Microbiol. Immunol. 2005, 49, 589−597. (10) Pauly, D.; Kirchner, S.; Stoermann, B.; Schreiber, T.; Kaulfuss, S.; Schade, R.; Zbinden, R.; Avondet, M. A.; Dorner, M. B.; Dorner, B. G. Analyst 2009, 134, 2028−2039. (11) Soelberg, S. D.; Stevens, R. C.; Limaye, A. P.; Furlong, C. E. Anal. Chem. 2009, 81, 2357−2363. (12) van Baar, B. L. M.; Hulst, A. G.; Roberts, B.; Wils, E. R. J. Anal. Biochem. 2002, 301, 278−289. (13) Hines, H. B.; Lebeda, F.; Hale, M.; Brueggemann, E. E. Appl. Environ. Microbiol. 2005, 71 (8), 4478−4486. (14) Callahan, J. H.; Shefcheck, K. J.; Williams, T. L.; Musser, S. M. Anal. Chem. 2006, 78, 1789−1800. (15) Dupuis, A.; Hennekinne, J. A.; Garin, J.; Brun, V. Proteomics 2008, 8, 4633−4636. (16) Duriez, E.; Fenaille, F.; Tabet, J. C.; Lamourette, P.; Hilaire, D.; Becher, F.; Ezan, E. J. Proteome. Res. 2008, 7, 4154−4163. (17) Duriez, E.; Goossens, P. L.; Becher, F.; Ezan, E. Anal. Chem. 2009, 81, 5935−5941. (18) Kalb, S. R.; Barr, R. Anal. Chem. 2009, 81, 2037−2042. (19) Kull, S.; Pauly, D.; Störmann, B.; Kirchner, S.; Stämmler, M.; Dorner, M. B.; Lasch, P.; Naumann, D.; Dorner, B. G. Anal. Chem. 2010, 82, 2916−2924. (20) McGrath, S. C.; Schieltz, D. M.; McWilliams, L. G.; Pirkle, J. L.; Barr, J. R. Anal. Chem. 2011, 83, 2897−2905. (21) Seyer, A.; Fenaille, F.; Féraudet-Tarisse, C.; Volland, H.; Popoff, M. R.; Tabet, J. C.; Junot, C.; Becher, F. Anal. Chem. 2012, 84, 5103− 5109. (22) Gallien, S.; Duriez, E.; Domon, B. J. Mass Spectrom. 2011, 46, 298−312. (23) Norrgran, J.; Williams, T. L.; Woolfitt, A. R.; Solano, M. I.; Pirkle, J. L.; Barr, J. R. Anal. Biochem. 2009, 393, 48−55. (24) Barr, J. R.; Moura, H.; Boyer, A. E.; Woolfitt, A. R.; Kalb, S. R.; Pavlopoulos, A.; McWilliams, L.; Schmidt, J. G.; Martinez, R. A.; Ashley, D. L. Emerging Infect. Dis. 2005, 11, 1578−1583. (25) Williams, J. P.; Green, B. N.; Smith, D. C.; Jennings, K. R.; Moore, K. A. H.; Slade, S. E.; Roberts, L. M.; Scrivens, J. H. Biochemistry 2005, 44, 8282−8290. H

dx.doi.org/10.1021/ac3028678 | Anal. Chem. XXXX, XXX, XXX−XXX