Rapid Quantification of Clostridial Epsilon Toxin in Complex Food and

Apr 30, 2012 - Multiple-reaction monitoring for multiplex detection of three bacterial ... for terror—characteristics, challenges and medical counte...
0 downloads 0 Views 878KB Size
Download PDF
Article pubs.acs.org/ac

Rapid Quantification of Clostridial Epsilon Toxin in Complex Food and Biological Matrixes by Immunopurification and Ultraperformance Liquid Chromatography-Tandem Mass Spectrometry Alexandre Seyer,† François Fenaille,† Cecile Féraudet-Tarisse,† Hervé Volland,† Michel R. Popoff,‡ Jean-Claude Tabet,§ Christophe Junot,† and François Becher*,† †

CEA, iBiTec-S, Service de Pharmacologie et d’Immunoanalyse, 91191 Gif-sur-Yvette, France Institut Pasteur, Unité des Bactéries Anaérobies et Toxines, 75724 Paris Cedex 15, France § Université Pierre et Marie Curie, Laboratoire de Synthèse, Structure et Fonction de Molécules Bioactives, CNRS UMR 7613, 75252 Paris, France ‡

S Supporting Information *

ABSTRACT: Epsilon toxin (ETX) is one of the most lethal toxins produced by Clostridium species and is considered as a potential bioterrorist weapon. Here, we present a rapid mass spectrometry-based method for ETX quantification in complex matrixes. As a prerequisite, naturally occurring prototoxin and toxin species were first structurally characterized by top-down and bottom-up experiments, to identify the most pertinent peptides for quantification. Following selective ETX immunoextraction and trypsin digestion, two proteotypic peptides shared by all the toxin forms were separated by ultraperformance liquid chromatography (UPLC) and monitored by ESI-MS (electrospray ionization-mass spectrometry) operating in the multiple reaction monitoring mode (MRM) with collision-induced dissociation. Thorough protocol optimization, i.e., a 15 min immunocapture, a 2 h enzymatic digestion, and an UPLC-MS/MS detection, allowed the whole quantification process including the calibration curve to be performed in less than 4 h, without compromising assay robustness and sensitivity. The assay sensitivity in milk and serum was estimated at 5 ng·mL−1 for ETX, making this approach complementary to enzyme linked immunosorbent assay (ELISA) techniques.

T

in mice, it passes through the intestinal barrier and disseminates via the circulation to several organs (kidney, brain), causing toxic shock and death.2 Its precise mode of action remains to be clearly established, but it exhibits toxicity toward neuronal cells via the glutamatergic system and edema in the brain.3−5 Although there are no available reports on the effects of purified toxins in humans, by extrapolating experiments performed on small animal models, it can be estimated that its lethal dose is ∼70−100 ng.kg−1 (given intravenously).6

oxins classified as category B by the Centers for Disease Control and Prevention (CDC) are considered as potential terrorist weapons as they could be used for smallscale attacks on the basis of their availability, ease of preparation, high toxicity, and/or the lack of therapies or vaccines approved for human use. Ricin, a natural protein from castor bean, staphylococcal enterotoxin B, and epsilon toxin (ETX) belong to this category.1 ETX is the third most potent toxin after botulinum and tetanus toxins secreted by the Gram-positive Clostridium species. This toxin, produced by C. perf ringens classified as types B and D, belongs to the family of aerolysin pore-forming proteins and is responsible for lamb dysentery and enterotoxaemia, a fatal disease for domestic animals. After ingestion © 2012 American Chemical Society

Received: March 30, 2012 Accepted: April 30, 2012 Published: April 30, 2012 5103

dx.doi.org/10.1021/ac300880x | Anal. Chem. 2012, 84, 5103−5109

Analytical Chemistry

Article

urea, analytical grade formic acid (FA), and HPLC-grade acetonitrile (AcN) were from Sigma-Aldrich (Sigma Chemical Co., St Louis, MO, USA). Nonfat cow milk was purchased from a local market. Murine monoclonal antibody PεTX7 directed against formalin-inactivated prototoxin was provided by the Laboratoire d’Etudes et de Recherches en Immunoanalyse (CEA, Saclay, France; manuscript in preparation). Safety Considerations. Epsilon toxin is a highly toxic protein. Even though low toxin amounts were needed for this work, it must be handled with strict adherence to safety rules for the handling of toxic substances. All contact with the substance should be avoided. Prototoxin Purification and Activation. The prototoxin was purified from an overnight culture of C. perf ringens type D strain NCTC2062 as previously described.28 It was further activated by the action of trypsin (enzyme/substrate ratio of 1/ 100) for 30 min at room temperature, leading to the fully active toxin. Amino benzamide coated on agarose was added to inhibit and remove protease. The purity of both prototoxin and toxin was estimated to be >90% by sodium dodecyl sulfatepolyacrylamide gel electrophoresis (SDS-PAGE). The concentration was determined using the BCA protein assay kit (Thermo Scientific, San Jose, CA). Sample Preparation. PεTX7 antiepsilon prototoxin antibodies were immobilized on protein G-coated magnetic beads (Dynal Biotech, Oslo, Norway) as previously described.18 IgGcoated magnetic beads (10 μL) were mixed with 500 μL of sample in an Eppendorf Protein LoBind Microcentrifuge Tube (Westbury, NY, USA) and incubated at 37 °C for 15 min. The beads were recovered using a magnetic particle concentrator for centrifuge tubes (Dynal Biotech) and washed three times with 500 μL of HEPES buffer pH 7.4 (50 mM) to remove nonspecifically bound species. After removing the washing solution, the beads were resuspended in 30 μL of 400 mM ammonium bicarbonate buffer pH 7.5 containing 0.05% of RapiGest SF Surfactant (Waters Corp., Milford, MA) and incubated at 95 °C for 10 min in a dry heating block. A total of 5 μL of 100 μg·mL−1 aqueous trypsin solution was added, and the enzymatic digestion was performed at 40 °C for 120 min. After digestion, magnetic beads were removed and 40 μL of H2O/AcN/FA (95:5:0.1 v/v/v) was added along with 12.5 μL of each of the labeled synthetic peptides (400 ng·mL−1 in H2O/ AcN/FA 95:5:0.1 v/v/v) for a final concentration of 50 ng·mL−1. Characterization of Epsilon Prototoxin and Toxin. Top-down and bottom-up experiments were performed as detailed in the supplementary data. Briefly, electrospray intact mass measurement and collision-induced dissociation (CID) experiments were performed on a quadrupole time-of-flight (QTOF) mass spectrometer (maXis, Bruker Daltonik, Bremen, Germany). MS/MS data were submitted to Biotools software and further inspected manually. Epsilon prototoxin and toxin tryptic digests were also analyzed by UHPLC-MS/MS in the data-dependent acquisition mode using an LTQ-Orbitrap instrument fitted with an Accela UHPLC system (Thermo Scientific, San Jose, CA). Data analysis was performed using the Sequest algorithm included in the Bioworks Browser (Version 3.3.1 SP1, Thermo). Further details are presented in the supplementary data, Supporting Information. UPLC-MS/MS Quantification. Targeted UPLC-MS/MS experiments were performed using an Acquity UPLC system coupled to a triple-quadrupole Xevo TQ MS (Waters Corp.).

ETX is synthesized as a poorly active single-chain protein called prototoxin (MW 32 980 Da).7 The prototoxin is activated by the action of proteases such as trypsin or αchymotrypsin either produced by C. perf ringens or present in the animal or human digestive tract. Experiments performed on recombinant ETX produced by Escherichia coli strains demonstrate that the activation of the prototoxin is achieved following the release of the 13 N-terminal and 23 C-terminal amino acid residues, thus resulting in a significant reduction in size (28 973 Da).8 It has also been shown that the C-terminal cleavage is essential for its heptamerisation, a feature common to pore-forming toxins.9 The threat of bioterrorism has accelerated the need for fast, sensitive, and specific assays for biological agents. To date, toxin quantification methods are mainly based on immunogenic interactions (such as enzyme linked immunosorbent assay (ELISA) or surface plasmon resonance (SPR) biosensor).10−14 Immunoassays commonly exhibit high sensitivity, while particular formats such as immunochromatographic strips can generate results in less than 1 h.15 However, they may suffer from cross-reactivity of the antibodies, especially when complex matrixes are to be analyzed. Such potential interference may lead both to matrix effects and selectivity issues. In particular, ELISA-based detection methods were reported for ETX, but comparison with other techniques and in vivo activity testing in the mouse showed marked inconsistencies and discrepancies.16,17 Mass spectrometry has recently emerged as a particularly useful tool in the quantification of bioterrorism-related agents. Its ability to detect and quantify toxins with high sensitivity and specificity in complex matrixes has already been demonstrated for ricin,18,19 staphylococcal enterotoxin B,20,21 anthrax edema factor,22 and Bacillus anthracis spores.23 Sensitivity can even reach a level similar to that of immunoassays, by combination of immunocapture for efficient sample purification, enzymatic digestion, and targeted quantification of the resulting proteotypic peptides by tandem mass spectrometry, most of the time with triple quadrupole instruments operating in the multiple reaction monitoring (MRM) mode.23−25 Nevertheless, the time inherently required for each step of such analytical flow may limit the usefulness of these MS-based approaches, especially when the time-to-result must be as short as possible. However, strategies have recently been proposed to decrease sample digestion or chromatographic separation time.26,27 Here, we report the first application of mass spectrometry to ETX and prototoxin quantification. Bottom-up and top-down approaches were first implemented to go deeper both into the structure of ETX and its activation process from the prototoxin. On the basis of these data, proteotypic peptides shared by all the ETX forms were selected and further quantified by ultraperformance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS) with a collision-induced dissociation (CID) approach in complex matrixes such as milk or serum in less than 4 h, thanks to an accelerated protocol.

■

EXPERIMENTAL SECTION Chemicals and Reagents. Sequencing grade modified trypsin was from Promega (Promega, Madison, WI, USA). Synthetic peptides, i.e., ALLTNDTQQEQ [13C6; 15N2]K and VTINPQGNDFYINNP[13C6; 15N2]K were synthesized by Bachem (Burgdorf, Switzerland). N-2-Hydroxyethylpiperazine-N′-2-ethanesulfonic acid (HEPES), sodium acetate, bovine serum albumin (BSA), 5104

dx.doi.org/10.1021/ac300880x | Anal. Chem. 2012, 84, 5103−5109

Analytical Chemistry

Article

Chromatographic separation of enzymatically digested samples was performed on an Aeris PEPTIDE 1.7u XB-C18 column (2.1 mm × 150 mm i.d., 1.7 μM particle size) (Phenomenex, Sydney, NSW, Australia). Mobile phase A consisted of 0.1% FA in ultrapure water and mobile phase B, of AcN containing 0.1% FA. After injection of 25 μL of sample, a gradient consisting of 95% phase A for 1 min, followed by a linear gradient to 60% phase B in 6 min at 600 μL.min−1 was applied. The column effluent was directly introduced into the electrospray source of the mass spectrometer, and analyses were performed in the positive ion mode. Source parameters were as follows: droplet evaporation temperature, 400 °C; desolvation and cone gas flow rates, (L/h) 800 and 50; capillary voltage, 3.5 kV; and cone voltage, 35 V. Mass resolution parameters, LM 1, HM 1, LM 2, and HM 2, were fixed at 2.5, 14, 2.5, and 13, respectively (arbitrary units, i.e., 3 Th fwhm at m/z 700 for a doubly charged ion). To improve detection specificity, two MRM transitions were monitored for each peptide (see MRM transitions and optimized conditions in Table 2).

considering the release of the 13 N-terminal residues and 23 Cterminal residues, and by matching the experimental mass values with the theoretical ones, obtained from known amino acid sequences. These assignments are summarized in Table 1. Thanks to the high resolving power and high mass accuracy provided by the instrument, average mass deviations below 13 and 4 ppm were obtained after external and internal calibration, respectively. Before activation by trypsin (Figure 1A), major species were found to result from N-terminal cleavages, with mainly the loss of 6 residues (loss of KEISNT; resulting mass 32306.6 Da) and to a lesser extent 13 amino acid residues (−KEISNTVSNEMSK; resulting mass 31531.0 Da). The intact prototoxin (32979.4 Da) unexpectedly only accounted for a minority part of the signal. In addition, their activated counterparts cleaved at their C-terminal side were also detected with rather intense corresponding signals (loss of 23 residues; resulting masses 30421.3 and 29748.7 Da) and are probably caused by natural or artifactual proteolysis during the production/purification steps which demonstrates the unstable nature of the prototoxin. After trypsin treatment (Figure 1B), all the detected forms had lost their 22 or 23 C-terminal amino acid residues with cleavages occurring after either K305 or K306. Only partial N-terminal cleavage was obtained, even if the fully activated and cleaved forms were observed as the most intense peaks at 28972.6 and 29100.5 Da. Altogether, these data suggest (i) that a more efficient and/or specific C-terminal cleavage may occur in vivo, since partial cleavage after K305 was only observed after trypsin treatment and not under more “natural” conditions, and (ii) the existence of an enzyme that also specifically releases the N-terminal KEISNT peptide. These observations are in good agreement with data from the literature and provide more insights into the ETX structure and activation process. The C-terminus cleavage, in contrast with the N-terminus one, is suspected to be essential for the activity of ETX. For instance, it has been observed that the LD50 values of the toxin cleaved at its C-terminal side with and without the N-terminal truncations are very close (380 and 490 ng·kg−1 of toxin per body weight in mice, respectively), while species still containing the C-terminal peptide, cleaved or not at their Nterminal side, show very low lethality (31 000 ng·kg−1 of toxin per body weight in mice for both).8 Results obtained by mass measurement of intact protein species were further confirmed by performing top-down experiments on epsilon prototoxin and toxin. Precursor ion selection needs a large isolation window of 100 Th to improve detection of specific fragment ions from minor protein forms and covered charge states +31, +32, and +33 for ETX and +33, +34, and +35 for the prototoxin. Table S1 (Supporting Information) shows b- and y-type fragment ions observed for the different prototoxin and toxin species, along with their respective mass deviation. Although CID of intact protein often yields less complete structural information than ETD (electron transfer dissociation),29,30 enough data were obtained to identify unambiguously the amino acid sequence of each species. In total, we report 71 distinct b-type and 42 y-type ions for the prototoxin and 42 b-type and 23 y-type ions for the toxin, with an average mass deviation better than 5 ppm. It should be mentioned that, as expected, most of these cleavages occur in the neighborhood of proline and/or acidic residues. Nand C-terminal cleavages observed in MS experiments were also successfully confirmed using this top-down approach (Table S1, Supporting Information). As an example, Figure S1 (Supporting Information) depicts the CID spectrum and

■

RESULTS AND DISCUSSION ETX is synthesized as a weakly active prototoxin, activated by the action of proteases either produced by C. perf ringens or already present in the host animal or human digestive tract. A comprehensive characterization of prototoxin and toxin structures and cleavages occurring upon protease activation was first performed by top-down and bottom-up experiments to drive selection of the most pertinent and reliable proteotypic peptides for further toxin quantification.8 Characterization of Epsilon Prototoxin and Toxin. Intact Protein Mass Measurement and Top-Down Experiments. ETX was purified as its inactive prototoxin form, which was further activated by incubation with trypsin. Electrospray ionization time-of-flight mass spectra of these two species were recorded in the positive ion mode and showed broad chargestate distributions ranging from ∼+20 to ∼+40 (Figure 1). Deconvolution of these signals toward neutral protein masses revealed the presence of numerous protein species. Identification of the most abundant forms was first done according to the protein truncations expected for trypsin treatment, i.e.,8

Figure 1. Deconvoluted ESI mass spectra recorded in the positive ion mode of epsilon prototoxin (A) before and (B) after trypsin activation. Raw ESI mass spectra are shown in the insets. 5105

dx.doi.org/10.1021/ac300880x | Anal. Chem. 2012, 84, 5103−5109

Analytical Chemistry

Article

Table 1. Amino Acid Sequence Assignment of Epsilon Prototoxin and Toxin Forms Including ppm Deviation from Top-down Experiments Performed with a Quadrupole Time-of-Flight (Q-TOF) Mass Spectrometera sequence

theoretical mass (Da)

error (ppm)b

32979.43 32306.56 31531.04 30421.28 29748.71

32979.56 32306.83 31530.96 30421.62 29748.89

3.9 8.3 2.6 10.9 5.9

30549.40 30421.42 29876.69 29748.56 29100.51 28972.62

30549.79 30421.62 29877.06 29748.89 29101.19 28973.01

12.8 6.4 12.4 10.8 23.3 13.5

deconvoluted observed mass (Da)

epsilon prototoxin KEISNTVSNEMSKKASYDN [...] VIPVDKKEKSNDSNIVKYRSLSIKAPGIK VSNEMSKKASYDN [...] VIPVDKKEKSNDSNIVKYRSLSIKAPGIK KASYDN [...] VIPVDKKEKSNDSNIVKYRSLSIKAPGIK KEISNTVSNEMSKKASYDN [...] VIPVDK VSNEMSKKASYDN [...] VIPVDK epsilon toxin KEISNTVSNEMSKKASYDN [...] VIPVDKK KEISNTVSNEMSKKASYDN [...] VIPVDK VSNEMSKKASYDN [...] VIPVDKK VSNEMSKKASYDN [...] VIPVDK KASYDN [...] VIPVDKK KASYDN [...] VIPVDK a

Theoretical and experimental deconvoluted masses are average masses. Theoretical average masses were calculated on the basis of amino acid sequences. bMass calibration was done externally. After internal mass calibration, the average deviation was ∼4 ppm for epsilon toxin and prototoxin.

Table 2. Mass Spectrometer Parameters for MRM Quantification of Epsilon Toxin and Prototoxin peptide

average theoretical mass (m/z)

ALLTNDTQQEQK

694.8 (2+)

ALLTNDTQQEQ[13C6; 15N2]K

698.7 (2+)

VTINPQGNDFYINNPK

917.5 (2+)

VTINPQGNDFYINNP[13C6; 15N2]K

921.5 (2+)

observed precursor ion (m/z)a 695.1 695.1 699.0 699.0 917.5 917.5 921.5 921.5

(2+) (2+) (2+) (2+) (2+) (2+) (2+) (2+)

observed product ion (m/z) 990.3 1091.4 998.3 1099.3 703.7 1181.3 707.5 1189.7

1+

(y8 ) (y91+) (y81+) (y91+) (y122+) (y101+) (y122+) (y101+)

collision energy Elab (eV)

Emcb

24 24 24 24 30 30 30 30

0.47 0.47 0.47 0.47 0.45 0.45 0.45 0.45

a

Cone voltage was set at 30 V. bEmc = Elab [Mt/(Mt + Mi)] with Elab, the collision energy (eV), Mt, the mass of the collision gas (i.e., nitrogen), and Mi, the mass of the ion.

become activated by the host proteases.31 Tryptic peptides containing residues that could undergo chemical modifications during sample handling (i.e., C, W, M) were avoided. Finally, on the basis of top-down and bottom-up results, two peptides common to all protein species were chosen for a broader detection (Table 2): VTINPQGNDFYINNPK (amino acid residues from 85 to 100) and ALLTNDTQQEQK (amino acid residues from 120 to 131). These peptides were detected with high sensitivity as [M + 2H]2+ ions, dissociated easily via CID (Figure S2, Supporting Information) and have good chromatographic behavior. The occurrence of deamidation was also checked, and despite N/Q residues in both peptides, no deamidated peptide was observed. To improve specificity further, two MRM transitions for each peptide were monitored. Their sequences were further submitted to BLAST similarity searches to ensure specificity toward ETX. Quantification Strategy and Method Development. The focus of this work was to develop an immuno-UPLC-MSMS approach for the fast and specific quantification of ETX at the low ng·mL−1 level in complex samples. All the method development was performed on the active toxin ETX, since the prototoxin could be readily activated in vivo or in a food matrix such as milk which contains high levels of proteases.31,32 The first step consists of the immunocapture of the protein using magnetic beads coated with antibodies directed against ETX and prototoxin. This approach has two advantages: the protein is specifically isolated and purified from complex samples and is also easily concentrated. An approximately 16fold increase in sensitivity was obtained thanks to sample

sequence coverage obtained for the prototoxin form at 32 306.6 Da. Bottom-up Experiments. Tryptic peptides identified from the bottom-up experiments performed on prototoxin and toxin are presented in Table S2 (Supporting Information). Under these conditions, sequence coverages of 93% for the prototoxin (23 peptides out of 25 expected with a mass above 500 Da and excluding the signal peptide) and 89% for the toxin (21 peptides out of 25) were obtained considering the Uniprot sequence entry Q57398. In both experiments, the N-terminal peptide was detected (i.e., tryptic peptide 34−45), which is consistent with the top-down results. It should be noted that the C-terminal tryptic peptides were still detected in the toxin digest (tryptic peptides 319−323 and 324−328), since peptides released after activation of the prototoxin were not removed before analysis. While intact protein mass measurement and top-down MS/MS experiments rapidly identify proteins and their potential modifications, analysis of epsilon prototoxin and toxin tryptic digests provides complementary results, giving access to the most intense proteotypic peptides suitable for protein quantification. Development of the Quantitative MRM Assay. Selection of the Proteotypic Peptides. C-terminal peptides would, if selected for the quantification assay, differentiate the inactive epsilon prototoxin from the active toxin. However, they either are too short for enough quantification specificity or show poor chromatographic behavior leading to low detection sensitivity. Moreover, all forms should preferentially be quantified since all either are already active or may readily 5106

dx.doi.org/10.1021/ac300880x | Anal. Chem. 2012, 84, 5103−5109

Analytical Chemistry

Article

concentration (starting from 500 μL and concentrating down to 30 μL before digestion). Enzymatic digestion was then performed, and the two released proteotypic peptides were monitored by UPLC-MS/MS in the MRM mode. The use of an internal standard is required for an accurate and reliable mass spectrometry-based quantification. To control all steps involved in sample preparation, addition of a labeled protein at the beginning would be the best way.33 However, addition of AQUA (absolute quantification) peptides is an alternative when the labeled protein is not available. Isotopically labeled counterparts (incorporating [13C6; 15N2] labeled Lys residues) of the two peptides selected for the quantification were thus added prior to UPLC-MS/MS analysis, and a calibration curve was constructed for each sample set to check the immunocapture and digestion yields. Optimization of the Immunocapture Step. In the case of a bioterrorism event, it is essential to ascertain rapidly whether suspicious samples are really contaminated. Therefore, increasing sample throughput while maintaining sensitivity and robustness is necessary. Immunocapture can be seen as time-consuming, despite possible technical improvement. For instance, the use of magnetic beads simplifies separation of the captured analyte from the biological matrix. Incubation time for analyte capture by the antibodies, which was fixed at 1 h at ambient temperature for other applications in our group,18 could be optimized and potentially shortened. In a recent study where the rapid detection of microcystins and nodularins by fluorescent immunochromatography is described, good sensitivity and reproducibility were obtained after a contact time of only 10 min at 20 °C.15 On the basis of this observation, milk samples spiked with ETX at 100 ng·mL−1 and 700 ng·mL−1 were incubated from 1 min to 1 h with Ig-G coated magnetic beads. To enhance the immunocapture efficiency, samples were incubated at 37 °C instead of room temperature. Captured proteins were digested using a conventional protocol (Figure S3, protocol 2, Supporting Information), before addition of internal standards as described in the Experimental Section. Figure 2A shows the results obtained for the MRM transition m/z 917.5 to m/z 703.7 of the peptide VTINPQGNDFYINNPK at 100 ng·mL −1 (dotted line) and 700 ng·mL−1(solid line) in milk (mean of two independent measurements). Below 5 min of incubation, the immunocapture was not efficient. Maximum yields for the target peptide were obtained after 5 min of incubation at 37 °C for both concentrations and were stable throughout the incubation period. To maintain assay robustness, the incubation time was fixed at 15 min. Similar results were obtained for the peptide ALLTNDTQQEQK (data not shown). Protein Denaturation and Digestion. The protein denaturation method, prior to tryptic digestion, has also to be carefully optimized. As ETX is activated by the action of proteases, it can be supposed that it is quite resistant to enzymatic digestion. Norrgran et al. have recently demonstrated that the use of a surfactant for protein denaturation has a decisive influence on tryptic digestion efficiency.26 They observed that, using RapiGest SF as denaturation agent, a tryptic digestion performed at 40 °C, and an enzyme-to-substrate ratio of 3, the digestion yields for ricin, another protein toxin, were ∼3fold higher than with other methods. To confirm these results on ETX, different denaturation methods, prior to overnight incubation with trypsin, were investigated, i.e., assisted by solvent (60% AcN), with urea (8 M) or using dithiothreitol followed by iodoacetamide. Similarly, parameters such as

Figure 2. Optimization of (A) immunocapture and (B) digestion times. (A) Peak area ratio of the MRM transition m/z 917.5 to m/z 703.7 of the peptide VTINPQGNDFYINNPK with the corresponding internal standard as a function of the immunocapture time for milk samples spiked with 100 ng·mL−1 (dotted line) and 700 ng·mL−1 (solid line) of epsilon toxin. (B) Peak area ratio of the MRM transition m/z 917.5 to m/z 703.7 of the peptide VTINPQGNDFYINNPK (solid line) and the MRM transition m/z 695.1 to m/z 990.3 of the peptide ALLTNDTQQEQK (dotted line) with their respective internal standards after various times of digestion of milk spiked with ETX at 700 ng·mL−1. Bars indicate the standard deviation of two independent determinations.

proteolysis temperature and enzyme-to-substrate ratio that provide optimal digestion efficiency were selected (40 °C and a ratio E/S of 3 at 500 ng·mL−1 of substrate, respectively). Best results were obtained using 10 min denaturation with 0.05% RapiGest SF at 95 °C, i.e., up to 3 times more efficient compared with the protocol using only ammonium bicarbonate (Protocol 2, Figure S3, Supporting Information). Since the whole process has to be accelerated, the minimum digestion time needed for the release of the two peptides was also evaluated. After immunocapture of milk spiked with ETX at 700 ng·mL−1 and a 10 min incubation at 95 °C with 0.05% RapiGest SF, various digestion times ranging from 30 min to overnight were tested (Figure 2B). After a 30 min incubation with trypsin, the maximum observed yield was obtained for the peptide ALLTNDTQQEQK, while 2 h were necessary for the peptide VTINPQGNDFYINNPK. This difference may be explained through the structural representation of ETX (PDB ID: 1UYJ chain C); both peptides are included in β-barrels, but the cleavage sites of the peptide ALLTNDTQQEQK seem to be more accessible.34 An incubation time of 2 h at 40 °C was selected. The digestion recovery of ETX and prototoxin was finally determined. A known quantity of proteins was digested, and the same molar amount of internal standard was added at the end of the reaction. By comparing areas of peptides released after 2 h of proteolysis and the corresponding internal standard, 5107

dx.doi.org/10.1021/ac300880x | Anal. Chem. 2012, 84, 5103−5109

Analytical Chemistry

Article

The reproducibility of the assay was investigated by repeatedly analyzing five independent milk samples spiked with ETX at 70 ng·mL−1 and 700 ng·mL−1. The mean CVs were 12.8% and 5.7% for the lower and the higher concentration, respectively. The accuracy was also evaluated and was estimated to be 99.7% ± 9.3% (Table S3, Supporting Information) for milk spiked with three concentrations of the toxin (18, 91, and 364 ng·mL−1). Altogether, these results demonstrate that ETX can be accurately measured using the developed approach. As mentioned previously, prototoxin proved to be rather sensitive to proteases present in food and biological samples, thus yielding fully active ETX. Consequently, one could expect that the samples to be analyzed would contain far lesser amounts of prototoxin than toxin. Nevertheless, the ability of the method to quantify the prototoxin was investigated. As previously performed for the toxin, the whole process was applied to milk samples spiked with three different concentrations of epsilon prototoxin (i.e., 19, 96, and 385 ng·mL−1). Results showed that the mean accuracy (using a toxin calibration curve) was 73.7% ± 2.6% (Table S3, Supporting Information). Such a poor accuracy is probably due to a lower affinity of the antibodies for the prototoxin. However, the limit of detection was similar to the one of the toxin (i.e., ∼5 ng.mL−1), which still allows the sensitive detection of this inactive form. Last, but not least, it should be mentioned that similar results in terms of detection sensitivity and linearity were obtained for the quantification of epsilon toxin in serum (Figure S6, Supporting Information).

global recoveries of peptides VTINPQGNDFYINNPK and ALLTNDTQQEQK were estimated to 90% and 65% for epsilon prototoxin and toxin, respectively. Optimization of the Chromatographic Separation. The analytical throughput can be increased using UPLC rather than conventional HPLC separation.27 Our UPLC method ensures a baseline separation for both peptides in 10 min, including column washing and reequilibration (Figure S4, Supporting Information), instead of 27 min using more conventional HPLC conditions. The first peptide (2.29 min; ALLTNDTQQEQK) was eluted in 6 s and the second (3.43 min; VTINPQGNDFYINNPK) in 10 s, while both peptides were eluted in 25 s by HPLC. Scan time was further adjusted to obtain at least 15 points per chromatographic peak for each MRM transition (i.e., 50 ms per transition). Finally, using optimized protocols for immunocapture and enzymatic digestion, in line with detection by UPLC-MS/MS, analysis of ETX can be achieved in less than 3 h and the whole quantification method including the calibration curve in less than 4 h. Quantification of Epsilon Prototoxin and Toxin in Complex Matrixes. To validate the method for the quantification of ETX in complex matrixes, different quantities of the toxin ranging from 5 to 1800 ng·mL−1 were spiked into milk (Figure 3) for further estimation of the method

■

CONCLUSION

■

ASSOCIATED CONTENT

We have developed for the first time a quantitative mass spectrometric analysis of all C. perf ringens ETX forms in complex samples (milk and serum). These different toxin forms were first thoroughly characterized by a combination of topdown and bottom-up experiments. The quantitative assay provides a rapid and sensitive detection tool (i.e., down to 5 ng·mL−1 in less than 4 h) for assessing the contamination level of environmental matrixes, which is of particular importance in the case of chemical, biological, radiological, and nuclear threat. The method can now be implemented in regulatory laboratories to complement results obtained in the field, using, for example, immuno-chromatographic strips. Our assay relies on antibody availability and affinity. Nevertheless, such selective extraction is necessary to reduce interferences from coeluting species in complex matrixes. Other selective capture agents could be evaluated such as aptamers,35 which are readily produced by chemical synthesis. Another improvement would be to provide an additional dimension of separation with, for instance, ion mobility,36 thus increasing the assay’s dynamic range and possibly avoiding the capture step.

Figure 3. Calibration curve obtained for epsilon toxin spiked in milk with concentrations ranging from 5 ng·mL−1 to 1800 ng·mL−1. Calibration was done by fitting a linear regression of the area of the MRM transition m/z 917.5 to m/z 703.7 for peptide VTINPQGNDFYINNPK, with the corresponding internal standard, versus the epsilon toxin concentration. Inset shows a focus on the low concentration range. Two independent replicate analyses were performed for each concentration. Red triangles correspond to quality control samples at 70 ng·mL−1 and 700 ng·mL−1 (n = 5).

characteristics. The limit of quantification was estimated at ∼5 ng·mL−1 (with a signal-to-noise ratio of 8) and the limit of detection at ∼3 ng·mL−1 (signal-to-noise ratio of 3). The coefficient of correlation also indicated good linearity (R2 = 0.9856) throughout the whole concentration range. As an example, chromatograms of two MRM transitions, corresponding to the two targeted peptides at the limit of quantification, is presented in Figure S5, Supporting Information. The ratio of both peptide areas was also monitored throughout the concentration range and proved to be stable at 1.5 ± 0.2, thus indicating the robustness of the accelerated digestion.

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: 33-1-69-08-13-15. Fax: 33-1-69-08-59-07. E-mail: [email protected]. 5108

dx.doi.org/10.1021/ac300880x | Anal. Chem. 2012, 84, 5103−5109

Analytical Chemistry

Article

Notes

(29) Tsybin, Y. O.; Fornelli, L.; Stoermer, C.; Luebeck, M.; Parra, J.; Nallet, S.; Wurm, F. M.; Hartmer, R. Anal. Chem. 2011, 83, 8919− 8927. (30) Kim, M. S.; Pandey, A. Proteomics 2012, 12, 530−542. (31) Bokori-Brown, M.; Savva, C. G.; Fernandes da Costa, S. P.; Naylor, C. E.; Basak, A. K.; Titball, R. W. FEBS J. 2011, 278, 4589− 4601. (32) D’Alessandro, A.; Scaloni, A.; Zolla, L. J. Proteome Res. 2010, 9, 3339−3373. (33) Brun, V.; Dupuis, A.; Adrait, A.; Marcellin, M.; Thomas, D.; Court, M.; Vandenesch, F.; Garin, J. Mol. Cell. Proteomics 2007, 12, 2139−2149. (34) Cole, A. R.; Gilbert, M.; Popoff, M.; Moss, D. S.; Titball, R. W.; Basak, A. K. Nat. Struct. Mol. Biol. 2004, 11, 797−798. (35) Bunka, D. H. J.; Stockley, P. G. Nat. Rev. Microbiol. 2006, 4, 588−596. (36) Swearingen, K. E.; Hoopmann, M. R.; Johnson, R. S.; Saleem, R. A.; Aitchison, J. D.; Moritz, R. L. Mol. Cell. Proteomics 2011, 11:M111.014985 .

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We are greatly indebted to Patricia Lamourette, Karine Moreau, and Marc Plaisance for production and provision of monoclonal antibodies.

■

REFERENCES

(1) http://www.bt.cdc.gov/agent/agentlist-category.asp. (2) Popoff, M. R. FEBS J. 2011, 278, 4602−4615. (3) Miyamoto, O.; Sumitani, K.; Nakamura, T.; Yamagami, S. I.; Miyata, S.; Itano, S.; Negi, T.; Okabe, A. Infect. Immun. 1998, 66, 2501−2508. (4) Miyamoto, O.; Sumitani, K.; Nakamura, T.; Yamagami, S. I.; Miyata, S.; Itano, T.; Negi, T.; Okabe, A. FEMS Microbiol. Lett. 2000, 189, 109−113. (5) Finnie, J. W.; Blumbergs, P. C.; Manavis, J. J. Comp. Pathol. 1999, 120, 415−420. (6) Paddle, B. M. J. Appl. Toxicol. 2003, 23, 139−170. (7) Hunter, S. E.; Clarke, I. N.; Kelly, D. C.; Titball, R. W. Infect. Immun. 1992, 60, 102−110. (8) Miyata, S.; Matsushita, O.; Minami, J.; Katayama, S.; Shimamoto, S.; Okabe, A. J. Biol. Chem. 2001, 276, 13778−13783. (9) Minami, J.; Katayama, S.; Matsushita, O.; Matsushita, C.; Okabe, A. Microbiol. Immunol. 1997, 41, 527−535. (10) Griffiths, G. D.; Newman, H.; Gee, D. J. J. Forensic Sci. Soc. 1986, 26, 349−358. (11) Shyu, H. F.; Chiao, D. J.; Liu, H.; Tang, S. S. Hybridoma Hybridomics 2002, 21, 69−73. (12) Soelberg, S. D.; Stevens, R. C.; Limaye, A. P.; Furlong, C. E. Anal. Chem. 2009, 81, 2357−2363. (13) Sasaki, T.; Terano, Y.; Shibata, T.; Kawamoto, H.; Kuzuguchi, T.; Kohyama, E.; Watanabe, T.; Ohyama, T.; Gemba, M. Microbiol. Immunol. 2005, 49, 589−597. (14) 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. (15) Khreich, N.; Lamourette, P.; Lagoutte, B.; Ronco, C.; Franck, X.; Créminon, C.; Volland, H. Anal. Bioanal. Chem. 2010, 397, 1733− 1742. (16) Uzal, F. A.; Kelly, K. R.; Thomas, R.; Hornitzky, M.; Galea, F. J. Vet. Diagn. Invest. 2003, 15, 94−99. (17) Layana, J. E.; Fernandez Miyakawa, M. E.; Uzal, F. A. Anaerobe 2006, 12, 204−206. (18) Duriez, E.; Fenaille, F.; Tabet, J. C.; Lamourette, P.; Hilaire, D.; Becher, F.; Ezan, E. J. Proteome. Res. 2008, 7, 4154−4163. (19) McGrath, S. C.; Schieltz, D. M.; McWilliams, L. G.; Pirkle, J. L.; Barr, J. R. Anal. Chem. 2011, 83, 2897−2905. (20) Callahan, J. H.; Shefcheck, K. J.; Williams, T. L.; Musser, S. M. Anal. Chem. 2006, 78, 1789−1800. (21) Dupuis, A.; Hennekinne, J. A.; Garin, J.; Brun, V. Proteomics 2008, 8, 4633−4636. (22) Duriez, E.; Goossens, P. L.; Becher, F.; Ezan, E. Anal. Chem. 2009, 81, 5935−5941. (23) Chenau, J.; Fenaille, F.; Ezan, E.; Morel, N.; Lamourette, P.; Gossens, P. L.; Becher, F. Anal. Chem. 2011, 83, 8675−8682. (24) Dubois, M.; Becher, F.; Herbert, A.; Ezan, E. Rapid Commun. Mass Spectrom. 2007, 27, 352−358. (25) Gallien, S.; Duriez, E.; Domon, B. J. Mass Spectrom. 2011, 46, 298−312. (26) Norrgran, J.; Williams, T. L.; Woolfitt, A. R.; Solano, M. I.; Pirkle, J. L.; Barr, J. R. Anal. Biochem. 2009, 393, 48−55. (27) Luna, L. G.; Williams, T. L.; Pirkle, J. L.; Barr, J. R. Anal. Chem. 2008, 80, 2688−2693. (28) Petit, L.; Gibert, M.; Gillet, D.; Laurent-Winter, C.; Boquet, P.; Popoff, M. R. J. Bacteriol. 1997, 179, 6480−6487. 5109

dx.doi.org/10.1021/ac300880x | Anal. Chem. 2012, 84, 5103−5109