Detection of Ricin in Complex Samples by Immunocapture and Matrix-Assisted Laser Desorption/Ionization Time-of-Flight Mass Spectrometry Elodie Duriez,† Franc¸ois Fenaille,† Jean-Claude Tabet,§ Patricia Lamourette,† Didier Hilaire,‡ Franc¸ois Becher,† and Eric Ezan*,† CEA, Service de Pharmacologie et d’Immunoanalyse, 91191 Gif-sur-Yvette, France, Laboratoire de Synthe`se, Structure et Fonction de Mole´cules Bioactives, CNRS UMR 7613, Universite´ Pierre et Marie Curie, 75252 Paris, France, and Centre d’Etudes du Bouchet, 91710 Vert Le Petit, France Received May 5, 2008
Ricin, the toxin component of Ricinus communis is considered as a potential chemical weapon. Several complementary techniques are required to confirm its presence in environmental samples. Here, we report a method combining immunocapture and analysis by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) for the accurate detection of different species of R. communis. Liquid environmental samples were applied to magnetic particles coated with a monoclonal antibody directed against the B-chain of the toxin. After acidic elution, tryptic peptides of the A- and B-chains were obtained by accelerated digestion with trypsin in the presence of acetonitrile. Of the 20 peptides observed by MALDI-TOF MS, three were chosen for detection (m/z 1013.6, m/z 1310.6 and m/z 1728.9, which correspond to peptides 161-LEQLAGNLR-169, 150-YTFAFGGNYDR-160, and 233SAPDPSVITLENSWGR-248, respectively). Their selection was based on several parameters such as detection sensitivity, specificity toward ricin forms and absence of isotopic overlap with unrelated peptides. To increase assay reproducibility, stable isotope-labeled peptides were incorporated during the sample preparation phase. The final assay has a limit of detection estimated at ∼50 ng/mL (∼0.8 nM) of ricin in buffer. No interference was observed when the assay was applied to ricin-spiked milk samples. In addition, several varieties of R. communis or from different geographical origins were also shown to be detectable. The present assay provides a new tool with a total analytical time of ∼5 h, which is particularly relevant in the context of a bioterrorist incident. Keywords: ricin • proteolysis • bioterrorism • mass spectrometry • antibody
Introduction Ricin has been considered as a chemical weapon since the beginning of the 20th century. Several historical events have confirmed the potential bioterrorist use of ricin, like the assassination of a Bulgarian journalist in 1998 and the discovery of ricin in a mailroom serving a U.S. Senator’s office in 2003.1 Ricin can be easily obtained from the seeds of the castor bean plant Ricinus communis, which is widely grown in subtropical and temperate regions.2 Reported ricin content of castor beans varies but is in the 1-5% (w/w) range,1 which represents high amounts of easy-to-obtain material. It has been estimated that in man, the lethal dose by inhalation (breathing in solid or liquid particles) and injection (into muscle or vein) is approximately 5-10 µg/kg, that is, 350-700 µg for a 70 kg adult.3 * To whom correspondence should be addressed: Phone, 33-1-69-08-7350; fax, 33-1-69-08-59-07; e-mail,
[email protected]; mailing address, Service de Pharmacologie et d’Immunoanalyse, Baˆt 136, CE Saclay, 91191 Gif-surYvette, France. † CEA, Service de Pharmacologie et d’Immunoanalyse. § Universite´ Pierre et Marie Curie. ‡ Centre d’Etudes du Bouchet.
4154 Journal of Proteome Research 2008, 7, 4154–4163 Published on Web 07/24/2008
For these reasons, ricin is included in Schedule 1 of the Chemical Weapons Convention (CWC).4 The ricin toxin obtained from the seeds of the castor bean plant is an ∼62 kDa glycoprotein, consisting of two disulfidelinked polypeptide chains (almost 30 kDa each) with different functions. The ricin A-chain is a highly efficient N-glycosidase and the ricin B-chain is a lectin, which promotes internalization in target cells. Once inside the cell, the A-chain acts as a ribosome-inactivating protein (RIP) by releasing a specific adenine base of the essential 28S rRNA and, thus, inhibits protein biosynthesis.5,6 Unambiguous identification of ricin requires evidence for the presence of these two chains and distinction from other proteins in complex matrices. As a consequence of its easy availability and high toxicity, many laboratories have studied the detection of ricin, through the development of analytical techniques to identify the toxin at low concentrations. So far, these approaches have mainly comprised immunoassays, which are the most sensitive detection techniques,7,8 activity testing with ribosomes,9 and assays involving the measurement of adenine release from synthetic substrates.10,11 However, these assays suffer from several 10.1021/pr8003437 CCC: $40.75
2008 American Chemical Society
research articles
Ricin Detection by Immunocapture and MALDI-TOF MS
Figure 1. Flowchart representation of the methodology developed for ricin detection.
drawbacks, resulting either from potential cross-reactivity of the antibodies used or the presence of other ribosomeinactivating proteins. To reach a higher degree of specificity, we have recently developed an analytical method involving the immunocapture of ricin by a monoclonal antibody directed against the B-chain, followed by the accurate quantification of the adenine released by the A-chain from an RNA template by liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS).12 This assay exhibited a limit of detection of 0.1 ng/mL (1.6 pM) of functional ricin.12 Nevertheless, despite improved specificity due to the “recognition” of the two ricin chains, a potential false-positive result is still possible through the nonspecific extraction of other RIPs. This is especially true when working on complex environmental samples. Advances in biological mass spectrometry now allow sensitive detection, accurate mass measurement, and structural analysis of a wide range of biomolecules. Thus, mass spec-
trometry techniques involving either electrospray ionization (ESI) or matrix-assisted laser desorption/ionization (MALDI) are now being increasingly developed for the detection and absolute quantification of proteins in biological environments. Previously published studies of ricin have demonstrated that the toxin can be unambiguously identified using either MALDI time-of-flight (TOF) MS or LC-ESIMS.13–16 However, these different methodologies focused only on the identification and/ or characterization of ricin from crude seed extracts, and none addressed the quantification and limit of detection of the toxin in complex environmental or food samples to highlight potential bioterrorist attacks. MALDI-TOF MS has several advantages compared with other ionization techniques, including a very high sensitivity (low femtomole range), ease of operation, and especially speed of analysis. Furthermore, this technique is very tolerant to sample contaminants (e.g., salts and detergents), which often impair analysis by ESIMS. For these reasons, an increasing number of reports describe the use of MALDI-TOF mass spectrometry as a quantitative tool for proteins.17–23 To obtain more specificity and sensitivity, an immunoaffinity isolation step can be used in line with the detection by MALDI-TOF MS. This combination of techniques constitutes a powerful tool to unambiguously detect trace levels of a given protein in complex samples. Indeed, the ability of mass spectrometry to obtain amino acid sequence data on affinity-captured peptides provides absolute specificity and avoids “false positives” generated by nonspecific binding.22 For example, this methodology has been successfully applied to the quantification of several proteins such as insulin growth factor 1 in biofluids19 and staphylococcal enterotoxin B (SEB) at low-nanogram levels in biological matrices.24 Although MALDI-TOF MS analysis of whole ricin has been described in previous studies,13,25 the limit of detection is too low to envision potential direct quantification of ricin by monitoring its native form. Ricin is an ∼62 kDa protein with a high level of structural heterogeneity, mainly due to the presence of many glycoforms.26 When analyzed by MALDI-TOF MS, ricin exhibits a very broad mass spectrum peak (due to both glycosylation heterogeneity, salt adduction and metastable decomposition) which compromises both detection sensitivity and specificity. One way to overcome this limitation is to work on enzymatic digests of the protein of interest.18,20 The goal of the present work was to develop an MS-based analytical method that allows the specific and sensitive detection of ricin in complex environmental samples. The present development was accomplished by extending the mass spectrometric immunoassay concept.27 Much attention was devoted to the optimization of sample preparation and analysis steps to provide a fast-response method, which is mandatory when dealing with bioterrorism issues. To the best of our knowledge, this is the first report of accurate detection of ricin in environmental samples using a proteomics-based approach.
Experimental Procedures Materials. Purified ricin samples were from two different sources. Assay development was performed using ricin D purified from the seeds of the castor bean R.communis by Dr. B. Beaumelle at the Departement Biologie-Sante´, Universite´ Montpellier, France.28 Additional ricin samples were obtained by purification of R. communis, R. communis impala, and R. communis zanzibariensis seeds (Centre d’Etudes du Bouchet, Vert Le Petit, France). Seeds from different varieties of R. communis beans were locally purchased, whereas milk samples Journal of Proteome Research • Vol. 7, No. 9, 2008 4155
research articles
Duriez et al.
Figure 2. MALDI-TOF mass spectrum of a tryptic digest of ricin D (corresponding to ∼1 pmol on-target). Digestion was performed in 80% acetonitrile with no preliminary disulfide bond reduction. This mass spectrum was calibrated internally by using the trypsin autolysis products present after a 2 h digestion at 37 °C. Peptide assignment was performed by matching the observed and the theoretical m/z ratios; a mass accuracy of 20 ppm was observed throughout the spectrum using this procedure. No internal standard was added.
were from local supermarkets. Ricin is a highly toxic protein and consequently requires strict adherence to safety rules for the handling of toxic substances. All contact with the substance should be avoided. Monoclonal antibodies directed against the ricin B-chain were provided by the Laboratoire d’Etudes et de Recherches en Immunoanalyse (CEA Saclay, France).12 The three stable isotope-labeled peptides of the A-chain ricin peptides [13C6;15N]LEQLAGNLR (m/z 1020.6), YTFA[13C6]FGGNYDR (m/z 1316.6), and SAPDPSVIT[13C6;15N]LENSWGR (m/z 1735.9) were synthesized by Bachem (Merseyside, U.K.). HEPES, bovine serum albumin (BSA), trifluoroacetic acid (TFA), and ammonium bicarbonate were from Sigma (Saint Quentin Fallavier, France). R-Cyano-4-hydroxycinnamic acid (CHCA) and HPLCgrade acetonitrile (ACN) were from Laserbiolab (Sophia-Antipolis, France) and SDS (Peypin, France), respectively. Protein Amino Acid Sequence Information. Sequence information was obtained from the Swiss-Prot database on ricin, R. communis agglutinin and trypsin: entries P02879, P06750 and P00761, respectively. Extraction of Ricin from Castor Seeds. Different varieties of castor seeds (large- and small-grain) were extracted to analyze ricin from the crude extract. Whole castor beans were cracked to remove the outer hull. Hulled beans (1.6 g) were then vigorously mixed with 3 mL of a 50 mM HEPES buffer, pH 7.3, containing 1 mg/mL of BSA (HEPES/BSA buffer) during ∼10 min using an Ultraturrax (Ika, Staufen, Germany), until a milky slurry was obtained. The bean pulp slurry was placed in an Eppendorf tube and further delipidated by five successive 10-min centrifugations at 14 000g and 4 °C. The supernatant was removed and separated into aliquots that were stored at -20 °C. Sample Preparation. Calibration standards were obtained by serially diluting ricin stock solution from R. communis in HEPES/BSA buffer, from 50 to 800 ng/mL. Milk samples spiked with ricin, as well as bean pulp slurries, were further diluted with HEPES/BSA buffer prior to the immunocapture step. An equimolar solution of internal standards (IS) was obtained by mixing [13C6;15N]LEQLAGNLR, YTFA[13C6]FGGNYDR, and SAP4156
Journal of Proteome Research • Vol. 7, No. 9, 2008
DPSV IT[13C6;15N]LENSWGR peptides, at a final concentration of 80 nM in 80% ACN. Immunocapture and Enzymatic Digestion. Ricin antibodies were immobilized on Dynabeads protein G magnetic beads (Dynal Biotech, Oslo, Norway) as previously described.12 Five microliters of IgG-coated beads were mixed with 500 µL of sample, and further incubated for 2 h at room temperature with gentle rotation. The beads were recovered using a magnetic particle concentrator for microcentrifuge tubes (Dynal MPCS, Dynal Biotech, Oslo, Norway) and washed twice with 500 µL of HEPES buffer (50 mM, pH 7.3) to remove nonspecifically bound species, and then three times with 500 µL of ultrapure water to remove remaining traces of buffer. Ricin elution was accomplished by incubating the beads with 35 µL of 0.1% TFA for 5 min at room temperature. The supernatant was transferred into a new tube and the pH further neutralized by adding 7 µL of 400 mM ammonium bicarbonate. Trypsin (4 µL of a 0.4 µg/µL aqueous solution) and the IS mixture (10 µL) were then added. The trypsin digestion was carried out under organic-aqueous conditions by addition of 194 µL of ACN to obtain a final concentration of 80%. After a 2-h incubation at 37 °C, the peptide mixture was concentrated to ∼20 µL by centrifugal evaporation. A 10 µL aliquot of 0.1% TFA was then added prior to extraction/concentration of tryptic peptides using C18 pipet tips (ZipTip, Millipore, Saint-Quentin en Yvelines, France). Peptide elution was performed using 5 µL of 20% ACN containing 0.1% TFA. One microliter of this solution was then spotted on the MALDI target and thoroughly mixed on-target with 0.5 µL of CHCA solution (10 mg/mL in 50% ACN containing 0.1% TFA). MALDI-TOF MS Analysis. Peptide analyses were performed on a Voyager-DE STR instrument (Applied Biosystems, Les Ulis, France) equipped with a pulsed nitrogen laser (337 nm). The instrument was operated in the delayed extraction reflector mode with a 20 kV acceleration voltage and an extraction delay of 140 ns. The laser power was carefully monitored so as to be high enough to have a good signal/noise ratio, but low enough so as to avoid detector saturation and nonlinear response. The optimized laser power was set at ∼1550 arbitrary units. Data
research articles
Ricin Detection by Immunocapture and MALDI-TOF MS e
Table 1. Tryptic Peptides from Purified Ricin Identified by MALDI TOF-MS and LC-ITMS
a N-glycosylation sites are represented in bold characters. b Disulfide-linked cysteines are underlined. c The free cysteine residue of peptide (167-180) is represented on a gray background. d Major glycoforms on different sites are represented in italic. e The MALDI peaks from the mass spectrum were assigned to tryptic peptides by matching measured to expected molecular weights with a 20 ppm mass accuracy.
were accumulated over the m/z ratio window of 500-4000 Th. All samples were prepared in duplicate and spotted, and spectra comprising 500 averaged laser shots each were acquired from two different regions of each spot to give four mass spectra per sample. Mass spectra were internally calibrated with trypsin autolysis peaks at m/z 842.51 and m/z 2211.10. To obtain more robust quantitative results, the average intensity of the first three monoisotopic peaks of the ricin targeted peptides [(161-169), (150-160), and (233-248)] and their labeled counterparts was taken into account for the calculation of the light/ heavy ratio. The interspot variability was 2-5%. LC-ESIMS/MS Conditions. To confirm peptide sequences, tryptic digests of ricin were also analyzed by LC-MS/MS. Chromatographic separations were performed on a Zorbax 300SB C18 HPLC column (2.1 × 150 mm, 5 µm, 300 Å, Interchim, Montluc¸on, France) mounted on an HPLC system. Mobile phases consisted of 0.1% formic acid (A) and ACN containing 0.1% formic acid (B). Peptides were eluted from the column at 300 µL/min using a 5-60% phase B gradient over 35 min. The HPLC eluent was directly connected to the ESI probe of a LCQ-Duo ion trap mass spectrometer (Thermo, San Jose, CA). The source conditions were as follows: capillary temperature, 275 °C; sheath gas flow, 80 arbitrary units; auxiliary gas flow, 20 arbitrary units; ESI spray voltage, 4.5 kV. The target was fixed at 1 × 107 ions and the automatic gain control (AGC) was turned on. The LCQ was operated in a data-
dependent mode, in which the instrument was set up to automatically acquire full scan over the 300-2000 m/z range, followed by ZoomScan (high resolution mass spectrum) and MS/MS spectra (with a normalized collision energy of 35%29 and an isolation width of 2.0 Th) for the two most intense ions. An additional experiment was performed using in-source voltage set to 60 V, in order to generate carbohydrate-specific product ions and further specifically identify glycopeptides.30
Results and Discussion Analytical Strategy. The goal of this study was to develop a method combining an immunocapture step and a mass spectrometric analysis for fast and unambiguous detection of ricin at low-nanogram levels in complex matrices. This approach consists of affinity isolation of a target protein from a given matrix along with detection by MALDI-TOF mass spectrometry (Figure 1). This involved different steps: (i) immunoaffinity isolation of ricin using the antibody-coated magnetic beads, (ii) acidic elution from the beads, (iii) addition of labeled peptides for accurate detection, (iv) trypsin digestion under organic-aqueous conditions to shorten digestion time, (v) reduction of mixture complexity by ZipTip fractionation, and (vi) precise MALDI-TOF MS detection through the monitoring of three different peptides. In the framework of our study, particular attention was paid to optimization of the whole analytical process in order to develop a fast-response method. Journal of Proteome Research • Vol. 7, No. 9, 2008 4157
research articles
Duriez et al.
Figure 3. MALDI-TOF mass spectra of the selected peptides, obtained after immunocapture and organic-digestion of buffer solutions containing 0, 50, 100 and 400 ng/mL of ricin D. Digestion was performed in 80% ACN at 37 °C. IS peaks are labeled with an asterisk (*). IS concentrations were set so as to generate a theoretical area ratio of 1:1 between unlabeled and labeled peptides for 400 ng/mL of ricin. Such expected ratios were observed for ions at m/z 1013.6 and m/z 1310.6, whereas it was higher for the ion at m/z 1728.9, probably because of biases inherent to the solubilization of the labeled peptide.
Figure 4. MALDI-TOF mass spectra of tryptic digest of ricin. Standard ricin D purified from R. communis (a) before and (b) after immunocapture, and ricin from a crude extract of R. communis (c) before and (d) after immunocapture.
Selection of Peptides for Ricin Detection. Tryptic peptides for the development of the quantitative assay were selected according to several criteria: specificity toward ricin forms, detection sensitivity, and absence of isotopic overlap with 4158
Journal of Proteome Research • Vol. 7, No. 9, 2008
unrelated tryptic peptides. The first step was to identify specific tryptic peptides showing the most intense signal in MALDITOF mass spectra. Figure 2 depicts the MALDI-TOF mass spectrum of a tryptic digest of a concentrated ricin solution
Ricin Detection by Immunocapture and MALDI-TOF MS
research articles
Figure 5. Intensity ratio of selected ricin tryptic peptides (a) m/z 1013.6/1310.6 and (b) m/z 1013.6/1728.9 obtained from the digestion of purified ricin toxin, or ricin toxin obtained by immunocapture of the toxin from a crude extract. Dark bars represent ricin from crude plant extracts and white bars represent purified ricin samples (n ) 2).
Figure 6. Intensity ratios m/z 1013.6/m/z 1020.6 obtained for immunopurified ricin samples of different origins and cultivars. Dark bars represent ricin from crude plant extracts and white bars represent purified ricin samples (n ) 2).
(corresponding to ∼1 pmol on-target). Peptide sequences were further confirmed by running an LC-MS/MS experiment in parallel. The identified peptides along with their respective
sequences are summarized in Table 1. With MALDI-TOF MS, the sequence coverage was 66% for the A-chain (i.e., 12 out of 16 peptides expected with a mass above 500 Da) and 58% for Journal of Proteome Research • Vol. 7, No. 9, 2008 4159
research articles the B-chain (i.e., 8 out of 13). This corresponded to 61% of the total sequence (Table 1). The digestion protocol used left the ricin disulfide bonds intact. The ricin peptide sequence contains 11 cysteine residues, that is, 5 disulfide bonds and one free cysteine residue (according to the Swiss-Prot database). Besides the expected disulfide-linked peptides, the corresponding peptides bearing free cysteine residues were also “artifactually” observed in the MALDI mass spectrum. Their presence is a consequence of the cleavage of disulfide-linked peptides that can easily occur under MALDI conditions, especially when CHCA is used as matrix.31–33 A number of small and large peptides of both A- and B-chains were not detected by MALDI-TOF, because they were either eluted or not retained by the ZipTip C18 phase, or they were outside in the m/z range monitored. Moreover, ricin contains four potential N-glycosylation sites, Asn10 and Asn236 for A-chain, and Asn95 and Asn135 for B-chain. None of the corresponding glycopeptides was detected by MALDI-TOF MS because of their low desorption/ionization efficiency or high mass or both. Nevertheless, it should be noted that glycopeptides bearing Asn10 from A-chain and Asn95 from B-chain were observed by LC-MS/MS (Table 1), and were shown to bear mainly the (GlcNAc2Man3Xyl1Fuc1) and (GlcNAc2Man6) motifs, respectively. This perfectly agreed with the previously published N-glycan structures.34,35 Two isoforms of ricin basically exist, namely, the D form and the E form which has been considerably less potent than the D form.36 The A-chains of ricin D and E are identical, while their B-chain differs by 15%.37 It has been suggested that ricin E is a gene recombination product of ricin D and R. communis agglutinin (RCA). Indeed, ricin D shares a high level of sequence homology with RCA, that is, 93% and 84% for the A- and B-chains, respectively,37 and much care must be taken to avoid false-positive identification of RCA as ricin. Among the peptides belonging to the A-chain and detected by MALDI-TOF MS, four were common to RCA and ricin D and E (m/z 1074.6, m/z 1172.5, m/z 1728.9, and m/z 2259.2). Among the eight detected peptides for the B-chain, two peptides are specific to ricin D alone (m/z 1862.0 and m/z 2277.2), three are common to ricins D and E (m/z 1390.7, m/z 2400.2, and m/z 2947.4), one (m/z 2410.1) to ricins D and RCA, while the others occur equally in ricins D, E and RCA (m/z 611.3, and m/z 2231.2). The disulfide-linked interchain peptide at m/z 2275.1 is also represented in ricins D, E and RCA. Thus, this peptide, although representative of the intact protein with both A- and B-chains, cannot be used for specific forensic identification of ricin. As mentioned earlier, another major criterion regarding peptide selection is sensitivity under MALDI-TOF MS conditions. Ricin-specific peptides detected with the highest sensitivity belonged to the A-chain and corresponded to the ions at m/z 1013.6 and m/z 1310.6. The most sensitive peptide from the B-chain was the one at m/z 1862.0, but it was far less sensitive than those of the A-chain (factor of 10). However, when the immunocapture step was implemented, we found that the isotopic pattern of an ion at m/z 1309.5, corresponding to a tryptic fragment of the Fc domain of the mouse antibody used in the assay, can interfere with the ion at m/z 1310.6 when low levels of ricin are present. Cross-linkage of the antibody to the protein G beads failed to reduce the intensity of this background ion. In consequence, only the ion at m/z 1013.6 (although not providing the lower limit of detection) was used for accurate detection purposes. Nevertheless, the ion at m/z 4160
Journal of Proteome Research • Vol. 7, No. 9, 2008
Duriez et al. 1310.6 was kept as confirmation ion only for high levels of ricin. An ion at m/z 1728.9, which accounts for a peptide sequence occurring in both ricin and RCA, was also used to estimate RCA contamination. The sequences of these three A-chain peptides were confirmed by LC-MS/MS experiments, and further submitted to similarity searches (BLAST) to ensure specificity toward ricin. No B-chain peptides were chosen for the ricin detection, since the B-chain is considered to be present owing to the immunocapture step with antibodies directed against it. Assay Development. 1. Optimization of Trypsin Digestion. The use of an internal standard (IS) is required for accurate and reproducible detection, especially when sample preparation involves numerous steps. Addition of labeled proteins at the beginning of sample preparation can be used to account for all sources of sample loss.38 However, the production of labeled proteins by the introduction of stable isotope residues in the whole protein backbone is a highly complex process. The alternative we have chosen is the synthesis of labeled counterparts (13C6, 15N) of the three above-mentioned tryptic peptides (m/z 1013.6, m/z 1310.6 and m/z 1728.9). Such an approach provides reliable data but relies on essentially complete digestion of the target proteins so that the moles of peptide produced are equivalent to the moles of protein.39,40 In this context, the ricin digestion was carefully optimized in order to obtain complete proteolysis. Published data indicate the efficiency and specificity of trypsin digestion of proteins in organic-aqueous solvent systems.41–43 Addition of organic solvents such as methanol or acetonitrile to buffers can improve digestion by unfolding and solubilizing proteins, while trypsin remains active under these conditions. The studies of Russell et al. and Strader et al. carefully compared the effectiveness of different solvent systems involving either methanol, isopropanol or acetonitrile for digesting standard proteins as well as complex protein extracts, and clearly demonstrated that a solvent containing 80% ACN was the best mixture in both cases.41,42 Such organic-aqueous solvent systems also enable higher trypsin specificity and digestion yields than those obtained with commonly used denaturants or chaotropes.42,43 For ricin, Ostin et al. have shown that a 1-h trypsin digestion performed in 50% methanol enabled identification within crude bean or seed extracts.16 Surprisingly, these authors only investigated the use of methanol or isopropanol as solvents, and did not evaluate the impact of acetonitrile, which is the best solvent according to the different papers above-mentioned. Therefore, we have studied the potential benefit of using such a solvent by monitoring under different digestion conditions the intensities of the three targeted ions at m/z 1013.6, m/z 1310.6 and m/z 1728.9. Ricin was diluted in different solutions (100% aqueous, 50% MeOH, 50% ACN, 80% MeOH, 80% ACN) and further digested by trypsin by incubating the mixture at 37 °C for 2 h. The optimized conditions described by Ostin et al. were also studied.16 The best conditions were 80% ACN, 6.4 µg/mL trypsin, and 2-h digestion (Supporting Information, Figure S1). No improvement was noted using more “standard” digestion conditions, that is, with urea denaturation and cysteine reduction/alkylation prior to an overnight trypsin digestion performed as already described.44 Interestingly, we found that the efficiency of the enzymatic digestion could be controlled by monitoring two peptides at m/z 1546.8 and m/z 2305.1. The latter m/z ratio corresponds to the peptide (150-169) bearing one missed cleavage site which generates the two peptides of interest (150-160) and (161-169) from the A-chain. The ion
Ricin Detection by Immunocapture and MALDI-TOF MS at m/z 1546.8 corresponds to the peptide (41-53) of the B-chain. These two peptides were never observed under optimal 80% ACN digestion conditions. 2. Assay Specificity and Sensitivity. To increase the sensitivity of the assay (∼400 ng/mL without any concentration step) and to supply additional specificity, immunocapture of ricin was realized with magnetic beads coated using affinity-purified anti-B-chain monoclonal IgG for immobilization.12 The recovery of this particular immunocapture step was around 50% at 400 ng/mL. The specificity of the immunocapture step was evaluated by incubating RCA (from 100 to 1000 ng/mL) in the presence or absence of ricin (400 ng/mL). Under such experimental conditions, no RCA-specific peptide was detected, whereas no decrease in ricin-specific peptides was shown even with saturating concentrations of RCA. We also checked that no nonspecific binding occurred when using magnetic beads coated with monoclonal or polyclonal antibodies directed against unrelated protein. Elution of ricin from the immunocapture beads was initially attempted with organic solvent (ACN), but we observed considerable release of polymer from the beads that was not eliminated by the ZipTip purification (data not shown). Conversely, elution with 5% formic or 0.1% TFA resulted in “clean” MALDI-TOF mass spectra. With 0.1% TFA, the recovery of ricin from the beads was almost 100% using a single elution. After ricin elution from beads, the solution was neutralized and ACN was added to a final concentration of 80% for the trypsin digestion. Then, an evaporation step yielded a concentrated tryptic digest, which was further desalted/concentrated using ZipTip C18 pipet tips. To lower ion suppression effects inherent to sample complexity, the ACN percentage in the elution buffer was carefully optimized to limit the number of peptides eluted. The tryptic peptides of interest were almost 100% eluted with a solvent containing only 20% ACN and 0.1% TFA. By comparison, an elution containing 50% ACN resulted in observation of a 3- to 5-fold increase in sensitivity (Supporting Information, Figure S2). Finally, under the optimal conditions, the limit of detection of ricin based on the intensity determination of peaks at m/z 1013.6, m/z 1310.6, m/z 1728.9 was 50 ng/mL with a 500 µL sample (Figure 3). Compared with existing proteomic approaches19,24 with the same methodology, this limit of detection has the same order of sensitivity, and it remains sufficient for the detection of toxic concentrations of ricin. Assay precision was evaluated by analyzing the quality control (QC) samples at 100, 300, and 500 ng/mL in triplicate over 3 days. The CVs at m/z 1013.6 were 7%, 3.6% and 2.7% at 100, 300 and 500 ng/mL, respectively. These data showed that ricin could be accurately measured using the developed mass spectrometric immunoassay. Detection of Ricin in Biological Samples. 1. Detection in Media Containing Potentially Interfering Components. To demonstrate the applicability of the current method to environmental samples, milk was spiked with ricin at 400 ng/mL and submitted to immunocapture and MALDI-TOF. Compared with ricin spiked in assay buffer, the intensity of three selected peptides was in the range of 85-110% and 110-125% for semiskim and skim milk, respectively. These results demonstrate that the presence in milk of endogenous lactose, a potent binder to the B-chain, did not inhibit the binding of ricin B-chain (lectin) to immunomagnetic beads. 2. Ricin Detection in Castor Beans of Different Varieties. To be suitable, the assay should be able to detect ricin toxin from different varieties of R. communis plants and from
research articles different geographic origins. This is a concern since it has been reported that horticultural varieties of R. communis behave differently when analyzed by liquid chromatography/electrospray ionization and capillary zone electrophoresis.26 For instance, one of the species, R. communis zanzibariensis, could be differentiated from other species such as R. communis impala or gibsonii. Although this was attributed to different patterns of isoforms, no study has demonstrated that the different species share common peptide profiles. We tested several different Ricinus species, namely, R. communis carmencita pink from Spain, R. communis zanzibariensis, R. communis gibsonii and R. communis impala from Tanzania, R. communis noori from Pakistan, R. communis black diamond from India, and several cultivars of R. communis zibima, R. communis jiaxing or R. communis yunnan from China. Their beans were extracted and the crude extract was submitted without further purification to the present assay. Purified ricins were also tested. In the case of complex samples such as seed extracts, the immunocapture step is of particular interest for obtaining specific and sensitive ricin detection. As an example, Figure 4 shows the mass spectra obtained before and after immunocapture for a purified ricin sample (Figure 4a,b) and a crude seed extract sample (Figure 4c,d). As may be seen from this figure, the immunocapture step substantially lowers the detection limit of ricin in crude extracts by removing the peptides originating from coextracted proteins (Figure 4c,d). An interesting finding is that all the most intense peptides found for the reference purified ricin were detected in all the extracted seeds (Supporting Information, Figure S3). These data indicate that the monoclonal antibody used for the immunocapture was able to recognize the ricin B-chain of the different cultivars and species. Moreover, it should be emphasized that the diagnostic peptides at m/z 1013.6, m/z 1310.6 and m/z 1728.9 were consistently observed in the different seed extracts, thus, underlining their usefulness for forensic issues. All the peptides observed both in purified ricin and crude plant extracts are summarized in Table S1 (Supporting Information). On the basis of these particular MALDI-TOF peptide mass fingerprints, no clear differences in terms of amino acid sequence are evident between ricins of various origins or cultivars. For instance, no ricin E specific peptides can be observed, which is probably due to the low detection sensitivity of the B-chain peptides. The intensity ratios between the ions of interest (i.e., m/z 1013.6/1310.6 and m/z 1013.6/1728.9 ratios) were determined in order to probe the robustness of the assay for the analysis of complex seed extracts (Figure 5). The m/z 1013.6/1310.6 ratio remained roughly constant in both seed extracts and purified ricin samples (Figure 5A), whereas the m/z 1013.6/1728.9 ratio was markedly lower in a few seed extracts (Figure 5B). The stability of the former ratio highlights the absence of interfering peptides in extracts and also confirms that no major amino acid sequence variants exists on these two peptides, whatever the seeds considered. There was more fluctuation in the ratio m/z 1013.6/1728.9 (Figure 5B). As mentioned earlier, the ion at m/z 1728.9 corresponds to a peptide that occurs equally in ricin and RCA; thus, a lower m/z 1013.6/1728.9 ratio reflects the presence of contaminating RCA in the extracts. Such nonspecific extraction was not previously observed in buffer solutions artificially contaminated with RCA (see section on assay specificity), and is therefore linked to the extreme complexity of some crude extracts. The presence of RCA in Journal of Proteome Research • Vol. 7, No. 9, 2008 4161
research articles these samples was further confirmed by the presence of other RCA-specific peptides (e.g., m/z 885.5, m/z 2239.3, and m/z 1533.7, corresponding to LEQLGGLR, FNVYDVSILIPIIALMVYR, and SNTDWNQLWTLR peptide sequences, respectively). In another set of experiments, we analyzed the different seed extracts at a theoretical ricin concentration of 2 µg/mL, assuming an average ricin content of beans of 2% (w/w). Immunocaptured ricin samples were spiked with a constant amount of the IS mixture prior to trypsin digestion. The m/z 1013.6/m/z 1020.6 (IS) ratios were then determined for all the seed extracts and further compared with those obtained using purified ricin samples at the same concentration (Figure 6). For a ricin concentration of 2 µg/mL, m/z ratios of purified varieties impala and zanzibariensis were ∼25% lower than those obtained for the two samples of R. communis (Figure 6, white bars). This discrepancy could be explained by a lower affinity of the antibody used for the immunocapture toward ricin varieties differing from R. communis, presumably as a consequence of structural differences (e.g., glycosylation). For each variety, this variability in binding affinity must be taken into account when developing a quantitation assay. That is why the present assay is only described as a detection tool with a LOD expressed as a minimum detectable amount of Ricin D from R. communis. The variability observed in the m/z 1013.6/m/z 1020.6 ratios (Figure 6, dark bars) between the different extracts could be also explained by differences in binding affinities, but a constant extraction ricin recovery could not be ascertained. The present data demonstrate nevertheless that seeds from different species and origins have ricin contents of the same order of magnitude, which is in good agreement with the 1-5% range often mentioned in the literature.1
Conclusion In a potential bioterrorist incident, the presence of ricin should be confirmed using a panel of complementary analytical methods. In addition to existing immunoassays and bioassays, we report here a proteomics-based methodology addressing both ricin structure confirmation and detection. In this work, an immunocapture step was implemented in order to concentrate the samples and to remove unrelated proteins potentially present in the media to be tested. This new method offers an alternative to existing methods such as immunological methods or bioassays. Although the limit of detection (50 ng/mL) of ricin from R. communis remains higher than existing methods, it is sufficient to assess toxic concentration of ricin in environmental and food samples. Compared with other approaches, an interesting feature is that the mass spectrometry detection provides a specificity increase due to the simultaneous monitoring of several characteristic ricin-specific peptides. Interestingly, the assay is able to detect a large number of R. communis species, as demonstrated by the result obtained with several different cultivars and varieties. Finally, the possible miniaturization of MALDI-TOF technology may mean that the assay could be adapted for use with a portable mass spectrometer.45,46
Supporting Information Available: Figure S1, MALDI TOF mass spectra showing digestion efficiency of ricin with different organic-aqueous solvent compositions: 80% ACN (a), 50% ACN (b), 50% MeOH16 (c), 100% aqueous (d). Peptides from A- and B-chains are labeled with (2) and (9), respectively. Peaks corresponding to peptides bearing a miscleavage are labeled with (•), IS peaks with an asterisk (*) and tryptic 4162
Journal of Proteome Research • Vol. 7, No. 9, 2008
Duriez et al. peptides with (T). Figure S2, peak intensity of the target peptides according to the acetonitrile percentage used for elution from ZipTips. After partial evaporation, the peptide mixture was extracted and concentrated using C 18 pipet tips. Peptide elution was performed using different acetonitrile percentages containing 0.1% trifluoroacetic acid (n ) 2). Figure S3, MALDI-TOF mass spectra of tryptic digest of ricin from a crude extract of R. communis (a), R. communis carmencita pink (b), R. communis gibsonii (c), R communis impala (d), R. communis black diamond (e), R. communis jiaxing 1 (f), R. communis jiaxing 2 (g), R. communis noori (h), R. communis yunnan (i), R communis zanzibariensis (j), R. communis zibima 1 (k), R. communis zibima 2 (l), R. communis zibima 3 (m), R. communis zibima 4 (n), R. communis zibima 5 (o), R. communis zibima 101 (p), R. communis zibima 108 (q), purified R. communis 1 (r), purified R. communis 2 (s), purified R. communis impala (t), purified R. communis zanzibariensis (u) at 2 µg/mL after immunocapture. Table S1, summary of ricin tryptic peptides observed in the different seed extracts. This material is available free of charge via the Internet at http://pubs.acs.org.
References (1) Audi, J.; Belson, M.; Patel, M.; Schier, J.; Osterloh, J. Ricin poisoning: a comprehensive review. JAMA, J. Am. Med. Assoc. 2005, 294 (18), 2342–2351. (2) Olsnes, S. The history of ricin, abrin and related toxins. Toxicon 2004, 44 (4), 361–370. (3) Bradberry, S. M.; Dickers, K. J.; Rice, P.; Griffiths, G. D.; Vale, J. A. Ricin poisoning. Toxicol. Rev. 2003, 22 (1), 65–70. (4) Convention on the prohibition of the development, production, stockpiling and use of chemical weapons and their destruction, www.opcw.org. Entry into force, April 29, 1997. (5) Endo, Y.; Tsurugi, K. RNA N-glycosidase activity of ricin A-chain. Mechanism of action of the toxic lectin ricin on eukaryotic ribosomes. J. Biol. Chem. 1987, 262 (17), 8128–8130. (6) Olsnes, S.; Kozlov, J. V. Ricin. Toxicon 2001, 39 (11), 1723–1728. (7) Ler, S. G.; Lee, F. K.; Gopalakrishnakone, P. Trends in detection of warfare agents. Detection methods for ricin, staphylococcal enterotoxin B and T-2 toxin. J.Chromatogr., A 2006, 1133 (1-2), 1–12. (8) Lubelli, C.; Chatgilialoglu, A.; Bolognesi, A.; Strocchi, P.; Colombatti, M.; Stirpe, F. Detection of ricin and other ribosome-inactivating proteins by an immuno-polymerase chain reaction assay. Anal. Biochem. 2006, 355 (1), 102–109. (9) Endo, Y.; Mitsui, K.; Motizuki, M.; Tsurugi, K. The mechanism of action of ricin and related toxic lectins on eukaryotic ribosomes. The site and the characteristics of the modification in 28 S ribosomal RNA caused by the toxins. J. Biol. Chem. 1987, 262 (12), 5908–5912. (10) Brigotti, M.; Barbieri, L.; Valbonesi, P.; Stirpe, F.; Montanaro, L.; Sperti, S. A rapid and sensitive method to measure the enzymatic activity of ribosome-inactivating proteins. Nucleic Acids Res. 1998, 26 (18), 4306–4307. (11) Hines, H. B.; Brueggemann, E. E.; Hale, M. L. High-performance liquid chromatography-mass selective detection assay for adenine released from a synthetic RNA substrate by ricin A chain. Anal. Biochem. 2004, 330 (1), 119–122. (12) Becher, F.; Duriez, E.; Volland, H.; Tabet, J. C.; Ezan, E. Detection of functional ricin by immunoaffinity and liquid chromatographytandem mass spectrometry. Anal. Chem. 2007, 79 (2), 659–665. (13) Darby, S. M.; Miller, M. L.; Allen, R. O. Forensic determination of ricin and the alkaloid marker ricinine from castor bean extracts. J. Forensic Sci. 2001, 46 (5), 1033–1042. (14) Na, D. H.; Cho, C. K.; Youn, Y. S.; Choi, Y.; Lee, K. R.; Yoo, S. D.; Lee, K. C. Capillary electrophoresis to characterize ricin and its subunits with matrix-assisted laser desorption/ionization timeof-flight mass spectrometry. Toxicon 2004, 43 (3), 329–335. (15) Fredriksson, S. A.; Hulst, A. G.; Artursson, E.; de.Jong, A. L.; Nilsson, C.; van.Baar, B. L. Forensic identification of neat ricin and of ricin from crude castor bean extracts by mass spectrometry. Anal. Chem. 2005, 77 (6), 1545–1555. (16) Ostin, A.; Bergstrom, T.; Fredriksson, S. A.; Nilsson, C. Solvent.assisted trypsin digestion of ricin for forensic identification by LCESI MS/MS. Anal. Chem. 2007, 79 (16), 6271–6278.
research articles
Ricin Detection by Immunocapture and MALDI-TOF MS (17) Gobom, J.; Kraeuter, K. O.; Persson, R.; Steen, H.; Roepstorff, P.; Ekman, R. Detection and quantification of neurotensin in human brain tissue by matrix-assisted laser desorption/ionization timeof-flight mass spectrometry. Anal. Chem. 2000, 72 (14), 3320–3326. (18) Helmke, S. M.; Yen, C. Y.; Cios, K. J.; Nunley, K.; Bristow, M. R.; Duncan, M. W.; Perryman, M. B. Simultaneous quantification of human cardiac alpha- and beta-myosin heavy chain proteins by MALDI-TOF mass spectrometry. Anal. Chem. 2004, 76 (6), 1683– 1689. (19) Nelson, R. W.; Nedelkov, D.; Tubbs, K. A.; Kiernan, U. A. Quantitative mass spectrometric immunoassay of insulin like growth factor 1. J. Proteome. Res. 2004, 3 (4), 851–855. (20) Alterman, M. A.; Kornilayev, B.; Duzhak, T.; Yakovlev, D. Quantitative analysis of cytochrome p450 isozymes by means of unique isozyme-specific tryptic peptides: a proteomic approach. Drug Metab. Dispos. 2005, 33 (9), 1399–1407. (21) Gelfanova, V.; Higgs, R. E.; Dean, R. A.; Holtzman, D. M.; Farlow, M. R.; Siemers, E. R.; Boodhoo, A.; Qian, Y. W.; He, X.; Jin, Z.; Fisher, D. L.; Cox, K. L.; Hale, J. E. Quantitative analysis of amyloid-beta peptides in cerebrospinal fluid using immunoprecipitation and MALDI-Tof mass spectrometry. Briefings Funct. Genomics Proteomics 2007, 6 (2), 149–158. (22) Jiang, J.; Parker, C. E.; Fuller, J. R.; Kawula, T. H.; Borchers, C. H. An immunoaffinity tandem mass spectrometry (iMALDI) assay for detection of Francisella tularensis. Anal. Chim. Acta 2007, 605 (1), 70–79. (23) Pan, S.; Rush, J.; Peskind, E. R.; Galasko, D.; Chung, K.; Quinn, J.; Jankovic, J.; Leverenz, J. B.; Zabetian, C.; Pan, C.; Wang, Y.; Oh, J. H.; Gao, J.; Zhang, J.; Montine, T.; Zhang, J. Application of targeted quantitative proteomics analysis in human cerebrospinal fluid using a liquid chromatography matrix-assisted laser desorption/ionization time-of-flight tandem mass spectrometer (LC MALDI TOF/TOF) Platform. J. Proteome. Res. 2008, 7 (2), 720–730. (24) Schlosser, G.; Kacer, P.; Kuzma, M.; Szilagyi, Z.; Sorrentino, A.; Manzo, C.; Pizzano, R.; Malorni, L.; Pocsfalvi, G. Coupling immunomagnetic separation on magnetic beads with matrix-assisted laser desorption ionization-time of flight mass spectrometry for detection of staphylococcal enterotoxin B. Appl. Environ. Microbiol. 2007, 73 (21), 6945–6952. (25) Woo, B. H.; Lee, J. T.; Na, D. H.; Lee, K. C. Sepharose-unbinding ricin E as a source for ricin A chain immunotoxin. J. Immunol. Methods 2001, 249 (1-2), 91–98. (26) Despeyroux, D.; Walker, N.; Pearce, M.; Fisher, M.; McDonnell, M.; Bailey, S. C.; Griffiths, G. D.; Watts, P. Characterization of ricin heterogeneity by electrospray mass spectrometry, capillary electrophoresis, and resonant mirror. Anal. Biochem. 2000, 279 (1), 23–36. (27) Nelson, R. W.; Krone, J. R.; Bieber, A. L.; Williams, P. Mass spectrometric immunoassay. Anal. Chem. 1995, 67 (7), 1153–1158. (28) Beaumelle, B.; Alami, M.; Hopkins, C. R. ATP-dependent translocation of ricin across the membrane of purified endosomes. J. Biol. Chem. 1993, 268 (31), 23661–23669. (29) Lopez, L. L.; Tiller, P. R.; Senko, M. W.; Schwartz, J. C. Automated strategies for obtaining standardized collisionally induced dissociation spectra on a benchtop ion trap mass spectrometer. Rapid Commun. Mass Spectrom. 1999, 13, 663–668. (30) Sullivan, B.; Addona, T. A.; Carr, S. A. Selective detection of glycopeptides on ion trap mass spectrometers. Anal. Chem. 2004, 76 (11), 3112–3118.
(31) Patterson, S. D.; Katta, V. Prompt fragmentation of disulfide-linked peptides during matrix-assisted laser desorption ionization mass spectrometry. Anal. Chem. 1994, 66 (21), 3727–3732. (32) Crimmins, D. L.; Saylor, M.; Rush, J.; Thoma, R. S. Facile, in situ matrix-assisted laser desorption ionization-mass spectrometry analysis and assignment of disulfide pairings in heteropeptide molecules. Anal. Biochem. 1995, 226 (2), 355–361. (33) Schnaible, V.; Wefing, S.; Resemann, A.; Suckau, D.; Bucker, A.; Wolf-Kummeth, S.; Hoffmann, D. Screening for disulfide bonds in proteins by MALDI in-source decay and LIFT-TOF/TOF-MS. Anal. Chem. 2002, 74 (19), 4980–4988. (34) Kimura, Y.; Hase, S.; Kobayashi, Y.; Kyogoku, Y.; Ikenaka, T.; Funatsu, G. Structures of sugar chains of ricin D. J. Biochem 1988, 103 (6), 944–949. (35) Kimura, Y.; Kusuoku, H.; Tada, M.; Takagi, S.; Funatsu, G. Structural analyses of sugar chains from ricin A-chain variant. Agric. Biol. Chem. 1990, 54 (1), 157–162. (36) Woo, B. H.; Lee, J. T.; Lee, K. C. Purification of Sepharoseunbinding ricin from castor beans (Ricinus communis) by hydroxyapatite chromatography. Protein Expression Purif. 1998, 13 (2), 150–154. (37) Araki, T.; Funatsu, G. The complete amino acid sequence of the B-chain of ricin E isolated from small-grain castor bean seeds. Ricin E is a gene recombination product of ricin D and Ricinus communis agglutinin. Biochim. Biophys. Acta 1987, 911 (2), 191– 200. (38) Brun, V.; Dupuis, A.; Adrait, A.; Marcellin, M.; Thomas, D.; Court, M.; Vandenesch, F.; Garin, J. Isotope-labeled protein standards: toward absolute quantitative proteomics. Mol. Cell. Proteomics 2007, 6 (12), 2139–2149. (39) Gerber, S. A.; Rush, J.; Stemman, O.; Kirschner, M. W.; Gygi, S. P. Absolute quantification of proteins and phosphoproteins from cell lysates by tandem MS. Proc. Natl. Acad. Sci. U.S.A. 2003, 100 (12), 6940–6945. (40) Barnidge, D. R.; Hall, G. D.; Stocker, J. L.; Muddiman, D. C. Evaluation of a cleavable stable isotope labeled synthetic peptide for absolute protein quantification using LC-MS/MS. J. Proteome Res. 2004, 3 (3), 658–661. (41) Russell, W. K.; Park, Z. Y.; Russell, D. H. Proteolysis in mixed organic-aqueous solvent systems: applications for peptide mass mapping using mass spectrometry. Anal. Chem. 2001, 73 (11), 2682–2685. (42) Strader, M. B.; Tabb, D. L.; Hervey, W. J.; Pan, C.; Hurst, G. B. Efficient and specific trypsin digestion of microgram to nanogram quantities of proteins in organic-aqueous solvent systems. Anal. Chem. 2006, 78 (1), 125–134. (43) Hervey, W. J.; Strader, M. B.; Hurst, G. B. Comparison of digestion protocols for microgram quantities of enriched protein samples. J. Proteome Res. 2007, 6 (8), 3054–3061. (44) Fenaille, F. Le; Mignon, M.; Groseil, C.; Ramon, C.; Riande, S.; Siret, L.; Bihoreau, N. Site-specific N-glycan characterization of human complement factor H. Glycobiology 2007, 17 (9), 932–944. (45) English, R. D.; Warscheid, B.; Fenselau, C.; Cotter, R. J. Bacillus spore identification via proteolytic peptide mapping with a miniaturized MALDI TOF mass spectrometer. Anal. Chem. 2003, 75 (24), 6886–6893. (46) Cornish, T. J.; Antoine, M. D.; Ecelberger, S. A.; Demirev, P. A. Arrayed time-of-flight mass spectrometry for time-critical detection of hazardous agents. Anal. Chem. 2005, 77 (13), 3954–3959.
PR8003437
Journal of Proteome Research • Vol. 7, No. 9, 2008 4163
