Identification of Ricin in Crude and Purified Extracts from Castor Beans

May 20, 2010 - Ricin is a toxic protein produced in the seeds of the castor bean plant. The toxicity of the protein and the ease in which it can be ex...
23 downloads 6 Views 631KB Size
Download PDF
Anal. Chem. 2010, 82, 5246–5252

Identification of Ricin in Crude and Purified Extracts from Castor Beans Using On-Target Tryptic Digestion and MALDI Mass Spectrometry Craig S. Brinkworth* Human Protection and Performance Division, Defence Science and Technology Organisation, Fishermans Bend, Victoria, Australia, 3207 Ricin is a toxic protein produced in the seeds of the castor bean plant. The toxicity of the protein and the ease in which it can be extracted from the seeds makes it a potential biological warfare agent. There has been extensive work in the development of analytical techniques that can identify the protein robustly and rapidly. On-target tryptic digestion and MALDI MS was used to distinguish ricin from bovine serum albumin and three other type 2 ribsome-inactivating proteins (RIPs), abrin, agglutinin (RCA120), and viscumin, using the peptide mass fingerprint. The sequence coverage obtained was enhanced using methanol-assisted tryptic digestion and was particularly useful for the detection of these toxins in complex matrixes. When used in conjunction with intact protein MALDI mass measurement, a positive identification of ricin (or any of the other RIPs) was achieved including confirmation of the integrity of the disulfide bond between the A and B chains. This applicability of this methodology was demonstrated by the identification of ricin in a typical “crude white powder” that may be illicitly produced in a clandestine lab. The analysis on the solubilized sample using this method can be undertaken in around an hour with minimal sample preparation. Ricin is a toxin produced by the castor bean plant that gained notoriety as the suspected poison used to murder a Ukranian secret service agent administered via injection by an umbrella tip.1 However, ricin can also be administered via inhalation of the aerosolized protein. The LD50 of ricin exposure via these two methods has been estimated as 5-10 µg/kg and 3-5 µg/kg in mice, respectively.2 The ease of production of the toxin from the castor bean seeds and the accessibility of the plants in several countries have led to the inclusion of ricin in the * To whom correspondence should be addressed. Address: Human Protection and Performance Division, Defence Science and Technology Organisation, 506 Lorimer Street, Fishermans Bend VIC Australia 3207. Telephone: + 61 3 9626 8421. Fax: + 61 3 9626 8342. E-mail: [email protected]. (1) Crompton, R.; Gall, D. Med. Leg. J. 1980, 48, 51–62. (2) Audi, J.; Belson, M.; Patel, M.; Schier, J.; Osterloh, J. JAMA 2005, 294, 2342–2351.

5246

Analytical Chemistry, Vol. 82, No. 12, June 15, 2010

Category B list of potential biological weapons produced by The United States Centers for Disease Control and Prevention (CDC).3,4 The most common methods of detection for ricin are immunoaffinity based assays due to their sensitivity and utility under aqueous conditions.5 This technology has been adapted and optimized for the detection of ricin in the field.6,7 These assays use antibodies that bind to sites on the ricin molecule and, in many cases, use fluorescence to detect this bound antibody. However, in practice, these assays are not always unambiguous8 because of false positives resulting from cross reactivity of other molecules with similar binding motifs. Thus, it would be advantageous to use additional methods utilizing orthogonal technologies to confirm the presence of ricin. Recently, we published a robust method for the detection of ricin using intact protein MALDI mass measurement.9 This method was initially developed to complement the immunological assays and to have two (or more) disparate technologies for a positive identification of chemicals defined as Schedule 1 by the Chemical Weapons Convention.10 The advantages of this method are that it detects the intact protein demonstrating the integrity of the disulfide bond between the A and B chains that is crucial for the toxicity of the protein and it is quick and relatively simple to perform. This bond is crucial for the toxicity of the toxin because the B chain is responsible for cell binding and internalization of the toxin allowing the enzymatically active A chain to enter the cell and inhibit cellular translation.11 The major limitation of this method is its inability to distinguish between molecules of similar molecular weight. This can only be used as a confirmatory (3) Waterer, G. W.; Robertson, H. Respirology 2009, 14, 5–11. (4) Coopman, V.; De Leeuw, M.; Cordonnier, J.; Jacobs, W. Forensic Sci. Int. 2009, 189, e13–20. (5) Ler, S. G.; Lee, F. K.; Gopalakrishnakone, P. J. Chromatogr., A 2006, 1133, 1–12. (6) Petrovick, M. S.; Harper, J. D.; Nargi, F. E.; Schwoebel, E. D.; Hennessy, M. C.; Rider, T. H.; Hollis, M. A. Lincoln Lab. J. 2007, 17, 63–84. (7) Dayan-Kenigsberg, J.; Bertocchi, A.; Garber, E. A. J. Immunol. Methods 2008, 336, 251–254. (8) Tate, J.; Ward, G. Clin. Biochem. Rev. 2004, 25, 105–120. (9) Brinkworth, C. S.; Pigott, E. J.; Bourne, D. J. Anal. Chem. 2009, 81, 1529– 1535. (10) Conference on Disarmament, The Convention on the Prohibition of the Development, Production, Stockpiling and Use of Chemical Weapons and on their Destruction, United Nations Document CD/1170, Geneva, August 1992. (11) Lord, M. J.; Jolliffe, N. A.; Marsden, C. J.; Pateman, C. S.; Smith, D. C.; Spooner, R. A.; Watson, P. D.; Roberts, L. M. Toxicol. Rev. 2003, 22, 53– 64. 10.1021/ac100650g Published 2010 by the American Chemical Society Published on Web 05/20/2010

test. In contrast, the use of peptide maps generated by proteolytic digestion of ricin can be used to identify the protein,12 and as demonstrated by Ostin and co-workers, the protein can be unambiguously identified using LC/MSMS.13 However, these methods require additional sample preparation steps that lead to longer analysis times. Therefore, a method that combines both the speed of intact protein mass measurement and the information provided by proteolytic digestion would be very advantageous for the rapid identification of ricin. On-target tryptic digestion is a method that has been applied for the rapid identification of microorganisms such as Bacillus spores and bacteria.14-16 The advantage of this technique is that it combines rapid sample preparation along with peptide mass fingerprinting/MSMS spectral data that unambiguously identifies the proteins of interest. Positive identification of Bacillus spores have been reported using tryptic digestion times as short as 5 min.14 This paper describes a similar technique that can be applied to the identification of ricin and can distinguish ricin from other type 2 ribsome-inactivating protein (RIP) toxins. EXPERIMENTAL SECTION Material. HPLC grade water and acetonitrile were from Merck (Darmstadt, Germany). Tween 80, purified ricin from Ricinus communis (castor bean) (now discontinued), purified abrin from Abrus precatorius, purified viscumin from Viscum album, purified agglutinin (RCA120) from Ricinus communis, purified essentially fatty acid free bovine serum albumin (BSA), and trifluoroacetic acid (TFA) were from Sigma (St. Louis, MO). Ricin, abrin, viscumin, and agglutinin are very toxic proteins and require appropriate safety measures. R-Cyano-4-hydroxycinammic acid (HCCA) (>99.6%) was from Bruker Daltonik Gmblt (Leizig, Germany). Trypsin Gold-Mass Spec grade was from Promega (Madison, WI). The mass spectra were acquired on an MTP 384 Ground Steel MALDI Target and an autoflex II TOF/TOF from Bruker Daltonik Gmblt (Leizig, Germany). A sample of crude castor bean extract was produced using an “open source” recipe. On-Target Digestion of Purified Ricin Under Various Solvent Conditions. A working solution of purified ricin (500 fmol/µL) was prepared in water by dilution of a ricin stock solution (3.4 mg/mL). Ricin working solution (1 µL) was deposited onto the target and the spot air-dried. The dried spot was layered with incubation solution (1 µL) and incubated in a humid environment for 30 min at ambient temperature. The incubation solution consisted of 0.1 µg/µL trypsin in equal volumes of 50 mM ammonium bicarbonate and water, methanol, isopropanol, or acetonitrile. Trifluoroacetic acid (1%, 0.5 µL) and matrix solution (saturated HCCA in 0.1% TFA/acetonitrile 2:1, 1 µL) were added to the spot, and the spot was allowed to dry. MALDI mass spectra were acquired on a Bruker autoflex II TOF/TOF mass spectrometer using an automated laser firing sequence. (12) Duriez, E.; Fenaille, F.; Tabet, J. C.; Lamourette, P.; Hilaire, D.; Becher, F.; Ezan, E. J. Proteome Res. 2008, 7, 4154–4163. (13) Ostin, A.; Bergstrom, T.; Fredriksson, S. A.; Nilsson, C. Anal. Chem. 2007, 79, 6271–6278. (14) Pribil, P. A.; Patton, E.; Black, G.; Doroshenko, V.; Fenselau, C. J. Mass Spectrom. 2005, 40, 464–474. (15) Yao, Z. P.; Afonso, C.; Fenselau, C. Rapid Commun. Mass Spectrom. 2002, 16, 1953–1956. (16) Russell, S. C.; Edwards, N.; Fenselau, C. Anal. Chem. 2007, 79, 5399– 5406.

On-Target Tryptic Digestion of Ricin, Agglutinin, Abrin, Viscumin, and BSA. Working solutions of purified ricin, agglutinin, abrin, viscumin, and BSA (1 pmol/µL) were prepared. Protein working solutions (1 µL) were deposited onto the target, and the spots air-dried. The dried spots were layered with incubation solution (0.1 µg/µL trypsin, 50 mM ammonium bicarbonate/methanol 1:1, 1 µL), and the spots were incubated in a humid environment for 30 min at ambient temperature. Trifluoroacetic acid (1%, 0.5 µL) and matrix solution (saturated HCCA in 0.1% TFA/acetonitrile 2:1, 1 µL) were added to the spot, and the spot was allowed to dry. MALDI mass spectra were acquired on a Bruker autoflex II TOF/TOF mass spectrometer using an automated laser firing sequence. MALDI Spot Preparation of the “Open Source” Sample (Intact Protein). The resuspended pellet was centrifuged (×14 100g, 5 min at 20 °C), and the supernatant was collected. Serial dilutions of ×2, ×5, ×50, ×100, and ×500 of the supernatant were prepared in a final concentration of 0.01% Tween 80. The spots were prepared by mixing the diluted supernatant (1 µL) and matrix solution (sat. HCCA in 0.1% TFA/acetonitrile 1:1, 1 µL) and applying the entire mixture to the target. The spots were allowed to dry. MALDI Spot Preparation of the “Open Source” Sample (On-Target Tryptic Digestion). Serial dilutions of ×5, ×50, ×100, and ×500 of the supernatant were prepared in water. The neat and diluted supernatant solutions (1 µL) were deposited onto the target, and spots were allowed to dry. The dried spots were layered with incubation solution (0.1 µg/µL trypsin, 50 mM ammonium bicarbonate/methanol 1:1, 1 µL), and the spots were incubated in a humid environment for 30 min. Trifluoroacetic acid (1%, 0.5 µL) and matrix solution (saturated HCCA in 0.1% TFA/acetonitrile 2:1, 1 µL) were added to the spot, and the spot was allowed to dry. Acquisition of Mass Spectral Data (Intact Protein). MALDI mass spectra were acquired on an autoflex II TOF/TOF mass spectrometer from Bruker Daltonik (Leipzig, Germany). The spectra were acquired in linear mode using the following settings: laser power, 20%; laser frequency, 50 Hz; ion source 1, 20.00 kV; ion source 2, 17.50 kV; lens, 8.50 kV; pulsed ion extraction, 350 ns; matrix suppression, gating (maximum, 6000 m/z); m/z detection: 10 000-140 000; range, high; sample rate, 0.1 GS/s; detector gain, 1.77× (turbo: off); high mass accelerator, 8.50 kV; electronic gain, enhanced; and real time smooth, high. An automated laser firing sequence was used to obtain the data that involved the accumulation of 3000 shots (100 shots per position) using a hexagonal pattern originating from the center of the spot. The shots were summed and smoothed using the following parameters: algorithm, Savitzky Golay; width, 100 m/z; and cycles, 3. The spectra obtained were externally calibrated. Acquisition of Mass Spectral Data (On-Target Tryptic Digestion). MALDI mass spectra were acquired on an autoflex II TOF/TOF mass spectrometer from Bruker Daltonik (Leipzig, Germany). The spectra were acquired in reflectron mode using the following settings: laser power, 17%; laser frequency, 50 Hz; ion source 1, 19.00 kV; ion source 2, 16.90 kV; lens, 8.50 kV; reflector, 20.00 kV; reflector 2, 9.50 kV; pulsed ion extraction, 100 ns; matrix suppression, deflection (400 m/z); m/z detection, 640-4000; range, low; sample rate, 2.0 GS/s; detector gain, 6.6×; Analytical Chemistry, Vol. 82, No. 12, June 15, 2010

5247

electronic gain, enhanced; and real time smooth, off. An automated laser firing sequence was used to obtain the data that involved the accumulation of 2000 shots (100 shots per position) using a spiral pattern originating in the center of the spot. The shots were summed, baseline subtracted using the TopHat algorithm, and smoothed using the following parameters: algorithm, Savitzky Golay; width, 0.2 m/z; and cycles, 1. The spectra obtained were externally calibrated. The peaklists were generated with the following parameters: peak detection algorithm, snap; signal to noise threshold, 4; relative intensity threshold, 0; minimum intensity threshold, 0; quality factor threshold, 75; and snap average composition, Averagine. Mass Spectral Data Searching: MSFit Protein Prospector. Peaklists from the on-target tryptic digestions were searched using MSFit in the Protein Prospector suite of programs (http:// prospector.ucsf.edu). The following parameters were used for the searches: database, SwissProt; digest used, Trypsin; max. # missed cleavages, 2; minimum matches, 4; sort type, score sort; min. parent ion matches, 1; MOWSE on, 1; MOWSE P factor, 0.4; report homologous proteins, interesting. The five top hits were displayed. RESULTS The in-solution tryptic digestion of proteins and the subsequent mass spectral anaylsis of the digests has been used as an effective way to identify proteins specifically and rapidly.17 The success of this type of analysis depends on the number of tryptic fragments that can be detected and is directly related to the efficiency of the enzymatic digestion used to produce the fragments.18 Solventassisted tryptic digestion has been used extensively to increase this efficiency particularly for hydrophobic peptides.19 This technique has been applied for the digestion of ricin and has demonstrated increased sequence coverage. These reports recommended the addition of either 50% methanol13 or 80% acetonitrile12 to the digestion solution. Can similar results be achieved using solvent-assisted on-target tryptic digestion of ricin? Figure 1 summarizes the mass spectra obtained from the ontarget digestion of ricin under a variety of solvent conditions: water, water/methanol (1:1), water/isopropanol (1:1), and water/ acetonitrile (1:1). The on-target digestion using water only produced 16 tryptic peptides that were recognized as peaks by the software. These are summarized in the table at the bottom of the figure. The peptides were from both the A and B chains, including peptides that were glycosylated and contained disulfide bonds. Five of the observed tryptic peptides are found in both ricin and aggluttinin (or RCA120). The sequence coverage achieved was comparable to that obtained when ricin is digested via the in-solution method, and a similar amount is spotted onto the target (Table 1 and Figure 1S in the Supporting Information). However, the in-solution digestion sample has to be treated prior to spot preparation in order for signal to be observed. Conversely, the on-target digest can be analyzed immediately after the spot is dried. One tryptic peptide of interest that was not observed in either the on-target or in-solution digestion mass spectra shown in Figure 1 (but has been reported13) was the disulfide linked tryptic peptide joining the A (17) Aebersold, R.; Mann, M. Nature 2003, 422, 198–207. (18) Yates, J. R., 3rd; Speicher, S.; Griffin, P. R.; Hunkapiller, T. Anal. Biochem. 1993, 214, 397–408. (19) Russell, W. K.; Park, Z. Y.; Russell, D. H. Anal. Chem. 2001, 73, 2682– 2685.

5248

Analytical Chemistry, Vol. 82, No. 12, June 15, 2010

Figure 1. Mass spectra obtained on the autoflex II TOF/TOF mass spectrometer from the on-target tryptic digestion of purified ricin (500 fmol) using a variety of solvent conditions. The dried spot of analyte was layered with 0.1 µg/µL trypsin in 25 mM ammonium bicarbonate (1 µL) in the solvent conditions shown on the figure. The peptides observed in the mass spectra are summarized in the table below as follows: ricin only peptides (normal text), glycosylated ricin peptides (bold text), and ricin/agglutinin peptides (bold and italic text). The numbers correspond to each of the peptides listed.

and B chains (Ta24-ss-Tb1). This peptide demonstrates that the disulfide between the A and B chains is intact. However, it is generally only reliably observed when the tryptic digest peptide mixture is chromatographically separated prior to mass spectral acquisition. The solvent-assisted on-target tryptic digestions produced mixed results depending on the solvent composition (Figure 1). In methanol/water and isopropanol/water, the software identified 14 (peaks 12 and 13 absent) and 13 (peaks 10, 12, and 13 absent) of the 16 tryptic peptides observed in the aqueous digest,

Table 1. Comparison between the Sequence Coverage Achieved with On-Target and In-Solution (with ZipTip Cleanup) Tryptic Digestion of Ricin (500 fmol) Using MALDI MS tryptic peptides trypsin digestion method

A chain

B chain

total protein sequence coveragea

on-target in-solution

10 12

6 5

54.10% 55.60%

a

The sequence coverage is determined on the processed protein.

respectively. However, upon closer inspection of the spectra, these peaks were present but were of insufficient quality to be picked by the software, but they were easily identifiable with a manual search. In contrast, the acetonitrile solvent-assisted on-target digestion produced only four identifiable tryptic peptides from ricin. This included three of the peptides found in both the ricin and aggluttinin protein. Thus, only one tryptic peptide (Figure 1, peptide 10) was specific to ricin, which is insufficient to confidently identify ricin. The discrepancy in the results is most likely explained by the robustness of the spots during the incubation period. The water and aqueous methanol drops remain hydrated throughout the deposition and subsequent incubation steps. Conversely, the drops that contained 50% acetonitrile were very prone to evaporation with the spots having evaporated extensively after the 30 min incubation. This reduced fluidity of the spot adversely affects the efficiency of the enzymatic digestion resulting in fewer observed tryptic peptides. Further experimentation was carried out to optimize the conditions for the reaction. Two experimental conditions were evaluated: incubation time and analyte/trypsin ratio. A series of experiments were run varying the incubation time from as little as 5 min to 1 h using 0.1 µg of trypsin in 25 mM ammonium bicarbonate (Figure S-2 in the Supporting Information). With the exception of peptide 2 (Figure 1), all the other peptides are observed after only 5 min of digestion using either an aqueous or methanol-assisted digestion. Peptide 2 was observed after 10 min of digestion under methanol assisted digestion and after 30 min using an aqueous digestion. No additional tryptic peptides were observed if the digestion was incubated for longer periods up to an hour. However, optimal peptide intensity appeared to be achieved after 30 min. Varying the analyte/trypsin ratio had a more profound result on the observed peptides (Figure S-3 in the Supporting Information). The data was collected using the method described above with an incubation time of 30 min but using more dilute solutions of trypsin. The data described previously used an analyte/trypsin ratio of 1:3 (w/w). A ratio of 1:1 (w/w) produced a mass spectrum containing the same number of tryptic

fragments of ricin. These data indicate that an excess of trypsin produces an abundance of tryptic peptides related to ricin and with no apparent suppression effects resulting from autolysis products of the trypsin. However, ratios of 5:1 and 10:1 (w/w) only produced two tryptic peptides (peptides 6 and 9) with very small absolute intensities. Ratios of 20:1 and 50:1 (w/w) produced no observable signal for any of the tryptic peptides of ricin. These results demonstrate that on-target tryptic digestion can be used as an identification method for ricin and that results similar to in-solution solvent-assisted digestion can be obtained with ontarget digestion. On the basis of the data collectively, the following experimental conditions would produce a robust and reproducible method: methanol-assisted digestion using an incubation time of 30 min with an analyte/trypsin ratio of greater than 1:1 (w/w). One advantage of on-target tryptic digestion is that it provides sequence information via the tryptic maps that is critical for correct protein assignment. This ability is invaluable to be able to differentiate ricin from other proteins that exhibit the same activity or physical properties. For instance, the intact protein MALDI mass measurement method uses the size of the protein for identification. Table 2 lists four proteins that are similar to ricin and could potentially result in a false positive hit. Three of the proteins are type 2 RIPs. Agglutinin is also isolated from the castor bean seed while the other two can be isolated from the seeds of the Indian licorice (abrin) and European mistletoe (viscumin). Each of these proteins have a similar 3D structure to ricin with agglutinin being the most diverse, consisting of two identical dimers. The other protein in the table is bovine serum albumin that, while very different in 3D structure, has a molecular weight similar to that of ricin. This protein is very abundant within the environment. Figure 2 demonstrates the ability of the on-target tryptic digestion method to differentiate spectra from analytes of similar molecular weight that appear similar using the intact Protein MALDI mass measurement. On the left of the figure are the intact protein MALDI mass measurement spectra for each of the proteins. With the exception of agglutinin, all the spectra have very similar ionization patterns. They each have a MH+ peak at ∼m/z 63 000 and the peaks corresponding to the (M + 2H)2+, (M + 3H)3+, and (M + 4H)+4 species. The abrin intact mass spectrum shows a slight variation as it appears that the MH+ is actually two peaks with similar m/z: 62 400 and 60 900. This is also reflected in the (M + 2H)2+ peak. Even though the masses are different on each of the spectra, they are similar enough due to insufficient peak resolution and mass accuracy to categorically distinguish which sample is ricin. Conversely, the intact protein mass spectrum of agglutinin has a MH+ and (M + 3H)3+ at m/z 127 200 and 42 600, respectively, both of which are absent in any of the other mass spectra. This allows

Table 2. Summary of the Proteins Used to Evaluate the Specificity of the On-Target Tryptic Digestion Method for Ricin protein

species

abrin agglutinin ricin serum albumin viscumin

Abrus precatorium (Indian licorice) Ricinus communis (castor bean) Ricinus communis (castor bean) Bos taurus (bovine) Viscum album (European mistletoe)

description type 2 ribosome: inactivating type 2 ribosome: inactivating type 2 ribosome: inactivating the main protein in plasma type 2 ribosome: inactivating

protein subfamily protein subfamily protein subfamily protein subfamily

Analytical Chemistry, Vol. 82, No. 12, June 15, 2010

5249

Figure 2. (Left) The mass spectra obtained on the autoflex II TOF/TOF mass spectrometer from the intact protein MALDI mass measurement of 500 fmol of purified ricin, RCA120, abrin, viscumin, and BSA. (Right) The mass spectra obtained on the autoflex II TOF/TOF mass spectrometer from the on-target tryptic digestion of purified ricin, agglutinin, abrin, viscumin, and BSA (1 pmol) using trypsin dissolved in water/methanol (1:1) buffered to pH 8. The resulting peaklists were searched against the SwissProt database using MSFit (Protein Prospector). Masses that matched the protein of interest are indicated by an asterisk. The text in each spectrum lists the protein accession number, identified peaks compared to total peaks submitted, and the overall sequence coverage. Peaks labeled with a underlined asterisk in the ricin spectrum represent peaks derived from ricin but not identified due to glycosylation. The abrin spectrum contains peptides from three forms of the protein indicated by asterisks, cross hatches, and caps. All spectra have been magnified ×3 from m/z 2500.

this spectrum to be identified as agglutinin. However, this information only would be insufficient to identify this protein unequivocally if the sample were an unknown mixture. On the right of the figure are summarized the mass spectra from ontarget tryptic digestions of the five proteins. These mass spectra each exhibit a different ionization pattern that makes distinguishing them easier. When the peaklists for each mass spectrum are submitted to a Peptide Mass FingerPrinting algorithm (MSFit, Protein Prospector), the protein is easily identified as the first hit accounting for more than a third of all the masses submitted (Figure 2 right annotation). The sequence coverage achieved for each of the proteins is in excess of 30% and does not include tryptic peptides that are glycosylated, such as the two indicated in the ricin example. There is enough sequence coverage to distinguish between ricin and agglutinin; 5250

Analytical Chemistry, Vol. 82, No. 12, June 15, 2010

however, each are reported as lesser hits in the other’s protein lists due to the five common tryptic peptides. In addition, the abrin search identifies the presence of three forms of the protein (b, c, and d). Interestingly, abrin-b and -d have a reported molecular weight of the unprocessed protein of around 60 kDa; abrin-c is about 2 kDa larger. This is about the same difference between two apexes in the intact protein mass spectrum of the MH+ species, suggesting that the intact protein method has sufficient resolution to indicate the presence of the larger form. Can a purely mass spectrometric method be used to identify ricin in a “typical” field sample? To evaluate this scenario, a crude ricin extract prepared by one of the open access methods was obtained and analyzed as described in the methods. The mass spectra obtained from the ×5 and ×500 serial dilutions in 0.01% Tween 80 using the intact protein method and the neat and ×500

of a crude castor bean extract.20 Interestingly, it appears that the tryptic digestion has been unable to cleave the first two residues of each of the peptides. This data is consistent with the intact protein data indicating that the sample is a crude castor bean extract. The mass spectrum of the on-target tryptic digestion of the supernatant diluted 500 times contains a greater number of peaks. The peaks at m/z 896.6, 1013.6, 1310.6, 1862.0, 3307.6, and 3675.7 are all consistent with tryptic peptides observed from ricin (Figure 1). In addition, the peaks at m/z 1074.6, 1728.9, and 2259.2 are consistent with the tryptic peptides found in both ricin and agglutinin. Two further peaks at m/z 1889.0 and 2380.1 represent tryptic peptides from agglutinin. Interestingly, the two peaks representing the defensins as observed in the neat sample are now absent in either the digested or nondigested forms. This suggests that either of these compounds are not very abundant in the crude extract or that they do not ionize in the presence of other peptides. Therefore, the combined data from both the intact protein MALDI mass measurement and the on-target tryptic digestion confirms that the sample is a crude castor bean extract and that it contains intact ricin.

Figure 3. Top two panes show the intact protein MALDI mass measurement spectra for the ×5 and ×500 dilution of the crude extract in 0.01% Tween 80. The bottom two panes show the on-target tryptic digestion mass spectra for the neat and ×500 dilution of the crude extract. The peaks representing the defensins are annotated on the neat spectrum. In the bottom pane, masses indicated in bold are tryptic peptides associated with ricin, normal text are those associated with ricin and agglutinin, and those in italics are associated with agglutinin.

serial dilution in water using the on-target tryptic digestion are shown in Figure 3. The intact protein mass spectrum of the ×5 dilution in 0.01% Tween 80 is dominated by the singly charged oligomers of the seed storage proteins (SSPs). The presence of the SSPs is a strong indicator that this is a crude castor bean extract, but no peaks are observed for any of the charge states expected for ricin. However, when the supernatant is diluted 500 times, the spectrum looks significantly different. The intensity of the peaks of the singly charged oligomers of the seed storage proteins are substantially reduced with the peak for the tetramer unobserved and that for the trimer less intense than in the ×5 dilution even though that part of the m/z scale has been multiplied 10 times. More importantly, a peak is now observed at m/z 63 700 and a narrower peak at 31 600 which is consistent with what would be expected for the MH+ and (M + 2H)2+ of ricin, suggesting that the extract may contain the protein. The on-target tryptic digestion of the neat supernatant produces only two peaks at m/z 2067.0 and 1979.9. These values are consistent with the MH+ of the defensins that have been shown to be characteristic

DISCUSSION Ricin is a toxic protein produced in the seeds of the castor bean plant and is a member of the type 2 RIPs family of proteins. This family of proteins share common structural motifs and sequence homologies with each other and other proteins that affect the ability of immunological assays and intact protein mass measurement to distinguish between the proteins. On-target tryptic digestion provides specific sequence information via tryptic maps that enabled the identification of ricin from the other RIPs investigated: abrin, agglutinin, and viscumin. Ricin was also able to be distinguished from BSA, an abundant protein of similar molecular weight. The protein sequence coverage obtained was enhanced using a methanol-assisted tryptic digestion which was particularly versatile for identifying the protein in complex matrixes. An automatic search algorithm was able to identify all five proteins from a database, demonstrating the potential of this method for high throughput sampling. This method would be applicable for the rapid analysis of a large number of samples to ascertain those that contain ricin, as would be required in a triaging exercise. To produce an unambiguous identification of ricin, for use in a criminal court or by the Organisation for the Prohibition of Chemical Weapons (OPCW), the method could easily be extended by the collection of MSMS mass spectral data for the identified tryptic peptides to confirm the amino acid sequence. On-target tryptic digestion MALDI MS in conjunction with intact protein MALDI mass measurement is a robust method that can be used to identify the intact ricin protein in a crude extract. The duality of methods has several advantages. First, only one instrumental platform is required for both analyses. Second, the same target can be used for both analyses at the same time, requiring only one auto acquisition file. This was achieved by performing the on-target tryptic digestion first and following the incubation period preparing the spots used for the intact protein mass measurement. Third, the sample preparation required for (20) Ovenden, S. P.; Fredriksson, S. A.; Bagas, C. K.; Bergstrom, T.; Thomson, S. A.; Nilsson, C.; Bourne, D. J. Anal. Chem. 2009, 81, 3986–3996.

Analytical Chemistry, Vol. 82, No. 12, June 15, 2010

5251

the crude sample analysis was carried out in about 1 h, and the only sample preparation required was sample dilution into the relevant buffers. This robustness is essential if the method is to be adaptable to the variety of samples. One of the great advantages of the on-target tryptic digestion is that the resulting digest can be analyzed without the requirement to remove salts beforehand. This saves a substantial amount of time particularly when there are a larger number of samples to be processed. Finally, each method provides complementary information. The on-target tryptic digestion identifies the ricin using the tryptic map while the intact protein MALDI mass measurement demonstrates that the disulfide bond between the A and B chains is still intact. This bond must remain intact to preserve the toxicity of the ricin toxin. The importance of this bond is demonstrated in the observed activity of the type 1 RIPs. These proteins consist only of an A chain and, thus, have difficulty entering into the cells. As a result, they are much less toxic than type 2 RIPs but become highly toxic if they are introduced into cells by other means.21 This method could also be applied for the identification of ricin from “white powder” incidents with no adaptation.

sufficient information to enable ricin to be distinguished from the other type 2 RIPs: abrin, viscumin, and agglutinin. The peaklist can be searched manually or using an automated Peptide Mass FingerPrinting algorithm such as MS-Fit (Protein Prospector). The method is robust and can be performed with minimal sample preparation and without cleanup following tryptic digestion. In conjunction with intact protein MALDI mass measurement, it was used to identify the presence of intact ricin in a crude extract produced using an “open source” recipe.

ACKNOWLEDGMENT I would like to thank Julie Peeler and Kim Taylor for the production of the crude castor bean extract. SUPPORTING INFORMATION AVAILABLE Further details are given in Figures S-1, S-2, and S-3. This material is available free of charge via the Internet at http://pubs.acs.org.

CONCLUSION On-target tryptic digestion can be used to identify ricin using the resulting tryptic map that is produced. These maps provide

Received for review March 11, 2010. Accepted May 12, 2010.

(21) Stirpe, F. Toxicon 2004, 44, 371–383.

AC100650G

5252

Analytical Chemistry, Vol. 82, No. 12, June 15, 2010