Anal. Chem. 2005, 77, 6140-6146
Detection of Botulinum Neurotoxin A in a Spiked Milk Sample with Subtype Identification through Toxin Proteomics Suzanne R. Kalb,† Michael C. Goodnough,‡ Carl J. Malizio,‡ James L. Pirkle,† and John R. Barr*,†
National Center for Environmental Health/Agency for Toxic Substances and Disease Registry, Centers for Disease Control and Prevention, 4770 Buford Highway, NE Atlanta, Georgia 30341-3724, and Metabiologics, 505 South Rosa Road, Madison, Wisconsin 53719
Botulinum neurotoxin (BoNT) causes the disease botulism, which can be lethal if untreated. Rapid determination of exposure to BoNT is an important public health goal. Previous work in our laboratory focused on the development of Endopep-MS, a mass spectrometry-based endopeptidase method for detecting and differentiating BoNT in buffer. This method can rapidly determine the presence of BoNT in a sample and differentiate the toxin type of BoNT present but does not yield additional information about the subtype. We now describe here the application of Endopep-MS to detect BoNT A in a spiked milk sample. This work also describes subtype identification achieved through mass spectrometric analysis of the protein toxin itself and does not require the presence of DNA from the toxin-producing bacteria. Tryptic digests of A1 and A2 subtypes of BoNT were analyzed by mass spectrometry, and peptides unique to either the A1 or A2 subtype were subjected to tandem mass spectrometry analysis to confirm their identities. Finally, subtype identification through mass spectrometric analysis was performed on BoNT A isolated from spiked milk. In its entirety, this method would allow for analysis of BoNT with toxin type identification in a few hours and subtype identification within 24 h.
Botulinum neurotoxin (BoNT) is produced by some species of the genus Clostridium, particularly Clostridium botulinum, Clostridium butyricum, and Clostridium baratii.1 BoNTs cause the disease known as botulism, which, in most cases, is contracted through consumption of food containing the bacterium and its associated toxin.2 Botulism in humans is most often caused by BoNT A, B, E, or F, and a majority of the foodborne cases in the United States have been caused by BoNT A.3 Other sources of botulism involve colonization of the bacteria in the gastrointestinal * Corresponding author. Phone-770-488-7848. Fax-770-488-4609. E-mail:
[email protected]. † Centers for Disease Control and Prevention. ‡ Metabiologics. (1) Schiavo, G.; Matteoli, M.; Montecucco, C. Physiol. Rev. 2000, 80 (2), 717766. (2) Centers for Disease Control and Prevention. Botulism in the United States: 1899-1996. Handbook for Epidemiologists, Clinicians and Laboratory Workers; Centers for Disease Control and Prevention: Atlanta, GA, 1998.
6140 Analytical Chemistry, Vol. 77, No. 19, October 1, 2005
tract of infants or immunocompromised individuals and wound botulism obtained through contact of the bacterium with a wound.2,4 Although the toxicity of BoNT in humans has not been studied directly, the toxicity can be estimated from primate studies. As little as 0.001 µg/kg of body weight is estimated to be toxic when administered intravenously or intramuscularly.5,6 Oral toxicity is estimated to be 1 µg/kg,5 and inhaled toxicity is 0.01 µg/kg.7 Because of its extreme toxicity, high availability, and ease of preparation, BoNT is considered a likely agent for bioterrorism.8 Treatment of botulism involves administration of an equine-based antineurotoxin immunoglobulin (IgG), and treatment is most effective when administered within 24 h of exposure.2,9 Therefore, rapid determination of exposure to BoNT is an important public health goal. We have previously reported on the development of the Endopep-MS method for BoNT identification and differentiation.10,11 This method utilizes the endoproteinase activity of the toxin and has proven successful for detecting all seven toxin types of BoNT (A-G). It involves the addition of BoNT to a peptide substrate that mimics the toxin’s natural target, synaptosomeassociated protein (SNAP-25) or vesicle-associated membrane protein 2 (VAMP-2), also known as synaptobrevin 2. BoNT cleaves the peptide substrate in a specific, toxin-dependent location. The (3) Sobel, J.; Tucker, N.; Sulka, A.; McLaughlin, J.; Maslanka, S. Emerg. Infect. Dis. 2004, 10 (9), 1606-1611. (4) Werner, S. B.; Passaro, D.; McGee, J.; Schechter, R.; Vugia, D. J. Clin. Infect. Dis. (Epub) 2000, 31 (4), 1018-1024. (5) Herrero, B. A.; Ecklung, A. E.; Streett, C. S.; Ford, D. F.; King, J. K. Exp. Mol. Pathol. 1967, 6, 84-95. (6) Scott, A. B.; Suzuki, D. Mov. Disord. 1988, 3 (4), 333-335. (7) Franz, D. R.; Pitt, L. M.; Clayton, M. A.; Hanes, M. A.; Rose, K. J. In Botulinum and Tetanus Neurotoxins: Neurotransmission and Biomedical Aspects; DasGupta, B. R., Ed.; Plenum Press: New York, 1993; pp 473476. (8) Arnon, S. S.; Schechter, R.; Inglesby, T. V.; Henderson, D. A.; Bartlett, J. G.; Ascher, M. S.; Eitzen, E.; Fine, A. D.; Hauer, J.; Layton, M.; Lillibridge, S.; Osterholm, M. T.; O’Toole, R.; Parker, G.; Perl, T. M.; Russell, P. K.; Swerdlow, D. L.; Tonat, K. Working Group on Civilian Biodefense. JAMA, J. Am. Med. Assoc. 2001, 285 (8), 1059-1070. (9) Caya, J. G.; Agni, R.; Miller, J. E. Arch. Pathol. Lab. Med. 2004, 128 (6), 653-662. (10) Barr, J. R.; Moura, H.; Boyer, A. E.; Woolfitt, A. R.; Kalb, S. R.; Pavlopoulos, A.; McWilliams, L. G.; Schmidt, J. G.; Martinez, R. A.; Ashley, D. L. Emerg. Infect. Dis. [serial on the Internet]. Oct 2005. Available from http:// www.cdc.gov/ncidod/EID/vol111no10/04-1279.htm. (11) Boyer, A. E.; Moura, H.; Woolfitt, A. R.; Kalb, S. R.; Pavlopoulos, A.; McWilliams, L.; Schmidt, J. G.; Barr, J. R. Anal. Chem. 2005, 77, 39163924. 10.1021/ac0511748 CCC: $30.25
© 2005 American Chemical Society Published on Web 09/03/2005
reaction mixture is then introduced into a mass spectrometer, which detects the presence of any peptides present in the mixture and accurately reports the mass of each peptide. The presence of peptide-cleavage products corresponding to their toxin-dependent location indicates the presence of a particular toxin type of BoNT. If the peptide substrate either remains intact or is cleaved at a site that is not toxin type-specific, then that toxin type is not present at a detectable level. Previous publications10,11 have demonstrated that this method is able to rapidly detect BoNT at levels comparable to or lower than levels detected with mouse bioassays, the current standard for BoNT analysis.12 We describe here a 30-min Endopep-MS reaction that detects BoNT A spiked into milk. Although the Endopep-MS reaction effectively identifies the toxin type of BoNT present in a sample, it cannot identify the toxin subtype. Identifying the subtype of BoNT can help determine whether multiple samples originate from a single source and might yield information about the source of the toxin for epidemiologic or forensic purposes. Typically, subtype identification is determined through DNA analysis via polymerase chain reaction (PCR)13-16 or more recently through real-time PCR (RT-PCR).17 These methods can only be used if the bacterium that produces the toxin is present. However, BoNT can potentially be present in a sample that does not contain the bacterium. In such a situation, subtype identification would not be possible using traditional, DNA-based methods. Here, we describe a method to identify the subtype using the toxin protein itself rather than bacterial DNA. This method involves tryptic digestion of the toxin and mass spectrometric analysis of the tryptic fragments. In its entirety, this method would allow analysis of BoNT with toxintype identification in a few hours and subtype identification within 24 h. EXPERIMENTAL SECTION Materials. Botulinum neurotoxin is very toxic and therefore requires appropriate safety measures. All neurotoxin was handled within a level 2 biosafety cabinet equipped with HEPA filters. BoNT A complex and polyclonal rabbit IgGs were obtained from Metabiologics (Madison, WI). The BoNT A complex was provided at 1 mg/mL total protein in 0.02 M sodium phosphate buffer, pH 7.0. The polyclonal rabbit IgGs were provided in 150 mM potassium phosphate, pH 7.4, at 4.61 mg/mL. Dynabeads Protein G were purchased from Dynal (Lake Success, NY) at 1.3 g/cm3 in phosphate-buffered saline (PBS), pH 7.4, containing 0.1% Tween20 and 0.02% sodium azide. Milk with 2% fat was obtained from retail sources and used as purchased. All chemicals were from Sigma-Aldrich (St. Louis, MO) except where indicated. Los Alamos National Laboratory (Los Alamos, NM) synthesized the HPLCpurified peptide substrate (biotin- KGSNRTRIDEANQRATRMLGGK-biotin). (12) Kautter, D. A.; Solomon, H. M. J. Assoc. Anal. Chem. 1977, 60, 541-545. (13) Szabo, E. A.; Pemberton, J. M.; Desmarchelier, P. M. Appl. Environ. Microbiol. 1992, 58 (1), 418-420. (14) Szabo, E. A.; Pemberton, J. M.; Desmarchelier, P. M. Appl. Environ. Microbiol. 1993, 59 (9), 3011-3020. (15) Fach, P.; Hauser D.; Guillou, J. P.; Popoff, M. R. J. Appl. Bacteriol. 1993, 75 (3), 234-239. (16) Fach, P.; Gibert, M.; Griffais, R.; Guillou, J. P.; Popoff, M. R. Appl. Environ. Microbiol. 1995, 61 (1), 389-392. (17) Lovenklev, M.; Holst, E.; Borch, E.; Radstrom, P. Appl. Environ. Microbiol. 2004, 70 (5), 2919-2927.
Tryptic Digests of A1 and A2 Subtypes of BoNT A. Type A neurotoxins were purified according to the methods listed in refs 18-20. A 20-µg aliquot of each subtype of BoNT A (purified neurotoxin) at 1 mg/mL was added to 20 µL of acetonitrile and incubated at room temperature for 30 min. Then, a 140-µL solution of 50 mM ammonium bicarbonate (tryptic digest buffer) and 20 µL of trypsin, diluted to 10 µg/mL in tryptic digest buffer, were added to produce a final solution of 100 µg/mL BoNT A. The solutions were incubated overnight at 37 °C. MALDI Analysis of Tryptic Digests. A 10-µL portion of each digestion was purified in a C18 minicolumn (C18 Zip Tips, Millipore, Bedford, MA) using the standard protocol obtained from Millipore. The peptides were eluted in 2 µL of 80% acetonitrile/0.1% trifluoroacetic acid (TFA) and mixed with 8 µL of R-cyano-4hydroxycinnamic acid (CHCA) at 5 mg/mL in 50% acetonitrile, 0.1% TFA, and 1 mM ammonium citrate (CHCA matrix). This mixture was then applied in duplicate at 0.5 µL/sample well to a 192-spot stainless steel matrix-assisted laser desorption/ionization (MALDI) plate (Applied Biosystems, Framingham, MA). Mass spectra of each sample well were obtained by scanning from 650 to 4500 m/z in MS positive-ion reflector mode on the Applied Biosystems 4700 Proteomics Analyzer (Framingham, MA). The instrument uses a nitrogen laser at 337 nm, and each spectrum is an average of 2400 laser shots. Peptides believed to originate from BoNT A were subjected to MS/MS analysis to confirm identity through sequence determination. LTQ-FT Analysis of Tryptic Digests. A 10-µL portion of each digestion was injected onto a PicoFrit ProteoPep II C18 column, 75-µm i.d. and 5 cm long (New Objective, Woburn, MA). After a sample-loading flow rate of 2 µL/min for 10 min, the peptides were eluted from the column at a flow rate of 200 nL/min using the following gradient conditions, where A is water with 0.25% formic acid and B is 80% acetonitrile and 0.25% formic acid: A ) 100% and B ) 0% at 0 min; A ) 10% and B ) 90% at 20 min; A ) 0% and B ) 100% at 22 min; A ) 0% and B ) 100% at 24 min; A ) 100% and B ) 0% at 30 min. Peptides eluting from the column were introduced into a 7-T LTQ-FT instrument (Thermo Electron, San Jose, CA) using a standard nanoelectrospray ion source interface. The instrument was operated in data-dependent acquisition mode to automatically switch between MS and MS/MS analysis. The FT-MS was used for MS acquisition with a resolution of 12 500 from 400 to 2000 m/z. Simultaneously, the linear ion trap was used for MS/MS analysis of the three most abundant ions in each MS scan. Automatic gain control was used to accumulate ions for FT-MS analysis with a target value of 1 000 000 and a target value of 2000 for MS/MS analysis. A collision energy of 30% was utilized in the linear ion trap to fragment the tryptic digest fragments. All MS/MS spectra were merged to a single file that was searched using the Mascot Search Engine (Matrix Science Inc., Boston, MA) against a NCBI nonredundant database indexed for only Clostridia organisms with oxidized methionine as a variable modification. Tryptic constraints were applied, allowing up to two (18) Malizio, C. J.; Goodnough, M. C.; Johnson, E. A. In Methods in Molecular Biology: Bacterial Toxins, Methods, and Protocols, Walker, J., Holst, O., Eds.; Humana Press: Totowa, NJ, 1999; pp 27-39. 1999. (19) Tse, C.; Dolly, J.; Hambleton, P.; Wray, D.; Melling, J. Eur. J. Biochem. 1982, 122, 493-500. (20) Woody, M.; DasGupta, B. J. Chromatogr. 1988, 430, 279-289.
Analytical Chemistry, Vol. 77, No. 19, October 1, 2005
6141
Figure 1. MALDI mass spectra of a tryptic digest of the A1 (A) and A2 (B) subtypes of BoNT A. Most significant differences between the two spectra are marked with asterisks. Table 1. Peptides Unique to Either A1 or A2 Toxin Subtypes of BoNT A Observed by MALDI Listed with Their Observed m/z Valuesa digest fragment
m/z obs of A1
sequence
G827-R835 S166-R176 Y882-R892 V381-R392 I565-R580 R806-K820 G1140-R1155 F212-R230
958.3 1272.3 1346.3 1474.4 1711.5 1733.5 1818.4 2055.6
GTLIGQVDR SFGHEVLNLTR YESNHLIDLSR VNYTIYDGFNLR IALTNSVNEALLNPSR RLEDFDASLKDALLK GSVMTTNIYLNSSLYR FATDPAVTLAHELIHAGHR
digest fragment
m/z obs of A2
sequence
R1064-K1069 G827-R835 L807-R815 S166-R176 V720-K730 I23-K36 G393-R410 F212-R230 I897-K914
823.3 1000.3 1050.3 1258.3 1328.3 1529.4 1937.4 2127.6 2153.6
RYIMIK GTLVLQVDR LKDFDASVR SFGHDVLNLTR VNTQIDLIREK IPNAGQMQPVKAFK GANLSTNFNGQNTEINSR FATDPAVTLAHELIHAEHR INIGDRVYYDSIKNQIK
a Residues that differ between subtypes are underlined and listed in boldface type. The amino acid sequences are listed and are predicted from the genomic sequences.
missed cleavages. Peptide searches were performed with an initial tolerance on mass measurement of 0.5 Da in MS mode and 1 Da in MS/MS mode. Sample Binding and Endopep-MS Reaction. The immobilization of IgG to the Dynabeads protein G was performed as described in the manufacturer’s protocol. IgG-coated beads (100 µL) were washed 5 times with successive amounts of 500 µL of PBS-Tween buffer. A 500-µL sample of casein buffer (10 g of casein in 1 L of PBS) was added to the beads, and the resulting mixture was incubated at 37 °C for 30 min. Casein buffer was removed from the beads, and 500 µL of 2% milk spiked with 2 µg of BoNT A was added to the beads. The sample was then incubated at 37 °C for 2 h without agitation. Supernatant was then removed from the beads, which were washed 3 times with successive amounts of 500 µL of PBS-Tween buffer followed by 2 washes with successive amounts of 500 µL of water. Following removal of all the supernatant, 19 µL of the reaction buffer consisting of 0.05 M Hepes (pH 7.3), 25 mM dithiothreitol, 20 6142
Analytical Chemistry, Vol. 77, No. 19, October 1, 2005
mM ZnCl2, and 1 mg/mL bovine serum albumin and 1 µL of peptide substrate at 1 nmol/µL were added to the beads. The sample was then incubated at 37 °C for 30 min. After incubation, the reaction supernatant was combined with 2 µL of a 1% TFA solution to quench the reaction. The resulting mixture was then purified with C18 minicolumns and subjected to MALDI MS analysis as described above. Tryptic Digestion of BoNT A Isolated from Spiked Milk. After removing all the reaction supernatant, a solution of 10 µL of acetonitrile with 1% TFA was added to the beads and incubated at room temperature for 30 min to elute BoNT A from the antibody-coated beads. The supernatant was removed from the beads, evaporated, and reconstituted in 15 µL of tryptic digest buffer and 5 µL of trypsin, diluted as above. Following overnight digestion, the peptides were purified with a C18 minicolumn, and analysis of the peptides proceeded as before, including MS/MS analysis to confirm toxin subtype.
Table 2. Peptide Sequences and Their Observed m/z from the Tryptic Digests of the A1 and A2 Subtypes Observed with the LTQ-FT Instrument with MS/MS Verificationa digest fragment
m/z obs of A1
peptide sequence
Q6-K22
1984.018 QFNYKDPVNGVDIAYIK
I37-K65 Q66-K83 I97-R104 M105-R112 G106-K127 V128-R144 S166-R176 N177-R186 F187-K211 L231-K243 V244-R263 F272-R282 L283-K290 D291-K300 S301-K313 Y320-K329 F330-K336 L334-K332 M333-K358 T364-K374 I375-K380 V381-R392 N393-K414 L415-K426 V456-K476
3356.691 2108.966 924.479 932.560 1620.827 1906.933 1272.670 1158.554 2717.299 1504.864 2351.087 1397.670 1138.594 861.468 1398.730 1112.548 836.488 1054.630 2025.004 1345.716 683.446 1474.733 2470.142 1506.799 2516.137
IHNKIWVIPERDTFTNPEEGDLNPPPEAK QVPVSYYDSTYLSTDNEK IYSTDLGR MLLTSIVR GIPFWGGSTIDTELK VIDTNCINVIQPDGSYR SFGHEVLNLTR NGYGSTQYIR FSPDFTFGFEESLEVDTNPLLGAGK LYGIAINPNRVFK VNTNAYYEMSGLEVSFEELR FIDSLQENEFR LYYYNKFK DIASTLNK SIVGTTASLQYMK YLLSEDTSGK FSVDKLK LKFDKLYK MLTEIYTEDNFVKFFK TYLNFDKAVFK INIVPK VNYTIYDGFNLR NTNLAANFNGQNTEINNMNFTK LKNFTGLFEFYK VNNWDLFFSPSEDNFTNDLNK
I565-R580 V581-K591 A596-K625 V688-K700 N702-K710 Y711-K719 V720-K729 E733-K743 A744-K758 N759-K771 A779-K785 L807-K815 L807-K820 G827-K839 V840-K854 L861-K870 N871-R881 Y882-R892 I897-K910 V903-K922 I923-K928 N929-K950 V975-R995 Y1000-K1027
1711.934 1355.652 3383.552 1401.795 1210.575 1107.620 1199.711 1203.586 1905.887 1493.724 803.445 1037.516 1577.843 1442.833 1676.886 1214.667 1270.748 1346.670 1559.843 2349.209 714.477 2680.323 2534.325 3417.738
IALTNSVNEALLNPSR VYTFFSSDYVK ATEAAMFLGWVEQLVYDFTDETSEVSTTDK VLTVQTIDNALSK NEKWDEVYK YIVTNWLAK VNTQIDLIRK EALENQAEATK AIINYQYNQYTEEEK NNINFNIDDLSSK AMININK LEDFDASLK LEDFDASLKDALLK GTLIGQVDRLKDK VNNTLSTDIPFQLSK LLSTFTEYIK NIINTSILNLR YESNHLIDLSR INIGSKVNFDPIDK VNFDPIDKNQIQLFNLESSK IEVILK NAIVYNSMYENFSTSFWIRIPK VSLNYGEIIWTLQDTQEIKQR YSQMINISDYINRWIFVTITNNRLNNSK
L1034-K1055 Y1065-K1069 Y1070-K1081 E1082-K1097 Y1121-R1130 G1131-K1136 G1140-R1155 K1163-R1174 N1175-K1186 L1192-1203 I1204-K1233 M1236-K1259
2467.313 722.424 1559.775 1836.934 1148.606 774.386 1818.906 1364.729 1432.791 1188.622 3210.741 2721.284
LIDQKPISNLGNIHASNNIMFK YIWIK YFNLFDKELNEK EIKDLYDNQSNSGILK YVDVNNVGIR GYMYLK GSVMTTNIYLNSSLYR KYASGNKDNIVR NNDRVYINVVVK LATNASQAGVEK ILSALEIPDVGNLSQVVVMKSKNDQGITNK MNLQDNNGNDIGFIGFHQFNNIAK
digest fragment
m/z obs of A2
peptide sequence
Q6-K22 I23-K33 I37-K65 Q66-K83 I97-R104 M105-R112 G106-K127
1984.018 1182.631 3356.691 2108.966 924.479 932.560 1620.827
QFNYKDPVNGVDIAYIK IPNAGQMQPVK IHNKIWVIPERDTFTNPEEGDLNPPPEAK QVPVSYYDSTYLSTDNEK IYSTDLGR MLLTSIVR GIPFWGGSTIDTELK
S166-R176 N177-R186 F187-K211 L231-K243 V244-R263 F272-R282 L283-K290 D291-K300 S301-K313 Y320-K329 F330-K336 L334-K332 M333-K358 T364-R374 I375-K386 D387-K392 G393-R410 L415-K426 V456-K476 A555-R564
1258.654 1158.554 2717.299 1504.864 2351.087 1397.670 1138.594 847.453 1412.746 1112.548 836.488 1054.630 2010.952 1373.722 1418.753 693.357 1936.911 1506.799 2517.121 1175.508
SFGHDVLNLTR NGYGSTQYIR FSPDFTFGFEESLEVDTNPLLGAGK LYGIAINPNRVFK VNTNAYYEMSGLEVSFEELR FIDSLQENEFR LYYYNKFK DVASTLNK SIIGTTASLQYMK YLLSEDTSGK FSVDKLK LKFDKLYK MLTEIYTEDNFVNFFK TYLNFDKAVFR INIVPDENYTIK DGFNLK GANLSTNFNGQNTEINSR LKNFTGLFEFYK VNNWDLFFSPSEDNFTNDLDK AQEFEHGDSR
V688-K700 N702-K710 Y711-K719 V720-K730 K733-K743 A744-K758 N759-K771 L772-K785 L807-R815 D816-K820 G827-R835 L836-K854 L861-K870 N871-K883 K884-R892 I897-K914 L915-K928
1400.811 1210.575 1095.584 1328.754 1202.638 1905.887 1493.724 1560.806 1050.558 587.377 1000.579 2132.124 1214.667 1463.847 1074.579 2154.119 1571.926
VLTVQTINNALSK NEKWDEVYK YTVTNWLAK VNTQIDLIREK KALENQAEATK AIINYQYNQYTEEEK NNINFNIDDLSSK LNESINSAMININK LKDFDASVR DVLLK GTLVLQVDR LKDEVNNTLSADIPFQLSK LLSTFTEYIK NIVNTSILSIVYK KDDLIDLSR INIGDRVYYDSIDKNQIK LINLESSTIEVILK
V975-R995 V996-R1022 S1026-R1033 L1034-K1051 Y1065-K1069 Y1070-K1081 E1082-K1097 Y1121-R1130 G1131-K1136 G1140-K1158 K1163-R1174 N1175-K1186 L1192-1203 I1204-K1223 M1236-K1259 L1260-R1268
2532.321 3320.725 950.542 1962.077 667.385 1559.775 1809.924 1162.622 774.386 2060.055 1365.676 1432.791 1188.622 2125.194 2723.289 1122.569
VSLNYGEIIWTLQDNKQNIQR VVFKYSQMVNISDYINRWIFVTITNNR SKIYINGR LIDQKPISNLGNIHASNK YIMIK YFNLFDKELNEK EIKDLYDSQSNSGILK YVDVNNIGIR GYMYLK GSVVTTNIYLNSTLYEGTK KYASGNEDNIVR NNDRVYINVVVK LATNASQAGVEK ILSALEIPDVGNLSQVVVMK MNLQDNNGNDIGFIGFHLYDNIAK LVASNWYNR
a The amino acid sequences are predicted from the genomic sequences of X52066 (A1) and X73423 (A2). Residues underlined and in boldface type differ between the subtypes.
Analytical Chemistry, Vol. 77, No. 19, October 1, 2005
6143
RESULTS AND DISCUSSION A botulism outbreak is a public health emergency, and early diagnosis and treatment of the disease are critical to minimize the risk of death.2 Prompt epidemiology and diagnostic laboratory testing can be important for identifying the source and preventing further cases of botulism.2 Also, identifying BoNT in a clinical sample or remnants of food consumed by a patient can confirm a diagnosis of botulism.2 Additionally, BoNTs are now considered potential biological warfare and terrorist agents.8 Information gained on the toxin may promote an understanding of a botulism outbreak and suggest possible sources. It may also suggest commonality between botulism outbreaks that occur in different locations and may aid the epidemiologic investigation. We have developed a method called Endopep-MS that can detect and differentiate the BoNT toxin type in a variety of clinical and food samples by employing antibody-coated beads.21 In its current form, this method does not allow for identification of the subtype. Therefore, we devised an additional procedure that can determine the subtype of BoNT A to supplement the EndopepMS method. Amino acid sequences of BoNT A1 and A2 are predicted from DNA sequences of A1 and A2 subtypes that are ∼90% homologous.22 We decided to exploit these differences in amino acid sequence to identify the subtype through tryptic digestion. Following the protocol for tryptic digestion of the purified neurotoxin, two MALDI mass spectra depicted in Figure 1 were obtained. In the mass range of 700-3000 m/z, the mass spectra appear similar because of the 90% amino acid homology. However, some significant differences exist in the spectra, and several are highlighted. These differences can be used to determine the subtype as A1 or A2. Many of the peptides in both samples were subjected to MS/MS analysis to obtain sequence information for verification of the peptide’s identity. The identity of the peptide was predicted from the genomic sequence of X52066 (A1) or X73423 (A2). Table 1 lists m/z values of peptides present in either the A1 or A2 subtype, but not both, as well as their respective peptide sequences determined by MS/MS with amino acid residue substitutions indicated. Peptides listed in this table can therefore be used to distinguish by MALDI analysis between the A1 and A2 subtypes of BoNT A. There are 18 peptides present in both digests, and these can be used to determine the presence of BoNT A. Although MALDI analysis is rapid and sensitive, coverage of the protein sequence was limited to approximately 25-30% due in part to the large size of BoNT A (150 kDa). Despite the low sequence coverage, MALDI analysis yields enough peptide information to confirm the subtype as A1 or A2 as seen in Table 1. However, higher coverage would result in greater confidence of subtype identification and perhaps help identify new or unknown subtypes. Analysis of the tryptic digests on an LTQ-FT mass spectrometer following chromatographic separation of the peptides resulted in much higher coverage (65-70%) of the protein sequence. Table 2 lists peptide sequences and their observed m/z from the tryptic digests of the A1 and A2 subtypes observed with the LTQ-FT instrument with MS/MS verification. Residues that are amino acid substitutions found in the A2 subtype (21) Kalb, S. R.; Moura, H.; Boyer, A. E.; McWilliams, L. G.; Pirkle, J. L.; Barr, J. R. In preparation. (22) Willems, A.; East, A. K.; Lawson, P. A.; Collins, M. D. Res. Microbiol. 1993, 144, 547-556.
6144 Analytical Chemistry, Vol. 77, No. 19, October 1, 2005
Figure 2. MALDI mass spectrum of Endopep-MS reaction of BoNT A spiked into milk. The peptide substrate is present at m/z 2911.5, and the toxin cleavage products at m/z 1215.7 (C-terminal product) and 1714.9 (N-terminal product) are marked with asterisks.
are underlined and in boldface type. The amino acid sequences are predicted from the genomic sequences of X52066 (A1) and X73423 (A2). MS data correspond well to the sequence predicted from DNA for both A1 and A2. This type of analysis shows there are numerous peptides present that can be used to identify the subtype as A1 or A2. These data indicate that identification of the subtypes of BoNT A is possible from the analysis of tryptic fragments obtained by various mass spectrometric techniques. Results from this study also agree with the previous report of mass spectrometry analysis of an A1 tryptic digest of the BoNT A complex.23 Following tryptic digestion of the BoNT A complex, which consists of the neurotoxin as well as other associated proteins, the tryptic fragments were analyzed by MALDI-MS and LC-MS/MS. Because van Baar et al.23 reported on analysis of the complex rather than only the purified neurotoxin, many peptides are reported there that are not present in this analysis. However, several peptides are prominent in both studies, namely, the peaks at m/z 1272.3, 1711.5, 1474.1, 1397.3, and 2863.7. Our work reports the first protein sequencing applied to A2 and the first protein comparison of two subtypes of the same toxin type. After determining that mass spectrometric analysis of tryptic digest fragments of BoNT A yields subtype-specific information, we performed an experiment to examine whether this process could be used to determine the subtype of BoNT A present in a milk sample. Beads coated with antibodies to BoNT A were combined with a milk sample spiked with BoNT A complex. The antibody-coated beads served to concentrate the toxin and remove it from the abundant proteins endogeneous in milk. Before tryptic digestion however, the beads were subjected to the Endopep-MS reaction as described in the Experimental Section. Mass spectrometric analysis of the reaction supernatant is depicted in Figure 2. The peptide substrate is present at m/z 2911.5 singly charged and at m/z 1456.8 doubly charged. Cleavage of the peptide substrate by BoNT A results in two peptides with the sequences biotin-KGSNRTRIDEANQ and RATRMLGGK-biotin. These peptides are present at m/z 1714.9 and 1215.7, respectively. The presence of these two peaks indicates that BoNT is in the milk sample and that the toxin type is BoNT A. Following the Endopep-MS reaction, a tryptic digest was performed on the eluent of BoNT A from the antibody-coated beads that were used in the Endopep-MS assay. A MALDI mass (23) van Baar, B. L. M.; Hulst, A. G.; de Jong, A. L.; Wils, E. R. J. J. Chromatogr., A 2002, 970, 95-115.
Figure 3. MALDI mass spectrum of a tryptic digest of BoNT A recovered from spiked milk. Peaks marked with an asterisk are unique for identification of the A1 subtype. Table 3. Peptides and Their Observed MALDI m/z Values from a Tryptic Digest of BoNT A Spiked into Milka digest fragment
m/z obs of BoNT A
sequence
Y821-R826 I97-R104 M105-R112 V720-R728 Y547-R554 F272-R282 I41-K65
843.2 924.2 932.3 1071.3 1130.3 1397.3 2863.7
YIYDNR IYSTDLGR MLLTSIVR VNTQIDLIR YTMFHYLR FIDSLQENEFR IWVIPERDTFTNPEEGDLNPPPEAK
digest fragment
m/z obs of A1
sequence
G827-R835 S166-R176 V381-R392 I565-R580 F212-R230
958.3 1272.3 1474.4 1711.5 2055.6
GTLIGQVDR SFGHEVLNLTR VNYTIYDGFNLR IALTNSVNEALLNPSR FATDPAVTLAHELIHAGHR
a Peptides are listed from BoNT A and from the A1 subtype alone. Residues underlined and in boldface type differ between the subtypes. The amino acid sequences are listed and are predicted from the genomic sequences.
spectrum of the digest is in Figure 3. Peptides corresponding to either the A1 or A2 subtype as well as the A1 subtype alone are listed in Table 3. It is apparent that this subtype of BoNT A complex is the A1 subtype rather than the A2 subtype. MS/MS spectra were acquired for several of the peaks present in only the A1 subtype (data not shown) to confirm the identity of these peaks. This experiment demonstrates the ability to identify the subtype of toxin isolated from a milk sample as either A1 or A2. CONCLUSIONS The Endopep-MS assay can be used to quickly determine the presence or absence of BoNT A in a food sample. This analysis can be performed in as few as 3 h, and our laboratory is currently working to establish LODs of BoNT A, B, E, and F spiked into serum, stool extract, and food products using this method. The need for a rapid, activity-based assay for BoNT is critical, and we are developing a method to detect all seven toxin types of BoNT in clinical samples that rivals or exceeds the mouse bioassay, which defines 1 unit as one mouse LD50.24 Previous publications regarding the Endopep-MS method list the LODs of BoNT A, B, E, and F spiked into buffer as 1.25 units/mL for A, B, and E and 6.25 units/mL for F using a 4-h incubation and MALDI mass spectrometry for analysis.10-11 Currently, the Endopep-MS method exceeds the mouse bioassay in speed, throughput, and specificity. (24) Hatheway, C. L.; Ferreira, J. L. In Natural Toxins II; Singh, B. R., Tu, A. T., Eds.; Plenum Press: New York 1996; pp 481-498.
Tryptic digestion of the A1 subtype of BoNT A produces a set of peptides that are indicative of BoNT A and detectable using a MALDI-TOF mass spectrometer or an ion trap instrument. Furthermore, a portion of the peptides are not produced upon tryptic digestion of the A2 subtype of BoNT A. By examining the tryptic digest of BoNT A, a toxin can be first identified as BoNT A and then further identified as subtype A1 or A2 or perhaps as another, yet unknown, A subtype. Combination of the EndopepMS reaction with tryptic digestion of the toxin demonstrates the ability first to isolate BoNT A from a milk sample and rapidly determine the presence of BoNT A using the Endopep-MS method and second to determine the identity of that isolated BoNT A as either A1 or A2, all within 24 h. It should be noted that the toxin subtype identification requires much higher levels of toxin than the toxin type identification; the current LOD of subtype identification as A1 for example is ∼2 µg. Historically, most food samples have not been analyzed quantitatively for BoNT; however, this level of toxin is comparable to some reports of BoNT A present in food samples.25 Additionally, if the level of toxin is found to be inadequate for subtype identification, the sample can be cultured to produce additional amounts of toxin. Continued research is important in investigating the possibility of additional subtypes of BoNT A that have not yet been identified. (25) Kalluri, P.; Crowe, C.; Reller, M.; Gaul, L.; Hayslett, J.; Barth, S.; Eliasbery, S.; Ferreira, J.; Holt, K.; Bengston, S.; Hendricks, K.; Sobel, J. Clin. Infect. Dis. 2003, 37 (11), 1490-1495.
Analytical Chemistry, Vol. 77, No. 19, October 1, 2005
6145
By using this approach, we plan to investigate whether supplementary information can yield new clues into other subtype differences in BoNT A. In addition, we plan to examine tryptic digests of the other six toxin types of BoNT. Through this work, we hope to gain more information, which may aid an epidemiologic or forensic investigation of a botulism outbreak. Knowledge of the subtype of the toxin may help determine whether cases of
6146
Analytical Chemistry, Vol. 77, No. 19, October 1, 2005
botulism are related. It may also prove to be important forensically to help identify a common source in the event of multiple botulism outbreaks. Received for review July 1, 2005. Accepted August 15, 2005. AC0511748
