ilable at ScienceDirect

Toxicon 57 (2011) 730–738

Contents lists ava

Toxicon

journal homepage: www.elsevier .com/locate/ toxicon

Distinguishing the cyanobacterial neurotoxin b-N-methylamino-L-alanine(BMAA) from other diamino acids

S.A. Banack a, J.S. Metcalf a,b, Z. Spá�cil c, T.G. Downing d, S. Downing d, A. Long a,P.B. Nunn e, P.A. Cox a,*

a Institute for Ethnomedicine, Box 3464, Jackson, WY 83001, USAbNational Center for Natural Products Research, School of Pharmacy, University of Mississippi, Oxford, MS 38677, USAcDepartment of Chemistry, University of Washington, Seattle, WA, USAdDepartment of Biochemistry & Microbiology, Nelson Mandela Metropolitan University, P.O. Box 7700, Port Elizabeth 6031, South Africae School of Pharmacy and Biomedical Sciences, University of Portsmouth, Portsmouth, UK

a r t i c l e i n f o

Article history:Received 23 November 2010Accepted 8 February 2011Available online 15 February 2011

Keywords:Amino acidMass spectrometryUHPLCNeurodegenerative diseaseEnvironmental toxinDiaminopimelic acid

* Corresponding author. Tel.: þ1 801 375 6214; faE-mail address: paul@ethnomedicine.org (P.A. CURL: http://www.ethnomedicine.org

0041-0101/$ – see front matter � 2011 Elsevier Ltddoi:10.1016/j.toxicon.2011.02.005

a b s t r a c t

b-N-methylamino-L-alanine (BMAA) is produced by diverse taxa of cyanobacteria, and hasbeen detected by many investigators who have searched for it in cyanobacterial blooms,cultures and collections. Although BMAA is distinguishable from proteinogenic amino acidsand its isomer 2,4-DAB using standard chromatographic and mass spectroscopy techniquesroutinely used for the analysis of amino acids, we studied whether BMAA could be reliablydistinguished from other diamino acids, particularly 2,6-diaminopimelic acid which hasbeen isolated from the cell walls of many bacterial species. We used HPLC-FD, UHPLC-UV,UHPLC-MS, and triple quadrupole tandem mass spectrometry (UHPLC-MS/MS) to differ-entiate BMAA from the diamino acids 2,6-diaminopimelic acid, N-2(amino)ethylglycine,lysine, ornithine, 2,4-diaminosuccinic acid, homocystine, cystine, tryptophan, as well asother amino acids including asparagine, glutamine, and methionine methylsulfonium.

� 2011 Elsevier Ltd. All rights reserved.

1. Introduction

The possible role of the cyanobacterial toxin b-N-meth-ylamino-L-alanine (BMAA) in triggering neurodegenerativediseases such as amyotrophic lateral sclerosis, Alzheimer’sdisease, and Parkinson’s disease in vulnerable individualshas resulted in an increased interest in this unusualamino acid (Bradley and Mash, 2009). BMAA is present inenvironmental samples and human tissues both as a freeamino acid, and as a protein-bound amino acid (Murchet al., 2004a, b). It has been hypothesized that its mis-incorporation into polypeptide sequences results in proteinmisfolding and collapse. It has also been hypothesized that

x: þ1 801 373 3593.ox).

. All rights reserved.

misincorporation of BMAA triggers protein collapsewhich islinked to various tangle diseases characterized by irrevers-ible progressive neurodegeneration (Murch et al., 2004a, b;Bradley and Mash, 2009).

Studies of laboratory animals have shown that evena low rate of misincorporation of standard protein aminoacids can trigger neurodegenerative disease (Lee et al.,2006).

Misincorporation of the non-protein amino acid L-can-avanine results in anomalous proteins and toxicity inherbivores which forage on Canavalia ensiformis [Fabaceae](Allende and Allende, 1964; Rosenthal, 1977). Proteins withmisincorporated L-Dopa have been shown to be moreresistant to proteolysis and prone to cross-linking andaggregation (Rogers and Shiozawa, 2008).

BMAA is specifically toxic to motor neurons (Rao et al.,2006) as an agonist at NMDA and AMPA receptors, inducesoxidative stress, andpromotes accumulation of extracellular

S.A. Banack et al. / Toxicon 57 (2011) 730–738 731

glutamate through its action on the cystine/glutamateantiporter (system Xc

�) (Liu et al., 2009). Furthermore,BMAA potentiates neuronal damage from other neurotoxicinsults (Lobner et al., 2007). It has been hypothesized thatvulnerable individuals may be at increased risk of progres-sive neurodegenerative illness if exposed to BMAA (Coxet al., 2003; Banack and Cox, 2003a). Given the relativeubiquity of cyanobacteria, which have successfully colo-nized habitats ranging from Antarctica to the deserts of theMiddle East, andwhich often are present as blooms,mats, orscums in freshwater, estuarine, and marine waterbodies,there is a possibility of exposure to BMAA in disparate partsof the globe (Cox et al., 2005, 2009; Metcalf et al., 2008).

Due to the possible clinical implications of BMAAexposure and the wide array of compounds produced bycyanobacteria, it is important to distinguish BMAA fromother similar molecules. To date BMAA has been routinelydistinguished from other proteinogenic amino acids as wellas its neurotoxic isomer 2,4-diaminobutyric acid (2,4-DAB)by reversed-phase HPLC methods (Banack et al., 2007,2010) and with the most selectivity by UHPLC with triplequadrupole tandem mass spectrometry (UHPLC-MS/MS)using distinct retention times, diagnostic product ions andcharacteristic ratios between selective reaction monitoringtransitions (Banack et al., 2010; Spá�cil et al., 2010).

Another BMAA isomer, N-2(amino)ethylglycine, wasanalyzed along with other diamino acids and amides,which could potentially interfere with BMAA analysis. Wefocused on compounds that might occur in complexphysiological matrices, such as 2,6-diaminopimelic acid,found in the cell walls of many bacterial species (Work andDewey, 1953).

2. Materials and methods

2.1. Chemicals

L-BMAA was synthesized by Dr. Peter Nunn andcompared to Sigma B-107 (St. Louis, MO). 2,6-diamino-pimelic acid (synonym: 2,6-diaminoheptanedioic acid) fromSigma (D1377)was compared to ResearchOrganics # 2045D(Cleveland, OH). N-2(amino)ethylglycine was from TCIAmerica (Portland, OR); Part # A1153. L-lysine, L-ornithinemonohydrochloride, L-asparagine, L-glutamine, L-cystine,L-tryptophan, L-2,4-diaminobutyric acid dihydrochloride,methanol and acetonitrile (LC-MS CHROMASOLV�) werepurchased from Sigma–Aldrich (St. Louis, MO) as Part #’sL5501, O2375, A0884, G8540, C7602, T0254, 32830, 34966,and 34967 respectively. Ammonium formate (NC9841158)was from Fisher Scientific. 2,4-diaminosuccinic acid [syno-nym meso-A,A-diaminosuccinic acid] was purchased fromResearch Organics #0255D (Cleveland, OH). L-homocystinewas obtained from MP Biomedicals (Solon, OH); Part# 521061580. Formic acid was purchased from Acros Orga-nics (Geel, Belgium, distributed by Fisher Scientific, NJ,# 270480025). Proprietary compositions of UHPLC eluent A:part #186003838, UHPLC eluent B: part # 186003839, andHPLC eluent A part # 052890 were obtained from WatersCorp. (Milford,MA).Waterwas purchased from FisherW6-4Optima LC/MS for use with the LC/MS/MS or purified in-house (Direct-Q uv3, Millipore Ltd, Bedford, MA) to 18.2 MU

quality for all other purposes. For comparison purposes, DL-methioninemethylsulfonium chloride (Sigma # 64382) wasanalyzed.

2.2. Standard preparation

All compounds were reconstituted in 20 mMHCl exceptfor L-homocystine and 2,4-diaminosuccinic acid whichwere initially dissolved in 0.5 M NaOH before being dilutedwith water or 20 mM HCl for analysis. Each compound wasanalyzed both as an acid hydrolysate prepared in a finalconcentration of 6.0 M HCl for 16 h at 110 �C and withouthydrolysis. Samples were derivatized with 6-amino-quinolyl-N-hydroxysuccinimidyl carbamate (AQC WatersAccQTag reagent, PN WAT052880) or with propyl chlor-oformate (EzFaast�, LC/MS Physiological (Free) Amino AcidKit KH0-7337) following standardized, validated protocols(Banack and Cox, 2003a; Esterhuizen and Downing, 2008).Standards and blanks were derivatized and analyzedseparately and in combinations to evaluate chromato-graphic separation.

2.3. Ultra high pressure liquid chromatography-Tandem massspectrometry (UHPLC-MS/MS)

AQC derivatives were analyzed using a triple quadru-pole instrument (Thermo Scientific Finnigan TSQ QuantumUltraAM, San Jose, CA) after separation by a Waters Acq-uity-UHPLC systemwith a Binary Solvent Manager, SampleManager and a Waters AccQTag Ultra column (part#186003837, 2.1 � 100 mm) at 55 �C. Separation was ach-ieved using gradient elution at 0.65 ml/min in aqueous 0.1%(v/v) formic acid (Eluent A) and 0.1% (v/v) formic acid inacetonitrile (Eluent B): 0.0 min ¼ 99.1% A; 0.5 min ¼ 99.1%A curve 6; 2 min ¼ 95% A curve 6; 3 min ¼ 95% A curve 6;5.5 min ¼ 90% A curve 8; 6 min ¼ 15% A curve 6;6.5 min ¼ 15% A curve 6; 6.6 min ¼ 99.1% A curve6; 8min¼ 99.1% A curve 6. Nitrogen gas was supplied to theheated electrospray ionization (H-ESI) probe with a nebu-lizing pressure of 40 psi and a vaporizer temperature of400 �C. The mass spectrometer was operated under thefollowing conditions: the capillary temperature was set at270 �C, capillary offset of 35, tube lens offset of 110, auxil-iary gas pressure of 35, spray voltage 3500 V, source colli-sion energy of 0 eV, and multiplier voltage of �1719 V. Adivert valve was used except during the selected ionmonitoring (SIM) scans. The second quadrupole was pres-surized to 1.0 m Torr with argon. Ion m/z 459 was isolatedin the first quadrupole filter as the precursor ion and sub-jected to collision induced dissociation (CID). Second stepmass filtering was performed using selective reactionmonitoring (SRM) of BMAA after CID in the collision cell.The followingm/z transitions were monitored: 459–119, CE21 eV; 459–289 CE 17 eV; 459–171 CE 38 eV. The resultantthree product ions originating from derivatized BMAA(m/z 119, 289, 171) were detected, after entering the thirdquadrupole and their relative abundances were quantified.Compound specific m/z transitions were added to ourstandard method, 459–258 CE 21 eV (BMAA), 459–188 CE38 eV (DAB) and 459–214 CE 35 eV (AEG) to increaseselectivity. A new SRM method was established to detect

S.A. Banack et al. / Toxicon 57 (2011) 730–738732

2,6-diaminopimelic acid usingm/z 531 as the precursor ionand its most abundant product ions for m/z transitions531–191 CE 21 eV, 531–171 CE 38 eV, 531–361 CE 21 eV,531–387 CE 21 eV. In order to detect AQC-derivatizedcompounds of different molecular mass and thereforedetermine their retention times, the instrument was run inSIM mode using m/z 303 (Asn), 317 (Gln), 375 (Trp), 459(BMAA, 2,4-DAB, AEG), 473 (Orn), 487 (Lys), 489 (2,4-dia-minosuccinic acid), 531 (2,6-diaminopimelic acid), 581(Cys) and 609 (hCys). Common product ions for eachmolecular ion were established by sequential direct infu-sion of AQC-derivatized compounds using collision ener-gies 17, 21, 38 eV. Common product ions for the isomersBMAA, 2,4-DAB, AEG, as well as 2,6-diaminopimelic acidwere also established by sequential direct infusion ofunderivatized compounds using collision energies 17, 21,38 eV. Methionine methylsulfoniumwas analyzed with theprecursor ion set at m/z 334 using SIM mode and by infu-sion as an underivatized amino acid.

2.4. Ultra high pressure liquid chromatography-massspectrometry (UHPLC-MS), propyl chloroformatederivatization

Chloroformate (EzFaast�) derivatized samples wereanalyzed with a Thermo Scientific Finnigan TSQ QuantumUltraAM after separation by aWaters Acquity-UHPLC systemwith a Binary Solvent Manager, Sample Manager, and a Phe-nomenex AAA-MS 250 � 2.0 mm, 4 mm amino acid column(PN: 00G-4402-B0) with a flow rate of 0.25 ml/min, 10 mMammonium formate in water (Eluent A), 10 mM ammo-nium formate in methanol, and the following gradient:0.0 min ¼ 32% A; 0.01 min ¼ 32% A curve 6; 13 min ¼ 17%A curve 6; 13.01 min ¼ 32% A curve 6; 17 min ¼ 32% A.The instrument was run in Q1MS mode with nitrogen gassupplied to the H-ESI probe with a nebulizing pressure of40 psi and a vaporizer temperature of 199 �C. The massspectrometer was operated under the following conditions:the capillary temperaturewas set at 270 �C, capillary offset of35, aux gas pressure of 50, spray voltage 5000 V, sourcecollision energy of 0 eV, and multiplier voltage of �1302 V.Selective ion monitoring mode was used to monitorm/z 243(Asn), 275 (Gln), 333 (BMAA, 2,4-DAB, AEG), 347 (Orn), 361(Lys), 497 (2,6-diaminopimelic), and 525 (hCys) ions usingtimed segments to optimize instrument scans.

2.5. Ultra high pressure liquid chromatography withultraviolet detection (UHPLC-UV)

Chromatographic separation by UHPLC-UV with eluents(A and B) purchased from Waters (composition proprie-tary) were used according to manufacturers specification.Separation was achieved on a Waters Acquity-Ultra HighPerformance system with a Binary Solvent Manager,Sample Manager, and Sample Organizer by reverse phaseover 9.5 min on a Waters AccQTag Ultra column (part#186003837, 2.1 � 100 mm) at 55 �C (Cox et al., 2009). Thegradient was run with a flow rate of 0.7 ml/min(0.0 min ¼ 0.1% B; 0.54 min ¼ 0.1% B curve 6;4.04 min ¼ 6.5% B curve 6; 6.04 min ¼ 9.1% B curve 7;7.74 min ¼ 21.2% B curve 6; 8.04 min ¼ 59.6% B curve 6;

8.64 min ¼ 59.6% B curve 6; 8.73 min ¼ 0.1% B curve 6;9.5 min ¼ 0.1% B curve 6). The Waters Acquity-Ultra HighPerformance TUV detector was set at 260 nm.

2.6. Ultra high pressure liquid chromatography with massspectrometry detection (UHPLC-MS), AQC derivatization

Samples were analyzed with the above UHPLC front-end with a post-column split of the flow delivering a flowrate of 0.3 ml/min to aWaters EMD 1000 single quadrupolemass spectrometer. Nitrogen gas was purified to 97.8%using a NitroFlow nitrogen generator (Parker Balston,Haverfill, MA) and supplied to the electrospray ionizationinterface (ESI) at 101 psi with desolvation at 500 l/h. Posi-tive ion mode was used to detect derivatized samples usingselective ion monitoring at the following m/z 303 (Asn),304 (Asp), 317 (Gln), 318 (Glu), 376 (Trp), 459 (BMAA, 2,4-DAB, AEG), 473 (Orn), 487 (Lys), 489 (2,4-diaminosuccinicacid), 531 (2,6-diaminopimelic acid), 581 (Cys), 609 (hCys)with a span of 0.2 Da and a dwell time of 0.5 s. The sourcetemperature and desolvation temperature were 130 �C and100 �C respectively. Voltages were as follows: capillary2.9 kV, cone 27 V, extractor 2 V, RF lens 0.7 V.

2.7. High-performance liquid chromatography withfluorescence-detection (HPLC-FD)

BMAA was separated from other amino acids usinga validated method (Banack and Cox, 2003a,b, Banack et al.,2007) by reverse-phase elution (Waters Nova-Pak C18column, 3.9 � 300 mm at 37 �C) on a Waters 1525 binaryHPLC pump and a Waters 717 Autosampler. The mobilephase consisted of 140 mM sodium acetate, 5.6 mM trie-thylamine, pH 5.2 (Eluent A), and 52% (v/v) aqueousacetonitrile (Eluent B) using a flow rate of 1.0 ml/min, anda 10 ml sample injection. The samples were eluted usinga 60 min gradient: 0.0 min ¼ 100% A; 2 min ¼ 90% A curve11; 5 min ¼ 86% A curve 11; 10 min ¼ 86% A curve 6;18 min ¼ 73% A curve 6; 30 min ¼ 57% A curve 10;35 min ¼ 40% A curve 6; 37.5 min ¼ 100% B curve 6;47.5 min ¼ 100% B curve 6; 50 min ¼ 100% A curve 6;60 min ¼ 100% A curve 6. Detection was achieved usinga Waters 2475 Multi l-Fluorescence Detector with excita-tion at 250 nm and emission at 395 nm. Selected chro-matographic peaks were collected from HPLC-FD and themass was verified by injection into UHPLC-MS.

3. Results

Thirteen compounds were analyzed using five differentmethods adapted from Banack et al. (2007), Esterhuizenand Downing (2008), and Spá�cil et al. (2010). In addition,underivatized preparations of BMAA, 2,6-diaminopimelicacid, and methionine methylsulfonium were analyzed.Furthermore, each compound was analyzed using AQCderivatization before and after acid hydrolysis on HPLC-FD,UHPLC-UV, UHPLC-MS, and UHPLC-MS/MS. No appre-ciable difference was seen in retention times in any ofthe analytical platforms between hydrolyzed and unhy-drolyzed amino acids with the exception of asparagine andglutamine. Following hydrolysis these compounds were

S.A. Banack et al. / Toxicon 57 (2011) 730–738 733

found to revert to their corresponding aspartic (m/z 304)and glutamic acids (m/z 318). Chromatographic peaks forthese two compounds on the HPLC shifted from 12.47 minand 13.97 min (asparagine and glutamine, respectively) to11.31 min and 12.39 min (aspartic acid and glutamic acid,respectively). On the UHPLC-UV, retention times shiftedfrom 2.14 min and 2.59 min (asparagine and glutaminerespectively) to 3.08 min and 3.46 min (aspartic acid andglutamic acid respectively).

BMAA is consistently and clearly separated by HPLC andthus distinguishable from other tested diamino acids andamides by distinct retention time. In the case of methodsbased on tandem mass spectrometry, the selectivity isincreased by additionally using the mass-to-charge ratio ofprecursor and product ions as well as characteristic ratiosbetween selected SRM traces (Spá�cil et al., 2010). Chro-matographic separation of BMAA as the sole analyte, versusseparation of BMAA within a mixture of thirteencompounds did not affect the retention time of BMAA, itsproduct ion yields or their ratios, when UHPLC-MS/MS was

Fig. 1. UHPLC/MS/MS of a non-hydrolyzed 6-aminoquinolyl-N-hydroxysuccinimidylnature of UHPLC/MS/MS using a precursor ion m/z 459 and selective reaction morespectively only N-2(amino)ethylglycine (AEG), b-N-methylamino-L-alanine (BMA4.88, and 5.03 respectively. Distinct product ions and product ion ratios from col6-aminoquinolyl-N-hydroxysuccinimidyl carbamate (AQC) derivatized 2,6-diaminopmonitoring of transitions m/z 531 to m/z 171, 360, 386, and m/z 190 respectively re

used. Such observation was consistent with expectationand indicates that the presence of these other compoundsin themixture does not interferewith the reliable detectionof BMAA.

3.1. Tandem mass spectrometry

All of the compounds tested were clearly distinguishedfrom BMAA using standard methods (Banack et al., 2007;Spá�cil et al., 2010) based on triple quadrupole tandemmass spectrometry (UHPLC-MS/MS). In these methodsparent ions of AQC-derivatized 2,6-diaminopimelic acid,lysine, ornithine, homocystine, glutamine, cystine, 2,4-diaminosuccinic acid, tryptophan, asparagine, and methi-onine methylsulfonium were all immediately excluded bythe first quadrupole during SRM analysis, due to a differentmass-to-charge ratio of their ions with only ions atm/z 459subjected to CID at the second quadrupole. Diamino acidisomers of BMAA all provide ions at m/z 459 e.g. 2,4-DABand N-2(amino)ethylglycine (AEG), but they eluted at

carbamate (AQC) derivatized mix of 12 diamino acids. 1A: Given the specificnitoring of six transitions m/z 459 to m/z 171, 289, 119, 214, 258, and 188A), and 2,4-diaminobutyric acid (2,4-DAB) are visible at retention time 4.77,lision-induced dissociation distinguish these isomers. 1B: UHPLC/MS/MS ofimelic acid (DPM) run using a precursor ion of m/z 531 and selective reactionveal two isomers present in the mix and distinct product ions.

Table 1Triple quadrupole mass spectrometry of injected and infused non-hydrolyzed AQC-derivatized compounds under mass spectrometric conditions suitable todetect BMAA. Compounds were separated based on retention time, product ions, and product ion ratios. The five most abundant product ions, with areasminimally 1% of most abundant peak, are listed in order of abundance. Product ions m/z 171 and m/z 145 are fragments of the AQC tag and are not uniquelyidentifying. Ion ratios represent the relative ratio of the 119 and 289 ion, respectively, in comparison with the most abundant ion 171 (100%).

Diamino acid m/z Injection Infusion Injection Injection

RT Common ions – derivatized BMAA ion ratio (%) ofdaughter ion 119 cf. 171a

BMAA ion ratio (%) ofdaughter ion 289 cf. 171a

Asparagine 303 1.68 171, 133, 286Glutamine 317 1.99 171, 145, 276, 128, 147Methinonine methylsulfonium 334 0.90 129, 128, 198, 266, 143Tryptophan 375 6.13 171, 188, 159, 145, 205BMAA 459 4.88 171, 289, 119, 145, 258 11.2 � 0.6 25 � 1.62,4-DAB 459 5.03 171, 376, 315, 145, 336 4.9 � 0.5 7.4 � 0.4N-2(amino)ethylglycine 459 4.77 102, 119, 145, 171, 214 15.2 � 0.9 19.2 � 0.5Ornithine 473 5.77 171, 145, 329, 133, 70Lysine 487 6.05 171, 145, 84, 127, 1472,4-diaminosuccinic acid 489 3.29 171, 319, 145, 471, 1752,6-diaminopimelic acid 531 4.98, 5.38 171, 145, 191, 361Cystine 581 6.04 171,411,241, 203Homocystine 609 6.08 171, 439, 306, 272, 145

a Mean � SE.

S.A. Banack et al. / Toxicon 57 (2011) 730–738734

distinctly different retention times (Fig.1A). In addition, theratios from collision-induced dissociation of product ionsdiffer between these amino acids and BMAA (Table 1).

In order to increase specificity, a unique product quali-fier ion was selected for BMAA (m/z 258) and 2,4-DAB(m/z 188) as used by Spá�cil et al. (2010). We identifieda unique product qualifier for AEG (m/z 214; Fig. 1A). Othercompounds were analyzed with UHPLC-MS/MS by moni-toring additional SRM transitions, which allowed for thedetection of these compounds at the different retentiontimes on the LC front-end (Table 1). A new method wasestablished to examine 2,6-diaminopimelic acid using

Fig. 2. UHPLC/MS/MS separation of non-hydrolyzed propyl chloroformate (EzFaassimultaneous selective ion monitoring (SIM) of Asn (m/z 247), Gln (m/z 275), BMAA

m/z 531 as the precursor ion. Two distinct 2,6-dia-minopimelic acid chromatographic peaks were observedusing AQC derivatization, which correspond to differentisomers. Product ions 171 and 145 (Table 1) were fragmentsof the AQC tag itself and were not uniquely identifying.Derivatized 2,6-diaminopimelic acid provides uniqueproduct ions compared to derivatized BMAA (191, 361,m/z 387; Fig. 1B).

We characterized product ions of underivatized BMAAand 2,6-diaminopimelic acid byMS/MSproduct ion analysisusing direct infusion of standards into the triple quadrupoleworking under our standard instrument parameters,

t) derivatized diamino acids showing the total ion count TIC (top) and the, 2,4-DAB, AEG (m/z 333), Orn (m/z 347), Lys (m/z 361), 2,6-DPM (m/z 497).

Fig. 4. UHPLC-MS separation of non-hydrolyzed 6-aminoquinolyl-N-hydrox-ysuccinimidyl carbamate (AQC) derivatized b-N-methylamino-L-alanine(BMAA) and 2,4-diaminobutyric acid (2,4-DAB) (top) detected at m/z 459 inselective ion recording (SIR) mode. 2,6-Diaminopimelic acid with a molecularweight ofm/z 531 can be detected using SIR on a different channel (bottom).

S.A. Banack et al. / Toxicon 57 (2011) 730–738 735

specifically collision energies of 17, 21, and 38 eV, which areused during the identification of derivatized BMAA. Theresultant product ions for underivatized BMAA were m/z102, 76, 88, and m/z 70 in order of abundance. Similarproduct ions were used for identification of underivatizedBMAA by S. Christensen in Hawaii (m/z 102, 88, 76, 73;Bidigare et al., 2009) and by E. Faassen in the Netherlands(m/z 102, 88, 76; Faassen et al., 2009). On the other hand,underivatized 2,6-diaminopimelic acid has a differentprecursor ion and produced completely different productions (m/z 128, 82, 55, and 150) when compared to theproduct ions used for identification of underivatized BMAA.

We also characterized product ions of AQC-derivatizedmethionine methylsulphonium, which is distinguishedfromBMAA in the triple quadrupole UHPLC-MS/MSbymass(m/z 334), retention time (0.9min), and unique product ionsfrom collision-induced dissociation (m/z 129, 128, 198, 266,143, in order of abundance). Underivatized methioninemethylsulphonium was infused and was shown to havea parentmass ofm/z 164 and unique product ions ofm/z 56,74, 84, and 28 following collision-induced dissociation.Although not unique, m/z 102 is the most abundant ionfollowing CID for underivatized methionine methyl-sulphonium. Researchers using methods to identify under-ivatized BMAA should be cautious of this potential overlap.

3.2. Mass spectrometry using propyl chloroformatederivatization

When using propyl chloroformate pre-column derivati-zation and UHPLC-MS, BMAA clearly separated from theeight compounds tested, based on retention time andmass-to-charge ratio of the precursor ion (Fig. 2). Of the BMAAisomers providing the same ion atm/z 333 (2,4-DAB andN-2

Fig. 3. UHPLC-UV separation of non-hydrolyzed 6-aminoquinolyl-N-hydroxysuccinimidyl carbamate (AQC) derivatized b-N-methylamino-L-alanine (BMAA), 2,6-diaminopimelic acid (DPM) and 2,4-diaminobutyricacid (2,4-DAB).

(amino)ethylglycine), differentiation based on retentiontimeswas sufficient to distinguish these compounds (Fig. 2)although different product ion ratios do occur as a result ofcollision-induced dissociation (Banack et al., 2010). BMAAelutes at 7.32minwith AEG eluting 25 s earlier and 2,4-DABeluting 1.27min earlier. Furthermore, homocystine elutes atleast 5 min after BMAA and requires a modification of thegradient used here to be resolved from other analytes thatelute during the re-equilibration phase of the run.

3.3. UHPLC-UV, UHPLC-MS and HPLC-FD separation

In the UHPLC, AQC-derivatized BMAA is separated from2,4-DAB, 2,6-diaminopimelic acid, lysine, ornithine, homo-cystine, 2,4-diaminosuccinic acid, asparagine, glutamine,tryptophan, cystine, and methionine methylsulfonium(Figs. 3 and 4; Table 2). Although N-2(amino)ethylglycine isclearly distinguished fromBMAAwith our currentHPLC and

Table 2Retention times of non-hydrolyzed AQC-derivatized diamino acids andamides showing chromatographic separation in high performance liquidchromatography with fluorescence detection and ultra high pressureliquid chromatography with ultraviolet and mass spectrometric detection.Double chromatographic peaks from 2,6-diaminopimelic acid and 2,4-diaminosuccinic acid were verified by peak collection and mass deter-mination on UHPLC-MS.

Diamino acid Retention time

HPLC-FD UHPLC-UV UHPLC-MS

Asparagine 12.47 2.14 2.19Glutamine 13.97 2.59 2.63Methylsulfonium methionine 14.89 1.26 1.312,6-diaminopimelic acid 19.70, 21.90 5.02, 5.40 5.06, 5.422.4-diaminosuccinic acid 22.84, 24.43 4.15 4.18Cystine 24.27 6.79 6.79N-2(amino)ethylglycine 29.20 5.04 5.08BMAA 30.03 5.04 5.082,4-DAB 32.24 5.17 5.20Ornithine 32.80 6.03 6.07Lysine 33.94 6.80 6.85Homocystine 34.49 7.77 7.82Tryptophan 35.73 8.16 8.27

Table 3An examination of literature citations that examined non-experimentalenvironmental and clinical samples for the presence of BMAA.

Publication Number ofdetectionmethods used

Method description

Vega and Bell,1967

4 Ionophoresis, PaperChromatography,Spectrophotometry, 1H NMR

Polsky et al.1972

1 Amino Acid Analyzer

Dossaji andBell, 1973

2 Ionophoresis, Amino AcidAnalyzer

Kisby et al.,1988

1 HPLC-FD

Duncan et al.,1989

1 Gas Chromatography-MS

Duncan et al.,1990

1 Gas Chromatography-MS

Duncan, 1991 1 Gas Chromatography-MSKisby et al.,

19921 HPLC-FD

Pan et al., 1997 1 Gas Chromatography-MSKhabazian

et al., 20021 HPLC-MS/MS

Wilson et al.,2002

1 HPLC-MS/MS

Banack andCox, 2003a

3 HPLC-FD, Thin LayerChromatography, GasChromatography-MS

Banack andCox, 2003b

1 HPLC-FD

Cox et al., 2003 2 HPLC-FD, HPLC-MSMurch et al.,

2004a2 HPLC-FD, HPLC-MS

Murch et al.,2004b

2 HPLC-FD, HPLC-MS

Cox et al., 2005 2 HPLC-FD, HPLC-MSBanack et al.,

20062 HPLC-FD, HPLC-MS

Banack et al.,2007

5 HPLC-FD, UHPLC-UV, UHPLC-MS,Amino Acid Analyzer, HPLC-MS/MS

EsterhuizenandDowning,2008

1 Gas Chromatography-MS

Johnson et al.,2008

4 HPLC-FD, UHPLC-UV, UHPLC-MS,HPLC-MS/MS

Metcalf et al.,2008

2 HPLC-FD, HPLC-MS/MS

Dietrich et al.,2008

1 Gas Chromatography-MS/MS

Rosén andHellenӓs,2008

1 HPLC-MS/MS

Kubo et al.,2008

1 HPLC-MS

Eriksson et al.,2009

1 HPLC-FD

Scott et al.,2009

1 HPLC-FD

Faassen et al.,2009

1 HPLC-MS

Pablo et al.,2009

2 HPLC-FD, HPLC-MS/MS

Cox et al., 2009 4 HPLC-FD, UHPLC-MS, AAA, HPLC-MS/MS

Craighead et al.,2009

1 UHPLC-MS/MS

Cheng andBanack, 2009

5 HPLC-FD, UHPLC-UV, UHPLC-MS,Amino Acid Analyzer, HPLC-MS/MS

Table 3 (continued)

Publication Number ofdetectionmethods used

Method description

Banack et al.,2009

1 HPLC-FD

Roney et al.,2009

1 HPLC-MS

Bidigare et al.,2009

2 HPLC-FD, HPLC-MS

Caller et al.,2009

1 UHPLC-MS/MS

Moura et al.,2009

1 1H NMR

Kr�uger et al.,2010

2 HPLC-MS, HPLC-MS/MS

Spá�cil et al.,2010

1 HPLC-MS/MS

Brand et al.,2010

2 HPLC-FD, HPLC-MS/MS

Jonasson et al.,2010

1 HPLC-MS/MS

Li et al., 2010 3 HPLC-MS, HPLC-MS/MS, UHPLC-MS/MS

S.A. Banack et al. / Toxicon 57 (2011) 730–738736

triple quadrupoleUHPLC-MS/MSmethods, it co-elutedwithBMAA in our current UHPLC-UV/MS method. Multipleanalytical platforms were used to verify the presence ofBMAA in all previous publications using UHPLC-UV orUHPLC-MS (Table 3); based on these criteria, BMAA waspresent in previously reported samples (Banack et al., 2007;Johnson et al., 2008; Cox et al., 2009; Cheng and Banack,2009).

High Performance Liquid Chromatography (HPLC) withAQC pre-column derivatization clearly separated thetwelve compounds tested from BMAA (Fig. 5, Table 2). Ofparticular interest, are the BMAA isomers; BMAA eluted at30 min with AEG eluting 1 min earlier and 2,4-DAB eluting2 min later. Thus, none of these diamino acids or amidescoelute with BMAA using HPLC-FD.

Fig. 5. HPLC-FD separation of non-hydrolyzed 6-aminoquinolyl-N-hydrox-ysuccinimidyl carbamate (AQC) derivatized diamino acids: asparagine (Asn),glutamine (Gln), AQC derivative peak (AQC derv), 2,6-diaminopimelic acid(DPM), cystine, 2,4-diaminosuccinic acid, N-2(amino)ethylglycine (AEG),b-N-methylamino-L-alanine (BMAA), 2,4-diaminobutyric acid (2,4-DAB),ornithine (Orn), homocystine (hCys), and tryptophan (Trp).

S.A. Banack et al. / Toxicon 57 (2011) 730–738 737

4. Discussion

We found that standard methods of amino acid analysisclearly distinguish BMAA from the twelve amino acidstested. Historically, quantitative analysis of amino acids hasbeen performed by pre- and post-column derivatization toincrease selectivity, column retention or chemical stability.Pre-column derivatization has been carried out in order toincrease sensitivity and to allow for faster analysis (Cohen,2001; Liu, 2001). Most commonly, pre-column derivatiza-tion of amino acids are performed with chloroformate andAQC (Cohen, 2005; Hu�sek, 2005; Josefsson, 2005). Similarly,post-column derivatization with ninhydrin has beena mainstay of amino acid analysis (Liu, 2001). In this exer-cise, we used commercially available AQC and chlor-oformate derivatization procedures, which allowed us toclearly distinguish BMAA from other diamino acids andamides after HPLC analysis. Furthermore, in the triplequadrupole tandemmass spectrometer, UHPLC-MS/MS, wewere able to clearly distinguish underivatized BMAA fromunderivatized 2,6-diaminopimelic acid which havedifferent product ions resulting from collision-induceddissociation. In all cases, the diamino acids were equallydistinguished before and after acid hydrolysis. The triplequadrupole UHPLC-MS/MS is particularly well-suited todetect analytes of low molecular weight within complexmatrices (Domon and Aebersold, 2006). In addition, the useof multiple analytical platforms (e.g. HPLC-FD, UHPLC-UV,UHPLC-MS, UHPLC-MS/MS, amino acid analyzer), can resultin increased confidence in the determination, verification,and quantification of BMAA, particularly within complexphysiological matrices (Table 3; Banack et al., 2007).

Acknowledgments

We thank L. Atherton, I. Cumming, P. and H. Henry, andK. Farkas for equipment support.

Conflict of interest

None.

References

Allende, C.C., Allende, J.E., 1964. Purification and substrate specificity ofarginylribonucleic acid synthetase from rat liver. J. Biol. Chem. 239,1102–1106.

Banack, S.A., Cox, P.A., 2003a. Biomagnification of cycad neurotoxins inflying foxes: implications for ALS-PDC in Guam. Neurology 61, 387–389.

Banack, S.A., Cox, P.A., 2003b. Distribution of the neurotoxic nonproteinamino acid BMAA in Cycas micronesica. Bot. J. Linn. Soc. 143, 165–168.

Banack, S.A., Murch, S.J., Cox, P.A., 2006. Neurotoxic flying foxes as dietaryitems for the Chamorro people, Marianas Islands. J. Ethnopharmacol.106 (1), 97–104.

Banack, S.A., Johnson, H.E., Cheng, R., Cox, P.A., 2007. Production of theneurotoxin BMAA by amarine cyanobacterium.Mar. Drugs 5,180–196.

Banack, S.A., Cox, P.A., Murch, S.J., 2009. Flying fox consumption andhuman neurodegenerative disease in Guam (Chapter 12). In:Fleming, T.H., Racey, P.A. (Eds.), Island Bats: Ecology, Evolution, andConservation. University of Chicago Press.

Banack, S.A., Downing, T.G., Spá�cil, Z., Purdie, E.L., Metcalf, J.S., Downing, S.,Esterhuizen, M., Codd, G.A., Cox, P.A., 2010. Distinguishing the cyano-bacterial neurotoxin b-N-methylamino-l-alanine (BMAA) from itsstructural isomer 2,4-diaminobutyric acid (2,4-DAB). Toxicon 56, 868–

879.Bidigare, R.R., Christensen, S.J., Wilde, S.B., Banack, S.A., 2009. Cyano-

bacteria and BMAA: possible linkage with avian vacuolar myelinop-athy (AVM) in the south-eastern United States. Amyotroph. LateralScler. 10 (S2), 71–73.

Bradley, W.G., Mash, D.C., 2009. Beyond Guam: the cyanobacteria/BMAAhypothesis of the cause of ALS and other neurodegenerative diseases.Amyotroph. Lateral Scler. 10 (S2), 7–20.

Brand, L.E., Pablo, J., Compton, A., Hammerschlag, N., Mash, D.C., 2010.Cyanobacterial blooms and the occurrence of the neurotoxin, beta-N-methylamino-L-alamine (BMAA), in South Florida aquatic food webs.Harmful Algae. doi:10.1016/j.hal.2010.05.002.

Caller, T.A., Doolin, J.W., Haney, J.F., Murby, A.J., West, K.G., Farrar, H.E.,Ball, A., Harris, B.T., Stommel, E.W., 2009. A cluster of amyotrophiclateral sclerosis in New Hampshire: a possible role for toxic cyano-bacterial blooms. Amyotroph. Lateral Scler. 10 (S2), 101–108.

Cheng, R., Banack, S.A., 2009. Previous studies underestimate BMAAconcentrations in cycad flour. Amyotroph. Lateral Scler. 10 (S2), 41–43.

Cohen, S.A., 2001. Amino acid analysis using precolumnderivatizationwith6-aminoquinolyl-N-hydroxysuccinimidyl carbamate. In: Cooper, C.,Packer, N., Williams, K. (Eds.), Methods in Molecular Biology, 159.Humana Press, Totowa, USA, pp. 39–47.

Cohen, S.A., 2005. Quantitation of amino acids as 6-aminoquinolyl-N-hydroxysuccinimidyl carbamate derivatives. J. Chromatogr. Lib. 70,242–267.

Cox, P.A., Banack, S.A., Murch, S.J., 2003. Biomagnification of cyanobacterialneurotoxins and neurodegenerative disease among the Chamorropeople of Guam. Proc. Natl. Acad. Sci. U S A 100, 13380–13383.

Cox, P.A., Banack, S.A., Murch, S.J., Rasmussen, U., Tien, G., Bidigare, R.R.,Metcalf, J.S., Morrison, L.F., Codd, G.A., Bergman, B., 2005. Diverse taxaof cyanobacteria produce b-N-methylamino-l-alanine, a neurotoxicamino acid. Proc. Natl. Acad. Sci. U S A 102 (14), 5074–5078.

Cox, P.A., Richer, R.,Metcalf, J.S., Banack, S.A., Codd,G.A., Bradley,W.G., 2009.Cyanobacteria and BMAA exposure from desert dust: a possible link tosporadic ALS among Gulf War veterans. ALS 10 (Suppl. 2), 109–117.

Craighead, D., Metcalf, J.S., Banack, S.A., Amgalan, L., Reynolds, H.V.,Batmunkh, M., 2009. Presence of the neurotoxic amino acids b-N-methylamino-l-alanine (BMAA) and 2,4-diamino-butyric acid (DAB)in shallow springs from the Gobi Desert. Amyotroph. Lateral Scler.10 (S2), 96–100.

Dietrich, D.R., Fischer, A., Michel, C., Hoeger, S.J., 2008: (Chapter 39). Toxinmixture in cyanobacterial bloomsda critical comparison of realitywith current procedures employed in human health risk assessment.In: Hudnell, K. (Ed.), Advances in Experimental Medicine and Biology:Cyanobacterial Harmful Algal Blooms. State of the Science of ResearchNeeds, 619, pp. 885–912.

Domon, B., Aebersold, R., 2006. Mass spectrometry and protein analysis.Science 312, 212–217.

Dossaji, S.F., Bell, E.A., 1973. Distribution of a-amino-b-methyl-aminopropionic acid in Cycas. Phytochemistry 12, 143–144.

Duncan, M.W., Kopin, I.J., Crowley, J.S., Jones, J.S., Jones, S.M., Markey, S.P.,1989. Quantification of the putative neurotoxin 2-amino-3-(meth-ylamino)propanoic acid (BMAA) in cycadales: analysis of the seedsof some members of the family Cycadaceae. J. Anal. Toxicol. 13,169–175.

Duncan, M.W., Steele, J.C., Kopin, I.J., Markey, S.P., 1990. 2-amino-3-(methylamino)-propanoic acid (BMAA) in cycad flour: an unlikelycause of amyotrophic lateral sclerosis and parkinsonism-dementia ofGuam. Neurology 40, 772–776.

Duncan, M.W., 1991. Role of the cycad neurotoxin BMAA in the amyo-trophic lateral sclerosis–parkinsonism dementia complex of theWestern Pacific. Adv. Neurol. 56, 301–310.

Eriksson, J., Jonasson, S., Papaefthimiou, D., Rasmussen, U., Bergman, B.,2009. Improving derivatization efficiency of BMAA utilizing AccQ-Tag� in complex cyanobacterial matrix. Amino Acids 36, 43–48.

Esterhuizen, M., Downing, T.G., 2008. b-N-methylamino-l-alanine (BMAA)in novel South African cyanobacterial isolates. Ecotoxicol. Environ.Saf. 71, 309–313.

Faassen, E.J., Gillissen, F., Zweers, J., Lürling, M., 2009. Determination ofthe neurotoxins BMAA (b-N-methylamino-l-alanine) and DAB (a-,d-diaminobutyric acid) by LC-MSMS in Dutch urban waters with cya-nobacterial blooms. Amyotroph. Lateral Scler. 10 (S2), 79–84.

Hu�sek, P., 2005. Quantitation of amino acids as chloroformatesda returnto gas chromatography. J. Chromatogr. Lib. 70, 2–38.

Johnson, H., King, S.R., Banack, S.A., Webster, C., Callanaupa, W.J., Cox, P.A.,2008. Cyanobacteria (Nostoc commune) used as a dietary item in thePeruvian highlands produce the neurotoxic amino acid BMAA.J. Ethnopharmacol. 118, 159–165.

S.A. Banack et al. / Toxicon 57 (2011) 730–738738

Jonasson, S., Eriksson, J., Berntzon, L., Spá�cil, Z., Ilag, L.L., Ronnevi, L.-O.,Rasmussen, U., Bergman, B., 2010. Transfer of a cyanobacterialneurotoxin within a temperate aquatic ecosystem suggests pathwaysfor human exposure. Proc. Natl. Acad. Sci. U S A 107, 9252–9257.

Josefsson, B., 2005. HPLC of amino acids as chloroformate derivatives.J. Chromatogr. Lib. 70, 199–228.

Khabazian, I., Bains, J.S., Williams, D.E., Cheung, J., Wilson, J.M.B.,Pasqualotto, B.A., Pelech, S.L., Andersen, R.J., Wang, Y.-T., Liu, L.,Nagai, A., Kim, S.U., Craig, U.-K., Shaw, C.A., 2002. Isolation of variousforms of sterol b-d-glucoside from the seed of Cycas circinalis;neurotoxicity and implications for ALS-parkinsonism dementiacomplex. J. Neurochem. 82, 516–528.

Kisby, G.E., Roy, D.N., Spencer, P.S., 1988. Determination of b-N-methyl-amino-L alanine (BMAA) in plant (Cycas circinalis L.) and animal tissueby precolumn derivatization of with 9-fluorenylmethyl chloroformate(FMOC) and reversed-phase high-performance liquid chromatog-raphy. J. Neurosci. Methods 26, 45–54.

Kisby, G.E., Ellison, M., Spencer, P.S., 1992. Content of the neurotoxinscycasin (methylazoxymethanol b-D-glucoside) and BMAA (b-N-methylamino-l-alanine) in cycad flour prepared by Guam Chamorros.Neurology 42, 1336–1340.

Krüger, T., Mönch, B., Oppenhäuser, S., Luckas, B., 2010. LC-MS/MSdetermination of the isomeric neurotoxins BMAA (b-N-methylamino-l-alanine) and DAB (2,4-diaminobutyric acid) in cyanobacteria andseeds of Cycas revoluta and Lathyrus latifolius. Toxicon 55, 547–557.

Kubo, T., Kato, N., Hosoya, K., Kaya, K., 2008. Effective determinationmethod for a cyanobacterial neurotoxin, b-N-methylamino-l-alanine.Toxicon 51, 1264–1268.

Lee, J.W., Beebe, K., Nangle, L.A., Jang, J., Longo-Guess, C.M., Cook, S.A.,Davisson, M.T., Sundberg, J.P., Schimmel, P., Ackerman, S.L., 2006.Editing-defective tRNA synthetase causes protein misfolding andneurodegeneration. Nature 443, 50–55.

Li, A., Tian, Z., Li, J., Yu, R., Banack, S.A., Wang, Z., 2010. Detection of theneurotoxin BMAA within cyanobacteria isolated from freshwater inChina. Toxicon 55, 947–953.

Liu, H., 2001. Measurement of blood plasma amino acids in ultrafiltratesby high-pressure liquid chromatography with automatic precolumno-phthaldialdehyde derivatization. In: Cooper, C., Packer, N.,Williams, K. (Eds.), Methods in Molecular Biology, 159. Humana Press,Totowa, USA, pp. 123–140.

Liu, X., Rush, T., Zapata, J., Lobner, D., 2009. b-N-methylamino-l-alanineinduces oxidative stress and glutamate release through action onsystem Xc

�. Exp. Neurol. 217, 429–433.Lobner, D., Piana, P.M.T., Salous, A.K., Peoples, R.W., 2007. b-N-methyl-

amino-l-alanine enhances neurotoxicity through multiple mecha-nisms. Neurobiol. Dis. 25, 360–366.

Metcalf, J.S., Banack, S.A., Lindsay, J., Morrison, L.F., Cox, P.A., Codd, G.A.,2008. Co-occurrence of beta-N-methylamino-L-alanine, a neurotoxicamino acid with other cyanobacterial toxins in British waterbodies,1990–2004. Environ. Microbiol. 10 (3), 702–708.

Moura, S., de Almeida Ultramari, M., de Paula, D.M.L., Yonamine, M.,Pinto, E., 2009. 1H NMR determination of b-N-methylamino-l-alanine

(L-BMAA) in environmental and biological samples. Toxicon 53,578–583.

Murch, S.J., Cox, P.A., Banack, S.A., 2004a. A mechanism for slow release ofbiomagnified cyanobacterial neurotoxins and neurodegenerativedisease in Guam. Proc. Natl. Acad. Sci. U S A 101 (33), 12228–12231.

Murch, S.J., Cox, P.A., Banack, S.A., Steele, J.C., Sacks, O.W., 2004b. Occur-rence of beta-methylamino-l-alanine (BMAA) in ALS/PDC patientsfrom Guam. Acta Neurol. Scand. 110 (4), 267–269.

Pablo, J., Banack, S.A., Cox, P.A., Johnson, T.E., Papapetropoulos, S.,Bradley, W.G., Buck, A., Mash, D.C., 2009. Cyanobacterial neurotoxinBMAA in ALS and Alzheimer’s disease. Acta Neurol. Scand. 120, 216–225.

Pan, M., Mabry, T.J., Cao, P., Moini, M., 1997. Identification of nonproteinamino acids from cycad seeds as N-ethoxycarbonyl ethyl esterderivatives by positive chemical-ionization gas chromatography-mass spectrometry. J. Chromatogr. A 787, 288–294.

Polsky, F.I., Nunn, P.B., Bell, E.A., 1972. Distribution and toxicity ofa-amino-b-methlyaminopropionic acid. Fed. Proc. 31, 1473–1475.

Rao, S.D., Banack, S.A., Cox, P.A., Weiss, J.H., 2006. BMAA selectivelyinjures motor neurons via AMPA/kainate receptor activation. Exp.Neurol. 201, 244–252.

Rogers, K.J., Shiozawa, N., 2008. Misincorporation of amino acidanalogues into proteins by biosynthesis. Int. J. Biochem. Cell Biol.40, 1452–1466.

Roney, B.R., Renhui, L., Banack, S.A., Murch, S., Honegger, R., Cox, P.A.,2009. Consumption of fa cai Nostoc soup: a potential for BMAAexposure from Nostoc cyanobacteria in China? Amyotroph. LateralScler. 10 (S2), 44–49.

Rosén, J., Hellenäs, K.-E., 2008. Determination of the neurotoxin BMAA (b-N-methylamino-l-alanine) in cycad seed and cyanobacteria by LC-MS/MS (liquid chromatography tandem mass spectrometry). Analyst 133,1785–1789.

Rosenthal, G.A., 1977. The biological effects and mode of action ofL-canavanine, a structural analogue of L-arginine. Q. Rev. Biol. 52,155–178.

Scott, P.M., Niedzwiadek, B., Rawn, D.F.K., Lau, B.P.-Y., 2009. Liquid chro-matographic determination of the cyanobacterial toxin b-N-methyl-amino-l-alanine in algae food supplements, freshwater fish, andbottled water. J. Food Prot. 72, 1769–1773.

Spá�cil, Z., Eriksson, J., Jonasson, S., Rasmussen, U., Ilag, L.L., Bergman, B.,2010. Analytical protocol for identification of BMAA and DAB in bio-logical samples. Analyst 135, 127–132.

Vega, A., Bell, E.A., 1967. a-amino-b-methyl-aminopropionic acid,a new amino acid from seeds of Cycas circinalis. Phytochemistry 6,759–762.

Wilson, J.M.B., Khabazian, I., Wong, M.C., Seyedalikhani, A., Bains, J.S.,Pasqualotto, B.A., Williams, D.E., Andersen, R.J., Simpson, R.J., Smith, R.,Craig, U.-K., Kurland, L.T., Shaw, C.A., 2002. Behavioral and neurologicalcorrelates of ALS-Parkinsonism dementia complex in adult mice fedwashed cycad flour. Neuromolecular Med. 1, 207–221.

Work, E., Dewey, D.L., 1953. The distribution of a,e-diaminopimelic acidamong various micro-organisms. J. Gen. Microbiol. 9, 394–409.