Reference materials for domoic acid, a marine neurotoxin - Springer Link

2 downloads 0 Views 632KB Size Report
Summary. Domoic acid was recognized recently [1, 2]to be a marine neurotoxin associated with shellfish harvested or cultured for use as human food. Evidence ...
Fresenius J Anal Chem (1990) 338:520-525

Fresenius' Joumal of

@ Springer-Verlag 1990

Reference materials for domoic acid, a marine neurotoxin* W. R. Hardstaff, W. D. Jamieson, J. E. Milley, M. A. Quilliam, and P. G. Sire Marine Analytical Chemistry Standards Program, Institute for Marine Biosciences, National Research Council of Canada, 1411 Oxford Street, Halifax, Nova Scotia, B3H 3Z1 Canada Summary. Domoic acid was recognized recently [1, 2]to be a marine neurotoxin associated with shellfish harvested or cultured for use as human food. Evidence about the occurrence of domoic acid and its importance to shellfish industries is reviewed. The preparation and certification of two reference materials for the determination of domoic acid, an instrument calibration solution (DACS-I, released by the Marine Analytical Chemistry Standards Program [MACSP] in May, 1989), and a tissue reference material (MUS-1, homogenized soft tissues ofMytilus edulis, released by the MACSP in August, 1989) are described. We believe these are the first certified standards or reference materials to be available for the determination of shellfish toxins, a problem of increasing importance to aquaculturists and the seafood industry, as well as to agencies concerned with the safety of food. The tissue homogenate preparation techniques we have developed and used may be of general interest for the preparation of other tissue reference materials for the determination of other organic compounds, since the sealed, fluid homogenate samples seem acceptably stable without being continuously frozen or refrigerated.

Introduction In the late Fall of 1987, an outbreak of poisoning associated with cultured blue mussels (Mytilus edulis) from eastern Prince Edward Island, Canada, was responsible for an intense chemical "manhunt" to identify the causative agent. Following 104 h of intense effort in our laboratories [1, 2], the toxin was identified as domoic acid 1, the structure of which is shown in Scheme 1. Domoic acid had not previously been associated with neurotoxicity, although it had been known for some time [3] as a marine natural product and is a member of a family of neurotoxic amino acids which includes kainic acid 2. Concurrently, work proceeded to identify the source of the intoxication, which was shown to be the pennate diatom Nitzschiapungens f. multiseries Hasle [4]. This is the first instance of human toxicity known to have been caused by a marine diatom. The significance of the problem was supported by the results of further monitoring of the affected area during the same period of the following year, when a bloom of the same diatom again * NRCC No. 31925

Offprint requests to: M. A. Quilliam

resulted in domoic acid contamination of mussels in the area [4]. Following the incident, the health authorities in Canada established a tentative regulatory limit of 20 gg/g (drained wet tissue basis) in the edible tissues of food products sold for human consumption. This limit is below the limit of detection of domoic acid by the AOAC mouse bioassay (ca. 40 ~tg/g, drained wet tissue basis, for detection of domoic acid symptoms; 120 pg/g for LDso) used widely for the detection of paralytic shellfish poisoning (PSP) toxins [5], and mandates the use of instrumental monitoring techniques. A pair of simple, rapid high performance liquid chromatography (HPLC) techniques was developed to aid in such work [6, 7]. One system is an isocratic separation, using a mobile phase composition of 10% aqueous acetonitrile, with 0.1% trifluoroacetic acid. The other, also an isocratic separation, uses a 12 % aqueous acetonitrile mobile phase, adjusted to pH 2.5 by the addition of 2% orthophosphoric acid. In both instances, detection is effected by taking advantage of the characteristic ultraviolet absorbance maximum of domoic acid at 242 nm. Either a single wavelength or a photodiode array detector may be used. At the same time, two alternate extraction schemes were developed. The first, based on extraction with boiling water and workup using an octadecylsilica solid phase extraction cartridge, was shown to give excellent recovery of domoic acid from contaminated tissue. This method also offers the potential for automation, useful in monitoring situations. The other method is based on the same sample extraction scheme as specified for the usual AOAC method for the monitoring of PSP by a mouse bioassay [5]. It is thus costeffective for laboratories which have a large workload of samples to be monitored for PSP toxins. However, this AOAC method (extraction with boiling 0.1 mol/1 HC1) does not completely recover the domoic acid ( 7 0 - 8 0 % yield), presumably due to its partial decomposition (vide infra). Because there was no bioassay of adequate sensitivity, domoic acid has become one of the first natural biotoxins associated with seafood to be monitored primarily by instrumental techniques. In view of the pressure on regulatory agencies in some countries to discontinue use of animal bioassays, we expect that monitoring of food for biotoxins will increasingly be done by instrumental methods, and that their wide use will cause new demands for the needed instrument calibration solutions or performance standards and certified reference materials. Our need to support extensive monitoring of the occurrence of domoic acid led to the

521

CH3 CH2"-'--/ ~CO 2

/__~,IH3j~C02"

~'~C02 H

I

c°2"

CH3 ,~",.~ ~ C 0 2 H CH3.~S~ H: C02H

~'~C02 I H

H

C02H CH~ CH3"~_~"~CO2H

H

~N/ "~C02H H

I

8 C02H --z

~C02H

CH3"-~

/ex5 ~Z. . ~ C 02H

~C02H I H 6

~N/ ~'qC02H ] H 9

Scheme 1. Structure of compounds development and production by our Marine Analytical Chemistry Standards Program of the fi~'st certified standards for the determination of any shellfish toxin.

from deionized water [1]. Drying in vacuo at 55 ° C for several days gave anhydrous domoic acid. The purity of this material was checked by HPLC with UV detection [6], mass spectrometry [10, 11], 300 MHz proton N M R [1], and infrared spectroscopy [12]. A fluorescent 9-fluorenylmethoxycarbonyl (FMOC) derivative was prepared and analyzed by HPLC with fluorescence detection [13]. All analyses indicated the absence of significant ( > 0.5%) impurities, except for small amounts of what has been shown to be a diastereomer of domoic acid (8 in Scheme 1) [14]. A portion of this characterized, purified domoic acid was used to prepare 500 ml of calibration solution, the solvent for which was acetonitrile/water (1:9, v/v). This calibration solution was thoroughly mixed and dispensed into clean pre-scored glass ampoules, which were then immediately flame-sealed, then packaged as sets of four ampoules. Each glass ampoule contains 0.5 ml of the solution of domoic acid. The presence of the diastereomeric material was not considered to be a major problem because pure domoic acid in solution has been found to gradually isomerize [6], and a mixture will inevitably result when any standard solution is stored. Since diastereomers have identical UV spectra, the relative molar response factors in HPLC with UV detection are identical and relative proportions can be recalculated at any time. Though domoic acid and its diastereomer do not resolve under some HPLC conditions, this does not present a problem and even simplifies analysis. Smaller quantities of domoic acid taken through further purification were used to make three highly accurate primary calibration solutions for which the concentrations were determined by mass and UV absorbance. Accurate dilutions of these primary calibration solutions were then used as reported below to quantify the domoic acid in DACS-1 by HPLC analysis.

Preparation of the standards 1 Instrument calibration solution DACS-1

2 Tissue reference material MUS-1

To meet emergency needs for reliable determinations of domoic acid during 1988 and early 1989, an interim standard was prepared by heartcutting the domoic acid peak from preparative scale HPLC runs and determining the concentration by UV absorbance (literature value, ~max: [2.63 _+ 0.02] x 104 at pH 7 [3]). The concentration of domoic acid thus determined for that solution was later confirmed when more highly purified domoic acid was prepared by colleagues (see acknowledgements), allowing a more exact determination of the extinction coefficient (emax = [2.60 _+ 0.02] x 104 at pH 7 [8]) and thus of the concentration of the domoic acid solution. Though this interim calibration solution was useful as a stopgap measure, trifluoroacetic acid from the HPLC mobile phase was present in the solution. Experiments showed that the resulting low pH (pH 2) caused slow decomposition of the domoic acid at room temperature, and that the reaction was more rapid in frozen specimens [6, 9], denying a conventional way to extend the shelf life of the solution. As colleagues developed methods for the large scale extraction of domoic acid fi-om contaminated shellfish, relatively large amounts of suitably pure compound became available to us and this material was used for the preparation of the DACS-1 instrument calibration solution. A large quantity (ca. 100 rag) ofdomoic acid was isolated from an aqueous extract of contaminated cultured blue mussels (Mytilus edulis) and crystallized as the dihydrate

While an instrument calibration solution such as DACS-I is essential for verification of chromatographic efficiency, detector response linearity and other instrument characteristics, it is not in and of itself sufficient to assure accurate determinations (for achieving complete statistical control of analytical data). For adequate QA/QC to be attained, the complete determination procedure (extraction, workup, separation or isolation, then measurement) needs to be verified and this can only be done through the use of suitable reference materials. In particular, it is important to ensure that the matrices of the sample and the reference are as similar as possible. Thus, the development was undertaken of a suitable reference material which would be acceptably stable in storage. As had been done for the development by colleagues elsewhere of the LUTS-I undefatted lobster hepatopancreas reference material for the determination of a suite of trace metals [15], preparation of this reference material was done by contract from the National Research Council of Canada to the Canadian Institute of Fisheries Technology, Technical University of Nova Scotia, Halifax. A preliminary feasibility study on a small batch of mussel tissue naturally contaminated with domoic acid was used to determine that an homogenized slurry of pre-cooked mussel tissue could be thermally sterilized to obtain a shelf-stable product suitable for this application. Analyses over a period of one year showed essentially constant levels of domoic acid

522 242

888588'

L

7,~ 4 8 8 ' ~

200

3002

250

308

Wavelength (nm)

ID

Fig. 1 .~ 288-

1888'

°

8

2

-

.

.

,

4

,

6

-

,

8

.

,

,

.

.

18

,

Isocratic elution HPLC chromatogram of DACS-1 using UV detection at 242 nm. An expansion of the chromatogram and the UV spectrum at the apex of peak 1 are shown in the insets. Peak identities 1 = domoic acid; 8 = diastereomer of domoic acid

12

Time (rain)

and no detectable deterioration of the material. However, since the level of domoic acid in this material was very high (> 700 gg/g, for a slurry of all the soft tissues), it was necessary to dilute this material with mussel tissue not containing detectable levels of domoic acid to achieve a target concentration in the range 100-150 gg/g. This target concentration approximates the LDs0 in the mouse bioassay [7] for extracts prepared by the AOAC method [5]. A large scale preparation was then performed as follows. Clean (7.6kg) and contaminated (3.8kg) pre-cooked mussels were removed from a - 30 ° C coldstore and thawed in a chilled room at 2 ° C. After thorough mixing, they were comminuted six times using a Comitrol 3600 food cutter (Urschel Laboratories Inc., Valparaiso, IN). After the first pass, 0.02% ethoxyquin antioxidant was added to the comminuted tissue and the moisture content (ca. 74%) was adjusted to 87% by adding distilled water. The resulting slurry was processed four times in a No-Bac Unitherm IV homogenizer (Cherry-Burrell Corporation, Cedar Rapids, IA) at successively increased pump/h0mogenizing valve pressures (1800/2500, 1800/3000, 2500/4000, 3000/4000 psi), the final pass being collected in 4-liter polyethylene bottles, flushed with nitrogen, capped and stored in a 2°C chilled room until further processing. The mussel homogenate in these bottles was thoroughly mixed, a bottletop dispenser was installed and 15-gram samples were placed into 1000 15 ml polypropylene vials (J-6034-05, Cole-Palmer Instrument Company, Chicago, IL). The vials were heat sealed with polyester/aluminum foil/polypropylene laminated film closures. The sealed samples were placed in racks, each holding 264 vials, and thermally processed in 2 batches in a steam retort (WSF Industries Inc., Tonawanda, NY) at 118.5°C for 22 min (3 standard deviations above the mean time required for sterilization determined during the preliminary study). A retort pressure of 125 kPa was maintained during the cooling cycle to counteract internal pressures in the vials and maintain seal integrity. After cooling the vials, inspecting the seals and installing the cap closures, they were heat sealed in individual 100 x 127 mm trilaminate retort pouches (Reynolds Metals Company, Richmond, Va) 1, then packaged as sets of four. 1 Requests for further information about this method of preparing shelf-stable homogenates of biological tissues should be addressed to Dr. M. A. Tung, Canadian Institute of Fisheries Technology, Technical University of Nova Scotia, P.O. Box 1000, Halifax, Nova Scotia, Canada, B3J 2X4

C e r t i f i c a t i o n procedures

1 DACS-1

Analyses of DACS-1 were performed in February, 1989, by HPLC with diode array detection [6] to ensure homogeneity and for quantitative determination of domoic acid. Calibration with the three primary standard domoic acid solutions gave a 1.3% relative standard deviation (RSD) (n = 18) for the response factor. A random selection of 30 of the sealed ampoules of DACS-1 were analyzed to determine the total concentration of domoic acid and its diastereomer to be 89.3 lag/ml with an RSD of 0.33%, showing good intersample homogeneity. These same analyses yielded values of 87.3 and 1.9 ~tg/ml for domoic acid and its diastereomer, respectively. With time, the ratio [domoic acid]/[domoic acid diastereomer] is expected to change gradually. Figure 1 presents a typical chromatogram of DACS-I. The HPLC conditions used for these analyses were as follows. A Hewlett-Packard model 1090M HPLC was used, equipped with a DR5 solvent delivery system, autosampler, variable volume injector, built-in HP 1040 DAD and HP79994 data system. The column was 25 cm × 4.6 mm i.d., packed with 5 ~tm LC-PAH (Supelco, Bellefonte, PA) [Vydac 201TP (Separations Group, Hesperia, CA) is equivalent]. Other conditions were column temperature, 55°C; injection volume, 10 I11;flow rate, 1.0 ml/min; mobile phase (isocratic, acetonitrile/water (10:90) with 0.1% v/v trifluoroacetic acid); detection by absorbance at 242nm with 10nm bandwidth. Taking into consideration the precision of calibration and analysis, as well as sample heterogeneity, the determined mean concentrations and their 95% confidence intervals were: for domoic acid alone, 87.3_+0.5 ~tg/ml; for the diastereomer alone, 1.9 _+0.4 pg/1; for domoic acid and the diastereomer together, 89 _+ 1 ~tg/1. These concentrations are suitable for determination by HPLC with UV detection. 2 MUS-1

The mass of tissue homogenate in each vial is approximately 15 g. Measurements on 22 vials gave a mean value of 15.00 g with a standard deviation of 0.16 g (1.1% RSD). The procedures used for the certification analyses of domoic acid in MUS-I were developed after an extensive investigation of different extraction and cleanup procedures [6]. Spike recovery experiments determined that recovery is

523 1

I~"

i Domoic Acid