Congener-independent immunoassay for microcystins and nodularins

Article abstract of DOI:10.1021/es011182f

Cyanobacteria (blue-green algae) (e.g., Microcystis and Nodularia spp.) capable of producing toxic peptides are found in fresh and brackish water worldwide. These toxins include the microcystin (MC) heptapeptides (>60 congeners) and the nodularin pentapep

Full text of DOI:10.1021/es011182f

Environ. Sci. Technol. 2001, 35, 4849-4856
Introduction
Congener-Independent Immunoassay
for Microcystins and Nodularins
W E R N E R J . F I S C H E R , † , ‡
I A N G A R T H W A I T E , §
C H R I S T O P H E R O . M I L E S , §
K A T H R YN M . R O S S , § J A M E S B . A G G E N , |
A . R I C H A R D C H A M B E R L I N , |
N E A L E R . T O W E R S , § A N D
D A N I E L R . D I E T R I C H * , ‡
Microcystins (MCs) and nodularins (including motuporin)
constitute a family of more than 60 structural variants of
monocyclic hepta- and pentapeptide toxins (1-3) produced
by cyanobacteria of several genera, such as Microcystis,
Anabaena, Hapalosiphon, Nodularia, Nostoc, Oscillatoria, and
Planktothrix (4). Acute poisoning of humans and animals
constitutes the most obvious problem from toxic cyano-
bacterial blooms (5), in severe cases leading to death (5-7).
In a recent incident, 60 patients of a hemodialysis unit in
Caruaru, northeast Brazil, died of acute liver failure in
February 1996 following acute exposure to toxin-contami-
nated dialysis water prepared from cyanobacteria-containing
lake water (6). In an earlier case involving a massive Anabaena
and Microcystis bloom in Itaparica Dam, Brazil, the resulting
contamination of drinking water with cyanobacterial toxins
was responsible for 2000 cases of acute gastroenteritis and
88 deaths, most of which occurred in children (8). In contrast
to the previously well-described acute intoxications, chronic
exposure, although less-well defined, is of even greater
concernsespecially in view of the high incidence of primary
liver cancer in China, which has been attributed to the tumor-
promoting effects of chronic consumption of drinking water
contaminated with cyanobacterial hepatotoxins (9-11). As
a result, there has been much interest in MC from national
and international safety committees, with increasing pressure
for stricter regulatory guidelines.
Nestle´ Research Center, Nestec Ltd., Vers-chez-les-Blanc,
1000 Lausanne 26, Switzerland, Toxinology and
Food Safety Research, AgResearch Ruakura, Hamilton,
Waikato, New Zealand, Department of Chemistry,
University of California, Irvine, California 92697-2025, and
Environmental Toxicology, University of Konstanz,
D-78457 Konstanz, Germany
Cyanobacteria (blue-green algae) (e.g., Microcystis and
Nodularia spp.) capable of producing toxic peptides are
found in fresh and brackish water worldwide. These toxins
include the microcystin (MC) heptapeptides (>60
congeners) and the nodularin pentapeptides (ca. 5
congeners). Cyanobacterial cyclic peptide toxins are
harmful to man, other mammals, birds, and fish. Acute
exposure to high concentrations of these toxins causes
liver damage, while subchronic or chronic exposure may
promote liver tumor formation. The detection of cyclic peptide
cyanobacterial toxins in surface and drinking waters has
been hampered by the low limits of detection required and
that the present routine detection is restricted to a few
of the congeners. The unusual â-amino acid ADDA (4E,6E-
3-amino-9-methoxy-2,6,8-trimethyl-10-phenyldeca-4,6-di-
enoic acid) is present in most (>80%) of the known toxic
penta- and heptapeptide toxin congeners. Here, we
report the synthesis of two ADDA-haptens, the raising of
antibodies to ADDA, and the development of a competitive
indirect ELISA for the detection of microcystins and
nodularins utilizing these antibodies. The assay has a limit
of quantitation of 0.02-0.07 ng/mL (depending on which
congeners are present), lower than the WHO-proposed
guideline (1 ng/mL) for drinking water, irrespective of the
sample matrix (raw water, drinking water, or pure toxin in
PBS). This new ELISA is robust, can be performed
without sample preconcentration, detects toxins in
freshwater samples at lower concentrations than does
the protein phosphatase inhibition assay, and shows very
good cross-reactivity with all cyanobacterial cyclic
peptide toxin congeners tested to date (MC-LR, -RR, -YR,
-LW, -LF, 3-desmethyl-MC-LR, 3-desmethyl-MC-RR, and
nodularin).
To protect consumers from adverse effects of cyanobac-
terial peptide toxins, the World Health Organization (WHO)
proposed a provisional upper limit for microcystin-LR (MC-
LR) of 1 ng/ mL in drinking water (12), whereas Health Canada
calculated a tolerable daily intake (TDI) of 0.013 µg of MC-LR
(kg of body weight)-1 day-1 (defined as a 60-kg adult
consuming 1.5 L of water per day, with an MC-LR content
of 0.5 ng/ mL water) (13). However, effective consumer
protection requires the sensitive and efficient detection of
the whole spectrum of cyanobacterial cyclic peptide toxin
congeners, many of which are as toxic as MC-LR, and
regulation should not be restricted to MC-LR alone. This,
however, also requires that the present methods for cyclic
peptide toxin analysis be able to quantify the individual
congeners with similar sensitivities and at concentrations
well below the proposed limits (because the toxic effects of
the various congeners are expected to be additive). The mouse
bioassay, routinely used to date, is able to detect toxicity
irrespective of the toxin congeners present. However, it has
several disadvantages: its sensitivity is quite low, so that it
is of use only for the detection of toxin concentrations that
could lead to acute intoxications; the intraperitoneal route
of administration may not appropriately parallel natural
exposures; and many animals and large samples are neces-
sary. The other two widely employed methods, chromato-
graphic detection (e.g., HPLC, HPLC-MS) (14, 15) or protein
phosphatase inhibition assay (16, 17), offer excellent sen-
sitivity but are expensive to perform and require highly skilled
personnel or demand extraction and up-concentration of
the cyanobacterial mass, associated with problems of re-
covery and quantification, respectively. While chromato-
graphic methods are capable of detecting and identifying
single congeners, routine quantification of all known con-
geners is almost impossible because new analogues, espe-
cially of microcystins, continue to be discovered (18-21).
Furthermore, chromatographic methods are ill-suited to the
detection of toxicity arising from the presence of low levels
of several congeners because the concentrations of the
individual components may be below the limits of detection.
* Corresponding author telephone: +49-7531-883518; fax: +49-
7531-883170; e-mail: Daniel.Dietrich@uni-konstanz.de.
† Nestec Ltd.
‡ University of Konstanz.
§
AgResearch Ruakura.
| University of California.
9
10.1021/es011182f CCC: $20.00
Published on Web 11/16/2001
2001 Am erican Chem ical Society
VOL. 35, NO. 24, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 4 8 4 9

FIGURE 1. Chemical structures of microcystins-LR, -RR, -YR, -LF, and -LW, 3-desmethylmicrocystins-LR and -RR, and nodularin.
Because enzyme-linked immunosorbent assays (ELISAs)
are quick, cheap, and easy to perform, ELISAs are ideal for
the establishment of a screening programsprovided that
robust assays of sufficient sensitivity and, especially, universal
cross-reactivity to the numerous toxin congeners are avail-
able. The development of a broad-spectrum ELISA with high
sensitivity and broad specificity for cyanobacterial cyclic
peptide toxin congeners, and its validation for analysis of
water, is therefore a key goal in the establishment of an
efficient and cost-effective screening procedure.
Several groups have developed immunoassays for mi-
crocystins using monoclonal (22-25) and polyclonal (26-
28) antibodies. However, all of these antibodies were raised
against specific microcystin congeners, and the assays
therefore tend to be sensitive to a relatively narrow range of
microcystin analogues. Such assays are not ideal for screening
because of the possibility of false negatives in the presence
of toxic congeners to which the assay is insensitive.
The most commonly encountered congener, microcystin-
LR (Figure 1), has the structure cyclo-D-alanine-L-leucine-
erythro-â-m ethyl-D-isoaspartic acid-L-arginine-ADDA-D-
isoglutamic acid-N-methyldehydroalanine (2). The charac-
teristic feature of both microcystins and nodularins (Figure
1) is the presence of ADDA (4E,6E-3-amino-9-methoxy-2,6,8-
trimethyl-10-phenyldeca-4,6-dienoic acid) (Figure 2), a hy-
drophobic â-amino acid (29). Structural variants of micro-
cystin commonly contain other L-amino acids at two
nonconserved sites in the peptide ring. Other structural
variants arise from the presence or absence of methyl groups
at the â-Me-Asp and N-methyldehydroalanine (Mdha)
residues, but such changes have little effect on the toxicity
of the molecules (30). Isomerization of the ADDA moiety, to
form 6Z-ADDA microcystin analogues, however, renders the
molecule essentially nontoxic. Thus, 6Egeometry in the ADDA
moiety is considered a prerequisite for toxicity in microcystin
congeners (31). Approximately 15% of the known micro-
cystins contain ADDA moieties that are modified by the
presence of a hydroxyl or acetoxy (instead of a methoxy)
group at C-9. Although such analogues are usually of similar
toxicity to their ADDA-containing homologues, they are not
common constituents of cyanobacteria and usually co-occur
with the more common and abundant unmodified ADDA-
derived congeners.
An ELISA derived from antibodies recognizing 6E-ADDA,
the common structural feature present in the toxic congeners
of microcystins and nodularins, would reduce the total cost
of cyanobacterial cyclic peptide toxin analysis and the
workload associated with safety testing. It would also
minimize the risk of water containing novel but toxic
microcystin or nodularin analogues being declared “toxin-
free”.
Here we report the synthesis of 6E-ADDA analogues for
use as haptens, the production of antibodies to these haptens,
and the use of these antibodies in the development of an
ELISA for cyanobacterial cyclic peptide toxins. Characteriza-
tion of the ELISA was performed and the assay validated for
analysis of toxic microcystins and nodularins in water. In
addition, preliminary experiments were conducted to com-
pare this new ELISA with the protein-phosphatase inhibition
assay (PP assay) for analysis of extracts from lakes and water
treatment plants from Switzerland. The technology described
in the following paragraphs, as well as the ELISA derived
thereof, was submitted for worldwide patenting (PCT WO
01/ 18059 A2, patent pending).
9
4 8 5 0 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 24, 2001

FIGURE 2. Synthesis of the immunizing hapten N-acetyl-ADDA-D-alanine (6) and the plate-coating hapten N-hemiglutaryl-ADDA methyl
ester (7) from N-Boc-ADDA methyl ester (1). The structure of ADDA is also shown for comparison along with the atom numbering system
for ADDA.
C18 column (86-200-C5). Solution pH measurements were
Materials and Methods
made with Whatman-type CF pH 0-14 indicator paper strips.
Materials. MC-RR was from Sigma, Germany; MC-LR, MC-
Inert atmosphere operations were conducted under nitrogen
YR, and nodularin were from Calbiochem, Germany; MC-
passed through a Drierite drying tube in oven- or flame-
LW and MC-LF were from Alexis, Germany; and 3-desmethyl-
dried glassware. Anhydrous tetrahydrofuran (THF) was
MC-LR (dmMC-LR) and 3-desmethyl-MC-RR (dmMC-RR)
distilled first from calcium hydride and then from potassium;
(32) were provided by Dr. Jussi Meriluoto (A° bo Akademi
anhydrous ether was distilled from potassium. Triethylamine
University, Turku, Finland). N,N-Carbonyldiimidazole and
and methylene chloride were purified by distillation from
ovalbumin (OVA) were from Sigma, St. Louis, MO. Bovine
calcium hydride. Diisopropylethylamine (DIEA) and DMF
serum albumin (BSA) was from Life Technologies, New
were dried over 3-Å molecular sieves. LHMDS (lithium
Zealand. Anti-sheep secondary antibody was from ICN,
hexamethyldisilazide, 1.0 M in THF) was purchased from
Germany. 3,3′,5,5′-tetramethylbenzidine (TMB) was from
Aldrich Chemical Co. HATU (N-[(dimethylamino)-1H-1,2,3-
Roche, Germany. Tween 20 was purchased from Roth,
triazolo[4,5-b]pyridin-1-ylmethylene]-N-methylmethanamin-
Germany. Cationized BSA (cBSA) was prepared by coupling
iumhexafluorophosphate N-oxide) was purchased from
diethylamine to BSA by means of 1-ethyl-3-(3-dimethyl-
PerSeptive Biosystems Inc. Reagents were used as purchased
aminopropyl)carbodiimide (EDAC) (33).
from Aldrich, Sigma-Aldrich, or Acros unless stated otherwise.
Hapten Synthesis. The hapten used for the immunization,
N-Acetyl-ADDA Methyl Ester (3). TFA (2 mL) was added
N-acetyl-D-alanyl-ADDA (6), was synthesized (Figure 2) from
to 31 mg (0.70 mmol) of N-Boc-ADDA methyl ester (1) in a
the previously reported (34) N-Boc-ADDA methyl ester (1)
flask. After 1 h, the TFA was removed under vacuum, and the
by replacing the Boc group with acetyl, followed by hydrolysis
residue was concentrated three times from toluene to remove
of the methyl ester and coupling of the resultant free acid
group with alanine methyl ester. Saponification of the alanine
the TFA. The resulting oil (2) was dissolved in 2.5 mL of
freshly distilled CH2Cl2 and cooled to 0 °C. Anhydrous
methyl ester group gave 6, suitable for carboxyl activation
triethylamine (28 mg, 0.28 mmol) was added to the solution,
and coupling to the carrier protein. The hapten used for the
plate-coater, N-hemiglutaryl-ADDA methyl ester (7), was
similarly synthesized (Figure 2) from 1 by replacement of
the Boc group with hemiglutaryl.
followed by freshly distilled acetic anhydride (0.141 g, 1.39
mmol). After 1 h, saturated NH4Cl (5 mL) was added. The
mixture was stirred for 20 min at 0 °C and then extracted
with EtOAc (3 × 5 mL). The extract was washed successively
with 50% saturated NH4Cl, 50% saturated NaCO3, and brine,
then dried (MgSO4), and concentrated under vacuum to give
a white solid. This solid was purified by flash chromatography
(EtOAc-hexane, 1:1) to give N-acetyl-ADDA methyl ester (3)
(25 mg, 93%) as a white solid. TLC Rf 0.23 (EtOAc-hexane,
1H NMR spectra were obtained on an Omega 500 (500
MHz), a General Electric GN-500 (500 MHz), or a Bruker
DRX (400 MHz) spectrometer, and chemical shifts are
reported relative to internal tetramethylsilane. Data (at 500
MHz unless specified) are reported as follows: chemical shift
[integral, multiplicity (s, singlet; d, doublet; t, triplet; q, quartet;
m, multiplet), coupling constant]. Multiplicities and coupling
constants for each resonance are recorded in numerical order.
Infrared (IR) spectra were obtained with a Perkin-Elmer
model 1600 series FTIR spectrophotometer. High-resolution
mass spectra were obtained at the UCI Mass Spectrometry
Center with a Micromass Autospec spectrometer. Melting
points (mp) were obtained with a Laboratory Devices Mel-
Temp melting point apparatus and are reported uncorrected.
Thin-layer chromatography (TLC) was performed on 0.25-
mm Merck precoated silica gel plates (60 F-254), and flash
chromatography was performed using ICN 200-400 mesh
silica gel. Normal-phase preparative HPLC was conducted
with a 25 × 100 mm Waters prep Nova-Pak HR silica cartridge
(WAT038511), and reversed-phase preparative HPLC was
conducted with a 25 × 100 mm Waters prep Nova-Pak HR
1
2:3). IR (thin film): 3330, 2919, 1731, 1654, 1454 cm-1. H
NMR (CDCl3, δ): 0.99 (3H, d, J ) 6.5 Hz), 1.20 (3H, d, J ) 7
Hz), 1.57 (3H, s), 2.02 (3H, s), 2.58 (1H, ddq, J ) 6, 6.5, 10 Hz),
2.67 (1H, dd, J ) 7.5, 14 Hz), 2.78 (3H, m), 2.79 (1H, dd, J )
5, 13.5 Hz), 3.17 (1H, ddd, J ) 5, 6, 7 Hz), 3.21 (3H, s), 3.65
(3H, s), 4.71 (1H, ddd, J ) 4.5, 5, 5.5 Hz), 5.37 (1H, d, J ) 9.5
Hz), 5.42 (1H, dd, J ) 6.5, 15.5 Hz), 6.18 (1H, d, J ) 15.5 Hz),
6.40 (1H, d, J ) 9 Hz), 7.25-7.15 (5H, m). HRMS (m/ z):
388.2505 (M + H)+ calcd for C23H34NO4, 388.2488.
N-Acetyl-ADDA (4). The methyl ester from the previous
step (3) (22 mg, 0.057 mmol) in THF (2 mL) was hydrolyzed
by addition to LiOH (1 M, 0.57 mL, 0.57 mmol). After being
stirred for 22 h, the mixture clarified. Water (5 mL) was added,
and the product was washed with hexane (2 × 5 mL). The
combined hexane washes were back-extracted with water (3
× 5 mL). The combined aqueous extracts were acidified with
NaHSO4 (1 M) and extracted with CH2Cl2 (3 × 5 mL). The
C
18 cartridge (WAT038510). Reversed-phase analytical HPLC
was conducted with a 4.6 × 250 mm Rainin Microsorb-MV
9
VOL. 35, NO. 24, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 4 8 5 1

CH2Cl2 extracts were washed with brine, filtered, and
concentrated to give N-acetyl-ADDA (4) (23 mg) as an oil
that was used without purification. TLC Rf 0.34 (HOAc-
EtOAc-hexane, 1:49:50). IR (thin film): 3295 br, 2923, 1713,
1640 cm-1. 1H NMR (CDCl3, δ): 0.99 (3H, d, J ) 6.5 Hz), 1.25
(3H, d, J ) 7 Hz), 1.58 (3H, s), 2.02 (3H, s), 2.57 (1H, ddq, J
) 6.5, 6.5, 9.5 Hz), 2.65 (1H, dd, J ) 7.5, 14 Hz), 2.76 (3H, m),
2.77 (1H, dd, J ) 5, 13 Hz), 3.17 (1H, ddd, J ) 5, 6.5, 6.5 Hz),
3.21 (3H, s), 4.71 (1H, ddd, J ) 5, 6, 10 Hz), 5.37 (1H, d, J )
9.5 Hz), 5.45 (1H, dd, J ) 6.5, 15.5 Hz), 6.18 (1H, d, J ) 15.5
Hz), 6.37 (1H, d, J ) 9.5 Hz), 7.25-7.15 (5H, m). HRMS (m/ z):
374.2325 (M + H)+ calcd for C22H32NO4, 374.2331.
TABLE 1. Midpoint Titers, Determined as 1/Antiserum Dilution
a
Giving 50% Maximum Absorbance, for Each Antiserum
sheep no.
immunogen
antibody titer
813
814
815
823
824
825
826
827
828
BSA-ADDA
BSA-ADDA
BSA-ADDA
cBSA-ADDA
cBSA-ADDA
cBSA-ADDA
OVA-ADDA
OVA-ADDA
OVA-ADDA
45 000
45 000
25 000
100 000
200 000
100 000
<1 000
<1 000
<1 000
N-Acetyl-ADDA-D-Alanyl Methyl Ester (5). N-Acetyl-
ADDA (4) (9 mg, 0.024 mmol) in DMF (0.6 mL) was added
to D-alanine methyl ester hydrochloride (17 mg, 0.12 mmol)
and HATU (14 mg, 0.036 mmol). The resulting solution was
cooled to 0 °C, and collidine (41 mg, 0.34 mmol) was added.
The solution was stirred at 0 °C for 2 h, then warmed to room
temperature, and stirred overnight. Water (5 mL) was added,
and the product was extracted with EtOAc (2 × 5 mL). The
extract was washed successively with saturated NaHCO3,
water, 1 M NaHSO4, water, and brine, and then dried (MgSO4)
and concentrated under vacuum. Flash chromatography
(EtOAc-hexane, 4:1) gave N-acetyl-ADDA-D-alanyl methyl
ester (5) (8 mg, 73%) as a white solid. TLC Rf 0.17 (EtOAc-
hexane, 3:2). IR (thin film): 3284, 3067, 2923, 1743, 1650,
1542 cm-1. 1H NMR (CDCl3, δ): 0.99 (3H, d, J ) 6.5 Hz), 1.23
(3H, d, J ) 7 Hz), 1.35 (3H, d, J ) 7 Hz), 1.58 (3H, s), 2.04 (3H,
s), 2.52 (1H, dq, J ) 4, 7 Hz), 2.59 (1H, ddq, J ) 6.5, 7, 9.5 Hz),
2.68 (1H, dd J ) 7.5, 14 Hz), 2.81 (1H, dd, J ) 4.5, 14 Hz), 3.19
(1H, ddd, J ) 5, 7, 7 Hz), 3.22 (3H, s), 3.75 (3H, s), 4.55 (1H,
dq, J ) 7, 7 Hz), 4.62 (1H, m), 5.39 (1H, d, J ) 9.5 Hz), 5.45
(1H, dd, J ) 6.5, 15.5 Hz), 6.18 (1H, d, J ) 15.5 Hz), 6.23 (1H,
d, J ) 7 Hz), 7.05 (1H, d, J ) 9 Hz), 7.27-7.17 (5H, m). HRMS
(m/ z): 459.2869 (M + H)+ calcd for C26H39N2O5, 459.2859.
N-Acetyl-ADDA-D-Alanine (6). To N-acetyl-ADDA-D-
alanyl methyl ester (5) (5 mg, 0.011 mmol) in THF (2 mL) was
added 1 M LiOH (0.10 mL, 0.10 mmol). After 50 min, water
was added, and the solution was washed with diethyl ether
(2 × 5 mL). The ethereal extracts were back-extracted with
water (3 × 3 mL), and the aqueous fractions were combined
and adjusted to pH 3 by addition of saturated citric acid. The
aqueous solution was extracted with EtOAc (2 × 5 mL), and
the EtOAc extracts were washed with water (2 × 5 mL) and
brine (5 mL), dried (MgSO4), and concentrated under vacuum.
The resulting solid was purified by preparative reversed-
phase HPLC (MeOH-aqueous TFA (0.2%), 7:3; Rt of product,
15.7 min) to give N-acetyl-ADDA-D-alanine (6) (4 mg, 85%)
as a white solid. TLC Rf 0.36 (HOAc-MeOH-CH2Cl2, 1:10:
89). IR (thin film): 3288 br, 2937, 1720, 1658, 1632 cm-1. 1H
NMR (DMSO-d6, δ): 0.94 (3H, d, J ) 7.0 Hz), 0.96 (3H, d, J
) 7.0 Hz), 1.19 (3H, d, J ) 7.0 Hz), 1.52 (3H, s), 1.82 (3H, s),
2.63 (1H, dd, J ) 7.0, 14.0 Hz), 2.73 (1H, dd, J ) 5.0, 14.0 Hz),
3.16 (3H, s), 3.22 (1H, ddd, J ) 5.5, 5.5, 6.5 Hz), 4.19 (1H, dq,
J ) 7.0, 7.5 Hz), 4.40 (1H, m), 5.38 (1H, d, J ) 10.0 Hz), 5.44
(1H, dd, J ) 6.5, 16.0 Hz), 6.05 (1H, d, J ) 16.0 Hz), 7.17 (3H,
d, J ) 7.5 Hz), 7.25 (2H, t, J ) 7.5 Hz), 7.60 (1H, d, J ) 9.0 Hz),
8.01 (1H, d, J ) 7.0 Hz). FAB MS (m/ z): 445.2695 (M + H)+
calcd for C25H37N2O5, 445.2702.
a
Italicized entry corresponds to the antiserum of highest titer selected
for further study.
centrated under vacuum. Purification by preparative TLC
gave the anhydride as a white solid. This was suspended in
EtOAc and washed with saturated NaHCO3. The organic layer
was concentrated under vacuum, and the residue was purified
by preparative TLC (EtOAc-hexane-HOAc, 70:29:1) to yield
7 as a colorless oil (3 mg, 42% yield). IR (thin film): 3286,
2926, 1733, 1648, 1534, 1453, 1198 cm-1. 1H NMR (400 MHz,
CDCl3, δ): 1.01 (3H, d, J ) 6.8 Hz), 1.22 (3H, d, J ) 7.2 Hz),
1.59 (3H, s), 2.00 (2H, m), 2.35 (2H, t, J ) 7.6 Hz), 2.44 (2H,
t, J ) 7.2 Hz), 2.58 (1H, m), 2.67 (1H, dd, J ) 7.6, 14.0 Hz),
2.79 (2H, m), 3.19 (1H, m), 3.22 (3H, s), 3.68 (3H, s), 4.71 (1H,
m), 5.38-5.44 (2H, m), 6.18 (1H, d, J ) 15.6 Hz), 6.56 (1H,
d, J ) 9.2 Hz), 7.17-7.28 (5H, m). HRMS (FAB NaI/ NBA,
m/ z): 460.2708 (M + H)+ calcd for C26H38NO6, 460.2699.
Conjugation of N-Acetyl-ADDA-D-Alanine (6) to BSA,
cBSA, and OVA. BSA (10.6 mg), cBSA (10.0 mg), and OVA (8.3
mg) were each dissolved in phosphate-buffered saline, pH
7.6 (PBS) (1000 µL). N,N-Carbonyldiimidazole (19.81 mg, 0.12
mmol) was dissolved in dry DMF (500 µL), and a portion of
the solution (100 µL) was added to N-acetyl-ADDA-D-alanine
(1.0 mg, 2.2 µmol) and allowed to stand for 90 min. DMF was
added to each of the protein solutions (BSA, 260 µL; cBSA,
260 µL; OVA, 280 µL) just prior to addition of the activated
ADDA derivative. The solution of the activated ADDA
derivative was then added to the protein solutions (40 µL
each to the BSA and cBSA, 20 µL to the OVA), and the
conjugation reaction was allowed to proceed at 4 °C for 16
h. The resulting conjugates were repeatedly diluted and then
concentrated by ultrafiltration (Filtron centrifugal ultrafil-
tration tubes, 10 kDa cutoff) until the calculated dilution of
unretained low molecular compounds was >10 (6).
Conjugation of N-Hem iglutaryl-ADDA Methyl Ester (7)
to OVA. A solution of N,N′-carbonyldiimidazole (0.46 mg) in
dry DMF (100 µL) was added, with swirling, to N-hemiglutaryl-
ADDA methyl ester (7) (0.27 mg). The solution was allowed
to stand at ambient temperature for 50 min and then added
to a solution of OVA (11.14 mg) in PBS (1 mL) at 4 °C with
stirring. The solution became turbid and DMSO (200 µL) was
added to aid solubility of the active ester. After being stirred
at 4 °C overnight, the protein solution was desalted on a
polyamide size-exclusion column (Econo-Pac 10DG, Bio-
Rad) according to the manufacturers instructions. The eluted
protein was then repeatedly diluted and then concentrated
by ultrafiltration (Filtron centrifugal ultrafiltration tubes, 10
kDa cutoff) until the calculated dilution of unretained low
molecular compounds was >103 to give OVA-ADDA-HG.
Im m unizations. Groups of three sheep were immunized
with N-acetyl-ADDA-D-alanine (6) coupled to either BSA
(BSA-ADDA), cBSA (cBSA-ADDA), or OVA (OVA-ADDA),
respectively (Table 1). For primary immunizations, immu-
nogens were prepared as water-in-oil suspensions by inject-
ing the conjugate, dissolved in PBS (1 mg of protein in 1 mL),
into 2.5 mL Freund’s complete adjuvant, followed by vortex
N-Hem iglutaryl-ADDA Methyl Ester (7). N-Boc-ADDA
methyl ester (1) (7 mg, 0.016 mmol) was added to TFA (1 mL,
15 mmol) and stirred at ambient temperature for 0.5 h. Excess
TFA was removed under vacuum, and the resulting TFA salt
was treated three times with toluene (2 mL) and concentrated
under vacuum. The resulting oil was suspended in CH2Cl2
(0.5 mL) and treated with glutaric anhydride (3.5 mg, 0.03
mmol). After being stirred at ambient temperature for 1 h,
the solution was partitioned between CH2Cl2 and saturated
NH4Cl (10 mL each). The organic layer was washed with
saturated NaHSO4 (2 × 10 mL), dried (Na2SO4), and con-
9
4 8 5 2 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 24, 2001

mixing. Immunogens for secondary and subsequent im-
munizations (up to eight boosts) were administered as
emulsions in Freund’s incomplete adjuvant, prepared as
above.
were harvested by filtration, and the intracellular toxins were
extracted with by sonication in methanol, the methanol was
evaporated, and the toxins were harvested by SPE as above.
Detailed descriptions of the samples, the extraction, and the
analytical methods and detailed analytical results are being
published elsewhere (Ho¨ ger et al., submitted for publication
in Environ. Sci. Technol.).
The immunogen emulsion (0.5 mL) was administered
intramuscularly into the semitendinosis muscle, 0.25 mL per
hind leg via a 20-gauge needle. Test bleeds (10 mL Vacutainer
no. 366430, Becton Dickinson, Franklin Lakes, NJ) were taken
from the jugular vein 14 days after the third and subsequent
immunizations, and antibody titers were determined by
ELISA. Larger volumes of blood for antisera production were
obtained by venipuncture and collected into blood bags
(blood bag 4R0001, Baxter, Deerfield, IL) under negative
pressure, from sheep identified as having high titers. A series
of three injections at four-weekly intervals was followed by
a rest period, with subsequent boosts at no less than four-
weekly intervals.
All animal manipulations were performed under the
authority of the AgResearch Ruakura Animal Ethics Com-
mittee, in accordance with the New Zealand Animal Protec-
tion (Codes of Ethical Conduct) Regulations 1987, and
Animals Protection Act (1960 and subsequent amendments).
ELISA. ELISA plates (NUNC MaxiSorp 1 no. 439454,
Denmark) were coated overnight with OVA-ADDA-hemi-
glutaryl (OVA-ADDA-HG) or BSA-ADDA conjugate in 0.05 M
sodium carbonate buffer at pH 9.6 (75 µL/ well, 2.5 µg/ mL)
at 20 °C. Unbound material was removed by aspiration. After
being washed with PBS, additional binding sites were blocked
by incubation with OVA or BSA, respectively (1% w/ v, 300
µL, 1 h, 20-25 °C). Plates were washed three times with PBS
and used immediately or stored at 4 °C for up to 7 days. In
the assay, sample or standard (50 µL) were added to the wells
together with antiserum (50 µL, at appropriate dilution, e.g.,
AB824 at 1/ 200 000). After incubation at 20-25 °C for 2 h,
wells were washed twice with PBS containing 0.05% Tween
20 (PBST) and twice with PBS. Anti-sheep secondary antibody
(ICN/ Cappel rabbit-anti-sheep-HRP) (100 µL, dilution 1/ 6000)
was then added to the wells and incubated for 2 h.
Subsequently, wells were aspirated and washed twice with
PBST and twice with PBS. Substrate solution (100 µL),
prepared by addition of TMB stock (110 µL, 10 mg/ mL in
DMSO) to sodium acetate buffer (11 mL, 0.1 M, pH 5.5)
containing 0.005% H2O2, was added and incubated for 15
min. The reaction was stopped by addition of H2SO4 (50 µL,
2 M), and absorbance was determined with a microplate
spectrophotometer at 450 nm.
To detect matrix effects, ELISA standards and samples
were diluted with the following matrixes: (a) PBS, (b) river
water (Waikato River, New Zealand), and (c) tap water
(Hamilton, New Zealand). Duplicates were then analyzed
over a dilution range spanning 6 orders of magnitude.
Assay Characterization. Optimal concentrations of assay
reagents were determined empirically by checkerboard
titrations. Assay standard curves were calculated using Logit
transformation of the data. The assay working range was
defined as the linear region at 20-80% of maximum
absorbance. Cross-reactivities in the assay of the microcystin
congeners MC-LR, -RR, -YR, -LW, -LF, dmMC-LR, dmMC-
RR, and nodularin (Figure 1) were calculated from the molar
concentration of analogue giving 50% inhibition (I50) of the
binding of the serum to the protein-ADDA solid phase and
are expressed relative to the I50 for MC-LR.
Extraction of Water Sam ples and Assay for the Activity
of Ser/Thr Protein Phosphatases-1 and -2A. Organic ex-
traction of the water and bloom samples from lakes and
water treatment plants of Switzerland, preparation of [32P]-
phosphorylase a and the PP assay were performed as
previously described in the literature (6, 7). Extracellular toxins
were harvested from the filtrates by solid-phase extraction
(SPE), and the toxins were eluted with methanol. Algal cells
Results and Discussion
Most previous immunoassays for microcystins (22, 25, 27)
were based on antibodies raised against MC-LR, coupled via
its carboxyl groups to either BSA, OVA, or keyhole limpet
haemocyanin (KLH) in the presence of EDAC. MC-LR
contains two carboxyl groups and is thus capable of semi-
random “shotgun” and multivalent coupling to the carrier
proteins. Furthermore, as microcystins are cyclic heptapep-
tides and nodularins are cyclic pentapeptides, the actual
epitope(s) for antibody production provided by coupling of
MC-LR or nodularin to proteins are difficult to control or
even define. These factors and the fact that there are currently
more than 60 known microcystin and nodularin congeners
(35) largely explain why antibodies raised against microcystin-
LR-protein conjugates have major limitations with regard
to cross-reactivity with other microcystin and/ or nodularin
analogues (22, 26). This is a major constraint in their routine
use for analysis and quantitation of cyanobacterial cyclic
peptide toxins in drinking water. In contrast, the presence
of 6E-ADDA (Figure 2) in at least 80% of the known strongly
toxic microcystin and nodularin congeners provides a unique
opportunity to use 6E-ADDA as the epitope for antibody
recognition. Such an antibody ought to recognize virtually
all the toxic microcystins and nodularins with approximately
equal sensitivity, regardless of their specific amino acid
composition. Moreover, an assay based on such antibodies
should be able to recognize the linear microcystins and
nodularins that can result from microbial degradation
(hydrolysis) of these cyclic peptides by microcystinases (37).
This ability is of major importance because linear tetra-
(ADDA-Glu-Mdha-Ala-OH) and hepta-(acyclo-MC-LR) pep-
tides arising from degradation of MC-LR are also potent
inhibitors of protein phosphatases (with IC50 values of 12
and 95 nM, respectively) (37). These degradation products,
although less toxic than MC-LR (IC50 ) 0.6 nM), therefore
have the potential to induce acute or chronic effects similar
to those of microcystins and nodularins. However, these
degradation products, although of importance for human
risk assessment, would most likely not be recognized by
currently available monoclonal or polyclonal antibodies. In
contrast, detection of such degradation products is to be
expected by an antibody that specifically recognizes 6E-
ADDA. Consequently, a method was developed for synthe-
sizing ADDA derivatives (haptens) that could be coupled to
carrier proteins such as BSA, cBSA, or OVA.
A methyl ester of a tertiary butoxycarbonyl-protected
ADDA precursor, N-Boc-ADDA methyl ester (1), was syn-
thesized (Figure 2) using established methodology (34).
Deprotection, followed by acetylation, ester hydrolysis,
coupling with alanine methyl ester, and finally, ester hy-
drolysis afforded N-acetyl-ADDA-D-alanine (6). This was
reacted with N,N-carbonyldiimidazole to form the activated
alanyl ester, which was then coupled to BSA, cBSA, and OVA
to afford protein conjugates containing a structurally intact
ADDA group.
Analysis of antibody titers (Table 1) indicated that sheep
immunized with OVA-ADDA responded only weakly but that
a strong immune response was generated by both BSA-ADDA
and cBSA-ADDA, with cBSA-ADDA giving the best response.
Titers were defined as the reciprocal of serum dilution giving
50% maximum absorbance in the indirect ELISA. The serum
from sheep 824, henceforth referred to as polyclonal antibody
9
VOL. 35, NO. 24, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 4 8 5 3

FIGURE 3. cELISA standard curves for cyanobacterial cyclic peptide toxin congeners. Data are shown without standard error of mean
of the individual congener values for improved clarity of the figure.
824 (AB824), had the highest titer as well as the highest
sensitivity when tested with OVA-ADDA-HG-coated 96-well
plates both in an indirect ELISA under limiting antibody
conditions and in a preliminary competitive ELISA (data not
TABLE 2. Cross-Reactivities, Limits of Quantitation, and
Working Ranges of ELISA for Selected Cyanobacterial Cyclic
Peptide Toxin Congeners
working
range
(ng/mL)
cross-reactivity
relative to
MC-LR (%)
shown). Therefore, AB824 was chosen for further charac-
terization of sensitivity and congener cross-reactivity.
Competition ELISAs (cELISA) were set up using BSA-ADDA
and OVA-ADDA-HG as plate coaters. In view of the low titer
of OVA-immunized animals and the high background
observed in cELISA employing the BSA-ADDA coater, all
further characterizations were carried out with OVA-ADDA-
HG as plate coater. The sensitivity and congener cross-
reactivity of AB824 to MC-LR, MC-RR, MC-YR, MC-LW, MC-
LF, dmMC-LR, dmMC-RR, and nodularin (Figure 1) was
tested in the cELISA over a range of dilutions (Figure 3).
Employing the conservative cutoff values of 20% and 80%
maximum absorbance, the ELISA had limits of quantitation
(LOQs) of 0.02-0.07 ng/ mL, with working ranges of up to
5-26 ng/ mL, depending on the congener (Figure 3, Table 2).
As demonstrated in Table 2, the cross-reactivity among the
congeners tested was excellent despite the problems of
determining cross-reactivities in the absence of certified
reference standards for the congeners. Some of the deviation
from 100% cross-reactivity can be explained by the facts that
(i) all commercially available congener standards contained
impurities (guaranteed purity g95%) and (ii) others have
found on reanalyzis that the actual quantities of the congener
supplied ranged between (20% of the quantity stated by the
supplier (personal communications). As cross-reactivities
were calculated on the toxin mass specified by the supplier,
cross-reactivites between 80 and 120% are most likely not
differentiable from 100%.
b
c
I80
congenera (ng/mL)
I50
MW (nM)
MC-LR
MC-RR
MC-YR
MC-LW
MC-LF
0.05 0.05-7.50
995.2 0.61
100
50
0.06 0.06-26.26 1038
0.02 0.02-6.87 1045
0.03 0.03-8.33 1025
1.22
0.37
0.52
167
118
108
157
80
0.02 0.02-15.14 986.2 0.57
dm MC-LR 0.02 0.02-8.76
981.2 0.39
dm MC-RR 0.07 0.07-9.05 1024
0.77
0.61
nodularin
0.06 0.06-4.5
825
100
a
b
c
For structures, see Figure 1. Lim it of quantitation. 50% inhibition
of binding.
reactivity observed to some congeners. This is of course an
even greater problem when the antibody epitope is in a more
variable region of the molecule, as is the case for presently
published antibodies raised against the complete microcystin
toxin conjugated to a protein carrier (22, 23, 27, 36, 38).
The effect of sample matrix on assay performance was
tested using PBS, water, and river water spiked with MC-YR.
The concentration curves for all three matrixes are depicted
in Figure 4 (Supporting Information and Ho¨ ger et al.,
submitted for publication in Environ. Sci. Technol.) and reveal
negligible matrix effects over a concentration range spanning
4 orders of magnitude (see also Table 3). Thus, sample
preparation may be omitted when using the cELISA for the
analysis of relatively clean water samples.
However, as the difference in the free energy of binding
associated with 50% cross-reactivity is very small, only minute
conformational or steric interactions of the ADDA moiety
could account for the more significant differences in cross-
The comparison of the intra-assay variability (n ) 2-4
repeats) of the competitive indirect ELISA with the PP assay
(16, 17) using extracts from environmental and treated water
samples from Swiss lakes and water treatment plants
9
4 8 5 4 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 24, 2001

FIGURE 4. Negligible matrix effects in aqueous samples analyzed without sample pretreatment. Standard curves of microcystin-YR in
tap (s) and river water (--) coincided with standard curves in PBS (- -).
ELISA should provide powerful tools for identification of novel
toxin analogues, toxin metabolites, bioconjugates, and
residues. The successful development of this assay highlights
TABLE 3. Negligible Matrix Effects for Aqueous Samples
Tested without Sample Pretreatment
the importance of careful consideration of hapten design
and selection in both the development of ELISAs for small
molecules and the control of antibody specificity. It also
illustrates the advantages to be gained from raising antibodies
against carefully chosen fragments of the analyte and the
benefits of incorporating synthetic organic chemistry into
the ELISA development process.
matrix
IC50 (% of PBS)
PBS
tap water
river water
100
100
97.7
demonstrated generally good correlation between the assays.
The ELISA generally gave higher values than the PPA but had
5-fold better sensitivity with much better repeatability across
a wide range of toxin concentrations (Supporting Informa-
tion). The competitive indirect ELISA displays both the
desired high cross-congener reactivity and high sensitivity.
Indeed, all congeners tested to date are detected with high
certainty (coefficient of variation ,20%), well below the 1
ng/ mL limit proposed by the WHO (12). In view of the fact
that not only MC-LR but also almost all microcystin and
nodularin congeners pose a serious human health risk and
a major problem for providers of drinking water, methods
for simple and reliable detection and quantification of all
toxic cyclic peptides of cyanobacterial origin are of great
interest.
Acknowledgments
We thank Hugo Eng, UCI Department of Chemistry, for
preparation of N-hemiglutaryl-ADDA methyl ester (7); Jussi
Meriluoto, A° bo Akademi University (Turku, Finland), for
supplying dmMC-LR and dmMC-RR; Stefan Ho¨ ger and
Bernhard Ernst, University of Konstanz (Germany), and Jan
Sprosen, AgResearch (Hamilton, New Zealand), for assisting
with testing of the ELISA. A.R.C. thanks the NIH for financial
support (GM57550).
Supporting Information Available
The new cELISA allows detection and quantitation of
numerous microcystin congeners, including nodularin, in
drinking water at levels well below those proposed by the
WHO, without any sample preparation or preconcentration
steps. The integrative nature of ELISA sums the total
microcystin content of the water, presenting the analyst with
a value representing the total concentration of MC and
nodularin hepatotoxins. Although only a limited range of
microcystin congeners have been tested to date, the data
presented suggest that this cELISA is a valuable tool for
routine analysis of raw and drinking water. Indeed, the ELISA
has been used to quantify cyclic peptide toxins in river and
reservoir water, cyanobacterial cultures, and blooms and has
been employed for the detection of microcystin contamina-
tion in rivers and lakes in both New Zealand and Switzerland
(data not presented). The wide cross-reactivity of the
antibodies will also be useful in chemical and biochemical
studies of cyanobacterial toxins, where applications such as
immunoaffinity chromatography, HPLC-ELISA, and TLC-
Figure showing log-log plot for intracellular and extracellular
concentrations of microcystins measured by ELISA and PPA.
This material is available free of charge via the Internet at
http:/ / pubs.acs.org.
Literature Cited
(1) de Silva, E. D.; Williams, D. E.; Andersen, R. J.; Klix, H.; Holmes,
C. F. B.; Allen, T. M. Tetrahedron Lett. 1992, 33, 1561-1564.
(2) Rinehart, K. L.; Namikoshi, M.; Choi, B. W. J. Appl. Phycol. 1994,
6, 159-176.
(3) Sivonen, K.; Jones, G. In Toxic Cyanobacteria in Water: A Guide
to Their Public Health Consequences, Monitoring and Manage-
ment; Chorus, I., Bartram, J., Eds.; E&FN Spon: London, 1999;
pp 41-111.
(4) Carmichael, W. W. Adv. Bot. Res. 1997, 27, 211-256.
(5) Schwimmer, M.; Schwimmer, D. In Algae, Man, and the
Environment; Jackson, D. F., Ed.; Syracuse University Press:
Syracuse, NY, 1968; pp 279-358.
9
VOL. 35, NO. 24, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 4 8 5 5

(6) Pouria, S.; de Andrade, A.; Barbosa, J.; Cavalcanti, R.; Barreto,
V.; Ward, C.; Preiser, W.; Poon, G.; Neild, G.; Codd, G. Lancet
1998, 352, 21-26.
(7) Jochimsen, E. M.; Carmichael, W. W.; An, J. S.; Cardo, D. M.;
Cookson, S. T.; Holmes, C. E.; Antunes, M. B.; de Melo Filho,
D. A.; Lyra, T. M.; Barreto, V. S.; Azevedo, S. M.; Jarvis, W. R. N.
Engl. J. Med. 1998, 338, 873-878.
(8) Teixeira, M.; Costa, M.; Carvalho, V.; Pereira, M.; Hage, E. Bull.
Pan Am. Health Org. 1993, 27, 244-253.
(9) Harada, K.; Oshikata, M.; Uchida, H.; Suzuki, M.; Kondo, F.;
Sato, K.; Ueno, Y.; Yu, S. Z.; Chen, G.; Chen, G. C. Nat. Toxins
1996, 4, 277-283.
(10) Ueno, Y.; Nagata, S.; Tsutsumi, T.; Hasegawa, A.; Watanabe, M.
F.; Park, H.-D.; Chen, G.-C.; Chen, G.; Yu, S.-Z. Carcinogenesis
1996, 17, 1317-1321.
(11) Yu, S.-Z. In Primary Liver Cancer; Tang, Z. Y., Wu, M. C., Xia,
S. S., Eds.; China Academic Publishers/ Springer: New York,
1989; pp 30-37.
(12) WHO. Guidelines for Drinking-Water Quality. Addendum to
Volume 2. Health Criteria and Other Supporting Information;
World Health Organization: Geneva, Switzerland, 1998.
(13) Kuiper-Goodman, T.; Gupta, H.; Combley, H.; Thomas, B. H.
In Toxic Cyanobacteria: Current Status of Research and Man-
agement; Steffensen, D. A., Nicholson, B. C., Eds.; Australian
Centre for Water Quality Research: Salisbury, Australia, 1994;
pp 67-73.
(22) Nagata, S.; Soutome, H.; Tsutsumi, T.; Hasegawa, A.; Sekijima,
M.; Sugamata, M.; Harada, K.-i.; Suganuma, M.; Ueno, Y. Nat.
Toxins 1995, 3, 78-86.
(23) Saito, S.; Nakano, Y.; Kushida, K.; Shirai, M.; Harada, K.-i.;
Nakano, M. Microbiol. Immunol. 1994, 38, 389-392.
(24) Tsutsumi, T.; Nagata, S.; Yoshida, F.; Ueno, Y. J. Toxicol. Sci.
1998, 23 (Suppl.), 400.
(25) Ueno, Y.; Nagata, S.; Tsutsumi, T.; Hasegawa, A.; Yoshida, F.;
Suttajit, M.; Mebs, D.; Pu¨ tsch, M.; Vasconcelos, V. Nat. Toxins
1996, 4, 271-276.
(26) An, J.; Carmichael, W. W. Toxicon 1994, 32, 1495-1507.
(27) Chu, F. S.; Huang, X.; Wei, R. D.; Carmichael, W. W. Appl. Environ.
Microbiol. 1989, 55, 1928-1933.
(28) McDermott, C. M.; Feola, R.; Plude, J. Toxicon 1995, 33, 1433-
1442.
(29) Rinehart, K. L.; Harada, K.-i.; Namikoshi, M.; Chen, C.; Harvis,
C. A.; Munro, M. H. G.; Blunt, J. W.; Mulligan, P. E.; Beasley, B.
R.; Dahlem, A. M.; Carmichael, W. W. J. Am. Chem. Soc. 1988,
110, 8557-8558.
(30) Carmichael, W. W. J. Appl. Bacteriol. 1992, 72, 445-459.
(31) Harada, K.-i.; Matsuura, K.; Suzuki, M.; Watanabe, M. F.; Oishi,
S.; Dahlem, A. M.; Beasley, V. R.; Carmichael, W. W. Toxicon
1990, 28, 55-64.
(32) Meriluoto, J.; Lawton, L.; Harada, K.-i. In Bacterial Toxins:
Methods and Protocols; Holst, O., Ed.; Humana Press: Totowa,
NJ, 2000; pp 65-87.
(14) Harada, K. In Toxic Microcystis; Watanabe, M., Harada, K.,
Carmichael, W., Fujiki, H., Eds.; CRC Press: Boca Raton, FL,
1996; pp 103-148.
(33) Hermanson, G. T. Bioconjugate Techniques; Academic Press:
San Diego, 1996.
(15) Meriluoto, J. Anal. Chim. Acta 1997, 352, 277-298.
(16) MacKintosh, C. In Protein Phosphorylation: A Practical Ap-
proach; Hardie, D. G., Ed.; Oxford University Press: Oxford,
1993; pp 197-229.
(17) Mountfort, D. O.; Kennedy, G.; Garthwaite, I.; Quilliam, M.;
Truman, P.; Hanna, D. J. Toxicon 1999, 37, 909-922.
(18) Namikoshi, M.; Yuan, M.; Sivonen, K.; Carmichael, W. W.;
Rinehart, K. L.; Rouhiainen, L.; Sun, F.; Brittain, S.; Otsuki, A.
Chem. Res. Toxicol. 1998, 11, 143-149.
(34) Humphrey, J. M.; Aggen, J. B.; Chamberlin, A. R. J. Am. Chem.
Soc. 1996, 118, 11759-11770.
(35) Chorus, I.; Bartram, J. Toxic Cyanobacteria in Water: A Guide
to Their Public Health Consequences, Monitoring and Manage-
ment; E&FN Spon: London, 1999.
(36) Chu, F. S.; Huang, X.; Wei, R. D. J. Assoc. Off. Anal. Chem. 1990,
73, 451-456.
(37) Bourne, D. G.; Jones, G. J.; Blakeley, R. L.; Jones, A.; Negri, A.
P.; Riddles, P. Appl. Environ. Microbiol. 1996, 62, 4086-4094.
(38) Metcalf, J. S.; Bell, S. G.; Codd, G. A. Water Res. 2000, 34, 2761-
2769.
(19) Sivonen, K.; Namikoshi, M.; Evans, W. R.; Carmichael, W. W.;
Sun, F.; Rouhiainen, L.; Luukkainen, R.; Rinehart, K. L. Appl.
Environ. Microbiol. 1992, 58, 2495-2500.
(20) Sivonen, K.; Namikoshi, M.; Evans, W. R.; Fardig, M.; Carmichael,
W. W.; Rinehart, K. L. Chem. Res. Toxicol. 1992, 5, 464-469.
(21) Sivonen, K.; Namikoshi, M.; Evans, W. R.; Gromov, B. V.;
Carmichael, W. W.; Rinehart, K. L. Toxicon 1992, 30, 1481-
1485.
Received for review August 1, 2001. Revised manuscript
received October 17, 2001. Accepted October 17, 2001.
ES011182F
9
4 8 5 6 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 24, 2001