Let's Talk About Tha Belladonna Plant

[Delta9]

998 J. Agric. Food Chem. 1989, 37, 998-1005
Occupational Exposure to Malathion. Am. Ind. Hyg. Assoc.
Frei, R. W.; Lawrence, J. F. In Chemical Derivatization in
Analytical Chemistry, Volume I: Chromatography; Plenum
Press: New York, 1981; p 175.
Reid, S. J.; Watts, R. R. A Method for the Determination of
Dialkyl Phosphate Residues in Urine. J . Anal. Toxicol. 1981,
5 , 126-132.
Shafik, M. T.; Enos, H. F. Determination of Metabolic and Hy-
drolytic Products of Organophosphorus Pesticides Chemicals
in Human Blood and Urine. J. Agric. Food Chem. 1969,17,
Shafik, T. M.; Bradway, D. E.; Enos, H. F.; Yobs, A. R. Human
Exposure to Organophosphorus Pesticides. A Modified Pro-
cedure for the Gas-Liquid Chromatographic Analysis of the
Alkyl Phosphate Metabolites in Urine. J Agric. Food Chem.
J . 1988,49,438-444.
1186-1189.
1973, 21, 625-629.
USEPA. Pesticide Assessment Guidelines, Subdivision U: Ap-
plicator Exposure Monitoring; National Technical Information
Service: Arlington, VA, 1987.
USEPA. Environmental Protection Agency: 40 CFR Parts 156
and 170: Worker Protection Standards for Agricultural Pes-
ticides. Fed. Regist. 1988, 53, 25970-26201.
van Heemstra-Lequin, E. A. H., van Sittert, N. J., Eds. Biological
Monitoring of Workers Manufacturing, Formulating and Ap-
plying Pesticides Toxicol. Lett. 1986, 33.
Wang, R. A,, Franklin, C. A., Honeycutt, R. C., Reinert, J., Eds.
Biological Monitoring for Pesticide Exposure; ACS Sympcsium
Series 382; American Chemical Society: Washington, DC, 1989.
Wolfe, N. L.; Zepp, R. G.; Gordon, J. A.; Baughman, G. L.; Cline,
D. M. Kinetics of Chemical Degradation of Malathion in Water.
Environ. Sci. Technol. 1977, 11, 88-93.
Received for review September 19,1988. Accepted March 9,1989.
Composition of Jimson Weed (Datura stramonium ) Seeds
Mendel Friedman* and Carol E. Levin
Bulk commercial grain, such as soybeans and wheat, may be contaminated by nongrain impurities,
including jimson weed seeds, that coexist with the crop to be harvested. The present study was undertaken
to determine the content of the major alkaloids of jimson weed seeds, atropine and scopolamine, as well
as protein, carbohydrate, fat, mineral, hemmagglutinin, and tannin. Combined GC-MS analysis of a
jimson weed seed extract revealed the presence of atropine and scopolamine plus possibly three additional
tropane-like alkaloids. An improved HPLC procedure showed that the alkaloid concentration in samples
obtained from different parts of the United States varied by as much as 50%: 1.69-2.71 mg/g for atropine
and 0.36-0.69 mg/g for scopolamine. The presence of a strongly fluorescent green compound of unknown
structure is also described. Baking experiments with jimson weed seed fortified wheat flour showed
that atropine and scopolamine largely survive bread-baking conditions. Jimson weed seeds do not contain
protease or amylase inhibitors. These observations provide a rational basis for relating seed composition
to biological effects in animals and for assessing the possible significance of low levels of the seeds in
food-producing animals and in the human diet.
The plant Datura stramonium was grown in England
in the 16th century from seeds that came from Constan-
tinopole, Turkey (Claus, 1961). The English presumably
imported the plant to the American colonies, as evidenced
by the fact that when English soldiers, who were sent to
quell Baconâ??s rebellion at Jamestown in Colonial Virginia,
inadvertently ate the plant as part of a salad in 1676, some
of them became ill and died. The name jimson weed or
Jamestown weed derives from this episode of fatal poi-
soning (Claus, 1961; Duke, 1984 Feenghaty, 1982; Oâ??Grady
et al., 1983). This and related reports of poisonings by
jimson weed seeds demonstrate that the plant exerts
pharmacological and toxicological effects in animals and
humans. In fact, Klein-Schwartz and Oderda (1984)
suggest that jimson weed abuse is a potentially serious
form of substance abuse in adolescents and young adults.
The most common symptoms of jimson weed ingestion are
altered perception of the environment, visual hallucina-
tions, mydriasis (dilation of the eye pupils), and tachy-
cardia (increase in heart rate) (Oâ??Grady et al., 1983). High
Western Regional Research Center, US. Department
of Agriculture-Agricultural Research Service, 800 Bu-
chanan Street, Albany, California 94710.
levels may cause depression of the central nervous system,
with symptoms ranging from lethargy to coma (Klein-
Schwartz and Oderda, 1984). Antidotes include the use
of the anticholinesterase drug physostigmine, charcoal to
slow down absorption, magnesium citrate to speed passage
through the intestinal tract, and ipecac to induce vomiting
(Om, 1975; Oâ??Grady et al., 1983).
The literature on jimson weed covers a variety of aspects
including agronomic and botanical (Broekaert et al., 1988;
Hagood et al., 1981; Kilpatrick et al., 1984; van De Velde
et al., 1988; Weaver, 1986), chemical and pharmaceutical
(Cordell, 1981; Duez et al., 1985; List and Spencer, 1976;
List et al., 1979), and medical toxicological (Day and
Dilworth, 1984; El Dirdiri, 1981; Fangauf and Vogt, 1961;
Feenghaty, 1982; Flunker et al., 1987; Gururaja and Khare,
1987; Keeler, 1981; Levy, 1977; Mahler, 1975; Mikolich et
al., 1975; Nelson et al., 1982; Shervette et al., 1979; Testa
and Fontanelli, 1988; Urich et al., 1982; Weintraub, 1960;
Williams and Scott, 1984; Worthington et al., 1981).
The objectives of this study were to (a) develop an im-
proved HPLC procedure for the analysis of atropine and
scopolamine in jimson weed seeds, (b) to demonstrate the
presence of known and unknown alkaloids in the seeds by
GC-MS analysis, (c) to measure the nutrient and antinu-
trient composition of the seeds, and (d) to assess the
This article not subject to U.S. Copyright. Published 1989 by the American Chemical Society
Composition of Jimson Weed
variation in the range of toxicant levels as affected by
growing locale i n order to estimate worst case concentra-
tions.
MATERIALS AND METHODS
Atropine and scopolamine were obtained from Sigma Chemical
Co. (St. Louis, MO). Apoatropine came from Adams Chemical
Co. (Round Lake, IL). Seed samples were obtained from the
Federal Grain Inspection Service (Kansas City, KS) and from
Valley Seed Co. (Fresno, CA). Samples were first picked through
to clean out any debris. They were then ground on a Wiley mill
using a coarse screen (No. 10).
Proximate Composition. Analyses for nitrogen, moisture, fat,
fiber, ash, carbohydrate, and mineral content were carried out
by standard procedures (AOAC, 1975).
Amino Acid Composition. Three analyses with samples
containing about 5 mg of protein (N X 6.25) were used to establish
the amino acid composition of the jimson weed seed protein: (a)
standard hydrolysis with 6 N HC1 for 24 h in evacuated sealed
tubes (Friedman et al., 1979); (b) hydrolysis with 6 N HC1 after
performic acid oxidation to measure half-cystine and methionine
content as cysteic acid and methionine sulfone, respectively; (c)
basic hydrolysis by barium hydroxide to measure tryptophan
content (Friedman and Cuq, 1988).
Tannin Content. Tannin content was determined by the
vanillin assay (Price et al., 1978) with 100 mg of jimson weed
samples extracted with 5 mL of MeOH in capped vials with
stirring for 20 min. The extracts were then centrifuged in a
Beckman microfuge. The assay was carried out a t 30 "C with
reagents previously warmed to this temperature. The tannin
content was calculated with the aid of a catechin standard curve.
Hemagglutination Assay. The assay for the presence of
hemagglutinins (lectins) was carried out with 150-mg samples of
jimson weed or soybean seed samples and human red blood cells,
as described previously for lima bean and soybean flours (Wallace
and Friedman, 1985; Friedman and Gumbmann, 1986).
Trypsin, Chymotrypsin, and a-Amylase Inhibition Assays.
Titration by previously described techniques (Friedman and
Gumbmann, 1986; Buoncore and Silano, 1986) demonstrated the
absence of inhibitors of digestive enzymes in jimson weed seed
extracts.
GC-MS Analysis of Jimson Weed Seed Extracts. The
samples were defatted by extracting with hexane in a Soxhlet
extractor for 18 h. The seed meal was air-dried and ground to
a fine powder on a UDY cyclone mill. The defatted samples (100
mg) were extracted with methanol in a Soxhlet extractor for 18
h. The volume was reduced to a few milliliters, which was then
filtered through a 0.45-wm membrane (Schleicher and Schuell).
The extract was bought up to a volume of 10 mL with methanol
and stored under refrigeration in an air-tight flask.
Preliminary studies showed that the same extracts of jimson
weed seeds used for HPLC analysis described below were suitable
for GC-MS analysis provided any HCl present was neutralized
as follows: 0.5 mL of the methanol solution was dried under a
stream of nitrogen. The residue was dissolved in 1.0 mL of 0.05
N NaOH. This solution was then mixed with 1 mL of chloroform,
the two phases were vigorously stirred, and the chloroform phase
containing the alkaloids was separated and used for analysis. The
neutralization step was necessary because any HC1 present was
found to degrade the packing of the GC capillary column.
To minimize the appearance of large numbers of peaks asso-
ciated with fatty acids and esters in the GC-MS chromatograms,
samples were defatted twice with hexane before extraction with
methanol.
HPLC Assay for Atropine and Scopolamine in Jimson
Weed Seeds. An Alumina A Sep-Pak (Waters) was conditioned
with 5 mL of chloroform. A 2-mL portion of the methanol extract
followed by 5 mL of chloroform was then passed through the
Sep-Pak. The eluants were combined and evacuated to dryness
with an aspirator. The residue was then taken up in 1 mL of
methanol containing 0.2 mg/mL cystamine as an internal
standard. This solution was used for analysis by HPLC.
A Beckman Instruments (San Ramon, CA) 334 HPLC system
with a 427 integrator and a 165 UV-vis variable-wavelength de-
tector was used. The column was a C8 Beckman Ultrasphere (4.6
X 250 mm) with a Beckman CB precolumn.
J. Agric. Food Chem., Vol. 37, No. 4, 1989 999
The mobile phase consisted of W36 water-methanol containing
0.02 M phosphate buffer (pH 3; 0.87 mL of 87% phosphoric acid
plus 0.77 g of monosodium phosphate/L) and 0.01 M dibutylamine
as a counterion. Solvent flow rate was 0.8 mL/min. Dibutylamine
was chosen as a counterion because it has lower absorptivity a t
the wavelength used (200 nm) and produced a more stable base
line than other amine modifiers, such as ethanolamine. The
mixture of water and methanol proved to be stable on the column
compared to other solvents such as acetonitrile. The low ab-
sorbtivity in the UV and the relatie low toxicity of this mobile
phase were additional benefits.
The following compounds were evaluated as possible internal
standards for the analysis of atropine and scopolamine by HPLC:
benzylamine, cystamine, diaminopropionic acid, histamine, hy-
d r o ~ yethylamine, hydroxytryptamine, nicotinamide, nicotinic acid,
nicotinic acid methyl ester, penicillamine, theobromine, theo-
phylline, and tyramine. Cystamine was selected as the internal
standard because its elution position, peak shape, and linear
relationship of peak areas of concentration were superior to the
other compounds tested.
Relative recovery of alkaloids was determined by spiking the
ground jimson weed samples with atropine and scopolamine
standards. The samples were then extracted and analyzed by
HPLC.
Baking Experiments. Unbleached, unbrominated, malted,
and enriched white wheat flour (Mellow Judith) was obtained from
Con Agra Inc. (Oakland, CA). Fresh cake yeast (Fleischman's)
was obtained a t a local market.
The recipe for one loaf of bread consisted of 183.1 g of flour,
106.8 g of water, 3.5 g of salt, 6.1 g of yeast, and 25 g (12% of dry
weight) of jimson weed seeds in the final mix. The jimson weed
seeds were mixed whole with the flour, and the mixture was then
ground in a Wiley mill. This overcame the difficulty of finely
grinding the fatty seeds alone. The ingredients were combined,
and the resulting dough was kneaded on a Hobart Model ClOO
kneader (Troy, OH) for a total of 8 min. The dough was allowed
to rise twice in a fermentation chamber (National Co., Lincoln,
NE) at 37 "C and 90% humidity for 45 min. The dough was
placed in 14.6 X 7.6 X 7.4 cm nonstick-coated pans and baked
35 min at 215 "C. The crust was separated from the crumb with
an electric knife. Both the crust and crumb were sliced, lyo-
philized, and ground in a Wiley mill with a 1-mm screen.
Green Fluorescent Compounds. In working with extracts
of jimson weed seeds, it was noted that there was present a
substance that under UV light (365 nm) exhibits an intense green
fluorescence in neutral or acidic solutions and a bright yellow
fluorescence in basic solution. The substance is present in such
quantity that the interior portion of freshly ground seeds fluoresce
without purification. The material is easily extracted with several
portions of methanol.
The material was chromatographed on thin layer precoated
silica gel plates (E. Merck). Twenty microliters of a methanol
extract (1 g extracted with 10 mL of hot methanol) was spotted
on plates 2 cm from bottom and allowed to run until the solvent
front was 15 cm from the origin. In the solvent ethyl acetate-
methanol-water (100:13.5:10), a large spot a t R, 0.05-0..10 was
observed. In the solvent ethyl acetate-formic acid-glacial acetic
acid-water (100:11:11:25), this green spot resolved itself into two
components, a smaller spot a t Rf 0.35 and a larger one at Rf 0.43.
RESULTS AND DISCUSSION
Seed Morphology and Contaminants. Seeds of the
25 species of Datura appear morphologically indisti-
guishable (J. Effengerger, Department of Food and Ag-
riculture, S t a t e of California, Sacramento, private com-
munication, May 16,1985). D. stramonium (jimsonweed)
and Datura ferox (Chinese thornapple), for example, have
seeds that are very similar i n appearance. Generally, a
genus name rather than that of a species should be used
unless a grow-out test is carried out with the seed i n
question.
The nature of t h e contaminants found t o be present i n
pure Datura spp. collected i n the field are summarized i n
Table I. These need to be removed manually, as was done
i n our studies, before the seeds can be used for composi-
1000 J. Agric. Food Chem., Vol. 37, No. 4, 1989 Friedman and Levin
Table I. Contaminant Seeds in Datura Jimson Weed Seeds
Collected in the Field"
no. of
seeds %
Datura spp. seeds, pure 1303 78.78
contaminant
broken seeds of Glycine max (soybean),
caryopses of Triticum spp. (wheat), soil
particles from microscopic to 7 mm long 243 14.69
possibly G. max. 39 2.36
Polygonum pennsylvanicum (Bigweed 24 1.45
Triticum spp. (wheat) 5 0.30
Sida spinosa (prickly mallow) 1 0.06
Digitaria sanguinalis (crabgrass) 1 0.06
Atriplex: s p p . (saltbush) 1 0.06
1654 100.00
"Test results by count. Some counts were made with an elec-
immature shriveled seeds of Fabaceae,
Abutilon spp. (Indian mallow) 35 2.12
ladysthumb)
Amaranthus spp. (pigweed) 2 0.12
tronic seed counter and are approximate.
Table 11. Comparison of Composition (%) of Jimson Weed
and Some Commonly Contaminated Grains
unpurified
jimson weed defatted full-fat whole wheat
material seed flour soy flour soy flour pastry flour
nitrogen 3.1 8.0 5.4 2.2
H20 7.7 11.2 9.0 11.1
fat 18.1 0.9 20.7 0.6
fiber 17.8 1.5 4.5 1.7
ash 6.6 5.9 4.9 1.8
carbohydrate 31.9 37.8 31.5 81.6
starch 1.1 0.7 0.6 57.5
sugar 2.1 13.9 10.6 2.4
reducing sugar 0.3 0 0 0
glucose 0.16 0.13 0.12 0.06
" Detection limit 0.1%.
tional and toxicological studies.
Proximate Composition. Table I1 shows the nitrogen,
moisture, fat, fiber, carbohydrate, and ash content of jim-
son weed seed flour, full-fat and defatted soy flour, and
whole wheat flour for comparison. The value for nitrogen
of 3.1 %, corresponding to a protein content of about 20%
(3.1 X 6.251, compares to 2.2% and 5.4% for wheat flour
and soy flours, respectively. Thus, protein content of
jimson weed is greater than that of wheat flour but much
lower than of soy flour. The table also shows that the fat
content of 18.1% in jimson weed seeds is similar to the
20.7% in soy flour. Wheat flour has a very low fat content.
The crude fiber content of jimson weed seed flour was
17.8% compared to 4.5% for soy and 1.7% for wheat flour.
The carbohydrate content of jimson weed seeds (31.9%)
is identical with that of soy flour (31.5%) but less than half
that of whole wheat flour (81.6%). A separate analysis for
starch content revealed that both jimson weed and soy
flours contain negligible amounts of this polysaccharide
compared to a 57.5% content in whole wheat flour. These
considerations suggest that, from a nutritional standpoint,
jimson weed seeds are a good source of protein and an
excellent source of fat and fiber.
Table I11 summarizes the content of minerals in jimson
weed seeds, full-fat raw soy flour, defatted soy flour, and
whole wheat flour. The data show that (a) the toxic trace
elements cadmium, mercury, and selenium were not de-
tected in any of the flours; (b) the iron content of jimson
weed seed flour is about 40 times greater than in wheat
flour and 14 and 17 times greater than in defatted and
full-fat soy flour, respectively; (c) the chromium and
Table 111. Mineral Content (ppm) of Jimson Weed Seeds
and Other Grains
unpurified
jimson weed defatted full-fat whole wheat
mineral seed flour soy flour soy flour pastry flour
cadmium" 0 0 0 0
calium 2310 2680 2250 480
chromium 2.2 0.22 0.22 0.10
copper 16.4 16.0 14.5 4.4
iron 1305 76.9 91.2 33.9
magnesium 2890 2890 2500 1630
magnanese 134 36.4 28.0 40.7
mercuryb 0 0 0 0
potassium 6540 2260 18700 4710
seleniumb 0 0 0 0
sodium 175 221 271 54
zinc 59.5 63.1 60.6 38.8
'Detection limit 0.05 ppm. *Detection limit 0.5 ppm.
Table IV. Free Amino Acid Content of Jimson Weed Seeds
elution content
amino acid mg/100 g mg/16 g N time, min
unknown" 51.0 230.0 40.2
histidine 8.3 37.0 56.0
arginine 12.4 5.6 77.0
y-aminobutyric (GABA) 16.6 74.6 53.5
a Calculated as leucine equivalents.
Table V. Amino Acid Content (g/16 g of N) of Defatted
Jimson Weed Seed Flour, Defatted Soy Flour, and
Commercial Wheat Flour"
amino jimson soy wheat
acid weed seed flour flour FAO'
ASP 7.74 11.74 4.11
Thr 3.14 3.58 2.54 4.0
Ser 4.03 4.90 4.39
Glu 13.05 18.59 33.1
Pro 3.32 5.17 11.5
G1Y 3.87 4.04 3.48
Ala 3.51 4.05 2.80
Val 3.62 5.20 4.07 5.0
Metc 1.38 1.19 1.62 3.5'
Ile 3.22 4.66 3.53 4.0
Leu 5.31 7.74 6.73 7.0
TYr 2.55 3.42 3.20 6.08
Phe 3.47 5.05 4.78
His 1.84 2.47 2.09
LYS 3.19 5.82 1.85 5.5
'4% 6.54 7.27 3.65
Tryd 0.51 1.13 0.55
Cysb 2.00 1.14 2.20
N content, %: defatted jimson weed seed flour, 3.56; defatted
soy flour, 8.00; wheat flour, 2.18. bDetermined as cysteic acid after
performic acid oxidation. Determined as methionine sulfone after
performic acid oxidation. d T ~ o separate determinations by ion-
exchange chromatography after hydrolysis by barium hydroxide.
eProvisional amino acid scoring pattern for an ideal protein (FAO,
1973). fCys + Met. gTyr + Phe.
manganese levels in jimson weed seeds are also greater
than in the three grain flours; and (d) copper, zinc, mag-
nesium, and calcium levels is the three flours do not differ
significantly.
Free and Protein Amino Acid Contents. Table IV
lists the free amino acid content of jimson weed seeds. The
seeds contained only histidine and arginine in the free form
along with two unknown amino acids. One of these eluted
in the same positions as y-aminobutyric acid. The second
elutes in the vicinity of cystine. It does not appear to be
y-glutamyl-L-aspartic acid that Ungerer et al. (1988) found
in jimson weed seeds. This dipeptide elutes in the same
position as methionine on our column. Elucidation of the
Composition of Jimson Weed J. Agric. Food Chem., Vol. 37, No. 4, 1989 1001
Table VI. Lectin Content of Jimson Weed Seeds Measured
by Agglutination
jimson activityb activity 2 years
weed seedsn (n = 2) later (n = 3)
0.63 0.63 f 0
0.80
2.50
0.63 1.7 & 0.3
1.25
0.08 5.6 f 1.0
0.10
0.10
0.04
See Table IX for origin of seeds. Minimum amount of sample
required to agglutinate human red blood cells (pg/50 pL). The
lower the number, the greater the hemagglutinating activity.
structure of the second ninhydrin-positive compound
awaits further studies.
Table V lists the amino acid composition of acid hy-
drolysates of defatted jimson weed seed flour, defatted soy
flour, and commercial wheat flour. The values of the
amino acid scoring pattern of the essential amino acids for
an ideal protein, as defined by the Food and Agricultural
Organization of the United Nations (FAO, 1973), are also
shown for comparison. The data show that the amino acid
pattern of jimson weed seeds, especially the essential amino
acids, falls between those of a legume such as soy and a
cereal such as wheat. The results show that jimson weed
seeds meet the provisional requirements for the sulfur
amino acids (Cys + Met) and for Tyr + Phe, but values
for Thr, Val, Ile, and Lys are below those of the FA0
pattern. In spite of these deficiencies, the amino acids of
jimson weed could contribute significantly to protein nu-
trition, if it were possible to remove the toxic alkaloids
either by chemical inactivation or through plant genetics.
Thus, it may be possible in the future to develop jimson
weed cultivars in which the genes controlling the biosyn-
thesis of the alkaloids are suppressed or eliminated. The
new varieties might then displace the ones currently
growing in the field. A possible approach would be to use
molecular biological techniques to insert DNA constructs
capable of generating antisense RNA specific for genes
involved in alkalid biosynthesis, resulting in blocked ex-
pression in the jimson weed plant (Walder, 1988). Instead
of presenting a problem to the consumer, the new cultivars
would contribute to the nutritional value of the grain. At
this stage, more information is needed about specific en-
zymes involved in the biosynthesis of the alkaloids (Cor-
dell, 1981) before it would be possible to control, via an-
tisense RNA or otherwise, the genes coding for these
compounds.
Hemagglutinin (Lectin) Content. Lectins, present
in many plant seeds, are carbohydrate-binding proteins
with differing sugar specificities. Normally, lectins are
thermally unstable and are partly or f d y denatured during
cooking of foods. Lectins from different sources may,
however, differ in heat stability. Inadequately cooked
legumes containing lectins may cause gastrointestinal
disturbances and adverse nutritional and toxic effects in
humans (Reaidi et al., 1981; Liener, 1988).
D. stramonium seeds are reported to contain a lectin
visualized by an immunocytochemical technique and
quantified by agglutination assays (Broekaert et al., 1988).
These authors also demonstrated a possible physiological
function for the jimson weed seed lectin in the plant, in-
volving mediation of cell-cell interactions.
Our agglutination assay of the jimson weed seed samples
listed in Table VI revealed that the minimum amount of
jimson weed seed flour required to cause agglutination of
Table VII. Tannin Content of Jimson Weed Seeds
mg tannin/ mg tannin/
samplea g seedb samplea g seedb
2.9 & 0.05 1 1.7 & 0' 4
2 1.4 f 0 5 2.3 f 0
3 1.9 f 0 7 4.5 f 0.02
bExpressed as catechin.
Average f standard deviation from two separate determinations.
human red blood cells ranged from 0.04 to 2.5 pg/50 pL
(the smaller the number, the more potent the hemagglu-
tinating activity).
The reason for the wide variation in lectin content is
unknown. An assay of the same three samples after a
2-year interval indicated an unchanged value for one and
a lower content for two (Table V). The lectins of jimson
weed seeds might be slowly inactivated during storage and
exposure to heat and sunlight under field conditions, ac-
counting for the variable results.
Tannin Content. Table VI1 lists the tannin contents
of seven jimson weed samples obtained from different
locations. The values ranged from 1.6 to 5.6 mg/g of seed.
The reasons for the more than &fold variability in tannin
content is not immediately apparent. Soil and climatic
conditions may affect the biosynthesis of tannins (Desh-
pande et al., 1984).
Tannins adversely affect the nutritional quality and
safety of foods (Griffiths, 1986; Deshpande et al., 1984),
presumably through chelation of essential trace elements
such as iron and through inhibition of proteolytic digestive
enzymes such as trypsin and chymotrypsin.
Atropine and Scopolamine Content. In their review
of toxic weed seed contamination in soybeans during
harvest, transport, and storage, List et al. (1979) state that
jimson weed seeds are probably the more prevalent and
most toxic compared to others such as castor (Ricinus
communis), cocklebur ( X a n t h i u m strumarium), corn
cockle (Agrostemma githago), cow cockle (Saponaria
uaccaria), crotolaria (Crotolaria spp.), morning glory
(Ipomoea spp.), nightshade (Atropa belladonna), and
pokewood (Phytolacca americana). These authors also
suggest that, aside from toxicity aspects, weed seed con-
tamination may also adversely affect the appearance, or-
ganoleptic, functional, and nutritional properties of grain.
The major toxic principles present in jimson weed seeds
are the alkaloids atropine and scoplamine. These sola-
naceous alkaloids are present in a number of other plants,
including Atropa belladonna (nightshade) and Hyo-
scyamus niger. The alkaloids (see structures) have a bi-
cyclic structure in which a five-membered pyrrolidine ring
is fused to a six-membered pyridine ring. Atropine
probably exists in nature as the optically active hyoscya-
mine isomer, which racemizes to the DL form (atropine)
during isolation. The major difference between scopola-
mine, also known as hyoscine, and atropine is the prescence
of an epoxide group in cis position to the N-bridge (Cordell,
1981):
'See Table IX for origin of seeds.
CH *OH CH 2 OH
ATROPINE (d 1 ) SCOPOLAMINE (d 1 )
HYOSCYAMINE ( 1 ) HYOSCINE ( 1 )
J. Agric. Food Chem., Vol. 37, No. 4, 1989 Friedman and Levin
1
1002
IO0
no
1 5 60
VI
40 z ,
MASS SPECTRUM of
PURE ATROPINE
M+ j ...
LI I! Pb 289 I
I I
? I !
0 -, , , , , I , ,. ~ , I ,
20 60 100 100 180 220 260 300
rn /z
Figure 1. Mass spectrum of atropine.
__ __ 94 100
I38 no
42 I
I MASS SPECTRUM of
PURE SCOPOLAMINE
- s v 60
108 z I ,
VI
? I
z 1 5 1 ' 136
P 1 &?- I A ! , 303
57 7 7
20
29
27
- - 2 $ y , ~ 1 , , , 0 1 - _ - -
20 60 100 140 180 220 260 300
m /z
Figure 2. Mass spectrum of scopolamine.
Figures 1 and 2 show the mass spectral fragmentation
patterns for authentic atropine and scopolamine obtained
from commercial sources. The figures show the parent
molecular ion peaks for both compounds: M+ = 289 for
atropine and M+ = 303 for scopolamine. The presence of
a parent peak in a mass spectrum generally defines the
molecular weight of a compound (Friedman, 1977). The
GC-MS chromatogram of a methanol extract of jimson
weed seeds is shown in Figure 3. Alkaloids are present
in the extract along with several other peaks associated
with fatty acid esters. Note the clear separation of the
atropine and scopolamine peaks a t scans 374 and 422,
respectively. These results indicate that GC-MS is useful
for demonstrating the presence of atropine and scopola-
mine in plant extracts. However, this technique was found
to be useful only for qualitative measurements. The me-
thod needs further study to assess its potential for a
quantitative assessment of the alkaloid content.
In addition to atropine and scopolamine, Figure 3 sug-
gests the presence of small amounts of three other akaloids
designated as X, Y, and Z. X appears to be atropine-like
in nature. Its molecular weight of 271 corresponds to the
molecular weight of atropine minus 18 (H,O), possible
apotropine. The mass spectrum of authentic apoatropine
was identical with that of X. Y, also atropine-like, has a
molecular weight of 303, which corresponds to the mo-
lecular weight of atropine plus 14 (CH, group). Z may be
scopolamine-like. Its molecular weight corresponds to that
of scopolamine plus 14 (CH,.group).
Extensive studies were carried out to develop a reliable
HPLC procedure for measuring atropine and scopolamine
in jimson weed seed extracts. Beneficial features of this
I
I GC-MS of EXTRACT of
JIMSON WEED SEEDS
0 200 400 600 BOO 1000
4: 1 9:16 1432 19:49 25:6
SCANS and ELUTION TIME in MINUTES
Figure 3. GC-MS analysis of a jimson weed seed extract showing
linear plots of elution time on the gas chromatograph and scan
numbers (scan rate 0.63 scans/s).
0 I I 1 I I
0 0.05 0.1 0.2 0.3 0.4 0 5
ALKALOID CONCENTRATION (rng/ml)
Figure 4. Plots of atropine and scopolamine concentrations
against the ratio of areas of the HPLC peaks for the two alkaloids
to the area for the internal standard cystamine.
method include (a) high sensitivity due to the use of a low
detection wavelength (200 nm), (b) use of a solvent system
with a stable base line even at this low wavelength, (c) use
of a new internal standard that is well resolved from the
complex matrix, and (d) rapid sample preparation using
solid-phase extraction (Sep-Pak). Since tropane alkaloids
are known to hydrolyze in aqueous solution, solid-phase
extraction has an additional advantage over the traditional
liquid-liquid extraction that utilizes acid-base solubility
(List and Spencer, 1976).
Figure 4 shows that the HPLC column responds linearly
to both atropine and scopolamine in the concentration
range shown. As little as 20 ng of alkaloids was detected
on the column. However, the response was not always
linear with concentration below 200 ng of total alkaloids.
Figure 5-7 demonstrate the clear separation of atropine
and scopolamine from each other and from the internal
standard cystamine. Spiking experiments, in which known
amounts of a mixture of atropine and scopolamine were
added to jimson weed flour and then reextracted, revealed
that the recovery of atropine ranged from 92 to 97% of
the amount added and of scopolamine from 87 to 92%
(Table VIII).
Table IX shows that the atropine content of seven seeds
obtained from different parts of the United States ranged
from 1.69 to 2.71 mg/g seed, with an average f SD = 2.27
Composition of Jimson Weed J. Agric. FoodChem., Vol. 37, No. 4, 1989 1003
I -
h E c
0
0
cy v
E
s
9
2
z
pc
Y
Y I e
i
n K
- E
E
0
0
cy
Y
B
s
s1
2
z
pc
0 5 10
ELUTION TIME (minutes)
Figure 5. Typical HPLC chromatogram for atropine and sco-
polamine applied to the column as a mixture. Conditions: 0.05
mg of atropine + 0.05 ma; of scopolamine + 0.2 mg of cyst-
amine/mL. A 20-pL sample of this solution was injected onto
the column. Flow rate: 0.8 mL/min. Solvent: 0.02 M phosphate
buffer; 0.01 M dibutylamine; 36% methanol; pH 3.0.
h C I
Y w l
I
c -
h ' P f r
0 5 10
ELUTION TIME (minutes)
Figure 6. Typical HPLC chromatogram of a jimson weed seed
extract. Conditions as in Figure 5.
f 0.36 mg/g. The corresponding range for scopolamine
was from 0.36 to 0.69 mg/g, with an average f SD = 0.53
f 0.13 mg/g. The atropine and scopolamine values varied
b y as much as 50'70, depending on origin. The cause of
t h i s variation is presently unknown but needs to be es-
tablished. These results are similar to those observed b y
List and Spencer (19761, List et al. (1979), and Mirzamotov
and Luftulin (1986) (who also report no significant dif-
ferences in the biosynthesis rate of the two alkaloids in D.
stramonium at different seasons of growth).
0 5 10
ELUTION TIME (minutes)
Figure 7. HPLC chromatogram of a methanol extract of a
defatted diet containing 4% jimson weed seeds. Conditions as
in Figure 4. This diet was used in animal feeding studies to assess
the toxicity of jimson weed seeds.
Table VIII. Recovery of Atropine and Scopolamine Added
to Jimson Weed Seed Floura
atropine scopolamine
material mg/g 70 rec mg/g % rec
jimson weed seed flour, 2.69 f 0.16 0.75 f 0.04
100 mg of jimson weed 3.09 f 0.13 97 0.90 f 0.04 88
control
seed flour + 0.51 mg
of atropine, 0.27 mg
of scopolamine
seed flour + 1.01 mg
of atropine, 0.54 mg
of scopolamine
seed flour + 2.02 mg
of atropine, 1.08 mg
of scopolamine
100 mg of jimson weed 3.55 f 0.21 96 1.19 f 0.06 92
100 mg of jimson weed 4.31 f 0.17 92 1.59 f 0.05 87
Average f standard deviation from two separate determina-
tions.
Table IX. Atropine and Scopolamine Content of Jimson Weed
Seeds Obtained from Different Locations
ratio of
mg/g jimson weed seeds to sample
no. location
1 Indianapolis, IN
2 Peoria, IL
3 Belle Chase, LA
4 Indianapolis, IN
5 Belle Chase, LA
6 Cedar Rapids, IA
7 Fresno, CA
atropine scopolamine scopolamine
1.69 f 0.09 0.36 f 0.03 4.7
2.07 f 0 0.59 f 0.02 3.5
2.09 f 0.06 0.51 f 0 4.1
2.26 f 0.08 0.51 f 0 4.4
2.41 f 0.05 0.69 f 0.02 3.5
2.68 f 0.04 0.39 f 0.05 6.9
2.71 f 0.02 0.66 f 0.05 4.1
av f SD 2.27 & 0.36 0.53 i 0.13 4.5 & 1.16
(I Average from two separate determinations f standard deviation.
Since jimson weed seeds contaminate soybeans, corn,
and wheat, i t is of paramount interest to find out whether
the tropane alkaloids survive the processing conditions to
which these grains may be subjected before consumption.
To obtain information on this question, we added jimson
weed seed flour to wheat flour, which was then baked into
bread. The bread was then separated into crust and crumb
1004 J. Agric. Food Chem., Vol. 37, No. 4, 1989 Friedman and Levin
Release upon Imbibition. Plant Physiol. 1988, 86, 569-574.
Buoncore, V.; Silano, V. Biochemical, Nutritional, and Toxico-
logical Aspects of a-Amylase Inhibitors from Plant Foods. Adv.
Exp. Med. Biol. 1986, 199, 483-507.
Claus, E. P. Pharmacognosy; Lea and Febiger: Philadelphia, 1961;
p 312.
Cordell, C. A. Introduction to Alkaloids-A Biogenetic Approach;
Wiley-Interscience: New York, 1981; p 94.
Crowley, J. F.; Goldstein, I. J. Datura Stramonium Lectin.
Methods Enzymol. 1982,83, 368-373.
Day, E. J.; Dilworth, B. C. Toxicity of Jimson Weed Seed and
Coca Shell Meal to Broilers. Poultry Sci. 1984, 63, 466-468.
Deshpande, S. S.; Sathe, S. K.; Salunkhe, D. K. Chemistry and
Safety of Plant Polyphenols. In Nutritional and Toxicological
Aspects o f Food Safety; Friedman, M., Ed., Plenum: New
York, 1984; pp 457-495.
Duez, P.; Chamart, S.; Hanocq, J.; Molle, L.; Vanhaelen, M.;
Vanhaelen-Fastre, R. Comparison between Thin-Layer Chro-
matography-Densitometry and HPLC for the Determination
of Hyoscyamine and Hyoscyne in Leaves, Fruit and Seeds of
Datura (Datura spp.). J . Chromatogr. 1985, 329, 415-421.
Dugan, G. M.; Gumbmann, M. R.; Friedman, M. Toxicology of
Jimsonweed Seeds. Submitted for publication, 1989.
Duke, J. A. CRC Handbook of Medicinal Herbs; CRC Press: Boca
Raton, FL, 1984; pp 161-162.
El Dirdiri, N. I.; Wafsi, I. A.; Adam, S. E. I.; Edds, G. T. Toxicity
of Datura Stramonium to Sheep and Goats. Vet. Hum.
Toxicol. 1981, 23, 244-246.
Fangauf, R.; Vogt, H. Toxicity Trials in Laying Hens and Chicks
with Datura Stramonium Seeds, a Common Contaminant of
Soya Bean Consignments. Arch. Geflugelk. 1961,25,167-171.
FAO. Energy and Protein Requirements. FA0 Nutritional
Meetings Report Series; No. 52; Food and Agricultural Or-
ganization of the United Nations: Rome, 1973.
Feenghaty, D. A. Atropine Poisoning: Jimsonweed. J. Emergency
Med. 1982,8, 139-141.
Flunker, L. K.; Damron, B. L.; Sundlor, S. F. Jimsonweed Seed
Contamination of Broiler Chick and White Leghorn Hen Diets.
Nutr. Rep. Int. 1987, 36, 551-556.
Friedman, M. Mass Spectra of Cysteine Derivatives. Adu. Exp.
Med. Biol. 1977,86A, 713-726.
Friedman, M.; Gumbmann, M. R. Nutritional Improvement of
Legume Proteins through Disulfide Interchange. Adv. Exp.
Med. Biol. 1986, 199, 357-389.
Friedman, M.; Cuq, J. L. Chemistry, Analysis, Nutrition, and
Toxicology of Tryptophan in Food. A Review. J. Agric. Food
Chem. 1988,36, 709-719.
Friedman, M.; Noma, A. T.; Wagner, J. R. Ion-exchange Chro-
matography of Sulfur Amino Acids on a Single-column Amino
Acid Analyzer. Anal. Biochem. 1979, 98, 293-305.
Fuahiki, T.; Dukuoka, S. S.; Kajiura, H.; Iwai, K. Atropine-non-
sensitive Feedback Regulatory Mechanism of Rat Pancreatic
Secretion in Response to Food Protein. J . Nutr. 1987, 11 7,
Griffiths, D. W. Inhibition of Digestive Enzymes by Polyphenolic
Compounds. In Nutritional and Toxicological Significance
of Enzyme Inhibitors in Foods; Friedman, M., Ed.; Plenum:
New York, 1986; pp 509-516.
Gururaja, K.; Khare, C. B. Datura Poisoning: A Case Report.
Med. J. Malays. 1987, 42, 68-69.
Hagood, E. S., Jr.; Bauman, T. T.; Williams, J. L., Jr.; Schreiber,
M. M. Growth Analysis of Soybean (Glycine max.) in Com-
petition with Jimsonweed (Datura stramonium). Weed Sci.
1981,29, 500-504.
Keeler, R. F. Absence of Arthrogryposis in Newborn Hampshire
Pigs from Sows Ingesting Toxic Levels of Jimsonweed During
Gestation. Vet. Hum. Toxicol. 1981, 23, 413-415.
Kilpatrick, B. L.; Wax, L. M.; Stroller, E. W. Competition of
Jimsonweed with Soybean. Agron J. 1984, 75, 833-836.
Klein-Schwartz, W.; Oderda, G. M. Jimsonweed Intoxication in
Adolescents and Young Adults. Am. J. Dis. Child. 1984,138,
737-739.
Levy, R. Jimson Seed Poison: A New Hallucinogen on the Ho-
rizon. Ann. Emerg. Med. 1977, 6, 58-61.
Liener, I. E. The Nutritional Significance of Lectins. JAOCS,
J . Am. Oil Chem. SOC. 1988,65, 518.
948-954.
Table X. Effect of Baking on Atropine and Scopolamine
Content of Jimson Weed Seed Fortified Bread
atropine scopolamine
mg/g mg/g
material mixed flour % rec mixed flour % rec
wheat flour + 12% 0.408 100 0.115 100
jimson weed seed
flour, unbaked
bread crumb 0.304 75 0.100 87.
bread crust 0.335 82 0.083 72
fractions. These were freeze-dried and milled into flours.
Table X shows the following recovery of atropine and
scopolamine from these flours compared to an unbaked
control of wheat plus jimson weed flours: for atropine,
bread crumb 75%, bread crust 82%; for scopolamine,
bread crumb 87%, bread crust 72%. The findings show
that both alkaloids in jimson weed seed flour largely
survived the high temperature of bread-baking, in both the
crumb and crust.
In a related study, List and Spencer (1976) describe the
fate of the two alkaloids during processing of jimson weed
contaminated soybeans. They report that extraction of
a 5050 mixture of soybeans and jimson weed seeds with
petroleum ether produced a meal and crude oil fractions.
Analyses of these fractions by a gas chromatographic
procedure showed that virtually all of the atropine and
scopolamine remained in the meal. They also found that
alkali refining effectively removed atropine added to soy-
bean oil.
Subchronic 90-day toxicity studies with laboratory rats
were conducted to establish safe levels of jimson weed
seeds (Dugan et al., 1989). Dose-related adverse effects
were observed at 0.5% or more in the diet. An unanswered
question is whether toxicological manifestations of jimson
weed seeds can be predicted on the basis of their atropine
and scopolamine contents or whether additional constit-
uents, such as lectins, unknown alkaloids, and the de-
scribed fluorescent compound(s), contribute to biological
activity. If the former is true, then a rapid assay for tro-
pane alkaloid content by the described HPLC method may
be useful for predicting the safety of jimson weed seeds.
Moreover, since dietary constituents and the process of
digestion and metabolism can be expected to modify the
adverse manifestations of atropine and scopolamine
(Fuahiki et al., 1987), the described compositional infor-
mation should contribute to a better understanding of
toxicological manifestations of jimson weed seeds when fed
as a part of a complete diet. These studies are expected
to help provide a basis for establishing maximum tolerance
levels in food and feed and in setting official grain stand-
ards.
ACKNOWLEDGMENT
We thank J. C. Halvorson and colleagues of the Federal
Grain Inspection Service (FGIS) for collecting the seeds
and for helpful advice, R. England and W. F. Haddon for
the mass spectra, F. F. Bautistia and G. McDonald for
excellent technical assistance, and M. R. Gumbmann for
valuable contributions.
Registry No. GABA, 56-12-2; His, 71-00-1; Arg, 74-79-3; ni-
trogen, 7727-37-9; starch, 9005-25-8; glucose, 50-99-7; cadmium,
7440-43-9; calcium, 7440-70-2; chromium, 7440-47-3; copper,
7440-50-8; iron, 7439-89-6; magnesium, 7439-95-4; manganese,
7439-96-5; mercury, 7439-97-6; potassium, 7440-09-7; selenium,
7782-49-2; sodium, 7440-23-5; zinc, 7440-66-6; atropine, 51-55-8
scopolamine, 51-34-3.
LITERATURE CITED
Broekaert, W. F.; Lambrechts, D.; Verbelen, J. P.; Peumans, W.
J. Datura Stramonium Agglutinin-Location in the Seed and
J. Agrlc. Food Chem. 1989, 37, 1005-1009 1005
List, G. R.; Spencer, G. F. Fate of Jimsonweed Alkaloids in
Soybean Processing. J. Am. Oil Chem. SOC. 1976,53,535-536.
List, R.; Spencer, G. F.; Hunt, W. H. Toxic Weed Seed Contam-
inants in Soybean Processing. J. Am. Oil Chem. Soc. 1979,
Lovett, J. V.; Levitt, J.; Duffield, A. M.; Smith, N. G. Allelopathic
Potential of Datura Stramonium (Thornapple). Weed Res.
1981,21, 165-170.
Mahler, D. The Jimsonweed High. JAMA, J. Am. Med. Assoc.
1975, 231, 138.
Mikolich, J. R.; Paulson, G. W.; Cross, C. J. Acute Anticholinergic
Syndrome due to Jimson Seed Ingestion. Ann. Intern. Med.
Mirzamotov, R. T.; Luftulin, K. L. Dynamics of the Accumulation
of Alkaloids in Datura Stramonium. Chem. Nut. Comp. (Engl.
Transl.) 1986,22(3), 359; Khim. Prir. Soedin. 1986,22(3), 381.
Nelson, P. D.; Mercer, H. D.; Essig, H. W.; Minyard, J. P. Jim-
sonweed Toxicity in Cattle. Vet. Hum. Toxicol. 1982, 24,
Oâ??Grady, T. C.; Brown, J.; Jacamo, J. Outbreak of Jimsonweed
Abuse among Marine Corps Personnel at Camp Pendleton.
Mil. Med. 1983,148, 732-734.
Orr, R. Reversal of Datura stramonium Delirium with Physo-
stigmine: Report of Three Cases. Anesth. Analg. 1975, 54,
Price, M. L.; Scoyoc, S. V.; Butler, L. G. A Critical Evaluation
of the Vanillin Reaction as an Assay for Tannin in Sorghum
Grain. J. Agric. Food Chem. 1978, 26, 1214.
Reaidi, G. B.; McPherson, L.; Bender, A. E. Toxicity of Red
Kidney Beans (Phaseolus Vulgaris). J. Sei. Food Agric. 1981,
Shervette, R. E.; Schydlower, M.; Lample, R. W. Jimson â??Locoâ??
Weed Abuses in Adolescents. Pediatrics 1979, 63, 520-523.
Testa, P.; Fontanelli, G. Intossicazione Acuta de Alcaloidi
56, 706-710.
1975,83, 321-325.
321-325.
158-161.
32,846-847.
Atropinosimili della Datura Metel. Clin. Ter. (Rome) 1988,125,
Ungerer, A.; Schmitz-Rourgeois, M.; Melan, C.; Goulanger, Y.;
Reinbolt, J.; Amiri, I.; De Barry, J. y-L-glutamyl-L-asparatate
(from Datura Stramonium) Induces Specific Deficits in
Long-term Memory and Inhibits Tritiated Glutamate Binding
on Hippocampal Membranes. Brain Res. 1988,446,205-211.
Urich, R. W.; Bowerman, D. L.; Levisky, J. A.; Pflug, J. L. Case
Report-Datura Stramonium: a Fatal Poisoning. J. Forensic
Sei. 1982, 27, 948-954.
van De Velde, H.; Demeyer, K.; Djaegere, R. Influence of Indo-
leacetic Acid and Dimethylallylaminopurine on Hyoscyamine
and Scopolamine Production in Datura Stramonium Vartatulu.
Acta Agron. Hung. 1988, 37, 55-64.
von Gottfried, W.; Schmidt, K. H. Chemical Analysis of (-)
Scopolamine Metabolism in Several Mammals. Z. Physiol.
Chem. 1968, 349, 741-752.
Walder, J. Antisense DNA and RNA: Progress and Prospects.
Genet. Deu. 1988,2,502-504.
Wallace, J. M.; Friedman, M. Inactivation of Hemagglutinins in
Lima Bean (Phaseolus Linatus) Flour by N-acetyl-L-cysteine,
pH, and Heat. Nutr. Rep. Int. 1985, 32, 743-748.
Weaver, S. E. Factors Affecting Threshold Levels and Seed
Production of Jimsonweed (Datura stramonium L.) in Soy-
beans (Glycine max. L.). Weed Res. 1986,26, 215-223.
Weintraub, S. Stramonium Poisoning. Postgrad. Med. 1960,28,
Williams, S.; Scott, P. The Toxicity of Datura Stramonium
(Thorn-apple) to Horses. N . 2. Vet. J. 1984, 32, 47.
Worthington, T. R.; Nelson, E. P.; Bryant, M. J. Toxicity of
Thornapple (Datura Stramonium) Seeds to the Pig. Vet. Res.
203-211.
364-367.
1981,108, 208-211.
Received for review September 1,1988. Accepted January 9,1989.
Red Squill Modified by Lactobacillus acidophilus for Rodenticide Use
Anthony J. Verbiscar,* Thomas F. Banigan, Rex E. Marsh, and Allen D. Tunberg
Red squill (Urginia maritima, Liliaceae) bulb and root preparations were treated with three strains
of Lactobacillus acidophilus, fortifying the cultures with dry milk and oat, wheat, and rice flour.
Lactobacillus growth with the production of a @-glucosidase converted bitter glucoside scilliroside to
its tasteless aglycon scillirosidin. These products were blended into rat diets at 0.03% scillirosidin levels,
and 95% of the female rats died. Although male rats usually ate more bait than the females, none ate
enough for a lethal dose of scillirosidin. The rats learned to avoid the baits if they did not die after
initial ingestion of these fast-acting rodenticides. Technical scillirosidin mixed into rat diets had a toxic
effect on female rats similar to the L. acidophilus treated red squill products.
Red squill is being investigated as a new economic crop
for the southwest United States where it grows well
(Gentry et al., 1987). The bulb, roots, and other plant parts
contain scilliroside [6@-(acetyloxy)-3@-(@-D-gluco-
pyranosyloxy)-8,14-dihydroxybufa-4,20,22-trienolide], a
highly toxic, emetic, and bitter bufadienolide glucoside
(Verbiscar et al., 1986a, b). Because of high toxicity and
the emetic safety factor, dried bulb powders have been
used in rat baits for centuries. However, rats and mice,
which are unable to vomit, may not eat a lethal amount
of red squill baits when first exposed, resulting in formu-
lation problems.
Anver Bioscience Design, Inc., Sierra Madre, California
91024 (A.J.V., T.F.B.), and Wildlife and Fisheries Biology,
University of California, Davis, California 95616 (R.E.M.,
A.D.T.).
Our initial attempts to improve acceptability involved
conversion of scilliroside to its aglycon scillirosidin, which
is tasteless and equally toxic. The aerobic fungus As-
pergillus niger was used as a source of @-glucosidase to
elicit this cleavage (Verbiscar et al., 1987). A. niger was
grown in extracts of red squill, producing the enzyme
necessary for the hydrolysis of scilliroside to scillirosidin.
The resulting aglycon extracts were administered orally
to Charles River rats. The scillirosidin aglycons were found
to be more toxic to female rats than to males, which is also
the case for scilliroside.
In addition to the A . niger study we tested 12 strains
of Lactobacillus bulgaricus and Lactobacillus acidophilus
on hand from a jojoba detoxification project. It seemed
reasonable that because Lactobacilli cleave lactose, a ga-
lactosylglucose, the active enzyme could also cleave the
glucose from scilliroside. The Lactobacilli are nontoxic and
microaerobic, which facilitates processing. The Lactobacilli
0021-8561/89/1437-1005$01.50/0 @ 1989 American Chemical Society

[Delta9]

J. A@. Food Chem. 1992, 40, 1617-1624 1617
Structural Relationships and Developmental Toxicity of Solanum
Alkaloids in the Frog Embryo Teratogenesis Assay-Xenopus
Mendel Friedman,',' J. R. Rayburn,t and J. A. Bantlet
Food Safety Research Unit, Western Regional Research Center, Agricultural Research Service,
U.S. Department of Agriculture, 800 Buchanan Street, Albany, California 94710, and Department of Zoology,
Oklahoma State University, Stillwater, Oklahoma 74078
Solanum plants produce potentially toxic alkaloids. As part of a program to improve safety of plant-
derived foods such as potatoes, we examined the relative embryotoxicities of 13 structurally different
compounds using the frog embryo teratogenesis assay-Xenopus (FETAX). Our purpose was to better
understand structural features governing the developmental toxicology of these compounds. We measured
the minimum concentrations needed to inhibit growth of the embryos, the median lethal concentration
of 96-h exposure (96-h LC50), and the concentration inducing gross terata in 50% of the surviving
animals [96-h EC50 (malformation)]. The following glycoalkaloids produced concentration-response
curves: a-chaconine, a-solanine, solasonine, and a-tomatine. All compounds were tested at equimolar
(0.005 and 0.015 mM) concentrations in order to develop a relative potency scale. The data showed
that (a) glycoalkaloids are more toxic than corresponding aglycons lacking the carbohydrate groups,
(b) for glycoalkaloids, the nature of the carbohydrate strongly influences potency, (c) the nitrogen of
the steroid is required for teratogenicity, (d) the orientation of the unshared electron pair associated
with the nitrogen atom does not affect potency, and (e) the presence of nitrogen in rings of nonsteroidal
alkaloids such as atropine, scopolamine, and ergonovine does not impart teratogenicity. The observed
structural effects should facilitate predicting developmental toxicities of compounds of dietary interest
without the use of live animals and provide information to guide selection of potato plants with a low
content of specific toxic alkaloids. The possible significance of the results to food safety is discussed.
INTRODUCTION
Alkaloids and glycoalkaloids are nitrogen-containing
secondary plant metabolites found in numerous plant
species. More than 20 structurally different alkaloids have
been recognized in potatoes and tomatoes and about 300
in other Solanaceae plants (Schreiber, 1979). The Solan-
aceae or "nightshade" family contains many plants im-
portant to animals and man, including potatoes (Solanum
tuberosum), tomatoes (Lycopersicon esculentum), cap-
sicum (Capsicum frutescens), eggplant (Solanum mel-
ongenu), tobacco (Nicotiana tabacum), black nightshade
(Solanum nigrum), and jimsonweed (Datura stramonium).
These plants produce antinutritional and toxic compounds
including glycoalkaloids, both during growth and after
harvest.
Relatively high concentrations of glycoalkaloids have
been found in Solanaceae plants consumed by man, such
as potatoes (Friedman and Dao, 1992). Levels are espe-
cially high in green and damaged potatoes and immature
green tomatoes. Glycoalkaloids are far more toxic to man
than to other animals studied, although there appears to
be considerable individual variation in the susceptibility
of animals and humans (Morris and Lee, 1984; Keeler et
al., 1991; Caldwell et al., 1991; Friedman, 1992; Friedman
and Henika, 1992). The toxicity may be due to adverse
effects on the central nervous system and to disruption of
cell membranes, adversely affecting the digestive system
and general body metabolism. The possible human
toxicity of the Solanum glycoalkaloids has led to the
* Author to whom correspondence should be addressed
t USDA.
t Oklahoma State University.
[telephone (510) 559-5615; FAX (510) 559-57771.
establishment of guidelines limiting the glycoalkaloid
content of new potato cultivars (Morris and Lee, 1984).
In order to decrease the biosynthesis of the most toxic
compound(s) in the plant, it is necessary to define a relative
toxicity scale to facilitate the design of suitable molecular
biology experiments. In a previous study, we evaluated
the embryotoxicity of several potato alkaloids in the frog
embryo teratogenesis assay-Xenopus (FETAX) as well as
the minimum concentration needed to inhibit the growth
of the embryos (Friedman et al., 1991). In terms of these
parameters, our results indicated that, overall, a-chaconine
was about 3 times more toxic than a-solanine and that
solanidine was not very toxic. This study (a) defines the
relative potencies at equimolar concentrations of a series
of structurally different Solanum alkaloids in the in vitro
frog embryo assay and (b) relates the results to reported
findings with hamsters and to the chemical structures of
the test compounds.
MATERIALS AND METHODS
Test Materials. Test compounds were obtained from Bio-
synth AG, Switzerland; Roth AG, Germany; and Sigma Chemical
Co., St. Louis, MO: atropine (Sigma), a-chaconine (Sigma lot
97F-7045), demissidine (Sigma lot 77F-0308), digoxigenin (Roth
7958), ergonovine (Sigma lot 94F-014511, scopolamine (Sigma),
a-solanine (Roth 5414), solanidine (Roth 5329), solasodine (Sigma
Lot 15F-40251), solasonine (Biosynth 2091), tomatidine (Sigma),
and tomatine (Sigma lot 76F-5031).
FETAX Tests. Seta of 25 embryos were placed in 60-mm
glass petri dishes with varying concentrations of the appropriate
test compound dissolved in FETAX solution. For each test
material, 6-10 concentrations were tested in duplicate. Four
control dishes of 25 embryos each were exposed to FETAX
solution alone, as described previously (Friedman et al., 1991).
For each test, probit analysis (Tallarida and Murray, 1980)
was used to determine the 96-hLC50 (median lethal concentration
This article not subject to U.S. Copyrlght. Published 1992 by the American Chemical Soclety
1618 J. Agric. FOW'Chem., Vol. 40, No. 9. 1992 Frledman et el.
6coqH a 8,O;H - a
CHPOH CH20H
ATROPINE SCOPOLAMINE
0
II HCH3
N - CH,
H
ERGONOVINE
H
DlGOXlGENlN
22aH, 25!3H-SOLANlD-!XN-O~R
Solanidine R =OH
a-Solanine R = 0-galactose - glucose I
rhamnose
a-Chaconine R I 0-glucose -rhamnose
I
rhamnose
Figure 1. Structures of compounds evaluated in the FETAX test.
or concentration causing 50% embryolethality), 96-h EC50
(malformation) (concentration inducing gross terata in 50 % of
the surviving embryos), and 95% confidence intervals. These
assays are used to define a teratogenic index (TI) of a test
compound equal to (LC50/EC50). The TI is useful in asseesing
teratogenicpotential (Bantle et al., 1990; Fort and Bantle, 1990).
Head-tail length was measured by following body contour. For
each test, the minimum concentration to inhibit growth (MCIG)
was calculated using the t test for grouped observations (P <
0.05).
Experimenta were conducted to compare developmental
toxicity at the same molar concentration of alkaloids. Compounds
were tested at 0.005 and 0.015 mM concentrations. The 0.005
mM concentration was selected because it was the highest
concentration at which all the alkaloids were soluble. The 0.015
mM concentration was selected because it was within the
developmental toxicity range of the most soluble of the alkaloids.
Four dishes of 25 embryos each were used at each concentration.
All other test conditions were the same as the concentration-
response testa described above.
RESULTS
Rankings Based on 96-h LC50 and EC50 Values.
Figure 1 depicts the structures of the compounds inves-
% NICOTINE
R J3Y 5 .
ii
(255)-228N-5a-SPIROSOLAN-OB-R
Tomatidine R = OH
Tometine R I 0-galactose -glucose -xylose
I
glucose
(25 R)-22aN-SPI ROSA L-5 EN-O&R
Solasodine R=OH
Solasonine R 0-galsctose -glucose
rhamnose I
Table I. Developmental Toxicity Ranking of Five Plant
Alkaloids Baaed on 96-h LCSO and ECSO Values
compound EC50," m M LC50," mM TI
a-solanine 0.010 (0.007-0.0132) 0.013 (0.012-0.014) 1.3b
solasonine 0.0057 (0.005-0.0067) 0.00633 (0.006-0.0067) 1.11
a-chaconine 0.0036 (0.0033-4.004) 0.004 (0.0034-0.0041) l.llb
tomatine 0.0018 (0.0017-0).002) 0.0019 (0.001&0.002) 1.05
nicotine 0.0028 (0.0025-0.0030) 0.869 (0.844-0.875) 31Oe
0: 95 5% confidence intervals in parentheses. b Data by Friedman et
al. (1991). Data by Dawson et al. (1988).
tigated in this study. Four glycoalkaloids were water
soluble enough to produce concentration-response curves.
These were a-solanine, a-solasonine, a-chaconine, and
a-tomatine. Their millimole LC50 and EC50 data are
compared from least to most developmentally toxic in
Table I and Figure 2. Molarity rather than concentration
(in mg/L) was used as a basis for comparison to account
for differences in molecular weight.
Rankings Based on Exposure at Highest Soluble
Concentrations. Figures 3-7 show rankings according
to mortality, malformation, and growth, respectively. For
each compound, 10% control mortality and 2.1 % control
Developmental Toxicity of Solanum Alkaloids in FETAX
0.015 I I
h
v t hl
0,010 2 e
d e:
b
u
E 0.005
2
0 u
0.000
4
9 6 - H R LC50
9 6 - H R EC50
Figure 2. Ranking of four plant alkaloids based on relative 96-h
LC50 and EC50 (malformation). The ranking is from lowest
developmental toxicity (highest LC50 and EC50) to highest
developmental toxicity (lowest LC50 and EC50). Bars represent
95 % confidence intervals.
5 90
-4 ?oo? * k 80
0 70
60
Q) 50
2
z
4
40
4
30
0
L 20
a,
A. l o
n
0.005 mM 0.015 mM
CONCENTRA TION
Figure 3. Ranking of plant alkaloids based on mortality.
Rankings are based on experimental minus controls and are
arranged from low mortality on the left to high mortality on the
right. Both tomatine (0.005 mM) and a-chaconine (0.015 mM)
caused 100% mortality and therefore do not appear on malfor-
mation and growth figures. ETOH refers to 0.1 % v/v ethanol
solvent controls. Bars are standard error of the mean of four
replicates.
malformation were deducted. The plots show net results
corrected for control values.
At 0.005 mM a-tomatine caused 100% mortality (90%
net mortality shown in Figure 3). a-Chaconine also caused
>50% mortality and malformation. Solasonine, however,
only caused 29% malformation and little mortality.
a-Solanine caused 11% net mortality and 17% net
malformation.
Digoxigenin and ergonovine both caused less than 5 %
net mortality and less than 10% net malformation.
Digoxigenin and ergonovine actually increased embryo
growth slightly compared to controlâ??s. This is the reason
for their negative reduction of growth (Figure 5).
G
0 100
(d 90
k
0
rcr â??O
.i
4
E 80
ri
60
50
z
-4
40
J. Agric. Food Chem., Vol. 40, No. 9, 1992 181#
0.005 mM 0.015 mM
CONCENTRATION
Figure 4. Rankings of plant alkaloids based on malformation.
Rankings are based on experimental minus controls and are
arranged from low mortality on the left to high mortality on the
right. ETOH refers to 0.1 % v/v ethanol solvent controls. Bars
are standard error of the mean of four replicates.
d
0
k
0
15 , , , , ( , ( , , , , , , , , , I /
1.
10 T
I
0.005 mM 0.015 mM
CONCENTRATION
Figure 5. Ranking of plant alkaloids based on growth inhibition.
Rankings are based on percent reduction of growth. In some
cases, embryos were longer than controls resulting in negative
growth reduction. However, these were not significantly different
from controls. ETOH refers to 0.1 % v/v ethanol solvent controls.
Bars are standard error of the mean of four replicates.
The aglycons tomatidine, solasodine, demissidine, and
solanidine all had between 10 and 30 â??3% mortality at 0,005
mM. Although they also showed a slight growth reduction,
they were not significantly different from controlâ??s.
a-Solanine, tomatidine, solanidine, and solasodine were
also tested at 0.015 mM. Tomatidine, solanidine, and
solasodine showed no noticeable increase in effect, and in
1620 J. Agric. Food Chem.. Vol. 40, No. 9, 1992
m w 3 - e:
2
1
Frledmen et al.
0
10 100
" 1 m 7
2 '
1 10 100 concentration ( mg/L)
z z 110 t 1
5 105 B
3 4 80 851 1 L 7 5 1 ' ' ' ' ' ' ' ' ' ' ' ' ' 0 20 40 60 80 100 120 140
PERCENT LC50
0 Malformation (Dash)
0 Mortality (Solid)
Figure 6. Solasonine concentration-response curves for mal-
formation, mortality, and growth. Increasing concentrations of
solasonine were tested in a solution of FETAX containing 0.1-
0.3% v/v ethanol. Solvent-only controls were also included. A:
0, mortality; 0, malformation. B: -, growth. Bars are for
standard error of the mean.
one case a slight decrease was seen. Also no effect was
seen on growth at 0.015 mM. a-Solanine caused 63% net
mortality, 94% net malformation, and significantly re-
duced growth.
It must be remembered that the lower the LC50 and
EC50, the higher the developmental toxicity (Figure 2).
For Figures 3-5, the higher the percent response at a given
concentration, the higher the toxicity. Each endpoint must
be considered separately when ranking toxicity.
Atropine and scopolamine did not significantly affect
mortality or malformation a t the concentrations we tested.
Three different range testa were performed, each testing
progressively higher concentrations. Atropine and sco-
polamine were not graphed in Figures 3-5 because their
toxicities were low.
Atropine has been tested up to 2.5 mg (8.63 mM), the
limit of solubility in FETAX. No test produced greater
than control mortality or malformation. Thus, all that
we can say is it has a higher solubility than the potato
alkaloids and its effects, as measured, are not different
from ergonovine and digoxigenin.
Scopolamine was tested up to 4 mg/mL (10 mM) without
reaching its limit of solubility. Mortality was the same as
the control's, and malformations were not significantly
different.
DISCUSSION
The frog embryo teratogenesis Assay-Xenopus
(FETAX) was used in this study because it allowed rapid
evaluation of the developmental toxicity. Compared to
other short-term assays such as fish embryos, FETAX is
faster because primary organogenesis is completed in only
4 days. Unlike fish assays, FETAX has undergone
extensive validation using compounds of known mam-
malian teratogenicity and has attained a predictive
EO
2 75 ' ' , " 1 ' ~ " ~ ~
0 20 40 60 80 100 120 140
PERCENT LC50
0 Malformation (Dash)
0 Mortality (Solid)
Figure 7. Tomatine concentration-response for mortality,
malformation, and growth. Increasing concentrations of tomatine
were tested in a FETAX solution in the presence of 0.143.3 % v/v
ethanol. Solvent-only controls were also included. The diluent
was FETAX solution. A 0, mortality; 0, malformation. B -,
growth. Bars are for standard error of the mean.
accuracy of nearly 90% (Bantle et al., 1989). The in vitro
frog embryo assays is, therefore, both time and cost
effective and correlates with extensive evaluations in live
mammals.
Early tests (Friedman et al., 1991) suggested that
microsomes used for metabolic activation in FETAX had
no effect on the developmental toxicity of potato alkaloids,
so microsomes were not employed in this study. Com-
pounds such as plant alkaloids can be ranked solely on the
bask of the percent malformation caused a t the same molar
concentration (Figure 4).
Developmental Toxicity of Plant Alkaloids and
Glycoalkaloids. Glycoalkaloids have the ability to induce
spina bifida (the defective closure of the vertebral column),
anencephaly (absence of part of the brain and skull), and
embryotoxicity (Sharma et al., 1978; Morris and Lee, 1984;
Keeler et al., 1991). Structural, stereochemical, and
electronic configurations of the alkaloid seem to be
paramount in influencing the teratogenic response. For
example, Brown and Keeler (1978) suggested that
22S,25R epimers with a basic nitrogen atom in the terminal,
non-steroidal F-ring, shared or unshared with ring E, and
with bonding capabilities a to the steroid plane, may be
teratogens (e.g., solasodine, jervine). In contrast,
22R,25S epimers with the unshared electron pair on
nitrogen having a j3 orientation (e.g., demissidine) should
not be teratogenic. Since the potato glycoalkaloids
a-chaconine and a-solanine have the 22R,25S stereo-
chemical configuration with the unshared electron pair
projecting above the plane of the steroid ring system (/3
projection), this hypothesis predicts that these compounds
should not be teratogenic. Limited studies summarized
by Keeler et al. (1991) demonstrated that this prediction
is not always realized, since both compounds turned out
to be teratogenic in hamsters and other species (see below).
Developmental Toxicity of Solanum Alkaloids in FETAX
Keeler et al. (1991) and Friedman et al. (1991) attempted
to explain the unexpected behavior of the two potato
glycoalkaloids by suggesting that other factors could
contribute to the teratogenic potencies. These include
(a) enzyme-catalyzed inversion of the unshared electron
from the ,3 to an a orientation, (b) the presence of small
amounts of the teratogenic 22S,25R epimers in potatoes
and in the glycoside preparations evaluated for terato-
genicity, (c) the presence of saponins, which may enhance
gastrointestinal absorption by their action as emulsifying
agents, (d) modification of teratogenicities by parallel
toxicities in the gastrointestinal tract and in the liver (Dalvi
and Jones, 1986; Caldwell et al., 1991), and (e) the
possibility that the true teratogens are not the native
alkaloids but metabolic transformation products. We will
evaluate the embryotoxicity-teratogenicity effects of the
structurally different alkaloids and the non-alkaloidal
steroid digoxigenin in light of these comments.
Digorigenin. Digoxigenin is the aglycon of the cardiac
glycoside digoxin, produced by the plant Digitalis lunata.
Such glycosides all have a lactone ring attached to the
pentanoperhydrophenanthrene steroidal nucleus (Fieser
and Fieser, 1959). Digoxigenin is used therapeutically in
congestive heart diseases.
This steroidal compound was included in our compar-
ative evaluation because, although structurally it resembles
the steroidal alkaloids, it lacks a nitrogen-containing six-
membered ring. Evidently, such a ring is required for
teratogenicity since digoxigenin was not teratogenic in frog
embryos.
Atropine and Scopolamine. Atropine and scopolamine
are produced by Solanaceae plants including Atropa
belladonna and Datura stramonium (Friedman and Levin,
1989). They both stimulate and depress the central
nervous system. Structurally, the tropane ring of both
consists of a piperidine ring fused to a tropane ring.
Scopolamine has an additional epoxide group built into
the tropane ring system (Figure 1).
Neither compound was toxic nor teratogenic to cultured
frog embryos. These results suggest that nitrogen-
containing ring systems of atropine-type compounds do
not have the ability to induce the formation of terata.
Ergonovine. Ergonovine, produced by the parasitic
fungus Claviceps purpurea, is the hydroxyisopropyl amide
derivative of lysergic acid. Ergot alkaloids are also
produced by morning glory (Ipomoea violaceae) plants
(Friedman and Dao, 1990).
Although not derived from a Solanum plant, lysergic
acid derivatives are widely distributed in nature, have
numerous medical uses, and are also abused as drugs (Lewis
and Elvin-Lewis, 1977). Our studies revealed that they
do not induce a teratogenic response in embryos, even
though structurally they have a nitrogen-containing six-
membered ring as part of a polycyclic ring system (Figure
1). Evidently, such a structure does not have the ability
to elicit a teratogenic response, at least not in frog embryos.
a-Chaconine, a-Solanine, Solanidine. The glycoalka-
loids a-chaconine and a-solanine are found in potatoes
(5'. tuberosum). They each have three carbohydrate
residues attached to the 3-OH group of the same aglycon
solanidine (Figure 1). Solanidine contains 27 carbon atoms
and one atom of tertiary nitrogen, which is part of the
bicyclic ring E and F.
Data from Renwick et al. (1984) tabulated by Keeler et
al. (1991) indicate that at an a-chaconine dose of 0.173
g/kg of body weight, 26% of pregnant hamsters died by
day 8 of gestation and 66% of the surviving dams had
deformed litters. The corresponding values for a-solanine
J. AQric. FOodChem., VOl. 40, NO. 9, 1992 1621
at a dose of 0.200 glkg were 8 and 57%, respectively.
Solanidine was not evaluated in hamsters. a-Chaconine
appears to be more toxic in terms of mortality and more
teratogenic than a-solanine. Combining mortality plus
teratogenicity gives a value of 92% for a-chaconine and
65 % for a-solanine. Analyzed differently, these numbers
show that of 100 litters, 25 were unaffected by either
mortality or teratogenicity when gavaged with a-chaconine,
versus 40 with a-solanine.
It should also be noted that high rates of malformation
were seen in the hamster study only when there was
significant maternal toxicity. The low TIS observed in
this study and in our previous work (Friedman et al., 1991)
suggest that malformations due to plant alkaloid toxicity
occur near the cytotoxic concentration.
In the present work, we compared the relative potencies
of several compounds at equimolar concentrations. Table
I and Figures 2-5 show that at 0.005 mM concentrations,
a-chaconine was about three times more toxic than
a-solanine interms of mortality, malformation, and growth
inhibition and that solanidine was much less toxic than
the two glycosides. Since a-chaconine and a-solanine differ
only in the nature of the Carbohydrate side chain attached
to the corresponding 3-OH group of solanidine, the nature
and possibly the number of the carbohydrate side chain
residues appears to be paramount in influencing terato-
genicity. This hypothesis is further reinforced by our
observation (unpublished results) that hydrolytic removal
of one or two carbohydrate residues resulta in a progressive
decrease in embryotoxicity.
Solasonine and Solasodine. Solasonine is produced by
a large number of Solanum plants including eggplants (S.
melongena) and a number of potato cultivars other than
S. tuberosum (Schreiber, 1979). Solasodine, the aglycon
of solasonine, contains one more oxygen atom than
solanidine. It has a spiroketal structure with a methyl
group in the a-position in ring F and a double bond in ring
B (Figure 1). Solasonine belongs to the 25R series. The
molecule contains the same three carbohydrate residues
found in a-solanine. The death rate of pregnant hamsters
gavaged with 0.47 glkg of solasonine for 8 days was 38 7% ;
the percent of formed litters from surviving dams was 0
(Keeler et al., 1991). The respective values for solasodine
gavaged with 0.200 glkg of body weight was 0 and 10, and
for solasodine gavaged with 1.4 glkg of body weight, 4 and
Qr) 3 1 .
Table I and Figures 2-6 show that solasonine's frog
embryotoxicity and teratogenicity overall falls between
those of a-solanine and a-chaconine. Thus, solasonine
affected malformations, mortality, and growth at molar
concentrations that were only about one-half those re-
quired for a-solanine to show the same effects. Since
solasonine has the same carbohydrate side chains attached
to the 3-OH group as a-solanine, differences in structural
features of the respective nitrogen-containing ring systems
may account for the observed differences in toxicities.
Solasodine was much less embryotoxic than the glyco-
side. Ita toxicity is comparable to that of the other
aglycons, solanidine and demissidine. The structure of
demissidine is identical to that of solanidine except that
the double bond of ring B is reduced.
Tomatine and Tomatidine. The steroidal glycoalkaloid
tomatine has been isolated from tomatoes (L. esculentum).
It has four carbohydrate residues attached to the 3-OH
group of the aglycon tomatidine (Figure 1). Structurally,
it is a spirosolane belonging to the 25s series, with a reverse
S configuration at the spiroatom 22 compared to the
spirosolane alkaloid, solasonine. Both tomatine and
1622 J. Agric. Food Chem., Vol. 40, No. 9, 1992 Friedman et al.
. .& .& * a,-
Figure 8. Effect of tomatine on Xenopus development, as seen
from top to bottom: control and larvae exposed to 2,2.4, and 2.6
mg/L, respectively. Note that pigmentation increases with
increasing concentration. Gut coiling becomes progressively loose
and embryo length is reduced as concentration increases. More
severely malformaed larvae died by 48 h A, side view of stage
46 (96-h) larvae: B, ventral view of stage 46 (96-h larvae).
solasonine serve as precursors for synthetic hormones such
as progesterone (Fieser and Fieser, 1959).
Keeler et al. (1991) reported that the aglycon tomatidine
was nontoxic and nonteratogenic when given as a gavage
to pregnant hamsters. The glycoside tomatine, by contrast,
was highly toxic though it induced no terata at a dose of
one-eighth the molar equivalent of tomatidine.
Table I and Figures 2-5 and 7 show that tomatine is
about 5 times more embryotoxic than a-solanine and
caused the greatest (100 % ) mortality compared to all other
compounds tested. Tomatine caused severe malforma-
tions similar to those induced by a-chaconine. These
embryos typically died by 48 h. Among the survivors,
malformations were generally scored as moderate (Figure
8). Abnormal pigmentation and loose gut coiling were the
most frequent abnormalities encountered. The aglycon
tomatidine strongly inhibited growth but mortality and
teratogenicity (malformation) were of the same order as
those resulting from the other aglycons.
Struct ure-Activity Relations hips. Digoxigenin has
the structural features of a steroid (Figure 1) but is not
a steroidal alkaloid, since it lacks a nitrogen atom. It
showed no effect on frog embryos, suggesting that nitrogen-
containing steroids are potential teratogens. This sug-
gestion is reinforced by the negative results with the
Solanum alkaloids atropine and scopolamine, which
contain a nitrogen atom as part of their bicyclic ring
systems, and the lysergic acid derivative ergonovine, which
has a six-membered piperidine ring attached to a poly-
cyclic, but not steroidal ring system (Figure 1).
Our studies also show that the teratogenicity/embry-
otoxicity of the steroidal aglycon is strongly increased by
the carbohydrate residues attached to the secondary 3-OH
groups of the corresponding glycosides. This increase is
not uniform, but depends on the nature and order of
attachment of the carbohydrate residues. Thus, the
relative potency of a-solanine, a trioside with three sugars,
glucose, galactose, and rhamnose, is significantly lower
than that of a-chaconine, a trioside in which the sugars
rhamnose, glucose, and galactose are attached to the same
aglycon. Other structural features also seem to influence
relative toxicities. For example, the observed activity of
solasonine in the frog embryo assay is intermediate
between that of a-solanine and a-chaconine, even though
solasonine has the same three carbohydrates as a-solanine.
Evidently, structural features in the steroidal part of the
glycoside also influence biological activity since both
a-solanine and a-chaconine have a tertiary nitrogen atom
as part of a bicyclic ring system, whereas solasonine
contains a spiroketal structure with both oxygen and
secondary nitrogen atoms. It is striking that tomatine,
which has four sugar residues (Figure 1) attached to a
spiroketal steroid, appears not to be teratogenic, although
it is highly toxic to the frog embryos. It is possible that
most embryos die before malformations can occur.
We can only speculate about possible biological mech-
anisms by which the carbohydrate side chains and the
nitrogen-containing rings of the aglycons influence toxicity.
One possibility is that the carbohydrate residues influence
biological activity by participating in binding to sugar
molecules associated with receptor sites of cell membranes.
Possibly, the behavior of glycoalkaloids is similar to that
of the glycoproteins called lectins (hemagglutinins), which
also exhibit a high degree of "sugar specificity" in their
ability to agglutinate red blood cells and other biological
manifestations (Liener, 1989). The toxic effects of these
molecules appear to be due to the ability of the carbo-
hydrate part to bind to specific receptor sites on the surface
of intestinal epithelial cells, resulting in cell damage and
interfering with the absorption of nutrients across the
intestinal wall. Incidentally, unlike potato glycoalkaloids,
heat-sensitive potato lectins are largely destroyed during
cooking.
In a previous study, Dawson et al. (1988) demonstrated
a high TI for nicotine (Table I). However, unlike potato
alkaloids, the potent teratogenicity of nicotine was largely
deactivated by liver microsomes. Figure 1 shows that the
structure of nicotine contains a tertiary nitrogen as part
of a five-membered ring, in analogy with solanidine. The
nicotine structure also contains a heterocyclic six-mem-
bered pyridine ring in analogy with ergonovine. Whether
the mechanisms of teratogenicity induced by nicotine
differs from that of the potato alkaloids and whether
deactivation of nicotine also takes place in vivo awaits
further studies.
The variable influence of the nitrogen-containing rings
of the steroidal part of the molecules on toxicity could
Developmental Toxlcity of Solanum AlkaloMs in FETAX
arise from differences in stereochemical features associated
with the structurally different rings and/or relative ba-
sicities of the secondary and tertiary nitrogen atoms.
Although the stereochemistry of the unshared electron
pairs on these nitrogens appears not to influence terato-
genicity, basicity (pK) could, if the nitrogens are involved
in binding to cell membrane receptor sites by charge
transfer and/or hydrogen bonding interactions. Further
studies are needed to explore these possibilities.
The justification for our attempt to relate structural
features of naturally occurring Solanum alkaloids to
observed biological activities deserves further comment.
With natural compounds, we are limited by the structures
provided by nature. Therefore, we may not be able to
show the rigorous structure-activity correlations common
with synthetic agricultural and medicinal compounds,
where a specific structural feature can be artificially and
systematically varied. Our attempt is similar to that of
Rosenkrantz and Klopman (1990), who related common
structural features of natural compounds associated with
biological activities to carcinogenicities.
FUTURE STUDIES
An unanswered question is whether the glycoalkaloids
would induce teratogenicity in pregnant mammals when,
as part of a normal diet, they are subject to interaction
with dietary nutrients and non-nutrients, digestion, ab-
sorption, transport, metabolism, and elimination. For
example, since we do not know whether the carbohydrate
residues associated with the steroid moieties are cleaved
by hydrochloric acid in the stomach or by digestive
enzymes, we do not know whether the glycosides or
hydrolysis products are the actual toxicants. Only toxicity
associated with oral ingestion of the pure glycoalkaloids,
hydrolysis products, or alkaloid-containing potato extracts
added to standard diets would give a realistic indication
of potential health hazards to animals and humans. Since
many of the reported animal studies were carried out by
administering the alkaloids or plant extracts by gavage
and/or injection, additional studies are needed to ascertain
whether the reported findings can be confirmed by oral
studies.
Another unanswered question is whether heat-labile
lectins and heat-resistant inhibitors of digestive enzymes
such as carboxypeptidase, chymotrypsin, and trypsin
which are present in potatoes (Friedman, 1992; Lisinska
and Leszczynski, 1989) modulate the biological effects of
the glycoalkaloids when consumed as part of a potato-
containing diet rather than individually. A related pos-
sibility is that the alkaloids themselves could act syner-
gistically in vivo since in vitro studies on the disruption
of cell membranes revealed strong synergism between
a-chaconine and a-solanine (Roddick et al., 1988). There-
fore, these considerations suggest that, although the
embryotoxicity/teratogenicity data we obtained with FE-
TAX generally parallel the reported data obtained with
pregnant hamsters, additional studies are needed to relate
the described in vitro frog embryo studies to teratoge-
nicities of the structurally different compounds in higher
animals when fed as part of a normal diet and to the
disruption of membrane potentials of frog embryos
(Blankemeyer et al., 1992). We expect that the in vitro
FETAX assay will be a helpful guide for such studies and
will t u n out to be of value both as a developmental toxicity
screen and as a useful model to study the mechanism of
teratogenesis and its prevention without the need of live
mammals.
Finally, since our data show that the aglycons lacking
carbohydrate side chains are much less toxic than the
J. Agric. Food Chem., Vol. 40, No. 9, 1992 1623
glycosides, blocking the glycosylation step in the biosyn-
thetic pathway should result in potatoes with significantly
lowered toxicity (Stapleton et al., 1991, 1992).
ACKNOWLEDGMENT
We thank Mendi Hull for the photographs of the
embryos, A. Stapleton for computer-drawing Figure 1, and
L. Bjeldanes, J. T. Blankemeyer, K. L. Stevens, and C. C.
Willhite for helpful comments.
LITERATURE CITED
Bantle, J. A.; Fort, D. J.; Dawson, D. A. Bridging the gap from
short-term teratogenesis assays to human health hazard
assessment by understanding common modes of teratogenic
action. Proc. Aquat. Toxicol. Symp. 1989,12, STP-1027,46-
58.
Bantle, J. A.; Fort, D. J.; Rayburn, J. R.; DeYoung, D. J.; Bush,
S. F. Further validation of FETAX evaluation of the
developmental toxicity of five known mammalian teratogens
and non-teratogens. Drug Chem. Toxicol. 1990,13,267-277.
Blankemeyer, J. T.; Stringer, B. K.; Rayburn, J. R.; Bantle, J. A.;
Friedman, M. Effect of potato alkaloids a-chaconine and
a-solanine on membrane potential of frog embryos. J. Agric.
Food Chem. 1992, in press.
Brown, D.; Keeler, R. F. Structure-activity relation of steroid
teratogens: 3. Solanidan epimers. J. Agric. Food Chem. 1978,
26,566-569.
Caldwell, K. A.; Grosjean, 0. K.; Henika, P. R.; Friedman, M.
Hepatic ornithine decarboxylase induction by potato glycoal-
kaloids in rats. Food Chem. Toxicol. 1991,29,531-535.
Dalvi, R. R.; Jones, R. D. H. Role of microsomal cytochrome
P-450 containing monoxygenases in the metabolism and
toxicity of certain alkaloids. Curr. Sci. 1986, 55, 558-561.
Dawson, D. A.; Fort, D. J.; Smith, G. L.; Newell, D. L.; Bantle,
J. A. Comparative evaluations of the developmental toxicity
of nicotine and cotinine with FETAX. Teratog., Carcinog.,
Mutagen. 1988,8, 329-338.
Fieser, L. F.; Fieser, M. Steroids; Reinhold: New York, 1959.
Fort, D. J.; Dawson, D. A.; Bantle, J. A. Development of a
metabolic activation system for the frog embryo teratogenesis
assay: Xenopus (FETAX). Teratog., Carcinog., Mutagen.
Friedman, M. Composition and safety evaluation of potato
berries, potato and tomato seeds, potatoes, and potato alka-
loids. In Evaluation of Food Safety; Finley, J. W., Robinson,
S. F., Armstrong, D. J., Eds.; ACS Symposium Series 484;
American Chemical Society: Washington, DC, 1992; pp 429-
462.
Friedman, M.; Dao, L. Effect of autoclaving and conventional
and microwave baking on the ergot alkaloid and chlorogenic
acid content of morning glory (Ipomoea tricolor Cav. cv.) seeds.
J. Agric. Food Chem. 1990,38,805-808.
Friedman, M.; Dao, L. Distribution of glycoalkaloids in potato
plants and commercial potato products. J. Agric. Food Chem.
Friedman, M.; Henika, P. R. Absence of genotoxicity of potato
alkaloids a-chaconine, a-solanine, and solanidine in the Ames
Salmonella and adult and fetal erythrocyte micronucleus
assays. Food Chem. Toxicol. 1992, in press.
Friedman, M.; Levin, C. E. Composition of Jimson toxic weed
seeds (Datura stramonium). J. Agric. Food Chem. 1989,37,
Friedman, M.; Rayburn, J. R.; Bantle, J. A. Developmental
toxicology of potato alkaloids in the frog embryo teratogenesis
assay - Xenopus (FETAX). Food Chem. Toxicol. 1991,29,
537-547.
Keeler, R. F.; Young, S.; Brown, D. Spina bifida, exencephaly,
and cranial bleb produced by the Solanum alkaloid solasodine.
Res. Commun. Chem. Pathol. Pharmacol. 1976,13,723-730.
Keeler, R. F.; Baker, D. C.; Gaffield, W. Teratogenic Solanum
species and the responsible teratogens. In Handbook of
Natural Toxins; Keeler, R. F., Tu, A. T., Eds.; Dekker: New
York, 1991; Vol. 6, pp 83-97.
1990,8, 251-263.
1992,40, 419-423.
998-1005.
1624 J. Agric. Food Chem., Vol. 40, No. 9, 1992
Lewis, W. H.; Elvin-Lewis, P. F. Medical Botany; John Wiley:
New York, 1977.
Liener, I. E. The nutritional significance of lectins. In Food
Proteins; Kineella, J. E., Soucie, W. G., Eds.; The American
Oil Chemists Society: Champaign, IL, 1989; pp 329-353.
Lisinska, G.; Leszczynski, W. L. Potato Science and Technology;
Elsevier: London, 1989.
Morris, S. C.; Lee, T. H. The toxicity and teratogenicity of
Solanacea glycoalkaloids, particularly those of the potato
(Solanum tuberosum). Food Technol. Aust. 1984, 36, 118-
124.
Renwick, J. H.; Claringbold, W. D. B.; Earthy, M. E.; Few, J. D.;
McLean, C. S. Neural-tube defects produced in Syrian hamsters
by potato glycoalkaloids. Teratology 1984,30, 371-378.
Roddick, J. G.; Rijnenberg, A. L.; Osman, S. Synergistic inter-
action between potato glycoalkaloids a-solanine and a-cha-
conine in relation to destablization of cell membranes: eco-
logical implications. J. Chem. Ecol. 1988, 14, 889-902.
Rosenkrantz, H. S.; Klopman, G. Natural pesticides present in
edible plants are predicted to be carcinogenic. Carcinogenesis
1990, 11, 34s353.
Schreiber, K. The steroid alkaloids of Solanum. In The Biology
and Taxonomy of the Solanaceae; Hawkes, J. G.,Ed.;Linnean
Society Symposium Series No. 7; Academic Press: New York,
1979; pp 193-202.
Frledman et 81.
Sharma, R. P.; Willhite, C. C.; Wu, M. T.; Salunkhe, D. K.
Teratogenic potential of blighted potato concentrate in rabbits,
hamsters, and miniature swine. Teratology 1978,18,55-62.
Stapleton, A,; Allen, P. V.; Friedman, M.; Belknap, B. L. Isolation
and characterization of solanidine glucosyltransferase from
potato sprouts. J. Agric. Food Chem. 1991, 39, 1187-1293.
Stapleton, A.; Allen, P. V.; Tao, H. P.; Belknap, W. R.; Friedman,
M. Partial amino acid sequence of potato solanidine UDP-
glucose glucosyltransferase purified by new anion exchange
and size exclusion media. Protein Expression Purif. 1992,3,
85-92.
Tallarida, R. J.; Murray, R. B. Manual of Calculations with
Computer Programs, Springer Verlag: New York, 1980, p 83.
Received for review October 28, 1991. Revised manuscript
received March 30, 1992. Accepted June 8, 1992.
Registry No. Atropine, 51-55-8; scopalamine, 51-34-3; nic-
otine, 5411-5; digoxigenin, 1672-46-4; ergonovine, 60-79-7; toma-
tidine, 77-59-8; tomatine, 17406-45-0; solanidine, 80-78-4; a-sola-
nine, 20562-02-1; a-chaconine, 20562-03-2; solasodine, 126-17-0;
solasonine, 19121-58-5.

[Delta9]

New Synthesis and Characterization of (+)-Lysergic Acid
Diethylamide (LSD) Derivatives and the Development of a
Microparticle-Based Immunoassay for the Detection of LSD and Its
Metabolites
Zhuyin Li, K. Goc-Szkutnicka, A. J. McNally,* I. Pilcher, S. Polakowski, S. Vitone, Robert S. Wu, and
S. J. Salamone
Drug Monitoring, Diagnostics Research and Product Development, Roche Diagnostic System, Inc., 1080 U.S.
Highway 202, Somerville, New Jersey 08876-3771. Received April 24, 1997X
In this paper are reported the synthesis and characterization of three LSD derivatives. On the basis
of several analytical characterization studies, the most stable derivative has been selected and a
procedure to covalently link the derivative to polystyrene microparticles through a carrier protein
has been developed. In addition, two new LSD immunogens have been synthesized and characterized,
and from these immunogens antibodies that recognize not only LSD but also several major LSD
metabolites have been generated. Using the selected derivative and antibody, a homogeneous
microparticle-based immunoassay has been developed for the detection of LSD in human urine with
the required sensitivity and specificity for an effective screening assay. The performance of this LSD
OnLine assay has been evaluated using the criteria of precision, cross-reactivity, correlation to the
Abuscreen LSD RIA and GC/MS/MS, assay specificity, and limit of detection.
INTRODUCTION
(+)-Lysergic acid diethylamide (LSD,1 Scheme 1) is a
highly potent drug that acts on the central nervous
system and alters sensory perception, states of consciousness,
and thought processes. By causing these altered
states, the drug produces severe visual and auditory
hallucinations. In addition to these physical effects, the
use of LSD has been and continues to be a problem for
drug and law enforcement agencies around the world (1-
4). Making the problem even more complex is the fact
that the detection of LSD in body fluids of users is
difficult because the quantities typically ingested are very
small (100-250 Ã?g/dose) (4). To date, limited research
has been conducted on the chemical properties of LSD.
It is, however, known that LSD has an inherent fluorescence,
which can be excited at 320 nm and emits at 445
nm. Additionally, under UV light irradiation, LSD can
undergo catalytic hydration at the C-9,10 double-bond
position. Once hydration occurs, the loss of fluorescence
at 445 nm is observed. In alkaline solution, LSD undergoes
an epimerization at the C-8 position, resulting in
partial formation of iso-LSD. LSD is also unstable under
prolonged heat exposure, but the mechanism of the
thermal decomposition is not yet fully understood (5, 6).
Under physiological conditions, LSD is rapidly and
extensively converted to several known and unknown
metabolic products. At present, several metabolites of
LSD in the human body, such as N-demethyl-LSD (nor-
LSD), 2-oxo-3-hydroxy-LSD, 13-hydroxy-LSD, and 14-
hydroxy-LSD, have been tentatively identified; yet only
one metabolite (nor-LSD) and the parent compound, both
excreted at 1% of the total dose, have been conclusively
identified (6-10).
Currently, the measurements of LSD and its metabolites
in biological fluids rely on radioimmunoassay methods
or HPLC fluorescence methods and very specialized
GC/MS/MS methods for confirmation (7-13). These
methods produce undesirable radioactive waste or require
extensive pretreatment of samples, very specialized
equipment, and highly trained personnel. There has
been much interest in recent years, because of the
reported increased abuse, to develop nonisotopic, highly
automated, homogeneous analytical methods to detect or
screen for LSD abuse. Due to the lower sensitivities of
nonisotopic immunoassays, the instability of LSD under
both physiological and nonphysiological conditions, and
the lack of information about the majority of the metabolites,
the development of nonisotopic immunoassays
to date has been a challenge. It is, therefore, extremely
important to systematically synthesize LSD derivatives
and study the stability of these derivatives for the
development of a nonisotopic immunoassay. A conjugate
procedure for making a LSD microparticle is also critical
for obtaining a highly sensitive LSD assay. In addition,
it is very important to generate antibodies that are
capable of recognizing not only LSD but also its major
metabolites. These antibodies must have low cross-
* Author to whom correspondence should be addressed.
X Abstract published in Advance ACS Abstracts, October 15,
1997.
1 Abbreviations: BTG, bovine thyroglobulin; DMSO, dimethyl
sulfoxide; DMF, N,N-dimethylformamide; CMC, N-cyclohexyl-
N¢-(2-morpholinoethyl)carbodiimide methyl-p-toluenesulfonate;
EDTA, ethylenediaminetetraacetic acid; ELISA, enzyme-linked
immunosorbent assay; GC/MS/MS, gas chromatography/tandem
mass spectrometry; KPi, potassium phosphate; NHBâ??H2O, Nhydroxybenzotriazole
hydrate; NHS, N-hydroxysuccinimide;
LSD, (+)-lysergic acid diethylamide; nor-LSD, N-demethyl-LSD;
NMR, nuclear magnetic resonance; SAMHSA, substance abuse
and mental health services administration; TEA, triethylamine;
THF, tetrahydrofuran; TLC, thin-layer chromatography; TNBS,
trinitrobenzenesulfonic acid.
Scheme 1. Structure of LSD
896 Bioconjugate Chem. 1997, 8, 896-905
S1043-1802(97)00059-1 CCC: $14.00 © 1997 American Chemical Society
reactivity to structurally related compounds that are not
substances of abuse.
In this paper, we describe the synthesis and characterization
of new LSD derivatives used in the development
of a microparticle-based OnLine immunoassay. We
also describe a new procedure to make a LSD microparticle.
Like other OnLine immunoassays, this assay
is based on the principle of the kinetic interaction of
microparticles in solution (KIMS) (14), in which the drug
content in the urine is directly proportional to the
inhibition of the microparticle agglutination. Four LSD
derivatives were synthesized, two of which (3 and 7,
Scheme 3) were used to prepare immunogens (4 and
8, Scheme 3) for antibody production and three of
which (6, 7, and 10, Schemes 2 and 3) were examined
for best stability. From this work the most stable
derivative was selected for preparing the conjugate. For
the development of this microparticle-based assay, the
selected LSD derivative was covalently coupled to a
carrier protein, and this conjugate was then covalently
linked to microparticles. These newly developed LSD
microparticles, together with the properly selected antibodies,
were then developed into a competitive displacement
immunoassay for LSD with a detection limit of 0.2
ng/mL LSD.
MATERIALS AND METHODS
Reagents. All solvents were obtained from Fisher
Scientific (Pittsburgh, PA) unless specified. All flash
grade silica gel and silica gel preparative TLC plates were
obtained from E. M. Science (Gibbstown, NJ). Protein
concentrations were determined by using the Bradford
protein assay reagents (15) purchased from Bio-Rad
(Hercules, CA), and 2-oxo-3-hydroxy-LSD was purchased
from Radian (Austin, TX). LSD, nor-LSD, 1-(3-aminobutyl)-
N,N-diethyl-D-lysergamide, and 1-(3-aminopropyl)-
N,N-diethyl-D-lysergamide (1a and 1b, Scheme 2) (16)
and Abuscreen RIA were prepared by Roche Diagnostic
Systems (Somerville, NJ). LSD antibodies for the
Abuscreen RIA were generated using a LSD analogue
derivatized through the indole nitrogen and conjugated
to BSA. Carboxylated polystyrene microparticles were
obtained from Bangs Laboratories (Carmel, IN). CMC,
NHBâ??H2O, ovalbumin, and other reagents were obtained
from Sigma (St. Louis, MO).
Instrumentation. Fluorescence measurements were
carried out by using an LS-5B luminescence spectrometer
(Perkin-Elmer, Norwalk, CT). The excitation wavelength
was set at 320 nm, and the emission wavelength was
measured at 445 nm. Light irradiation was performed
using a 20 W desk-top fluorescent light. Proton nuclear
magnetic resonance (1H NMR) spectra were recorded at
400 MHz on an XL-400 NMR spectrometer (Varian,
Palo Alto, CA); coupling constants are given in hertz (Hz),
and CDCl3 was used as the solvent. The abbreviations
used are as follows: s, singlet; d, doublet; t, triplet; m,
multiplet. The OnLine immunoassay was performed
using an Olympus AU800 automated analyzer (Olympus,
Lake Success, NY).
Synthesis of LSD Labels. Synthesis of 1-(3-Aminobutyl)-
N,N-diethyl-d-lysergamide and 1-(3-Aminopropyl)-
N,N-diethyl-d-lysergamide (2a and 2b, Scheme 2).
A solution of 900 mg (1.7 mmol) of 1a in 25 mL of
methanol was treated with 0.385 mL (12.3 mmol) of
anhydrous hydrazine and stirred at room temperature
overnight. The reaction mixture was concentrated at
reduced pressure. The residue was treated with 25 mL
of a mixture of 9:1 methylene chloride/methanol, and the
insoluble solids were filtered off. The filtrate was chromatographed
on 200 g of silica gel using 25% methanol
in methylene chloride as an eluent to elute front-running
impurities, followed by 2% triethylamine/25% methanol
in methylene chloride as an eluent to elute the product
to yield 563 mg (83%) of 2a as a yellow amorphous solid:
1H NMR (400 MHz, CDCl3) â?° 1.17 (3H, t, J ) 7.1), 1.24
(3H, t, J ) 7.1), 1.50-1.60 (2H, m), 1.82-1.92 (2H, m),
2.59 (3H, s), 2.62-2.92 (6H, m), 3.02-3.10 (1H, m), 3.18-
3.26 (1H, m), 3.37-3.57 (6H, m), 3.84-3.92 (1H, m), 4.08
(2H, t, J ) 6.8), 6.33 (1H, s), 6.79 (1H, s), 7.11-7.20 (3H,
m); MS, m/e 394 (M+); HR-EI MS calcd for M+ 394.2733,
found 394.2731. Likewise, 2b was obtained using an
analogous procedure in 80% yield.
Synthesis of 1-[[[(4-Isothiocyanatophenyl)carbonyl]-
amino]butyl]-N,N-diethyl-d-lysergamide (3, Scheme 3). A
solution of 370 mg (0.94 mmol) of 2a in 15 mL of
anhydrous THF was cooled to 0 °C and treated with a
solution of 190 mg (0.96 mmol) of 4-isothiocyanatobenzoyl
chloride (17) in 5 mL of anhydrous THF and then
stirred at 0 °C for 30 min and then at room temperature
overnight; the reaction was driven to completion by
adding 0.14 mL (1.0 mmol) of triethylamine and stirred
at room temperature for 2 h. The reaction mixture was
concentrated at reduced pressure. The residue was
dissolved in methylene chloride, washed with H2O, dried
over anhydrous sodium sulfate, and concentrated at
reduced pressure. The residue was chromatographed on
200 g of silica gel using 3% methanol in methylene
chloride as an eluent to yield 325 mg (62%) of 3 as a tan
amorphous solid: 1H NMR (400 MHz, CDCl3) â?° 1.17 (3H,
t, J ) 7), 1.24 (3H, t, J ) 7), 1.55-1.65 (2H, m), 1.80-
2.05 (4H, m), 2.61 (3H, s), 2.63-2.75 (1H, m), 2.85-2.95
(1H, m), 3.04-3.12 (1H, m), 3.18-3.28 (1H, m), 3.35-
3.58 (4H, m), 4.13 (2H, t, J ) 6.4), 5.97-6.03 (1H, m),
6.35 (1H, s), 6.79 (1H, s), 7.10-7.20 (3H, m), 7.24 and
7.67 (4H, AA¢ BB¢q, J ) 8.4); MS, m/e 555 (M+); [R]D)+
47.5° (c 0.91%; CHCl3).
Synthesis of 4¢-[[2,5-Dioxo-1-pyrrolidinyl)oxy]carbonyl]-
[1,1¢-biphenyl]-4-carbonyl Chloride (5, Scheme 2). A
mixture of 2.0 g (8.2 mmol) of 4,4¢-biphenyldicarboxylic
acid in 40 mL of anhydrous THF was treated with 5.0
mL (55.0 mmol) of oxalyl chloride followed by 0.02 mL of
anhydrous DMF. The reaction was stirred at room
temperature for 10 min and then heated to reflux for 90
min. The reaction was then concentrated at reduced
pressure to a yellow oil. This was recrystallized from a
mixture of THF and ether to yield 1.67 g (73%) of the
diacid chloride as yellow needles: 1H NMR (200 MHz,
CDCl3) � 7.75 and 8.22 (8H, AA¢ BB¢q, J ) 8); MS, m/e
278 (M+).
A solution of 1.67 g (6.0 mmol) of 1,1¢-biphenyl-4,4¢-
dicarbonyl chloride in 65 mL of anhydrous THF was
treated with 710 mg (6.17 mmol) of N-hydroxysuccinimide,
followed by 0.835 mL (6.0 mmol) of triethylamine.
The reaction was stirred at room temperature for 2 h,
after which time it was filtered to remove triethylamine
HCl. The filtrate was concentrated at reduced pressure
to yield 2.0 g (93%) of 5 as a pale yellow solid: IR (CHCl3)
1775, 1742 cm-1; 1H NMR (400 MHz, CDCl3) â?° 2.94 (4H,
br s), 7.77 (4H, d, J ) 8.5), 8.23-8.27 (4H, m); MS, m/e
357 (M+).
Synthesis of 1-[3-[[[4¢-[[(2,5-Dioxo-1-pyrrolidinyl)oxy]-
carbonyl][1,1¢-biphenyl]-4-yl]carbonyl]amino]propyl]-N,Ndiethyl-
d-lysergamide (6, Scheme 2). A solution of 850
mg (2.375 mmol) of 5 in 65 mL of anhydrous THF under
argon was cooled to 0 °C in an ice bath and then treated
with a solution of 900 mg (2.365 mmol) of 2b in 50 mL of
anhydrous THF and 0.6 mL (4.3 mmol) of triethylamine
added dropwise over a 20 min period. The reaction
mixture was stirred at 0 °C for 1 h and then warmed to
room temperature, with stirring, for 1 h. The mixture
LSD Derivatives Bioconjugate Chem., Vol. 8, No. 6, 1997 897
was concentrated at reduced pressure, and the residue
was dissolved in 100 mL of methylene chloride. The
solution was washed with 100 mL of H2O, 100 mL of
saturated sodium bicarbonate solution, and 100 mL of
saturated brine solution, dried over anhydrous sodium
sulfate, and concentrated at reduced pressure to a brown
residue. This was chromatographed on a short column
of 100 g of silica gel using first methylene chloride as an
eluent, then 9:1 methylene chloride/ether as an eluent
to elute front-running impurities, and then 14:1 methylene
chloride/isopropyl alcohol as an eluent to elute the
product to yield 650 mg (39%) of 6 as a cream-colored
solid: IR (CHCl3) 1772, 1743 cm-1; 1H NMR (400 MHz,
CDCl3) â?° 1.12-1.22 (6H, m), 2.18-2.22 (2H, m), 2.52 (3H,
s), 2.60-2.68 (1H, m), 2.72-2.82 (1H, m), 2.93 (4H, s),
3.00-3.10 (2H, m), 3.38-3.58 (7H, m), 3.85-3.92 (1H,
m), 4.24-4.32 (2H, m), 5.63-5.71 (1H, m), 6.41 (1H, s),
6.86 (1H, s), 7.20-7.25 (3H, m), 7.36 and 7.57 (4H, AA¢
BB¢ q, J ) 8.8), 7.75 and 8.21 (4H, AA¢ BB¢ q, J ) 8.8);
MS, m/e 702 (M+). An almost equal amount of the dimer
was also obtained as a yellow oil: IR (CHCl3) 3445, 2780,
1653, 1627 cm-1; 1H NMR (CDCl3) â?° 1.16 (3H, t, J ) 7.2),
Scheme 2. Synthesis of LSD Derivatives
898 Bioconjugate Chem., Vol. 8, No. 6, 1997 Li et al.
1.21 (3H, t, J ) 7.1), 2.15-2.23 (2H, m), 2.51 (3H, s),
2.56-2.64 (1H, m), 2.77 (1H, t, J ) 11), 2.96-3.03 (1H,
m), 3.03-3.09 (1H, m), 3.35-3.47 (5H, m), 3.47-3.55 (2H,
m), 3.80-3.86 (1H, m), 4.20-4.30 (2H, m), 5.90-5.96 (1H,
m), 6.85 (1H, s), 7.17-7.25 (3H, m), 7.45 and 7.54 (4H,
AA¢ BB¢q, J ) 8.1); MS, m/e 967 (M+H).
Scheme 3. Preparation of LSD Protein Conjugates
LSD Derivatives Bioconjugate Chem., Vol. 8, No. 6, 1997 899
Synthesis of 1-[[5-[8b-9,10-Didehydro-8-[(diethylamino)
carbonyl]ergolin-6-yl]-1,5-dioxopentyl]oxy]-2,5-pyrrolidinedione
(7, Scheme 3). A solution of 200 mg (0.65
mmol) of nor-LSD in 10 mL of anhydrous THF, under
argon, was treated with 161 mg (0.65 mmol) of 5-[(2,5-
dioxo-1-pyrrolidinyl)oxy]-5-oxopentanoyl chloride (20, 21),
followed by 0.2 mL (1.4 mmol) of anhydrous triethylamine.
The reaction mixture was stirred at room temperature
for 30 min and then concentrated at reduced
pressure. The residue was dissolved in methylene chloride,
washed with H2O and saturated aqueous sodium
bicarbonate solution, dried over anhydrous sodium sulfate,
and concentrated at reduced pressure to yield 330
mg (98%) of 7 as a yellow amorphous solid: UV (CH3-
OH) �max 308 ( ) 8980); IR (KBr) 3396, 1814-1739, 1634,
1628 cm-1; 1H NMR (400 MHz, CDCl3) â?° 1.08-1.18 (3H,
m), 1.25-1.38 (3H, m), 2.05-2.25 (2H, m), 2.43-3.25 (6H,
m), 2.83 (4H, s), 3.28-3.60 6H, m), [4.26 (d, J ) 13.6)
(major) and 5.02 (d, J ) 13.6) (minor)] (1H, rotamers),
[4.73-4.81 (minor) and 5.25-5.33 (major)] (1H, m, rotamers),
6.38 (1H, m), [6.90 (major) and 6.95 (minor)] (1H,
s, rotamers), 7.08-7.30 (3H, m), 8.00 (1H, m); MS, m/e
521 (M+H).
Synthesis of 8â??-6-(3-Aminopropyl)-9,10-didehydro-N,Ndiethylergoline-
8-carboxamide (9, Scheme 2). Alkylation
of nor-LSD with iodopropylphthalimide was carried out
according to the procedure of Marzoni and Garbrect (16).
The resulting phthalimide derivative (820 mg, 1.65 mmol)
in 30 mL of methanol was treated with 0.4 mL (12.7
mmol) of anhydrous hydrazine and stirred at room
temperature overnight. The reaction mixture was filtered
and concentrated at reduced pressure. The residue
was chromatographed on 100 g of silica gel using 2%
triethylamine/15% methanol in methylene chloride as an
eluent. Fractions containing product were combined and
rechromatographed on 150 g of silica gel using 2%
triethylamine/15% methanol in chloroform as an eluent
to yield 560 mg (93%) of 9 as a yellow amorphous solid:
IR (CHCl3) 3479, 1663, 1624 cm-1; 1H NMR (400 MHz,
CDCl3) â?° 1.17 (3H, t, J ) 7), 1.25 (3H, t, J ) 7), 1.70-
1.80 (1H, m), 1.83-1.93 (1H, m), 1.90 (2H, s), 2.60-2.73
(2H, m), 2.79-2.87 (1H, m), 2.90-2.99 (2H, m), 2.99-
3.09 (1H, m), 3.13-3.20 (1H, m), 3.25-3.33 (1H, m),
3.35-3.52 (6H, m), 6.28 (1H, s), 6.88 (1H, s), 7.10-7.17
(2H, m), 7.17-7.23 (1H, m), 8.20 (1H, br s); MS, m/e
366 (M+).
Synthesis of 1-[[[4¢-[[[3-[8â??-9,10-Didehydro-8-[[(diethylamino)
carbonyl]ergolin-6-yl]propyl]amino]carbonyl]-1,1¢-
biphenyl]-4-yl]carbonyl]oxy]-2,5-pyrrolidinedione (10,
Scheme 2). A solution of 1.32 g (3.7 mmol) of 5 in 50 mL
of anhydrous methylene chloride under argon was cooled
to 0 °C and treated with a solution of 535 mg (1.46 mmol)
of 9 in 50 mL of anhydrous methylene chloride added
dropwise over a 30 min period. After the addition was
completed, the reaction mixture was washed with a
saturated aqueous sodium bicarbonate solution, dried
over anhydrous sodium sulfate, and concentrated under
vacuum. The residue was chromatographed on 100 g of
silica gel using 5% isopropyl alcohol as an eluent.
Fractions containing product were combined and concentrated
at reduced pressure to a yellow solid. The solid
was redissolved in ether and concentrated five times to
remove residual isopropyl alcohol to yield 280 mg (28%)
of 10 as a yellow solid: IR (CHCl3) 3479, 1773, 1743, 1636
cm-1; 1H NMR (400 MHz, CDCl3) â?° 0.86 (3H, m), 1.11
(3H, t, J ) 6.9), 1.82-1.92 (1H, m), 1.93-2.10 (1H, m),
2.70-2.88 (4H, m), 2.93 (4H, s) 3.08-3.20 (2H, m), 3.22-
3.38 (4H, m), 3.40-3.60 (4H, m), 3.60-3.70 (1H, m),
3.98-4.08 (1H, m), 6.35 (1H, s), 6.93 (1H, s), 7.15-7.25
(3H, m), 7.61 and 7.95 (4H, AA¢ BB¢q, J ) 8), 7.67 and
8.19 (4H, AA¢ BB¢q, J ) 8.3), 8.00 (1H, s), 9.05 (1H, br
s); MS, m/e 688 (M+H); [R]D ) +9° (c 0.355%, CHCl3).
Preparations of Protein Conjugates. Synthesis of
1-[[[(4-Isothiocyanatophenyl)carbonyl]amino]butyl]-N,Ndiethyl-
d-lysergamide-BTG (4, Scheme 3). A solution of
698 mg of BTG in 20 mL of 50 mM KPi, pH 7.5, was
cooled to 0 °C and treated with 58 mL of DMSO, added
dropwise, very slowly over a 2 h period. The mixture was
treated with a solution of 90 mg (0.16 mmol) of 3 in 2
mL of DMSO, added dropwise very slowly. The reaction
mixture was stirred at room temperature for 18 h, poured
into a dialysis bag of 50 kDa molecular weight cutoff, and
dialyzed 108-fold in 50 mM KPi, pH 7.5. The resulting
conjugate was filtered through a 0.22 Ã?m sterile filter to
yield 116 mL of the LSD-BTG immunogen 4. The protein
concentration was determined to be 5.3 mg/mL. The
degree of drug substitution on the BTG protein was
determined by the ability of remaining uncoupled lysine
residues to react with TNBS (18, 19). Unmodified BTG,
at the same concentration as the conjugate, was treated
in the same manner with TNBS to provide a control. This
procedure produced a yellow complex with an absorbance
maximum at 325 nm and was used to calculate the drug
substitution expressed as percent modification. The
assay showed a 63.8% modification of available lysines
on BTG.
Synthesis of 5-[8â??-9,10-Didehydro-8-[(diethylamino)-
carbonyl]ergolin-6-yl]-1,5-dioxopentyl-BTG (8, Scheme 3).
A solution of 700 mg of BTG in 13 mL of 50 mM KPi, pH
7.5, was cooled to 0 °C and treated with 13 mL of DMSO,
added dropwise, very slowly. After the addition was
complete, a solution of 84 mg (0.16 mmol) of 7 in 1 mL of
DMSO was added dropwise very slowly. The reaction
mixture was stirred at room temperature for 18 h, poured
into a dialysis bag of 50 kDa molecular weight cutoff, and
dialyzed 106-fold in 50 mM KPi, pH 7.5. The resulting
conjugate was filtered through a 0.22 Ã?m sterile filter to
yield the LSD-BTG conjugate 8. The protein concentration
was determined to be 12.1 mg/mL. The TNBS assay
showed a 45% modification of available lysines in BTG.
LSD Derivatives Exposed to Fluorescent Light,
Oxygen, and Different Solution pH Values. All LSD
derivatives (6, 7, and 10) were placed in quartz cuvettes,
and the irradiation experiments were conducted under
the following conditions: LSD derivatives were dissolved
in DMSO at a concentration of 1 mg/mL and further
diluted to 2.6 mM in 10 mM KPi buffer, pH 7.5, containing
0.09% NaN3, 5 mM EDTA, and 0.1% Tween 20. The
solutions were kept at 2-8 °C in the dark for 3 days to
ensure a complete hydrolysis of the NHS ester (e.g. 10a,
Scheme 2). This is indicated by a complete disappearance
of the starting material by TLC examination (data
not shown). For the stability to light study, samples were
exposed to a 20 W desk fluorescent light at room
temperature. The distance between the fluorescent light
source and the experimental samples was 15 cm. The
experimental samples were subjected to irradiation for
various times, and then fluorescence from the samples
was measured, with �ex ) 320 nm and �em ) 445 nm.
Triplicate measurements were performed under each
condition. Coefficients of variation were found to be
e1.5% for the analytical method. From these data we
established that >9% loss in fluorescent intensity represented
significant decomposition of the tested compounds.
This change of 9% represents three standard
deviations from the mean, which is a >99% confidence
interval. Corresponding LSD derivatives kept in the
dark, at room temperature, were used as controls. These
experiments were designed to investigate the effect of
indoor light on the stability of the C-9,10 double-bond
900 Bioconjugate Chem., Vol. 8, No. 6, 1997 Li et al.
position of LSD derivatives and to rank the light stabilities
of these derivatives.
For stability to oxygen and solution pH studies,
samples were placed in amber glass bottles and saturated
with either oxygen or argon. Subsequently, these experimental
samples were incubated at 45 °C for 10 days.
Samples kept at 2-8 °C in the dark with argon were
classified as the control. These experiments were conducted
to explore the effect of oxygen, pH, and heat on
the stability of the C-9,10 double-bond position of LSD
derivatives and to rank the stability of these derivatives
under these conditions.
Antibody Generation. Several goats were placed on
a modified immunization program, as described by
Vaitukaitis (22) using LSD immunogens (4 or 8). Briefly,
immunogen 4 or 8 was mixed with Freundâ??s adjuvant,
and 1 mg of the immunogen containing complete Freundâ??s
was injected into multiple sites across the back of
each goat. At week two, each goat continued to receive
1 mg of the immunogen containing incomplete Freundâ??s.
This injection was repeated twice at 1 week intervals,
followed by a monthly injection of 0.5 mg of the immunogen
mixed with incomplete Freundâ??s adjuvant for a
period of 6 months.
The individual animals were monitored for antibody
titer and for cross-reactivity with LSD, nor-LSD, and
2-oxo-3-hydroxy-LSD by an ELISA method. Specifically,
a selected derivative was covalently coupled to a carrier
protein (ovalbumin). Polystyrene 96 well microtiter
plates were coated with 50 Ã?L of a 1.6 Ã?g/mL LSD-ovalbumin
conjugate in PBS buffer (50 mM KPi, pH 7.2, containing
150 mM NaCl) and allowed to incubate for 2 h
at room temperature or overnight at 2-8 °C. The plates
were washed with PBS buffer and blocked with 1% BSA
in PBS buffer. Fifty microliters of LSD, nor-LSD, or
2-oxo-3-hydroxy-LSD diluted in 1% BSA/PBS buffer at
various concentrations or 50 Ã?L of 1% BSA/PBS buffer
without the drug as a control was added into each well.
Fifty microliters of the appropriate antiserum in 1% BSA/
PBS buffer was then added to each well. The plates were
incubated for 1 h at 37 °C and then washed with PBS/
Tween 20 buffer. Anti-goat-alkaline phosphatase conjugate
and p-nitrophenyl phosphate were then used to
generate a detection signal. Two criteria were used to
select antibodies: (1) affinity of antibodies as estimated
by IC50; (2) inhibition of solid-phase antibody binding by
soluble LSD and its major metabolites, namely, nor-LSD
and 2-oxo-3-hydroxy-LSD. Once several animals were
selected from an immunogen, a pool of antiserum was
made and used to develop the immunoassay.
Preparation of LSD-Ovalbumin Conjugate (11,
Scheme 3). Seventeen and a half milliliters of ovalbumin
solution at 25 mg/mL in 50 mM KPi buffer, pH 7.5,
was cooled in an ice bath, and to this was slowly added
10 mL of DMSO. Seven and a half miliigrams of selected
LSD derivative was then dissolved in 1.5 mL of anhydrous
DMSO to make a 5 mg/mL solution. The LSD
derivative solution was added dropwise into the ovalbumin
solution with stirring, and stirring was continued
for 18 h at room temperature. The resulting LSDovalbumin
conjugate was dialyzed 1012-fold using 30 kDa
molecular weight cutoff dialysis bags. The final total
protein concentration of LSD-ovalbumin conjugate was
determined according to the Bradford protein assay (15).
Characterization of LSD-Ovalbumin Conjugate.
To determine the concentration of the noncovalently
bound LSD in the LSD-ovalbumin conjugate, the following
experiments were conducted: 1.5 mL of 12.5 mg/
mL LSD-ovalbumin conjugate in KPi buffer, pH 7.5, was
heat stressed at 45 °C in a white nontransparent polyethylene
container (HDPE) for 24 h. It was then immediately
mixed with 1.5 mL of 40% DMF in 10 mM KPi,
pH 7.5. The material was then left at room temperature
for 2 h, placed in a ovalbumin precoated Centricon filter
(Amicon, Beverly, MA) with a molecular weight cutoff of
30 kDa, and centrifuged at 600g for 2 h. The resulting
filtrate was then analyzed by fluorescence spectroscopy.
Completely hydrolyzed LSD-biphenyl-NHS (6) solutions
with concentrations ranging from 200 to 3200 ng/mL
were used as fluorescence standard, and a linear regression
method was used to generate a calibration curve.
The concentration of LSD in the filtrate was derived from
the standard curve. Filtrate from ovalbumin treated according
to the same method as the LSD-ovalbumin conjugate
was used as the control for background measurements.
A method was also designed to measure the total
number of LSD derivatives per ovalbumin in the LSDovalbumin
conjugate. The LSD-ovalbumin conjugate
was diluted to a concentration of 0.125 mg/mL of total
protein, and the fluorescence at 445 nm from the diluted
conjugate solution was measured. Ovalbumin treated
according to the same method as the LSD-ovalbumin
conjugate was used as the control for a blank measurement.
Completely hydrolyzed LSD-biphenyl-NHS (6)
solution concentrations ranging from 200 to 3200 ng/mL
were again used as fluorescence standards to generate a
calibration curve. Total concentration of LSD molecules
per ovalbumin was estimated from the calibration curve.
Preparation of the LSD Microparticle (12, Scheme
3). Ten milliliters of carboxyl-modified microparticle
(10% solids) was first washed by centrifugation at 10000g
with 0.1% Tween 20 in H2O. To each milliliter of
particles was added 20 mL of 0.1% Tween 20 in water,
the mixture was centrifuged and decanted, and the
particles were subsequently resuspended. This process
was repeated five times, and the microparticle concentration
was then adjusted to 3% (w/v) with a 0.1% Tween
20 solution. One and two-tenths milliliters of NHB (25
mg/mL, 0.37 mmol), previously dissolved in DMSO, was
then added slowly to the 30 mL of microparticle suspension,
under rapid stirring conditions, and the suspension
was stirred for 10 min at 25 °C. To this suspension was
added 1.7 mL of a freshly prepared CMC solution (50 mg/
mL, 0.34 mmol), and the mixture was stirred slowly for
3 h at 25 °C. The material was then washed according
to the method of centrifugation described above. The
washed, activated microparticles (45 mL at 2%) were
immediately mixed with LSD-ovalbumin/ovalbumin mixture
at different molar ratios (total protein concentration
was fixed at 3.1 mg/mL) diluted in 50 mM sodium
bicarbonate buffer, pH 8.6, and this mixture was allowed
to stir for 15 h at 25 °C. The resulting LSD microparticles
were then washed according to the method of
centrifugation described above using a wash solution of
10 mM KPi buffer, pH 7.5, containing 0.09% NaN3 and
0.1% Tween 20. The washed microparticle was then
resuspended in this buffer at 1.0% solids (w/v).
Development of the LSD Immunoassay. The LSD
immunoassay contains three reagents: (1) the antibody
reagent, which was made by placing the titered antibody
in a solution of 50 mM HEPES, pH 6.5, containing 0.1%
BSA protein, 0.5% NaCl, and 0.09% NaN3; (2) a reaction
buffer containing 50 mM PIPES, pH 7.0, with 2-3%
PEG, 2% NaCl, and 0.09% NaN3; and (3) a LSD microparticle
reagent, diluted from 1% stock solution to 0.2%
solids in a buffer containing 10 mM KPi, pH 7.5, 0.09%
NaN3, and 0.1% Tween 20. In addition, LSD calibrators
at concentrations between 0 and 1 ng/mL in normal urine
containing 0.09% NaN3 were used. The concentration of
LSD Derivatives Bioconjugate Chem., Vol. 8, No. 6, 1997 901
these LSD standards were verified by GC/MS/MS method.
The antibody concentration was adjusted so that the
agglutination of the LSD microparticles was inhibited
proportionally to the LSD concentration in the calibrators.
The light scattering difference between different
calibrators was also maximized in the calibration range
(0-1 ng/mL) to obtain maximum sensitivity.
Cross-reactivity to structurally related compounds was
conducted as follows: Normal human urine samples were
spiked with the structurally related compound of LSD
at various concentrations and tested as unknowns in the
OnLine assay. The percent cross-reactivity of a structurally
related compound was determined using the concentration
of the compound that provided displacement
equivalent to 0.5 ng/mL (1.5 nM) LSD.
One thousand presumed negative urine specimens
were obtained from a large drug abuse screening laboratory.
These samples had been previously screened and
found to be negative for the SAMHSA five panel (cannabinolds,
opiates, cocaine metabolite, amphetamine, and
phencyclidine). At the time of analysis, these samples
were simultaneously screened for LSD with the Abuscreen
RIA and OnLine assay. The Abuscreen RIA has a
>99.5% accuracy rate and was used as a reference
method in addition to GC/MS/MS method. In addition,
LSD positive samples were supplied by Dr. R. Foltz and
Dr. D. Kuntz (Northwest Toxicology, Salt Lake City, UT).
These samples had been previously screened positive by
the Abuscreen RIA and were subsequently confirmed by
GC/MS/MS. These samples were received frozen and
were stored at -20 °C until the day of analysis. For analysis,
a qualitative screening assay was performed as
follows: a single point calibration standard was used, and
its absorbance value was assigned as the cutoff value. A
positive result was reported if the sample absorbance
value was greater than or equal to the absorbance of the
cutoff calibrator.
RESULTS AND DISCUSSION
Synthesis and Stability Studies of the LSD Derivatives.
A detailed analytical study of the synthesized
LSD derivatives was necessary to determine which
derivatives should be used to develop the LSD OnLine
assay. The effect of exposure to fluorescent light on the
LSD derivatives is shown in Figure 1. The C-9,10 double
bond of LSD as previously reported can undergo photocatalytic
hydration (5). A potential structure change or
instability at this position was indicated by a change in
fluorescence intensity when compared to a control. It
was demonstrated that the fluorescence intensity of each
compound decreased as the exposure time to fluorescent
light increased. These results suggested that the order
of stability of the LSD derivatives to photon-catalyzed
hydration is LSD-biphenyl-COOH (hydrolyzed from 6) >
nor-LSD-biphenyl-COOH (hydrolyzed from 10) > nor-
LSD-aliphatic-COOH (hydrolyzed from 7). It is hypothesized
that the improved stability of 6 may come from
the presence of biphenyl group at the N-1 position of the
LSD molecule, where the biphenyl moiety can effectively
stack on the LSD and exclude water molecules from
interaction, thus stabilizing the photolabile C-9,10 center.
Such hydrophobic stacking interaction has been documented
(23, 24). When the biphenyl modification was
at position 6 of the LSD molecule, the biphenyl linker
could not effectively stack on the LSD and the protection
efficiency was reduced. The aliphatic moiety at position
6 would be predicted to offer little or no protection effect
to the LSD molecule, and indeed, this molecule had the
least stability under these conditions. The LSD-biphenyl
compound (6) was chosen for the construction of the
conjugate used in the nonisotopic immunoassay. This
compound displayed a 50-60% decomposition after having
been exposed to the described conditions for 4 days.
To ensure good assay stability, this issue was addressed
by placing the microparticle reagent containing this
derivative in a nontransparent polyethylene container.
An accelerated stability study indicated that this container
protects the reagent from light to ensure <10%
loss at normal room light conditions for 1 year. Further
data to emphasize the need not to have decomposition
are reflected in cross-reactivity studies of antibodies
generated from LSD immunogen (4) to completely photodecomposed
LSD-biphenyl-COOH (6). It was found
that the cross-reactivity of this decomposed material was
<20% (data not shown), which would cause serious loss
in assay sensitivity if the derivative was allowed to
decomposed.
Under heating conditions, oxygen may have an effect
on the C-9,10 double bond (25). However, it appears from
our experiments that oxygen did not affect the stability
of the C-9,10 double bond for the LSD derivatives. Since
protons can catalyze the hydration of the C-9,10 double
bond (5, 25), we explored the effect of pH on the hydration
of different LSD derivatives at high temperatures. Solution
pH was chosen to be 6.0 or 7.5, because this is the
most acceptable pH range for immunochemical reactions.
At pH 6.0, in the presence or absence of oxygen, after 10
days at 45 °C, compound 10 lost 20% of its fluorescence
intensity, while fluorescence intensity changes for compounds
6 and 7 were negligible. No significant changes
in fluorescence intensity were observed for all three
compounds when they were kept at pH 7.5 for 10 days
at 45 °C. The results suggested that the LSD derivatives
were more stable in pH 7.5 buffer than in pH 6.0 buffer.
Synthesis of LSD Immunogen and the Generation
of LSD Antibodies. Since only 1% of ingested LSD
is excreted in urine and the typical ingested dose is very
low, it is prudent to generate LSD antibodies that can
recognize not only LSD but also LSD metabolites, yet
avoid other undesirable ergot alkaloids compounds.
Figure 1. Effect of fluorescent light irradiation on the stability
of LSD derivatives. Structural changes of LSD derivatives at
the C-9,10 position were indicated by changes in fluorescence
intensity. Condition: 2.6 mM LSD derivatives in 10 mM KPi,
pH 7.5, buffer containing 5 mM EDTA, 0.09% NaN3, and 0.1%
Tween 20. All experiments were conducted at 25 °C. Corresponding
LSD derivatives kept in the dark at 25 °C were used
as controls. The Y axis represents the fluorescence intensity of
irradiated LSD derivatives when compared to their controls. The
X axis is exposure time to fluorescent light. Error bar represents
( 2 SD.
902 Bioconjugate Chem., Vol. 8, No. 6, 1997 Li et al.
Selectivity of an antibody to a hapten can be directed
such that the antibody preferentially recognizes the part
of the molecule that is farthest away from the attachment
of the hapten to a carrier protein (26). It was assumed
that this will allow the antibody to tolerate changes to
the hapten near its point of attachment to the carrier
protein. Under this hypothesis, we designed two immunogens
(4 and 8). These immunogens directed antibodies
to be less specific regarding changes near the indole ring
of LSD (27) or at position 6 of LSD; therefore, crossreactivity
toward 2-oxo-3-hydroxy-LSD, 13-hydroxy-LSD,
14-hydroxy-LSD, and nor-LSD would be predicted to be
higher for antibodies generated with immunogen 4.
Antibodies generated from immunogen 8 would be expected
to have high cross-reactivity toward 2-oxo-3-
hydroxy-LSD and nor-LSD.
LSD-biphenyl-NHS (6) was used to construct the LSDovalbumin
conjugate label for the ELISA method. The
binding between the LSD-ovalbumin conjugate label and
the antibodies generated from immunogen 4 can be
displaced with LSD (100%), nor-LSD (20%), and 2-oxo-
3-hydroxy-LSD (50%). When antibodies generated from
immunogen 8 were used, the binding between the LSDovalbumin
conjugate label and antibodies could be displaced
by LSD and had high cross-reactivity to nor-LSD
(40%) but had <20% cross-reactivity to 2-oxo-3-hydroxy-
LSD. From the cross-reactivity studies, antibodies raised
against immunogen 4 were selected, and, from the
derivative stability studies, derivative 6 was chosen for
development of a LSD immunoassay.
Preparation and Characterization of LSD-Ovalbumin
Conjugate. For a LSD assay, a 1:1 molar ratio
between LSD-biphenyl-NHS (6) and ovalbumin was used
to synthesize the LSD-ovalbumin conjugate stock solution.
Since the LSD derivative is a highly hydrophobic
compound, it can be trapped in ovalbumin or absorbed
on the ovalbumin surface noncovalently. This unbound
LSD derivative could be gradually released from the
conjugate and could cause poor stability. The unbound
LSD derivative could react with titered antibody and
reduce the sensitivity of the LSD assay. Therefore, it
was important to develop a reproducible dialysis procedure
to remove the unbound LSD from the LSD-ovalbumin
conjugate. It was also important to establish a
method to measure the noncovalently bound LSD derivatives
on the conjugate so that methods could be selected
that prevent this from occurring. To accomplish this, the
conjugate was denatured at 45 °C and then in 20% DMF
solvent to release any noncovalently bound LSD derivative.
Under these mild denaturing conditions, no protein
precipitation was observed. Our results indicated that
free LSD derivative accounted for <0.3% of total LSD
derivative loaded on ovalbumin when an extensive
dialysis (1012-fold) procedure was used. This procedure
was necessary to obtain a stable OnLine LSD assay
reagent.
Methods were also developed to quantify the number
of the LSD molecules per ovalbumin molecule in the
Figure 2. Representative concentration response curve on the
Olympus AU800 analyzer of the Abuscreen LSD OnLine assay.
Absorbance (ABS) is expressed as the milliabsorbance multiplied
by a factor of 10. Error bar represents ( 1 SD.
Table 1. Qualitative Precision of Abuscreen LSD OnLine
Assaya
a The Olympus AU 800 analyzer multiplies the milliabsorbance
units by a factor of 10 to report results. Precision was determined
to be <3.0% at 0 ng/mL (blank) urine.
Table 2. Cross-Reactivities of Abuscreen OnLine LSD
Assay
LSD Derivatives Bioconjugate Chem., Vol. 8, No. 6, 1997 903
LSD-ovalbumin conjugate stock solution. Since amino
groups at the protein surface are used in the coupling,
normally, the degree of drug substitution can be determined
by the ability of remaining uncoupled amine
residues that react with TNBS. However, since the LSD
substitution ratio was low (mean < one LSD per ovalbumin),
poor results were obtained using the TNBS
method. It was found that the number of LSD molecules
in the conjugate could be estimated directly using the
LSD fluorescence intensity. The fluorescence from ovalbumin
at the same concentration as the conjugate was
insignificantly small compared to the total intensity.
Using completely hydrolyzed LSD-biphenyl-NHS (6) as
a standard, the conjugates have been shown to contain
0.6-0.8 LSD molecules per ovalbumin molecule. Because
coupling of LSD to ovalbumin would change the fluorescence
quantum yield from LSD, this method only provided
an estimate of the loading of LSD molecules per
ovalbumin and was useful for quality control purposes
(data not shown).
Synthesis and Characterization of LSD-Microparticle
Conjugate. Several different molar ratios of
the drug protein conjugate to microparticle were evaluated
to determine the optimal ratio that produced the
best dose response curve. In the development of a
microparticle-based immunoassay, it is important that
proper agglutination occurs in the absence of free antigen.
To accomplish this, proper amounts of drug protein
conjugate must be coupled to each microparticle such that
an equivalence point can be reached, allowing the crosslinking
of microparticles by antibody. Excess antigen or
excess antibody in the system will prevent the formation
of the large aggregates produced by cross-linking.
To establish the proper substitution of drug onto
microparticle, LSD was first covalently coupled to the
ovalbumin protein (stock conjugate) followed by mixing
of the LSD-ovalbumin conjugate with ovalbumin at
various molar ratios and coupling the LSD-ovalbumin/
ovalbumin mixture to the microparticle. Each coupled
microparticle was then titrated against the antibody to
determine the performance of each LSD-ovalbumin/
ovalbumin molar ratio. The ratio that gave the greatest
dose response curve and the lowest nonspecific binding
(agglutination rate in the absence of antibody) was
selected. This was determined to be a molar ratio of 1:8
(LSD-ovalbumin/ovalbumin).
Development of LSD Assay. Using an endpoint
analysis reading at 520 nm, a dose response curve was
generated with various concentrations of LSD as shown
in Figure 2. The light scattering difference measured by
light transmission from 0 ng/mL to the cutoff concentration
of LSD (0.5 ng/mL) was >130 milliabsorbances (mA);
the overall difference from 0 to 1.0 ng/mL was >240 mA.
Table 1 shows that the qualitative intra-assay (n ) 20)
and interassay (n ) 100) precision had CVs of <5%.
Table 2 illustrates the cross-reactivity of the LSD
OnLine assay to structurally related compounds of LSD.
As expected, this assay had low cross-reactivity to iso-
LSD (2.4%, molar concentration); the cross-reactivity to
nor-LSD was 25% and to 2-oxo-3-hydroxy-LSD was 32%.
The cross-reactivities of other structurally related compounds
that are undesirable to detect, such as serotonin,
tryptophan, ergotamine, egonovine, and others, were
<0.002%. Finally, the limit of detection (LOD) of the
assay was determined by performing 20 replicate assays
on the 0 ng/mL calibrator. Two standard deviations
below the mean yields a LOD of <0.2 ng/mL LSD.
Table 3 shows the correlation of the OnLine LSD
screening assay with RIA and GC/MS/MS methods. GC/
MS/MS confirmed LSD positive clinical samples were
used to study patient correlation. Eighty-one positive
samples were tested in the OnLine LSD assay. The
distribution of LSD concentration in these samples is also
shown in Table 3. Twenty percent of the samples
Table 3.
904 Bioconjugate Chem., Vol. 8, No. 6, 1997 Li et al.
contained <0.5 ng/mL of LSD according to GC/MS/MS
data. Due to the high cross-reactivity to major LSD
metabolites, all of these samples were positive by OnLine
LSD assay. One thousand presumptive negative samples
were also tested; 993 were negative and 7 were positive.
All of the presumptive negative samples were found to
be negative by RIA. These seven samples that were
OnLine positive and RIA negative were found to be GC/
MS/MS negative.
CONCLUSION
A homogeneous microparticle-based immunoassay has
been developed for the detection of LSD in human urine
with the required sensitivity and specificity. Three major
issues were considered when this assay was developed:
(1) the stability of the LSD derivatives; (2) the stability
of LSD microparticles; (3) the cross-reactivity of antibodies.
Light, temperature, and solution pH can alter the
structure of LSD at the C-9,10 double-bond position.
Therefore, it is desirable to prepare and select a derivative
that generates a stable LSD microparticle which is
able to withstand long-term storage. On the basis of
stability studies, we have selected LSD-biphenyl-NHS (6)
as the best derivative for the development of the LSD
OnLine assay. Besides a stable LSD derivative, LSD
microparticles free of noncovalently bound LSD are
necessary for the OnLine technology to obtain the
required assay sensitivity and reagent stability. The
conjugation and dialysis procedures reported here have
allowed us to minimize unbound LSD in the LSD microparticle
reagent and to achieve the targeted sensitivity
and stability of the immunoassay. Due to the extent of
LSD in vivo metabolism and low ingestion dosage, the
concentration of parent compound (LSD) in urine is
extremely low. To overcome this, we designed and
selected an immunogen using a LSD analogue derivatized
through the indole nitrogen and conjugated the
derivative to BTG. The antibody generated by this
immunogen has demonstrated broad reactivity toward
LSD and several LSD metabolites. All of these factors
were essential in the successful development of the LSD
OnLine assay, which demonstrated excellent clinical
sensitivity.
ACKNOWLEDGMENT
We thank Dr. D. Kuntz and Dr. R. Foltz for providing
the GC/MS/MS data for all of the LSD-positive samples
used in this study and Mr. E. Nowaswiat for providing
compounds 1a and 1b. We also thank Dr. L. Arabshahi,
Ms. L. Allison, Dr. K. Savoca, and Dr. K. Schwenzer for
their technical help during the development of this assay.
LITERATURE CITED
(1) Bonner, R. (1992) Drug Detection Report, Vol. 1, p 5, Pace
Publications, Washington, DC.
(2) National Institute on Drug Abuse. (April 13, 1993) Annual
National High School Senior Survey, Rockville, MD.
(3) National Narcotics Intelligence Consumers Committee
(NNICC). (June 1991) The NNICC Report 1990: The Supply
of Illicit Drugs to The United States, Drug Enforcement
Administration, Washington, DC.
(4) Gold, M. S. (1994) The epidemiology, attitudes, and pharmacology
of LSD use in the 1990s. Psychiatr. Ann. 24, 124-
26.
(5) Hoffmann, A. (1975) LSDsA Total Study (D. Siva Sankar,
Ed.) pp 107-140, PHD Publications, Westburg, NY.
(6) Foltz, R. B., and Foltz, R. L. (1989) Lysergic acid diethylamide
(LSD). In Advances in Analytical Toxicology (R. C.
Baselt, Ed.) Vol. 2, pp 140-158, Year Book Medical Publishers,
Chicago, IL.
(7) Nelson, C. C., and Foltz, R. L. (1992) Chromatographic and
mass spectrometric methods for determination of lysergic acid
diethylamide (LSD) and metabolites in body fluids. J. Chromatogr.
580, 97-109.
(8) Nelson, C. C., and Foltz, R. L. (1992) Determination of
lysergic acid diethylamide (LSD), iso-LSD, and N-deSmethyl-
LSD in body fluids by gas chromatography/tandem mass
spectrometry. Anal. Chem. 64, 1578-85.
(9) Lim, H. K., Andrenyak, D., Francom, P., and Foltz R. L.
(1988) Quantification of LSD and N-demethyl-LSD in urine
by gas chromatography /resonance electron capture ionization
mass spectrometry. Anal. Chem. 60, 1420.
(10) Cai, J., and Henion, J. (1996) Elucidation of LSD in vitro
metabolism by liquid chromatography and capillary electrophoresis
coupled with tandem mass spectrometry. J. Anal.
Toxicol. 20, 27-37.
(11) Diagnostic Product Corp. COAT-A-COUNT LSD assay,
1989.
(12) Roche Diagnostic Systems, Abuscreen RIA LSD package
insert, June 1993.
(13) Peel, H. W., and Boynton, A. L. (1980) Analysis of LSD in
urine using radioimmunoassay-excretion and storage effects.
Can. Soc. Forensic Sci. J. 13, 23-28.
(14) Looney, C. E. (1984) High-sensitivity light scattering
immunoassays. J. Clin. Immunoassays 7, 90-95.
(15) Bradford, M. M. (1983) A rapid and sensitive method for
the quantitative determination of microgram quantities of
protein utilizing the principal protein-dye binding, Anal.
Biochem. 72, 245-254.
(16) Marzoni, G., and Garbrect, W. L. (1987) N¢-Alkylation of
dihydro lysergic acid. Synthesis, 651-653.
(17) Ziegler, K., Frimmer, M., Mullner, S., and Fasold, H. (1984)
3-Isothiocyanato-benzamido[3H] cholate, a new affinity label
for heportocellular membrane proteins responsible for the
update of both bile acids and phalloidin. Biochim. Biophys.
Acta 773, 11-12.
(18) Goldfarb, A. R. (1966) A kinetic study of the reactions of
amino acids and peptides with trinitrobenzesulfonic acid.
Biochemistry 5, 2570-2574.
(19) Snyder, S. L., and Sobocinski, P. Z. (1975) An improved
2,4,6-trinitrobenzenesulfonic acid method for the determination
of amines. Anal. Biochem. 64, 284-288.
(20) Janda, K. D., Schloeder, D., Benkovic, S. J., and Lerner,
R. A. (1988) Introduction of an antibody that catalyzes the
hydrolysis of an amide bond. Science 241, 1188-1191.
(21) Figii, I., Lerner, R. A., and Janda, K. D. (1991) Enantiofacial
protonation by catalytic antibodies. J. Am. Chem. Soc.
113, 8528-8529.
(22) Vaitukaitis, J. (1981) Production of antisera with small
doses of immunogen: multiple intradermal injections. Methods
Enzymol. 73B, 46-52.
(23) Wiley, R. A., and Rich, D. H. (1993) Peptidomimetics
derived from natural products. Med. Res. Rev. 13, 327-384.
(24) Desai, M. C., Vincent, L. A., and Rizzi, J. P. (1994)
Importance of parallel vectors and â??hydrophobic collapseâ?�of
the aligned aromatic rings: discovery of a potent substance
P antagonist. J. Med. Chem. 37, 4263-4266.
(25) Morrison, R. T., and Boyd, R. N. (1981) Organic Chemistry,
3rd ed., Allyn and Bacon Inc., Boston, MA.
(26) Erlanger, B. F. (1980) The preparation of antigenic hapten
carrier conjugates: a survey. Methods Enzymol. 70, 85-104.
(27) Ratcliffe, W. A., Fletcher, S. M., Moffat, A. C., Ratcliffe, J.
G., Harland, W. A. and Levitt, T. E. (1977) Radioimmunoassay
of lysergic acid diethylamide (LSD) in serum and urine
by using antisera of different specificities. Clin. Chem. 23 (2),
169-174.
BC9700594
LSD Derivatives Bioconjugate Chem., Vol. 8, No. 6, 1997 905

[Delta9]

494 J. Agric. Food Chem. 1991, 39, 494-501
Mutagenicity of Toxic Weed Seeds in the Ames Test: Jimson Weed
(Datura stramonium), Velvetleaf (Abutilon theophrasti), Morning Glory
(Ipomoea spp.), and Sicklepod (Cassia obtusifolia)
Mendel Friedman' and Philip R. Henika
Western Regional Research Center, Agricultural Research Service, US. Department of Agriculture,
800 Buchanan Street, Albany, California 94710
Commercial grain, such as soybean and wheat, may be contaminated with nongrain impurities such as
toxic weed seeds that coexist with harvested crops. The present study investigated the genotoxic
potential of seeds from four nongrain sources: jimson weed (Datura stramonium), velvetleaf (Abutilon
theophrasti), sicklepod (Cassia obtusifolia), and morning glory (Ipomoea spp.). Mutagenic responses
of methanolic extracts of these seeds were determined by using four bacterial strains (TA98, TA100,
TA102, and TA2637) with and without microsomal activation. Relative potencies were compared by
using the following parameters derived from dose-response curves: (a) mutagenic potency or revertants
per milligram equivalent seeds (RIMES), which is defined as the maximal mutagenic response produced
by 1 mg of seeds, and (b) mutagenic potential, which relates minimal effective dose (MED) with the
estimated number of seeds required to produce a significant mutagenic response. Although the seed
extracts elicited responses in all four bacterial strains, TA102 was the most sensitive. The following
numbers of toxic weed seeds per 15 g of grain are estimated to constitute a minimal effective dose with
TA102 and microsomal activation: morning glory, 1; sicklepod, 6; velvetleaf, 50; jimson weed, 566.
These results show that morning glory and sicklepod seeds contain high levels of mutagens. Possible
sources of mutagens and possible reasons for the observed variation of relative mutagenic potency are
discussed. These observations provide a rational basis for relating seed composition to genotoxic effects
and for assessing the possible safety of low levels of weed seeds in the diets of food-producing animals
and in human diets.
INTRODUCTION
Commercial grain shipments may contain nongrain
contaminants, including the seeds from plants that co-
existed with the harvested crops. Some of these seeds
may contain moderately or highly toxic components and
may not be readily separable during the normal cleaning
process. Such seeds have been documented to present
serious problems, including illness and death of livestock
and concern over food safety for humans (Dugan et al.,
1989).
Although it is generally recognized that certain problem
weed seeds contain toxic principles, species variability in
identity and concentration of toxins is unknown. The
extent of human and animal exposure to the toxic
principles is also unknown, as are the genotoxic effects of
toxic weed seeds in humans and animals. The current
genotoxicity database is insufficient for possible risk
assessment and the setting of tolerance standards for these
food contaminants.
Two basic kinds of information need to be integrated
to establish maximum tolerance levels for toxic seeds in
grain: (1) qualitative and quantitative compositional data
on the variation of the toxic principles found in each seed
species and (2) potency data in animals for the known and
unknown toxic principles, alone and in combinations as
found in whole seeds.
The main objective of this study was to screen extracts
of seeds from four toxic weeds for potential mutagenicity
by using the Ames test (point mutation in bacteria). The
overall goal is to assess the possible significance of geno-
toxicity for the purpose of setting tolerance levels of the
weed seeds in edible grain such as wheat and soybeans.
This study complements previously described analytical,
compositional, chemical, nutritional, and toxicological
studies of some of these same seeds (Crawford and Fried-
man, 1990; Crawford et al., 1990; Dugan et al., 1989; Fried-
man and Levin, 1989; Friedman and Dao, 1990; Friedman
et al., 1989).
MATERIALS AND METHODS
Materials. Seed samples of sicklepod (Cassia obtusifulia),
jimsonweed (Datura stramonium), morning glory (Ipomoeapur-
purea), and velvetleaf (Abutilon theophrasti) were obtained from
the Federal Grain Inspection Service (Kansas City, MO) or the
Valley Seed Co. (Fresno, CA). Seedsamples were picked through
to remove debris and then ground in a Wiley mill, as previously
described (Friedman and Levin, 1989; Dugan et al., 1989). Afla-
toxin BI, danthron, emodin, rutin, and quercetin were purchased
from Sigma Chemical Co. (St. Louis, MO). Extraction solvents
employed were hexane and chloroform (glass distilled; EM
Sciences, Cherry Hill, NJ), methanol (HPLC grade; Fischer
Scientific, Fair Lawn, NJ), and Milli-Q deionized water.
Seed Extraction Methodology. The extraction scheme used
was modified from one described by van der Hoeven et al. (1983)
for vegetable extraction. Our extraction procedure started with
two hexane extractions instead of petroleum ether extraction to
remove seed oils which can interfere with the scoring of Ames
test plates. The deoiled seed samples were subsequently ex-
tracted with chloroform, methanol, and Milli-Q water. All
extractions were performed in a Soxhlet apparatus starting with
400 mL of solvent and 5-15 g of seed sample weighed in a What-
man cellulose thimble (33 X 94 mm). Solvent boilingwas adjusted
with a rheostat so that the numbers of fills of the thimble were
as follows: hexane, 20 fills in 3 h; chloroform, 20 fills in 3 h;
methanol, 40 fills in 6 h; Milli-Q water, 20 fills in 6 h.
The extracts were evaporated in a rotary evaporator from 400
to 100 mL or less. They were divided into two samples of equal
volume and evaporated to dryness in preweighed 50-mL boiling
flasks. Evaporation temperatures were as follows: 45 "C, hex-
ane; 45 "C, chloroform; 45 "C, methanol; and 65 "C Milli-Q water.
The first sample was weighed, reconstituted in DMSO at 45 "C
This article not subject to US. Copyright. Published 1991 by the American Chemical Society
Mutagenicity of Toxic Weed Seeds
for 1 h, and left unhydrolyzed. The second sample was weighed
and then hydrolyzed in a boiling water bath for one 1 h with 2.5
mL of 2 N HCl and 5.0 mL of ethanol. AcidIethanol hydrolysis
has been frequently used for the hydrolysis of flavonoid glyco-
sides (Brown and Dietrich, 1979). We observed a 93 70 conversion
of pure rutin to quercetin (data not shown).
The hydrolysate was reconstituted in 25 mL of ethanol and
evaporated three times: first at 70 "C, then in the water bath
warming from 70 to 100 "C, and finally at 100 "C. In a duplicate
series, unhydrolyzed samples from the methanol extract were
subjected to boiling, ethanol reconstitution, and ethanol evap-
oration for the purpose of verifying that heat alone was not
responsible for mutagen formation.
Unhydrolyzed and hydrolyzed samples were reconstituted in
a desired volume of DMSO at 45 "C for 1 h. Particulate DMSO-
insoluble material was removed by filtration through a single
layer of Schleicher and Schuell595 filter paper. DMSO-insoluble
oils were removed by separation in a pipet. The reconstituted
samples were checked for UV absorbance in the 260-400-nm
range.
Mutagenicity Assays with Seed Extracts. The Ames
Salmonella /microsome assay was performed by using Salmonella
typhimurium tester strains TA98, TA100, TA102, and TA2637
kindly provided by Dr. Bruce Ames (University of California,
Berkeley, CA). The methodology employed was previously used
in related studies (MacGregor and Friedman, 1977; Friedman
and Smith, 1984; Friedman et al., l980,1982,1990a,b; MacGre-
gor et al., 1980, 1989) and is described in detail by Maron and
Ames (1983). The DMSO-soluble seed extract (0.1 mL) was added
to culture tubes containing 2.0 mL of top agar and 0.1 mL of
tester strain. If microsomal activation was required, 0.5 mL of
"high" S9 mix was added to the culture tubes after addition of
bacteria. The cultures were then plated. The S9 liver homoge-
nate was prepared from the livers of 200 male Wistar rats induced
with Aroclor 1254 (Simonsen Laboratories, Gilroy, CA). The
tester strains and S9 used in this study were all taken from the
same in-house lot to compare revertant per plate responses within
strains. The extracts were checked for sterility by plating 0.1-
mL high dose samples with 2.0 mL of top agar.
Negative (DMSO) controls were run with each experiment.
The range of revertant colonies per plate (RIP) for three to five
experiments used to generate the figures and tables for each
strain were as follows: TA98 with S9,43 f 5; TA98 without S9,
36 f 3; TAlOO with S9,159 f 17; TAlOO without S9,156 f 18;
TA102 with S9,491 f 49; TA102 without S9,378 f 23; TA2637
with S9,50 * 4; and TA2637 without S9, 26 * 3.
The ranges of RIP responses for the positive controls with S9
were as follows: aflatoxin B1, 0.3 pglplate, TA98,459 f 36; afla-
toxin B1,0.3 pg/plate, TA100,951 f 57; danthron, 45 rg/plate,
TA102,1413 f 171; and emodin, 30 pglplate, TA2637,322 f 25.
Sparse pinpoint colonies of ulawn" growth and the presence of
DMSO-insoluble material were taken as indicators of toxic or
excessive doses, respectively.
Duplicate plates from tester strains dosed with seed extracts
were scored for revertant colonies by hand-counting with a
dissecting microscope for TA98 and TA2637 and by a combination
of hand-counting and Artek Model 980 automated scoring for
TAlOO and TA102. Each individual TAlOO and TA102 exper-
iment was scored by hand-counting four to six plates from the
lowest value to higher values for TAl00 (highest manual count,
1072) and TA102 (highest manual count, 962). The remaining
plates were scored by the Artek Model 980 counter with settings
of 0.2 mm for colony size, 5.3 for area, and 6.0 for sensitivity.
Regression analysis with the "Ames fit" model (see below) was
used to find manual versus Artek slopes in seven individual
experiments with TAlOO and TA102. Slope values for TAlOO
ranged from 1.21 to 1.50. Slope values for TA102 ranged from
2.21 to 3.44. Slope values were accepted for P zero slope of P <
0.05. Artek values times accepted slope value plus intercept was
used to estimate revertants per plate.
Since exogenous histidine has been shown to influence RIP
counts (Maron and Ames, 1983; Friedman et al., 1990a), histi-
dine concentration in the hydrolyzed methanol extracts was
measured by using an amino acid analyzer. The highest value
was 0.0025 mg/mL hydrolyzed sicklepod methanol extract.
J. Agric. Food Chem., Vol. 39, No. 3, 1991 495
Table I. Effects of Hydrolysis on Mutagenic Potency of
Methanol Extracts from Seeds of Morning Glory,
Sicklepod, Velvetleaf, and Jimson Weed Seeds Using TA102
with S9.
% uv
h drolysis 0.001X dilution
recovery after absorbance mutagenic
yield, o f ethanol of 38 Me (.e"%2?
seed mg treatment stock solutiona S9;
unhydrolyzed 1107 (3 0.68' 16
heated 952 108 0.6W 19
hydrolyzed 1095 82 0.43d 1225,1283'
hydrolyzed 1086 67 0.25d 534,536e
unhydrolyzed 437 (3 0.38' 20
heated 365 85 1.45' 49
hydrolyzed 533 478 0.41h 415
hydrolyzed 413 85' 0.72h 625,378'
morning glory
sicklepod
velvetleaf
unhydrolyzed 984 (3 0.31JC 9
heated 883 82 0.31Jh 2
hydrolyzede 941 78 1.14" 432
hydrolyzede 915 80 1.801.k 1160
jimson weed
unhydrolyzed 563 (3 0.45, ns
hydrolyzed 439 86 0.77 8
4 MES = milligram equivalent seeds. * RIMES = revertants per
milligram equivalent seeds. Peak = 332.5 nm. Peak = 286.5 nm.
e Duplicate values. 1 Peak = 280.5 nm. E Started with 15 g of aeeds,
filtered. h Peak = 282 nm. i Started with 6 g of seed, not filtered.
j Peak = 260 nm. Stock solution diluted 0.01X peak = 320 nm.
Addition of this amount of pure histidine did not influence colony
counts in any of the strains tested (data not shown).
RESULTS AND DISCUSSION
Mutagenic Potencies. The Ames fit microcomputer
method of Moore and Felton (1983) was used to evaluate
significant mutagenicity from dose-response curves. Dose
was defined in terms of milligram equivalent seeds (MES)
and the response observed was revertants per plate (R/
P). The slopes calculated from the Ames fit and used for
statistical analysis were RIP per MES or "mutagenic
potency".
Milligram equivalent seeds (MES) relates amounts of
starting seed material to milligram dosage of extract added
to Ames test plate such that MES = W X 1000 X 0.1 mL/
2V, where MES is milligrams equivalent seeds; W is the
weight of starting material (usually 5-15 g); 1000 is
milligrams per gram; 0.1 mL is the volume of extract added
t o the Ames plate; 2 is the factor which accounts for
dividing the sample into two parts for unhydrolyzed and
hydrolyzed treatment; and Vis the reconstituted volume
after evaporation of extract (usually 2-20 mL).
Table I shows the effects of hydrolysis on mutagenic
potencies of four seed extracts, major UVabsorbance peaks
(used as a concentration check in duplicate extracts), and
the calculated mutagenic potencies. Table I1 shows
examples of calculated Ames fit data for quercetin and
extracts of two seeds.
The Ames fit method calculates two probabilities (P).
First, the Pfor linearity of the dose-response curve defines
a linear portion of the curve for P > 0.05. When Plinearity
is calculated to be <0.05, t h e computer program warns
that the "model does not fit". This usually happens when
a toxic dose decreases the revertants per plate (RIP)
response at the peak of the dose-response curve (e.g., Table
11, toxic dose of quercetin, 50 pglplate) or when colony
counts are too high for accurate manual vs Artek Model
980 correlations (e.g., Table 11, sicklepod, 19 MES/plate).
The second probability value signifies a positive, linear,
496 J. Agrlc. Food Chem., Vol. 39, No. 3, 1991
Table 11. Ames Fit Data with Quercetin and Hydrolyzed
Methanolio Extracts from Jimson Weed and Sicklepod in
TA102 with Microsomes.
Friedman and Henlka
no. P
dose: revertants of P (zero slope
sample pg/plate per platec doses (linearity) slope) ( R l p g )
quercetin 0 467,412
0.4 491,451
0.8 504,465
1.6 557,504
3.1 MED 612,551
6.3 hs 852,710
12.5 1041,952
25.0 1378,1160
50.0 t 1319,1182
jimson
weed
sicklepod
0 648,612
0.3d 793,631
0.6d 631,725
1.2d 665,634
2.4d 604,670
4.8d 839,771
9.5d 816,692
38.W 962,157
0 476,557
0.15d 529,488
0.3**d 661,747
0.6d 832,744
2.4d 1193,1106
4.8d 2239,2019
9.5d 2611,3439
19.W 3395,2684
19.0**d 793,809
1.2d 977, c
3
4
5
6
7
8
9
3
4
5
6
7
8
9
3
4
5
6
7
8
9
0.99 0.25 56.3
1 0.0515 55.0
0.98 0.0079' 44.8
0.98 0.0006. 52.5 hs
0.89 0.0001* 45.1
0.34 0' 33.9
0.0019 mnf 0 17.7
0.67 0.81 80.0
0.61 1 -2.1
0.70 0.91 -13.1
0.33 0.0544 25.8
0.35 0.0572 12.6
0.31 0.0137' 7.9''
0.54 0.0062* 5.3
0.10 0.0332* 625''
0.19 0.0055* 506
0.22 0.0022' 416
0.0994 0.0004' 255
0.128 0' 320
0.70 O* 269
0.0134 mnf 0 147
a Abbreviations: R = revertants; MED = minimum effective
dose-see text; MES = milligramequivalent seeds: hs = highest slope
value observed; t = toxic dose; mnf = model does not fit linearity
parameter P < 0.05; ** = minimal effective dose and highest slope
value (mutagenic potency-see text) are identical; c = contaminated
plate; * = accepted slope value for P zero slope C 0.05 with a linear
fit. b MES per plate for jimson weed. Contaminated plate. dMES/
plate.
mutagenic response. We chose P = zero slope < 0.05 as
the cutoff point for comparison of seed extract mutagenic
potencies. The slopes chosen for tabulation were the
highest slope value and dose or mutagenic potency and
the minimal effective dose (MED) or lowest dose observed
for a significant mutagenic response. Quercetin's highest
accepted slope value in TA102 with S9 was found to be
52.5 revertants/pg at 6.3 pg/plate, and its MED was 3.1
pg/plate with a slope of 44.8 revertants/pg (Table 11).
Jimson weed and sicklepod data demonstrate conditions
where the mutagenic potency or the number of revertants
per milligram equivalent seed (RIMES) can be equal to
MED and where one seed extract is much more potent
than another (i.e., Table 11, sicklepod has anR/MES value
of 625 versus jimson weed with a corresponding value of
7.9 in TA102 with S9).
The dose range used to compare the methanol seed
extracts for mutagenic response was 19-0.019 milligram
equivalent seeds (doses reduced by 0.5X intervals). Doses
of 38-150 milligram equivalent seeds for this group of seed
extracts produced revertant per plate values that were
not accepted by the Ames fit model either because colony
counts were too high for accurate manual vs Artek
correlations or the doses were toxic to the bacteria.
The minimal effective dose was also used to estimate
the number of seeds required for a significant mutagenic
response. Thus, if there are 151 seeds per gram of jimson
weed and the percent required for MED is 25 5% (percent
required for MED inTAlO2 with S9 equalsRIME.9 divided
by starting dose of 75 RIMES times 100) then, 0.25 X
2265 (number of seeds in 15 g of starting material) = 566
seeds required for MED. This calculated mutagenic
Table 111. Mutagenic Activity of the Hydrolyzed Methanol
Extracts of Jimson Weed, Morning Glory, Sicklepod, and
Velvetleaf Seeds in the Ames Salmonella/Microsome Test
mutagenic estimated
potentialb % no. of
strain, mutagenic (derived required ~eeds/l6 g
S9 -or from MEB r uired
seed +?TA) %% and MESd) M%S. for%ED*
morning glory
sicklepod
velvetleaf
jimson weed
98, -
98, +-
100, -
100, +
102, -
102, +
2637, -
2637, +-
98, -
98, +
100, -
100, +
102, -
102, +
2637, -
2637, +
98, -
98, +
100, -
100, +
102, -
102, +
2637, -
2637, +
98, -
98, +-
100, -
100, +
102, -
102, +-
2637, -
2637, +-
18
36
72
112
316
1225
11
53
855
358
11
68
174
625
621
200
9
3
22
19
158
432
5
4
3
7
0
2
0
8
0
0
1.2
1.2
1.2
1.2
0.6
0.15
2.4
1.2
0.04
0.08
4.8
1.2
1.2
0.3
0.15
0.04
2.4
4.8
4.8
4.8
1.2
1.2
2.4
9.5
9.5
4.8
nd
38.0
ns
19.0
ns
ns
1.6
1.6
1.6
1.6
0.8
0.2
3.2
1.6
0.1
0.2
12.6
3.2
3.2
0.8
0.4
0.1
6.3
12.6
12.6
12.6
3.2
3.2
6.3
25.0
12.6
6.3
ns
50.0
ns
25.0
ns
ns
I
7
7
7
4
1
14
7
1
2
91
23
23
6
3
1
100
200
200
200
50
50
100
400
285
143
ns
1132
ns
566
ns
ns
RIMES = revertants per milligram equivalent seeds. For def-
inition, see text under Mutagenic Potencies. MED = minimal
effective dose. MES = milligram equivalent seeds. e Based on num-
ber of seeds per gram: 151 for jimson weed, 29 for morning glory, 48
for sicklepod, and 106 for velvetleaf. f ns = not significant.
potential allows for visual field inspection of possible mu-
tagenic seed contamination.
Factors Influencing Mutagenicity. Maron and
Ames (1983) recommend the following bacterial strains
for the detecting of several classes of mutagens: TA98 for
frameshift mutagens; TAlOO for base pair substitutions
at guanine-cytosine (G-C) base pairs; TA102 for frame-
shift mutations at adenine-thymidine (A-T) base pairs;
and TA2637 for frameshift mutagens at a site with several
cytosine residues (similar in specificity to TA97). We used
TA2637 instead of TA97 because we were unable to observe
reliable spontaneous reversion with TA97 strain. Previous
work with naturally occurring flavonoids, flavones, and
anthraquinones has shown that TA98 is most sensitive for
detecting mutagenic flavonoids (MacGregor 1984), TAlOO
is most sensitive for detecting mutagenic flavones (Elliger
et al., 1953), and TA2637 is most sensitive for detecting
mutagenic anthraquinones (Tikkanen et ai., 1983).
Our experimental system for mutagenicity studies of
weed seed extracts generated a series of chemically
undefined products. Previous studies and our present
results (Tables 1-111 and Figures 1-4) collectively define
major factors that can affect mixtures and the interpre-
tation of mutagenicity data in the Ames test. These factors
include the following: (1) mutagen glycosylation [e.g., hy-
drolysis of nonmutagenic rutin produces a mutagen: quer-
cetin (Figure 5a; Brown, 198O)l; (2) mutagen toxicity to
5'. typhimurium (e.g., emodin is toxic at doses above
approximately 30 pg/plate; Tikkanen et al., 1983); (3)
Mutagenicity of Toxlc Weed Seeds
900
800
700
600
500
400
300
200
100
75
sa
TA 98
I -59 + 59 -59
IIMSONWEED MORNING SICKLESPOD VELVETLEAF
GLORY
Figure 1. Mutagenic potencies of methanolic extracts of four
toxic weed seeds in revertants per milligram equivalent seed (R/
MES) in strain TA98, with and without microsomal activation.
structural requirements in a related series (e.g., hydroxy-
anthraquinones; Westendorf et al., 1990); (4) the presence
of antimutagens or S9 inhibitors (see discussion on sick-
lepod); (5) the presence of more than one mutagen (see
discussion on sicklepod) and (6) the possible contamination
of seeds with exogenous mutagens, e.g., mycotoxin (Fried-
man et al., 1982), herbicides sprayed on seeds, and mu-
tagens trapped in dirt or dust. Minor factors that were
checked include the (1) presence of histidine, (2) geno-
toxic artifacts generated by extraction, and (3) mutagens
generated by heat (Friedman et al., 1990b).
This combination of conditions and factors creates a
large number of experiments. Rangefinder studies were
used to determine which conditions significantly influence
mutagenic potency and minimal effective dose.
First, methanol extraction accounted for nearly all mu-
tagenic potency in all four seed sources. For example,
with TA102 plus S9, hydrolyzed methanol extracts as
compared to hexane, chloroform, or water extracts pro-
duced 98% of total activity for sicklepod, morning glory,
and velvetleaf and 61 % for jimson weed (data not shown).
Strain TA102 was selected for this comparison because it
gave positive results with all four seeds and its mutagenic
response was the least variable (Figures 1-4). Second,
using TA102 plus S9, we found that hydrolysis increased
mutagenic potency in the methanol extracts of morning
glory (58X), sicklepod (48X), velvetleaf (14X), and jimson
weed (approximately 4X-estimated because unhydro-
lyzed activity was not significant) (Tables I and 111). Third,
the increases were found not to be due to heat or
evaporation; Le., heating unhydrolyzed methanol extracts
J. Agric. Food Chem., Vol. 39, No. 3, 1991 497
TA 100
+59 -59
MORNING
GLORY
+59 -59
SICKLESPOD
+59 -59
ELVETLEAF
Figure 2. Mutagenic potencies of methanolic extracts of four
toxic weed seeds in revertanta per milligram equivalent seed (R/
MES) in strain TA100, with and without microsomal activation.
with evaporation temperatures did not produce similar
increases in mutagenic potency. Finally, the presence of
S9 in the Ames test reaction mixture either increased or
decreased mutagenic potency.
We expected variability in mutagenic potencies from
different seed lots due to potent variation in mutagen
concentration or composition. We therefore carried out
our experiments with a single lot of each seed source.
Extraction and mutagenic data for each seed source and
their interpretation are elaborated below.
Jimson Weed (D. stramonium) Seeds. Of the four
seed extracts tested for mutagenicity, jimson weed was
found to have the least amount of mutagenic potency and
potential. Activities from the methanol extracts ranged
from not significant (i.e., TA2637 with and without S9,
Table 111) to a comparatively weak response to TA98 with
S9 of 7 revertants/milligram equivalent seeds (Table 111).
This activity was in contrast to, for example, sicklepod,
which, in TA98 with S9, generated 358 RIMES. Table I11
also shows estimates of the number of seeds required per
15 g of starting material for a minimal effective dose
(MED). In TA98 with S9, methanol extracts of jimson
weed required an estimated 143 seeds for MED, whereas
sicklepod only required an estimated 2 seeds. A visual
scan of histograms in Figures 1-4 also demonstrates that
jimson weed is comparatively inactive compared to other
seeds. This inactivity also suggests that genotoxic artifacts
are not formed during the extraction procedure.
Jimsonweed seeds containing the toxic tropane alkaloids
atropine and scopolamine occasionally contaminate com-
mercial grain (Friedman and Levin, 1989). A recent 90-
490 J. Agric. FoodChem., Vol. 39, No, 3, 1991 Friedman and Henika
1225
1125
1025
92:
825
725
625
52s
425
325
225
125
100
75
50
25
0
TA 102
+59 -59 +59 .59 +59 -59 +59 - 5 9
JIMSONWEED MORNING SICKLESPOD VELVETLWF
GLORY
Figure 3. Mutagenic potencies of methanolic extracts of four
toxic weed seeds in revertants per milligram equivalent seed (R/
MES) in strain TA102, with and without microsomal activation.
day feeding study in rats showed that jimson weed seeds
produce adverse physiological effects, i.e., weight loss and
changes in plasma enzymes, but no observable clastogenic
(chromosome-damaging) activity in the bone marrow mi-
cronucleus test (Dugan et al., 1989). Negative micronu-
cleus test results have also been observed with scopolamine
given intraperitoneally to mice (data not shown). Atropine
and scopolamine were also negative in the Ames test in
TA98 (Glatt et al., 1983; McCann et al., 1975; unpublished
results). Atropine failed to induce @-galactosidase gene
expression in the umu gene test system using S. typh-
imurium strain TA1535/pSK 1002 (Nakamura et al.,
1987). Jin-fu et al. (1988), however, demonstrated a small
but significant increase in chromosome damage by sco-
polamine in human lymphocyte cell cultures. This effect
could arise from alkylation of DNA or other sensitive site
by the epoxide group of scopolamine, as illustrated in
Figure 5b. Our results suggest that jimson weed seeds are
toxic due to the presence of tropane alkaloids, whose ste-
reochemistry is still being elucidated (Schmidt and Honig-
berg, 1989), but have negligible genotoxicity in bacteria
and in vivo in bone marrow cells.
Velvetleaf (A. theOPhr8Stfi Seeds. The methanol
extracts of velvetleaf seeds were most active in strain
TA102, producing 423 RIMES with a minimal effective
dose (MED) of 1.2 MES and a requirement of 50 seeds/l5
g of starting material for MED (Table 111). These activities
required hydrolysis (e.g., a 40X increase vs unhydrolyzed
extracts, Table I) but produced variable responses with
S9. For example, S9 increased mutagenic potency 2X that
in TA102, did not influence activity in TAlOO or TA2637,
700
600
500
400
300
200
100
0
TA 2637
+s9 - 5 9 +59 -59 +59 - 5 9 +59 - 5 9
JIMSONWEED MORNING SICKLESPOD VELVETLEAF
GLORY
Figure 4. Mutagenic potencies of methanolic extracts of four
toxic weed seeds in revertants per milligram equivalent seed (R/
MES) in strain TA2637, with and without microsomal activation.
and decreased activity 3 X in TA98 (Figures 1-4). The
magnitude of differences in these responses indicates that
S9 had little influence on the active component(s) in vel-
vetleaf seeds.
Hydrolysis of the methanol extract produced DMSO-
insoluble material which was separated by filtration. The
reason for the precipitation is not known. Unhydrolyzed
and hydrolyzed extracts produced different UV absor-
bance spectra, indicating that chemical changes occurred
during hydrolysis (Table I).
Plant phenolics with alleopathic and mutagenic poten-
tial are reported to occur in the coats of velvetleaf seeds
(Paszkowski and Kremer, 1988). The components (in
relative order of concentration) were delphinidin = quer-
cetin > (+)-catechin = myricetin > (-)-epicatechin and
cyanidin. The condensed tannin, delphinidin, is negative
in the Ames assay but positive in vitro for micronuclei
clastogenicity in V79 cells (Ferguson et al., 1985). Cya-
nidin, (+)-catechin, and (-)-epicathechin are negative in
the Ames test. However, quercetin and myricetin are
strong mutagens in TA98 (MacGregor, 1984; Friedman
and Smith, 1984).
Of the components reported to occur in velvetleaf seed
methanol extracts, quercetin appears to be the most likely
candidate for an active mutagenic compound. If this is
true, our data imply that TA102 with an MED of 1.2 MES
should appear more sensitive for detecting quercetin than
TA98, with an MED of 4.8 MES (Table 111). We tested
this hypothesis by determining the following MED values
in micrograms per plate for pure quercetin with S9: TA98,
1.6; TA100,3.1; TA102,3.1; and TA2637,3.1. Thus, even
Mutagenicity of Toxic Weed Seeds J. Agk. FoodChm., Vol. 39, No. 3, 1991 499
Hydrolysis HO &OH (4
HO &OH
OH 0-rutinose
I I
O H 0 OH 0
QUERCETIN RUTIN
CHLOROGENIC ACID CAFFEIC ACID OUlNlC ACID
T f Cvtochrome f&u
"a - I + DNA*NH2
0
DNA ADDUCT
EMODIN
THIOL ADDUCT
Figure 5. Possible mechanisms of genotoxicity of some compounds in toxic weed seeds. (a) Hydrolysis of the nonmutagenic glycoside
rutin to the mutagenic aglycon quercetin. (b) Postulated alkylation of an amino group of DNA by an epoxide group of scopolamine
to form a DNA adduct. The epoxide group can be opened at either side of the ring to form two stereoisomers. (c) Hydrolysis of
nonmutagenic but genotoxic chlorogenic acid to caffeic and quinic acids. (d) Postulated oxidation of the methyl group of emodin by
cytochrome P-450 to a reactive, resonance-stabilized carbonium ion intermediate which then combines with an amino group of DNA
to form a DNA adduct. (e) Postulated transformation of biologically active emodin to an inactive thiol adduct.
though TA102 was the most sensitive strain for velvetleaf
seeds (Table 111), TA98 remains the most sensitive strain
for the detection of pure quercetin mutagenicity with a
MED of 1.6 Mg/plate. However, our data do not rule out
quercetin as a contributor (either by itself or by rutin hy-
drolysis; Figure 5a) to velvetleaf seed mutagenicity.
Furthermore, some plant phenolics such as quercetin have
been suspected as carcinogens. However, in vivo results
are contradictory (MacGregor, 1984). Other phenolic
compounds are reported to have antimutagenic properties
(Smith and Rosin, 1984; Shinohara et al., 1988).
Morning Glory (Ipomoea spp.) Seeds. Of the four
seed extracts tested for mutagenic potency, we considered
morning glory methanol extracts most potent because they
were positive in all four strains and required the fewest
average number of seeds for mutagenic potency (calculated
from Table 111). The seeds' highest mutagenic potency
was found in TA102 with S9, where an estimated 1 seed/
15 g of starting material was required for a MED of 0.15
MES. All strains required hydrolysis (Table I) and S9 for
maximal mutagenic potency (Figures 1-4).
Chlorogenic acid, a major component of morning glory
seeds, has a UV absorbance maximum of 328 nm in eth-
anol (Friedman et al., 1989). In our DMSO-soluble
extracts, we observed a peak at 332.5 nm in unhydrolyzed
extracts which was lost upon hydrolysis, possibly indicating
hydrolysis of chlorogenic acid to caffeic and quinic acids
(Figure 5c).
In a rangefinder experiment with doses of up to 75 MES
given intraperitoneally, we found weight loss but no
evidence for micronuclei formation in mice (data not
shown). The genotoxicity data of the major component
500 J. Agrlc. Food Chem., Vol. 39, No. 3, 1991
of morning glory seeds (chlorogenic acid) did not correlate
well with our findings. Chlorogenic acid and its hydrol-
ysis products (caffeic and quinic acids) were found to be
negative for the Ames test but did produce chromosome
aberration in Chinese hamster ovary (CHO) cells (Fung
et al., 1988; Stich et al., 1981). This rules out these plant
phenolics as the mutagen principles in our study. Dietary
flavonoids such as quercetin and chlorogenic acid may
have the potential of producing intestinal injury (Canada
et al., 1989). It is also noteworthy that chlorogenic acid
is rapidly destroyed under the influence of food-processing
conditions such as autoclaving and conventional and
microwave baking (Friedman and Dao, 1990).
A second group of chemicals, i.e., the ergot alkaloids,
that occur in seeds of the Heavenly Blue variety of morning
glory and which are much more heat-resistant than chlo-
rogenic acid (Friedman and Dao, 1990) have apparently
not been tested for mutagenicity in the Ames test. Thus,
we cannot assign a possible genotoxic principle that may
be responsible for the high mutagenicity we find for
morning glory seed extracts.
Sicklepod (C. obtusifolia) Seeds. Sicklepod seed
extracts are reported to be toxic to muscle tissues (Lewis
and Shibamoto, 1989; Putnam et al., 1988) and to possess
mutagenic anthraquinones such as emodin, chrysophanic
acid, and physcion (Crawford et al., 1990). In this study,
sicklepod methanol extracts were highly mutagenic, with
the most sensitive strains being TA98 without S9 (MED
= 0.04 MES) and TA2637 without S9 (MED = 0.15 MES).
These activities require only an estimated 1-3 seeds/l5
g of starting material (Table 111). The S9 (microsome)
requirements were ambiguous with TA98 and TA2637
activity inhibited byS9 (2.5X in TA98 and 3.1X inTA2637)
(Figures 1 and 2). Sicklepod methanol extracts also
required hydrolysis for maximal activity (a 15X increase
in activity vs unhydrolyzed, Table I). During extraction
with 15 g of seeds, a black residue formed on the boiling
flask. The filtered sample gave identical UV absorbance
maxima as a duplicate sample with starting material but
with no filtration. Filtered and unfiltered samples had
similar mutagenic potencies, indicating that the muta-
genic component did not precipitate out. Finally, UV ab-
sorbance profiles of unhydrolyzed and hydrolyzed extracts
did change, indicating chemical changes during hydrol-
ysis.
In the Ames test, anthraquinones are positive in strains
TA2637 and TAlOO (TA100 is less sensitive than TA2637)
and negative in TA98 and TA102 (Bachmann et al., 1979;
Tikkanen et al., 1983). The extracts were strongly positive
in TA98 and TA102, indicating that components other
than, or in addition to, anthraquinones may be responsible
for mutagenic potency. The decrease in mutagenic potency
in TA98 and TA2637 implies inhibition of microsomal
fraction by an antimutagen. In studies with rhubarb
extracts, Van der Hoeven et al. (1983) were able to assign
emodin as the active component in strain TA1537, a non-
plasmid parent strain of TA2637 (Maron and Ames, 1983).
However, we were not able to give a similar assignment
due to the complex nature of sicklepod extract mutage-
nicity data. Figure 5d illustrates a possible mechanism
for the mutagenicity of emodin (Tanaka et al., 1987) and
Figure 5e the trapping of the postulated electrophilic
intermediate by a nucleophilic thiol, thus preventing DNA
adduct formation (Friedman, 1984).
Our data and related studies suggest that exposure to
sicklepod seeds carries with it a risk for genotoxicity as
well as for myotoxicity. By themselves, mutagenic an-
thraquinones are not thought to pose a high cancer risk
Friedman and Henika
because of their inability to covalently bind Salmonella
and rat liver DNA (Bosch et al., 1987). However, further
compositional analyses of sicklepod seed extracts are
needed to provide a chemical basis for risk assessment
(Crawford and Friedman, 1990; Crawford et al., 1990).
CONCLUSIONS
The observed wide variability in the genotoxicity of
extracts of different toxic weed seeds should, together with
additional compositional-pharmacological-toxicological
data, be taken into account in setting maximum tolerance
levels for the contamination of commercial grain and other
foods and feeds with these seeds. The derived quantitative
parameters, milligram equivalent seeds (MES) and min-
imum effective dose (MED), should facilitate comparison
of relative genotoxicities of various grain samples. These
parameters merit adoption by other investigators so that
the relative potency index can be extended to many other
naturally occurring and processinginduced food ingre-
dients. Such an index would help relate consumption of
foods and feeds containing toxic weed seed to the safety
and health of animals and humans. A better understand-
ing of the chemical bases and molecular mechanisms of
toxic, mutagenic, and antimutagenic responses is also
needed to facilitate predicting genotoxicities and relative
risks to animal and human health from compositional data
of complex foods, feeds, toxic weeds, and other materials
(Deshpande et al., 1984; Bradfield and Bjeldanes, 1991;
De Flora et al., 1989; Loprieno et al., 1991; MacGregor,
1984; Friedman, 1980; Weisburger, 1991).
ACKNOWLEDGMENT
J. S. Felton for helpful comments.
LITERATURE CITED
Bachmann, M.; Luthy, L.; Schlatter, C. Toxicity and mutage-
nicity of molds of Apergillus glaucus group. Identification of
physcion and three related anthraquinones as main toxic
constituents from Aspergillus cheualieri. J. Agric. Food Chem.
1979,27, 1342-1347.
Bosch, R.; Friedrich, U.; Lutz, K.; Brocker, E.; Bachmann, M.;
Schlatter, C. Investigations on DNA binding in rat liver and
in Salmonella and on mutagenicity in the Ames test by emo-
din, a natural anthraquinone. Mutat. Res. 1987, 188, 161-
168.
Bradfield, C. A.; Bjeldanes, L. F. Modification of carcinogen
metabolism by indolylic autolysis products of Brassica ole-
racae. In Nutritional and Toxicological ComequencesofFood
Processing; Friedman, M., Ed.; Plenum: New York, 1991;
Chapter 13.
Brown, J. P. A review of the genetic effects of naturally occurring
flavonoids, anthraquinones and related compounds. Mutat.
Res. 1980, 75, 243-277.
Brown, J. P.; Dietrich, P. S. Mutagenicity of plant flavonols in
the Salmonella/mammalian microsome test. Activation of
flavonol glycosides by mixed glycosidases from rat cecal
bacteria and other sources. Mutat. Res. 1979, 66, 223-240.
Canada, A. T.; Watkins, W. D.; Nguyen, T. D. The toxicity of
flavonoids to guinea pig erythrocytes. Toxicol. Appl. Phar-
macol. 1989,99, 357-361.
Crawford, L.; Friedman, M. Effects of low levels of dietary toxic
weed seeds on the relative size of rat liver and levels and
function of cytochrome P-450. Toxicol. Lett. 1990,54, 175-
181.
Crawford, L.; McDonald, G. M.; Friedman, M. Composition of
sicklepod (Cassia obtusifolia) toxic weed seeds. J. Agric. Food
Chem. 1990,38,2169-2175.
De Flora, S.; Benicelli, D. S.; Izzotti, A.; Cesarone, C. F. Role of
glutathione and N-acetylcysteine as inhibitors of mutagenesis
and carcinogenesis. In Absorption and Utilization of Amino
We thank Paul V. Allen for graphics display of data and
Mutagenicity of Toxlc Weed Seeds
Acids; Friedman, M., Ed.; CRC: Boca Raton, FL, 1989; Vol.
3, Chapter 3.
Deshpande, S. S.; Sathe, S. K.; Salunkhe, D. K. Chemistry and
safety of plant polyphenols. In Nutritional and Toxicological
Aspects of Food Safety; Friedman, M., Ed.; Plenum: New
York. Adv. Exp. Med. Biol. 1984,177, 457-495.
Dugan, G. M.; Gumbmann, M. R.; Friedman, M. Toxicology of
Jimsonweed (Datura stramonium) seeds. Food Chem. Tox-
icol. 1989, 27, 501-510.
Elliger, C. A.; Henika, P. R.; MacGregor, J. T. Mutagenicity of
flavones, chromones, and acetophenones in Salmonella typh-
imurium. Mutat. Res. 1984, 135, 77-86.
Ferguson, L. R.; van Zijl, P.; Holloway, W. D.; Jones, W. T.
Condensed tannins and induced micronuclei in cultured V79
Chinese hamster cells. Mutat. Res. 1985, 158, 89-95.
Friedman, M. Chemical basis for biological effects of wool
finishing treatments. In Proceedings of the Sixth Znterna-
tional Wool Textile Research Conference; Pretoria: South
Africa, 1980; Vol. V, pp 323-348.
Friedman, M. Sulfhydryl groups and food safety. In Nutritional
and Toxicological Aspects ofFood Safety; Friedman, M., Ed.;
Plenum: New York. Adv. Exp. Med. Biol. 1984,177,31-63.
Friedman, M.; Dao, L. Effect of autoclaving and conventional
and microwave baking on the ergot alkaloid and chlorogenic
acid contents of morning glory (Ipomoea tricolor Cav. cv.)
Heavenly Blue seeds. J. Agric. Food Chem. 1990,38,805-808.
Friedman, M.; Levin, C. E. Composition of jimsonweed (Datura
stramonium) seeds. J. Agric. Food Chem. 1989,37,998-1005.
Friedman, M.; Smith, G. A. Inactivation of quercetin mutage-
nicity. Food Chem. Toxicol. 1984, 22, 535-539.
Friedman, M.; Diamond, M. J.; MacGregor, J. T. Mutagenicity
of textile dyes. Environ. Sci. Technol. 1980, 14, 1145-1146.
Friedman, M.; Wehr, C. M.; Schade, J. E.; MacGregor, J. T.
Inactivation of aflatoxin B1 mutagenicity by thiols. Food
Chem. Toxicol. 1982,20, 887-892.
Friedman, M.; Dao, L.; Gumbmann, M. R. Ergot alkaloid and
chlorogenic acid content of morning glory (Ipomoea spp.) seeds.
J. Agric. Food Chem. 1989,37, 708-712.
Friedman, M.; Wilson, R. E.; Ziderman, I. I. Effect of heating on
mutagenicity of fruit juices in the Ames test. J. Agric. Food
Chem. 1990a, 38,740-743.
Friedman, M.; Wilson, R. E.; Ziderman, I. I. Mutagen formation
in heated wheat gluten, carbohydrates, and gluten-carbohy-
drate blends. J. Agric. Food Chem. 1990b, 38, 1019-1028.
Fung, V. A.; Cameron, T. P.; Hughes, J. J.; Kirby, P. E.; Dunkel,
V. C. Mutagenic activity of some coffee flavor ingredients.
Mutat. Res. 1988,204, 219-228.
Glatt, H.; Jung, R.; Oesch, F. Bacterialmutagenicity investigation
of epoxides: drugs, drug metabolites, steroids, and pesticides.
Mutat. Res. 1983, 1 1 1 , 99-118.
Jin-fu, Y.; Yi-shou, Y.; Wei-yu, W.; Gui-xian, X.; Ming-sheng, C.
Mutagenicity and teratogenicity of chloropromazine and sco-
polamine. Chin. Med. J. 1988, 101, 339-345.
Loprieno, N.; Boncristiani, G.; Loprieno, G. An experimental
approach to identifying the genotoxic risk by cooked meat
mutagens. In Nutritional and Toxicological Consequences
of Food Processing; Friedman, M., Ed.; Plenum: New York,
1991; Chapter 9.
Lewis, D. C.; Shibamoto, T. Effects of Cassia obtusifolia (sick-
lepod) extracts and anthraquinones on muscle mitochondrial
function. Toxicon 1989, 27, 519-529.
MacGregor, J. T. Genetic and carcinogenic effects of plant fla-
vonoids: an overview. In Nutritional and Toxicological
Aspects of Food Safety; Friedman, M., Ed.; Plenum: New
York. Adv. Exp. Med. Biol. 1984, 177, 497-526.
MacGregor, J. T.; Friedman, M. Nonmutagenicity of tetrabro-
mophthalic anhydride and tetrabromophthalic acid in the
J. Agrlc. FoodChem., Vol. 39, No. 3, 1991 501
Ames Salmonella microsome mutagenicity assay. Mutat. Res.
MacGregor, J. T.; Diamond, M. J.; Mazzeno, L. W.; Friedman,
M. Mutagenicity testa for fabric finishing agents in Salmonella/
Typhimurium: fiber-reactive dyes and cotton flame retar-
dants. Enuiron. Mutagen. 1980, 2, 405-418.
MacGregor, J. T.; Tucker, J. D.; Ziderman, I. I.; Wehr, C. M.;
Wilson, R. E.; Friedman, M. Nonclastogenicity in mouse bone
marrow of fructose/lysine and other sugar/amino acid brown-
ing products with in vitro genotoxicity. Food Chem. Toxicol.
Maron, D. M.; Ames, B. N. Revised methods for the Salmonella
mutagenicity test. Mutat. Res. 1983, 113, 173-215.
McCann, J.; Choi, E.; Yamasaki, E.; Ames, B. N. Detection of
carcinogens and mutagens in the Salmonella/microsome test:
assay of 300 chemicals. Proc. Natl. Acad. Sei. U.S.A. 1975,
Moore, D.; Felton, J. S. A microcomputer program for analyzing
Ames test data. Mutat. Res. 1983, 119, 95-102.
Nakamura, S.; Oda, Y.; Shimada, T.; Oki, I.; Sugimoto, K. SOS-
inducing activity of chemical carcinogens and mutagens in
Salmonella typhimurium TA1535/pSK 1002: examination
of 151 chemicals. Mutat. Res. 1987, 192, 239-246.
Paszkowski, W.; Kremer, R. J. Biological activity and tentative
identification of flavonoid components in velvetleaf (Abuti-
lon theophrasti Medik.) seed coats. J. Chem. Ecol. 1988,14,
Putnam, M. R.; Boosinger, T.; Spano, J.; Wright, J.; Wiggins, A.;
Dâ??hdrea, G. Evaluation of Cassia obtusifolia (sicklepod) seed
consumption in Holstein calves. Vet. Hum. Toxicol. 1988,30,
3 16-3 18.
Schmidt, W. F.; Honigberg, I. L. Textbook correction: confor-
mation of hyoscyamine. Am. J. Pharm. Educ. 1989,53, 48-
49.
Shinohara, K.; Kuroki, S.; Miwa, M.; Kong, Z. L.; Hosada, H.
Antimutagenicity of dialyzates of vegetables and fruits. Ag-
ric. Biol. Chem. 1988, 52, 1369-1375.
Stich, H. F.; Rosin, M. P. Naturally occurring phenolics as an-
timutagenic and anticarcinogenic agents. In Nutritional and
Toxicological Aspects of Food Safety; Friedman, M., Ed.;
Plenum: New York. Adu. Exp. Med. Biol. 1984, 177, 1-30.
Stich, H. F.;Rosin, M. P.; Wu, C. H.;Powrie, W. D. Acomparative
genotoxicity study of chlorogenic acid (3-0-caffeoylquinic acid).
Mutat. Res. 1981, 90, 201-212.
Tanaka, H.; Morroka, N.; Haraikawa, K.; Ueno, Y. Metabolic
activation of emodin in the reconstituted cytochrome P-450
system of the hepatic microsomes of rata. Mutat. Res. 1987,
Tikkanen, L.; Matsushima, T.; Natori, S. Mutagenicity of an-
thraquinones in the Salmonella preincubation test. Mutat.
Res. 1983, 116, 297-304.
Van der Hoeven, J. C.; Lagerwij, W. J.; Bruggeman, I. M.; Vor-
agen, F. J.; Koeman, J. Mutagenicity of extracts of some
vegetables commonly consumed in the Netherlands. J . Agric.
Food Chem. 1983,31, 1020-1026.
Weisburger, J. J. Carcinogens in our food and cancer prevention.
In Nutritional and Toxicological Consequences of Food
Processing; Friedman, M., Ed.; Plenum: New York, 1991;
Chapter 12.
Westendorf, J.; Marquardt, H.; Poginsky, B.; Dominiak, M.;
Schmidt, J. Genotoxicity of naturally occurring hydroan-
thraquinones. Mutat. Res. 1990,240, 1-12.
1977,56,81-84.
1989,27, 715-721.
72,5135-5139.
1573-1582.
176, 165-170.
Received for review May 2, 1990. Revised manuscript received
September 24, 1990. Accepted October 24, 1990.

[MiddleburgsBum]

wow, whats all that crap ^^^^

[jadeius]

REALLY THOUGH, i don't think anyone wants to read all that....

anyways, i had some old antispasmodic elixir that contained full belladonna extracts and phenobarbital...it's rather weird stuff...the phenobarbital relaxes you WAY OUT...but the belladonna extracts make it trippier, i still don't recommend it though, because i think i've successfully linked that stuff to unbearable stomach pains....i know for sure it causes a very unpleasant dry mouth, much worse than weed would ever do...and i just kept feeling like i was on the verge of tripping, never could get anything real good out of it...but then i didn't want to take a large dose either, because as dick says, the lethal dosage of the stuff is variable and really close to a get high doseage...not worth it, not at all

[MiddleburgsBum]

well, I'll see

[nakedgunner]

wow that just took me 2 hours to read and i dont understand it

[NightProwler]

holy guacemolie that was a big post

[andruejaysin]

Took me 2 minutes just to page through it, anyone actually read the thing?