Potent Bacterial Mutagens Produced by Chlorination of Simulated Poultry Chiller Water

Article abstract of DOI:10.1021/jf950076s

Hypochlorite treatment of a simulated food-processing mixture produces 3,4-dichloromaleimide and 3,3-dichloro-4-(dichloromethylene)-2,5-pyrrolidinedione (C₅HCl₄NO₂). The tetrachloro compound and two analogs, which can be synthesized from citraconic anhydride and itaconic anhydride, are direct-acting Ames mutagens in Salmonella typhimurium TA100 tester strain. These novel five-carbon cyclic imides are structurally similar to 3-chloro-4-(dichloromethyl)-5-hydroxy-2(5H)-furanone (MX), the principal mutagenic compound present in paper pulp bleaching liquors. Molecular structure analysis of the mutagens was based on X-ray crystallography, ¹³C NMR, and mass spectrometry of synthetic chlorinated imides with identical mass spectra and gas chromatographic retention indices. The tetrachloroimide accounts for much of the mutagenicity of the dichloromethane-extractable pH 2 fraction from chlorination of a simulated food-processing system consisting of chicken frankfurters. In the Ames TA100 tester strain it has a molecular mutagenicity of 1450 revertants/ nmol without microsomal activation, making it the second most potent mutagen reported from a chlorination process.

Full text of DOI:10.1021/jf950076s

256
J. Agric. Food Chem. 1996, 44, 256−263
P oten t Ba cter ia l Mu ta gen s P r od u ced by Ch lor in a tion of Sim u la ted
P ou ltr y Ch iller Wa ter
William F. Haddon,* Ronald G. Binder, Rosalind Y. Wong, Leslie A. Harden, Robert E. Wilson,
Mabry Benson, and Kenneth L. Stevens
Western Regional Research Center, Agricultural Research Service, U.S. Department of Agriculture,
800 Buchanan Street, Albany, California 94710
Hypochlorite treatment of a simulated food-processing mixture produces 3,4-dichloromaleimide and
3,3-dichloro-4-(dichloromethylene)-2,5-pyrrolidinedione (C5HCl4NO2). The tetrachloro compound and
two analogs, which can be synthesized from citraconic anhydride and itaconic anhydride, are direct-
acting Ames mutagens in Salmonella typhimurium TA100 tester strain. These novel five-carbon
cyclic imides are structurally similar to 3-chloro-4-(dichloromethyl)-5-hydroxy-2(5H)-furanone (MX),
the principal mutagenic compound present in paper pulp bleaching liquors. Molecular structure
analysis of the mutagens was based on X-ray crystallography, 13C NMR, and mass spectrometry of
synthetic chlorinated imides with identical mass spectra and gas chromatographic retention indices.
The tetrachloroimide accounts for much of the mutagenicity of the dichloromethane-extractable
pH 2 fraction from chlorination of a simulated food-processing system consisting of chicken
frankfurters. In the Ames TA100 tester strain it has a molecular mutagenicity of 1450 revertants/
nmol without microsomal activation, making it the second most potent mutagen reported from a
chlorination process.
Keyw or d s: Poultry chiller water; bacterial mutagens; chlorine disinfection byproducts; chlorinated
imides
INTRODUCTION
In poultry processing, carcass temperatures are re-
duced to a mandated value of 40 °F in a chiller bath to
which chlorine is added as a disinfectant. This paper
reports the results of a study to identify, at elevated
chlorination levels and in a simulated processing mix-
ture, compounds that may contribute to the observed
bacterial mutagenicity of chlorinated poultry chiller
water.
The chlorinated organic compounds that are produced
during the chlorination of drinking water have been
extensively studied for over 15 years, and hundreds of
disinfection byproducts of chlorination have been identi-
fied. From a toxicology standpoint, one of the more
interesting aspects of these studies was the surprising
discovery that a single compound, 3-chloro-4-(dichlo-
romethyl)-5-hydroxy-2(5H)-furanone (MX, 5) can ac-
count for 30-50% of the bacterial mutagenicity of
drinking water, as measured in the Ames/Salmonella
assay, when chlorination is carried out in the presence
of humic material (Holmbom et al., 1981; Maier et al.,
1987; Holmbom, 1990). In food processing, chlorination
is also widely employed. Chlorination has special
importance as a food safety measure in the processing
of the several billion chickens produced yearly in the
United States. The chlorination of poultry chiller water
is especially effective in minimizing the levels of Sal-
monella typhimurium, Campylobacter jejunei, and other
bacterial pathogens (National Research Council, 1988).
Previous studies have established a dose-dependent
Ames-mutagenic response for chlorination of poultry
chiller water in several bacterial test strains (Masri,
1986). Although the levels of mutagenic activity are
very low under actual plant conditions, identification
of the specific bacterial mutagens in poultry chiller
water would be desirable, both to aid in the risk
assessment of the chlorination process and to evaluate
more specifically the effect of processing variables,
including the extent of water reuse. Alternatives to
chlorine disinfection procedures for poultry have been
described (Kim et al., 1994), but use of Cl2, where the
active chlorinating agent is hypochlorous acid, is cur-
rently the most widely used method.
MATERIALS AND METHODS
Ch lor in a tion . Two types of chlorinated samples were
investigated. Initially, we chlorinated water suspensions of
homogenized chicken frankfurters because this mixture could
be prepared easily and reproducibly. Once the mutagenic
compounds had been identified, their presence was confirmed
in chlorinated homogenates prepared from white and dark
meats and skin of whole chicken. In both cases, artificially
high levels of chlorine were employed to enhance the produc-
tion of mutagens.
Stock hypochlorite solutions with about 4% available chlo-
rine as hypochlorous acid were prepared by bubbling ultrapure
Cl2 (g) into 1.2 N NaOH until a rapid drop to pH 11 occurred.
The solutions were stored until needed, at which time chlorine
concentrations (free, combined available and total available
chlorine) were determined by titration (Schade et al., 1990).
Chlorination of Chicken Frankfurter Homogenates. Com-
mercial chicken frankfurters containing approximately 50%
solids were purchased locally and frozen until used. Two
frankfurters were slightly thawed and diced into approxi-
mately 1 cm cubes. The diced cubes and an equal weight of
high-purity water (prepared with Milli-Q UV-Plus, Millipore
S.A., Molsheim, France) were homogenized in a Model 17105
Omni mixer (OCI Instruments, Waterbury, CT) to yield a
suspension consisting of about 25% solids. Chlorine stock
solution was added to this suspension until the amount of
chlorine was equal, in weight, to the frankfurter solids. The
pH was then adjusted to 7.2. The mixture was allowed to
stand overnight at room temperature with gentle shaking;
after 12 h, the homogenate was filtered and adjusted to pH
11.
* Author to whom correspondence should be ad-
dressed (e-mail haddon@pw.usda.gov).
This article not subject to U.S. Copyright. Published 1996 by the American Chemical Society

Chlorination of Simulated Poultry Chiller Water
J. Agric. Food Chem., Vol. 44, No. 1, 1996 257
the beginning of the separation. This early eluting material
has not yet been characterized.
Reversed Phase HPLC. Fraction S7 from silica chromatog-
raphy was chromatographed on an HPLC column (Zorbax, C8,
250 mm × 21.2 mm) with 20 mL/min of H2O/CH3CN in a linear
gradient from 70:30 to 10:90, over 20 min, and held at the final
concentration for 10 min. The active fraction (7-9 min) was
diluted with an equal volume of water and extracted seven
times with DCM. After the extract was concentrated, the
material was again chromatographed on the same column at
16 mL/min with a linear gradient of H2O/CH3CN from 80:20
to 10:90, over 30 min, and held at the final concentration for
10 min. The active fraction, designated RP2, eluted between
11 and 13 min and was again diluted with an equal volume of
water and extracted with 7 portions of DCM. The extract was
concentrated by removing the DCM with a stream of dry
nitrogen, redissolved in 150 µL of DCM, and analyzed by
combined gas chromatography-mass spectrometry (GC-MS).
Normal Phase HPLC. Fraction RP2 was chromatographed
on a normal phase column (silica, 4.5 mm × 250 mm) with 1
mL/min of hexane/isopropyl alcohol in a linear gradient from
98:2 to 94:6, over 12 min, and held at the final concentration
for 3 min. Each solvent reservoir contained 0.5% acetic acid.
The active fraction NP1 eluted between 7.6 and 8.6 min. The
solvent was removed by freeze-drying in a Savant Speed Vac
concentrator until dry. A portion was taken for Ames testing,
and the remainder was dissolved in DCM and a portion taken
for GC-MS.
GC-MS. GC-MS analyses were carried out on a VG 7070/
HS mass spectrometer (Fisons Ltd., Manchester, England) and
HP 5830 gas chromatograph (Hewlett-Packard, Palo Alto, CA)
using 0.25 mm i.d., 0.25 µm, 30 m DB-5 or DB-5MS (5%
phenyl, 95% methyl silicone) capillary columns (J &W Scien-
tific, Folsom, CA) with splitless injection. Electron ionization
(EI) spectra were recorded at 70 eV ionizing voltage. For
negative ion mass spectra we used electron capture negative
ionization (ECNI) with methane reagent gas. Retention index
values were assigned by interpolating between the retention
times of n-alkane standards. All GC runs used linear tem-
perature programs, usually 3 °C/min from 60 to 280 °C, and
splitless injection at 220 °C. The ion source temperature was
180 °C, and transfer lines to the mass spectrometer were
maintained at 230 °C. Samples were usually concentrated 10-
fold prior to GC injection.
P r ep a r a tion of Ch lor in a ted Im id es. Synthetic mixtures
were examined by gas chromatography with a 30 m, 0.25 µm
DB-5 column (J &W Scientific). The reported percent composi-
tions for synthetic mixtures analyzed by GC are not precise
as GC response factors were unknown and the response was
often nonlinear with respect to the amount of sample injected.
See Figure 2 for structures. UV spectra were determined in
1,4-dioxane using an HP diode array spectrometer (Model
8451A, Hewlett-Packard).
F igu r e 1. Procedure for extraction and fractionation of
simulated poultry chiller water derived from chicken frank-
furters or homogenized chicken. The designated fractions were
active in TA100(-S9). DCM, dichloromethane; IPA, isopropyl
alcohol; CH3CN, acetonitrile; HOAC, acetic acid.
Chlorination of Chicken Homogenate. Whole chickens were
obtained from a local commercial chicken processing plant,
prior to immersion cooling (Tsai et al., 1987). Chickens were
manually deboned, and the meats and skin, except the wings,
tail, and abdominal fat, were mashed with a meat chopper
(Model 84181D, Hobart Mfg. Co., Troy OH) until the mixture
was homogeneous. Dry ice was added to maintain the mixture
at 2-5 °C during chopping. About 20 g of the meat mixture
and an equal weight of water were homogenized with the Omni
mixer, and the homogenate was chlorinated in the same
manner as for the frankfurter mixture. A portion of the
homogenate was not chlorinated and served as a control.
Isola tion of Mu ta gen ic F r a ction s. Figure 1 shows the
overall isolation scheme of the mutagenic material from
chlorinated simulated poultry chiller water. Ames/Salmonella
assays were carried out after each extraction and on all
chromatographic fractions. The pH 11 filtrates from chlorina-
tion were extracted three times with 0.2 volume amounts of
reagent grade dichloromethane (DCM); these were designated
alkaline extracts and were saved. The aqueous phase was
acidified to pH 2 and again extracted with DCM as above. The
extracts were combined and designated as the DCM-extract-
able acidic fraction (Schade et al., 1990). A control sample
was extracted in the same manner and found to be nonmu-
tagenic by Ames assay.
CAUTION: Du r in g th e ch lor in a tion r ea ction s, Cl2 a n d
HCl m a y ven t fr om th e r ea ction con ta in er .
3,4-Dichloro-1H-pyrrole-2,5-dione (3,4-Dichloromaleimide,
DCMI) (1). The method of Relles (1972) was used to prepare
dichloromaleic anhydride. Initial crystallization from heptane
gave product with mp 117.5-120.0 °C. Sublimation and
resublimation provided pure product: mp 119.3-120.3 °C [lit.
mp 118-120 °C (Relles, 1972)]; 13C NMR (CDCl3) δ 135.7 (C-
3, 4), 157.9 (C-2, 5). Dichloromaleic anhydride was converted
to dichloromaleimide by heating with urea and NaCl at 120
°C for 20 min (Chow and Naguib, 1984). The reaction mixture
was partitioned between ether and water. The ether solution
was dried with sodium sulfate, treated with decolorizing
carbon, filtered, and concentrated. Addition of hexane and
boiling out of some ether led to crystallization of dichloro-
maleimide: mp 175.0-177.0 °C [lit. mp 174-177 °C (Chow
and Naguib, 1984)]; 13C NMR (CDCl3) δ 134.1 (C-3, 4), 162.0
(C-2, 5).
Molecular Size Chromatography. The pH 2 fraction of the
original chlorination mixture was chromatographed in eight
equal portions on a molecular size exclusion column (Biobeads
S-X1, 5 cm × 100 cm) using DCM at 4 mL/min. The effluent
was monitored at 254 nm, and 20 mL fractions were collected.
Mutagenic activity appeared in fractions designated BB81-86
eluting between 6 h 10 min and 7 h 30 min.
Silica Gel Chromatography. Fraction BB83 was subse-
quently chromatographed on a normal phase column (silica
gel, 10 mm × 250 mm) with hexane/isopropyl alcohol (99:1)
at 5 mL/min. The most Ames-active fraction, designated S7,
eluted between 6 and 7.5 min. There was some additional
mutagenicity seen in a relatively nonpolar fraction eluting at
3-(Dichloromethylene)-2,5-pyrrolidinedione (Compound B)
(2). Hydrogenation of 5.5 mg of compound E (4) (vide infra)
in benzene solution with 7 mg of 5% rhodium on charcoal
catalyst yielded 4.0 mg of white solid with RI of 1515 on DB-
5: 1H NMR (CDCl3) δ 3.47 (s, 2H), (CH2), 8.19 (s, 1H); 13C NMR

258 J. Agric. Food Chem., Vol. 44, No. 1, 1996
Haddon et al.
tained via sublimation and crystallization from toluene, mp
149.0-149.8 °C [lit. mp 147-148.5 °C (Earl et al., 1978); 142-
1
154 °C (Hine et al., 1988)]; H NMR (CDCl3) δ 2.08 (s, CH3),
7.58 (bs, NH); 13C NMR (CDCl3) δ 9.0, 134.4, 138.8, 164.7,
168.4; RI on DB-5, 1228.
A saturated solution of Cl2 in 55 mL of CCl4 was added to
a solution of 8.3 g of 3-chlorocitraconimide in 115 mL of CH2-
Cl2, and the mixture was put in indirect sunlight on the
laboratory roof for 1.5 h. At hourly intervals thereafter, 3.5,
4.5, and 3.7 g of Cl2 was added and the mixture was put in
direct sunlight. Evaporation of solvent left 14.2 g of white
semisolid, seemingly 88% pure by GC (RI of 1470 on DB-5). A
portion dissolved in hot heptane upon cooling deposited 93%
pure 3-methyl-3,4,4-trichloro-2,5-pyrrolidinedione: 1H NMR
(CDCl3) δ 2.02, 8.23; 13C NMR (CDCl3) δ 20.4, 71.1, 84.3, 165.6,
168.8.
A solution of 88% 3-methyl-3,4,4-trichloro-2,5-pyrrolidinedi-
one in CCl4 along with 35% molar excess of Cl2 was exposed
to direct sunlight (HCl evolution). Over the course of 5 days,
more Cl2 was added to the mixture and the chlorination
reaction was continued until less than 20% of the original
compound remained unreacted. After CCl4 was removed, the
reaction product was dissolved in ethyl acetate and extracted
with dilute NaOH solution. The alkaline solution was acidified
with HCl and extracted to provide a mixture containing
compound E (4), DB-5 RI 1717. Multiple chromatographies
on a 2 cm column of silica gel H in a 15 mL filter funnel with
eluting solvent sequence CCl4, benzene, and ether gave a 80-
85% concentrate of E, which was dissolved in ether and diluted
with CCl4. After evaporation of ether, crystals formed. Sev-
eral recrystallizations yielded 3,3-dichloro-4-(dichloro-
methylene)-2,5-pyrrolidinedione: mp 174.0-175.0 °C; UV λmax
250 nm (13 500); 1H NMR (CDCl3) δ 8.41; 13C NMR (CDCl3) δ
73.7, 127.9, 144.1, 159.6, 164.9.
F igu r e 2. Chemical structures of Ames-active chlorinated
imides 1-4 from poultry chiller water and of MX (5), previ-
ously isolated from paper bleach liquors and from water that
has been chlorinated in the presence of humic material
(Holmbom, 1981, 1990).
(CDCl3) δ 37.2, 124.2, 131.6, 164.4, 170.2. After preparation
of a larger amount, the compound was dissolved in ether and
diluted with either CCl4 or benzene. After evaporation of the
ether, 3-(dichloromethylene)-2,5-pyrrolidinedione crystallized;
mp 192.0-192.5 °C; UV λmax 242 nm (14 200).
3-Chloro-4-(dichloromethylene)-2,5-pyrrolidinedione (Com-
pound D) (3). Itaconic anhydride was chlorinated as described
by Schreiber et al. (1961). Vacuum distillation of the product
gave a mixture containing chloromethylmaleic anhydride and,
as indicated by GC-MS, an isomeric compound that is most
likely 3-chloromethylenesuccinic anhydride. The mixture was
refluxed at 150 °C under vacuum for an hour and then
chlorinated as before. These procedures were repeated until
a mixture containing anhydrides with three or four chlorine
atoms was obtained.
Two grams of anhydride mixture, 5 g of urea, and 5 g of
NaCl were stirred with a spatula for 15 min at 125 °C. Product
was partitioned between ethyl acetate and water. Solvent was
removed from the ethyl acetate extract, and the residue was
taken up in ether and filtered through a 1 cm column of oven-
dried silica gel H in a 15 mL filter funnel. Ether was
evaporated, and the 0.9 g of yellowish liquid residue was
dissolved in CCl4 and applied to a 15 mm silica gel H column
and eluted with 10 mL portions of CCl4, benzene, and ether.
After removal of benzene from the benzene eluate, there was
0.2 g of liquid in which crystals formed in a few days. The
crystals were washed with CCl4, dissolved in ether, and
recrystallized from CCl4 after evaporation of ether to yield 20
mg of white solid. A sample recrystallized from benzene had
mp 153.0-154.5 °C, UV λmax 246 nm (14 200), and RI of 1645
on DB-5: 1H NMR (CDCl3) δ 5.09 (s), δ 8.86 (bs); 13C NMR
(CDCl3) δ 51.8, 125.7, 139.6, 162.6, 168.1.
3,3-Dichloro-4-(dichloromethylene)-2,5-pyrrolidinedione (Com-
pound E) (4). Citraconimide was synthesized according to the
procedure of Earl et al. (1978) from citraconic anhydride and
ammonium acetate in acetic acid. Product recrystallized from
CCl4 was a white solid: mp 104.5-105.3 °C [lit. mp 103.5-
105.5 °C (Earl et al., 1978)]; 1H NMR (CDCl3) δ 2.07 (d, 3H, J
) 2 Hz), 6.35 (t, 1H, J ) 2 Hz), 8.19 (bs, 1H); 13C NMR (CDCl3)
δ 10.7, 128.3, 148.9, 171.1, 172.1; RI on DB-5, 1084.
X-r a y Cr ysta llogr a p h ic An a lysis. Tables of anisotropic
thermal parameters with their estimated standard deviations
for the non-hydrogen atoms and observed-calculated structure
factors have been deposited at the Cambridge Crystallographic
Data Center.
Crystal Data of D (3): C5H2Cl3NO2, M ) 214.4, monoclinic,
space group P21/c, a ) 6.929 (1), b ) 5.837 (4), c ) 19.911 (2)
Å, â ) 98.77 (2)°, V ) 795.9 Å3, Z ) 4, F(000) ) 424, µ(Cu KR)
) 102.8 cm-1, and Dcalc ) 1.79 g cm-3; final R ) 0.056 (100
parameters), Rw ) 0.067 for 978 unique reflections, the final
average parameter shift is (0.01σ, and the difference Fourier
synthesis excursions are within (0.4 Å-3
.
Crystal Data of E (4): C5HCl4NO2, M ) 248.9, monoclinic,
space group P21/c, a ) 6.455(1), b ) 17.754(4), c ) 7.697(2) Å,
â ) 103.30(2)°, V ) 858.4 Å3, Z ) 4, F(000) ) 488, µ(Cu KR)
) 125.3 cm-1 and Dcalc ) 1.93 g cm-3; final R ) 0.054 (109
parameters), Rw ) 0.062 for 1039 unique reflections, the final
average parameter shift is (0.01σ, and the difference Fourier
synthesis excursions are within (0.3 Å-3
.
Data Collection and Structure Refinement for 3 and 4.
Single crystals of both compounds suitable for X-ray investiga-
tion were obtained by slow evaporation from a mixture of ether
and CCl4. Intensity data were measured in the range of (3°
e 2θ e 114°) on a Nicolet R3 diffractometer with graphite
monochromatized Cu KR radiation (λ ) 1.5418 Å) by the θ-2θ
scan technique with variable scan speed (4-30° min-1) at room
temperature. The lattice constants were refined by least-
squares fit to setting angles of 20 independent reflections (10°
e 2θ e 25°) measured on the diffractometer. Two standard
reflections were monitored periodically for crystal and instru-
ment stability; no significant change in their intensities was
noted during the course of the experiment. The intensity data
were corrected for background Lorentz polarization effects, and
absorption by the empirical method, but not for secondary
extinction. The crystal structure was solved by direct methods
and refined by a blocked-cascade full-matrix least-squares
procedure. The function minimized was [∑(ω|F(obs)| -
2
|F(calcd)|)2], where ω ) [σ2|F(obs)| + 0.001|F(obs)| ]-1. Unique
reflections with the criteria of (|F(obs)| g 3σ|F(obs)|) were
included in the structure refinement calculation. Scattering
factors were from the International Tables for X-ray Crystal-
lography (1974); those of nitrogen, oxygen, and chlorine were
Citraconimide was chlorinated (Earl et al., 1978) to provide
3,4-dichloro-3-methylsuccinimide, which reacted with hot wa-
ter to yield 3-chlorocitraconimide. Pure compound was ob-

Chlorination of Simulated Poultry Chiller Water
J. Agric. Food Chem., Vol. 44, No. 1, 1996 259
corrected for anomalous dispersion. Atomic parameters of all
non-hydrogen atoms were refined anisotropically, and the
hydrogen position was located in subsequent difference Fourier
maps and included with invariant idealized values in the
respective structure-factor calculation. All structure computa-
tions were accomplished with the SHELXTL (Sheldrick, 1981)
program package.
Nu clea r Ma gn etic Reson a n ce Sp ectr oscop y. NMR
spectra were run in CDCl3 with TMS as an internal standard
on a Bruker ARX400 spectrometer operating at 100.62 MHz
for carbon and 400.13 MHz for proton. One pulse experiments
were run for both nuclei at a 22° pulse for carbon at a 2.3 s
repetition rate and at a 90° pulse for proton at a 7-8 s
repetition rate. Carbon-carbon connectivity studies of 4 were
conducted using a solution of 4 (23 mg in CDCl3) with 1 mg
(0.0029 mmol) of chromium acetylacetone. A one-pulse experi-
ment was run with a 59° pulse at a 1.57 s repetition rate for
572 217 scans. The one-pulse experiment was utilized rather
than an INADEQUATE experiment because of its increased
sensitivity. A relaxation reagent, Gd[fod]3, was added to
decrease the relaxation times of the carbons, so that the one-
pulse experiment could be repeated with no relaxation delay.
Mu ta gen esis Assa ys. Samples were tested for mutage-
nicity by the Ames/Salmonella assay, using the standard plate
incorporation method (Maron and Ames, 1983). S. typhimu-
rium strain TA100, without S-9 mix (-S9), was chosen for
these assays on the basis of its reported sensitivity to mu-
tagens in poultry chiller water (Masri, 1986). Dose-response
curves for the test compounds were determined from four
concentrations in the linear range, selected by preliminary
range-finding experiments. Positive control plates contained
0.4 µL/plate methyl methanesulfonate. Because of apparent
increased stability of test solutions, 1,4-dioxane was used as
solvent for final measurements on synthesized compounds;
ethanol was used as solvent to follow mutagenicity during
chromatographic purification. Positive mutagenic response
was confirmed by duplicate plating in the case of DCMI, and
triplicate plating for the other imides. Mean molar mutage-
nicities, expressed as revertants per nanomole, were calculated
from the slopes of the dose-response curves as determined
by linear regression.
F igu r e 3. (A, top) Total ion current (TIC) chromatogram from
GC-MS of Ames-active fraction RP2 from reversed phase
chromatography of simulated poultry chiller water from
chicken frankfurters (see Figure 1). (B, bottom) Electron
ionization spectrum from peak at scan 818, RI 1299, identified
as 3,4-dichloromaleimide (1). GC retention time is scan
number × 1.957 s/scan.
there were several compounds, of possibly related
structure, present at higher retention index, including
a compound with four chlorines at scan 1134. The
known electrophilic behavior of 1 (Lynch and Crovetti,
1972), its use as a reagent for determination of protein
thiol and amino groups (Smith, 1987), and its patented
use as an insecticide and acaracide in J apan increased
our interest in these related compounds and suggested
that additional purification of fraction RP2 should be
carried out.
P r esen ce of F ive-Ca r b on Ch lor in a t ed Im id es.
Normal Phase HPLC Separation. Normal phase chro-
matography of RP2 on silica (hexane/isopropyl alcohol)
gave a broad peak at 7.20 min with mutagenic activity.
Acidifying the hexane/IPA with 0.5% acetic acid im-
proved the chromatographic separation, and ultimately
we obtained several well-defined chromatographic peaks,
as shown in Figure 4A. The Ames-active fraction was
under the peak at 7.90 min (λmax 249 nm). Our collected
fraction NP1 (7.6-8.6 min) contained all of the Ames
activity of the mixture; it and the adjacent nonmu-
tagenic fraction NP2 (8.6-9.7 min) were subjected to
GC-MS, yielding the TIC chromatograms of Figure
4B,C. DCMI was present in NP1 at scan 561 (RI 1299).
Mass Spectra of Five-Carbon Chlorinated Imides. In
fraction NP1 three peaks had mass spectra resembling
the spectrum of DCMI, and these components were
designated compounds B (scan 802, RI 1511), D (scan
940, RI 1640), and E (scan 1019, RI 1717). The
corresponding mass spectra are shown in Figure 5A-
C. Isotope patterns around m/z 136 (B), 170 (C), and
204 (E) suggested the presence of two, three, and four
chlorine atoms, respectively. For E, we assigned the
4-Cl isotopic cluster at m/z 204 to the M - CHNO ion,
analogous to the m/z 122 peak of DCMI. Loss of CHNO
is characteristic in the mass spectra of unsaturated
cyclic imides (Mitchell and Waller, 1970). These com-
pounds were completely absent in the adjacent inactive
HPLC fraction NP2 shown as the TIC trace in Figure
4C. Compound C (spectrum not shown) appeared as a
shoulder on the trailing edge of D and gave a mass
spectrum similar to that of compound B. We believe
that B and C arose from decomposition of D in the
injection port and transfer line into the mass spectrom-
eter. When NP1 was introduced into the mass spec-
CAUTION: Sever a l of th e com p ou n d s d escr ibed in
th is w or k , p a r ticu la r ly com p ou n d
E (4), h a ve been
sh ow n to be h igh ly m u ta gen ic in TA100 w ith ou t m i-
cr osom a l a ctiva tion . Un til th ese com p ou n d s h a ve been
eva lu a ted for p ossible ca r cin ogen icity a n d oth er bio-
logica l a ctivity, w e r ecom m en d th a t p r eca u tion s be
exer cised in th eir h a n d lin g a n d d isp osa l. Com p ou n d s
D (3) a n d E (4) m a y be d ea ctiva ted by tr ea tm en t w ith
a qu eou s solu tion s of sod iu m bisu lfite.
Ch em ica ls. Dichloromethane, acetonitrile, isopropyl alco-
hol, and hexane used in the chromatographic separations were
all of spectroscopic grade. The water was purified by reverse
osmosis and filtered through a biofilter before use.
RESULTS
Id en tifica tion of 3,4-Dich lor om a leim id e (DCMI,
1). GC-MS analysis of the reversed phase HPLC
mutagenic fraction, RP2, revealed a complex mixture,
represented by the partial TIC chromatogram of Figure
3A. Several chlorinated components could be identified
from their mass spectra, including 2-chlorocyclohexanol
(scan 411, RI 1019) and pentachlorophenol (RI 1747).
Authentic 2-chlorocyclohexanol having the same mass
spectrum and RI was nonmutagenic in TA100 and was
suspected to be an impurity in the solvents (Iwaoka et
al., 1994). The occurrence of a compound identified from
its mass spectrum (Figure 3B) as 3,4-dichloromaleimide
at scan 818, RI 1290, was notable. Its identity was
confirmed by comparison with synthetic DCMI, which
gave the same mass spectrum and RI. Although the
mutagenicity of 1 was too low to account for the
observed mutagenicity of the simulated chiller water,

260 J. Agric. Food Chem., Vol. 44, No. 1, 1996
Haddon et al.
F igu r e 4. (A) UV chromatogram (254 nm) from normal phase
HPLC separation of TA100-active fraction RP2 from simulated
chiller water from chicken frankfurters. (B) TIC chromato-
gram from GC-MS of TA100-active fraction NP1 from normal
phase HPLC. (C) TIC chromatogram for adjacent inactive
fraction NP2. (D) TIC for equivalent active fraction NP1 from
simulated chiller water based on homogenized chicken. The
peak at scan 1254 is dibutyl phthlate. GC retention time is
scan number × 2.162 s/scan.
F igu r e 5. (A-C) Positive EI mass spectra (70 eV) of five-
carbon chlorinated imides from GC-MS of fraction NP1;
compound B (scan 802 of Figure 4B, RI 15.11); compound D
(scan 940, RI 16.40); and compound E (scan 1019, RI 1717).
(D) ECNI mass spectrum of compound E from fraction NP1.
(E) Reference EI mass spectrum of synthetic 3-dichloro-4-
(dichloromethylene)-2,5-pyrrolidinedione (4), RI 1715. All
retention index values are for a DB5-MS GC column (see
Materials and Methods).
trometer without GC separation via the direct insertion
probe, compound E appeared to be the only chlorinated
imide, thus indicating that D may also be an artifact;
DCMI may not have been detected in the direct probe
experiment because of its greater volatility.
For E, accurate mass measurement of the weak (0.5%
relative abundance) ions at m/z 247 (( 10 ppm accuracy)
was consistent with C5HCl4NO2 for the composition of
M+, but we could not unambiguously assign the molec-
ular formula at this level of accuracy. Negative ion
mass spectra provided additional confirmation for the
putative molecular weight of 247 for E; thus, in its ECNI
spectrum, Figure 5D, we assigned the 3-Cl isotopic
cluster at m/z 211 to (M - HCl)-. Similar arguments
suggested that B and D were 2- and 3-Cl imides with
monoisotopic molecular weights of 179 and 213, respec-
tively.
Chlorinated Imides in Chicken Homogenate, Simu-
lated Processing Water. Chlorination of the simulated
processing water from homogenized chicken showed the
same mutagenicity profile during the extraction and
purification. The TIC chromatogram of Figure 4D
represents the mutagenic fraction from the normal
phase HPLC separation, equivalent to fraction NP1.
Compound E was easily detected in the sample, but
compounds B and D were less prominent. These results
provided further evidence that B and D may be artifacts
of the analytical procedures used to isolate them.
Syn th esis a n d Str u ctu r e An a lysis of Com p ou n d s
B, D, a n d E. The mass spectra did not allow assign-
ment of the double-bond position and thus failed to
distinguish between the various chlorine-containing
imides that could be derived from itaconic, citraconic,
or glucaconic imides. Therefore, we prepared synthetic
samples of the suspected mutagens. After several
reactions starting from citraconic or itaconic anhydrides,
we obtained chlorinated imides with identical mass
spectra and retention indices for B, D, and E of 1506,
1642, and 1711, respectively. The mass spectrum of
synthetic E is shown in Figure 5E. Synthetic samples
of compounds D and E were highly mutagenic in TA100.
X-ray Crystallography of Compounds D and E. Single-
crystal analyses unequivocally established that com-
pounds D (3) and E (4) have the 4-(dichloromethylene)-
2,5 pyrrolidinedione structures indicated in Figures 2
and 6. These two compounds, and compound B (2),
whose structure was assigned from NMR, retain the
exocyclic dichloromethylene group and are therefore
chlorinated imides of itaconic acid. The five-membered
ring for both 3 and 4 is planar with a maximum
deviation of (0.01 Å, suggesting that the ring nitrogen
may have sp2 hybridization. Preliminary molecular
orbital calculations for 4 (data not shown) also indicated
a planar ring for the minimum energy structure. These
results are in agreement with force field calculations
and X-ray crystallographic analysis of the structurally
similar R-chloro and R,R-dichloro-3-(chloromethyl)-2-
pyrrolidinones (as their N-tosylates) (Rachita and Slough,
1993). Although the crystal structure of 5 has not been
determined, single-crystal analysis of the MX dimer
showed that it was nearly planar (LaLonde et al., 1990).
13C NMR of Chlorinated Imides. Attempts to ascer-
tain the structure of 3 and 4 based on NMR correlation
data proved to be difficult since it was apparent from
an examination of published 13C NMR spectra that the
chemical shifts of the vinyl carbons would not indicate
the extent of chlorine substitution. However, for chlorine-

Chlorination of Simulated Poultry Chiller Water
J. Agric. Food Chem., Vol. 44, No. 1, 1996 261
1
Ta ble 1. 13C CMR Sh ifts (CDCl3) a n d J cc Obser ved for 4
C
3
shift (ppm)
73.7
1J cc (Hz)
connectivity
53
56
3-4
3-2
4
127.8
52
66
102
3-4
4-5
4-6
6
5
2
144.2
160.1
165.3
102
66
4-6
4-5
3-2
56
and 1450 revertants/nmol, respectively. Assessment of
1 as mutagenic was problematic, as the dose-response
was nonlinear above 1 µg/plate and definitely toxic at
2.5 µg/plate. Under the more liberal criteria of Maron
and Ames (twice the spontaneous revertants in the
linear range), all four chloroimides are mutagenic. The
assigned molar mutagenicities (Table 2) for 1 and 2 were
28 and 0.24 revertants/nmol, respectively. Above 200
µg/plate, 2 was toxic in TA100. For compound E the
mutagenic response in 1,4-dioxane was about twice that
observed when ethanol was used as solvent. In ethanol,
4 gave 800 revertants/nmol on the basis of two inde-
pendent dose-response assays (data not shown).
F igu r e 6. Molecular structures for 3-chloro-4-(dichlorometh-
ylene)-2,5-pyrrolidinedione (3) and 3,3-dichloro-4-(dichloro-
methylene)-2,5-pyrrolidinedione (4) determined from X-ray
crystallographic analysis.
DISCUSSION
The purpose of this work was to identify chemical
structures of mutagens that occur when chlorination is
carried out in poultry chiller water at levels higher than
those which are usually employed in poultry processing
plants. Previous work showed mutagenic activity was
generated in poultry chiller water, containing 1160 ppm
of total solids, treated with chlorine at 1000 ppm (Masri,
1986). Similarly, this study utilized a 1:1 ratio of total
solids to chlorine, but, to facilitate the isolation of active
compounds, the total solids content was increased to
25%. The use of homogenized chicken frankfurters
allowed us to carry out the chlorination reaction on a
mixture that simulates the physical and chemical condi-
tions in food processing plants but has defined and
reproducible composition and history. The subsequent
experiment with unchlorinated chickens provided an
important confirmation of the occurrence of 4. However,
evidence suggests that other mutagens are produced by
chlorination, and one cannot assume that the relative
amounts of different chlorine disinfection byproducts
would be independent of the extent of chlorination
(Masri, 1986). While the identification of the highly
mutagenic compound 4 is significant, it will be impor-
tant to reduce the chlorination level to assess its possible
contribution to mutagenic activity under more realistic
simulated processing conditions.
Compound E (4) is a second example, the first being
MX (5), in which a significant amount of mutagenic
activity of a very complex mixture of chlorination
byproducts can be attributed to a single compound
(Holmbom et al., 1990). At 1450 revertants/nmol in
TA100, the mutagenic potency of 4 in TA100 approaches
that of MX (5000 revertants/nmol), making it the second
most potent bacterial mutagen from chlorination so far
reported in the literature (LaLonde et al., 1991b). For
MX and its structural analogs, the propensity to un-
dergo electrophilic reactions appears to be the best
predictor of mutagenic potency (LaLonde et al., 1992;
Tuppurainen et al., 1992). The activity of 4 is therefore
in accord with the known electrophilic reactivity of 1
substituted aliphatic carbons, the chemical shifts of
mono-, di-, and trisubstituted carbons fall within distinct
ranges. The substituted carbon in monochloroalkanes
has a shift in the range δ 40-50 (δ 40.2 for chloro-
ethane, δ 47.0 for 1-chloropropane, δ 44.6 for 1-chlo-
robutane), in dichloroalkanes in the range δ 70-80 (δ
69.3 for 1,1-dichloroethane, δ 75.0 for 1,1-dichloropro-
pane, δ 71.0 for 1,1-dichloro-3,3-dimethylbutane), and
in trichloroalkanes in the proximity of δ 95 (δ 95.0 for
1,1,1-trichloroethane). From the position of the chemi-
cal shift of the aliphatic carbon (δ 73.7) for 4, it is
apparent that the compound has a dichloro aliphatic
carbon rather than a trichloro substituted carbon.
These data thus favor an exocyclic double bond for
compound 4.
To confirm the 13C NMR assignments for the chlori-
nated carbons, an attempt was made to establish
carbon-carbon connectivity. However, the relaxation
times of the nonprotonated carbons in these compounds
were exceedingly long, making it difficult to obtain
carbon NMR spectra. Because of this, a one-pulse
experiment to look at the one-bond C-C couplings was
chosen for its increased sensitivity, rather than an
INADEQUATE experiment to establish the carbon-
carbon connectivity for the tetrachloroimide 4. From
the observed couplings (Table 1) it was possible to
establish, unambiguously, the connectivity in agreement
with structure 4 of Figure 2 and hence the shift
assignments.
Mu ta gen icity in TA100. Compound E (4) was the
most potent mutagen of the synthetic chlorinated
imides, giving positive response in Ames/Salmonella
strain TA100 without microsomal activation. This
activity was substantially decreased by S9 microsomes
(data not shown). Table 2 compares the mutagenic
response of the four Ames-active chloroimides. Under
the criteria of Prival and Dunkel (1989) 3 and 4 are
direct-acting mutagens with molar mutagenicities of 7.7

262 J. Agric. Food Chem., Vol. 44, No. 1, 1996
Haddon et al.
Ta ble 2. Mu ta gen icity of Syn th esized Ch lor in a ted Im id es in S. typh im u r iu m TA100a
dose, µg/plate
molar mutagenicity
[slope (rev/nmol)]
compound
DCMIc
(nmol/plate)
revertants/plate
rb
1 (6)
0.5 (3)
0.125 (0.75)
351; 365
244; 254
233; 201
28
0.975
solvent control
(0)
179; 178
Bd
125 (694)
50 (278)
10 (56)
280; 318; 325
190; 230; 218
142; 151; 128
0.24
7.7
0.971
0.987
0.95e
D
E
10 (47)
5 (23)
1 (4.7)
451; 533; 513
337; 344; 347
188; 153; 165
0.1 (0.4)
0.025 (0.1)
687; 590; 623
260; 295; 278
1450e
solvent control
(0)
152; 148; 148
Ef
0.1 (0.4)
0.05 (0.2)
0.025 (0.1)
806; 875; 920
451; 514; 515
383; 333; 361
solvent control
(0)
193; 221; 158
a
b
The positive control, methyl methanesulfonate, gave >1200 revertants/plate in all assays at 0.4 µL/plate. Correlation coefficient
from linear regression. c Assay 1. Assay 2. e Data combined from assays 2 and 3. f Assay 3.
d
toward thiol and amino groups (Lynch and Crovetti,
1972; Smith, 1987). The structural similarity between
4 and 5 is noteworthy.
addition because of the resonance effect of directly
bonded Cl on the stability of the positive charge at the
exocyclic carbon, the expected reaction center for 1,4
Michael addition in 4. However, the resonance versus
inductive effect may have reduced importance because,
as with MX (LaLonde et al., 1992; Tuppurainen et al.,
1992), the lowest unoccupied molecular orbital (LUMO)
that appears to be associated with the electrophilicity
of these molecules is considerably delocalized (data not
shown). In addition, for 4, the ring nitrogen appears
to contribute to the LUMO via a hybridized sp2 orbital,
as reflected in the nearly planar structure of the
molecule. These findings are in accord with frontier
orbital MO calculations on maleimide (Mendez, 1992),
which indicate that all four of the maleimide ring
carbons can participate in nucleophilic addition, in
agreement with the known chemistry of maleimide
(J oseph-Nation et al., 1972). From these perspectives,
a study of the reactivity of 4 and its analogs and their
computed electronic properties should be of considerable
interest. Some of the reaction byproducts observed in
the course of attempts to synthesize 3 and 4 appeared,
from their mass spectra, to be isomeric with 3 (two
compounds) and, in one case, with 4. These compounds,
when isolated, should provide additional insight into the
correlation of structure with mutagenic activity for this
interesting new class of bacterial mutagens.
In view of the very low occurrence of 4, it was not
possible to isolate sufficient quantities of the mutagen
for structure analysis beyond mass spectrometry, which
could not determine the location of the double bond or
distinguish between imides having a five- or six-
membered ring. The crucial structural question was the
choice between structures having an endo and exo
double bond in 2-4. These alternative structures would
be related to itaconic acid (exocyclic) or citraconic acid
(endocyclic double bond). Since there are no known
exocyclic analogs of the MX family of compounds,
existing structure-activity relationships established for
the chlorinated hydroxyfuranones did not provide guide-
lines for choosing between the various alternative endo-
and exocyclic structures or the six-membered ring
structures that would be consistent with the elemental
formulas determined from mass spectrometry. The
X-ray crystallographic data on synthesized compounds
with identical mass spectra and retention index values
established unequivocally that the tetrachloroimide, 4,
and the trichloroimide, 3, both had an exocyclic double
bond and are thus derivatives of itaconic rather than
citraconic acid. The assignment of structure for 2 could
be made directly from 13C NMR of the synthesized
compound.
ACKNOWLEDGMENT
In our analyses for compound E in simulated process-
ing water, the amounts of 2 and 3 were variable and
were somewhat dependent on GC-MS injector and
transfer line temperatures, suggesting that they may
be artifacts of the analysis.
We recognize Buenafe Molyneux, Virginia Randall,
and J ohn Schade for contributing expert technical
assistance. We thank J ane Robens, Lee Tsai, and M.
S. Masri for valuable discussions.
The approximately 200-fold reduction in mutagenicity
for 3 compared to 4 is consistent with the data for
analogs of MX, in which substitution of H for the Cl
bonded to the exocyclic carbon or the OH in 5 reduces
the mutagenic activity by 8 and 20 times, respectively
(LaLonde et al., 1991a,b). As a direct-acting mutagen
in TA100, 4 is about one-third as potent as MX. The
lower activity for 4 compared to MX may reflect a
reduced ability of C(6) in 4 to undergo 1,4 Michael
LITERATURE CITED
Chang, L. W.; Daniel, F. B.; DeAngelo, A. B. DNA strand
breaks induced in cultured human and rodent cells by
chlorohydroxyfuranonessmutagens isolated from drinking
water. Teratog., Carcinog., Mutagen. 1991, 11, 103-114.
Chow, Y. L.; Naguib, M. A. [2+2] Photocycloadditions of
dichloromaleimide and dichloromaleic anhydride to cyclic
olefins. J . Chem. Soc., Perkin Trans. 1 1984, 172-174.

Chlorination of Simulated Poultry Chiller Water
J. Agric. Food Chem., Vol. 44, No. 1, 1996 263
Earl, R. A.; Clough, F. W.; Townsend, L. B. Chemical inves-
tigations of citraconimide. J . Heterocycl. Chem. 1978, 15,
1479-1483.
Hine, J .; Hahn, S.; Hwang, J . The relationship between
Brønstead acidities of imides and their hydrogen-bonding
acidities toward oxygen bases. J . Org. Chem. 1988, 53,
884-887.
Masri, M. S. Chlorinating poultry chiller water: the generation
of mutagens and water re-use. Food Chem. Toxicol. 1986,
24, 923-930.
Meier, J . R.; Knohl, R. B.; Coleman, W. E.; Ringhand, H. P.;
Munch, J . W.; Kaylor, W. H.; Streicher, R. P.; Kopfler, F. C.
Studies on the potent mutagen 3-chloro-4(dichloro-
methyl)-5-hydroxy-2(5H)-furanone: aqueous stability, XAD
recovery and analytical determination in drinking water and
in chlorinated humic acid solutions. Mutat. Res. 1987, 189,
363-373.
Mendez, F.; Galvan, M.; Garritz, A.; Vela, A.; Gazquez, J . Local
softness and chemical reactivity of maleimide: nucleophilic
addition. J . Mol. Struct (THEOCHEM) 1992, 81-86.
Mitchell, E. D.; Waller, G. R. Mass spectrometric analysis of
N-substituted cyclohexene-1,2-dicarboximides. Org. Mass
Spectrom. 1970, 3, 519-531.
Holmbom, B. R.; Voss, R. H.; Mortimer, R. D.; Wong, A.
Isolation and identification of an Ames-mutagenic compound
in Kraft chlorination effluents. Tappi 1981, 64, 172-174.
Holmbom, B.; Kronberg, L.; Backlund, P.; Langvik, V.-A.;
Hemming, J .; Reunanen, M.; Smeds, A.; Tikkanen, L.
Formation and properties of 3-chloro-4(dichloromethyl)-5-
hydroxy-2(5H)-furanone, a potent mutagen in chlorinated
waters. In Water Chlorination: Chemistry, Environmental
Impact and Health Effects; Jolley, R. L., Ed.; Lewis: Chelsea,
MI, 1990; Vol. 6, pp 125-135.
National Research Council: Poultry Inspection: The Basis for
a Risk-Assessment Approach; National Academy Press:
Washington, DC, 1987.
International Tables for X-ray Crystallography; Kynoch
Press: Birmingham, U.K., 1974; Vol. 4.
Iwaoka, W.; Hagi, Y.; Umano, K.; Shibamoto, T. Volatile
chemicals identified in fresh and cooked breadfruit. J .
Agric. Food Chem. 1994, 42, 975-976.
J ansson, K.; Maki-Paakanen, J .; Vaittinen, S.-L.; Vartiainen,
T.; Komulainen, H.; Tuomisto, J . Cytogenetic effects of
3-chloro-4-(dichloromethyl)-5-hydroxy-2(5H)-furanone (MX)
in rat peripheral lymphocytes in vitro and in vivo. Mutat.
Res. 1993, 229, 25-28.
J oseph-Nation, P.; Mendoza, V.; Garcia, E. Structure and
proton magnetic resonance study of 3-(N′-aziridinyl)succin-
imides. J . Org. Chem. 1972, 37, 3950-3952.
Kim, J .-W.; Slavik, M.; Pharr, M. D.; Raben, D. P.; Lobsinger,
C. M.; Tsai, S. Reduction of Salmonella on post-chill chicken
carcasses by trisodium phosphate treatment. J . Food Saf.
1994, 14, 9-17.
LaLonde, R. T.; Perakyla, H.; Ishiguro, Y.; Clardy, J .; Brinen,
L. S. Structure of an anhydride related to a mutagenic
component of drinking water, 3-chloro-4-(dichloromethyl)-
5-hydroxy-2(5H)-furanone. Chem. Res. Toxicol. 1990, 3,
404-405.
LaLonde, R. T.; Cook, G. P.; Perakyla, H.; Dence, C. W. Effect
of mutagenicity of the stepwise removal of hydroxyl group
and chlorine atoms from 3-chloro-4-(dichloromethyl)-5-
hydroxy-2(5H)-furanone: 13C NMR chemical shifts as de-
terminants of mutagenicity. Chem. Res. Toxicol. 1991a , 4,
35-40.
LaLonde, R. T.; Cook, G. P.; Perakyla, H.; Bu, L. Structure-
activity relationships of bacterial mutagens related to
3-chloro-4-(dichloromethyl)-5-hydroxy-2(5H)-furanone: an
emphasis on the effect of stepwise removal of chlorine from
the dichloromethyl group. Chem. Res. Toxicol. 1991b, 4,
540-545.
LaLonde, R. T.; Leo, H.; Perakyla, H.; Dence, C. W.; Farrell,
R. P. Associations of the bacterial mutagenicity of haloge-
nated 2(5H)-furanones with their MNDO-PM3 computed
properties and mode of reactivity with sodium borohydride.
Chem. Res. Toxicol. 1992, 5, 392-400.
Prival, M. J .; Dunkel, V. C. Reevaluation of the mutagenicity
and carcinogenicity chemicals previously identified as “False
Positives” in the Salmonella typhimurium mutagenicity
assay. Environ. Mol. Mutagen. 1989, 13, 1-24.
Rachita, M. A.; Slough, G. A. Ruthenium(II) catalyzed ring
closure of prochiral R-chloro-N-tosyl amides: a diastereo-
selectivity study. Tetrahedron Lett. 1993, 43, 6821-6823.
Relles, H. M. Dichloromaleimide chemistry. II. A thionyl
chloride-pyridine method for the conversion of maleimides
to dichloromaleimides. J . Org. Chem. 1972, 37, 3630-3637.
Schade, J . E.; Tsai, L.-S.; Tong, L.; Wilson, R.; MacGregor, J .
T. Extraction of mutagens from chlorinated poultry chiller
water. J . Food Sci. 1990, 55, 635-639.
Schreiber, J .; Leingruber, W.; Pesaro, M.; Trellfall, T.; Eschen-
moser, A. 65. Synthese des Colchicins. Helv. Chim. Acta
1961, 44, 540-597.
Sheldrick, G. M. SHELXTLsan integrated system for solving
refining and displaying crystal structures from diffraction
data. University of Go¨ttingen, Germany, 1981.
Smith, T. L. Method for detecting molecules containing amine
or thiol groups. U.S. Pat. 4 680 272 J uly 14, 1987.
Tsai, L.-S.; Mapes, C. J .; Huxsoll, C. C. Aldehydes in poultry
chiller water. Poult. Sci. 1987, 66, 983-989.
Tuppurainen, K.; Lotjonen, S.; Laatikainen, R.; Vartiainen,
T. Structural and electronic properties of MX compounds
related to TA100 mutagenicity. A semiempirical molecular
orbital QSAR study. Mutat. Res. 1992, 266, 181-188.
Watanabe, S. Equimolar reaction of showdomycin with thiols.
J . Antibiot. 1970, 23, 313-314.
Wiedhopf, R. M.; Trumbull, E. R.; Cole, J . R. Antitumor agents
from J atropha macrorhiza (Euphobiaceae) I: isolation and
characterization of jatropham. J . Pharm. Sci. 1973, 62,
1206-1207.
Received for review February 3, 1995. Revised manuscript
received J uly 10, 1995. Accepted September 19, 1995.X
Lynch, D. M.; Crovetti, A. J . Reactions of dichloromaleimides
with alcohols, phenols, and thiols. J . Heterocycl. Chem.
1972, 9, 1027-1032.
Maron, D. M.; Ames, B. N. Revised methods for the Salmonella
mutagenicity test. Mutat. Res. 1983, 113, 173-215.
J F950076S
X Abstract published in Advance ACS Abstracts, No-
vember 1, 1995.