Retro-diels-alder reaction: Possible involvement in the metabolic activation of 7-oxabicyclo[2.2.1]hepta-2(3),5(6)-diene-2,3-dicarboxylates and a phosphonate analog

Article abstract of DOI:10.1021/tx950127f

A discontinuous structure-activity relationship signaled a change in mode of action and led to the discovery of a possible novel metabolic activation mechanism. The toxicity of the herbicide endothal (exo,exo-7- oxabicyclo[2.2.1]heptane-2,3-dicarboxylic acid) to mice (ip LD₅₀ = 14 mg/kg) is attributed to the inhibition of protein phosphatase 2A (PP2A) at the cantharidin binding site. The potency is reduced by the introduction of a 2,3- or 5,6-double bond. Surprisingly, high toxicity (ip LD₅₀'s = 15-50 mg/kg) is restored in oxabicyclohepta-2(3),5(6)-dienes substituted in the 2- and 3-positions with bis(methyl carboxylate), bis(ethyl carboxylate), and diethyl phosphonate/ethyl carboxylate, whereas the dicarboxylic acid, bis(tert-butyl carboxylate), and bis(dimethyl phosphonate) are inactive. The diene adducts do not inhibit the cantharidin binding site of PP2A. Two observations provided an alternative working hypothesis that the active but not the inactive diene adducts are protoxicants: GC analyses revealed that selected bicyclic dienes readily undergo thermal dissociation by retro- Diels-Alder reactions to liberate disubstituted acetylenes; the liberated acetylenes have mouse ip LD₅₀'s of 8-25 mg/kg. Apparent exceptions to this hypothesis are that bicyclic dienes with bis(tert-butyl carboxylate) and bis(dimethyl phosphonate) substituents are not toxic, yet their corresponding acetylenes are quite toxic. These apparent anomalies are resolved by finding that only the toxic bicyclic dienes readily react with albumin and 4- nitrobenzenethiol and that their low-toxicity analogs are much less reactive. Albumin can be replaced by hemoglobin but not by myoglobin or chymotrypsin in reaction with a bicyclic diene indicating the importance of the free thiol group. Diethyl oxabicycloheptadienedicarboxylate readily reacts with GSH to give two products, which are also formed from the corresponding acetylene, identified as the cis and trans isomers of the GSH-acetylene conjugate. This is the first proposal, to our knowledge, that a retro-Diels-Alder-type reaction is involved in the metabolic activation of a toxicant.

Full text of DOI:10.1021/tx950127f

Chem. Res. Toxicol. 1996, 9, 241-246
241
Retr o-Diels-Ald er Rea ction : P ossible In volvem en t in
th e Meta bolic Activa tion of
7
-Oxa bicyclo[2.2.1]h ep ta -2(3),5(6)-d ien e-2,3-d ica r boxyla tes
a n d a P h osp h on a te An a log
Mahmoud Mahajna, Gary B. Quistad, and J ohn E. Casida*
Environmental Chemistry and Toxicology Laboratory, Department of Environmental Science,
Policy, and Management, University of California, Berkeley, California 94720-3112
Received J uly 17, 1995X
A discontinuous structure-activity relationship signaled a change in mode of action and
led to the discovery of a possible novel metabolic activation mechanism. The toxicity of the
herbicide endothal (exo,exo-7-oxabicyclo[2.2.1]heptane-2,3-dicarboxylic acid) to mice (ip LD50
)
14 mg/kg) is attributed to the inhibition of protein phosphatase 2A (PP2A) at the cantharidin
binding site. The potency is reduced by the introduction of a 2,3- or 5,6-double bond.
Surprisingly, high toxicity (ip LD50’s ) 15-50 mg/kg) is restored in oxabicyclohepta-2(3),5(6)-
dienes substituted in the 2- and 3-positions with bis(methyl carboxylate), bis(ethyl carboxylate),
and diethyl phosphonate/ethyl carboxylate, whereas the dicarboxylic acid, bis(tert-butyl
carboxylate), and bis(dimethyl phosphonate) are inactive. The diene adducts do not inhibit
the cantharidin binding site of PP2A. Two observations provided an alternative working
hypothesis that the active but not the inactive diene adducts are protoxicants: GC analyses
revealed that selected bicyclic dienes readily undergo thermal dissociation by retro-Diels-
Alder reactions to liberate disubstituted acetylenes; the liberated acetylenes have mouse ip
LD50’s of 8-25 mg/kg. Apparent exceptions to this hypothesis are that bicyclic dienes with
bis(tert-butyl carboxylate) and bis(dimethyl phosphonate) substituents are not toxic, yet their
corresponding acetylenes are quite toxic. These apparent anomalies are resolved by finding
that only the toxic bicyclic dienes readily react with albumin and 4-nitrobenzenethiol and that
their low-toxicity analogs are much less reactive. Albumin can be replaced by hemoglobin but
not by myoglobin or chymotrypsin in reaction with a bicyclic diene indicating the importance
of the free thiol group. Diethyl oxabicycloheptadienedicarboxylate readily reacts with GSH to
give two products, which are also formed from the corresponding acetylene, identified as the
cis and trans isomers of the GSH-acetylene conjugate. This is the first proposal, to our
knowledge, that a retro-Diels-Alder-type reaction is involved in the metabolic activation of a
toxicant.
In tr od u ction
PP2A site (Table 1). Two observations led to a working
hypothesis that the toxic bicyclic adducts are protoxi-
cants. First, they undergo thermal dissociation under
GC conditions to the corresponding disubstituted acety-
lenes by retro-Diels-Alder reaction (Figure 1). Second,
the toxicity of dimethyl acetylenedicarboxylate (8, 9)
approximates that of the corresponding diene adduct. The
research described here indicates that several oxabicy-
cloheptadienes derived from toxic disubstituted acety-
lenes undergo metabolic conversion back to the acetylene
derivatives in a biologically catalyzed retro-Diels-Alder-
type reaction and that this constitutes a new bioactiva-
tion mechanism.
exo,exo-7-Oxabicyclo[2.2.1]heptane-2,3-dicarboxylic acid
endothal) (1) is an important herbicide (1) and is also
toxic to mice (ip LD50 ) 14 mg/kg) (2, 3), where it acts as
an inhibitor of protein phosphatase 2A (PP2A) (4, 5) at
(
1
3
the [ H]cantharidin binding site (6, 7). Further studies
on endothal analogs as toxicants and inhibitors of the
cantharidin site revealed a discontinuous structure-
activity relationship (Table 1), suggesting a possible
change in mode of action within the series. The toxicity
and inhibitory activity are reduced by the introduction
of a 2,3- or 5,6-double bond (2, 3, 7). It was therefore
surprising to find that several 2,3-disubstituted 7-oxa-
bicyclo[2.2.1]hepta-2(3),5(6)-dienes, with double bonds in
both positions, are of high toxicity and do not act at the
Exp er im en ta l P r oced u r es
Gen er a l. 1H and 13C NMR spectra were recorded at 300 and
7
5 MHz, respectively, with a Bruker WM-300 spectrometer
*
Author whom correspondence should be addressed at the Envi-
using tetramethylsilane as the internal reference. GC/chemical
ionization mass spectrometry (CI-MS) with selected ion moni-
toring (SIM) involved a Hewlett-Packard 5890 gas chromato-
graph coupled to a 5971A mass spectrometer, a DB-5 fused-
silica capillary column (30 m × 0.25 mm id, J and W Scientific,
Folsom, CA), an injection port temperature of 175 °C (to prevent
thermal decomposition for the oxabicycloheptadienes, which was
complete at 250 °C), and a temperature program of 70 to 250
ronmental Chemistry and Toxicology Laboratory, Department of
Environmental Science, Policy, and Management, 114 Wellman Hall,
University of California, Berkeley, CA 94720-3112. telephone: (510)
6
42-5424. FAX: (510) 642-6497. E-mail: ectl@nature.berkeley.edu.
X
Abstract published in Advance ACS Abstracts, December 1, 1995.
Abbreviations: BSA, bovine serum albumin; CI-MS, chemical
1
ionization mass spectrometry; EGTA, [ethylenebis(oxyethylenenitrilo)]-
tetraacetic acid; FAB, fast atom bombardment; HRMS, high-resolution
mass spectrometry; I50, concentration for 50% inhibition; PP2A, protein
phosphatase 2A; PSCP, 2-phenyl-4H-1,3,2-benzodioxaphosphorin 2-ox-
ide; SIM, selected ion monitoring.
°
C over 20 min. This method was modified in two ways for the
analysis of furan: the column temperature was 50 °C, and the
0
893-228x/96/2709-0241$12.00/0 © 1996 American Chemical Society
+
+

2
42 Chem. Res. Toxicol., Vol. 9, No. 1, 1996
Mahajna et al.
Ta ble 1. Str u ctu r e-Activity Rela tion sh ip s for Toxicity
a n d Rela tive Bin d in g Affin ity for P P 2A of En d oth a l a n d
Rela ted Com p ou n d s
Oxa bicycloh ep ta n es (1-3) a n d -h ep ten es (4-7). Com-
pounds 1-5 were available from previous syntheses in this
laboratory (2, 3). Oxabicyclohept-2(3)-enes 6 and 7 were
prepared by partial reduction of oxabicycloheptadienes 9 and
1
9
2, respectively, described in the following. More specifically,
(10 mmol) in acetone (20 mL) was partially reduced by stirring
for 18 h over 10% (w/w) palladium on carbon (210 mg) under 1
atm of hydrogen. Filtration and solvent evaporation gave an
oily product, which was recrystallized from hexane [mp 51-52
C, lit. mp 51-52 °C (13)] (71%). Compound 7 was synthesized
in 82% yield from 12 by the same procedure, but using ethyl
substituents
°
no.
R1
R2
LD50a (mg/kg) PP2A IC50b (µM)
acetate.
2
,3-Disubstituted 7-Oxabicyclo[2.2.1]heptanes
1c C(O)OH
2
3
C(O)OH
C(O)OEt
C(O)OEt
14
110
50
0.59
1.4
2.2
Oxa bicycloh ep ta d ien es (8-13) (Ta ble 2). To prepare
dicarboxylic acid 8, furan (12 mmol) and bis(trimethylsilyl)
acetylenedicarboxylate (10 mmol) were stirred at 100 °C over-
night in a 3-oz., thick-walled glass pressure reaction apparatus
C(O)OH
C(O)OEt
2
,3-Disubstituted 7-Oxabicyclo[2.2.1]hept-5(6)-ene
C(O)OH
(Lab Glass, Inc., Vineland, NJ ). Dry ether was added, and after
4
C(O)OH
55
4.1
cooling to 0 °C and filtration, the clear solution was evaporated
and the oil residue of bis(trimethylsilyl) oxabicycloheptadiene-
dicarboxylate was hydrolyzed by the addition of methanol, which
was evaporated to give 8 (42%). The general procedure to
prepare 9 (13) was also used for 10 and 11. Phosphonate 12
was made by treating ethyl (diethoxyphosphonyl)propynoate
2
,3-Disubstituted 7-Oxabicyclo[2.2.1]hept-2(3)-enes
5
6
7
C(O)OH
C(O)OMe C(O)OMe
C(O)OEt P(O)(OEt)2
C(O)OH
>250
>250
>250
>10
>10
>10
2
,3-Disubstituted 7-Oxabicyclo[2.2.1]hepta-2(3),5(6)-dienes
8
9
0
C(O)OH
C(O)OMe C(O)OMe
C(O)OEt C(O)OEt
C(O)OH
>250
>10
>10
>10
(18) (4 mmol) with excess furan (6 mL) at 120 °C under pressure
15
as before for 36 h. Unreacted furan was removed by distillation,
ether was added, and after filtration and solvent evaporation
the oil was chromatographed on a silica gel column with
n-hexane-ether (1:1) to obtain 12 (58%). A similar procedure
converted 19 to 13 (39%).
1
21
a
Mouse ip 72-h data for 1-5 from earlier studies (2, 3) and for
b
3
6
-10 from this study. Inhibition of [ H]cantharidin binding site
c
of PP2A in mouse liver cytosol from this study. The herbicide
endothal.
Toxicity to Mice. LD50 values were determined 72 h after
ip administration of the test compounds to male albino Swiss-
Webster mice (18-23 g), using Me SO as the carrier vehicle (50
2
µL).
3
Com p etition for [ H]Ca n th a r id in Bin d in g Site (6, 7).
Mouse liver cytosol was prepared at 20% (w/v) equivalent in 50
mM imidazole hydrochloride (pH 7.0) containing 1 mM EDTA,
1
mM EGTA, 100 µM N-ethylmaleimide, and 10 µM 2-phenyl-
F igu r e 1. Retro-Diels-Alder reaction of diethyl 7-oxabicyclo-
[
2.2.1]hepta-2(3),5(6)-diene-2,3-dicarboxylate.
4H-1,3,2-benzodioxaphosphorin 2-oxide (PSCP) (an esterase
inhibitor) (14). The competitive binding assays were carried out
3
with cytosolic protein (250 µg) and 5 nM [ H]cantharidin in the
preceding buffer (500 µL); the inhibitors were added in buffer
or acetone (<1% final concentration) and incubated for 120 min
at 37 °C before filtration through poly(ethylenimine)-treated
glass fiber filters. Nonspecific binding was determined with 10
µM unlabeled cantharidin.
Rea ctivity w ith 4-Nitr oben zen eth iol. The relative reac-
tivities were compared for the various bicyclic adducts and
acetylene derivatives with 4-nitrobenzenethiol. The acetylene
or bicyclic diene derivative (8-19) was added individually as
an acetonitrile solution to 4-nitrobenzenethiol in 200 mM
sodium phosphate (pH 7.4) buffer containing 0.1 mM EDTA (to
minimize thiol autoxidation) at 25 °C. The final concentrations
were 120 µM for the acetylene or diene derivative, 120 µM for
nitrobenzenethiol, and 2% for acetonitrile. The extent of
reaction was determined spectrophotometrically as the decrease
in absorbance at 410 nm due to the disappearance of 4-ni-
trobenzenethiolate ion (15) after 200 s for the bicyclic dienes
and after 10 s for the acetylene derivatives.
F igu r e 2. Synthesis of 2,3-disubstituted 7-oxabicyclo[2.2.1]-
hepta-2(3),5(6)-dienes (8-13) by Diels-Alder reaction of furan
and disubstituted acetylenes (14-19) and partial reduction of
two products to 2,3-disubstituted 7-oxabicyclo[2.2.1]hepta-2(3)-
enes (6, 7).
sample for analysis was 10 µL withdrawn from the 8-mL head
space above a reaction mixture in a vial sealed with a septum.
Fast atom bombardment/high-resolution mass spectrometry
(
(
FAB/HRMS) was performed with a Kratos MS-50 instrument
Department of Chemistry, University of California at Berkeley).
HPLC utilized a Merck LiChrospher 100 RP-18 (5-µm) reverse-
phase column on a Waters Model 600E solvent delivery system
coupled to a Model 994 photodiode array detector: 0.1% tri-
fluoroacetic acid for 5 min, linear gradient of 0 to 60% methanol
in water with constant 0.1% trifluoroacetic acid over 20 min,
and finally a linear gradient to 100% methanol over an ad-
ditional 5 min, each at 1.5 mL/min (monitoring at 220 nm).
Ch em ica ls. Ca u tion : Dimethyl acetylenedicarboxylate, a
widely used reactant solvent, is a lachrymator and primary skin
irritant (8, 9). Care therefore should be taken in studies using
this chemical and its analogs. The oxabicycloheptadienes (8-
P r od u ct s fr om R ea ct ion s w it h N-Acet ylcyst ein e a n d
GSH. The products from reactions of diethyl oxabicyclohepta-
dienedicarboxylate and the corresponding acetylene with N-
acetylcysteine were isolated and identified (Figure 3). A solution
of 10 (or 16) (20 mg) and N-acetylcysteine (1 molar equiv) in
100 mM phosphate buffer (pH 7.4, 2 mL) was incubated for 1 h
at 37 °C and then analyzed by HPLC. The same two products
were found in each case. The starting materials gave t
of 25.2 and 27.6 min for 10 and 16, respectively, while the t
R
values
’s
1
3) and oxabicyclohept-2(3)-enes (6 and 7) were synthesized as
R
shown in Figure 2 and detailed in the following. The purity of
each of these compounds was estimated to be >97% on the basis
of 1H and C NMR integration. Acetylene derivatives 14
for the N-acetylcysteine conjugates (20 and 21) were 24.3 and
25.7 min, respectively (ratio 3:1). Conjugates 20 and 21 were
isolated by HPLC in amounts of 3-8 mg each. NMR and MS
13
(monopotassium salt) and 17 were from Aldrich Chemical Co.
data for the isomers are given for 20 followed by those for 21 in
1
(Milwaukee, WI) and the others were synthesized by reported
parentheses. NMR (CD
OCH CH
3
OD): H δ 1.25 (1.28) (3H, t, HC(O)-
procedures, i.e., 15 and 16 (10), 18 (11), and 19 (12).
2
3
), 1.32 (1.35) (3H, t, SC(O)OCH
2
CH
3
), 1.99 (1.97) (3H,
+
+

Retro-Diels-Alder in Metabolic Activation
Ta ble 2. 13C a n d 1H NMR Sp ectr a l Ch a r a cter istics of 2,3-Disu bstitu ted 7-Oxa bicyclo[2.2.1]h ep ta -2(3),5(6)-d ien esa
substituents
13C NMR (ppm)
Chem. Res. Toxicol., Vol. 9, No. 1, 1996 243
no.
R1
R2
C1
C2
C3
C4
C5
C6
C7
C8
other
8
9
0
1
2
C(O)OH
C(O)OH
C(O)OMe
C(O)OEt
C(O)O-t-Bu
P(O)(OEt)2
86.1
85.0
84.8
85.0
84.7
157.5
152.9
152.4
153.1
153.4
157.5
152.4
153.1
153.1
158.8
86.1
85.0
84.8
85.0
86.9
144.2
143.2
142.9
143.1
142.9
144.2
143.2
142.9
143.1
142.9
166.3
163.1
162.8
163.2
162.4
166.3
163.1
162.8
163.2
C(O)OMe
C(O)OEt
C(O)O-t-Bu
C(O)OEt
52.3 (2Me)
61.0, 13.9 (2Et)
82.3, 28.0 (2t-Bu)
61.2, 13.7 [C(O)OEt],
1
1
1
6
2.2, 15.8 [P(O)(OEt)2]
1
3
P(O)(OMe)2
P(O)(OMe)2
substituents
87.3
158.2
158.2
87.3
143.1
143.1
53.0 (4Me)
1H NMR (ppm)
H6
no.
R1
R2
H1
H4
H5
other
9
C(O)OMe
C(O)OEt
C(O)O-t-Bu
C(O)OEt
C(O)OMe
C(O)OEt
C(O)O-t-Bu
5.68 (s)
5.68 (s)
5.55 (s)
5.68 (s)
5.68 (s)
5.68 (s)
5.55 (s)
5.79 (s)
7.22 (s)
7.22 (s)
7.16 (s)
7.17 (d)
7.22 (s)
7.22 (s)
7.16 (s)
7.21 (d)
3.83 (s) (2Me)
4.29 (q) 1.3 (t) (2Et)
1.48 (s) (2t-Bu)
1
1
1
0
1
2
P(O)(OEt)2
4.26 (q), 1.32 (t) [C(O)OEt],
4
.15 (dq), 1.30 (t) [P(O)(OEt)2]
1
3
P(O)(OMe)2
P(O)(OMe)2
5.81 (s)
5.81 (s)
7.14 (s)
7.14 (s)
3.83 (m) (4Me)
a
CDCl3 solutions in all cases except 8 in CD3OD. Abbreviations are as follows: s, singlet; d, doublet; t, triplet; q, quartet; m, multiplet.
s, C(O)CH
3
), 3.14-3.42 (3.13-3.46) (2H, m, SCH
2
), 4.14 (4.22)
Ta ble 3. Str u ctu r e-Activity Rela tion sh ip s for Toxicity
a n d Rela tive Bin d in g Affin ity for P P 2A of
2,3-Disu bstitu ted
(
(
2H, q, HC(O)OCH ), 4.29 (4.32) (2H, q, (SC(O)OCH ), 4.64 (4.55)
2
2
1H, q, NCH), 5.96 (6.45) (1H, s, HCdC); 13C δ 14.17 (14.32),
7
-Oxa bicyclo[2.2.1]h ep ta -2(3),5(6)-d ien es a n d
Disu bstitu ted Acetylen es
1
4.46 (14.52), 22.37 (22.43), 34.02 (34.95), 52.92 (54.06), 61.91
(61.97), 63.59 (63.92), 116.15 (122.95), 149.91 (148.38), 164.98
compound no.
LD50 (mg/kg)a,b
(165.40), 166.77 (166.44), 172.52 (172.73), 173.36 (173.24). FAB/
+
HRMS [MH ] calcd for C13
H
19NO
7
S 334.0960, found 334.0963
substituents
bicyclic acetylene bicyclic acetylene
(
334.0960).
The products from reaction of dimethyl oxabicycloheptadi-
enedicarboxylate with GSH were isolated for characterization
Figure 3). Reaction of 9 (20 mg) with GSH (1 molar equiv) was
R
1
R
2
diene
deriv
diene
deriv
C(O)OH
C(O)OMe
C(O)OEt
C(O)O-t-Bu C(O)O-t-Bu
C(O)OEt
C(O)OH
C(O)OMe
C(O)OEt
8
9
10
11
12
13
14
15
16
17
18
19
>250
15
21
>200
50
>250
8
12
27
10
(
carried out in 100 mM phosphate buffer (pH 7.4, 2 mL) for 1 h
at 37 °C, with the isolation of two products (22 and 23) in a 3:1
ratio and analyses as before. Data for the conjugates are given
P(O)(OEt)
P(O)(OMe)
2
P(O)(OMe)
2
2
>200
25
for 22 followed by those for 23 in parentheses. t
R
: 18.4 (19.5)
), 2.57 (2.56)
), 3.69 (3.74) (3H,
), 3.92 (3.90) (2H, s, H ), 4.03 (4.03)
), 4.56 (4.59) (1H, q, H ), 6.03 (6.49) (1H, s, H
6.93 (26.98), 32.35 (32.41), 34.14 (35.34), 41.84 (41.84), 52.38
a
Mouse ip at 72 h. b The IC50 for each compound is >10 µM for
1
min. NMR (CD
2H, t, H ), 3.08-3.39 (3.09-3.40) (2H, m, H
s, H ), 3.83 (3.87) (3H, s, H
1H, t, H
3
OD): H δ 2.17 (2.19) (2H, s, H
b
3
the [ H]cantharidin binding site of protein phosphatase 2A (PP2A)
in mouse liver cytosol.
(
c
f
i
h
e
1
3
(
a
d
g
); C δ
as the vehicle (50 µL). At 5-60 min after treatment, the
animals were sacrificed by cervical dislocation and the liver was
homogenized in water (2 mL) at 5 °C. The homogenate was
centrifuged (10000g, 30 min), the supernatant fraction was
extracted with dichloromethane (2 mL), and a 1-µL aliquot was
analyzed by GC/MS.
2
(
(
52.38), 53.28 (53.55), 53.52 (53.94), 53.73 (54.66), 115.82
122.83), 149.94 (148.04), 165.55 (165.95), 167.34 (167.12),
1
71.42 (171.38), 171.84 (171.89), 172.61 (172.07), 174.41 (174.21).
+
FAB/HRMS [MH ] calcd for C16
23 3
H N O10S 450.1182, found
4
50.1182.
R ea ct ivit y in t h e P r esen ce of Albu m in a n d Ot h er
Resu lts
P r otein s. The bovine serum albumin (BSA) used was fraction
V from initial cold ethanol precipitation (96-99% pure) or heat
shock (98-99% pure) (Sigma Chemical Co., St. Louis, MO).
Incubation mixtures consisted of 100 mg (1.67 µmol) of BSA in
Str u ctu r e-Activity Rela tion sh ip s (Ta bles 1 a n d
3). The dicarboxylic acid endothal (1) is the most toxic
of the oxabicycloheptanes examined and is also the most
potent at the PP2A binding site. The toxicity and
inhibitory potency are reduced upon conversion of 1 to
its ethyl and diethyl esters (2 and 3). Introduction of a
1
00 mM phosphate buffer (pH 7.4, 1 mL), to which was added
the test compound (1.67 µmol unless indicated otherwise) in 10%
methanol in buffer (1 mL) followed by incubation for 1 h at 37
°
C. Analysis involved extraction with dichloromethane (2 mL)
and GC/MS on a 1-µL aliquot. t
values (minutes) for the test
compounds and products were as follows: 9 (10.0), 10 (11.6),
1 (12.8), 15 (4.4), 16 (6.5), 17 (9.1), and furan (1.8; head-space
analysis). Phosphonates 12 and 13 are unstable on GC, yielding
their acetylene derivatives with t ’s (minutes) of 12.7 for 12 and
8 and 12.1 for 13 and 19. Accordingly, the phosphonyl
derivatives were analyzed by HPLC on a direct aliquot (10 µL)
of the incubation mixture, resulting in t ’s (minutes) of 12 (25.1),
3 (18.4), 18 (26.5), and 19 (19.5).
In another set of experiments with bicyclic adduct 9, BSA
5
,6-double bond to give 4 reduces the potency, and
R
introduction of a 2,3-double bond to yield 5-7 results in
the loss of toxic and inhibitory properties. On an overall
basis for compounds 1-7, the toxic compounds interact
and the nontoxic ones do not with the PP2A binding site.
The combination of 2,3- and 5,6-double bonds in the
2(3),5(6)-diene series alters the potency, but not always
with the expected total loss of activity. The dicarboxylic
acid 8 is neither a toxicant nor an inhibitor at the PP2A
site, perhaps due to the 2,3-double bond. It therefore was
surprising to find high toxicity for the corresponding
dimethyl and diethyl esters (9 and 10). Further synthe-
sis and testing established that the diethyl phosphonate/
ethyl carboxylate 12 is toxic, but the bis(tert-butyl
carboxylate) 11 and bis(dimethyl phosphonate) 13 are
not, i.e., an unusual structure-activity pattern.
1
R
1
R
1
was compared with bovine hemoglobin, myoglobin, and R-chy-
motrypsin (1.67 µmol) in the presence and absence of the
esterase inhibitor PSCP (14) (10 µmol, 10-min preincubation)
added in Me SO (100 µL). The reaction and analytical condi-
2
tions were as before except that the incubation time was 30 min.
Meta bolism in Mice. Mice as before were administered a
2
test compound (9-11 and 15-17) ip at 250 mg/kg, with Me SO
+
+

2
44 Chem. Res. Toxicol., Vol. 9, No. 1, 1996
Mahajna et al.
Ta ble 4. Su bstitu en t Effects on Rela tive Rea ctivity w ith 4-Nitr oben zen eth iol a n d BSA of 2,3-Disu bstitu ted
7
-Oxa bicyclo[2.2.1]h ep ta -2(3),5(6)-d ien es a n d Disu bstitu ted Acetylen es
relative reactivitya
substituents
compound no.
bicyclic diene acetylene deriv bicyclic diene acetylene deriv bicyclic diene acetylene deriv
4-nitrobenzenethiolb
albuminc
R1
R2
C(O)OH
C(O)OH
C(O)OMe
C(O)OEt
C(O)O-t-Bu
P(O)(OEt)2
8
9
10
11
12
13
14
15
16
17
18
19
-
-
-d
C(O)OMe
C(O)OEt
C(O)O-t-Bu
C(O)OEt
+++
++
-
+++
+++
++
++
+
+++
+++
-
+++
+++
-
+++
++
+
-
+
-
P(O)(OMe)2 P(O)(OMe)2
a
Extent of reaction in pH 7.4 phosphate (%): +++, 90-100; ++, 60-80; +, 20-40; -, <5. b Reaction for 200 s with bicyclic dienes and
c
for 10 s with acetylene derivatives at 25 °C. Reaction for 1 h at 37 °C. Analysis for loss of 12 and 13 by HPLC and of the other compounds
except for dicarboxylic acids 8 and 14) by GC/MS. No furan was found in the head space above the reaction mixture.
d
(
Ta ble 5. Rela tive Rea ctivities of BSA, Hem oglobin ,
Acetylene derivatives 14-19 corresponding to dienes
Myoglobin , a n d Ch ym otr yp sin w ith Dim eth yl
8
-13 were also examined for toxicity and binding site
7
-Oxa bicyclo[2.2.1]h ep ta -2(3),5(6)-d ien ed ica r boxyla te
interaction. Acetylenedicarboxylic acid 14 is much less
toxic than the carboxylic and phosphonic acid ester
derivatives 15-19, which have ip LD50’s of 8-27 mg/kg.
The toxicity of the bicyclic dienes (8-13) and acetylene
derivatives (14-19) does not correlate with their potency
at the PP2A binding site. Lethal doses of the oxabicy-
cloheptanes and oxabicycloheptene (1-4) lead to hepa-
tohemia (visually evident), whereas lethal doses of the
oxabicycloheptadienes (9, 10, and 12) and acetylene
derivatives (15-19) do not.
Rea ctivity w ith 4-Nitr oben zen eth iol (Ta ble 4).
Nitrobenzenethiol was used as a convenient model for
biological thiols since it combines high reactivity with
ease of monitoring the reaction rates using a single
method for all compounds. The acetylene derivatives
Alon e a n d in th e P r esen ce of a n Ester a se In h ibitor
relative reactivitya
proposed
protein
albumin
hemoglobin
myoglobin
alone
PSCP
reactive site
+++
+++
(
+++
+++
-
cysteine
cysteine
none
chymotrypsin
+++
-
serine
a
Reaction for 30 min at 37 °C. Extent of reaction in pH 7.4
phosphate (%): +++, 80-90; ++, 60-70; (, 10-15; -, <5.
Rea ctivity in th e P r esen ce of Albu m in (Ta ble 4).
Upon incubation with BSA, bicyclic dienes 9 and 10 react
rapidly, 12 reacts more slowly, and 11 and 13 do not
react, paralleling their relative reactivities with nitroben-
zenethiol. Acetylenes 15, 16, and 18 react rapidly with
BSA, 19 reacts more slowly, and 17 does not react,
following the pattern of the corresponding dienes, except
for the bis(dimethyl phosphonate) 19. Interestingly,
bicyclic dienes 9, 10, and 12, which react readily with
BSA, are toxic, whereas adducts 11 and 13, which do not
react with BSA, are not toxic (Table 3).
Further information is available on the nature of the
reaction of BSA with bicyclic dienes. The analogous
oxabicyclohept-2(3)-enes (6 and 7) are not reactive (data
not given). Furan is detected by head-space analysis
from 9 and 10 with BSA, but not without BSA. The
reaction of diene adducts 9 and 10 is complete in 1 h with
equimolar compound to BSA and in 24 h with a 20:1
molar ratio. Acetylenes 15 and 16 (5 molar equiv) react
completely with BSA in 1 h. In the reaction of bicyclic
dienes 9-13 with BSA, the corresponding acetylene
derivatives 15-19 are never observed. In each case, the
substrate is stable for the designated time in the absence
of BSA, and the two preparations of BSA examined do
not differ in their reactivities.
(15-19) react with nitrobenzenethiol at least 20-fold
faster than the corresponding bicyclic dienes (9-13). The
reactivity pattern in each series follows the general order
bis(methyl carboxylate) g bis(ethyl carboxylate) > diethyl
phosphonate/ethyl carboxylate > bis(tert-butyl carboxy-
late) > bis(dimethyl phosphonate) . dicarboxylic acid.
P r od u cts fr om Rea ction s w ith N-Acetylcystein e
a n d GSH (F igu r e 3). Bicyclic dienes such as the bis-
(methyl carboxylate) 9 and bis(ethyl carboxylate) 10
readily react with cysteine, N-acetylcysteine, and GSH,
but not with serine, lysine, or N-acetyllysine on the basis
of comparative losses analyzed by GC of these substrates
in the organic extracts (data not given). The reaction of
bicyclic diene 10 with N-acetylcysteine yields conjugates
2
0 and 21 in a 1:1 ratio, and furan is also detected by
GC/MS. The reaction of the corresponding acetylene (16)
with equimolar N-acetylcysteine also gives conjugates 20
and 21 (identified by coinjection on HPLC), but in a 1:3
ratio. Bicyclic diene 9 reacts with equimolar GSH at pH
7
.4 to give not only furan but also conjugates 22 and 23
in a 3:1 ratio and a quantitative yield based on HPLC
analysis. Geometrical isomer assignments for 20-23,
isolated by HPLC from the bicyclic diene reactions, are
based on differences in the 1H chemical shift for the
olefinic proton, i.e., this proton in the cis isomers (20 and
Rea ctivity in th e P r esen ce of Oth er P r otein s
(Ta ble 5). Albumin, hemoglobin, and chymotrypsin
almost completely degrade bicyclic diene 9, whereas
myoglobin is much less active or inactive. Furan was
detected with hemoglobin but not with myoglobin. Heat-
denatured albumin and hemoglobin do not release furan
from dienes. Chymotrypsin is no longer effective when
treated with the esterase inhibitor PSCP.
Meta bolism in Mice. The reactivities of the dimethyl
(9) and diethyl (10) oxabicycloheptadienedicarboxylates
and the analogous acetylenes (15 and 16, respectively)
are established by the failure to recover them from the
livers of mice 5-60 min after ip administration at 250
mg/kg or even from liver homogenates fortified as in a
recovery analysis. The corresponding experiments with
2
2) is shielded by about 0.5 ppm compared with that in
the trans isomers (5.96 and 6.03 ppm for 20 and 22,
respectively, versus 6.45 and 6.49 ppm for 21 and 23,
respectively). In 13C NMR, the olefinic carbon bonded
to sulfur is shielded by 6 ppm in the cis isomers (20 and
2
1
1
2) compared with the trans isomers (21 and 23), i.e.,
16.15 and 115.82 ppm, respectively, versus 122.19 and
22.83 ppm, respectively. The mixture of GSH adducts
derived from 9 is not toxic to mice ip at 250 mg/kg
(equivalent to >120 mg/kg for the parent 9).
+
+

Retro-Diels-Alder in Metabolic Activation
Chem. Res. Toxicol., Vol. 9, No. 1, 1996 245
the tert-butyl diene led to recovery of the parent com-
pound (11) but not its acetylene derivative (17), whereas
the acetylene was recovered on direct administration.
Total loss of the bicyclic dienes and acetylenes with bis-
(methyl carboxylate) substituents (9 and 15) probably is
not due to esteratic cleavage since a 30-min pretreatment
with PSCP at 100 mg/kg does not lead to recovery of the
administered esters.
Discu ssion
The discontinuous structure-activity relationships
reveal two types of toxic bicyclic adducts. One set
consists of the oxabicycloheptane derivatives and an
oxabicyclohept-5(6)-ene analog, which probably poison as
PP2A inhibitors. The second set contains selected 2,3-
disubstituted oxabicycloheptadienes with markedly al-
tered substituent requirements and toxicity not associ-
ated with PP2A inhibition. Attention therefore was
focused on the oxabicycloheptadienes because their mode
of action is unknown. GC analyses provided a clue to
the difference between the two sets in that the first series
of toxic adducts is stable on GC, whereas the bicyclic
dienes readily undergo thermal dissociation by retro-
Diels-Alder reaction to liberate the corresponding acety-
lenedicarboxylates.
F igu r e 3. Reactions of diethyl 7-oxabicyclo[2.2.1]hepta-2(3),5-
(
6)-diene-2,3-dicarboxylate (10) and diethyl acetylenedicarboxy-
late (16) with N-acetylcysteine and of dimethyl 7-oxabicyclo-
2.2.1]hepta-2(3),5(6)-diene-2,3-dicarboxylate (9) with GSH.
[
Dimethyl acetylenedicarboxylate (15) also reacts directly with
GSH. Protons in structures 22 and 23 are designated as a-i.
force is the formation of the relatively stable leaving
group (furan), since similar reactions do not occur for the
2
,3-dehydro derivatives (5-7). Addition of N-acetylcys-
teine and GSH to the acetylenedicarboxylate is analogous
to known additions of cysteine to acetylenedicarboxamide
(an antibiotic produced by Streptomyces chibaensis, ef-
Dimethyl acetylenedicarboxylate is of relatively high
toxicity to rats with both oral and dermal treatments (8,
fective in the control of rice bacterial leaf blight) (16) and
of GSH to the triple bond of 4-hydroxy-3-methyl-2-(2-
propynyl)-cyclopent-2-en-1-one (an insecticide hydrolysis
product) (17). Reactions of GSH with the bicyclic dienes
and acetylene derivatives in this study may serve as
detoxification mechanisms.
The bicyclic dienes are readily degraded by two pro-
teins with free thiol substituents (albumin and hemo-
globin), but not by two others lacking a free thiol
(myoglobin and chymotrypsin) (18), provided that the
latter two proteins are assayed with PSCP. The corre-
sponding dimethyl oxabicyclohept-2(3)-enedicarboxylate
and a diethyl phosphonate/ethyl carboxylate analog do
not react with albumin, indicating that the loss upon
incubation with this protein is not attributable to the
carboxylic acid ester substituent, to the 2,3-double bond,
or to nonspecific binding. Chymotrypsin readily degrades
dimethyl oxabicycloheptadienedicarboxylate, but this
reaction is blocked by PSCP, consistent with ester
hydrolysis by this serine protease.
Earlier studies in this laboratory established the
importance of albumin in the bioactivation of an orga-
nophosphorus delayed neurotoxicant (19), in “enzymatic”
detoxification of the methylcarbamate insecticide car-
baryl (20), and in reaction with the decachloro miticide
dienochlor in which derivatization occurs (21). Albumin
may also play a role in the action of the bicyclic dienes
with methyl and ethyl carboxylate substituents. Thus,
BSA at a concentration approximating that in plasma
readily reacts with these dienes and the corresponding
acetylene derivatives are never observed, consistent with
the greater reactivity of the purported product than of
the substrate. Albumin, which contains a single free thiol
(18), degrades several equivalents of the bicyclic adducts
and corresponding acetylenes so that the derivatized thiol
site in the macromolecule may be regenerated, as with
an enzymatic reaction, even though the derivatives are
more stable with small molecules such as the GSH and
N-acetylcysteine conjugates.
9
). The present study uses mice to compare the sub-
stituent effects on the toxicity of the bicyclic dienes and
the corresponding acetylene derivatives. There is a clear
relationship for LD50’s of the diene adducts and acety-
lenes in three cases [bis(methyl carboxylate), bis(ethyl
carboxylate), and diethyl phosphonate/ethyl carboxylate],
and the dicarboxylic acids are nontoxic as tested in both
series, leading to the hypothesis that toxic bicyclic dienes
are protoxicants for the acetylene derivatives. An ap-
parent anomaly, that oxabicycloheptadienes with bis(tert-
butyl carboxylate) and bis(dimethyl phosphonate) sub-
stituents have low toxicity but their acetylene counterparts
are very toxic, might be due to the structural specificity
of a bioactivation site, for which BSA was examined as a
possible model. Importantly, only the toxic bicyclic
dienes readily react with BSA, and the low-toxicity bis-
(tert-butyl carboxylate) and bis(dimethyl phosphonate) do
not. A similar relationship is indicated for reactions in
vivo or with liver homogenate since toxic bicyclic dienes
with bis(methyl carboxylate) and bis(ethyl carboxylate)
substituents are very reactive, while the nontoxic bis-
(tert-butyl carboxylate) analog is more stable and does
not yield detectable amounts of di-tert-butyl acetylene-
dicarboxylate (which, in contrast to its methyl and ethyl
analogs, would be detected if it was formed). The
protoxicant hypothesis thus applies to the bicyclic dienes
rapidly reacting with bis(methyl carboxylate), bis(ethyl
carboxylate), and diethyl phosphonate/ethyl carboxylate
substituents, but not to the less reactive bis(tert-butyl
carboxylate), bis(dimethyl phosphonate) and dicarboxylic
acid analogs, with possible stabilization by steric and
electronic effects.
Proposed reactions and products for the bicyclic dienes
and acetylene derivatives with N-acetylcysteine and GSH
involve nucleophilic attack of the sulfur on the double
bond followed by the departure of furan (Figure 3). The
nucleophilicity of the thiol group is important since
cysteine, N-acetylcysteine, and GSH react readily, but
serine, lysine, and N-acetyllysine do not. The driving
The target site(s) for the toxic oxabicycloheptadienes
may prove to be as complex, and possibly the same, as
+
+

2
46 Chem. Res. Toxicol., Vol. 9, No. 1, 1996
Mahajna et al.
J . E. (1990) Endothall thioanhydride: structural aspects of
unusually high mouse toxicity and specific binding site in liver.
Chem. Res. Toxicol. 3, 318-324.
(
(
(
(
(
4) Li, Y.-M., and Casida, J . E. (1992) Cantharidin-binding protein:
identification as protein phosphatase 2A. Proc. Natl. Acad. Sci.
U.S.A. 89, 11867-11870.
5) Li, Y.-M., Mackintosh, C., and Casida, J . E. (1993) Protein
3
3
phosphatase 2A and its [ H]cantharidin/[ H]endothall thioanhy-
dride binding site. Biochem. Pharmacol. 46, 1435-1443.
6) Graziano, M. J ., Waterhouse, A. L., and Casida, J . E. (1987)
Cantharidin poisoning associated with specific binding site in
liver. Biochem. Biophys. Res. Commun. 149, 79-85.
7) Graziano, M. J ., Pessah, I. N., Matsuzawa, M., and Casida, J . E.
(1988) Partial characterization of specific cantharidin binding
sites in mouse tissues. Mol. Pharmacol. 33, 706-712.
8) Clark, B., Furlong, J . W., Ladner, A., and Slovak, A. J . M. (1984)
Dermal toxicity of dimethyl acetylene dicarboxylate, N-methyl
pyrrolidone, triethylene glycol dimethyl ether, dioxane and te-
tralin in the rat. IRCS Med. Sci. 12, 296-297.
9) Klain, G. J ., Bonner, S. J ., and Bell, W. G. (1988) Metabolic
alterations induced by topical dimethylacetylenedicarboxylate.
Fundam. Appl. Toxicol. 10, 730-735.
F igu r e 4. Retro-Diels-Alder-type reaction of dimethyl 7-oxa-
bicyclo[2.2.1]hepta-2(3),5(6)-diene-2,3-dicarboxylate with a thiol-
containing protein and addition of GSH to the bicyclic adduct
and the liberated dimethyl acetylenedicarboxylate. The initial
reaction with BSA, observed as substrate loss (and furan
liberation in the case of the bis(methyl and bis(ethyl carboxy-
lates)), also occurs for bicyclic adducts with 2,3-substituents of
bis(ethyl carboxylate) and diethyl phosphonate/ethyl carboxy-
late, but not of dicarboxylic acid, bis(tert-butyl carboxylate) and
bis(dimethyl phosphonate). The purported acetylene intermedi-
ates are not observed in reactions with BSA, presumably
because they undergo further reaction faster than they are
formed.
(
(
10) Huntress, E. H., Lesslie, T. E., and Bornstein, J . (1963) Dimethyl
acetylenedicarboxylate. Org. Synth. Coll. Vol. 4, 329-330.
11) Hall, R. G., and Trippett, S. (1982) The preparation and Diels-
Alder reactivity of ethyl (diethoxyphosphinyl)propynoate. Tetra-
hedron Lett. 23, 2603-2604.
(
(
(
(
12) Seyferth, D., and Paetsch, J . D. H. (1969) Tetramethyl acetylene-
diphosphonate and dimethyl chloroacetylenephosphonate and
their reactions with cyclopentadiene, 1,3-cyclohexadiene, and
diazomethane. J . Org. Chem. 34, 1483-1484.
13) Diels, O., and Alder, K. (1931) Synthesen in der hydroaromatis-
chen Reihe; XII. Mitteilung. “Dien-Synthesen” sauerstoffhaltiger
Heteroringe.2.Dien-Synthesen des Furans. Ann. Chem. 490, 243-
257.
those for dimethyl acetylenedicarboxylate, including but
not restricted to acetyl-CoA carboxylase and glucose-6-
phosphate dehydrogenase (9). We propose that the
toxicity of the oxabicycloheptadienedicarboxylates in-
volves metabolic activation with liberation of the highly
reactive acetylenedicarboxylates. This is the first case,
to our knowledge, of a bioactivation mechanism involving
a retro-Diels-Alder-type reaction (Figure 4).
14) Casida, J . E., Baron, R. L., Eto, M., and Engel, J . L. (1963)
Potentiation and neurotoxicity induced by certain organophos-
phates. Biochem. Pharmacol. 12, 73-83.
(15) Segall, Y., Kimmel, E. C., Dohn, D. R., and Casida, J . E. (1985)
-Substituted 2-halopropenals: mutagenicity, detoxification and
3
formation from 3-substituted 2,3-dihalopropanal promutagens.
Mutat. Res. 158, 61-68.
(
16) Yoneyama, K., Sekido, S., and Misato, T. (1978) Antagonistic
mechanism of sulfhydryl compounds on cellocidin activity. J .
Antibiot. 31, 1065-1066.
Ack n ow led gm en t. The project described was sup-
ported by Grant PO1 ES00049 from the National Insti-
tute of Environmental Health Sciences, NIH. We thank
our laboratory colleagues Susan Sparks for mouse toxic-
(
17) Tomigahara, Y., Shiba, K., Isobe, N., Kaneko, H., Nakatsuka, I.,
and Yamada, H. (1994) Identification of two new types of S-linked
conjugates of Etoc in rat. Xenobiotica 24, 839-852.
3
ity assays and Charles Laidley for the [ H]cantharidin
(18) J ocelyn, P. C. (1972) Biochemistry of the SH Group, Academic
Press, London.
binding studies.
(
(
(
19) Eto, M., Oshima, Y., and Casida, J . E. (1967) Plasma albumin as
a catalyst in cyclization of diaryl o-(R-hydroxyl)tolyl phosphates.
Biochem. Pharmacol. 16, 295-308.
Refer en ces
20) Casida, J . E., and Augustinsson, K.-B. (1959) Reaction of plasma
albumin with 1-naphthyl N-methylcarbamate and certain other
esters. Biochim. Biophys. Acta 36, 411-426.
(
1) Tomlin, C., Ed. (1994) The Pesticide Manual: A World Compen-
dium, 10th ed., pp 390-391, British Crop Protection Council, The
Bath Press, Bath, U.K.
21) Fruetel, J . A., Sparks, S. E., Quistad, G. B., and Casida, J . E.
(
2) Matsuzawa, M., Graziano, M. J ., and Casida, J . E. (1987)
Endothal and cantharidin analogues: relation of structure to
herbicidal activity and mammalian toxicity. J . Agric. Food Chem.
(
1994) Adducts of dienochlor miticide with glutathione, glu-
tathione S-transferases, and hemoglobins. Chem. Res. Toxicol. 7,
87-494.
4
3
5, 823-829.
(3) Kawamura, N., Li, Y.-M., Engel, J . L., Dauben, W. G., and Casida,
TX950127F
+
+