Formation, solvolysis, and transcarbamoylation reactions of bis(S- glutathionyl) adducts of 2,4- and 2,6-diisocyanatotoluene

Article abstract of DOI:10.1021/tx960201+

During our ongoing studies of the reactions of toluene diisocyanate (2,4- and 2,6-diisocyanatotoluene, TDI) in vivo, it became apparent that reactive form(s) of these diisocyanates reach(es) the circulatory system after passage through the respiratory system. Based on recent work by others regarding the transcarbamoylation reactions of monoisocyanates, we hypothesized that the reactive form could be masked as an S- thiocarbamoylglutathione adduct of one or more of the isocyanato moieties. In this study, the glutathione adducts of 2,4- and 2,6-diisocyanatotoluene were synthesized under physiological conditions. Bis adducts were the major products when near-equimolar amounts of glutathione and the individual diisocyanato compounds were mixed at physiological pH, and were formed in high yield. Little to no mono adducts formed under these reaction conditions. The masses of the bis adducts were confirmed by electrospray mass spectrometry (MS), and ¹H NMR analysis strongly suggested that the thiol of the cysteine residue of glutathione was the nucleophile in each case. The rates of solvolysis of the two bis adducts in aqueous buffer under conditions of physiological temperature and pH were determined, and electrospray MS analysis showed that the corresponding mono(glutathionyl)-TDIs were formed in these reactions. Incubation in vitro of each of the bis(glutathionyl)-TDI adducts with a 12 amino acid peptide (Thr-Cys-Val-Glu-Trp-Leu-Arg-Arg-Tyr- Leu-Lys-Asn) at pH 7.5 resulted in transfer of one mono(glutathionyl)- toluylisocyanato moiety to the peptide as detected by HPLC and on-line electrospray MS analyses. In both the solvolysis and transfer experiments, the 2,4-TDI-derived bis(glutathionyl) adduct reacted most quickly, while both the bis(glutathionyl)-2,6-TDI adduct and its transfer product with the peptide were more stable than their 2,4-TDI-derived counterparts. The results indicate high stoichiometry in formation and ready transfer to nucleophilic sites of protein, and suggest that the isocyanato moiety of both 2,4- and 2,6-TDI may be regenerated in vivo from their bis(glutathionyl) adducts. As a consequence, the thiol status of particular tissues may be a contributing factor to individual TDI toxicity susceptibility, and a mechanism by which toxicity at sites distant to the initial point of contact may be proposed.

Full text of DOI:10.1021/tx960201+

424
Chem. Res. Toxicol. 1997, 10, 424-431
F or m a tion , Solvolysis, a n d Tr a n sca r ba m oyla tion
Rea ction s of Bis(S-glu ta th ion yl) Ad d u cts of 2,4- a n d
2,6-Diisocya n a totolu en e
Billy W. Day,†,‡,§ Ruhzi J in,† Dina M. Basalyga,† J ean A. Kramarik,† and
Meryl H. Karol*,†,§
Department of Environmental & Occupational Health, Department of Pharmaceutical Sciences, and
University of Pittsburgh Cancer Institute, University of Pittsburgh, Pittsburgh, Pennsylvania 15238
Received December 11, 1996X
During our ongoing studies of the reactions of toluene diisocyanate (2,4- and 2,6-diisocy-
anatotoluene, TDI) in vivo, it became apparent that reactive form(s) of these diisocyanates
reach(es) the circulatory system after passage through the respiratory system. Based on recent
work by others regarding the transcarbamoylation reactions of monoisocyanates, we hypoth-
esized that the reactive form could be masked as an S-thiocarbamoylglutathione adduct of
one or more of the isocyanato moieties. In this study, the glutathione adducts of 2,4- and
2,6-diisocyanatotoluene were synthesized under physiological conditions. Bis adducts were
the major products when near-equimolar amounts of glutathione and the individual diisocy-
anato compounds were mixed at physiological pH, and were formed in high yield. Little to no
mono adducts formed under these reaction conditions. The masses of the bis adducts were
1
confirmed by electrospray mass spectrometry (MS), and H NMR analysis strongly suggested
that the thiol of the cysteine residue of glutathione was the nucleophile in each case. The
rates of solvolysis of the two bis adducts in aqueous buffer under conditions of physiological
temperature and pH were determined, and electrospray MS analysis showed that the
corresponding mono(glutathionyl)-TDIs were formed in these reactions. Incubation in vitro
of each of the bis(glutathionyl)-TDI adducts with a 12 amino acid peptide (Thr-Cys-Val-Glu-
Trp-Leu-Arg-Arg-Tyr-Leu-Lys-Asn) at pH 7.5 resulted in transfer of one mono(glutathionyl)-
toluylisocyanato moiety to the peptide as detected by HPLC and on-line electrospray MS
analyses. In both the solvolysis and transfer experiments, the 2,4-TDI-derived bis(glutathionyl)
adduct reacted most quickly, while both the bis(glutathionyl)-2,6-TDI adduct and its transfer
product with the peptide were more stable than their 2,4-TDI-derived counterparts. The results
indicate high stoichiometry in formation and ready transfer to nucleophilic sites of protein,
and suggest that the isocyanato moiety of both 2,4- and 2,6-TDI may be regenerated in vivo
from their bis(glutathionyl) adducts. As a consequence, the thiol status of particular tissues
may be a contributing factor to individual TDI toxicity susceptibility, and a mechanism by
which toxicity at sites distant to the initial point of contact may be proposed.
Studies of guinea pigs exposed to TDI vapors have shown
formation of adducts localized predominantly to the
apical surface of the respiratory tract epithelium (3), as
well as adduction of soluble proteins in bronchoalveolar
lavage (BAL) fluid (4). We also identified TDI-adducted
hemoglobin, both in BAL and in intact erythrocytes in
peripheral blood (5). The mechanism by which these
highly reactive diisocyanates are transported across the
epithelial layer of the respiratory tract, into the blood,
and through the erythrocyte membrane to react with
hemoglobin is unknown, but the presence of carbamoyl-
ated hemoglobin implies that a reactive form of TDI
survives the transport process. Studies by others on the
biochemistry of monoisocyanato compounds have shown
that, after carbamoylation of the cysteine sulfur of
glutathione (GSH), subsequent transfer of the isocyanato
moiety to protein thiols can occur (6,7). Thus, GSH
adducts of isocyanato compounds may serve as masked,
reactive forms of isocyanates, and may contribute to
toxicities at sites other than that of original absorption
into the body (8).
In tr od u ction
Diisocyanatotoluene (TDI)1 is the most prevalent cause
of occupational asthma in the Western world (1). Indi-
viduals are most often exposed to vapors of the 4:1 molar
mixture of the 2,4-isomer (1; see Scheme 1) and the 2,6-
isomer (2), and 5-10% develop respiratory hypersensitiv-
ity to the chemical (2).
In order to understand the mechanism(s) by which TDI
exerts its asthmogenic effects, we have been performing
studies to determine the tissues and soluble proteins of
the respiratory tract with which these isocyanates react.
* To whom correspondence should be addressed at the Department
of Environmental & Occupational Health, Center for Environmental
& Occupational Health & Toxicology, University of Pittsburgh, 260
Kappa Dr., Pittsburgh, PA 15238. Telephone: (412) 967-6530. FAX:
(412) 967-6611. E-mail: mhk@vms.cis.pitt.edu.
† Department of Environmental & Occupational Health.
‡ Department of Pharmaceutical Sciences.
§ University of Pittsburgh Cancer Institute.
X Abstract published in Advance ACS Abstracts, March 15, 1997.
1
Abbreviations: BAL, bronchoalveolar lavage; BCNU, N,N′-bis(2-
chloroethyl)-N-nitrosourea; CEIC, 2-chloroethyl isocyanate; DAD, diode
array detector; GSH, reduced glutathione; HBSS, Hank’s balanced salts
solution; MIC, methyl isocyanate; RADS, reactive airways disorder
syndrome; tR, retention time; TDI, toluene diisocyanate.
In the current study, we investigated the reactions of
TDI with GSH to determine if reasonably stable, mono-
meric or dimeric thiocarbamoyl adducts would form with
S0893-228x(96)00201-9 CCC: $14.00 © 1997 American Chemical Society

Toluene Diisocyanate-Glutathione Adducts
Chem. Res. Toxicol., Vol. 10, No. 4, 1997 425
Sch em e 1
mmol) dissolved in 10 mL of 0.1 M NH4HCO3 (final pH 7.73).
Neat 1 or 2 (175 µL, 1.2 mmol each) was added to individual
flasks containing the GSH solutions and shaken at ambient
temperature for 10 min, and then centrifuged at 1000 rpm for
10 min. Residual unreacted TDIs were removed (droplets on
the upper surfaces of the solutions) by pipet; the remaining
solutions were again centrifuged to pellet the small amount of
white precipitate that formed in each reaction. The superna-
tants were decanted to borosilicate test tubes and lyophilized
in a Speedvac, in each case yielding a white powder. The yield
of 3 was 355 mg (68%), while that of 8 was 345 mg (66%). Each
of the bis adducts yielded both 1H NMR (Varian XL-200 and
Bruker AM500 spectrometers) and pneumatically-assisted elec-
trospray MS (vide infra) spectral data consistent with the
proposed structures (data shown under Results). Reactions
performed at 37 °C gave identical results.
the tripeptide, and if these would transfer isocyanato
equivalents to other molecules. We found that bis(S-
glutathionyl) adducts of 2,4- and 2,6-TDI readily form
without enzymatic assistance under physiological condi-
tions (3 and 8; see Scheme 1), and that their rates and
mechanism of solvolysis allow transfer of one mono(glu-
tathionyl)toluylisocyanato moiety to a different cysteine-
containing peptide.
Exp er im en ta l P r oced u r es
Ca u tion : TDI is a skin and mucus membrane irritant and
should be handled in a chemical vapor hood with appropriate
protective equipment. Its hydrolysis products are aromatic
amines, potential human carcinogens.
Ma ter ia ls. GSH and Thr-Cys-Val-Glu-Trp-Leu-Arg-Arg-Tyr-
Leu-Lys-Asn (MHC antigen H-2Kb) were purchased from Sigma
Chemical Co. (St. Louis, MO). 2,4- and 2,6-TDI were gifts from
Bayer Chemical Corp. (Pittsburgh, PA). Trifluoroacetic acid
(sequencing grade), C2H3O2H, and 2H2O were purchased from
Aldrich Chemical Co. (Milwaukee, WI). Solvents of the highest
available purity were obtained from Fisher Scientific (Pitts-
burgh, PA).
HP LC An a lyses. Analytical HPLC was performed on a
Hewlett Packard 1090 Series II LC equipped with a Hewlett
Packard 1040 UV-vis diode array detector (DAD), both con-
trolled by a Hewlett Packard Series 300 Chemstation computer.
All separations were performed on a 10 µm particle size, 250 ×
4 mm, Hibar LiChrosorb C18 column (E. Merck Darmstadt,
Germany) using a 2 mL/min flow rate of the following mobile
phase and gradient: solvent A, H2O containing 0.1% (v/v)
CF3CO2H; solvent B, CH3CN containing 0.1% (v/v) CF3CO2H;
0-7 min, 16% B; 7-30 min, linear change to 100% B; 30-35
Syn th esis of BisGS-2,4-TDI a n d BisGS-2,6-TDI Ad -
d u cts. Two flasks were each charged with GSH (307 mg, 1

426 Chem. Res. Toxicol., Vol. 10, No. 4, 1997
Day et al.
min, linear return to 16% B. Column effluent was monitored
at 250 nm, and spectra over the range of 200-400 nm were
collected every 1.28 s.
in CH3CN-H2O in both acidic (apparent pH 2) and near-
neutral (apparent pH 6.8) conditions. HPLC analysis
indicated formation of over 30 products in each case. The
early eluting products were characterized by ionspray MS
(vide infra). Yields of these products were low (<20%
overall), and were accompanied by a large array of side-
products of limited aqueous solubilities. By virtue of
their hydrophobicity in C18 HPLC analyses, insolubility
in aqueous systems, and multiple high quasi-molecular
ion signals (i.e., >1500 amu) in MS experiments (data
not shown), we concluded they were various polymers of
the reactants.
We then performed the syntheses in aqueous bicarbon-
ate, pH 7.7, to simulate reactions that might occur in the
lung. The diisocyanates 1 and 2 were individually added,
neat, to bicarbonate solutions containing 0.8 molar equiv
of GSH. Rapid reactions ensued, each yielding es-
sentially one product. Reactions at 25 °C and at 37 °C
gave product profiles and yields that were indistinguish-
able. No attempts were made to optimize yields.
Ra tes of BisGS-2,4-TDI a n d BisGS-2,6-TDI Solvolysis.
The kinetics of the solvolyses of 3 and 8 in aqueous medium
(see Scheme 1) were determined as follows. Compounds 3 and
8 (4.55 mg each) were separately dissolved in 1 mL of Hank’s
balanced salts solution (HBSS), and the resulting solutions were
maintained at constant temperature (25 or 37 °C) in a jacketed
water bath. At selected time points, 20 µL of each sample
solution was removed and analyzed by HPLC and then by
electrospray MS when necessary (vide infra). The integral
under each chromatographic peak was converted to its percent-
age of all peaks in the chromatogram, and the molar concentra-
tion that this percentage represented was calculated from the
initial concentration of the bis adduct. The compounds were
thus determined to convert first to the intermediate 2-GS-
thiocarbamoyl-4(or 6)-isocyanato adducts (4 and 9, respectively),
followed by hydrolysis to the 2-GS-thiocarbamoyl-4(or 6)-amino
adducts (5 and 10, respectively).
Ma ss Sp ectr om etr y. A Perkin Elmer/Sciex API I mass
spectrometer with an atmospheric pressure ionization source
and an articulated IonSpray interface linked in tandem with
glass capillary tubing to the Hewlett-Packard HPLC/DAD
system was employed to determine the molecular masses of 3
and 8 and decomposition products (4-7 and 9-12, respectively).
The collected and lyophilized LC fractions were redissolved in
25-50 µL of 1:1 H2O-CH3CN containing 0.1% CF3CO2H, 5-20
µL of which was injected directly through the LC without a
column and introduced into the ion source of the mass spec-
trometer in 1:1 H2O-CH3CN containing 0.1% CF3CO2H flowing
at 40 µL/min. High-purity air was used for nebulization at an
operating pressure of 40 psi. High-purity N2 heated to 55 °C
was used as the curtain gas, flowing at 0.6 L/min. The ionspray
interface was maintained at 5 kV and the orifice voltage at 70
V. The quadrupole was scanned over the required range (i.e.,
gm/ z 70 to em/ z 2400) in 9-11 s per scan at a resolution of
m/ z 0.1. Twenty microliters of the crude time ) ∼0 and time
) 3 day solutions of 3 and 8 prepared at 4.55 mg/mL in HBSS
was also separately injected into the flowing 1:1 H2O-CH3CN/
CF3CO2H stream described above and analyzed by MS.
Tr an scar bam oylation Reaction s. The MHC antigen H-2Kb
peptide (0.5 mg, 316 nmol) was incubated in two separate
reactions with 10-fold molar excesses of 3 and 8 (2.5 mg, 3.16
µmol). Each reaction was carried out in 0.5 mL of 0.1 M
NH4HCO3 (final pH 7.73) at 25 °C for 24 h. Aliquots (15 µL) of
the reaction mixtures were withdrawn at given time points and
immediately stored at -70 °C until analyzed. For “time zero”
determinations, 1 mM solutions of 3, 8, GSH, and the peptide
were individually prepared in 0.1% aqueous CF3CO2H. The
appropriate mixtures of three components were prepared, or the
reaction aliquots were thawed and immediately injected for
HPLC-DAD/UV-MS analysis. GSH was detected with a moni-
toring wavelength of 214 nm, toluyl-containing materials at 250
nm, and the 12-mer peptide and any products containing it at
280 nm. A 5 µL portion of each of the reaction mixture aliquots
was analyzed in each HPLC-DAD/UV-MS run, and the vial
containing the remainder of each aliquot was immediately
stored at -70 °C. The column, flow rate, and MS parameters
used in the analysis of the solvolysis reactions were employed
with the following changes: for the LC gradient, time ) 0-15
min isocratic, 86% solvent A; linear change to 30 min, 10%
solvent A; linear change to 35 min, 86% solvent A; for the mass
spectrometer, N2 curtain gas flow was set at 0.65 L/min and
the quadrupole was scanned from m/ z 250 to m/ z 2300 in 10.29
s/scan at a resolution of m/ z 0.1.
The identity of each of the major products was deter-
mined by electrospray mass spectrometry and proton
NMR spectroscopy (Figure 1). Each of the two major
products yielded ion signals of m/ z 789, consistent with
[M+H]+ quasi-molecular ions of bis(glutathionyl) addition
products of the diisocyanato compounds (compounds 3
and 8 in Scheme 1). The proton NMR spectra of these
adducts exhibited identical signals for the two attached
glutathionyl moieties, and the aromatic signals gave no
indication of unsymmetrical substitution. Further, the
signals for the cysteinyl protons were shifted downfield
relative to unsubstituted GSH, while the signals for the
remaining glutathione-derived protons were essentially
unchanged. Because of the masses detected, the sym-
metries evident in the NMR spectra, and the lowered
energies of the cysteinyl proton signals, it was clear that
the products were the bis(thiocarbamoyl) structures
assigned to 3 and 8 as shown in Scheme 1.
Solvolysis of BisGS-2,4-TDI a n d BisGS-2,6-TDI
Ad d u cts. The solvolyses of 3 and 8 in a medium
mimicking physiological pH and ion strength (Hank’s
balanced salts solution, no pH indicator) were followed
over time by HPLC-DAD/UV-electrospray MS analyses.
The reactions of the bis adducts followed the paths shown
in Scheme 1 when incubated at 25 or 37 °C. Example
HPLC chromatograms obtained during the course of
these incubations are shown in Figure 2, and the kinetics
of the solvolysis reactions are depicted in Figure 3.
Solvolysis of the 2,4-isomer 3 was more rapid than that
of the 2,6-positional isomer 8. Each reaction followed
pseudo-first-order kinetics in the initial phases of the
reactions. Over the entire course of the reactions, disap-
pearance of 3 and 8 followed apparent pseudo-second-
order kinetics. For the 2,4-isomer 3, the instantaneous
first-order rate constant kapp,unimolecular at 25 °C was 4.7 ×
10-5 s-1. Overall, solvolysis of 3 at 25 °C followed second-
order kinetics with a kapp,bimolecular of 0.013 M-1 s-1. For
8, kapp,unimolecular at 25 °C was 4.7 × 10-6 s-1, while
kapp,bimolecular was 1.2 × 10-3 M-1 s-1. The rate of conver-
sion of 3 was accelerated by the studied increase in
temperature to 37 °C, yielding a kapp,unimolecular of 7.4 ×
10-5 s-1 and a kapp,bimolecular of 0.022 M-1 s-1
.
Resu lts
Ma ss Sp ectr om etr ic An a lysis of Solvolysis P r od -
u cts. HPLC fractions from solvolysis reactions were
collected and analyzed individually by direct injection
ionspray mass spectrometry. Aliquots from the later
time points of the reaction mixtures were also analyzed
Syn th esis of BisGS-TDI Ad d u cts. We initially
attempted the syntheses of GS-TDI adducts with a
method described useful for preparation of glutathione-
monoisocyanate adducts (6). Reactions were performed

Toluene Diisocyanate-Glutathione Adducts
Chem. Res. Toxicol., Vol. 10, No. 4, 1997 427
F igu r e 1. Proton NMR spectra of 3 (A) and 8 (B). Spectra were recorded at ambient temperature at 500.121 MHz in 9:1:0.009
(v/v/v) H2O-2H2O-CF3CO2H. Similar shifts and coupling were observed when spectra were recorded at 200.057 MHz in 199:1 (v/v)
2H2O-CF3CO2H, and at 200.058 MHz in C2H3O2H using standard decoupling (solvent suppression and assignment of coupling
constants) routines. Shifts and multiplicities for the Cys R signals were determined to be δ 4.58 (dd, 2H, J ) 5.3 and 2.9 Hz) for 3
and δ 4.54 (dd, 2H, J ) 5.4 and 3.1 Hz) for 8 from the 200 MHz spectra obtained in C2H3O2H.
by on-line LC-MS to verify results from the collected
fractions. As stated, the products from the synthetic
reactions performed in aqueous-only medium consisted
almost entirely of the bis adducts 3 and 8. Our earlier
but aborted attempts at synthesis in mixed organic-
aqueous media had given us indications of other possible
structures, including GSH, that could be formed during
solvolysis (i.e., 4-7 and 9-12). Each of these products
was detected over the course of the solvolysis reactions
as their [M+H]+ (4, 5, 9, 10, GSH), [M+Na]+ (4, 6, 9,
11), [M+NH4]+ (5, 10), and/or [M+2H]2+ (7, 12) quasi-
molecular ions. We found no firm indications of higher
order products (i.e., arylamino group addition to a freed
isocyanato group) in these reactions, although there were
a few minor but unexplained hydrophobic products noted
during the solvolysis of 8 (see Figure 2B).

428 Chem. Res. Toxicol., Vol. 10, No. 4, 1997
Day et al.
F igu r e 3. Time dependencies of the solvolyses of the bis(glu-
tathionyl) adducts 3 and 8 (closed symbols) and formation of
their corresponding mono(glutathionyl) mono(isocyanato) prod-
ucts 4 and 9 (open symbols) in a physiological bicarbonate buffer
(Hank’s balanced salts solution): circles, 8 f 9 at 25 °C;
triangles, 3 f 4 at 25 °C; diamonds, 3 f 4 at 37 °C.
8 under physiological conditions (Scheme 2). Reactions
of the 2,4-isomer 3 will be discussed first.
The MHC H-2Kb peptide was detected by LC-electro-
spray-MS as its doubly charged quasi-molecular ion
[retention time (tR) ca. 27 min, m/ z 791.5, [M+2H]2+],
while GSH (tR 4 min, m/ z 308), 3 (tR ) 14 min, m/ z 789),
and the intermediate mono(glutathionyl) mono isocyan-
ate 4 (tR 8 min, m/ z 482) were detected as their [M+H]+
ions. At time zero, the bis adduct and the MHC peptide
were evident. After 15 min, GSH, 3, and the unmodified
MHC peptide were detected. The transcarbamoylation
product of 3 [GS-2,4-TDI-peptide (13), tR 25 min]
appeared at the 30 min time point as its doubly charged
quasi-molecular ion [m/ z 1032, [M+2H]2+], as did 4.
GSH, 3, and the unreacted peptide were also evident at
this time. All five of these analytes were detected up to
and including the 2 h time point of the reaction. After 6
h of reaction, neither 3 nor 4 was detected, but GSH, 13,
and the unmodified MHC peptide were still evident. The
latter three analytes were all detected at the 12 and 24
h time points, but the signal of 13 showed a time-
dependent decay in intensity. An example chromatogram
after 30 min of reaction is shown in Figure 4A.
The reaction between 8 and the MHC peptide under
the same conditions was analyzed as above. Under these
conditions, compound 8 (tR 7 min) eluted more quickly
than did 3. This was also the case with its mono(glu-
tathionyl) mono(isocyanato) intermediate solvolysis prod-
uct 9 (tR 5 min). From the 15 min time point up to and
including the 2 h time point of this reaction, GSH, 8, 9,
and the unmodified 12-mer peptide were detected. After
6 h of reaction, 9 was not detected, but GSH, 8, the 12-
mer peptide, and the transcarbamoylation product GS-
2,6-TDI-peptide (14) (tR 25 min; again, m/ z 1032,
[M+2H]2+) were detected. At the 12 and 24 h points,
these four analytes were all detected, with the signal for
8 showing a time-dependent decay in intensity, but the
signal for the transcarbamoylation product 14 remained
relatively constant. An example chromatogram after 12
h of reaction is shown in Figure 4B.
F igu r e 2. Example HPLC chromatograms during the course
of solvolyses of 3 and 8 at 25 °C in a physiological bicarbonate
buffer (Hank’s balanced salts solution). Peaks are labeled
according to Scheme 1, and the principal cationized molecular
ions noted for each analyte in pneumatically-assisted electro-
spray experiments are listed.
Tr a n sca r ba m oyla tion by BisGS-TDI Ad d u cts.
The appearance of 4, 6, 9, and 11 during incubation of 3
and 8 in physiological medium suggested that glu-
tathione-TDI adducts could act as masked but reactive
intermediates, and might serve as protoxins capable of
transferring isocyanato equivalents to other biological
molecules. A commercially available major histocompat-
ibility complex-derived 12 amino acid peptide (Thr-Cys-
Val-Glu-Trp-Leu-Arg-Arg-Tyr-Leu-Lys-Asn, MHC anti-
gen H-2Kb) containing one free cysteine sulfhydryl was
incubated separately with 10-fold molar excesses of 3 and
Note that the structure of 4 in Scheme 1 was originally
assumed based on the more rapid solvolysis of 3 vs that
of 8. The more rapid transfer reaction of 3 vs that of 8
further supports this assumption, as well as the assumed
structure 13 given in Scheme 2.

Toluene Diisocyanate-Glutathione Adducts
Chem. Res. Toxicol., Vol. 10, No. 4, 1997 429
Sch em e 2
in guinea pigs exposed to vapors of 14C-TDI (12). In a
guinea pig model of TDI asthma, TDI adducts have been
found localized to the apical surface of the upper respira-
tory tract epithelium by immunohistochemical methods
(3). Additionally, TDI-adducted proteins can be isolated
from the BAL (4) and TDI-adducted hemoglobin from
circulating erythrocytes of such guinea pigs (5). The
latter finding suggests that a reactive form of TDI
reaches the circulation and penetrates the erythrocyte
membrane to carbamoylate hemoglobin. The present
study sought to investigate a potential mechanism by
which such a reaction could occur by addressing the
hypothesis that, under physiologic conditions, TDI may
react reversibly with sulfhydryl moieties, and, in the
presence of appropriate nucleophiles, the product(s) of
such reactions could react further to yield transcarbam-
oylation products.
Glutathione is the major intracellular non-protein thiol
that is protective against reactive electrophiles. Its
concentrations in epithelial lining fluid can reach 10 mM
(13). Due to its free sulfhydryl group, GSH may serve
as a nucleophile, an activity facilitated by enzymatic
catalysis of glutathione transferases. Conjugation of
electrophilic moieties with GSH is generally considered
a detoxification process. Such S-linked adducts are
excreted, or are transformed into the corresponding
cysteine or N-acetylcysteine conjugates before elimination
(14). If the conjugation process is reversible, however,
such adducts may be transformed back to the original
electrophilic species, and thus remain toxic (15). From
the current study, we propose that reaction of glu-
tathione-TDI adducts also involves subsequent libera-
tion of the original electrophile; this is a consequence of
the relative lability of the thiocarbamoyl bond linking the
alkyl sulfur to the isocyanato carbon (6, 7, 14-16).
Reactions of monoisocyanates with GSH, N-acetylcys-
teine, and cysteine have been studied. Under physiologic
conditions, methyl isocyanate (MIC) reacts with GSH to
form S-(N-methylcarbamoyl)glutathione (7). Rats in-
jected ip with MIC excrete large amounts of the carbam-
oylated GSH adduct in the bile (7) and its N-acetylcys-
teinyl transformation product in the urine (17), with the
latter accounting for 25% of the administered MIC.
F igu r e 4. Example HPLC chromatograms obtained during the
transcarbamoylation reactions of the MHC antigen H-2Kb
peptide Thr-Cys-Val-Glu-Trp-Leu-Arg-Arg-Tyr-Leu-Lys-Asn with
3 (A, 30 min of reaction) and 8 (B, 12 h of reaction) in NH4CO3-
buffered medium at 25 °C. Peaks are labeled with the analyte
contained therein and with the base ion in the pseudo-molecular
ion clusters detected in pneumatically-assisted electrospray
experiments.
Discu ssion
TDI has been associated with several pulmonary
disorders, including sensory and pulmonary irritation (9),
reactive airways disorder syndrome (RADS) (10), hyper-
sensitivity pneumonitis (11), and asthma (1). Although
there is evidence that each disorder is dependent on the
TDI exposure concentration and, for some individuals,
on the frequency of exposure, the molecular mechanisms
involved in each of these syndromes remain largely
unknown.
The initial step in chemical sensitization is believed
to be reaction of the chemical with an endogenous protein
(1). Animal models have been employed to identify the
macromolecules with which TDI reacts following inhala-
tion exposure. There is evidence that laminin is adducted
The current study demonstrated rapid reaction of GSH
with TDI under physiological conditions, possibly favored
by the carbonate-buffered medium in which reactions
were performed (18). The reactions did not require
enzymatic assistance from glutathione transferase. A

430 Chem. Res. Toxicol., Vol. 10, No. 4, 1997
Day et al.
similar situation has been found with MIC; the resulting
MIC carbamate thioester adducts, however, decompose
rapidly (7). We observed extremely rapid formation but
comparatively slower solvolysis of and transcarbamoyl-
ation by the TDI-GSH adducts as compared to the rates
observed by others for MIC-GSH adducts. Additionally,
the 2,4 diisocyanato adduct 3 solvolyzed more rapidly
(about 10-fold) than the 2,6 adduct 8. The higher
reactivity of 3 is no doubt due to the steric availability
of the 4-position of the toluyl moiety, and adds credence
to our hypothesis shown in Schemes 1 and 2 that the
thiocarbamoyl system at this position solvolyzes before
its counterpart at the 2-position.
cysteine-containing, MHC antigen-derived peptide as a
model target for such transfers (20). The proteins
involved in the initial stages of sensitization to TDI are
not known, but are believed to engage in antigen pre-
sentation to appropriate T lymphocytes (21). We noted
that the bis adducts transcarbamoylated this peptide
under physiological conditions, with the bis 2,4-TDI-
GSH adduct 3 demonstrating a greater rate of transfer
relative to the bis 2,6-TDI-GSH adduct 8. These results
suggest a potential relevance of the thiocarbamoylation
reaction in human sensitization to TDI.
Glutathione conjugation may offer protection against
isocyanate toxicity, or may enhance the toxicity of such
xenobiotics. Detection of transcarbamoylation reactions
of TDI suggests that regeneration of the reactive elec-
trophilic isocyanato functionality may occur in the pres-
ence of an appropriate second nucleophile. Precedence
for such activity has been shown with N,N′-bis(2-chloro-
ethyl)-N-nitrosourea (BCNU), an antitumor agent that
also possesses toxic properties. BCNU degrades in
aqueous solution to yield reactive alkylating intermedi-
ates that are at least partly responsible for its antitumor
activity, and to carbamoylating moieties, notably 2-chlo-
roethyl isocyanate (CEIC), which may be the mediator
of some of BCNU’s toxicity. Davis et al. (6) identified
the GSH conjugate of CEIC in rats following ip injection
of BCNU, suggesting that these thiocarbamoyl adducts
contribute to the toxicity of the parent drug. With MIC,
survivors of the catastrophic accident in Bhopal, India,
in 1984 have displayed numerous toxicities even though
the chemical exposure was likely restricted to an inhala-
tion route. The reports of cardiovascular, gastrointesti-
nal, and reproductive effects in the Bhopal victims may
well be ascribable to transcarbamoylation of distant
proteins. Our detection of the transcarbamoylation reac-
tions of 3 and 8 with the MHC antigen H-2Kb peptide
suggests that the toxicity of TDI can be regenerated in
the presence of appropriate nucleophiles, and may occur
at sites distant to the original location of TDI exposure.
It must also be considered that enhanced toxicity may
result simply from the rapid reaction of GSH with TDI;
intracellularly, such reaction would immediately result
in a lowering of cellular thiol levels. GSH protects cells
not only from electrophiles but also from damaging free
radicals by its intimate involvement in the maintenance
of cellular redox balance. A consequence of TDI exposure
may thus be toxicity not only due to the isocyanate
functionality but also as a result of reduced levels of GSH
and the consequent oxidative stress. Recent work has
also suggested a more insidious action that is less
dependent on thiol isocyanate stoichiometry: the isocy-
anato and diisocyanato GSH adducts may, via the site-
directing glutathionyl moiety, irreversibly inactivate
enzymes in the GSH-regenerating pathway (16, 22).
The reaction kinetics determined and shown in Figure
3 indicate that the thiocarbamoyl adducts 3 and espe-
cially the 2,6-adduct 8 are stable enough to diffuse from
their site of formation. Several implications can be
drawn from these data. First, the possibility exists that
TDI can be transported as a bis carbamoylated GSH
adduct to sites distant from the initial site of isocyanate
contact. As such, the data offer an explanation for our
previous finding of TDI-adducted hemoglobin inside red
blood cells following inhalation exposure of guinea pigs
to TDI vapor (5). Second, the isocyanato functionality
can be regenerated from the bisGSH adduct, resulting
in carbamoylation of a second nucleophile, and perhaps
To our knowledge, reaction products of diisocyanates
with GSH have not been reported. We observed high
yields of bis adducts 3 and 8 and their slow but efficient
transfer of isocyanato equivalents to the 12-mer peptide
in carbonate-buffered aqueous media. Explanation of the
preferred reaction of isocyanates with GSH and their
subsequent transfer of isocyanato moieties has been
offered as being due to preferential isocyanate reaction
with the soft nucleophile GSH versus reaction with the
hard nucleophile water (hydrolysis) (15, 16). For ex-
ample, preferred reaction of the GSH-MIC adduct S-(N-
methylcarbamoyl)glutathione with cysteine in aqueous
media has been shown (7). In the present case, rapid
and essentially complete reaction of 1 and 2 with GSH
to form only the bis adducts 3 and 8 can be attributed to
the TDIs’ insolubilities in purely aqueous systems. Their
dissolution into aqueous phases requires cosolvent or
physical mixing assistance, and their hydrolysis to di-
amines requires catalytic assistance (e.g., carbonate or
bicarbonate ions) (18). We thus envision that reaction
of 1 and 2 with GSH occurs at the organic-aqueous
interface to form 4 (or, more likely, its 2-isocyanato-4-
thiocarbamoyl isomer) and 9, which then react rapidly
with GSH (as per the hard/soft argument) to form 3 and
8. Our view of this reaction mechanism is strengthened
by the fact that we consistently observed 1 and 2
remaining as water-insoluble droplets at the end of the
synthetic reactions.
Although reactions in vivo and in vitro of monoisocy-
anates with sulfhydryl-containing peptides and proteins
have been reported (7, 16), there is little information on
the protein specificity of such reactions. For example,
incubation of [1-14C]MIC with rat erythrocyte membrane
proteins for 15 min at 37 °C results in a 59% reduction
in the free sulfhydryl content of the membrane proteins,
and MIC administered ip to rats results in a 27%
decrease in the free thiol content of erythrocyte mem-
branes 24 h after treatment (19). Carbamoylation of
proteins in vitro by the S-linked glutathione conjugate
of MIC has been shown by Pearson et al. (7), who found
that Cys-6 of reduced oxytocin is the preferred site of
carbamoylation, with secondary reaction sites at Cys-1
and Tyr-2. Bovine serum albumin, which contains one
free sulfhydryl, has also been shown to be carbamoylated
after incubation with GSH-MIC (7). These findings
indicate that S-conjugates of monoisocyanates can donate
their masked isocyanato moiety to nucleophilic sites on
peptides and proteins and that the preferred target
residue is cysteine and, since such thiocarbamate esters
are formed reversibly under physiological conditions,
suggest that the toxicity of isocyanates may in part be a
result of such carbamoylation processes.
To evaluate the transcarbamoylation potential of the
bisGS-TDI adducts, we selected a 12 amino acid, single

Toluene Diisocyanate-Glutathione Adducts
Chem. Res. Toxicol., Vol. 10, No. 4, 1997 431
(8) Varma, D. R., and Guest, I. (1993) The Bhopal accident and
methyl isocyanate toxicity. J . Toxicol. Environ. Health 40, 513-
529.
(9) Sangha, G. K., and Alarie, Y. (1979) Sensory irritation by toluene
diisocyanate in single and repeated exposures. Toxicol. Appl.
Pharmacol. 50, 533-547.
(10) Brooks, S. M., Weiss, M. A., and Bernstein, I. L. (1985) Reactive
airways dysfunction syndrome (RADS): persistent asthma syn-
drome after high-level irritant exposures. Chest 88, 376-384.
(11) Yoshizawa, Y., Ohtsuka, M., Noguchi, K., Uchida, Y., Suko, M.,
and Hasegawa, S. (1989) Hypersensitivity pneumonitis induced
by toluene diisocyanate: sequelae of continuous exposure. Ann.
Int. Med. 110, 31-34.
(12) Kennedy, A. L., and Brown, W. E. (1992) Isocyanates and lung
disease: experimental approaches to molecular mechanisms. In
Occupational Medicine: State of the Art Reviews, Vol. 7, no. 2,
pp 301-329, (Harley and Belfus, Inc., Philadelphia, PA.
(13) Cantin, A. M., North, S. L., Hubbard, S. C., and Crystal, R. G.
(1987) Normal alveolar epithelial lining fluid contains high levels
of glutathione. J . Appl. Physiol. 63, 152-157.
(14) Baillie, T. A., and Slatter, J . G. (1991) Glutathione: A vehicle for
the transport of chemically reactive metabolites in vivo. Acc.
Chem. Res. 24, 264-270.
(15) Baillie, T. A., and Kassahun K. (1994) Reversibility in glutathione-
conjugate formation. Adv. Pharmacol. 27, 163-181.
(16) Kassahun, K., J ochheim, C. M., and Baillie, T. A. (1994) Effect of
carbamate thioester derivatives of methyl- and 2-chloroethyl
isocyanate on glutathione levels and glutathione reductase activ-
ity in isolated rat hepatocytes. Biochem. Pharmacol. 48, 587-
594.
(17) Slatter, J . G., Rashed, M. S., Pearson, P. G., Han, D. H., and
Baillie, T. A. (1991) Biotransformation of methyl isocyanate in
the rat. Evidence for glutathione conjugation as a major pathway
of metabolism and implications for isocyanate-mediated toxicities.
Chem. Res. Toxicol. 4, 157-161.
causing toxicity and/or the formation of immunogens.
Such transfers would be favored at sites of altered micro-
pH, as the thiocarbamates are most stable at ca. pH 5.
As a consequence, the recognized toxicity of TDI to the
lung not only may result from direct inhalation of TDI
but also may reflect the acidic pH of the respiratory tract
and thus the stability of thiocarbamoyl-TDI adducts in
the lung. Third, the rapid and direct reaction of TDI with
GSH and the transfer of the isocyanate moiety to protein
relate to the possible sites of isocyanate toxicity following
an inhalation exposure, as well as the possibility of
oxidant/antioxidant status contributing to individual
susceptibility to these important industrial chemicals.
Ack n ow led gm en t. This research was supported in
part by PHS Grant ES 05651 from the National Institute
of Environmental Health Sciences. We thank Dr. Fu-
Tyan Lin at the Department of Chemistry for acquiring
the 500 MHz NMR spectra.
Refer en ces
(1) Mapp, C. E., Saetta, M., Maestrelli, P., Di Stefano, A., Chitano,
P., Boschetto, P., Ciaccia, A., and Fabbri, L. M. (1994) Mecha-
nisms and pathology of occupational asthma. Eur. Respir. J . 7,
544-554.
(2) Karol, M. H. (1986) Respiratory effects of inhaled isocyanates.
Crit. Rev. Toxicol. 16, 349-379.
(3) Karol, M. H., J in, R., and Lantz, R. C. (1997) Immunohistochemi-
cal detection of toluene diisocyanate (TDI) adducts in pulmonary
tissue of guinea pigs following inhalation exposure. Inhal. Toxicol.
9, 63-83.
(4) J in, R., Day, B. W., and Karol, M. H. (1993) Toluene diisocyanate
protein adducts in the bronchoalveolar lavage of guinea pigs
exposed to vapors of the chemical. Chem. Res. Toxicol. 6, 906-
912.
(5) Day, B. W., J in, R., and Karol, M. H. (1996) Reactions of toluene
diisocyanate isomers with guinea pig hemoglobin in vitro and in
vivo. Chem. Res. Toxicol. 9, 568-573.
(6) Davis, M. R., Kassahun, K., J ochheim, C. M., Brandt, K. M., and
Baillie, T. A. (1993) Glutathione and N-acetylcysteine conjugates
of 2-chloroethyl isocyanate. Identification as metabolites of N,N′-
bis(2-chloroethyl)-N-nitrosourea in the rat and inhibitory proper-
ties toward glutathione reductase in vitro. Chem. Res. Toxicol. 6,
376-383.
(7) Pearson, P. G., Slatter, J . G., Rashed, M. S., Han, D. H., and
Baillie, T. A. (1991) Carbamoylation of peptides and proteins in
vitro by S-(N-methylcarbamoyl)glutathione and S-(N-methylcar-
bamoyl)cysteine, two electrophilic S-linked conjugates of methyl
isocyanate. Chem. Res. Toxicol. 4, 436-444.
(18) Berode, M., Testa, B., and Savolainen, H. (1991) Bicarbonate-
catalyzed hydrolysis of hexamethylene diisocyanate to 1,6-diami-
nohexane. Toxicol. Lett. 56, 173-178.
(19) Bhattacharya, B. K., Sharma, S. K., and J aiswal, D. K. (1996)
14C]methyl isocyanate to erythrocyte membrane
Binding of [1-
proteins. J . Appl. Toxicol. 16, 137-138.
(20) Schneck, J ., Munitz, T., Coligan, J . E., Maloy, W. L., Magulies,
D. H., and Singer, A. (1989) Inhibition of allorecognition by an
H-2Kb-derived peptide is evidence for a T-cell binding region on
a major histocompatibility complex molecule. Proc. Natl. Acad.
Sci. U.S.A. 86, 8516-8520.
(21) Garssen, J ., Nijkamp, F. P., Van de Vliet, H., and Van Loveren,
H. (1994) A role for cellular immunity in the induction of airway
hyperresponsiveness induced by small molecular weight com-
pounds. Toxicol. Lett. 72, 151-154.
(22) J ochheim, C. M., and Baillie, T. A. (1994) Selective and irrevers-
ible inhibition of glutathione reductase in vitro by carbamate
thioester conjugates of methyl isocyanate. Biochem. Pharmacol.
47, 1197-1206.
TX960201+