Applicability of a modified Edman procedure for measurement of protein adducts: Mechanisms of formation and degradation of phenylthiohydantoins

Article abstract of DOI:10.1021/tx000247+

Adducts to N-terminal valine residues in hemoglobin (Hb) are used for monitoring in vivo doses of electrophiles and are quantitated by means of a modified Edman procedure, the "N-alkyl Edman procedure". In the reaction with pentafluorophenyl isothiocyanate, N-alkylated valines cyclize and detach from the protein as pentafluorophenylthiohydantoins (PFPTHs) much more efficiently than do unsubstituted N-terminal valine residues. The mechanisms of this reaction, and of possible degradation reactions, have been studied with model compounds using phenyl- and pentafluorophenyl isothiocyanate. The rapid cyclization to N-alkylvaline-PTHs occurs as a consequence of the influence of substituents on ring formation. This facilitated cyclization favors a direct attack by the thiocarbamoyl nitrogen atom on valine-C-1, and is also observed to occur slowly at unsubstituted N-terminal valines. Such cyclization is favored in protic solvents. Under alkaline conditions and in the presence of air, hydrolytic and oxidative processes give rise to degradation products. The PTH derivatives of N-alkylvaline are less apt to undergo such reactions than are the corresponding derivatives of unsubstituted valine. We conclude that the presence of an N-substituent exerts a greater influence on the cyclization process than the structure of the amino acid or of the Edman reagent. For adducts of different structures, the method has broad applicability, for which the limits, however, are not yet explored. The knowledge from the studies is valid not only for the N-alkyl Edman procedure, but also, to some extent, for the classical Edman degradation reaction. The oxidative side reaction gave rise to the invention of a novel synthesis route for insertion of nucleophiles at carbon-5 in thiohydantoins. The present investigation provides a basis for the N-alkyl Edman procedure, facilitating new toxicological applications.

Full text of DOI:10.1021/tx000247+

570
Chem. Res. Toxicol. 2002, 15, 570-581
Ap p lica bility of a Mod ified Ed m a n P r oced u r e for
Mea su r em en t of P r otein Ad d u cts: Mech a n ism s of
F or m a tion a n d Degr a d a tion of P h en ylth ioh yd a n toin s
Per Rydberg,† Bjo¨rn Lu¨ning,‡ Carl Axel Wachtmeister,† Lars Eriksson,§ and
Margareta To¨rnqvist*,†
Department of Environmental Chemistry, Department of Organic Chemistry, and Department of
Structural Chemistry, Arrhenius Laboratory, Stockholm University, S-106 91 Stockholm, Sweden
Received December 6, 2000
Adducts to N-terminal valine residues in hemoglobin (Hb) are used for monitoring in vivo
doses of electrophiles and are quantitated by means of a modified Edman procedure, the “N-
alkyl Edman procedure”. In the reaction with pentafluorophenyl isothiocyanate, N-alkylated
valines cyclize and detach from the protein as pentafluorophenylthiohydantoins (PFPTHs) much
more efficiently than do unsubstituted N-terminal valine residues. The mechanisms of this
reaction, and of possible degradation reactions, have been studied with model compounds using
phenyl- and pentafluorophenyl isothiocyanate. The rapid cyclization to N-alkylvaline-PTHs
occurs as a consequence of the influence of substituents on ring formation. This facilitated
cyclization favors a direct attack by the thiocarbamoyl nitrogen atom on valine-C-1, and is
also observed to occur slowly at unsubstituted N-terminal valines. Such cyclization is favored
in protic solvents. Under alkaline conditions and in the presence of air, hydrolytic and oxidative
processes give rise to degradation products. The PTH derivatives of N-alkylvaline are less apt
to undergo such reactions than are the corresponding derivatives of unsubstituted valine. We
conclude that the presence of an N-substituent exerts a greater influence on the cyclization
process than the structure of the amino acid or of the Edman reagent. For adducts of different
structures, the method has broad applicability, for which the limits, however, are not yet
explored. The knowledge from the studies is valid not only for the N-alkyl Edman procedure,
but also, to some extent, for the classical Edman degradation reaction. The oxidative side
reaction gave rise to the invention of a novel synthesis route for insertion of nucleophiles at
carbon-5 in thiohydantoins. The present investigation provides a basis for the N-alkyl Edman
procedure, facilitating new toxicological applications.
N-alkylated with a radioactively labeled 2-hydroxyethyl
moiety from ethylene oxide was released spontaneously
as a phenylthiohydantoin (PTH)1 under the conditions
(pH >7) employed for the coupling reaction between
phenyl isothiocyanate (PITC) and protein (8). This PTH
could be separated by extraction from unmodified N-
terminal valine residues, as well as from the rest of the
protein.
This observation led to the development of the so-called
“N-alkyl Edman” procedure for mass spectrometric (MS)
quantitation of Hb adducts (9). To attain the required
sensitivity and reproducibility, considerable development
was necessary (9-11). Because of its usefulness, this
method has been applied in a number of laboratories for
research purposes, dose monitoring, and hygienic surveil-
lance (12-17).
In tr od u ction
It has been demonstrated earlier that in vivo electro-
philic compounds can be monitored by measuring the
products (adducts) of their reaction with proteins, in
particular hemoglobin (Hb) (1-5). Important nucleophilic
sites in Hb which are reactive under physiological condi-
tions are the imidazole nitrogen atoms in histidine
residues, sulfur atoms in cysteine and methionine resi-
dues, oxygen atoms in carboxyl groups and in hydroxyl
groups in tyrosine and serine residues, and the R-nitro-
gen atoms in the N-terminal valine residue of all four
chains of human Hb (5, 6).
In attempts to develop a simple and sensitive proce-
dure for analysis of adducts to N-terminal valine residues
in Hb, a number of methods were tested, including the
Edman degradation procedure used for protein sequenc-
ing (7). It was observed that the N-terminal valine
The method has functioned well despite the incomplete
knowledge concerning the mechanism of formation and
1
* Correspondence should be addressed to this author at the Depart-
ment of Environmental Chemistry, Arrhenius Laboratory, Stockholm
University, S-106 91 Stockholm, Sweden. Fax: +46-8-152561, e-mail:
margareta.tornqvist@mk.su.se.
For abbreviations of the compounds studied, see Figure 2. Other
abbreviations: ATZ, anilinothiazolinone; BCPS, bis-(4-chlorophenyl)-
sulfone; DDQ, 2,3-dichloro-5,6-dicyano-p-benzoquinone; 2,4-DNF, 2,4-
dinitrophenylhydrazine; EtOAc, ethyl acetate; MeOH, methanol; MITC,
methyl isothiocyanate; PITC, phenyl isothiocyanate; PTH, phenyl-
thiohydantoin; PFPITC, pentafluorophenyl isothiocyanate; PFPTH,
pentafluorophenylthiohydantoin.
† Department of Environmental Chemistry.
‡ Department of Organic Chemistry.
§ Department of Structural Chemistry.
10.1021/tx000247+ CCC: $22.00 © 2002 American Chemical Society
Published on Web 04/15/2002

Formation and Degradation of Phenylthiohydantoins
Chem. Res. Toxicol., Vol. 15, No. 4, 2002 571
F igu r e 1. The N-alkyl Edman reaction.
detachment of the valine derivative. Modifications sub-
sequently introduced, e.g., alterations in the conditions
employed for isolation and further derivatization of the
detached valine derivative (18, 19), have not involved
changes in the basic reaction parameters. A more detailed
understanding of the underlying mechanism is required,
above all for studies on previously unknown adducts (how
widely applicable is this procedure?) as well as for its
further development.
Due to its sensitivity and the possibility of identifying
adducts by mass spectrometric techniques, this method
can also be utilized for the identification and quantitation
of background carcinogens (20, 21), an area in which
investigation of Hb adducts often has advantages over
studying DNA adducts. Problems with artifacts in con-
nection with studies of adducts from background car-
cinogens are minimized in the case of the N-alkyl Edman
procedure, since modified valine cannot be incorporated
during synthesis of the globin protein (22). The identi-
fication of genuine valine adducts in Hb is also facilitated,
since the PFPTH derivatives contain this residue from
the globin.
ment of the rate of the cyclization resulting from N-
substitution.
During these studies, it was observed that prolonged
reaction times could lead to partial degradation of the
products formed, i.e., the N-alkylvaline-PTHs/PFPTHs
and Val-PTH/PFPTH. The conditions leading to such
degradation were investigated and the degradation prod-
ucts characterized. Relative degradation rates were
determined under alkaline conditions, in the presence
and absence of molecular oxygen.
Since the studies were performed primarily on free
amino acids (and their derivatives), which exhibit higher
pKa values than the corresponding N-terminal residues,
reaction media with a higher apparent recorded pH than
formamide were used (10, 13).
This paper deals with the chemistry involved in the
N-alkyl Edman procedure. The increased knowledge
regarding both formation and degradation of the prod-
ucts, the thiohydantoins, is required for application to a
broadened range of adducts, including previously un-
known ones.
A brief description of the N-alkyl Edman procedure is
presented in Figure 1. A sample of the globin (isolated
from red blood cells by acid precipitation) is dissolved in
formamide, and pentafluorophenyl isothiocyanate (PF-
PITC) is then added, together with a small amount of
aqueous 1 M NaOH in order to obtain a near-neutral
solution. The mixture is maintained at room temperature
overnight, after which the temperature is raised to 45
°C for a couple of hours (23). The pentafluorophenyl-
thiohydantoin (PFPTH) derivatives of the terminal N-
alkylvaline residues are released in high yield by this
procedure and subsequently isolated by extraction.
The main goal of the present study was to elucidate
why the N-alkyl Edman procedure selectively leads to
detachment of N-substituted valine residues (as PTHs/
PFPTHs) from the globin, whereas unmodified valine
requires acidification to effectuate such detachment. This
difference has also been observed at the amino acid
level: N-substituted valines, e.g., N-alkyl-, -hydroxy-
alkyl-, and -phenylvalines, give rise directly to PTHs/
PFPTHs, whereas valine does not give rise to these cyclic
products after the initial carbamoylation (24). It was thus
assumed that the major explanation for the selectivity
observed could not be the reagent (PITC or PFPITC) or
the structure of the substituent or the leaving group (the
protein residue), but that the determining factor is
probably the presence of an N-substituent.
Exp er im en ta l P r oced u r es
Ch em ica ls. Structures of compounds 1-131 and abbrevia-
tions of compounds 14-24 are given in Figure 2. 2,3-Dichloro-
5,6-dicyano-p-benzoquinone (DDQ, >98%) was obtained from
Merck; 2,4-dinitrophenylhydrazine (2,4-DNF, ∼99%) and phenyl
isothiocyanate (PITC, purum) were obtained from Fluka. Bis-
(4-chlorophenyl)-sulfone (BCPS,1 >99%) was obtained from
Sigma-Aldrich. L-Valine (Val, 19), L-valinamide (ValNH2, 20),
N-methyl-D,L-valine (MeVal, 21), sodium cyanoborohydride
(NaBH3CN, >90%), and 5-isopropyl-3-phenyl-2-thiohydantoin
(Val-PTH, 5) were obtained from Sigma. N-Methylvalylleucyl-
anilide (MeValLeu-NHæ, >99%, 22) and valylleucylserine
[ValLeuSer (H-Val-Leu-Ser-OH), 95%, 23) were obtained from
Bachem (Bubendorf, Switzerland). 18O2 (96%) was obtained from
ICON (Marion, NY). (2H3)Acetonitrile (99.8% 2H), (2H)chloroform
(99.8% 2H), deuterium chloride (99.5% 2H, 20% in 2H2O),
deuterium oxide (99.9% 2H, 2H2O), (2H8)2-propanol (99% 2H),
and (2H4)methanol (99.8% 2H) were obtained from CIL (Andover,
MA). 5-Isopropyl-1-methyl-3-phenyl-2-thiohydantoin (MeVal-
PTH, 6), 5-isopropyl-3-pentafluorophenyl-2-thiohydantoin (Val-
PFPTH, 12), and 5-isopropyl-1-methyl-3-pentafluorophenyl-2-
thiohydantoin (MeVal-PFPTH, 13) were synthesized as described
earlier (24). Phenylthiourea (14) and N-methyl-N′-phenylthio-
urea (15) were synthesized by reacting PITC with an excess of
aqueous concentrated ammonia and methylamine, respectively,
according to (25). All other chemicals and solvents were of
analytical grade.
In str u m en ta tion , Meth od s for An a lysis, a n d Ch a r a c-
ter iza tion . 1H and 13C NMR spectra were recorded on a J EOL
GSX 270 instrument at 270 MHz. All solvents used were fully
deuterated; TMS was added as internal standard in chloroform,
acetonitrile, and methanol. The alkaline hydrolysis experiments
Therefore, the present study has focused primarily on
comparative studies of N-methylvaline and valine as
models for N-substituted and unsubstituted valine resi-
dues, respectively, in Hb. Findings with these amino
acids were also confirmed by limited studies on peptides.
On the basis of our findings, a mechanism, described
previously, is proposed to explain the general enhance-
2
were performed in 0.5 M Na2HCO3 in H2O/(2H8)2-propanol [1:1
(v/v)] or in 0.5 M Na2CO3 in 2H2O/(2H8)2-propanol [1:1 (v/v)] at
45 or 30 °C using sodium 3-(trimethylsilyl)propanesulfonate as

572 Chem. Res. Toxicol., Vol. 15, No. 4, 2002
Rydberg et al.
F igu r e 2. The phenylthiocarbamoyls (PTCs), phenylthiohydantoins (PTHs), and other chemicals studied.
internal standard. The conditions for each experiment are
described below.
152 parameters was refined with full matrix least-squares
methods. The wR2(all data) ) 0.2162 and R1(observed) )
0.0876. These high R values are to a great extent dependent on
the large number of weak reflections.
Methods and instrumentation for GC-MS [GLC-(EI, NCI, and
PCI)-MS] were as described earlier (24). The LC-MS/thermo-
spray (TS) system comprised a Finnigan TSQ 700 triple-stage
quadrupole mass spectrometer from Finnigan (San J ose, CA),
connected to a Finnigan Mat TSP2 vaporization unit. The LC
system components were a Waters 600 multisolvent delivery
system connected to a Supelcosil LC-18 column (26 × 250 mm),
a Waters 486 MS tunable absorbance detector (λ ) 254 nm, 2H2
lamp), and a Waters 600 MS controller unit. Other parameters
for the LC-MS system were as follows: evaporation tempera-
ture, 112 °C; ionization chamber temperature, 230 °C; flow rate
) 1 mL/min; eluent A ) 20% aqueous acetonitrile, 0.05 M NH4-
HCO3 buffered to pH 5 with acetic acid; eluent B ) acetonitrile,
gradient ) 0-100% B in 15 min, then 100% B; loop, 20 µL; scan
range, 110-350 m/z, 530 ms/scan. The positive-ion mode (PI)
was employed.
The LC-MS/electrospray (ESP) system comprised a Waters
2690 separation module interfaced with a Quatro-II triple-stage
quadrupole mass spectrometer from Micromass. The MS was
fitted with an atmospheric pressure ionization (API) source,
which was used in the electrospray mode. Mobile phase: H2O/
methanol (MeOH) (1:1), isocratic flow at 10 µL/min. The ion-
source temperature was 120 °C, the capillary voltage was 3.5
kV, and the cone voltage varied between 25 and 140 V. Nitrogen
was used as drying gas at a flow rate of 250 L/h. Both the PI
mode and the negative-ion mode (NI) were used.
Crystallographic data (excluding structure factors) for struc-
ture 8 in this paper have been deposited at the Cambridge
Crystallographic Data Center as supplementary publication no.
CCDC 137560. Copies of these data can be obtained, free of
charge, on application to CCDC, 12 Union Rd., Cambridge CB2
1EZ,U.K.(fax: +441223336033or e-mail: deposit@ccdc.cam.ac.uk).
Kinetic measurements under anaerobic alkaline conditions
gave rate constants at 45 °C ((0.1 °C) under pseudo-first-order
conditions in the thermostated cell compartment of a Hitachi
U-3210 spectrophotometer. The rate of the hydrolysis reaction
of 5 and 6 was calculated from the change of absorbance at 270
nm.
Degradation studies under aerobic alkaline conditions gave
more than one product, and kinetic measurements were there-
fore performed by HPLC, on a Shimadzu LC-4A as described
above: Supelcosil LC-18 column (26 × 250 mm), detection (λ )
254 nm), flow rate ) 0.7 mL/min, loop 100 µL (normally 10-50
µg/compound injected); eluents: A ) 20% aqueous acetonitrile,
buffered at pH 6.9 with 0.02 M phosphate; B ) acetonitrile,
gradient ) 0-100% B in 35 min, 100% B in 15 min. The
degradation rates of 5, 6, 12, and 13 were obtained under
pseudo-first-order conditions by comparing the absorbances of
the compounds with the absorbances of the respective reference
compound.
Purification of 24 was performed by HPLC, on a Shimadzu
LC-4A connected with a Kromasil LC-18 column (250 × 10 mm)
and a Shimadzu SPD-2AS detector (λ ) 214 nm, D2 lamp). Flow
rate ) 2.5 mL/min, loop 0.7 mL eluted with 2% aqueous
acetonitrile buffered with 0.02% TFA.
The crystallographic investigations of 8 were carried out on
a Siemens STOE/AED4 diffractometer equipped with a graphite
monochromator. Thin plates were obtained by recrystallization
from MeOH: space group P-1(2), a ) 6.87(1) Å, b ) 11.82(1) Å,
c ) 9.057(3) Å, R ) 85.62(5)°, â•• ) 86.12(6)°, and γ ) 76.05(6)°.
The thin single crystals yielded a total of 1148 reflections with
547 significant reflections with I g 2σ(I). The structure was
solved by direct methods, and the final model with a total of
TLC was performed using silica gel 60 f-254 plates (SiO2,
Merck); spots were developed with UV, ninhydrin, and/or
potassium permanganate. Melting points were determined on
a Bu¨chi 535 instrument. Mikrokemi AB, Uppsala, Sweden,
performed elementary analyses. Measurements of pH were
carried out on an Orion EA 920 pH-meter equipped with a Ross
8130 glass electrode.
St a r t in g Ma t er ia ls: Syn t h esis of Sod iu m N-P h en yl-
th ioca r ba m oyl-L-va lin a te (P TC-Va l-Na , 1), N-P h en ylth io-
ca r ba m oyl-L-va lin e (P TC-Va l, 2), a n d N-P h en ylth ioca r -
ba m oyl-L-va lin e Meth yl Ester (P TC-Va l-OMe, 3). Com-
pound 1 was synthesized according to Edman (26). Synthesis
of 2 and 3: To an ice-cold, stirred solution of 1 (100 mg, 0.36

Formation and Degradation of Phenylthiohydantoins
Chem. Res. Toxicol., Vol. 15, No. 4, 2002 573
mmol) in EtOAc (10 mL) was added 1 M aqueous NaHSO4 (5
mL). The organic phase was dried (Na2SO4) and divided into
two equal parts. One aliquot was used to obtain NMR spectra
of the acid 2, and the other aliquot was methylated with an
etheral solution of CH2N2 to yield 3 for NMR analysis. No
further characterizations were performed. 1: 1H NMR (in
acetonitrile-2H3, 25 °C): δ 0.88, 0.95 [dd, 6H, J ) 7.6 Hz, CH3-
(γ,γ′)], 2.23 [m, 1H, J ) 4.7, 7.6 Hz, CH(â)], 4.70 [dd, 1H, J )
4.7, 7.6 Hz, CH(R)], 7.01-7.62 (m, 5H, ar.), 8.27 [d, 1H, J )
7.5, NH(Val)], 10.42 (br, 1H, N′H). 13C NMR (in acetonitrile-
2H3, 25 °C): δ 19.50, 19.75 [CH3(γ,γ′)], 32.48 [CH(â)], 65.45 [CH-
(R)], 118.20, 124.64, 129.14, 141.40 (ar.), 178.22 (CO), 181.75
(CS). 2: 1H NMR (in chloroform-2H, 25 °C): δ 0.92, 0.95 (dd,
6H, 6.7 Hz, CH3(γ,γ′)], 2.32 [m, 1H, 4.5, 6.7 Hz CH(â)], 5.14
[d+d, 1H, J ) 4.5, 8.7 Hz CH(R)], 6.52 [d, 1H, J ) 8.7 Hz,
NH(Val)], 7.25-7.46 (m, 5H, ar.), 8.76 (s, 1H, N′H), 9.80 (br,
1H, COOH). 13C NMR (in chloroform-2H, 25 °C): δ 18.13, 18.65
[CH3(γ,γ′)], 31.02 [CH(â)], 62.38 [CH(R)], 127.33, 129.14, 130.18,
135.79 (ar.), 176.02 (CO), 180.38 (CS). 3: 1H NMR (in chloroform-
2H, 25 °C): δ 0.90, 0.93 (d+d, 3+3H, J ) 7.0 Hz, CH3(γ,γ′)],
2.26 [m, 1H, J ) 4.9, 7.0 Hz, CH(â)], (s, 3H, CH3-O), 5.12 [d+d,
1H, J ) 4.9, 8.5 Hz, CH(R)], 6.63 [d, 1H, J ) 8.5 Hz, NH(Val)],
7.28-7.48 (m, 5H, ar.), 8.35 (s, 1H, NH). 13C NMR (chloroform-
2H, 25 °C): δ 18.33, 18.56 [CH3(γ,γ′)], 31.25 [CH(â)], 52.20
(OCH3), 62.66 [CH(R)], 124.82, 127,26, 130.12, 135.91 (ar.),
176.34 (CO), 180.59 (CS).
Syn th esis of N-Meth ylva lylleu cylser in e (MeVa lLeu Ser ,
24). ValLeuSer (23, 101 mg, 0.302 mmol) was dissolved in water
(8 mL), and 36% aqueous formaldehyde (22 µL, 0.29 mmol) was
added with stirring at room temperature. After 15 min, NaBH3-
CN (21 mg, 0.30 mmol, calcd on 90% purity) was added, and 18
h later, the reduction was stopped by acidification to pH ∼2 with
TFA (CAUTION!).2 Purification of the crude product was
performed with semipreparative HPLC as described above. Four
peaks were obtained and fractions collected, retention time
[compound (yield)]: 15 min [23 (32.9 mg, 34%) obtained as a
white solid], 16 min [unidentified product, partially coeluted
with 23 (6.9 mg)], 20 min [24 (36.1 mg, 36%) obtained as a white
solid], 26 min [N-(dimethyl)-ValLeuSer (36.2 mg, 35%) obtained
as a white solid]. Analysis by LC-MS(ESI) m/z [(fragment, ion),
rel int.]: compound 23: 318 [(M+H)+, 5%], 213 [(M-105)+, (M-
Ser)+ 3%], 102 [(M-216, M-LeuSer)+, 14%], 72 [(M-246,
H2NCHCHMe2)+, 9%)]. Compound 24: 332 [(M+1+), 37%], 227
[(M-105, M-Ser)+, 6%], 86 [(M-246, MeNCHCHMe2)+, 78%].
N-(Dimethyl)-ValLeuSer: 346 [(M+1)+, 11%], 100 [(M-246,
Me2NCHCHMe2)+, 34%]. The identity of compound 24 was also
confirmed by performing the N-alkyl Edman procedure with
PITC as reagent (24). The product formed, MeVal-PTH [char-
acterized by TLC (SiO2, Tol/EtOAc 2:1) and by GC-MS (PCI)],
was pure (Val-PTH could not be detected).
precipitate formed was dissolved in EtOAc (2 × 5 mL). Evapora-
1
tion gave a crystalline solid identified as 5 by H NMR and l3C
NMR; the data obtained were in agreement with published
spectra (24).
Com p a r a tive Stu d y of th e Yield s of 5, 6, 12, a n d 13
Obta in ed u p on Rea ctin g MeVa l (21), MeVa lLeu -NHæ (22),
MeVa lLeu Ser (24), a n d Va lLeu Ser (23) w ith P ITC a n d
P F P ITC in Aqu eou s 0.5 M Na HCO3/1-P r op a n ol [2:1 (v/v)].
(A) P r ep a r a tion of Stock Solu tion s. Stock solution A: to a
solution of 23 (110 mg, 347 µmol) in water (2.2 mL) was added
BCPS (440 µg, internal standard) dissolved in 1-propanol (1.0
mL). Stock solution S-24: compound 24 (4.3 mg, 13 µmol) was
dissolved in water (0.86 mL, giving 5.0 mg/mL, 15 mM). Stock
solution S-22: compound 22 (4.8 mg, 15 µmol) was dissolved in
1-propanol (0.96 mL, giving 5 mg/mL, 16 mM). Stock solution
S-21: compound 21 (6.3 mg, 48 µmol) was dissolved in water
(4.3 mL, giving 1.5 mg/mL, 11 mM).
(B) P r ep a r a tion of Sa m p les, Gen er a l P r oced u r e. To six
glass vials (2 mL) each containing 1.25 mL of 0.5 M aqueous
NaHCO3/1-propanol [2:1 (v/v)] was added stock solution A (150
µL containing 5.0 mg, 16 µmol of 23, and 19 µg of BCPS)
together with 100 µL of stock solution S-24 (double sample) and
S-24 diluted 10 and 100 times with water (2+2 samples). This
dilution series gave molar ratios between 24 and 23 of ap-
proximately 1/10, 1/100, and 1/1000 (see Table 1), respectively.
The procedure described above was also applied for preparation
of dilution series from stocks S-21 and S-22 (totally 6+6
samples, S-22 was diluted with 1-propanol). Two control samples
were prepared from stock solution A and PITC or PFPITC.
(C) Der iva tiza tion . The samples were preheated (46 °C) and
divided into two equal groups; one group was reacted with PITC
(20 µL, 17 µmol/sample) and the other with PFPITC (25 µL, 17
µmol/sample), all reactions being performed with stirring. After
70 min, the reactions were stopped by fast-chilling to 0 °C (ice
bath), followed by extraction with toluene (2 × 1 mL). The
samples were dried with Na2SO4, analyzed by GC-MS (PCI, as
described above), and quantitated on the basis of standards
[containing 5, 6, 12, and 13 at concentrations of 100, 20, 4, 0.8,
and 0.16 µg/mL and BCPS (19 µg/mL in all standards)].
Syn th esis a n d Cycliza tion of P ota ssiu m N-Meth yl-N-
(p h en ylth ioca r ba m oyl)-D,L-va lin a te (P TC-MeVa l-K, 4) a t
p H >7. A solution of MeVal (7.8 mg, 59 µmol) in 0.1 M aqueous
KOH (590 µL) was evaporated to dryness, and the residue was
2
dissolved in H2O (1 mL), evaporated to dryness, and suspended
in acetonitrile-2H3 (1 mL). The suspension was again evaporated
and suspended in acetonitrile-2H3 (1.5 mL). PITC (7 µL) was
added and the suspension stirred on a vortex mixer (5 min) and
1
filtered. An aliquot (0.7 mL) was directly used to obtain the H
2
NMR spectrum of 4 at 37 °C (TMS as internal standard). H2O
(0.2 mL) was added, and after another 10 min, the spectrum
obtained was identical with that of MeVal-PTH (6) described
earlier (24). Analysis of 4: 1H NMR (acetonitrile-2H3, 37 °C) δ
0.87, 0.95 [d+d, 3+3H, J ) 6.6 Hz, CH3(γ,γ′)], 2.43 [m, 1H, J )
6.6, Hz, CH(â)], 3.37 (s, 3H, N-CH3), 3.55 [br 1H, CH(R)], 7.59
(1H, N′H), 7.00-7.58 (m, 5H, ar.).
Cycliza tion to Th ioh yd a n toin s
Acid -Ca ta lyzed Cycliza tion of P TC-Va l (2) to Va l-P TH
(5). Concentrated aqueous 2HCl (50 µL, 600 µmol) was added
to a solution of 1 (5.0 mg) in 1 mL of (2H4)methanol containing
sodium 3-(trimethylsilyl)propanesulfonate as internal standard.
Degr a d a tion a n d Oxid a tion of P THs
I
After removal of precipitated NaCl, four H NMR spectra of the
Alk a lin e An a er obic Hyd r olysis of Va l-P TH (5) a n d
MeVa l-P TH (6). Samples were prepared by dissolving com-
pounds 5 or 6 in 1-propanol (1 mg/mL), and these solutions were
freshly used. Buffers used were A Sørensen Glycin:NaOH [0.100
M + 0.100 M, 1:1 (v/v); measured at pH 11.2 at 22 °C, giving
pH 10.6 at 45 °C according to Clark (27)] and B Tris-Cl [0.100
M; measured at pH ) 8.3 at 22 °C, giving pH 7.7 at 45 °C
according to Clark (27)]. Hydrolysis was started by adding
aliquots (30 µL) of the sample solutions to the preheated buffers
(3 mL, 45 °C, giving 10 µg/mL 5 and 6 in buffer solutions A and
B containing 1% propanol) in sealed spectrometer cuvettes
under argon. The half-times of the hydrolysis reactions were
obtained from measurements of the UV absorption at 270 nm.
Alk a lin e Aer obic Hyd r olysis of Va l-P TH (5) a n d MeVa l-
P TH (6) a n d LC-MS of Hyd r olysis P r od u cts. Compound 5
solution were recorded; the first, immediately at 25 °C, the
second and third, 10 and 20 min later, respectively. Finally, the
temperature was raised to 55 °C, and the last spectrum was
recorded 15 min later. The first spectrum is consistent with that
of the free acid 2. The last spectrum is identical to that of 5
(24). No intermediate anilinothiazolinone (ATZ) could be de-
tected during the progress of the cyclization reaction.
Cycliza tion of N-P h en ylth ioca r ba m oyl-L-va lin e Meth yl
Ester (P TC-Va l-OMe, 3) a t p H ∼8. Compound 3 (20 mg) was
suspended in 0.5 M aqueous NaHCO3/1-propanol [2:1 (v/v), 5
mL]. After a few minutes at room temperature, the white
2
CAUTION! Acidification must be performed in a well-ventilated
hood since HCN is released.

574 Chem. Res. Toxicol., Vol. 15, No. 4, 2002
Rydberg et al.
(5.0 mg) in 5 mL of 0.5 M aqueous NaHCO3/1-propanol [2:1 (v/
v)] was kept at room temperature with p-nitroanisole as internal
reference. After 4 days, an aliquot (300 µL) was neutralized with
1 M HCl (50 µL), then diluted in eluent toluene/EtOAc [1.55
mL, 1:1 (v/v)], and analyzed by LC-MS. Analysis by LC-MS-
(TS) [(fragment) m/z (rel int.)]: phenylthiourea (14) eluted at
8.50 min [(M+H+) 153 (100), (M-33)+ 119 (26)]; PTC-Val (2)
eluted at 11.02 min [M+ missing, (M-16)+ 236 (20), (M-93)+
159 (100), (M-134)+ 118 (19)]; Val-PTH (5) eluted at 14.40 min
[(M+H)+ 235 (100)]; p-nitroanisole eluted at 15.28 min; Val-
PTH-R-OPr (7) eluted at 17.51 min [(M+H)+ 293 (100), (M-
59)+ 233 (53), (M-168)+ 124 (20), (M-177)+ 115 (53)] (see below
for more detailed analysis of 7). The retention times for 2, 5,
and 14 were in agreement with reference compounds analyzed
under identical conditions.
129.23, 129.45, 132.29 (ar.), 171.34 (CO), 183.66 (CS). UV
(ethanol, r.t.) λmax ) 273.1 nm, ꢀ273.1 ) 14 500.
Alk a lin e Oxid a tion of 5 by Oxygen Ga s in MeOH. To a
solution of 5 (50 mg, 0.21 mmol) in dry MeOH (10 mL) was
added 0.5 mL of 0.5 M NaOMe (0.25 mmol) in MeOH. Oxidation
was induced by a gentle stream of oxygen bubbles through the
solution. The progress of the reaction was monitored by TLC
[SiO2, toluene/EtOAc 1:1 (v/v)] and by iodometric titration of
aliquots (10 × 0.3 mL) (with 10 mM thiosulfate). After 8 h, the
remaining solution (7 mL) was diluted with diethyl ether (20
mL) and extracted with H2O (3 × 15 mL). Evaporation of the
etheral phase gave a colorless syrup, which was then chro-
matographed on SiO2 [20 × 130 mm, heptane/EtOAc 1:1 (v/v)].
After fractionation and evaporation, the solid residue was
recrystallized in CHCl3/heptane (1:6, v/v) to give 8 as white
needles (31 mg, 0.12 mmol, calculated yield 80%). Analysis: 1H
NMR (chloroform-2H, 25 °C): δ 1.02, 1.10 [d+d, 3+3H, J ) 7.0,
CH3(γ,γ′)], 2.37 [q, 1H, J ) 7.0 Hz, CH(â)], 3.36 (s, 3H, OCH3),
7.25-7.53 (m, 5H, ar.). 13C NMR (chloroform-2H, 25 °C): δ 15.46,
16.14 [CH3(γ,γ′)], 34.60 [CH(â)], 52.08 (OCH3), 93.43 [CH(R)],
128.18, 129.24, 129.46, 132.29 (ar.), 170.96 (CO), 183.83 (CS).
GC-MS(PCI): m/z 265 [(M+1)+, 100], 293 [(M+29)+, 38], 305
[(M+41)+, 13]. GC-MS(NCI): m/z 264 (M-, 9), 247 [(M-17)-,
4], 232 [(M-32)-, 100]; mp 152-153.5 °C. The structure of
compound 8 was determined by X-ray diffraction.
MeVal-PTH (6, 5.0 mg) was hydrolyzed as above in 0.5 M
Na2CO3/1-propanol [2:1 (v/v), 48 °C, 4 days]. An aliquot (300
µL) was analyzed as above. Analysis by LC-MS(TS) [(fragment)
m/z (rel int.)]: N-methyl-N′-phenylurea (18) eluted at 9.43 min
[151 (100)]; N-methyl-N′-phenylthiourea (15) eluted at 11.28 min
[(M+H)+ 167 (100)]; p-nitroanisole eluted at 15.22 min.
Alk a lin e Aer obic Hyd r olysis a n d Rela tive Ra tes of
Hyd r olysis of P h en ylth ioh yd a n toin s 5 a n d 6 a n d P en -
ta flu or op h en ylth ioh yd a n toin s 12 a n d 13. Preparation of
samples (general method): compound 5 (2 mg) and p-nitroani-
sole (5 mg) were dissolved in 1-propanol (1.0 mL) in a Pyrex
tube (8 mL) with a screw cap. The tube was heated on a
thermostat-controlled water bath at 45 °C and the hydrolysis
started by the addition of a freshly prepared, preheated solution
of 0.5 M aqueous NaHCO3/1-propanol [2:1 (v/v), 2.0 mL] and
vigorous shaking. The following samples and references were
used: Compound 6 (2.5 mg) with biphenyl (0.5 mg), 12 (2.5 mg)
with biphenyl (0.5 mg), 13 (2.5 mg) with p-nitroanisole (5 mg).
Analysis: Aliquots (100 µL) of the hydrolysis samples were
dissolved in 20% aqueous acetonitrile, buffered to pH 6.9 with
0.02 M phosphate, and injected onto a Shimadzu HPLC. The
half-lives (t1/2) were calculated from the pseudo-first-order rate
constants (k), inferred from log concentration (log A) vs time
plots.
Alk a lin e Oxid a tion of 6 by Oxygen Ga s in MeOH.
Compound 6 (52 mg, 0.21 mmol) was oxidized by a slow stream
of oxygen as described for 5 above. After 8 h, the remaining
solution (70%) was diluted with H2O (14 mL) and extracted with
diethyl ether (3 × 15 mL). The combined etheral phases were
dried with Na2SO4 and further purified by preparative TLC
[SiO2, toluene/EtOAc 1:1 (v/v)] to yield MeVal-PTH-R-OMe (11,
11 mg, 0.039 mmol, calculated yield 27%) and MeVal-PTH-R-
OH (10, 10 mg, 0.038 mmol, calculated yield 26%), both as
colorless syrups. A small amount N-methyl-N′-phenylthiourea
(15) was also obtained (1 mg, 0.006 mmol, calculated yield 4%).
Analysis of 10: 1H NMR (chloroform-2H, 25 °C): δ 1.00, 1.20
[d+d, 3+3H, J ) 7.0, CH3(γ,γ′)], 2.37 [q, 1H, J ) 7.0 Hz, CH(â)],
3.28 (s, 3H, NCH3), 3.47 [s, 1H, OH(R)], 7.24-7.52 (m, 5H, ar.).
13C NMR (chloroform-2H, 25 °C): δ 15.25, 15.95 [CH3(γ,γ′)],
29.25, 29.68 [CH(â), splitted 2.3:1 ratio], 34.49 (NCH3), 89.65
[CH(R)], 128.32, 129.10, 129.27, 132.72 (ar.), 172.07 (CO), 181.69
(CS). Analysis of 11: 1H NMR (chloroform-2H, 25 °C): δ 0.98,
1.18 [d+d, 3+3H, J ) 7.0, CH3(γ,γ′)], 2.36 [q, 1H, J ) 7.0 Hz,
CH(â)], 3.24, 3.27 (s+s, 3H+3H, NCH3, OCH3), 7.2-7.6 (m, 5H,
ar.). 13C NMR (chloroform-2H, 25 °C): δ 15.17, 15.95 [CH3(γ,γ′)],
28.95, 29.68 [CH(â), splitted 1.2:1 ratio], 34.49 (NCH3), 52.46
(OCH3) 95.03 [CH(R)], 128.35, 129.18, 129.47, 131.10 (ar.),
169.83 (CO), 182.50 (CS).
1H NMR St u d ies of Alk a lin e Aer obic Hyd r olysis of 5
a n d 6. General method: Val-PTH (5, 3 mg) was dissolved in
2-propanol-2H8 (350 µL); the hydrolysis was initiated by addition
of either 0.5 M Na2CO3 in 2H2O (350 µL) or 0.5 M Na2HCO3 in
2
2H2O (350 µL; 0.5 M Na2HCO3 in H2O was prepared by addition
2
of 1 equiv of 2HCl in H2O to a solution of Na2CO3). The samples
were then hydrolyzed in NMR tubes in a water bath at 45 °C.
Samples of MeVal-PTH (6) were prepared as described above.
The hydrolysis of 5 and 6 was studied under the following
conditions: p2H ∼9, 30 °C (t ) 20 min and t ) 30 min); p2H ∼9,
45 °C (t ) 30 min and t ∼18 h); p2H 11, 45 °C (t ) 18 h and t )
60 h).
Oxid a tion of 5 by DDQ. DDQ (98.0 mg, 4.31 mmol) was
added to a solution of 5 (101 mg, 0.431 mol) and triethylamine
(300 µL) in H2O/dioxane [10 mL, 1:1 (v/v)]. The solvent was
evaporated after 2 h, and the solid residue was fractionated on
a silica column [toluene/EtOAc 2:1 (v/v), TLC Rf ) 0.61] to give
9 (41 mg, 0.16 mmol, 37%) as a colorless syrup. Analysis: MS
(EI, direct inlet): m/z 260 [M+, 24], 259 [(M-1)+, 98], 250 [(M-
10)+, 13], 217 [(M-43)+, 100], 216 [(M-44)+, 47], 135 [(M-125)+,
36]. GC-MS(EI): m/z 258 [(M-2)+, 8], 232 [(M-28)+, 100]. GC-
MS(PCI): m/z 260 [M+, 100], 288 [(M+28)+, 30], 300 [(M+40)+,
10]. GC-MS(NCI): m/z 258 [(M-2)-, 23], 232 [(M-28)-, 100].1H
NMR (chloroform-2H, 25 °C): δ 1.14, 1.31 [d+d, 3+3H, J ) 7.0
Hz, CH3(γ,γ′)], 2.58 [sep, 1H, J ) 7.0 Hz, CH(â)], 7.24-7.65 (m,
5H, ar.). 13C NMR (chloroform-2H, 25 °C): δ 15,64, 16.99 [CH3-
(γ,γ′)], 36.45 [CH(â)], 64.04 [CH(â)], 113.28 [CN], 128.01, 129.50,
130.01, 131.66 (ar.), 166.44 (CO), 182.90 (CS). Compound 9 was
also obtained as the main product of oxidation of 5 in acetonitrile
with oxygen in the presence of KCN and triethylamine. It is
therefore evident that DDQ acted not only as an oxidant but
also as a source of cyanide ion.
P r ep a r a tive Alk a lin e Aer obic Hyd r olysis of 5. A solution
of 5 (0.98 g, 4.2 mmol) in 0.5 M aqueous NaHCO3/1-propanol
[250 mL, 1:1 (v/v)] was stirred for 50 h at 47 °C. The solution
was evaporated to dryness, and the residue was extracted with
acetone (3 × 50 mL). Removal of acetone yielded a crystalline
solid (618 mg) that was chromatographed on a silica column
(40 × 220 mm) with heptane/EtOAc [1:1 (v/v)] as eluent.
Compound 7 (38 mg) eluted first, followed by 5 (160 mg) and
14 (yield not determined). Compound 7 was further purified
from
a highly lipophilic impurity (revealed by HPLC) by
extracting with heptane (15 mL) and repeated crystallizations
from heptane/CHCl3 [6:1 (v/v)] to give pure 7 (18 mg, 60 µmol,
1.4%) as white crystals, mp 185.8-186.6 °C (100% purity on
HPLC). Analysis: C15H19N2O2S: C 61.25%, H 6.55%, N 9.2%.
Found: C 61.83, H 6.57, N 9.61. 1H NMR (chloroform-2H, 25
°C): δ 0.95 (t, 3H, J ) 7.4 Hz, CH2-CH3), 1.02, 1.10 [d+d, 3+3H,
J ) 6.8, 7.0 Hz, CH3(γ,γ′)], 1.65 (m, 2H, J ) 7.0 Hz, -CH2CH3),
2.35 [m, 1H, J ) 6.9 Hz, CH(â)], 3.40, 3.55 (2 × m, 1+1H, J )
6.9, 7.1 Hz, OCH2), 7.25-7.53 (m, 5H, ar.). 13C NMR (chloroform-
2H, 25 °C): δ 10.54 (CH2CH3), 15.54, 16.05 [CH3(γ,γ′)], 22.78
(CH2CH3) 34.81 [CH(â)], 66.16 (OCH2), 92.94 [CH(R)], 128.21,
Oxygen Ga s (18O2) Oxid a tion of 5 a n d 6 in MeOH w ith
a n Equ im ola r Am ou n t of Na OMe. Val-PTH (10.3 mg, 44
µmol) and MeVal-PTH (10.9 mg, 44 µmol) were dissolved in dry

Formation and Degradation of Phenylthiohydantoins
Chem. Res. Toxicol., Vol. 15, No. 4, 2002 575
Ta ble 1. Rela tive Yield s of Ed m a n Der iva tives fr om N-Meth yla ted Va lin e a s th e F r ee Am in o Acid or a s th e N-Ter m in u s
of a Tr ip ep tid e
relative yields (%) of
[ValLeuSer]/
ratio of yields
of MeVal-PTH/Val-PTH
or (MeVal-PFPTH/Val-PFPTH)
added amount
(nmol/sample)
[compound] MeVal-PTH MeVal-PFPTH Val-PTH Val-PFPTH
compound
molar ratio
(6)
(13)
(5)
(6)
MeValLeuSer
MeValLeuSer
MeValLeuSer
MeValLeu-NHæ
MeValLeu-NHæ
MeValLeu-NHæ
MeVal
MeVal
MeVal
control (ValLeuSer)
1510
151
15.1
1560
156
15.6
1140
114
11.4
15800
10.5
105
1050
10.1
101
1010
13.9
139
96
96
64
57
56
91
101
89
94
69
74
128
0.46
0.40
0.41
0.40
0.41
0.41
0.40
0.41
0.40
0.37
0.052
0.045
0.037
0.062
0.043
0.044
0.062
0.043
0.041
0.046
210 (1100)
230 (1200)
160 (2600)
240 (2500)
260 (2100)
300 (2100)
300 (1100)
300 (2200)
280 (3100)
270 (1700)
95
105
111
120
124
112
1390
0.21a
(33 nmol)
0.17b
(27 nmol)
a
The background formation of 6 was estimated to be 33 nmol, corresponding to 0.21% (based on the added amount of 23, assuming
b
100% yield). The background formation of 13 was somewhat lower, i.e., 0.17%.
MeOH (2 mL) in two septum-equipped tubes (6 mL). The
solutions were bubbled with argon and evacuated. A portion of
18O2 (4 mL, 160 µmol/tube) was added to the tubes, and then
0.5 M NaOMe in MeOH (0.1 mL, 50 µmol) was added. The tubes
were shaken vigorously and then kept at ambient temperature
for 18 h. Purifications were performed by preparative TLC as
described above for compound 6. Compound 5 gave only one
product (8) without incorporation of 18O, while compound 6 gave
three products (10, 11, and 15). Analysis on LC-MS showed that
only compound 10 had incorporated 18O (50%). Analysis of 10:
LC-MS[ES, NI, 30 V (cone voltage)]: m/z 265 [18O, (M-1)-, 7],
263 [16O, (M-1)-, 11], 192 [18O, (M-74)-, 13], 190 [16O, (M-
74)-, 18], 130 [18O, (M-136)-, 71], 128 [16O, (M-136)-, 100].
LC-MS[EI, pos. mode, 30 V (cone voltage)]: m/z 267 [18O,
(M+1)+, 39], 265 [16O, (M+1-), 42]. Analysis of 11: GC-MS-
(PCI): m/z 319 [(M+41)+, 8], 307 [(M+29)+, 37], 279 [(M+1)+,
100], 247 [(M-31)+, 64], 235 [(M-43)+, 11]. GC-MS(NCI): m/z
278 (M-, 1), 246 [(M-32)-, 100].
proximately 1500 nmol) gave values that were out of the
linear range of the calibration curve, whereas the values
for samples containing the lowest concentrations (around
15 nmol) were high because of impure 23. A byproduct
from PFPITC which partly coeluted with BCPS, and gave
rise to common, interfering ions, disturbed measurement
of compounds 12 and 13. Despite these uncertainties
(estimated to amount to a maximum error of approxi-
mately 30%), the data in Table 1 permit conclusions to
be drawn concerning the influence of N-alkylation on the
relative yields of the different thiohydantoins and con-
cerning the differences between the phenyl and pen-
tafluorophenyl reagents.
There was a pronounced selectivity in the yields of 6
and 13 obtained from the N-methylated model com-
pounds as compared to the low yields of 5 and 12 formed
from 23. The ratio of the yields of 6/5 was approximately
250, while the corresponding ratio of 13/12 was even
higher, around 2000. This selectivity was shown to be
independent of whether the N-substituted valine was
present as the free amino acid or as part of a valyl
peptide. However, the used reagent, PFPITC, could
explain the 10-fold lower yield of Val-PFPTH compared
to Val-PTH, leading to an approximately 10 times higher
ratio of 13/12 than of 6/5. The inherent effect of a
substituent on the R-nitrogen atom on cyclization is also
reflected in the equilibrium between the open (PTC) and
ring-closed (PTH) forms of the amino acid in protic
solvents. While the PTH of N-methylvaline (6) was
formed rapidly without discernible intermediates, Val-
PTH was formed only slowly in an equilibrium that
favors the open valinate form (PTC-Val anion) in alkaline
solution and the cyclized PTH at low pH (see Figure 3).
The free acid form of N-(phenylthiocarbamoyl)valine
(PTC-Val, 2, see Figure 2) was sufficiently stable to
permit methylation by diazomethane to produce the
methyl ester (3, see Figure 2). This methyl ester (3)
cyclized to Val-PTH (5) within minutes at room temper-
ature in aqueous 1-propanol at pH 8. Attempts to prepare
the corresponding N-methyl-N-(phenylthiocarbamoyl)-
valine (cf. 4) resulted in the PTH derivative (6). However,
this compound could be prepared by reacting the potas-
sium salt of N-methylvaline with PITC in an aprotic
solvent (dry acetonitrile). The potassium N-methyl-N-
(phenylthiocarbamoyl)-D,L-valinate (PTC-MeVal-K, 4)
formed was stable under aprotic conditions, although
addition of a catalytic amount of water or alcohol induced
rapid cyclization to MeVal-PTH (6).
Resu lts
Cycliza tion to P TH a n d P F P THs. To assess the
influence of N-methylation of globin N-termini on the rate
of thiohydantoin formation, amino acids and peptides
were employed as models for normal and N-methylated
proteins, derivatized under the originally used conditions
developed for the N-alkyl Edman procedure. The yields
of MeVal-PTH (6) and MeVal-PFPTH (13), formed from
MeVal (21), MeValLeu-NHæ (22), and MeValLeuSer (24)
upon reaction with PITC and PFPITC, were compared
with the yields of Val-PTH (5) and Val-PFPTH (12)
formed from ValLeuSer (23) under the same conditions.
To determine the yields of analytes (5, 6, 12, and 13),
duplicate dilution series for each N-methylated model
compound (21, 22, and 24) were mixed with fixed
amounts of 23 and an internal standard (see Table 1).
The samples in one of these dilution series were reacted
with PITC and the other with PFPITC in mildly alkaline
aqueous 1-propanol at 47 °C. After these reactions, the
samples were purified by extraction and quantitated
employing GC-MS(PCI), utilizing the internal standard
and reference compounds. Unfortunately, the commercial
tripeptide (23) was found to contain an impurity corre-
sponding to N-methylvaline in peptide or as amino acid
resulting in unrealistically high yields of 6 and 13 in the
most dilute samples. The presence of this impurity
(estimated to be ∼0.21 mol %; probably formed from 24)
has been corrected for in Table 1.
In the evaluation of the yields of measured products,
some problems were encountered: the samples contain-
ing the highest concentrations of 21, 22, and 24 (ap-

576 Chem. Res. Toxicol., Vol. 15, No. 4, 2002
Rydberg et al.
F igu r e 3. Formation of Val-PTH (5) and N-(R)-Val-PTH (e.g., 6).
F igu r e 4. Proposed mechanisms for alkaline hydrolysis and oxidation of Val-PTH (5) and MeVal-PTH (6).
Degr a d a tion of P THs a n d P F P THs (Alk a lin e Hy-
d r olysis a n d Oxid a tion ). When Val-PTH (5) was
digested in aqueous 1% 1-propanol at pH 8 (buffer A) or
pH 11 (buffer B) in the absence of air, PTC-Val anion (1)
was the only product observed, with t1/2 values of 320 and
5.0 min, respectively (see alkaline hydrolysis in Figure
4). The PTC-Val anion (1) obtained was found to be very
stable toward further hydrolysis to valine and aniline [t1/2
∼ 8000 min (5.5 days) at pH 11] (as monitored by TLC
and ninhydrin) and could be reconverted to Val-PTH (5)
by acidification. Under the same conditions, MeVal-PTH
(6) disappeared more slowly than 5 [t1/2 ∼ 6000 min (4.2
days) at pH 8 and 70.5 min at pH 11] with concomitant
formation of N-methylvalinanilide (17, λmax ) 238 nm;
see Figure 4). The end-products of alkaline anaerobic
hydrolysis of 5 and 6 reflected the equilibria between the
open PTCs (1 and 4) and the PTHs (5 and 6). As expected,
compound 17 could not be reconverted to MeVal-PTH.
If the alkaline solution of 5 in aqueous 1-propanol was
exposed to air (see alkaline oxidative hydrolysis in Figure
4), appreciable amounts of phenylthiourea (14) and a
nonpolar component were present in addition to the anion
of PTC-Val (1) (7, cf. below).Under similar conditions,
MeVal-PTH (6) was slowly converted to N-methyl-N′-
phenylthiourea (15). N-Methyl-N′-phenylurea (18) was
also obtained, through the oxidation of 15 by the hydro-
gen peroxide formed in the reaction (28). Upon determin-
ing the relative rates of hydrolysis under aerobic condi-
tions (pH 8, 45 °C), the following t1/2 values and retention
times in the HPLC system [t1/2 (min); retention time
(min)] were obtained: 5 [460; 25], 6 [4500; 29], 12 [330;
34], and 13 [3900; 36]. As shown (Figure 4), the hydro-
phobic compound 7 (with a retention time of 32 min) was
formed in connection with the hydrolysis of 5, and a
corresponding Val-PFPTH compound was assumed to be
formed from 12 (with a retention time 43 min; but the
identity of this product was not verified). On the other
hand, hydrolysis of 6 and 13 under these conditions did
not result in any hydrophobic products.
The NMR spectrum of 7 showed the presence of a
propoxy group, the location of which could not be as-
signed. An oxidation product analogous to 7 could be

Formation and Degradation of Phenylthiohydantoins
Chem. Res. Toxicol., Vol. 15, No. 4, 2002 577
step performed in anhydrous strong acid, with concomi-
tant cyclization. The only atom acting as a nucleophile
under such acidic conditions is the sulfur of the phen-
ylthiocarbamoyl group, resulting in the release of a
thermodynamically unfavorable anilinothiazolinone
(ATZ1)(see Val-ATZ in Figure 6). Step 3: The ATZ is
converted to the more stable phenylthiohydantoin (PTH)
in aqueous acid. These reactions have been thoroughly
studied by several groups (7, 30, 34-36) and their
chemistry and applications reviewed in detail by Allan
(37).
Ba r ett’s Mod ifica tion of th e Ed m a n Degr a d a tion
Rea ction . In the modified procedure of Barrett (29), the
coupling of PITC to the peptide is accomplished in the
same manner as in the original method. After removal
of excess PITC, the temperature is raised to 75 °C, which
leads to cleavage of the first peptide bond with direct
formation of the phenylthiohydantoin (PTH). This cleav-
age is strongly accelerated by the addition of an aqueous
triethylammonium acetate buffer with a pH around 7.
This buffer acts at the same time as a base and as a
proton donor, both of which are necessary for bond
cleavage. In this case, the anilino nitrogen of the thio-
carbamoyl peptide must act as the nucleophile, in a
manner similar to our proposal for the N-alkyl Edman
reaction (cf. Figure 6, path b; see below). This route would
also explain the phenomenon of “preview” in the classical
Edman degradation reaction, well-known, but scarcely
mentioned in the literature (38, 39).
F igu r e 5. An ORTEP plot of molecule 8.
prepared in nearly quantitative yield by bubbling oxygen
for a few hours at room temperature through a solution
of Val-PTH (5) in dry MeOH containing an equimolar
amount of sodium methoxide. This crystalline product
was demonstrated by X-ray crystallography to be 5-iso-
propyl-5-methoxy-3-phenylthiohydantoin (Val-PTH-R-
OMe, 8 in Figure 5). Formation of peroxide was also
detected during the reaction, whereas when oxygen was
bubbled through the alkalized solvent alone, the peroxide
test was negative.
When MeVal-PTH (6) was oxidized in dry MeOH as
described above, three products were obtained, viz., 11,
15, and, surprisingly, 10. Upon addition of water, 10 was
converted to 15. When 18O2 was employed for the oxida-
tion of 5 and 6 in MeOH/sodium methoxide, 18O was
incorporated only into 10, in 50% yield.
A strong tendency to cyclize at pH 8 is also exhibited
by PTC-Val methyl ester (3; see Figure 6), in which the
negative charge of the carboxyl group has been elimi-
nated. This rapid cyclization of 3 implies that the anilino
nitrogen atom actually functions as the nucleophile,
giving rise directly to the PTH (path b in Figure 6). An
alternative attack by sulfur would lead to a short-lived
thiazolinone intermediate that would hydrolyze to yield
the valinate anion (1, path a in Figure 6). This route is
not very probable, because of the state of the equilibrium
between the open form (1) and Val-PTH (5) under the
mildly alkaline conditions of this reaction. The fact that
3 gives rise immediately to 5 in high yield implies that
attack by the nitrogen atom is the predominant reaction
mechanism in the case of carbamoylated N-substituted
as well as unsubstituted valines and valyl peptides. The
difference in the cyclization rates of N-substituted and
normal valine derivatives is best explained by the gem-
dialkyl effect (discussed below), not by a change in the
reaction mechanism.
Cycliza tion to Hyd a n toin s. Irrespective of the sub-
stituent present [an alkyl group from methyl to octyl,
isopropyl, 2-hydroxyalkyls, a phenyl group, etc. (10, 24,
40)], N-substitution enhances cyclization to the N-sub-
stituted valine-PTH/PFPTH under the conditions em-
ployed in the N-alkyl Edman procedure. This effect is
observed both with valyl peptides and with free amino
acids (see Table 1).
It is reasonable to assume that the rapid release of
N-alkylated PTH under the mild conditions of the N-alkyl
Edman procedure reflects acceleration of a reaction route
which is normally minor. Kirby and Blagoeva found that
alkyl-substituted hydantoic acids undergo cyclization as
much as 105 times more rapidly (31) than do their
unalkylated analogues (32). These investigators ascribe
this acceleration to the Thorpe-Ingold effect (41), i.e.,
release of steric strain during ring formation.
Production of urea derivatives requires the formation
of 2-oxoisovaleric acid (16), which was indeed identified
by reaction with 2,4-DNF to give the 2,4-dinitrophenyl-
hydrazone [(characterized by LC-MS/NI: (M-1)-, m/z )
295 (100% int.)].
When Val-PTH or Me-Val-PTH was digested for as
long as 60 h in aqueous (2H8)2-propanol containing 0.25
M sodium deuterium carbonate or disodium carbonate,
respectively, hydrolysis could be monitored by observing
1
selected H resonances. Immediately after the addition
of sodium deuterium carbonate, the â-hydrogen signal
(2.30 ppm in 5 and 2.55 ppm in 6) was transformed from
a multiplet into a well-defined septet, indicating removal
of the R-proton. An additional weak septet indicating the
presence of 5a and 6a (Figure 4) appeared rapidly at 3.05
ppm during the hydrolysis of 5 and at 3.65 ppm during
the hydrolysis of 6. Both of these septets disappeared as
the hydrolysis approached completion.
Discu ssion
In the present paper, the driving forces involved in the
N-alkyl Edman reaction were examined. An increased
understanding of this reaction can be attained by con-
sidering the present experimental data with previously
published mechanistic studies on the classical Edman
degradation (7, 26) and Barrett’s modifications (29) as
well as Davies and Mohammed’s investigations on race-
mization (30) and Blagoeva and Kirby’s work on hydan-
toins (31-33). The stability of the derivatives formed and
side-reactions such as hydrolysis and oxidation will also
be discussed.
Th e Cla ssica l Ed m a n Degr a d a tion Rea ction . The
classical Edman degradation approach to peptide se-
quencing (7, 26) consists of the following three reaction
steps: Step 1: The coupling of a protein or peptide with
PITC, in an aqueous buffer at pH >7. Step 2: A cleavage

578 Chem. Res. Toxicol., Vol. 15, No. 4, 2002
Rydberg et al.
F igu r e 6. Formation of Val-PTH (5) directly from PTC-Val-OMe (3).
Kirby has also investigated the conditions under which
the anilino nitrogen atom can act as a nucleophile to
attack a negatively charged carboxylate anion, albeit with
rather weak electrophilic character. The possible mech-
anisms underlying this reaction, in particular the se-
quence of proton transfer, are discussed by Blagoeva and
Kirby in some detail (31). Finally, Blagoeva et al. consider
stereo-electronic factors involved in the cyclization step.
These studies by Blagoeva and co-workers differ from
our considerations in two essential respects: (a) these
other investigations deal with the formation and stability
of hydantoins, not of thiohydantoins; and (b) they do not
involve protein derivatives. With respect to (a), it is
evident that a thiocarbamoyl nitrogen atom should be
at least as good a nucleophile as the corresponding
carbamoyl nitrogen, since sulfur is less electronegative
than oxygen. Concerning (b), the effects of alkyl substit-
small, but not negligible, nucleophilic character of the
anilino nitrogen atom becomes decisive, leading directly
to the product. In agreement with this proposed mech-
anism for cyclization, it has been observed that in
connection with the application of the N-alkyl Edman
procedure to globin, the formation of Val-PFPTH can be
observed at a low rate [approximately 0.1% of the
unsubstituted N-terminal valines are detached as PF-
PTH (9)]; at the same time, high yields (>50%) of
PFPTHs are obtained from N-substituted valines (with
an abundance of 1 in 106-108 termini) (8, 9).
Irrespective of which amino acid forms the N-terminal,
N-alkylation is expected to enhance the rate of cyclization
to the corresponding N-alkyl-PTH derivatives after initial
carbamoylation. This has been demonstrated to occur in
the case of N-terminal N-methylglycine (44), as well as
several other N-alkylated N-termini in our present
study.4
In this context, it can be mentioned that isocyanates
also form stable carbamoyl adducts to N-terminal valine
in Hb. These adducts are cyclized and detached as
hydantoins after acidification of the reaction mixture in
a procedure similar to the classical Edman degradation
(45), a procedure which has been employed in studies of
exposure to isocyanates or their precursors (46-49).
Solven t Effects. Both the polar protein and the more
unpolar PITC/PFPITC Edman reagent must dissolve in
the solvent utilized for the coupling reaction. The solvent
systems that have been tested in this respect are mix-
tures of water and organic solvents such as 1-propanol,
pyridine, THF, and dioxane (10). In the case of the
modified Edman reaction, formamide is now used both
in order to decrease products from side-reactions and
because formamide is a good solvent for globin under the
reaction conditions employed. Moreover, good yields can
be obtained with formamide as the solvent (10).
uents on hydantoins and thiohydantoins on cyclization
rates would in general not be expected to be dependent
on the nature of the leaving groups.
An alternative explanation put forward by Bruice and
Pandit (42) is the “Reactive Rotamer” effect. This hy-
pothesis states that alkylation should produce a decrease
in the concentration of unfavorable rotamers, thereby
increasing the rate of cyclization. Modern computational
methods have shown, however, that the concentration of
a reactive rotamer is a less important rate-determining
factor than its enthalpy. The results collected and
interpreted by Parill and Dolata (43) have been referred
to as the “facilitated transition” hypothesis.
The rate of the cyclization reaction can also be en-
hanced by using MITC1 instead of PITC. This can be
attributed to the higher nucleophilicity of the attacking
thiocarbamoyl nitrogen atom in the N-(methyl)thiocar-
bamoyl peptide in comparison to the corresponding
phenyl compound. The facile cyclization then gives rise
to a new free N-terminal residue susceptible to attack
by MITC. It was observed that MITC partially degraded
an unalkylated tripeptide (Val-Leu-Ser) during coupling,
whereas PITC formed stable carbamoyl products under
the same conditions.3
If the solvent is changed from a protic, polar solvent
(e.g., propanol/water) to an aprotic solvent (e.g., aceto-
nitrile, THF, or Me2SO), the carbamoylation step should
be accelerated because the nitrogen atom is solvated to
a lesser degree and, thus, more nucleophilic. However,
proteins such as globin generally precipitate in aprotic
solvents. On the other hand, the cyclization reaction itself
The cyclization of N-alkyl-N-(phenylthiocarbamoyl)-
valine, even when still attached to the rest of the protein,
thus seems to be induced by a stereo-electronic effect
favoring a reaction path that is also present in the normal
Edman reaction procedure as well. In this reaction, the
3
Rydberg, P., et al., unpublished findings.
To¨rnqvist, M., unpublished findings.
4

Formation and Degradation of Phenylthiohydantoins
Chem. Res. Toxicol., Vol. 15, No. 4, 2002 579
will also be affected by the choice of solvent, since polar,
protic solvents such as water have been reported to
facilitate cyclization by providing more extensive solva-
tion of the anion of the carboxyl group in the transition
state (32). In the present studies with free amino acids,
we observed this effect with PTC-MeVal anion (4). In the
aqueous solvent system, cyclization of this compound to
MeVal-PTH (6) was too rapid to be followed by chromato-
graphic methods (24), whereas in dry acetonitrile 4 was
stable until water or ethanol was added.
favor a cleavage of the oxygen-oxygen bond. An alterna-
tive explanation might be incorporation of the 18OH-
arising from homolytic cleavage of the hydrogen peroxide
anion.
This proposed mechanism, initiated by the formation
of an R-carbanion, is supported by the finding that the
R-proton in both Val-PTH and MeVal-PTH exchanges
with the solvent in slightly alkaline medium. According
to Davies and Mohammed (30), such exchange is ac-
companied by complete racemization of the R-carbon
center.
Degr a d a tion a n d Oxid a tion of P THs. Since weakly
alkaline aqueous alcoholic media have been employed for
the N-alkyl Edman reaction, particularly for preparation
of standards from N-alkylvalines (9), it was of interest
to determine the stability of PTHs in such media.
Furthermore, in develomental work it has been shown
that high pH during preparation and purification of
PFPTHs leads to decreased yields.4 To study the stability
of PTH derivatives, they were maintained at pH ∼8 or
higher for prolonged periods of time. Both in the absence
and in the presence of air, the relative rate of disappear-
ance of N-methylvaline-PTH (MeVal-PTH, 6) was roughly
10-20 times lower than that of the corresponding deriva-
tive of unsubstituted valine (5), in agreement with the
high rates of cyclization of the corresponding N-alkyl-N-
(phenylthiocarbamoyl)valinates. The pentafluorophenyl
derivatives (12, 13) were only slightly less stable than
the corresponding phenyl compounds (5, 6). The products
formed during hydrolysis of 5 and 6 included the thiourea
derivatives 14 and 15, respectively, formation of which
must involve an oxidation step. In the case of degradation
of Val-PTH (5), an additional nonpolar component con-
taining an alkoxy group was also formed.
When degradation of 5 was performed in aqueous
MeOH, this alkoxy compound was identified by X-ray
diffraction as 5-isopropyl-5-methoxy-3-phenyl-2-thiohy-
dantoin (8). This compound could be obtained in almost
quantitative yield by performing the oxidation with
molecular oxygen in nonaqueous basic solution (dry
MeOH). Various nucleophiles, e.g., cyanide ion, can be
incorporated into the 5-position. When N-methylvaline-
PTH (6) was oxidized under nonaqueous basic conditions,
an analogous 5-methoxy compound (11) was formed
simultaneously with a hydroxy compound (10) which
could be converted to the thiourea derivative (15) by
water. Even under these nonaqueous conditions, a sub-
stantial amount of the thiourea derivative (15) was
obtained.
To determine whether other oxidants can replace
oxygen, DDQ1 was tested. In the presence of this com-
pound under nonaqueous conditions, the cyano compound
(9) was formed from 5. This formation of 9 indicates that
the cyanide ion produced by degradation of DDQ can
execute a nucleophilic attack on the imine (5a , Figure 4)
formed during the oxidation. This was verified by bub-
bling oxygen through a solution of Val-PTH (5) in
acetonitrile in the presence of potassium cyanide and
triethylamine. Under these conditions, 9 was obtained
in high yield.
The much more rapid degradation of Val-PTH than of
N-alkyl-Val-PTH under the conditions employed for the
N-alkyl Edman reaction examined here is in favor of the
N-alkyl compounds, although prolonged reaction times,
in general, should be avoided. Our results indicate that,
to obtain optimal results, the reaction should preferably
be performed under anaerobic conditions.
Con clu sion s
The occurrence and levels of several types of electro-
philic compounds/intermediates in vivo have been exam-
ined by the N-alkyl Edman procedure, for specific de-
tachment and analysis of adducts to N-terminal valines
in Hb. In the present study, insight into the mechanism
of the key step, i.e., the favored detachment, through
cyclization, to Edman derivatives (PTHs and PFPTHs)
of N-alkylvalines, and also mechanisms involved in
degradation of the desired products were studied. On the
basis of our study, combined with results from other
studies, some general conclusions could be drawn:
• The general enhancement of the rate of cyclization/
cleavage observed for N-substituted valines in compari-
son to unsubstituted valine is related to the gem-dialkyl
effect. As a result of this effect, the rate of cyclization/
cleavage depends mainly on the presence or absence of
an N-substituent, less on the nature of the N-substituent,
the amino acid, and the isothiocyanate reagent employed.
However, the rate of the cyclization/cleavage reaction for
unsubstituted valyl peptides is in the order MITC > PITC
> PFPITC. This phenomenon probably reflects the higher
nucleophilicity of the methylthiocarbamoyl nitrogen atom
in comparison to the same atom in reagents containing
electron-withdrawing phenyl and pentafluorophenyl
groups.
We propose a mechanism for the oxidation of Val-PTH
(5) and MeVal-PTH (6), shown in Figure 4, which
involves the addition of oxygen to the carbanion moiety
R to the carbonyl group (50). After a proton shift, the
hydrogen peroxide anion is eliminated to form a resonance-
stabilized imine or immonium ion (5a and 6a ) that finally
accepts a nucleophile. The CH signal in the 1H NMR
spectrum of the isopropyl group in Val-PTH (5) was
temporarily shifted from 2.3 to 3.1 ppm during the course
of the reaction, indicating formation of an intermediate
imine. Analogous to this, a weak CH signal was tempo-
rarily recorded at 3.6 ppm during the oxidation of MeVal-
PTH (6), indicating a shift from its normal location (2.4
ppm).
• A predicted limitation of the N-alkyl Edman proce-
dure is that coupling/cyclization is prevented by blocking
the terminal nitrogen atom from being thiocarbamoy-
lated, e.g., when this nitrogen atom is tertiary or when
it is substituted with acyl groups (known blocking
agents).
• Protic, polar solvents (e.g., H2O, alcohols, and forma-
mide) facilitate cyclization, while aprotic solvents reduce
the rate of this reaction.
Incorporation of 18O into 10 can be explained on the
basis of the formation of a charged intermediate (6a in
Figure 4) caused by the N-methyl group of 6. Retarded
loss of a hydrogen peroxide anion from the ring would

580 Chem. Res. Toxicol., Vol. 15, No. 4, 2002
Rydberg et al.
(7) Edman, P., and Henschen, A. (1975) Sequence determination. In
Protein Sequence Determination (Needleman, S. B., Ed.) pp 232-
279, Springer-Verlag, Berlin.
• Under the conditions employed here, the Edman
derivatives formed are racemic.
• A high pH and the presence of oxygen, especially for
longer periods of time, lead to degradation of the Edman
derivatives both of the model N-alkylvaline used in these
studies, i.e., N-methylvaline, and of unsubstituted valine.
However, the Edman derivatives of N-methylvaline were
found to be degraded 10-20-fold more slowly than the
corresponding derivative of unsubstituted valine.
(8) J ensen, S., To¨rnqvist, M., and Ehrenberg, L. (1984) Hemoglobin
as a dose monitor of alkylating agents: Determination of alky-
lation products of N-terminal valine. In Individual Susceptibility
to Genotoxic Agents in the Human Population (de Serres, F. J .,
and Pero, R. W., Eds.) pp 315-320, Plenum Press, New York.
(9) To¨rnqvist, M., Mowrer, J ., J ensen, S., and Ehrenberg, L. (1986)
Monitoring of environmental cancer initiators through hemoglobin
adducts by a modified Edman method. Anal. Biochem. 154, 255-
266.
With respect to various physicochemical and chemical
factors that affect the yield of analytes and their analyti-
cal properties, the protocol for the N-alkyl Edman
procedure (9, 23) results in high yields of derivatives
formed from low molecular weight adducts with a low
level of undesirable products, e.g., little background
formation of unalkylated valine hydantoins. In practice,
when the N-alkyl Edman procedure is used for the
identification of adducts, the possible limitations with
regard to applicability (see above) as well as the stability
and potential degradation of adducts must be considered.
In real-life applications, particularly when the levels of
the adducts being examined are low, it becomes impor-
tant to avoid the formation of byproducts, e.g., due to
oxidation or hydrolysis. From these points of view, the
enhanced stability of N-alkylvaline-PTH compared to
Val-PTH is advantageous and no doubt contributes to the
usefulness of the N-alkyl Edman procedure. The conclu-
sions drawn from the mechanistic studies presented in
this paper provide a basis for further development of the
N-alkyl Edman procedure and its applications to new
areas of toxicology.
(10) To¨rnqvist, M. (1989) Monitoring and cancer risk assessment of
carcinogens, particularly alkenes in urban air. Thesis, Depart-
ment of Radiobiology, Stockholm University, Stockholm.
(11) To¨rnqvist, M. (1996) Monitoring of human exposure to xenobiotics;
Identification and quantification of cancer initiators in vivo. In
Mass Spectrometry in Biomolecular Sciences (Caprioli, R. M.,
Malorni, A., and Sindona, G., Eds.) pp 381-395, Kluwer Academic
Publishers, Dordrecht.
(12) Tates, A. D., Grummt, T., To¨rnqvist, M., Farmer, P. B., van Dam,
F. J ., van Mossel, H., Schoemaker, H. M., Osterman-Golkar, S.,
Uebel, C., Tang, Y. S., Zwinderman, A. H., Natarajan, A. T., and
Ehrenberg, L. (1991) Biological and chemical monitoring of
occupational exposure to ethylene oxide. Mutat. Res. 250, 483-
497.
(13) To¨rnqvist, M., Magnusson, A. L., Farmer, P. B., Tang, Y. S.,
J effery, A. M., Wazneh, L., Beulink, G. D. T., van der Waal, H.,
and van Sittert, N. J . (1992) Ring test for low-levels of N-(2-
hydroxyethyl)valine in human hemoglobin. Anal. Biochem. 203,
357-360.
(14) Bergmark, E., Calleman, C. J ., He, F., and Costa, L. G. (1993)
Hemoglobin adducts in humans occupationally exposed to acryl-
amide. Toxicol. Appl. Pharmacol. 120, 45-54.
(15) Fidder, A., Noort, D., de J ong, A. L., Trap, H. C., de J ong, L. P.,
and Benschop, H. P. (1996) Monitoring of in vitro and in vivo
exposure to sulfur mustard by GC/MS determination of the
N-terminal valine adduct in Hb after a modified Edman degrada-
tion. Chem. Res. Toxicol. 9, 788-792.
(16) Thier, R., Lewalter, J ., Kempkes, M., Selinski, S., Bruning, T.,
and Bolt, H. M. (1999) Haemoglobin adducts of acrylonitrile and
ethylene oxide in acrylonitrile workers, dependent on polymor-
phisms of the glutathione transferases GSTT1 and GSTM1.
Occup. Environ. Med. 56, 197-202.
(17) Begemann, P., Sram, R. J ., and Neumann, H. G. (2001) Hemo-
globin adducts of epoxybutene in workers occupationally exposed
to 1,3-butadiene. Arch. Toxicol. 74, 680-687.
(18) Passingham, B. J ., Farmer, P. B., Bailey, E., Brooks, A. G. F.,
and Yates, D. W. (1988) 2-Hydroxyethylation of haemoglobin in
man. In Methods for Detecting DNA Damaging Agents in Hu-
mans: Applications in Cancer Epidemiology and Prevention
(Bartsch, H., Hemminki, K., and O’Neill, I. K., Eds.) pp 279-
285, IARC, Lyon, France.
Ack n ow led gm en t. We owe thanks to Dr. Roland
Stenutz for valuable discussions and for performing the
NMR measurements; to Lena Liedgren for NMR mea-
surements; to Anna Malmva¨rn, B. Sc., for skillful labora-
tory work; to Prof. Lars Ehrenberg for valuable discus-
sions; and to Dr. Hans Helleberg, Ioannis Athanasiadis,
J ohanna Minten, B. Sc., and Dr. Thomas Alsberg for their
skillful mass-spectrometric analyses. This work was
supported financially by the Stockholm University Center
for Research on Natural Resources and the Environment,
the Swedish Natural Science Research Council, the
Swedish Environmental Protection Agency, The Founda-
tion for Strategic Environmental Research, and the
Swedish Council for Working Life and Social Research.
(19) Hindsø Landin, H., Osterman-Golkar, S., Zorcec, V., and To¨rn-
qvist, M. (1996) Biomonitoring of epichlorohydrin by hemoglobin
adducts. Anal. Biochem. 240, 1-6.
(20) To¨rnqvist, M., and Kautiainen, A. (1993) Adducted proteins for
identification of endogenous electrophiles. Environ. Health Per-
spect. 99, 39-44.
(21) Tareke, E., Rydberg, P., Karlsson, P., Eriksson, S., and To¨rnqvist,
M. (2000) Acrylamide: A cooking carcinogen? Chem. Res. Toxicol.
13, 517-522.
(22) Kautiainen, A., Osterman-Golkar, S., and Ehrenberg, L. (1986)
Misincorporation of alkylated amino acids into hemoglobinsa
possible source of background alkylations. Acta Chem. Scand. B
40, 453-456.
(23) To¨rnqvist, M. (1994) Epoxide adducts to N-terminal valine of
hemoglobin. Methods Enzymol. 237, 650-657.
(24) Rydberg, P., Lu¨ning, B., Wachtmeister, C. A., and To¨rnqvist, M.
(1993) Synthesis and characterization of N-substituted valines
and their phenyl-thiohydantoins and pentafluorophenyl-thiohy-
dantoins. Acta Chem. Scand. 47, 813-817.
(25) Weith, W. (1875) CHBEAM Chem. Ber. 8, 1524.
(26) Edman, P. (1950) Methods for determination of the amino acid
sequence in peptides. Acta Chem. Scand. 4, 283-298.
(27) Clark, W. M. (1923) in The determination of hydrogen ions,
Williams & Wilkins, Baltimore.
(28) Hepner, B., and Frenkenberg, H. (1932) J . Prakt. Chem. 134,
249-256.
Refer en ces
(1) Ehrenberg, L., Hiesche, K. D., Osterman-Golkar, S., and Wenn-
berg, I. (1974) Evaluation of genetic risks of alkylating agents:
Tissue doses in the mouse from air contaminated with ethylene
oxide. Mutat. Res. 24, 83-103.
(2) Osterman-Golkar, S., Ehrenberg, L., Segerba¨ck, D., and Ha¨ll-
stro¨m, I. (1976) Evaluation of genetic risks of alkylating agents.
II. Haemoglobin as a dose monitor. Mutat. Res. 34, 1-10.
(3) Neumann, H. G., Baur, H., and Wirsing, R. (1980) Dose-response
relationship in the primary lesion of strong electrophilic carcino-
gens. Arch. Toxicol. Suppl. 3, 69-77.
(4) Skipper, P. L., and Tannenbaum, S. R. (1990) Protein adducts in
the molecular dosimetry of chemical carcinogens. Carcinogenesis
11, 507-518.
(5) To¨rnqvist, M., Fred, C., Haglund, J ., Helleberg, H., Paulsson, B.,
and Rydberg, P. (2002) Protein adducts: quantitative and quali-
tative aspects of formation, analysis and application. J . Chro-
matogr. (in press).
(6) Osterman-Golkar, S., and Ehrenberg, L. (1980) Alkylation of
macromolecules for detecting mutagenic agents. Teratog., Car-
cinog., Mutagen. 1, 105-127.
(29) Barrett, G. C., and Penglis, A. A. E. (1985) Edman stepwise
degradation of polypeptides: A new strategy employing mild basic
cleavage conditions. Tetrahedron Lett. 26, 4375-4378.

Formation and Degradation of Phenylthiohydantoins
Chem. Res. Toxicol., Vol. 15, No. 4, 2002 581
(30) Davies, J . S., and Mohammed, K. A. (1984) Chiral analysis of the
reaction stages in the Edman method for sequencing peptides. J .
Chem. Soc., Perkin Trans. 2, 1723-1727.
(31) Blagoeva, I. B., Pojarlieff, I. G., and Kirby, A. J . (1984) Intramo-
lecular nucleophilic attack on carboxylate by ureide anion.
General acid-base catalysis of the alkaline cyclisation of 2,2,3,3-
tetramethylhydantoic acid. J . Chem. Soc., Perkin Trans. 2, 745-
751.
(32) Blagoeva, I. B., Pojarlieff, I. G., and Taschev, D. T. (1989)
Intramolecular nucleophilic attack by urea nitrogen. Reactivity-
selectivity relationships for the general acid-base-catalyzed
cyclisations of ureido acids and esters. J . Chem. Soc., Perkin
Trans. 2, 347-353.
(41) Beesley, R. M., Ingold, C. K., and Thorpe, J . F. (1915) The
formation and stability of spiro-compounds. Part I. Spiro-
compounds from cyclohexane. J . Chem. Soc. 107, 1080-
1107.
(42) Bruice, T. C., and Pandit, U. K. (1960) The effect of geminal
substitution ring size and rotamer distribution on the intramo-
lecular nucleophilic catalysis of the hydrolysis of monophenyl
esters of dibasic acids and the solvolysis of the intermediate
anhydrides. J . Am. Chem. Soc. 82, 5858-5865.
(43) Parill, A. L., and Dolata, D. P. (1996) The “facilitated transition”
hypothesis as an explanation for the gem-dialkyl effect. J . Mol.
Struct. (THEOCHEM) 370, 187-202.
(33) Blagoeva, I. B., Kirby, A. J ., Koedjikov, A. H., and Pojarlieff, I.
G. (1999) Multiple changes in rate-determining step in the acid-
and base-catalyzed cyclizations of ethyl N-(p-nitrophenyl)hydan-
toates caused by methyl substitution. Can. J . Chem. 77, 849-
859.
(34) Bethell, D., Metcalfe, G. E., and Sheppard, R. C. (1965) Kinetics
and mechanism of the Edman reaction. Chem. Commun. 10, 189-
190.
(44) Chang, J . Y. (1978) A novel Edman-type degradation: Direct
formation of the thiohydantoin ring in alkaline solution by
reaction of Edman-type reagent with N-monomethyl amino acids.
FEBS Lett. 91, 63-68.
(45) Abderhalden, E., and Brockmann, H. (1930) The determination
of the constitution of proteines and polypeptides. Biochem. Z. 225,
386-408.
(46) Sriramachari, S., and Chandra, H. (1997) The lessons of Bhopal
(toxic) MIC gas disaster scope for expanding global biomonitoring
and environmental specimen banking. Chemosphere 34, 2237-
2250.
(35) Tarr, G. E. (1977) Improved manual sequencing methods. Methods
Enzymol. 67, 335-357.
(36) Toriba, A., Adzuma, K., Santa, T., and Imai, K. (2000) Develop-
ment of an amino acid sequence and D/L configuration determi-
nation method of peptide with a new fluorescence Edman reagent,
7-methylthio-4-(2,1,3-benzoxadiazolyl) isothiocyanate. Anal. Chem.
72, 732-739.
(37) Allan, G. (1981) Determination of peptide sequences by the Edman
degradation. In Sequencing of Proteins and Peptides (Work, T.
S., and Burdon, R. H., Eds.) pp 174-228, Elsevier Science
Publishers B.V., Amsterdam.
(38) Blomba¨ck, B., Blomba¨ck, M., Hessel, B., and Iwanaga, S. (1967)
Structure of N-terminal fragments of fibrinogen and specificity
of thrombin. Nature 215, 1445-1448.
(39) Sundby, F. (1972) The amino acid sequence of human glucagon.
FEBS Lett. 21, 315-319.
(40) Eide, I., Hagemann, R., Zahlsen, K., Tareke, E., To¨rnqvist, M.,
Kumar, R., Vodicka, P., and Hemminki, K. (1995) Uptake,
distribution, and formation of hemoglobin and DNA adducts after
inhalation of C2-C8 1-alkenes (olefins) in the rat. Carcinogenesis
16, 1603-1609.
(47) Angerer, J ., Goen, T., Kra¨mer, A., and Ka¨fferlein, H. U. (1998)
N-methylolcarbamoyl adducts at the N-terminal valine of globin
in workers exposed to N,N-dimethylformamide. Arch. Toxicol. 72,
309-313.
(48) Mra´z, J ., and Bouskova´, S. (1999) 2,4-Toluenediisocyanate and
hexamethylenediisocyanate adducts with blood proteins: assess-
ment of reactivity of amino acid residues in vitro. Chem.-Biol.
Interact. 117, 173-186.
(49) Sabbioni, G., Hartley, R., Henschler, D., Ho¨llrigl-Rosta, A.,
Koeber, R., and Schneider, S. (2000) Isocyanate-specific hemo-
globin adduct in rats exposed to 4,4′-methylenediphenyl diisocy-
anate. Chem. Res. Toxicol. 13, 82-89.
(50) Gersmann, H. R., and Bickel, A. F. (1971) Autoxidation of ketones
and esters in basic solution. J . Chem. Soc. (B) 11, 2230-2237.
TX000247+