Covalent Modification of Proteins and Peptides by the Quinone Methide from 2-tert-Butyl-4,6-dimethylphenol: Selectivity and Reactivity with Respect to Competitive Hydration

Article abstract of DOI:10.1021/jo962088y

Covalent modification of cellular nucleophiles by electrophilic metabolites has been shown to be an important pathway in the toxicological activity of many xenobiotic compounds. The p-quinone methide (4-allylidene-2,5-cyclohexadien-1-one, QM) oxidative me

Full text of DOI:10.1021/jo962088y

1820
J . Org. Chem. 1997, 62, 1820-1825
Cova len t Mod ifica tion of P r otein s a n d P ep tid es by th e Qu in on e
Meth id e fr om 2-ter t-Bu tyl-4,6-d im eth ylp h en ol: Selectivity a n d
Rea ctivity w ith Resp ect to Com p etitive Hyd r a tion
Paul G. McCracken, J udy L. Bolton,*,† and Gregory R. J . Thatcher*
Department of Chemistry, Queen’s University, Kingston, Ontario K7L 3N6, Canada, and
Department of Medicinal Chemistry & Pharmacognosy, College of Pharmacy, University of Illinois
at Chicago, Chicago, Illinois 60612
Received November 7, 1996X
Covalent modification of cellular nucleophiles by electrophilic metabolites has been shown to be
an important pathway in the toxicological activity of many xenobiotic compounds. The p-quinone
methide (4-allylidene-2,5-cyclohexadien-1-one, QM) oxidative metabolites of the 4-alkylphenols are
a chemical family whose toxicology has been well described. The reactivity of the QM from 2-tert-
butyl-4,6-dimethylphenol (BDMP-QM) with respect to hydration has been studied in the 3 < pH
< 11 range in order to examine selectivity for competitive reaction with biological nitrogen
nucleophiles. The reactivity of BDMP-QM was examined with human hemoglobin A (HbA),
angiotensin-III (AIII), amino acids, and amines. Nucleophilic selectivity (kNu/kH O) and reactivity
2
(kNu) in aqueous solution (7 < pH < 8) for addition of these biological nucleophiles to BDMP-QM
demonstrate (i) that adduct formation is competitive with hydration at low nucleophile concentra-
tions and (ii) that proteins, peptides, and tyrosine show enhanced reactivity toward QMs, relative
to simple amino acids and related compounds. All adducts were characterized by electrospray mass
spectrometry (ESMS). These results provide strong evidence supporting the toxicity of QM
metabolites via facile alkylation of nucleophilic nitrogen sites in proteins and peptides, even when
present in low concentration in aqueous solution under physiological conditions.
In tr od u ction
capacity for redox chemistry.1 Consequently, reactions
of QMs in biological systems are characterized by non-
enzymic Michael additions at the exocyclic methylene
carbon generating benzylic adducts with biological
nucleophiles.1a The description of a QM as a resonance-
stabilized carbocation emphasizes the importance of the
cationic, aromatic resonance contributor and the role of
inter- and intramolecular interactions in regulating
electron deficiency at the methylene carbon and hence
QM reactivity toward nucleophiles.6,7
The quinone methide (QM) oxidative metabolites of the
4-alkylphenols form a chemical family whose toxicology
has been well reported.1 Roles for QM metabolites have
been demonstrated in several endogenous biochemical
processes including melanization, sclerotization, and
lignin formation.1,2 In addition, QMs have been impli-
cated as the potential ultimate cytotoxins responsible for
the biological effects of antitumor drugs, antioxidants,
and several naturally occurring hydroxylated aromatic
compounds.1 They are also of interest as synthetic
intermediates in cyclization reactions,3 as selective DNA
alkylators,4 and as side products of prodrug strategies.5
In all cases, the mechanism of toxicological action is
proposed to be via QM alkylation of biological nucleo-
philes. Indeed, convincing evidence has been presented
to show that QM metabolites could become covalently
attached to cellular proteins.1
Reaction of QMs with various nucleophiles, although
often under nonaqueous, nonphysiological conditions, has
been well studied.8 While these findings demonstrate the
wide-ranging reactivity of QMs toward O, S, and N
nucleophiles, the issue of chemical selectivity is crucial.
A QM metabolite may have one of three fates as an
alkylating agent: (1) the QM may react with and not
escape from the redox protein in which it was formed;
(2) the QM may undergo hydration; and (3) the QM may
escape to alkylate a biological nucleophile.9 The efficacy
of QM trapping by glutathione and the isolation of QM-
SG conjugates from rat hepatocytes implies high selectiv-
ity toward S nucleophiles.1 But what of nitrogen nucleo-
The structural difference between a QM and a quinone
is the replacement of one carbonyl by a methylene or
substituted methylene group. This substitution results
in a much more reactive electrophile and a diminished
† University of Illinois at Chicago.
X Abstract published in Advance ACS Abstracts, February 15, 1997.
(1) (a) Peter, M. G. Angew. Chem., Int. Ed. Engl. 1989, 28, 555. (b)
Thompson, D. C.; Thompson, J . A.; Sugumaran, M.; Moldeus, P. Chem.-
Biol. Interact. 1993, 86, 129.
(6) Richard, J . P.; Amyes, T. L.; Bei, L.; Stubblefield, V. J . Am. Chem.
Soc. 1990, 112, 9513.
(7) Hulbert, P. B.; Grover, P. L. Biochem. Biophys. Res. Commun.
1983, 117, 129.
(2) (a) Sugumaran, M. FEBS Lett. 1991, 233, 4. (b) Monks, T. J .;
Hanzlik, R. P.; Cohen, G. M.; Ross, D.; Graham, D. G. Toxicol. Appl.
Pharmacol. 1992, 112, 2.
(3) (a) Angle, S. R.; Turnbull, K. D. J . Am. Chem. Soc. 1989, 111,
1136. (b) Angle, S. R.; Arnaiz, D. O.; Boyce, J . P.; Frutos, R. P.; Louie,
M. S.; Mattson-Arnaiz, H. L.; Rainier, J . D.; Turnbull, K. D.; Yang, W.
J . Org. Chem. 1994, 59, 6322.
(8) (a) Turner, A. B. Q. Rev., Chem. Soc. London. 1964, 18, 347. (b)
Leary, G.; Miller, I. J .; Thomas, W.; Woolhouse, A. D. J . Chem. Soc.,
Perkin Trans. 2 1977, 1737. (c) Wagner, H.-U.; Gompper, R. In The
Chemistry of Quinonoid Compounds; Patai, S., Ed.; Wiley: New York,
1974; pp 1145. (d) Tajima, K.; Yamamoto, K.; Mizutani, T. Biochem.
Pharmacol. 1985, 34, 2109. (e) Wan, P.; Barker, B.; Diao, L.; Fischer,
M.; Shi, Y.; Yang, C. Can. J . Chem. 1996, 74, 465.
(4) Chatterjee, M.; Rokita, S. E. J . Am. Chem. Soc. 1994, 116, 1690.
(5) (a) Thomson, W.; Nicholls, D.; Irwin, W. J .; Al-Mushadani, J .
S.; Freeman, S.; Karpas, A.; Petrik, J .; Mahmood, N.; Hay, A. J . J .
Chem. Soc., Perkin Trans. 1 1993, 1239. (b) Routledge, A.; Walker, I.;
Freeman, S.; Hay, A.; Mahmood, N. Nucleosides Nucleotides 1995, 14,
1545.
(9) In addition, QM solutions made by chemical oxidation of phenols
often contain a small amount of dimerized material. Dimerization as
a result of oxidation of hindered phenols has been well described: (a)
Cook, C. D.; Nash, N. G.; Flanagan, H. R. J . Am. Chem. Soc. 1955, 77,
1783. (b) Cook, C. D.; Norcross, B. E. J . Am. Chem. Soc. 1956, 78, 3797.
(c) Bauer, R. H.; Coppinger, G. M. Tetrahedron 1963, 19, 1201.
S0022-3263(96)02088-9 CCC: $14.00 © 1997 American Chemical Society

Covalent Modification of Proteins and Peptides
J . Org. Chem., Vol. 62, No. 6, 1997 1821
Sch em e 1
Sch em e 2
nucleophile solutions to achieve a final concentration of 0.2
mM QM and no more than 10% MeCN (v:v). Reactions were
allowed to proceed overnight at room temperature.
philes in aqueous solution? The observation of myoglobin
(Mb) alkylation by 2-tert-butyl-6-methyl-4-methylene-2,5-
cyclohexadienone (BDMP-QM, Scheme 1, R1 ) But, R2
) Me), despite the absence of cysteine residues in this
protein, demonstrates the ability of a QM to form adducts
2 by alkylation of nitrogen nucleophiles.10 But, what is
the relative reactivity and selectivity of a QM toward the
amine nucleophiles of peptides and proteins, in competi-
tion with background hydration?
Ch r om a togr a p h y. Aliquots (400 µL) of the product mix-
tures were analyzed by reversed-phase HPLC at 280 nm using
a 2.1 mm × 25 cm Ultrasphere C18 column (Beckman) using a
linear gradient, from 3% B to 90% B in 45 min (A ) 0.1% TFA,
B ) MeCN, 0.085% TFA) at 0.2 mL/min, on a Shimadzu LC-
10A gradient HPLC with SPD-10AV UV detector. All experi-
ments were run in triplicate, and standard errors ranged from
1 to 5%. Results were reproducible between batch prepara-
tions of BDMP-QM within 5%. The identities of the putative
conjugates of BDMP-QM with nucleophile observed in HPLC
chromatograms were confirmed by peak collection and subse-
quent ESMS analysis, as well as by chromatographic equiva-
lence with synthetic standards in selected instances (Scheme
2). The presence of BDMP phenol as well as small (e5% of
total products by integration of peak areas) amounts of dimer9
were observed consistently throughout all experiments.
Syn th esis a n d Ch a r a cter iza tion of P r od u cts.
2-ter t-Bu tyl-6-m eth yl-4-(h ydr oxym eth yl)ph en ol (1). Five
hundred mL of DDW was added to 500 mL of a 2.5 mM MeCN
solution of BDMP-QM (1.25 mmol), and the mixture was
stirred for 8 h. The MeCN was then removed under vacuum,
and the remaining aqueous solution was extracted with two
volumes of diethyl ether. The ether extracts were combined
and washed with two volumes of brine and dried with sodium
sulfate and then the ether removed under vacuum. The
product was purified using flash chromatography (hexanes/
EtOAc) and obtained as 200 mg of white powder (1 mmol,
82%); mp 93-95 °C; 1H NMR (200 MHz, DMSO-d6) δ 1.34 (9H,
s, But), 2.15 (3H, s, ArMe), 4.31 (2H, d, J CHOH ) 5.50 Hz, benzyl
CH2), 4.86 (1H, t, J OHCH ) 5.50 Hz, benzyl OH), 6.86 (1H, s,
ArH), 6.96 (1H, s, ArH), 7.92 (1H, s, phenol OH) ppm; 1H NMR
(400 MHz; CDCl3) δ 1.40 (9H, s, But), 2.24 (3H, s, ArMe), 4.55
(2H, s, benzyl CH2), 4.78 (1H, s, OH), 7.01 (1H, d, J meta ) 1.59
Hz, ArH), 7.12 (1H, d, J meta ) 1.98 Hz, ArH) ppm; 13C NMR
(50 MHz, CD3OD) J -mod δ 16.68 (ArMe), 30.31 (But Me), 35.62
(But C), 65.65 (benzyl CH2), 124.98 (ArC), 126.08 (ArC), 128.82
(ArC), 133.13 (ArC), 138.14 (ArC-4), 154.32 (ArC-1) ppm; CI+
GC/MS (isobutane) m/z 195 (2) (M + H)+, 194 (15) M+, 193
(100) (M - H)+ (C12H18O2).
Selectivity is intimately linked with the toxicity of the
alkylphenols, which has been correlated with the life-
times of their corresponding QMs.11-13 The lifetime of a
QM in aqueous solution is governed by its rate of
hydration to form benzyl alcohol 1. BDMP-QM was
reported as a highly hepatotoxic QM, possibly because
the lifetime of this QM, intermediate relative to other
QMs, allows escape from the parent redox protein, but
then subsequent reaction with cellular nucleophiles to
form adducts (2). A comparative study of the reactivity
and selectivity of BDMP-QM toward hydration and
toward alkylation of the nitrogen nucleophiles of proteins,
peptides, amino acids, and related compounds is required.
A pH-rate profile, 3.5 < pH < 10.2, has been obtained
for BDMP-QM using various oxygen buffers. Nucleo-
philic selectivity and reactivity in aqueous solution (7 <
pH < 8) for reaction of BDMP-QM with human hemo-
globin A (HbA), angiotensin-III (AIII), trityrosine tri-
peptide (Tyr3), amino acids, and amines have been
examined using kinetic analysis and product identifica-
tion by electrospray mass spectrometry (ESMS). The
results provide strong evidence supporting a toxicity
pathway for QM metabolites via facile alkylation of
nucleophilic nitrogen sites in proteins and peptides.
Exp er im en ta l Section
BDMP -NR-Tyr (3). A 20 mM solution of tyrosine (Tyr)
was prepared in 50 mL of water, pH ) 7.4. A QM solution in
MeCN (2.5 mM, 4 × 50 mL) was added in aliquots, and the
MeCN was removed after each addition under vacuum. The
product mixture was extracted with two volumes of ether and
then purged of organic solvent by bubbling nitrogen through
the solution. The remaining aqueous phase was passed
through a C-18 extraction cartridge (Baker). The cartridge
was washed with water, and the adducts were then eluted with
MeOH. The MeOH was removed under vacuum, and the
residue was redissolved in MeOH for chromatographic puri-
fication on a silica gel column with EtOAc/MeOH (1:4) as
eluant. Final purification was carried out by HPLC using an
Ultrasphere ODS column (10 × 250 mm, Beckman) with a flow
rate of 3.5 mL/min and the same mobile-phase composition
described above for analytical work: 1H NMR 200 MHz
(DMSO-d6) δ 1.31 (9H, s, But), 2.11 (3H, s, ArMe), 2.79 (1H,
dd, J ab ) 21 Hz, J R ) 5.91 Hz, Tyr benzyl CHa), 2.82 (1H, dd,
Ma ter ia ls. All chemicals were purchased from Aldrich
(Milwaukee, WI) or Sigma (St. Louis, MO) unless stated
otherwise. BDMP-QM was synthesized as described previ-
ously by lead oxidation of the corresponding phenol.12 HPLC
solvents used were from BDH (Toronto, ON). Purified human
Hb A in the CO form (HemAzero) was obtained from Hemosol
Inc. (Etobicoke, Ont.). Distilled deionized water was purified
using the Barnstead MegaPure MP-1 water purification
system.
Alk yla tion s for Nu cleop h ilic Selectivity Stu d y. Stock
solutions of nucleophiles were adjusted to pH 7.4 and diluted
in water to achieve final molar ratios Nu:QM of 20:1-1:5.
Stock solution concentrations were determined spectrophoto-
metrically where possible. The concentration of MeCN solu-
tions of BDMP-QM was determined spectrophotometrically,
12
using ꢀ286 ) 22 400 M-1 cm-1
.
BDMP-QM was added to the
J ba ) 21 Hz, J R ) 5.91 Hz, Tyr benzyl CHb), 3.14 (1H, t, J R
)
(10) Bolton, J . L.; Le Blanc, J . C. Y. Biol. Mass Spec. 1993, 22, 666.
(11) Bolton, J . L.; Valerio, L. G.; Thompson, J . A. Chem. Res. Toxicol.
1992, 5, 816.
(12) Filar, L. J .; Winstein, S. Tetrahedron Lett. 1960, 25, 9.
(13) Thompson, D. C.; Perera, K.; Krol, E. S.; Bolton, J . L. Chem.
Res. Toxicol. 1995, 8, 323.
5.91 Hz, Tyr CHR), 3.54 (1H, dd, J ab ) 28 Hz, BDMP benzyl
CHa), 3.59 (1H, dd, J ba ) 28 Hz, BDMP benzyl CHb), 6.64 (2H,
d, J ortho ) 8.39 Hz, Tyr ArH), 6.77 (1H, s, BDMP ArH), 6.91
(1H, s, BDMP ArH), 7.0 (2H, d, J ortho ) 8.39 Hz, Tyr ArH),

1822 J . Org. Chem., Vol. 62, No. 6, 1997
McCracken et al.
8.08 (s, 1H, BDMP phenol OH), 9.24 (s, 1H, Tyr phenol OH)
ppm; 1H NMR 200 MHz (D2O) δ 1.18 (9H, s, But), 2.00 (3H, s,
ArCH3), 2.57 (2H, d, J R ) 6.6 Hz, Tyr benzyl CH2), 3.08 (1H,
t, J R ) 6.8 Hz, Tyr CHR), 3.26 (1H, d, J ab ) 12.5 Hz, benzyl
CHa), 3.45 (1H, d, J ba ) 12.5 Hz, benzyl CHb), 6.39 (2H, d,
J ortho ) 8.4 Hz, Tyr ArH), 6.75, (1H, d, J meta ) 1.9 Hz, BDMP
ArH), 6.77 (2H, d, J ortho ) 8.4 Hz, Tyr ArH), 6.90 (1H, d, J meta
) 2 Hz, BDMP ArH) ppm; ES+ MS m/z 358 (100), MH+
(C21H28O4N).
BDMP -TRIS (4). The QM adduct of tris(hydroxymethyl)-
aminomethane (TRIS) was synthesized and purified as de-
scribed above for 3: 1H NMR (200 MHz, D2O) δ 1.35 (9H, s,
But), 2.21 (3H, s, ArMe), 3.81 (6H, s, TRIS CH2), 4.16 (2H, s,
benzyl CH2), 7.15 (1H, d, J meta ) 1.96 Hz, ArH), 7.26 (1H, d,
J meta ) 1.96 Hz, ArH) ppm; 13C NMR (50 MHz, CD3OD) J -mod
δ 16.48 (ArMe), 30.11 (But Me), 35.75 (But C), 59.23 (benzyl
CH2), 61.03 (Tris C), 67.57 (Tris CH2), 123.53 (ArC), 126.67
(ArC), 127.97 (ArC), 131.59 (ArC), 138.78 (ArC-4), 156.17 (ArC-
1) ppm; UV-vis λmax (H2O) 272 nm; ES+ MS m/z 298 (100)
MH+ (C16H28O4N).
F igu r e 1. pH-rate profile for BDMP-QM hydrolytic decom-
position from data in Table 1, at 22 °C, I ) 0.1, substrate
concentration 45 µM, formate (9), acetate (0), phosphate (2),
borate (4), carbonate (b). The line drawn is fit to the equation,
+
-
+
-
+
ko ) kH [H3O ] + kH O[H2O] + kOH [OH ], where kH
)
O
O
Rechromatography of 1, 3, and 4 using the analytical (2.1
mm × 25 cm) C18 HPLC column confirmed the identities of
the corresponding putative adducts from the kinetic experi-
ments based on identical retention times.
3
2
3
200 M-1 s-1, kOH ) 50 M-1 s-1, for Kw ) 10-14 M, kH ) 2.7
-
O
× 10-4 M-1 s-1. The latter value was used to corres2pond to
previous estimates.11
ESMS. Peaks collected from HPLC experiments were
analyzed directly. All samples were delivered to the Fisons
VG Quattro ESMS probe using a Harvard syringe pump at
8.3 µL/min and were scanned at 84 m/z per s. Alkylated
protein product mixtures were acidified and diluted to 20 µM
in protein, 5% (v:v) MeCN, 2% (v:v) HOAc. Ion evaporation
produced an “envelope” series of ions for each protein species.
The identity of the individual charge state of each ion in each
series as well as the molecular mass of each protein species
was computed using the algorithm of Mann et al.14 The
addition of BDMP-QM to a potential nucleophile results in a
net increase in molecular weight of 177 amu.
Nu cleop h ilic Selectivity. Nucleophilic selectivities kNu
/
kH were determined for each nucleophile using eq 16 as the
O
2
slopes in the plots shown in Figure 4. The ratio A2/A1 is the
A2ꢀ1/A1ꢀ2 ) kNu[Nu]/kH O[H2O] (1)
2
F igu r e 2. Buffer dilution plot of pseudo-first-order rate
constant, kobs, vs [borate], at pH 8.36 (2), 8.63 (4), 8.79 (9),
8.98 (0), 9.12 ((), I ) 0.1, substrate concentration 45 µM, at
22 °C. This plot is representative of those obtained for the other
buffers employed.
peak area ratio of nucleophilic adduct 2 to benzyl alcohol 1
from C18 HPLC analysis of product mixtures at 280nm, and
ꢀ1/ꢀ2 is the ratio of extinction coefficients at 280 nm. The
extinction coefficient ꢀ1 for the BDMP-QM hydration product
in water was determined to be ꢀ1 ) 0.68 mM-1 cm-1 using a
pure synthetic sample (vide supra). Extinction coefficients for
nucleophilic adducts ꢀ2 were determined additively. For
instance, for the AIII-QM adduct, ꢀAIII ) 0.91 mM-1 cm-1 from
a pure synthetic sample and ꢀAIII-QM was determined as ꢀAIII-QM
kobs, vs total buffer concentration were found to be linear for
all reactions of BDMP-QM studied, yielding rate constants
kcat. (M-1 s-1) from the slope of the correlation (Figure 2). Plots
of kcat. vs the fraction of buffer as free base, fB, were used to
-
) 0.91 + 0.68 ) 1.6 mM-1 cm-1
.
For Tyr, ꢀTyr ) 1.4 mM-1
estimate the general base, kA , and general acid, kAH, contribu-
cm-1
.
tions of the buffer to catalysis (eqs 2 and 3, Figure 3). The
pH-rate profile (log ko vs pH) was fitted to the standard
relationship (eq 4).
Hyd r a tion Kin etics. All buffers were prepared with
distilled, deionized water. The ionic strength of diluted buffers
was maintained at I ) 0.1 M with NaCl for monobasic anions
and with Na2SO4 for dibasic anions. Buffers of formic acid
(pH 3.5-4.4), acetic acid (pH 4-5.3), phosphoric acid (pH 5.6-
7.6), boric acid (pH 8.4-9.1), and carbonic acid (pH 9.3-10.2)
were used to determine the pH-rate profile. A freshly
prepared solution of QM in MeCN (30 µL) was added to 970
µL of buffer in a quartz cuvette, producing a QM concentration
of approximately 45 µM. The disappearance of BDMP-QM
was followed, under pseudo-first-order conditions, by monitor-
ing the decrease in absorbance at 286 nm12 using a Hewlett-
Packard 8452A diode array spectrophotometer. Rate constants
were determined for triplicate reactions run for at least 6 half-
lives. The temperature was maintained at 22 °C with a
thermostated cell holder. The pH of each buffer was measured
using a Beckman Φ32 pH meter with a Beckman 39848
electrode. Experiments were carried out at six different buffer
concentrations in order to extrapolate to zero buffer and obtain
the buffer-independent rate constant ko (s-1) at each pH value.
Plots of the experimental pseudo-first-order rate constants,
kobs ) ko + kcat.
(2)
(3)
(4)
kcat. ) kA-[A-] + kAH[AH]
ko ) kH O+[H O+] + kH O[H2O] + kOH-[OH-]
3
3
2
Resu lts a n d Discu ssion
Hyd r a tion . The pH-rate profile for the addition of
water to BDMP-QM was obtained by monitoring disap-
pearance of the QM by UV in various oxygen buffers. The
profile obtained is similar in appearance to that previ-
ously described for a QM15 (Figure 1). There are three
distinctive regions corresponding to dominant acid-
catalyzed (pH < 5), pH-independent (5 < pH < 9), and
base-catalyzed reaction pathways (pH > 9) (Figure 1,
(15) Hemmingson, J . A.; Leary, G. J . Chem. Soc., Perkin Trans. 2
1975, 1584.
(14) Mann, M.; Meng, C. K.; Fenn, J . B. Anal. Chem. 1989, 61, 1702

Covalent Modification of Proteins and Peptides
J . Org. Chem., Vol. 62, No. 6, 1997 1823
Ta ble 1. Ra te Con sta n ts for BDMP -QM Br ea k d ow n Extr a p ola ted to Zer o Bu ffer
buffer
formate
3.75
acetate
4.76
0.34 × 0.04
phosphate
7.20
borate
9.24
0.22 × 0.02
carbonate
10.33
a
pKa
b
kA /q (M-1 s-1
)
0.41 × 0.03
0.26 × 0.01
3.91 × 0.4
-
formate
acetate
phosphate
102ko (s-1
borate
carbonate
pH
102ko (s-1
)
pH
102ko (s-1
)
pH
)
pH
102ko (s-1
)
pH
102ko (s-1
)
3.50
3.84
4.06
4.27
8.60 ( 0.38
4.84 ( 0.09
3.58 ( 0.19
2.77 ( 0.19
4.00
4.34
4.69
5.10
5.29
3.03 ( 0.14
2.49 ( 0.20
1.64 ( 0.10
1.64 ( 0.06
1.49 ( 0.50
5.56
5.95
6.16
6.36
6.86
7.63
1.51 ( 0.02
1.46 ( 0.02
1.39 ( 0.03
1.30 ( 0.02
1.21 ( 0.30
1.10 ( 0.10
8.36
8.63
8.79
8.98
9.12
1.26 ( 0.02
1.26 ( 0.03
1.31 ( 0.02
1.37 ( 0.02
1.35 ( 0.03
9.57
9.81
9.98
10.03
10.19
1.46 ( 0.70
1.51 ( 0.25
1.36 ( 0.38
2.03 ( 0.50
2.54 ( 0.50
a
pKa values taken from: Dawson, R. M. C.; Elliott, D. C.; Elliott, W. H.; J ones, K. M.; Data for Biochemical Research, 2nd ed.; Oxford
b
University Press: London, 1969. q ) the number of basic sites on the buffer anion.
detection and assessed to be present in similar quantities
independent of added nucleophile. All adduct peaks
observed in the HPLC chromatograms and used for
quantification of kNu were isolated and characterized as
the 1:1 QM-nucleophile adducts by ESMS. The hydra-
tion product (1) was identified in the HPLC chromato-
gram by comparison to an authentic sample.
Benzylation of nucleophiles by QMs is well-docu-
mented.16 Contrary to QMs derived from catechols, there
is no competitive o-quinone reaction for QMs derived from
simple phenols. Amino acids possessing nucleophilic side
chains may be alkylated at the NR and/or side chain N,
O, or S. The QM conjugates of Tyr (3) and Tris (4) were
characterized by NMR, since competitive O-alkylation
must be considered. Comparison of the aromatic regions
1
of the H NMR spectra of 1, 3, 4, and BDMP confirms
addition to the methylene C of BDMP-QM (i.e., benzy-
lation). That QM alkylation occurs at the N of Tris rather
than one of the oxygens is evidenced by the equivalence
of the six methylene protons of Tris in the isolated adduct
F igu r e 3. Plot of kcat. versus fraction of buffer in base form
(fB) for phosphate buffer, as representative example of plots
for all buffers employed.
Table 1). Plots of kobs vs total buffer concentration yielded
rate constants kcat for each pH studied (Figure 2). Plots
of kcat. versus fraction of buffer in base form clearly
indicate participation of the base form in breakdown of
BDMP-QM for each oxygen buffer employed (Figure 3).
It is probable that the rate constant for buffer base
catalysis, kA-, obtained from these plots describes mixed
general base and nucleophilic catalysis, since a Bronsted
plot of log (kA-/q) versus [pKa + log (p/q)] (p ) the
number of acidic protons on the buffer acid, q ) the
number of basic sites on the buffer anion) for these
buffers shows considerable scatter (not shown), compat-
ible with a mechanism other than pure general base
catalysis (Table 1). It is not unreasonable, at the
relatively high concentrations of buffer used, that buffer
anions compete with water for nucleophilic addition to
BDMP-QM. Indeed, comparison of kA- values for buffer
catalysis with second-order rate constants obtained for
nucleophilic addition of nitrogen nucleophiles (Table 2)
to BDMP-QM supports this hypothesis. For example,
the second-order rate constant, kNu, for nucleophilic
addition by Tris is lower than the buffer rate, kA-, for
some buffers (Tables 1 and 2). However, a stable buffer
adduct has not been identified by HPLC analysis of the
hydration products.
1
and by the observed methylene chemical shifts. The H
NMR of 3 is incompatible with alkylation at a Tyr ring-
C. Moreover, the relative upfield shift of the benzylic
protons of the adduct (3.3, 3.5 ppm) is definitive for the
N-adduct (cf: 1 4.55, 4.31 ppm; 4 4.2 ppm; lysine-NR-
adduct 3.88, 4.01 ppm;17 histidine-NR-adduct 3.88, 4.02
ppm17). The spectroscopic data are compatible with the
absence of any adduct formation with NR-acetyl-L-Tyr
ethyl ester (vide infra).
Reactions with the “nitrogen nucleophiles,” peptides,
amino acids, and related species, were allowed to proceed
to completion, and product ratios for adduct formation
versus hydration (Scheme 1) were measured by HPLC
with UV-detection to yield selectivity ratios (kNu/kH O).
2
The value of kH was determined spectrophotometri-
2O
cally,11 and absolute values for kNu were thus derived.
The more rigorous derivation of kNu was applied using
varying nucleophile concentrations (40 µM to 4 mM), as
employed by Richard et al. (Table 2, Figure 4).6 The first
point of interest is the high selectivity for all nitrogen
nucleophiles studied over reaction with water, which
ranges from a factor of 104 to a factor of 106 (Table 2).
The second-order rate constant, for example, for reaction
of BDMP-QM with His was measured as 17 M-1 s-1, as
compared to 2.7 × 10-4 M-1 s-1 for hydration.
Nitr ogen Nu cleop h ile Selectivity. In order to
examine reaction of BDMP-QM in amino acid, peptide,
and protein solutions, unbuffered solutions were em-
ployed with pH measured before and after reaction at 7
< pH < 8. Since this is in the pH-independent region
for hydration, small variations in pH are not of concern
(Figure 1).
The heptapeptide angiotensin III (AIII, Arg-Val-Tyr-
Ile-His-Pro-Phe) was chosen as a readily available,
representative oligopeptide containing various potential
nucleophilic nitrogen sites. The reactivity of AIII toward
(16) (a) Lewis, M. A.; Graff, D. A.; Bolton, J . L.; Thompson, J . A.
Chem. Res. Toxicol. 1996, 9, 1368. (b) Bolton, J . L.; Comeau, E.;
Vukomanovich, V. Chem.-Biol. Interact. 1995, 95, 279.
The major side products were isolated by HPLC and
monitored in each reaction mixture by HPLC with UV-
(17) Bolton, J . L.; Thompson, J . A. Unpublished results.

1824 J . Org. Chem., Vol. 62, No. 6, 1997
McCracken et al.
Ta ble 2. Secon d -Or d er Ra te Con sta n ts for Ad d u ct F or m a tion of BDMP -QM w ith Nu cleop h iles
nucleophile
H2O
Tris
His
6.04, 9.33
17.3 ( 0.9
HisNAc
NDc
10.1 ( 0.5
LysNAc
ND
18.6 ( 2.8
Arg
8.99, 12.84
17.2 ( 1.0
Agm
ND
23.7 ( 3.8
Cys
8.35, 10.46
3320 ( 66
8.3b
a
pKa
15
kNu (M-1 s-1
)
2.7 × 10-4 e
3.73 ( 0.09
d
nucleophile
Ile
9.76
7.9 ( 4.7
Trp
9.44
14.1 ( 2.8
Trpa
10.2
3.48 ( 0.56
Tyr
9.11, 10.13
45.0 ( 3.0
Tyra
9.3, 10.4
4.62 ( 0.97
Tyr3
ND
274.6 ( 10.7
AIII
ND
123.9 ( 7.8
a
pKa
kNu (M-1 s-1
)
d
a
pKa values taken from Dawson, R. M. C.; Elliott, D. C.; Elliott, W. H.; J ones, K. M.; Data for Biochemical Research, 2nd ed.; Oxford
b
University Press: London, 1969. Good, N. E.; Winget, G. D.; Winter, W.; Connolly, T. N.; Izawa, S.; Singh, R. M.M. Biochemistry 1966,
5, 467. c Not determined. Derived from kNu/kH values from product studies. e Reference 11.
d
2O
F igu r e 5. ESMS mass spectrum of AIII-QM adduct. The
adduct was collected from HPLC separation of the product
mixture and analyzed directly as described in the Experimen-
tal Section. The adduct is predicted to have an empirical
formula C58H82N12O10, FW ) 1106 (M). The unmodified peptide
has FW ) 931.
is identified as the 1:1 AIII-QM adduct (Figure 5).
Consideration of the chromatogram of the reaction
products shows that this adduct represents >95% of
AIII-QM adducts present in solution. Furthermore,
reference to the reactivity trend, Agm > Arg > LysNAc
>
HisNAc > HisNAc > TyrNAc (Table 2), strongly suggests that
the N-terminus is the primary site of modification.
Further characterization is in progress using a combina-
tion of proteolytic digests and LC/ESMS/MS.
Tryptophan (Trp) was studied as an amino acid with
side chain nitrogen and Tyr because of its presence in
AIII. Trp is poorly reactive, whereas surprisingly, free
Tyr has the highest reactivity toward BDMP-QM of all
the free amino acids and derivatives studied, with the
exception of cysteine (Cys) (Table 2, Figure 4). The
unlikely possibility that the reactivity of AIII was due to
the side chain phenolic nucleophile was ruled out by the
absence of any adduct formation with NR-acetyl-L-Tyr
ethyl ester. Alkylation of Tyr itself occurs at N as
confirmed by the synthesis, purification and character-
ization of the Tyr-QM adduct (3). Tyramine (Tyra), the
decarboxylated analog of Tyr, has an activity that is
reduced 10-fold compared to the parent amino acid.
Reduced reactivity is also seen with tryptamine (Trpa),
relative to Trp. However, the tripeptide Tyr3 was exam-
ined and found to have remarkably high reactivity toward
BDMP-QM: 6-fold greater than Tyr and twice that of
AIII.
F igu r e 4. (a) Plot of extent of adduct formation with respect
to nucleophile concentration: A2ꢀ1/A1ꢀ2 versus [nucleophile].
Nucleophiles used: (a) AIII ((), (b) Agm (9), (c) His (0), (d)
LysNAc (2), (e) Arg (4), (f) HisNAc (O), (g) tris (b). Best fit lines
are shown for each data series. (b) Plot of extent of adduct
formation with respect to nucleophile concentration: A2ꢀ1/A1ꢀ2
versus [nucleophile]. Nucleophiles used: (h) Tyr3 (0), (i) Tyr
(9), (j) Trp (O), (k) Ile (b), (l) Tyra (4), (m) Trpa (2). Best fit
lines are shown for each data series.
BDMP-QM is substantial. In comparison to its con-
stituent nitrogenous-nucleophilic amino acid side chains,
modeled by NR-acetyl-L-lysine (LysNAc), NR-acetyl-L-his-
tidine (HisNAc), and arginine (Arg), AIII is more reactive
by factors of 7, 10, and 7, respectively (Table 2, Figure
4). Moreover, agmatine (Agm), which is a good analogue
of the nucleophilic N-terminus of AIII, is only one-fifth
as reactive as AIII toward BDMP-QM. Addition of
BDMP-QM to AIII leads to a small amount of precipita-
tion (<8% w/w), as is observed to varying degrees on
addition of QMs to many peptides and proteins. The
ESMS chromatogram of the recovered, redissolved pre-
cipitate shows a mixture of AIII and 1:1, 1:2, and 1:3
adducts with BDMP-QM in an ion current intensity
ratio of 18:8:3:1. However, the major adduct peak
isolated from HPLC separation of the reaction solution
BDMP-QM is poorly hydrophilic, suggesting the pos-
sibility of a local concentration effect in which the local
concentration of BDMP-QM with hydrophobic amino
acids or with hydrophobic interiors of peptides and
proteins would be greatly increased by aggregation.
Thus, aggregation would lead to increased reaction rates.
(18) Freudenberg, K.; Werner, H. K. Chem. Ber. 1964, 97, 579.

Covalent Modification of Proteins and Peptides
J . Org. Chem., Vol. 62, No. 6, 1997 1825
QM mixtures at molar ratios from 1:1 to 1:8 gave spectra
similar to that reported for Mb.10 At a 1:1 ratio, it was
found that one unit of QM is added selectively to the R
chain of Hb A (Figure 6). At higher concentrations of
BDMP-QM, a second unit is added to the â chain. This
result demonstrates selectivity for the HbA R chain by
BDMP-QM and an apparent reactivity trend: HbA R
chain > HbA â chain > Mb. The simplest explanations
for this selectivity are again based upon either (a)
reactivity of protein side chains and catalytic motifs or
(b) hydrophobic binding provided by the protein tertiary
structure. The R and â chains of human HbA possess
one and two cysteine residues, respectively, whereas Mb
possesses no cysteine. In separate experiments, cysteine
was determined to be more reactive toward BDMP-QM
adduct formation than any of the nitrogen nucleophiles
assayed by at least a factor of 10 (Table 2). The high
reactivity of protein-thiol toward BDMP-QM may ex-
plain the high reactivity demonstrated by HbA compared
to Mb. However, the intrinsic reactivity of cysteine does
not explain the selectivity of BDMP-QM for the R chain,
which was observed in the ESMS studies. There are, of
course, several examples of proteins binding small hy-
drophobic arenes in the vicinity of reactive lysine resi-
dues, for example: (i) HbA, which catalyzes p-nitrophenyl
acetate (PNPA) hydrolysis,22 and (ii) serum albumin,
which catalyzes PNPA hydrolysis and the Kemp elimina-
tion reaction of a benzisoxazole.23
F igu r e 6. ESMS spectrum of HbA after alkylation by
equimolar BDMP-QM at pH 7.4. Spectrum shows the forma-
tion of an adduct of 177 Da, corresponding to the addition of
one QM, to HbAR in preference to HbAâ. b ) HbAR, Mr
)
15122; O ) (HbAR+ BDMP-QM), Mr ) 15 299; 9 ) HbAâ, Mr
) 15 861.
This possibility was examined using isoleucine (Ile), the
amino acid with the most hydrophobic side chain. How-
ever, in this case, relatively low reactivity was observed.
The rate of addition of alcohols to QMs has been shown
to depend to a substantial extent on the pKa of the
nucleophile.18 However, no such simple correlation is
seen with the amine nucleophiles studied herein. The
selectivity of BDMP-QM for the primary amine Tris is
104-fold that for water, and selectivities increase up to 2
× 105 for Tyr, 5 × 105 for AIII, and 1 × 106 for Tyr3.
Peptide glycation is a biologically highly important
protein modification in which reaction is site-specific. In
this case, high rates of N-glycation have been observed
when His is proximal to the N-terminus residue, possibly
from intramolecular catalysis.19,20 Similar site-specific
catalysis is possible for peptide reaction with QMs.
Hydration studies demonstrated the importance of
specific acid and specific and general base catalysis for
BDMP-QM and indicate a possible role for general acid/
base catalysis by certain amino acid side chains. How-
ever, catalytic motifs are not obvious in the reactive
peptides studied (AIII and Tyr3). Alternatively, in the
case of poorly hydrophilic BDMP-QM, local concentra-
tion effects from hydrophobic binding of amine with
BDMP-QM may influence observed reactivity. The
hydrophobic aliphatic amino acid, Ile, was found to be
poorly reactive toward BDMP-QM (vide supra); thus, the
very high reactivity of AIII, Tyr, and Tyr3 may suggest
special hydrophobic binding with phenolic side chains via
a favored arene-arene interaction.21
Con clu sion s
Previous studies have shown wide-ranging reactivity
of QM electrophiles with biological nucleophiles under
varying reaction conditions. This study demonstrates,
for the first time, the quantitative nucleophilic reactivity
and selectivity of biological nitrogen nucleophiles toward
adduct formation with a QM, in neutral aqueous solution.
A full analysis of BDMP-QM hydration in the range 3
< pH < 11 provides an essential background to this
study. The observed selectivity for nitrogen nucleophiles
is sufficiently high for adduct formation to compete with
hydration, even at low concentrations of nucleophile.
Furthermore, the observed reactivity of proteins (HbA,
Mb), peptides (AIII, Tyr3), and certain amino acids (e.g.,
Tyr) is anomalously high, suggesting that specific binding
and/or catalytic motifs may be important in directing QM
reactivity. Such specific motifs in peptides and proteins
would be significant in dictating the cellular targets for
the toxicological action of QMs.
BDMP-QM, a representative 2,4,6-trialkylphenol oxi-
dative metabolite, forms 1:1 covalent adducts with amino
acids, peptides, and proteins via reaction with nitrogen
nucleophiles. The reactivity and selectivity for these
nucleophiles is very high relative to background hydra-
tion and consistent with protein alkylation as an impor-
tant mode of toxicological action of these phenols and
their QM metabolites.
The addition of up to three units of BDMP-QM to Mb
has been previously demonstrated using ESMS on un-
buffered solutions.10 Further experiments were con-
ducted at equimolar concentrations of BDMP-QM for
both Mb and HbA at pH 7.4 to compare the reactivities
and selectivities for adduct formation over hydration. C18
HPLC analysis of product mixtures indicated that 1 was
formed in the presence of Mb. Conversely, 1 was not
observed in the reaction of BDMP-QM with HbA. This
qualitative result demonstrates that Hb is more reactive
toward BDMP-QM than Mb. ESMS analysis of HbA:
Ack n ow led gm en t. The authors thank the NIH
(Grant No. ES06216) and Queen’s University for finan-
cial support of this research and Dr. Gordon Adamson
of Hemosol, Inc., for the generous gift of HemAzero
human HbA.
(19) Bai, Y.; Ueono, H.; Manning, J . M. J . Protein Chem. 1989, 8,
299.
J O962088Y
(20) Mori, N.; Bai, Y.; Ueno, H.; Manning, J . M. Carbohydr. Res.
1989, 189, 49.
(21) Williams, V. E.; Lemieux, R. P.; Thatcher, G. R. J . J . Org. Chem.
1996, 61, 1927 and references therein.
(22) Elbaum, D.; Wiedenmann, B.; Nagel, R. L. J . Biol. Chem. 1982,
257, 8454.
(23) Hollfelder, F.; Kirby, A. J .; Tawfik, D. S. Nature 1996, 383, 61.