Alkylation of amino acids and glutathione in water by o-Quinone methide. Reactivity and selectivity

Article abstract of DOI:10.1021/jo0006627

o-Quinone methide (1) has been produced in water both thermally and photochemically from (2-hydroxybenzyl)trimethylammonium iodide (2). Michael addition reactions of 1 to various amines, and sulfides, including amino acids and glutathione have been carried out, obtaining alkylated adducts (3-16) in fairly good to quantitative yields. The reaction rate and selectivity of 1 toward nitrogen and sulfur nucleophiles, in competition with the hydration reaction, have been investigated at different pH by laser flash photolysis technique. The observed reactivity spans 7 orders of magnitude on passing from water (k_Nu = 5.8 M^-1 s^-1) to the most reactive nucleophile (2.8 × 10⁸ M^-1 s^-1, 2-mercaptoethanol under alkaline conditions). These are the first direct reaction rate measurements of nucleophilic addition to the parent o-quinone methide (1). Competition experiments provided strong kinetic support to the involvement of free 1 as an intermediate in both thermal and photochemical reactions. Furthermore, several alkylation adducts regenerate 1 either by heating (9, 10, 13, and 14) or by irradiation (9, 11-13, 16). Such a thermal and photochemical reversibility of the alkylation process opens a new perspective for the use and application of such adducts as o-QM molecular carriers.

Full text of DOI:10.1021/jo0006627

J . Org. Chem. 2001, 66, 41-52
41
Alk yla tion of Am in o Acid s a n d Glu ta th ion e in Wa ter by o-Qu in on e
Meth id e. Rea ctivity a n d Selectivity
Emilia Modica, Riccardo Zanaletti, Mauro Freccero,* and Mariella Mella
Dipartimento di Chimica Organica, Universit a` di Pavia, V.le Taramelli 10, 27100 Pavia, Italy
freccero@chifis.unipv.it
Received May 3, 2000
o-Quinone methide (1) has been produced in water both thermally and photochemically from (2-
hydroxybenzyl)trimethylammonium iodide (2). Michael addition reactions of 1 to various amines,
and sulfides, including amino acids and glutathione have been carried out, obtaining alkylated
adducts (3-16) in fairly good to quantitative yields. The reaction rate and selectivity of 1 toward
nitrogen and sulfur nucleophiles, in competition with the hydration reaction, have been investigated
at different pH by laser flash photolysis technique. The observed reactivity spans 7 orders of
-
1
-1
8
magnitude on passing from water (kNu ) 5.8 M
s
) to the most reactive nucleophile (2.8 × 10
-
1
-1
M
s , 2-mercaptoethanol under alkaline conditions). These are the first direct reaction rate
measurements of nucleophilic addition to the parent o-quinone methide (1). Competition experiments
provided strong kinetic support to the involvement of free 1 as an intermediate in both thermal
and photochemical reactions. Furthermore, several alkylation adducts regenerate 1 either by heating
(
9, 10, 13, and 14) or by irradiation (9, 11-13, 16). Such a thermal and photochemical reversibility
of the alkylation process opens a new perspective for the use and application of such adducts as
o-QM molecular carriers.
In tr od u ction
gested by Rokita11 and others,12 o-QM could cross-link
two biologically useful molecules, such as nucleic bases
or peptides and proteins. If the alkylation process is
achieved under mild (ideally, biological) reaction condi-
tions, it could be used in a number of biomolecular
applications.
Quinone methides (QMs) have been proposed as in-
termediates in a large number of chemical and biological
processes. The reactivity of QMs is mainly due to their
electrophilic nature, which can also be directly correlated
to their toxicological properties. QMs have been proposed
1
2
Numerous investigations on nucleic acid base alkyla-
as intermediates in the biosynthesis of lignin and
enzyme inhibition.1
,3-7
4-7
tions with o-QMs have appeared in the past decade,
Among hydrolase inhibitors,
mainly published by the Rokita’s group.1
1,13-15
In fact
5
QMs have been recently used as covalent â-lactamase,
phosphatase,4,6 and ribonuclease A7 inactivators. It has
these reactive intermediates have been used as DNA
alkylating agents and cross-linkers. Recently, the same
author used parent o-QM 1, generated by fluoride
induced desilylation of O-(tert-butyldimethylsilyl)-2-(bro-
momethyl)phenol at 37-50 °C in aqueous DMF or
been suggested that quinone methides play a key role in
the chemistry of several classes of antibiotic drugs and
antitumor compounds such as mitomycin C and anthra-
8
cyclines.9
,10
acetonitrile, to alkylate dC,14 dA, and dG.
15
15
o-QM (1) is the prototype of more complex o-quinone
More recently Turnbull investigated the possibility of
alkylation of the phosphodiester functional group by
p-QMs in the presence of a Brønsted acid.16 He concluded
that in the absence of an acid catalyst the potential
alkylation of DNA through phosphodiester groups is
unlikely.
methide-like structures. With simple modifications sug-
*
Corresponding author. Tel: +39 0382 507668. Fax: +39 0382
07323.
1) Peter, M. G. Angew. Chem., Int. Ed. Engl. 1989, 28, 555 and
references therein.
2) Lignins: Occurrence, Formation, Structure and Reactions; Sar-
kanen, K., V.; Ludving, C., Eds.; Wiley: New York, 1971.
3) McDonald, I. A.; Nyce, P. L.; J ung, M. J .; Sabol, J . S. Tetrahedron
Lett. 1991, 32, 887.
4) Wang, Q.; Dechert, U.; J irik, F.; Withers, S. G. Biochem. Biophys.
Res. Commun. 1994, 200, 577.
5) Cabaret, D.; Adediran, S. A.; Garcia Gonzales, M. J .; Pratt, R.
F.; Wakselman, M. J . Org. Chem. 1999, 64, 713.
6) Myers, J . K.; Cohen, J . D.; Widlanski, T. S. J . Am. Chem. Soc.
995, 117, 11049.
7) Stowell, J . K.; Widlanski, T. S., Kutateladze, T. G.; Raines, R.
T. J . Org. Chem. 1995, 60, 6930.
8) (a) Han, I.; Russell, D. J .; Kohn, H. J . Org. Chem. 1992, 57, 1799.
b) Li, V.-S.; Kohn, H. J . Am. Chem. Soc. 1991, 113, 275. (c) Tomasz,
5
(
(
QM reactions with various nucleophiles, often under
(
1
7,18
nonaqueous conditions, have been studied.
More
(
(
(11) Zeng, Q.; Rokita, S. E. J . Org. Chem. 1996, 61, 9080.
(12) Nakatani, K.; Higashida, N.; Saito, I. Tetrahedron Lett. 1997,
38, 5005.
(13) Chatterejee, M.; Rokita, S. E. J . Am. Chem. Soc. 1994, 116,
1690.
(14) Rokita, S. E.; Yang, J .; Pande, P.; Shearer, J .; Greenberg, W.
A. J . Org. Chem. 1997, 62, 3010.
(
1
(
(
(
(15) Pande, P.; Shearer, J .; Yang, J .; Greenberg, W. A.; Rokita, S.
E. J . Am. Chem. Soc. 1999, 121, 6773.
M.; Das, A.; Tang, K. S.; Ford, M. G. J .; Minnock, A.; Musser, S.;
Waring, M. J . J . Am. Chem. Soc. 1998, 120, 11581.
(16) (a) Zhou, Q.; Turnbull, K. D. J . Org. Chem. 1999, 64, 2847. (b)
Zhou, Q.; Turnbull, K. D. J . Org. Chem. 2000, 65, 2022.
(17) Leary, G.; Miller, I. J .; Thomas, W.; Woolhouse, A. D. J . Chem.
Soc., Perkin Trans. 2 1977, 1737.
(18) Gardner, P. D.; Sarrafizadeh Rafsanjani, H.; Rand, L. J . Am.
Chem. Soc. 1959, 81, 3364.
(
9) (a) Egholm, M.; Koch, T. H. J . Am. Chem. Soc. 1989, 111, 8291.
b) Gaudiano, G.; Frigerio, M.; Bravo, P.; Koch T. H. J . Am. Chem.
Soc. 1990, 112, 6704.
10) (a) Angle, S. R.; Rainier, J . D.; Woytowicz C. J . Org. Chem. 1997,
2, 5884. (b) Angle, S. R.; Yang, W. J . Org. Chem. 1992, 57, 1092.
(
(
6
1
0.1021/jo0006627 CCC: $20.00 © 2001 American Chemical Society
Published on Web 12/14/2000

4
2
J . Org. Chem., Vol. 66, No. 1, 2001
Modica et al.
recently, a few investigations have focused on alkylation
of cellular nucleophiles in water, such as amino acids and
peptides, but only moderately reactive QMs such as 2,6-
di-tert-butyl-4-methylene-2,5-cyclohexadienone (BHT-
QM) and 2-tert-butyl-6-methyl-4-methylene-2,5-cyclo-
Sch em e 1
hexadienone (BDMP-QM) have been used in these
studies.1
9,20
The comparison of the results concerning the
alkylation of purine amines by BHT-QM, obtained by
Thompson’s group,21 to those obtained by Rokita’s group
(which used o-QM 1) seems to suggest that bulky p-QM
intermediates react with deoxynucleosides with lower
selectivity. If we assume the reactivity-selectivity prin-
ciple to be operative, the above finding is surprising,
because the prototype o-QM 1 is expected to be more
reactive than the alkyl-substituted analogues above.15
This aspect of the o-QM reactivity/selectivity is peculiar
and, from our point of view, it should be more thoroughly
investigated. Low reactivity of crowded QMs has mainly
been attributed to the shielding of the carbonyl oxygen
from solvent interactions (including hydrogen bonding)
tion in the presence of biological nucleophiles. Both
thermal and photochemical (steady state and laser)
activation processes of the salt were successful in water.
We presently report a study on the reactivity and selec-
tivity of 1, in Michael addition reactions to nitrogen,
oxygen, and sulfur nucleophiles of amino acids in com-
petition with hydration under acidic, neutral, and basic
conditions. The aim of this work is to measure and
rationalize the electrophilic nature of 1 toward simple
bifunctional and polyfunctional biological nucleophiles.
To the best of our knowledge neither adduct isolation
nor direct measurement of alkylation rates have previ-
ously been reported for the reaction of the parent o-QM
1 with free amino acids and peptides in water. The data
obtained from product distribution analysis, competition
alkylation experiments, and laser flash photolysis (LFP)
kinetic measurements will contribute to those investiga-
tions focused on QM-based enzyme inhibition and DNA
alkylation studies.
by the hydrophobic substituents at the 2- and 6-posi-
tions.2
0,21
Other intrinsic structural features, pointed out
2
2
by Soucek, such as the distortion of the quinonoid ring,
and electronic delocalization in benzoannelated QMs23
play an important role in lowering the reactivity of 2,6-
disubstituted and other QMs in comparison to unsubsti-
tuted QMs.
Although the research on enzyme inhibition and DNA
cross-linking by QMs seeks a rationalization of the
covalent labeling process, Wakselman notes that, “there
is only a limited amount of data in the literature on the
5
lifetime of quinone methides in aqueous solution” in the
presence of biological nucleophiles. Surprisingly, reported
data are related only to 2,6-disubstituted QMs, and only
Resu lts
Alk yla tion by Th er m a lly Gen er a ted o-QM (1). The
chosen precursor was (2-hydroxybenzyl)trimethylammo-
nium iodide (2) easily available from the phenol. When
this compound was heated at 38 °C and 80 °C in aqueous
solution, o-QM 1 was generated and trapped by N, O,
and S nucleophiles (see Scheme 1).
reasonable hypotheses have been made on the reactivity
of 1.15
Due to their synthetic utility and biological importance,
several methods for o-QM generation (i.e., high-temper-
ature dehydration of o-hydroxybenzyl alcohols and amino
2
4a-c
thermal extrusion of small molecules,24a,d
derivatives,
oxidation of phenols,24a,e and photochemical deprotonation
We initially investigated the alkylation of n-PrNH2,
of o-hydroxybenzyl alcohols25 or Mannich bases12) have
t-BuNH , aniline, piperidine, morpholine, and 2-mercap-
toethanol by quinone methide 1 obtaining addition
products (3-8) (Scheme 2) in good to quantitative yields
see Table 1). The study was then extended to the product
distribution (9-16, Scheme 2) arising from the alkyla-
tion, by the electrophile 1, of free amino acids such as
glycine (Gly), L-serine (Ser), L-cysteine (Cys), L-lysine
2
been published. Unfortunately, only a few of these
methods are suitable for studying reaction dynamics
under mild conditions using water as solvent. We thought
(
that an elimination reaction from a benzylammonium
salt26 could be the opportune strategy for o-QM genera-
(
Lys), L-tyrosine (Tyr), and glutathione (Glu) in water
(
19) McCracken, P. G.; Bolton, J . L.; Thatcher, G. R. J . J . Org. Chem.
997, 62, 1820.
20) (a) Bolton, J . L.; Turnipseed, S. B.; Thompson, J . A. Chem.-
1
solution from slightly acidic to basic conditions (5 <
pH < 12).
(
Biol. Interact. 1997, 107, 185. (b) Bolton, J . L.; Valerio, L. G.;
Thompson, J . A. Chem. Res. Toxicol. 1992, 5, 816. (c) Thompson, D.
C.; Perera, K.; Krol, E. S.; Bolton, J . L. Chem. Res. Toxicol. 1995, 8,
Glycine, serine, and lysine were selectively alkylated
at the N-atom(s) to give adducts 9, 10 and 12, 13. Lysine
was alkylated both at chain amino (Nꢀ) and the R-amino
groups (NR) to give adducts 12 and 13, respectively. No
adducts arising from an O-alkylation of the carboxy or
of the hydroxy (in the case of serine) groups were
3
2 3( 2. 1) Lewis, M. A.; Graff Yoerg, D.; Bolton, J . L. Thompson, J . A.
Chem. Res. Toxicol. 1996, 9, 1368.
22) Velek, J .; Koutek, B.; Musil, L.; Vasickova, S.; Soucek, M.
Collect. Czech. Chem. Commun. 1981, 46, 873.
23) (a) Musil, L.; Koutek, B.; Pisova, M.; Soucek, M. Collect. Czech.
(
(
Chem. Commun. 1981, 46, 1148-1159. (b) Koutek, B.; Pisova, M.;
Krupicka, J .; Lycka, A.; Snobl, D.; Soucek, M. Collect. Czech. Chem.
Commun. 1982, 47, 1645.
1
detected, by H NMR or by HPLC analysis, in the
reaction mixture. Under neutral or slightly acidic condi-
tions o-hydroxybenzyl alcohol (17), resulting from the
hydration of 1, was always present as a byproduct.
(
24) (a) Gr u¨ nanger, P. Methoden Org. Chem. 1979, VII/ 3b, 395. (b)
Dorrestijn, E.; Kranenburg, M.; Ciriano, M. V.; Mulder, P. J . Org.
Chem. 1999, 64, 3012. (c) Qiao, G. G.-H.; Lenghaus, K.; Solomon, D.
H.; Reisinger, A.; Bytheway, I.; Wentrup, C. J . Org. Chem. 1998, 63,
806. (d) Yato, M.; Ohwada, T.; Shudo, K. J . Am. Chem. Soc. 1990,
12, 5341. (e) Bolon, D. A. J . Org. Chem. 1970, 35, 3666.
3 2 3
Increasing the pH of the solution (by NaHCO /Na CO
9
1
(
25) (a) Wan, P.; Barker, B.; Diao, L.; Fisher, M.; Shi, Y.; Yang, C.
(26) (a) Gutsche, C. D.; Chun Nam, K. J . Am. Chem. Soc. 1988, 110,
6153. (b) Blad e´ -Font, A.; de Mas Rocabayera, T. J . Chem. Soc., Perkin
Trans. 1 1982, 841. (c) Breuer, E.; Melumad, D.; Tetrahedron Lett.
1969, 23, 1875.
Can. J . Chem. 1996, 74, 465. (b) Diao, L.; Cheng, Y.; Wan, P. J . Am.
Chem. Soc. 1995, 117, 5369. (c) Brousmiche, D.; Wan, P. Chem.
Commun. 1998, 491.

Alkylation of Amino Acids and Glutathione in Water
J . Org. Chem., Vol. 66, No. 1, 2001 43
Ta ble 1. P r od u cts fr om P r ep a r a tive Exp er im en ts, Gen er a tin g o-QM 1 Th er m a lly fr om 2, in th e P r esen ce of Am in es,
Th iols, Am in o Acid s, a n d Glu ta th ion e
conditions
a
(
pH, temp/°C,
alkylation adducts
(%, yield)b
hydroxybenzyl alcohol
b
nucleophiles
reaction time/h)
(17) (%, yield)
n-PrNH2
t-ButNH2
PhNH2
piperidine
morpholine
HO(CH2)2SH
glycine
pH 12, 80 °C, 1 h
pH 12, 80 °C, 1 h
CH3CN, 80 °C, 12 h
pH 12, 80 °C, 1 h
pH 12, 80 °C, 1 h
pH 7, 80 °C, 3.5 h
pH 7, 80 °C, 12 h
pH 10, 80 °C, 2 h
pH 6, 80 °C, 15 h
pH 7, 80 °C, 15 h
pH 10, 80 °C, 2 h
pH 7, 80 °C, 2 h
3 (65)
4 (87)
5 (81)
6 (84)
7 (>99)
8 (95)
-
-
-
-
-
-
70
-
36
21
4
9 (29)
9 (99)
serine
10 (60)
10 (74)
10 (84)
11 (98)
11 (10)
11 (99)
cysteine
-
-
-
pH 8, 38 °C, 36 h
pH g 10, 80 °C, 0.5 h
side-chain, R-amino
12 (-), 13 (-)
12 (10), 13 (30)
12 (17), 13 (32)
12 (33), 13 (40)
12 (35), 13 (39)
15 (-), 14 (-)
15 (1.6), 14 (57)
15 (1.9), 14 (60)
15 (5.5), 14 (58)
15 (12), 14 (64)
15 (18), 14 (53)
16 (84)
lysine
pH 5.0, 80 °C, 3 h
pH 6.0, 80 °C, 3 h
pH 7.0, 80 °C, 1 h
pH 7.6, 80 °C, 1 h
pH 9.6, 80 °C, 1 h
pH 7.0, 80 °C, 1 h
pH 8.5, 80 °C, 1 h
pH 9.0, 80 °C, 1 h
pH 9.5, 80 °C, 1 h
pH 10, 80 °C, 1 h
pH 12, 80 °C, 1 h
pH 7.0, 80 °C, 0.5 h
pH 7.8, 38 °C, 24 h
-
39
23
12
-
32
23
17
14
10
8
tyrosine
glutathione
-
-
16 (20)
a
pH was maintained constant by use of the following buffers: CH3COOH/CH3COO- (pH 5.0), KH2PO4/Na2HPO4 (pH 6.0-7.8), H3BO3/
NaOH (pH 8.0-9.0), NaHCO3/Na2CO3 (pH 9.5-10.4). pH 12.0 was adjusted by addition of NaOH 0.1 M solution. pH was measured in
reference solutions, which were prepared to be identical to the corresponding solution from preparative irradiation. Yield has been
b
determined by HPLC analysis of the reaction mixture, after 10-fold dilution, comparing peak areas to standard solution (at known
concentration) prepared from purified adducts.
Sch em e 2
The chemoselectivity in the alkylation of lysine and
tyrosine was also a function of the solution pH. Specifi-
cally, R-amino/side chain product ratios (13/12 and 14/
1
5) rose gradually from basic to neutral conditions (see
Table 1). L-Tyrosine was not reactive under neutral
conditions and after 1 h at 80 °C 17 was the only product
detected in the reaction mixture. Nevertheless, two
adducts, arising from a N-alkylation (14) and from an
O-alkylation (15), were isolated and identified at higher
pH (8.5 by borate buffer) at 80 °C. Similar to lysine, the
chemoselectivity of the tyrosine alkylation, evaluated by
the 14/15 ratio, is strongly affected by pH (see data in
Table 1).
Cysteine and a cysteine-containing peptide such as
glutathione were both alkylated, with complete selectiv-
ity, at the S atom, by 1 (pH 7 by KH PO /Na HPO buffer,
2 4 2 4
at 80 °C) and gave adducts 11 and 16 in quantitative
yields. Under mildly basic conditions (38 °C and pH 7.8),
only a small amount of the substrate was alkylated, but
a higher pH (pH g 10) allowed a quantitative transfor-
mation. No adducts arising from an N-alkylation of these
substrates or from a competitive hydration were in any
1
case detected in the crude reaction mixture by H NMR
or by HPLC analysis. Adduct 16 is the first example, to
the best of our knowledge, of a fully characterized o-QM-
glutathione type adduct.
The structures of the adducts 9-16 were firmly
buffer) led to improved yields of alkylation products
established on the basis of spectroscopic evidences, in
particular on 13C NMR chemical shifts, and H- C short
and long-range correlation experiments (HSQC and
HMBC, see Experimental Section).
1
13
(Table 1). Thermal decomposition of 2 was inefficient at
pH lower than 5, and under such conditions neither
alkylation adducts arising from 1 nor 17 were detected.

4
4
J . Org. Chem., Vol. 66, No. 1, 2001
Modica et al.
Ta ble 2. P r od u cts fr om P r ep a r a tive Exp er im en ts,
precursor (20%) was obtained. The 12/13 ratio was not
constant with the irradiation time. Specifically, it dropped
from 0.38 to 0.25 increasing the irradiation time from 1
min up to 2 h. With cysteine (after 1 h of irradiation of 2
at pH 7), adduct 11 and 17 were present in almost
equimolar amount, together with unreacted starting
material 2. In preparative irradiation the ratio 11/17
changed with time. In the initial stage this ratio was very
high (in fact, after a few minutes no hydroxybenzyl
alcohol could be detected), and the ratio increased to
approximately 1 after 1 h.
Gen er a tin g o-QM 1 P h otoch em ica lly (a t 254 n m ) fr om 2,
in th e P r esen ce of F r ee Am in o Acid s a n d Glu ta th ion e in
Wa ter
conditions
pH, λ ) 254 nm,
nucleophiles reaction time/h)
alkylation
adducts
(%, yield)
hydroxybenzyl
a
(
alcohol (17)
b
(%, yield)b
glycine
serine
pH 2, hν, 2 h
pH 7, hν, 2 h
pH 2, hν, 2 h
pH 7, hν, 2 h
pH 10, hν, 2 h
pH 2, hν, 1 h
9 (-)
9 (25)
10 (-)
10 (9)
10 (84)
11 (20)
6
75
45
77
-
55
-
40
-
cysteine
The alkylation of glutathione to give 16 proceeded
efficiently under photochemical conditions (86% yield).
A preparative photochemical alkylation of cysteine or
glutathione under basic conditions was not feasible
because such substrates in the anionic form absorbed at
254 nm precluding excitation of 2a .
pH 7, hν, 0.16 h 11 (10)
pH 7, hν, 1 h
pH 12, hν, 2 h
pH 7, hν, 2 h
11 (45)
12 (16), 13 (64)
16 (86)
lysine
glutathione
12
a
Constant pH was maintained by use of the following buffers:
KH2PO4/Na2HPO4 (pH 6.0-7.8), Na2CO3/NaHCO3 (pH 9.5-10.4).
pH 2.0 and 12.0 were adjusted by addition of HCl and NaOH 0.1
M solutions, respectively. The pH was measured in reference
solutions, which were prepared to be identical to the corresponding
P h otoch em ica l a n d Th er m a l Rea ctivity of th e
Alk yla tion Ad d u cts. As mentioned in the previous
paragraph, it was noted that alkylation adducts/17 and
12/13 ratios (in preparative photochemical experiments)
depended on the irradiation time. The above ratios were
also different from those obtained in thermal experi-
ments. This evidence suggested that a photoinduced
elimination process could have taken place from adducts
b
solution from preparative irradiation. Yield has been determined
by HPLC analysis of the reaction mixture, after 10-fold dilution,
comparing peak areas to standard solution (at known concentra-
tion) prepared from purified adducts.
The decomposition of the ammonium salt 2 was also
explored in the absence of added nucleophiles. Under
slightly basic conditions (pH e 10) the only product
detected was 17. At pH > 10 the thermal decomposition
of 2 produced a mixture of insoluble oligomers which
could not be fully characterized. The spectroscopic prop-
9
-13 and 16 with reformation of o-QM 1. In fact, the
irradiation at 254 nm of 9, 11-13, and 16 in diluted
aqueous solution (concentration 10 M and pH e 7) at
-3
low conversion afforded 17 and the corresponding free
amino acid (see Table 3). Surprising and noteworthy is
the high photoreactivity of adducts 12 and 13, which both
efficiently decompose under acid conditions (3 < pH <
1
erties (IR and H NMR, see Experimental Section) of this
mixture are consistent with its formation from the
anionic polymerization of 1 initiated by a Michael addi-
tion on 2a through the intermediate 18 (according to
Scheme 1). No detectable amounts of dimers (or trimers)
arising from either a Diels-Alder dimerization of 127 or
a Michael addition between 1 and the phenolic OH group
of the precursor 228 were detected at pH lower than 10.
Alk yla tion by P h otoch em ica lly Gen er a ted o-QM.
We found that o-QM 1 could also be generated photo-
chemically from (2-hydroxybenzyl)trimethylammonium
iodide (2) or from the corresponding dipolar ion 2a under
basic conditions by irradiation at λ ) 254 nm (see Scheme
7
) to give 17 (see data in Table 3) with high and very
similar quantum yields (Φ ) 0.78 and Φ ) 0.73, respec-
tively). Under basic condition (pH 12) 12 and 13 were
still photoreactive, though with lower quantum yields.
The photogenerated 1 could be trapped by adding cys-
-
3
teine at low concentration (10 M). In particular, the
quantum yield of o-QM generation from 12 (Φ ) 0.38)
was three times higher than that one from 13 (Φ ) 0.13).
The higher photofragmentation efficiency of 12 in basic
solution, rationalizes the chemoselectivity (13:12 ) 4:1)
observed in the preparative photochemical alkylation of
lysine.
1
). Irradiation of 2 in the presence of glycine and under
acidic conditions (pH 2 by 0.01 M HCl) afforded 17 in a
low yield (6%) leaving 2 largely (86%) unreacted. No
alkylation adduct was detected. When the pH of the
solution was increased to 7, the precursor was quanti-
tatively transformed, and the alkylated adduct 9 ac-
counted for 25% of the product mixture (Table 2).
Likewise, irradiation of 2 at pH 2 in the presence of
serine afforded 17 (45%), unreacted starting material
As far as thermal stability is concerned, all the adducts
alkylated at the R-amino acid groups, i.e., 9, 10, 13, and
4, decomposed at 80 °C under neutral conditions within
a few hours, affording 17 (see Table 3) and the corre-
sponding free amino acid. The other adducts (8, 11, 12,
and 15) were stable in water at pH e 7, and 80 °C for at
least 2 h.
1
Absolu te Ra te Con sta n ts for th e Alk yla tion of
o-QM As Mea su r ed by LF P . Laser flash photolysis
LFP) was used to investigate the formation, as inter-
mediate, of o-QM 1 and to measure its alkylation rate
constants with S (thiol) and N (amine) nucleophiles,
including amino acids and glutathione. A Nd:YAG laser
(42%) and no alkylation adduct. The preparative photo-
chemical alkylation of serine to give 10 became progres-
sively more efficient when we passed from neutral to
basic conditions (see data in Table 2). At pH 10, adduct
(
1
0 was the only product detected and isolated (84% yield).
In the case of lysine at pH 12, a mixture of adducts 12
and 13 (16% and 64% yield, respectively) and unreacted
(
power 2-8 mJ , pulse duration 10 ns, λexc ) 266 nm) was
used for the measurements, and the transient absorption
was recorded between 300 and 550 nm. As a diluted
solution of 2 in water was flashed, a transient with
absorption maximum at 400 nm was detected (Figure 1).
This absorption was attributed to 1 on the basis of the
following evidences: (a) the absorption maximum was
very close to that produced from hydroxybenzyl alcohol
(
27) (a) Letulle, M.; Guenot, P.; Ripoll, J .-L. Tetrahedron Lett. 1991,
3
2, 2013. (b) Gardner, P. D.; Sarrafizadeh Rafsanjani, H.; Brandon,
R. L. J . Am. Chem. Soc. 1959, 81, 5515. (c) Cavitt, S. B.; Sarrafizadeh
Rafsanjani, H.; Gardner, P. D. J . Org. Chem. 1962, 27, 1211.
(
28) (a) Chiba, K.; Hirano, T.; Kitano, Y.; Tada, M. Chem. Commun.
1
6
999, 691. (b) Inoue, T.; Inoue, S.; Sato, K. Bull. Chem. Soc. J pn. 1990,
3, 1062.

Alkylation of Amino Acids and Glutathione in Water
J . Org. Chem., Vol. 66, No. 1, 2001 45
Ta ble 3. Th er m a l a n d P h otoch em ica l (a t 254 n m ) Rea ctivity of th e Alk yla ted Ad d u cts
conditions
pH, thermal or
photochemical activation,
reaction time/h)
unreacted
a
(
alkylation
adduct
b
alkylated
adducts
17 or 11
(%, yield)
(%, yield)c
c
9
pH 3, 80 °C, 1 h
pH 7, 80 °C, 1 h
pH 12, 80 °C, 1 h
pH 7, 80 °C, 3 h
pH 3, 80 °C, 1 h
pH 7, 80 °C, 1 h
pH 12, 80 °C, 1 h, + cysteine
pH 3, 80 °C, 1 h
pH 7, 80 °C, 1 h
pH 12, 80 °C, 1 h, + cysteine
pH 7, 80 °C, 3 h
pH 3, hν, 0.08 h
pH 7, hν, 0.16 h
100
96
97
70
100
100
95
100
80
83
65
75
86
90
30
80
-
-
17 (3)
17 (2)
17 (28)
-
1
1
0
2
-
d
d
11 (5)
-
1
1
3
17 (20)
11 (17)
17 (32)
17 (25)
17 (14)
11 (10)
17 (68)
17 (16)
17 (99)
17 (99)
11 (10)
17 (99)
17 (99)
11 (15)
17 (19)
4
9
d
pH 12, hν, 0.16 h, + cysteine
pH 3, hν, 0.16 h
1
1
1
2
pH 7, hν, 0.16 h
pH 3, hν, 0.16 h
pH 7, hν, 0.16 h
-
-
-
-
-
78
d
d
pH 12, hν, 0.16 h, + cysteine
pH 3, hν, 0.16 h
1
1
3
6
pH 7, hν, 0.16 h
pH 12, hν, 0.16 h, + cysteine
pH 7, hν, 0.25 h
a
Constant pH was maintained by use of the following buffer: KH2PO4/Na2HPO4 (pH 6.0-7.8). pH 3.0 and 12.0 were adjusted by
addition of HCl (0.01 M) and NaOH (0.1 M) solutions, respectively. The pH was measured in reference solutions, which were prepared
b
-3
to be identical to the corresponding solution from preparative irradiation. Concentration of the alkylation adducts 9, 11-13 were 10
M; 10-fold smaller than in preparative photochemical experiments (in Table 2). c Yield has been determined by HPLC analysis of the
reaction mixture, without further dilution, comparing peak areas to standard solution (at known concentration) prepared from purified
adducts. Cysteine was added as o-QM trap, to avoid polymerization of 1, which occurs at pH 12 in absence of efficient nucleophiles.
d
Photogeneration of 1 from 2 at various pHs showed
that its formation was more efficient at pH 7, but it
remains high also under alkaline conditions. After 50
laser shots, 17 was detected in the solution as the only
product. In the presence of the nucleophiles, the corre-
sponding alkylation adducts were identified by compari-
son with authentic samples obtained from thermal
preparative experiments, as described above.
o-QM 1 was highly reactive (Table 4), particularly with
sulfur and nitrogen nucleophiles. On the other hand
hydration of o-QM was a much slower process (by at least
4
5
1
0 -10 times) under neutral conditions, but the rate of
this reaction increased under both acid and alkaline
conditions.
The direct measurement of the second-order rate
constants of the alkylation and hydration reactions by
LFP allowed us a quantitative evaluation and comparison
of the reactivity and selectivity of 1 toward free amino
acids and the solvent at different pH. Plots of the pseudo-
first-order rate constant vs nucleophile concentration
yielded linear correlations. The second-order rate con-
stants (kNu, see Table 4) for the addition reaction were
obtained from the slopes of these plots.
F igu r e 1. Transient absorption spectrum of o-QM 1, following
2
7
66 nm excitation of an aqueous solution of 2 (0.4 mM) at pH
, recorded 0.2 (b), 0.5 (O), 1.0 (9), and 2 ms ([) after the
laser pulse.
o-QM lifetime was not reduced by addition of acetate,
chloride, and perchlorate ions in 0.1 M concentration. In
particular o-QM became slightly more stable increasing
the ionic strength to 0.1 M by NaCl [τ ) 3.1 ((0.05) ms].
Furthermore, the determination of the 14/15 product
ratio, at pH 12, afforded the opportunity to calculate the
second-order rate constant for the alkylation of the
both by high-temperature dehydration in the gas phase27a
and by photolysis in acetonitrile solution,25 (b) it showed
a 40 nm blue shift with respect to that reported by Wan
for a phenyl-substituted o-QM,2 and (c) its lifetime was
reduced by the addition of N and S nucleophiles. o-QM 1
was a long-lived species in neat acetonitrile (τ > 1 s, not
measurable with a fast technique like LFP) but in water
5a
hydroxy group of tyrosine in the anionic form: kNu
)
5
-1 -1
(
(
pH 7) it decayed with a medium-long lifetime [τ ) 2.30
(0.05) ms], which was unaffected by the presence of
2.3 × 10 M
s
. This value was obtained by assuming
that the absolute rate constant for the alkylation of the
tyrosine amino group was identical to that of the glycine
oxygen, in accordance with what has been reported by
Wan.2
5b
2
NH .

4
6
J . Org. Chem., Vol. 66, No. 1, 2001
Modica et al.
Ta ble 4. Secon d Or d er Ra te Con sta n ts (kNu ), Mea su r ed in Wa ter by LF P , for th e Alk yla tion Rea ction s by o-QM 1
o-QM kNu
BDMP-QM
relative (to water)
o-QM reactivity scale
relative (to water)
BDMP-QM reactivity scale
alkylated substratesa
pKa
(M
-1
s
-1
)
kNu (M
-1 -1
s
)b
c,d
2.7 × 10-4
H2O (I ) 0.1 M)
H2O (I ) 0.0)
-
5.8
7.8
1
1
c,e
-
15.74f
-
4 g
6 h
5
4
5
OH
6.3 × 10
1.4 × 10
5.5 × 10
1.1 × 10
1.3 × 10
2.3 × 10
7.1 × 10
6.9 × 10
5.9 × 10
50
200
1.1 × 10
1.8 × 10
+
5
5
H3O
2.4 × 10
7.4 × 10
10.68i
-
-
-
-
-
-
9.5 × 10
4
-
-
-
-
-
-
n-PrNH2 (pH 12.0)
t-BuNH2 (pH 12.0)
piperidine (pH 12.0)
morpholine (pH 12.0)
Et3N (pH 12.0)
10.34j
5
1.9 × 10
4
k
6
5
11.24
8.78
10.9
9.68
10.8
9
9.17
-
2.2 × 10
k
6
5
4.0 × 10
l
5
5
1.2 × 10
m
5
5
glycine (pH 12.0)
lysine (pH 12.1)
1.2 × 10
n
5
5
4
18.6 ( 2.8
1.0 × 10
6.9 × 10
n
o
.16
5
4
HO(CH2)2SH (pH 6.9)
1.9 × 10
2.8 × 10
2.3 × 10
6.9 × 10
1.3 × 10
1.3 × 10
9.5 × 10
-
-
-
3.3 × 10
-
-
-
-
8
7
HO(CH2)2S (pH 12.1)
4.8 × 10
-
p
p
5 q
5
4
Tyr-O (pH 12.0)
9.99
8.94
8.2r
-
3.6 × 10
5
5
Tyr-NH2 (pH 12.0)
cysteine (pH 6.8)
cysteine (pH 12.2)
glutathione (pH 7.1)
45.0 ( 3
1.2 × 10
1.7 × 10
5
4
7
3320 ( 66
2.2 × 10
1.2 × 10
8
7
-
-
1.9 × 10
-
-
8.72r
5
1.6 × 10
5
a
In brackets is reported the pH of the solutions stabilized by KH2PO4/Na2HPO4 buffer (pH 6.8-7.1). pH 12.0 was adjusted by addition
of NaOH 0.1 M. From ref 19. Second-order rate constant for the hydration process (kH O) at pH 7 has been calculated from the measured
b
c
2
d
pseudo-first-order rate constant and the H2O concentration in pure water. kH2O in a solution with constant (I ) 0.1 M) ionic strength by
NaCl. Identical value has been obtained using NaClO4. kH2O in a solution of pure water (I ) 0.0). Reference 52. g Obtained monitoring
e
f
-
the lifetime of 1 as a function of HO concentration, adding NaOH, and keeping the ionic strength constant (I ) 0.1 M) with NaCl.
h
+
Obtained monitoring the lifetime of 1 as a function of H3O concentration, adding HCl, and keeping the ionic strength constant (I ) 0.1
M) with NaCl. Reference 53. Reference 54. Reference 55. Reference 56. Reference 57. n Reference 34. o Reference 58. Reference
i
j
k
l
m
p
q
3
5, 36. Calculated kNu value, from 14/15 product ratio, assuming the alkylation rate to the N atom identical to the alkylation rate of
r
glycine. Reference 59.
Ta ble 5. Com p etition Alk yla tion Exp er im en ts In volvin g P r NH2/t-Bu NH2, Mor p h olin e/P ip er id in e, a n d t-Bu NH2/OH- a s
Cou p les of Nu cleop h iles
second-order
adducts ratio
competing
nucleophiles
couple of
adducts
rate constant ratios
hνb
∆c
(kNu′/kNu′′)
a
n-PrNH2
t-BuNH2
morpholine
piperidine
3/4
4.6 ((0.3)
4.9 ((0.1)
5.0
1.7
1.8
6/7
1.5 ((0.2)
1.6 ((0.2)
1.6 ((0.1)
1.7 ((0.1)
d
t-BuNH2 (pH 12)
4/17
-
d
OH (pH 12)
a
From absolute second-order rate constants (kNu) in Table 4. b Photochemical competition experiments were performed at 254 nm, at
low conversion, in the presence of a 10-fold excess of nucleophiles, which were in equimolecular ratio (for more detail, see Experimental
Section, competition experiments). The values were obtained as an average of five different experiments. c Thermal competition experiments
at low conversion (see Experimental Section, competition experiments). The values were obtained as an average of three different
d
experiments. pH was adjusted by NaOH solution (0.1 M).
In the absence of nucleophiles the decay of o-QM 1
consistently followed a pseudo-first-order kinetic. This
kinetic behavior provided convincing evidence that the
dimerization of 1, through a [4 + 2] Diels-Alder reaction,
was not significant under these conditions. At pH > 10,
the decay rate of 1 became a linear function of the
precursor (2a ) concentration. This suggested that under
alkaline conditions the dimerization process was occur-
ring via Michael addition of the precursor 2a on o-QM
alkylation experiments involving a few couples of nu-
cleophiles, generating 1 with both a thermal and photo-
chemical activation process. We used n-PrNH
2
/t-BuNH
2
,
-
morpholine/piperidine, and t-BuNH
2
/OH , at pH 12, in
equimolecular ratio, as couples of nucleophiles. We
generated 1 both under thermal and photochemical
conditions (at low conversion <10%). The product ratios
of the thermal and photochemical alkylations were very
similar to each other. Moreover, their values were almost
identical to the ratios calculated from the measured (by
LFP) absolute rate constants. The results are sum-
marized in Table 5.
(1) (see Scheme 1). This kinetic evidence fits with the
formation of a mixture of oligomers in preparative
experiments under the same conditions.
When a diluted solution of adducts 9-13 and 16 was
flashed at 266 nm, o-QM (1) was generated as transient
with an identical spectrum as that in Figure 1. This
unequivocally supports that the alkylation adducts 9-13
and 16 are also photoprecursors of 1.
Com p etition Alk yla tion Exp er im en ts. Although
quinone methides have been proposed as the intermedi-
ates involved in thermally induced decomposition reac-
tions of phenolic Mannich base methiodides25 no direct
experimental proof has been produced so far. To produce
such an evidence, we carried out a set of competition
Discu ssion
Other studies have used the approach of benzylam-
monium thermal elimination26 and the photochemically
2
5
induced deprotonation of o-hydroxybenzyl alcohols and
(2-hydroxybenzyl)dimethylamine12 to achieve o-QM gen-
eration. Nevertheless, we feel that some aspects of our
investigation are novel and they could be useful for
further applications involving o-QM 1 as biological alky-
lating agent. They will be discussed in more detail in the
following paragraphs.

Alkylation of Amino Acids and Glutathione in Water
J . Org. Chem., Vol. 66, No. 1, 2001 47
(keeping the laser power and the absorbance of both
solutions constant) and then multiplying it by the
quantum yield determined by Wan.2
5b
Comparison of o-QM photochemical generation from
1
2
2
and (2-hydroxybenzyl)dimethylamine (used by Saito)
demonstrates that both precursors have identical ef-
ficiency at pH 7, each showing quantum yield Φ ∼ 1.
Nevertheless, our o-QM photochemical generation from
2
remains advantageous for two reasons: (i) lower
nucleophilicity of the precursor and (ii) higher quantum
yield under alkaline conditions. The scarce nucleophilic
character of 2 prevents its alkylation by 1, while the
opposite occurs with Saito’s amine.30 Both aspects make
salt 2 an ideal o-QM photoprecursor for LFP kinetic
measurements of alkylation process of several nucleo-
philes even under basic conditions.
The fact that product ratio in both thermal and
photochemical alkylations was almost identical to the
ratio calculated from the absolute rate constants mea-
sured by LFP (see Table 5) provides compelling kinetic
evidence that o-QM 1 is the actual intermediate in both
processes. Therefore, an alternative mechanism involving
direct displacement of trimethylamine from the salt 2,
or 2a , is unlikely.
F igu r e 2. Decay curves for o-QM 1 generated from 2 (upper
trace) and from hydroxybenzyl alcohol (17) (lower trace) at 400
nm, in water.
o-QM (1) Gen er a tion . Our investigation has revealed
that o-QM 1 can be generated not only in organic
solvents, but also in water by a thermally induced
elimination at 80 °C at pH > 5, using salt 2 as precursor.
Generation of 1 from 2 is easier under slightly basic
conditions as revealed from more efficient alkylation
reactions. In fact, we obtained higher alkylated product
yields in shorter reaction times at pH g 8. Thus, the
results of preparative experiments can be rationalized
by assuming that compound 2a is the actual precursor
of 1 under basic conditions (pH 12), where its concentra-
tion can be significant. Noteworthy, the activation of the
substrate (Scheme 1) can be thermally achieved under
almost physiological conditions (38 °C and pH 7.8) where
o-QM can be trapped by an efficient nucleophile such as
cysteine.
Rea ctivity of o-QM a s Alk yla tin g Agen t. The
absolute second-order rate constants (kNu) for alkylation
reactions by QMs present in the literature are sporadic
1
9
and have often been derived from product ratios. One
could also be tempted to apply such an approach to 1.
However, due to the variable thermal and photochemical
instability of some o-QM alkylation adducts, such an
approach could give erroneous results. On the other hand,
we have shown that the generation of o-QM 1 starting
from 2 by LPF afforded a way to measure its kNu with
several nucleophiles. The above technique allows a
quantitative analysis of the reactivity of 1 and, in
principle, of other highly reactive QMs.
Since proteins, peptides, purine, and pyrimidine nu-
cleobases all contain basic functionalities and the acidity
(2) ) 8.74],29 they could induce
of 2 is not negligible [pK
a
o-QM reactivity (see data in Table 4) spans seven
formation of 1 from 2 (through 2a ) by base catalysis.
Thus, it is reasonable to consider 2 a leading candidate
for alkylation of biological nucleophiles by o-QM 1
through a pH-activated process.
-
1
magnitude orders on passing from water (kNu ) 5.8 M
-
1
8
-1
s
-1,
s
) to the most reactive nucleophiles (2.8 × 10 M
8
-1 -1
2
-mercaptoethanol; 1.3 × 10 M
s
Cys; under strongly
alkaline conditions). Thiols are the only nucleophiles that
can be quantitatively alkylated also under neutral or
slightly acidic conditions (3 < pH < 7). In fact, under such
Passing to the photogeneration of 1 from 2, it should
be stressed that it was not unexpected. Indeed, Wan had
previously studied the photoreactivity of substituted
5
hydroxybenzyl alcohols,25 and even more recently Saito
conditions they still react with kNu higher than 1 × 10
-
1
-1
M
s
2
(see data in Table 4). Unprotonated NH groups
used (2-hydroxybenzyl)dimethylamine as a photoprecur-
are also highly reactive toward 1, showing kNu from
sor of 1, trapping it by electron rich alkenes in a hetero
5
6
-1 -1
1
2
1.1 × 10 to 2.3 × 10 M
s .
Diels-Alder reaction. Nevertheless, with LFP of 2, we
managed to achieve, for the first time, a full spectroscopic
characterization of 1 in water (see Figure 1). LFP of
hydroxybenzyl alcohol, as stated by Wan, “gave a long
Our results show that o-QM 1 is the most reactive
alkylating agent in water among QMs.
In fact, o-QM 1 appears to be much more reactive than
2
-tert-butyl-6-methyl-4-methylene-2,5-cyclohexadien-
lived species but the spectrum was much weaker and not
1
9
as well-defined”.2
5b
one (BDMP-QM), which has been defined as “highly
reactive” in previous investigation. The former reacts
2
We were able to record the full
2
0a
spectrum of 1, because its quantum yield from 2 was 4.3
4
3
.1 × 10 , 1.3 × 10 and 39-fold faster than the latter
times higher [Φ ) 0.98 ((0.02) at pH 7] than that from
-
the alcohol (Φ ) 0.23).25b The decay traces of 1 generated,
toward H
2
O, OH , and cysteine, respectively (see Table
4
for more details).
Even the highly electrophilic 4-[bis(trifluoromethyl)-
methylene]cyclohexa-2,5-dienone31 studied by Richard is
respectively, from hydroxybenzyl alcohol and from salt
2
(Figure 2), clearly indicated the latter as a more
efficient precursor of 1 than the alcohol. The above
quantum yield was calculated from the ratio of the
absorption of 1 (at time zero) obtained by flashing a
solution of 2 and a solution of hydroxybenzyl alcohol
(30) o-QM (1) lifetime is at least 10 times shorter if it is generated
starting from (2-hydroxybenzyl)dimethylamine. The lifetime is also
linearly dependent on the precursor concentration. This kinetic
evidence show that (2-hydroxybenzyl)dimethylamine competes for
o-QM (1) more efficiently than water, even at low concentration
-4
(29) Epstein, J .; Plapinger, R. E.; Michel, H. O.; Cable, J . R.;
(∼10 M).
(31) Richard, J . P.; Toteva, M. M.; Crugeiras, J . J . Am. Chem. Soc.
2000, 122, 1664.
Stephani, R. A.; Hester, R. J .; Billington, C., J r.; List, G. R. J . Am.
Chem. Soc. 1964, 86, 3075.

4
8
J . Org. Chem., Vol. 66, No. 1, 2001
Modica et al.
6
0- and 130-fold slower than o-QM 1 in the addition to
negligible. Second-order rate constants suggest that
unprotonated side-chain and R-amino groups compete for
o-QM at the same rate.
6
-1 -1 32
thiolates (propanethiolate, kNu ) 4.6 × 10 M
s ) and
primary amines (ethylamine, kNu ) 4.1 × 10 M s-1),
respectively. At present no other data on o-QM alkylation
rates are available in the current literature. Our results
in addition to those results obtained for BDMP-QM1
and 4-[bis(trifluoromethyl)methylene]cyclohexa-2,5-di-
enone31 represent, to our knowledge, the only kinetic data
present in the literature on QMs reactivity in water.
The fact that o-QM reactivity with water is the highest
of all known QMs may have an important consequence
in limiting the pharmacological side effects of o-QM-based
drugs. In fact, it is commonly believed that the shorter
QM half-lives the lower their hepatotoxic effects.1
Selectivity of o-QM in Alk yla tion Rea ction s. Al-
though o-QM 1 is the most reactive among QMs, it is also,
surprisingly, a quite selective alkylating agent. Compar-
ing the second-order rate constants measured by LFP
3
-1
32
Although a fairly good alkylation selectivity of lysine
by o-QM can be achieved under neutral condition (NR/
Nꢀ ) 3 at pH 6.0), its selectivity is at least 10 times
smaller than the BDMP-QM selectivity (NR/Nꢀ ) 30 at
pH 7.4), where no side chain alkylation adduct was
detected.20 Similarly to lysine, the product ratio 14/15
from the alkylation of tyrosine decreases with the in-
crease of the pH of the solution (see in more detail data
9,20
in Table 1). In fact, the phenolic and amino acidic pK
a
of
3
5,36
the tyrosine are respectively 9.99 (OH)
and 8.94
9,20b,c
+ 36
3
(NH ).
The observed competition between the anionic form of
5
-1 -1
tyrosine (kNu ) 2.3 × 10 M
s ) and the unprotonated
5
-1 -1
NH
2
group of the amino acid (kNu ) 6.9 × 10 M
s )
toward 1 provides the only example of O-alkylation by
QMs with amino acids. Moreover, it stands in striking
contrast with the fully selective alkylation at the nitrogen
(Table 4) we can compile a nucleophilicity scale toward
-
-
o-QM: RS . R
substituents) > HO . H
scale can be used also to rationalize the selectivity of 1
in alkylation reactions.
3
N > PhO > RSH ≈ RNH
2
(bulky R
-
-
-
33
O g Cl , AcO , ROH. Such a
atom of tyrosine by BDMP-QM reported by Bolton,
2
Thatcher, and Thompson.1
9,20
This reactivity trend is not
a peculiarity of tyrosine. In fact, o-QM 1 alkylates 2a in
Obviously, the pH of the solution has an important role
in controlling the selectivity of the alkylation, particularly
when the competing nucleophiles have different basicity,
as in the case of the alkylation of Nꢀ versus NR amino
groups of lysine and amino versus hydroxy groups of
tyrosine. The selectivity enhancement of o-QM, for the
R-amino versus the ꢀ-amino group of lysine, on passing
from basic to neutral conditions, can be rationalized by
the different degree of protonation of the former relative
the absence of both N and S nucleophiles, confirming that
the PhO nucleophilic system competes efficiently with
the solvent and N nucleophiles (under basic conditions)
for o-QM 1.
-
It is well-known that in an aqueous intracellular
environment the strongest nucleophile with quinones and
1
QMs is the cysteinyl SH group. Actually, UV-spectro-
scopic evidences for the reaction between cysteine and
various quinones,37 and the characterization of the cys-
to the latter (pK
a
(2) ) 9.16, pK
a
(3) ) 10.81).34 The above
teinyl-Dopa conjugates support such a high nucleophi-
38
explanation for such a selectivity of the alkylation
reaction of lysine by BDMP-QM (at pH 7.4) has also been
suggested by Thompson.20 A convincing support concern-
ing the role of selective protonation in the control of the
selectivity comes from the comparison of second-order
licity. However, to the best of our knowledge, adduct 16
is the first example of a fully characterized o-QM-
glutathione like adduct, as the addition of the thiol group,
in cysteine-containing peptides, to o-QM-like structures
has not been documented so far. Only recently few p-QM
glutathionyl-SH adducts have been isolated from in vitro
rate constants for the alkylation of n-PrNH
2
(kNu )
5
-1 -1
) and glycine (kNu ) 6.9 × 10 M s-1).
5
-1
39
5
.5 × 10 M
s
experiments. Thus, it should be appreciated that through
Such rate constants have been measured under strongly
the photochemical activation method it is possible to
achieve the alkylation of cysteine and glutathione even
under acidic conditions, even if the same results cannot
be obtained through a thermal reaction. The latter
observation is consistent with the finding by Wan that
the efficiency of the o-QM formation is controlled by the
alkaline conditions (pH 12) where protonation of both
nucleophiles (pK
) 10.6854 and 9.68, respectively) is
58
a
(
32) Richard, J . P.; Amyes, T. L.; Bei, L.; Stubblefield, V. J . Am.
Chem. Soc. 1990, 112, 9513.
33) Second-order rate constants for Cl , AcO , ROH alkylation
-
-
(
deprotonation of the singlet excited state of the phenol
O
reactions by 1 were too small, in comparison to kH2 , to be measured
in water solution.
5a
(pK
a
) 2.5),2 which is much more acidic than its own
(
(
34) Ellenbogen, E. J . Am. Chem. Soc. 1952, 74, 5198.
35) (a) Kim, K.; Cole, P. A. J . Am. Chem. Soc. 1998, 120, 6851. (b)
ground state.
Castro, E. A.; Pavez, P.; Santos, J . G. J . Org. Chem. 1999, 64, 2310.
36) Ishihama, Y.; Oda, Y.; Asakawa, N. J . Pharm. Sci. 1994, 83,
500.
(
(45) Baires, S. V.; Ivanov, V. B.; Ivanov, B. E.; Zyablikova, T. A.;
Efrimov, Y. Y. Bull. Acad. Sci. USSR (Engl. Transl.) 1987, 36, 1881.
(46) Maroni, P.; Cazaux, L.; Tisnes, P.; Zambeti, M. Bull. Soc. Chim.
Fr. 1980, 2, 179.
1
1
(37) Mason, H. S.; Peterson, E. V. Biochim. Biophys. Acta 1965, 111,
34.
(
(
38) Ito, S.; Prota, G. Experentia 1977, 33, 1118.
39) Thompson, D. C.; Perera, K.; Krol, E. S.; Bolton, J . L. Chem.
(47) Schepartz, A.; Breslow, R. J . Am. Chem. Soc. 1987, 109, 1814.
(48) Steevens, J . B.; Pandit, U. K. Tetrahedron 1983, 39, 1395.
(49) Hodgkin, J . H. Aust. J . Chem. 1984, 37, 2371.
(50) Ranganathan, S.; Singh, W. P. Tetrahedron Lett. 1988, 29, 1435.
(51) Bax, A.; Summers, M. F. J . Am. Chem. Soc. 1986, 108, 2093.
(52) Harned, R. Trans. Faraday Soc. 1940, 36, 973.
(53) Battye, P. J .; Ihsan, E. M.; Moodie, R. B. J . Chem. Soc., Perkin
Trans. 2 1980, 741.
(54) Hine, J .; Chou, Y. J . Org. Chem. 1981, 46, 649.
(55) Castro, E. A.; Cubillos, M.; Santos, J . G. J . Org. Chem. 1994,
59, 3572.
(56) Emly, M.; Leussing, D. L. J . Am. Chem. Soc. 1981, 103, 628.
(57) Li, N. C.; White, J . M.; Yoest, R. L. J . Am. Chem. Soc. 1956,
78, 5218.
Res. Toxicol. 1995, 8, 323.
40) Bordwell, F. G.; Cripe, T. A.; Hughes, D. L. Nucleophilicity,
(
Basicity, and the Brønsted Equation in Nucleophilicity; Harris, J . M.;
McManus, S. P., Eds.; American Chemical Society: Washington, DC,
987; Chapter 9, pp 137-153.
(
1
41) Pearson, R. G. J . Am. Chem. Soc. 1986, 108, 6109.
(
42) Et
3
N (∆GoNu ) -1.1 kcal mol ) and morpholine are 1.4 and 4
-1
o
times, respectively, more reactive than PrNH
2
(∆G Nu ) +2.4 kcal
-1
-
o
mol ), even if the latter is less solvated by water. TyrO (∆G Nu
∼
-1
-
4.6 kcal mol ; TyrO- solvation energy was assumed to be identical
-
-
o
to PhO solvation energy) is 3.6 times as reactive as HO (∆G Nu
)
1
41
-
4.4 kcal mol- ), and they have similar solvation.
(
(
43) In preparation.
44) Pochini, A.; Puglia, G.; Ungaro, R. Synthesis 1983, 11, 906. (b)
(58) Eldin, S.; J encks, W. P. J . Am. Chem. Soc. 1995, 117, 4851.
(59) Patel, H. M. S.; Williams, D. L. H. J . Chem. Soc., Perkin Trans.
2 1990, 37.
Stedman, E. J . Chem. Soc. 1927, 1904.

Alkylation of Amino Acids and Glutathione in Water
J . Org. Chem., Vol. 66, No. 1, 2001 49
anol, 1.9 × 10 M s-1 glutathione, kNu ) 9.5 × 10 M
5
-1
5
-1
Both product distribution analysis of lysine (and ty-
rosine) in alkylation reaction and LFP measurements
show, without doubt, that 1 is less selective and much
more reactive than BHP-QM and BDMP-QM, at least in
the alkylation of amino acids. As far as selectivity is
concerned, our results show an unexpected different
behavior of o-QM 1 toward amino acids and nucleosides.
In fact, as also mentioned in the Introduction, Rokita
stated that o-QM shows a “greater selectivity in nucleo-
-
1
-1 -1
s
), while kH O is 5.8 M
s .
2
Currently, we are investigating in more detail the role
of hydrogen-donor solvents in the activation of o-QM, and
in the solvation of the nucleophile.43
Alk yla ted Ad d u cts a s o-QM Molecu la r Ca r r ier s.
In our study we have been able to identify in the o-QM
alkylation adducts at the NH
2
amino acidic group (9, 10,
1
3, and 14) a new class of compounds which are capable
1
5
side modification.” Our result suggests that, in addition
to cysteine thiols, protein alkylation by o-QM 1 should
occur not only at the N-terminal residue but also at the
side-chain groups of lysine and tyrosine. This remarkable
difference between o-QM and BDMP-QM chemical reac-
tivity suggests that the former could find different
applications as an enzyme inhibitor in comparison to the
latter.
The rate of addition of 1 to water, amines, and sulfides
a
does not correlate with the pK of the conjugated acid of
the nucleophiles, as apparent from the data collected in
Table 4. Thus, a simple parallelism between basicity
of a thermal generation of o-QM from neutral to basic
conditions. Such a reactivity contrasts with the thermal
stability of all the other products obtained in this study
and suggests that the above compounds and similar
peptide like adducts (where the o-QM is linked to the
N-terminal amino acid) could be used as o-QM molecular
carriers. The peculiar thermal behavior of 9, 10, 13, and
-
1
4 is probably due to the presence, of a proximal -COO
to the phenolic OH group, which may catalyze (through
an intramolecular deprotonation) the decomposition to
o-QM. Work is still in progress in our laboratory to clarify
this aspect, and to decrease the activation temperature
(measured by the pK
a
values) and nucleophilicity (mea-
43
of the above presursors.
sured by the second-order rate constant of the alkylation
process) of the substrates is not possible. The lack of
The photoreactivity of all isolated alkylation adducts,
including those with cysteine and glutathione, shows that
the photoalkylation by o-QM 1 is a reversible process.
This general photoreactivity explains (i) the higher
amount, in comparison to thermal alkylation, of hydroxy-
benzyl alcohol formed as byproduct and (ii) the different
selectivity of lysine alkylation in steady-state preparative
irradiations. Adducts photoreactivity provides a wider
and milder approach, in comparison to the thermal one,
to efficiently generate alkylating agents.
a
linearity in the plot ln(kNu) vs pK is expected for different
4
0
classes of nucleophiles (S, N, and O nucleophiles), but
it is surprising within a more homogeneous nucleophile
family such as alkylamines. Solvation of the nucleophile
by the protic medium is not the main controlling factor
of nucleophile reactivity toward 1, because there is no
correlation between stabilization of nucleophile by hydra-
tion (as described by ∆GoNu ) and ln(kNu). Therefore, it
is likely that other factors related to the peculiar ortho
geometry of 1, such as steric effects and electrostatic
repulsion between the o-QM oxygen atom and the nu-
cleophile could also be important. In fact, o-QM 1 appears
to be slightly sensitive to the steric requirements of the
41
42
Actually, photoactivation of o-QM-amino acid and
o-QM-oligopeptides conjugates such as o-QM-glutathione
(16) is certainly one of the most interesting finding of
the present study. It suggests that any alkylation adduct
of biological nucleophiles is a potential candidate in
delivering, under irradiation, a highly reactive alkylating
agent such as o-QM 1.
nucleophile, as PrNH
t-BuNH , and twice less reactive than piperidine.
Alk yla tion vs Hyd r a tion . Th e Role of Acid a n d
Ba se Ca ta lysis. The second-order rate constants (kNu
2
is 5 times more reactive than
2
)
+
-
Con clu sion
3
with H O and OH demonstrate the importance of the
acid and base catalysis in the hydration process of o-QM.
The above kNu values have been measured without
buffers, controlling the pH of the solution by HCl or
NaOH addition. Therefore the acid- and base-catalyzed
hydration process has been studied without the interfer-
ence of buffer acids or buffer anions. The o-QM rate vs
pH dependence is qualitatively similar to other QMs
The investigation described above has provided a
useful quantitative characterization of the electrophilic
character of o-QM 1 under neutral, acid, and basic
aqueous conditions. Several interesting findings of our
study can be summarized as follows:
(i) We reported one the most efficient thermal and
photochemical approaches to o-QM 1 generation in water,
starting from the same precursor (2).
1
9
+
previously described, such as BDMP-QM.
with 1 22 times faster than OH . The reactivity of OH
H
3
O reacts
-
-
(
ii) Thermal generation from 2 is achievable under
almost” physiological conditions by pH activation to give
alkylation products in good yields.
iii) Salt 2 shows properties that make it a more
in water is always lower than that of unprotonated
amines (even in the case of the bulky t-BuNH ) and
2
“
thiols. Thus, it is not surprising that in alkaline aqueous
solution (pH 10) hydroxybenzyl alcohol has never been
(
-3
preferable o-QM photoprecursor than o-hydroxybenzyl
alcohols or (2-hydroxybenzyl)dimethylamines for kinetic
measurements.
detected, even at low nucleophile concentration (10 M).
The hydration of 1 is competitive under acidic conditions,
as shown qualitatively by thermal and photochemical
preparative experiments, for two different reasons. These
reasons are acid catalysis, which increases the hydration
rate, and protonation of the N nucleophiles. The only
substrates which can compete efficiently with water at
pH lower than 7 are cysteine and cysteine-containing
peptides. In fact, the SH group shows second-order rate
(iv) This study is the first quantitative investigation
on the electrophilic character of parent o-QM 1. To our
knowledge there are no studies either on the selectivity
or reactivity of a quinone methide with an ortho geometry
as alkylating agent of amino acids and peptides.
(v) Quantitative comparison (based on absolute second-
order rate constants) of o-QM and BDMP-QM reactivity
5
-1 -1
constants higher than 10 M
s
, under neutral condi-
5
-1 -1
19,20
tions (cysteine, kNu ) 1.3 × 10 M
s
; 2-mercaptoeth-
(provided mainly by Thompson, Bolton, and Thatcher)

5
0
J . Org. Chem., Vol. 66, No. 1, 2001
Modica et al.
shows the huge reactivity gap between 2,6-disubstituted-
p-QMs and o-QM 1.
trimethylammonium iodide (2) (50 mg, 0.17 mmol) in 5 mL of
NaHCO
NaHCO
3
/Na
2
CO
3
buffer adjusted at pH 10 (2.6 mL of a
CO 0.1 M) was heated at
3
0.1 M and 2.4 mL of a Na
2
3
(vi) o-QM should alkylate the side-chains of lysine and
8
(
0 °C for 2 h, cooled, and then purified by preparative HPLC
eluent, H O: CH CN ) 85:15 + 0.1% CF COOH, flux ) 4.0
mL/min, t ) 32.0 min). To exchange the trifluoroacetate anion
tyrosine containing peptides much more efficiently than
other less reactive QMs, which are capable of linking
mainly the N-terminal residues of peptides.
2
3
3
R
with chlorine anion the product was dissolved in 1 mL of a
0.1 M HCl solution. The solution was stirred, and the solvent
was evaporated at room temperature in a vacuum (0.1 mmHg).
(vii) Several alkylation adducts regenerate 1 by heating
or by irradiation. Thus, they represent a new and a broad
class of compounds which could be used as molecular
alkylating carriers which generate o-QM by pH-triggered
processes. Additional studies are being carried out in our
laboratory in the attempt to modify such precursors with
the aim of activating them in even milder conditions
3
6.7 mg (99% yield) of adduct 9 was obtained as a colorless
-
1
oil. IR (Nujol) (cm ) 3400-2400 (broad), 1681, 1557, 1206,
1
1
(
143. H NMR (D
m, 2 H, aromatics), 7.30-7.45 (m, 2 H, aromatics). C NMR
δ 47.55, 48.06, 116.38, 117.88, 121.38, 132.39, 132.69, 155.96,
70.82. Anal. Calcd for C Cl: C, 49.67; H, 5.56; N, 6.44;
2
O) δ 3.90 (s, 2 H), 4.35 (s, 2 H), 6.90-7.00
13
1
9 3
H12NO
(photochemically using a longer wavelength and ther-
Cl, 16.29. Found: C, 49.75; H, 5.59; N, 6.40; Cl, 16.35.
mally at lower temperature).
N-(2-Hyd r oxyben zyl)ser in e Hyd r och lor id e (10). A solu-
tion of L-serine (220 mg, 2.10 mmol) and (2-hydroxybenzyl)-
trimethylammonium iodide (2) (100 mg, 0.34 mmol) in 5 mL
Thus, we feel confident that such results will facilitate
further study and hopefully result in the use of o-QM as
covalent protein linker and enzyme inhibitor.
of NaHCO
and 2.4 mL of a Na
cooled, and then purified by preparative HPLC (eluent, H
CH CN ) 85:15 + 0.1% CF COOH, flux ) 4.2 mL/min, t
3
/Na
2
CO
3
buffer pH 10 (2.6 mL of a NaHCO
3
0.1 M
2
CO 0.1 M) was heated at 80 °C for 2 h,
3
2
O:
Exp er im en ta l Section
3
3
R
)
2
7 min). Trifluoroacetate anion of the adduct was replaced by
Gen er a l Meth od s. Melting points are uncorrected. El-
emental analyses were performed using a standard analyzer.
Infrared spectra were recorded as KBr disks, Nujol, or films
chlorine anion according to the procedure described above for
glycine adduct to give 71.0 mg (84% yield) of adduct 10 as
colorless crystals, which decomposes at T > 150 °C. IR (Nujol)
1
13
13
on a FT spectrophotometer. H, C, C-DEPT, and 2D-
correlated NMR spectra were recorded in D O solutions (unless
-
1
(
cm ) 3417 (very broad), 2520 (broad), 1688, 1597, 1203, 1143.
2
1
H NMR (D
.30-4.40 (m, 2 H), 6.90-7.00 (m, 2 H, aromatics), 7.30-7.45
m, 2 H, aromatics). 13C NMR δ 46.97, 59.65, 62.15, 116.36,
17.80, 121.43, 132.38, 132.68, 155.98, 171.32. Anal. Calcd for
Cl: C, 48.48; H, 5.70; N, 5.66; Cl, 14.31. Found: C,
2
O) δ 3.98-4.05 (m, 1 H), δ 4.05-4.10 (m, 2 H),
otherwise stated) on a 300 MHz spectrometer. Chemical shifts
are reported in ppm. Protons were correlated by decoupling
4
(
1
1
13
and H- C COSY experiments (HSQC and HMBC). Reaction
products were separated and quantified analytically by reverse-
phase (Intersil ODS-2, 5 µm, column dimension: φ ) 4.6 mm,
length ) 250 mm, φ ) 10.0 mm, length ) 250 mm, Micro-
Column) HPLC chromatography using a variable-wavelength
detector.
10 4
C H14NO
4
8.50; H, 5.69; N, 5.69; Cl, 14.35.
S-(2-Hyd r oxyben zyl)cystein e Hyd r och lor id e (11). A
solution of L-cysteine (50 mg, 0.41 mmol) and (2-hydroxyben-
zyl)trimethylammonium iodide (2) (100 mg, 0.34 mmol) in 5
Ma ter ia ls. Chemicals and solvents were purchased from
Aldrich and Sigma and used without further purification. All
aqueous solutions were made with distilled water.
mL of NaHCO
for 30 min, cooled, and then purified by preparative HPLC
eluent ) H O:CH CN ) 85:15 + 0.1% CF COOH, flux ) 4.0
mL/min, t ) 18.0 min). Trifluoroacetate anion of the adduct
3 2 3
/Na CO buffer (pH 10) was heated at 80 °C
(
2
3
3
(
2-Hyd r oxyben zyl)tr im eth yla m m on iu m Iod id e (2). (2-
R
Hydroxybenzyl)trimethylammonium iodide was prepared, in
almost quantitative yield, from 2-(dimethylaminomethyl)-
was replaced by chlorine anion according to the procedure
described above for glycine adduct to give 89.1 mg (99% yield)
of adduct 11 as colorless crystals, which decomposes at T >
4
4
phenol. Using the general protocol described in the literature,
the Mannich base (50 g, 33 mmol) was treated with methyl
iodide (4.27 mL, 69 mmol) in acetonitrile (15 mL) for 2 h. After
addition of 100 mL of diethyl ether, the insoluble quaternary
ammonium salt was filtered. The product was crystallized from
ethanol: diethyl ether ) 8:2, to give 8.29 g (92% yield) of white
-
1
1
45 °C. IR (Nujol) (cm ) 3395, 1694, 1598, 1207, 1188, 1149.
1
H NMR (D
(
3
2
O) δ 2.95 (dd, J ) 15.2 Hz, J ) 8.3 Hz, 1 H), 3.13
dd, J ) 15.2 Hz, J ) 4.1, 1 H), 3.76 (d, J ) 13.2 Hz, 1 H),
.83 (d, J ) 13.2 Hz, 1 H), 4.20 (dd, J ) 8.3 Hz, J ) 4.1 Hz, 1
4
5
1
H), 6.90-7.00 (m, 2 H, aromatics), 7.15-7.25 (m, 2 H,
crystals, mp ) 169 ( 0.5 °C (literature 169-170 °C ). H NMR
aromatics). 13C NMR δ 30.90, 31.59, 52.10, 118.73, 121.57,
(
D
2
O) δ 2.97 (s, 9 H, NMe
3 2
), 4.35 (s, 2 H, CH Ph), 6.88-6.95
1
24.86, 130.15, 131.8, 154.62, 171.43. Anal. Calcd for C10
NO SCl: C, 45.54; H, 5.35; N, 5.31; S, 12.16; Cl, 13.44.
Found: C, 45.49; H, 5.30; N, 5.25; S, 12.13; Cl, 13.40.
14
H -
(m, 2 H, aromatics), 7.25-7.35 (m, 2 H, aromatics).
4
6
46
47,48
49
50
3
Adducts 3, 4, 5,
6, and 7 were identified by
comparison of their spectroscopic properties with published
data.
E-N-(2-Hydr oxyben zyl)-L-lysin e Hydr och lor ide (12) an d
N-(2-Hyd r oxyben zyl)-L-lysin e Hyd r och lor id e (13). A solu-
tion of L-lysine (144 mg, 0.99 mmol) and (2-hydroxybenzyl)-
trimethylammonium iodide (2) (60 mg, 0.20 mmol) in 5 mL of
2
-(2-Hyd r oxy-eth ylsu lfa n ylm eth yl)p h en ol (8). A solu-
tion of 2-mercaptoethanol (40 mg, 0.51 mmol) and (2-hydroxy-
benzyl)trimethylammonium iodide (2) (100 mg, 0.34 mmol) in
NaHCO
3 2 3
/Na CO buffer adjusted at pH 9.6 was heated at 80
5
mL of KH
2
PO
4
/Na
2
HPO
4
buffer adjusted at pH 7.0 (2.0 mL
HPO 0.1 M) was
°
C for 1 h and then cooled. Two alkylation adducts 12 and 13
of a KH PO
2
4
0.1 M and 3.0 mL of a Na
2
4
were isolated and purified by preparative HPLC (eluent, H
CH CN ) 90:10 + 0.1% CF COOH, flux ) 4.0 mL/min, t
4.1 min and t
2
O:
heated at 80 °C for 3.5 h and then cooled. The solution was
acidified with 0.5 M HCl and extracted with diethyl ether. The
organic phase was washed with water and dried over MgSO .
4
3
3
R
)
1
R
) 12.2 min). Trifluoroacetate anion of the
adducts was replaced by chlorine anion according to the
procedure described above for glycine adduct to give adducts
12 (22.8 mg, 35% yield), and 13 (25.4 mg, 39% yield) as
hydrochlorides both as colorless crystals, which both decom-
posed at T ) 155 °C and 147 °C, respectively.
Evaporation of the solvent followed by column chromatography
on silica gel (eluent, diethyl ether:hexane ) 6:4) gave 59.5 mg
-
1
(
1
1
95%, yield) of 8 as colorless oil. IR (Nujol) (cm ) 3330 (broad),
1
596, 1456, 1247. H NMR (CDCl
3
) δ 2.20-2.60 (very broad,
H), 2.70 (t, J ) 6.8 Hz, 2 H), 3.80 (t, J ) 6.8 Hz, 2 H), 3.90
-1
1
(
s, 2 H), 6.00-6.75 (very broad, 1 H), 6.70-6.85 (m, 2 H,
12. IR (Nujol) (cm ) 3389, 1747, 1603. H NMR (D
(m, 2 H), 1.75 (m, 2 H), 1.90 (m, 2 H), 3.05 (t, J ) 8.0 Hz, 2 H,
CH N), 3.97 (t, J ) 6.0 Hz, 1 H, CHCOO), 4.20 (s, 2 H, CH
Ph), 6.90-7.00 (m, 2 H, aromatics), 7.15-7.25 (m, 2 H,
aromatics). 13C NMR δ 21.35 (CH
), 24.73 (CH ), 29.18 (CH ),
46.22 (CH ), 46.59 (CH ), 52.88 (CH), 115.38 (CH), 117.27 (C),
120.40 (CH), 131.31 (CH), 131.57 (CH), 154.85 (C), 172.38 (C).
2
O) δ 1.50
1
3
aromatics), 7.10-7.25 (m, 2 H, aromatics). C NMR δ 31.67,
3.46, 61.32, 116.70, 120.65, 123.12, 129.01, 130.59, 154.70.
Anal. Calcd for C S: C, 58.67; H, 6.56; S, 17.40. Found:
3
2
2
-
9
12 2
H O
C, 58.60; H, 6.52; S, 17.35.
N-(2-Hyd r oxyben zyl)glycin e Hyd r och lor id e (9). A solu-
tion of glycine (64 mg, 0.85 mmol) and (2-hydroxybenzyl)-
2
2
2
2
2

Alkylation of Amino Acids and Glutathione in Water
J . Org. Chem., Vol. 66, No. 1, 2001 51
stereogenic center is far in 14 and close in 15 to the CH . We
Anal. Calcd for C13
H
22
N
2
O
3
Cl
2
: C, 48.01; H, 6.82; N, 8.61; Cl,
2
2
1.28. Found: C, 48.11; H, 6.80; N, 8.59; Cl, 21.30.
confirmed the structures by HSQC and HMBC experiments.
-1
1
1
3. IR (Nujol) (cm ) 3572, 1735, 1615. H NMR (D
2
O) δ 1.40
S-(2-Hyd r oxyben zyl)glu ta th ion e (16). A solution of glu-
tathione (50 mg, 0.16 mmol) and (2-hydroxybenzyl)trimethyl-
ammonium iodide (2) (45 mg, 0.15 mmol) in 5 mL of KH PO /
(
m, 2 H), 1.65 (m, 2 H), 1.95 (m, 2 H), 2.95 (t, J ) 7.1 Hz, 2 H,
CH N), 3.85 (t, J ) 6.0 Hz, 1 H, CHCOO), 4.29 (s, 2 H, CH
Ph), 6.90-7.00 (m, 2 H, aromatics), 7.25-7.40 (m, 2 H,
2
2
-
2
4
Na HPO buffer pH 7.0 (2.0 mL of a KH PO 0.1 M and 3.0
2
4
2
4
1
3
aromatics). C NMR δ 21.23 (CH
2
), 26.09 (CH
), 59.07 (CH), 115.40 (CH), 116.64 (C),
20.47 (CH), 131.58 (CH), 131.89 (CH), 155.09 (C), 171.40 (C).
Cl : C, 48.01; H, 6.82; N, 8.61; Cl,
1.28. Found: C, 48.09; H, 6.77; N, 8.55; Cl, 21.33.
The structure of the adducts 12 and 13 have been firmly
2
), 28.49 (CH
2
),
mL of a Na HPO4 0.1 M) was heated at 80 °C for 30 min,
2
3
1
8.82 (CH
2
), 45.93 (CH
2
cooled, and then purified by preparative HPLC (eluent, H O:
2
CH CN ) 80:20 + 0.1% CF COOH, flux ) 4.0 mL/min, t
3
3
R
)
Anal. Calcd for C13
2
H
22
N
2
O
3
2
10.37 min). Trifluoroacetate anion of the adducts was replaced
by chlorine anion according to the procedure described above
for glycine adduct to give adduct 16 (57 mg, 84% yield).
1
3
-1
based on spectroscopic evidences, in particular on C NMR
chemical shifts, DEPT and H- C long-range correlation
experiments (HMBC). In more detail, on passing from 12 to
Colorless crystals mp ) 51 ( 1 °C. IR (Nujol) (cm ) 3410
1
13
(broad), 1732, 1652, 1547, 1244. 1H NMR (D O) δ 2.11 (dt,
2
J ) 7.1 Hz, 6.7 Hz, 2 H, H-3), 2.44 (t, J ) 7.1 Hz, H-4, 2 H),
2.72 (dd, J ) 14.3 Hz, J ) 8.7 Hz, 1 H, H-12), 2.91 (dd, J )
14.3 Hz, J ) 5.2 Hz, 1 H, H-12′), 3.67 (d, J ) 13.0 Hz, 1 H,
H-14), 3.73 (d, J ) 13.0 Hz, 1 H, H-14′), 3.89 (s, 2 H, H-10),
3.97 (t, J ) 6.7 Hz, 1 H, H-2), 4.41 (dd, J ) 8.7 Hz, J ) 5.2
Hz, 1 H, H-7), 6.81-6.88 (m, 2 H, H-17, H-19, aromatics),
1
3 both the carbon atoms at the R position (C-2) and that at
position C-6 (see Scheme 2 for numbering) of the amino acid
moiety experience a diagnostic chemical shift change. C-2 is
more deshielded (59.07 ppm) in the adduct 13 than the
corresponding carbon atom in the adduct 12 (52.88 ppm). Vice
versa, C-6 in the adduct 13 is more shielded (38.82 ppm) than
the corresponding carbon atom in the adduct 12 (46.59 ppm).
This difference is likely due to the fact that alkylated amino
group gives rise to higher deshielding effect than the corre-
7.10-7.17 (m, 1 H, H-18, aromatic), 7.17-7.20 (m, 1 H, H-20
aromatic). 13C NMR δ 25.30 (C-3), 30.17 (C-14), 30.70 (C-4),
32.24 (C-12), 40.90 (C-10), 52.06 (C-2), 52.83 (C-7), 115.71 (C-
17), 120.50 (C-19), 124.39 (C-15), 128.97 (C-18), 130.76 (C-20),
153.53 (C-16), 171.40 (C-1), 172.64 (C-8 and C-11), 174.03 (C-
5). Anal. Calcd for C H N O ClS: C, 45.38; H, 5.38; N, 9.34;
2
sponding NH .
After a full assignment of H and 13C spectra by HSQC, the
1
1
7
24
3
7
identity of 12 and 13 have been definitely established by long
range H- C coupling detected through a HMBC protocol,
Cl, 7.88; S, 7.13. Found: C, 45.29; H, 5.35; N, 9.30; Cl, 7.89;
S, 7.09.
1
13
51
since each one exhibits a unique connectivity between the
benzylic protons (H-8) and lysine carbon atoms (C-6 in 12 or
C-2 in 13).
N-(2-Hyd r oxyben zyl)-L-tyr osin e Hyd r och lor id e (14)
a n d O-(2-Hyd r oxyben zyl)-L-tyr osin e Hyd r och lor id e (15).
A solution of L-tyrosine (87 mg, 0.48 mmol), and (2-hydroxy-
benzyl)trimethylammonium iodide (2) (50 mg, 0.18 mmol) in
HSQC and HMBC experiments were used to fully assign
every signal in both 1H and 13C spectra. In the alkylated
product 16 (Scheme 2), the connectivity between the benzylic
carbon (C-14) and protons (H-14) with C-12 and H-12 of
peptide cysteine moiety was established by the combined used
of HSQC and HMBC technique.
Oligom er s. We isolated an oligomeric mixture as a pale
5
mL of NaHCO
for 1 h. Then, two alkylated products 14 and 15 were isolated
and purified by preparative HPLC (eluent ) H O:CH CN )
0:20 + 0.1% CF COOH, flux ) 4.0 mL/min, t ) 10.3 min
and t ) 20.5 min, respectively). Trifluoroacetate anion of the
3
/Na
2
CO
3
buffer (pH 10) was heated at 80 °C
-1
yellow powder, mp ) 177-185 °C. IR (KBr) (cm ) 3261
1
(
(
(
broad), 1607, 1453, 1256. H NMR (DMSO) δ 2.80-3.10
2
3
+
NMe
3
), 3.60-3.90 (CH
2
O), 4.10-4.30 (CH
2
N), 6.50-7.10
8
3
R
aromatics), 9.4-10.2 (m, broad, phenolic OH). The integral
R
ratio of the signals at δ 3.60-3.90 and 4.10-4.30 was 6.5. This
value suggest that the average of the o-QM units involved in
the oligomeric structures is 6-7.
adducts was replaced by chlorine anion according to the
procedure described for glycine adduct, to give 14 (37 mg, 64%
yield), 15 (7 mg, 12% yield) both as hydrochloride, and 17 (2
P r ep a r a tive P h otoch em ica l Rea ction s. Gen er a l P r o-
ced u r e. A phosphate buffer solution (10 mL, adjusted at pH
7.0) of (2-hydroxybenzyl)trimethylammonium iodide (28 mg,
mg, 10% yield, t
R
) 11.5 min).
1
4. White crystals, mp ) 114 ( 1 °C followed by decomposi-
-1
1
tion. IR (Nujol) (cm ) 3290, (very broad), 1730, 1518. H NMR
O) δ 3.06 (dd, J ) 14.7 Hz, J ) 7.5 Hz, 1 H, H-3), 3.15 (dd,
J ) 14.7 Hz, J ) 6.6 Hz, 1 H, H-3′), 3.99 (dd, J ) 7.5 Hz, J )
.6 Hz, 1 H, H-2), 4.15 (d, J ) 13.2 Hz, 1 H, H-11), 4.22 (d,
-
2
0
.1 mmol, 10 M) and cysteine (24 mg, 0.2 mmol) was placed
(D
2
in a quartz tube (20 mL capacity) which was then sealed with
a rubber serum cap and purged with nitrogen. The sample was
then irradiated at 254 nm for 1 h at room temperature, by
means of two 15 W low-pressure mercury lamps. After irradia-
tion and 10-fold dilution, the alkylation product 11 was
identified and quantified by HPLC analysis by comparison
with a solution containing an authentic sample (obtained from
thermal experiments) in a known concentration.
6
J ) 13.2 Hz, 1 H, H-11′), 6.73-6.89 (m, 4 H, aromatics), 6.99-
7
7
6
1
.14 (m, 2 H, aromatics), 7.14-7.18 (m, 1 H, aromatic), 7.22-
.29 (m, 1 H, aromatic). 13C NMR δ 34.39 (C-3), 46.35 (C-11),
0.21 (C-2), 115.26 (C-14), 115.74 (C-6 and C-8), 116.24 (C-
2), 120.42 (C-16), 125.23 (C-4), 130.49 (C-5 and C-9), 131.54
(
1
C-15), 131.76 (C-17), 154.86 (C-7), 154.97 (C-13), 170.88 (C-
). Anal. Calcd for C16 Cl: C, 58.35; H, 5.60; N, 4.33;
Cl, 10.95. Found: C, 58.30; H, 5.62; N, 4.27; Cl, 10.85.
The other reactions with glycine, serine, lysine, and glu-
tathione were similarly carried out.
H18NO
4
-
1
P h otoch em ica l Sta bility of th e Ad d u cts. Gen er a l P r o-
ced u r e. A phosphate buffer solution (10 mL, adjusted at pH
7.0) of ꢀ-N-(2-hydroxybenzyl)-L-lysine hydrochloride (12) (3.3
1
5. Colorless oil. IR (Nujol) (cm ) 3270, (very broad), 1725,
1
1
2
502. H NMR (D O) δ 2.90 (dd, J ) 14.2 Hz, J ) 7.7 Hz, 1 H,
H-3), 3.15 (dd, J ) 14.2 Hz, J ) 5.3 Hz, 1 H, H-3′), 3.82 (s, 2
H, H-11), 3.89 (dd, J ) 7.7 Hz, J ) 5.3 Hz, 1 H, H-2), 6.79-
6
7
6
-3
mg, in 10 mL phosphate buffer, 10 M) was placed in a quartz
tube (20 mL capacity) which was then sealed with a rubber
serum cap and purged with nitrogen. The sample was then
irradiated at 254 nm for 10 min at room temperature. After
irradiation without further dilution, hydroxybenzyl alcohol (17)
was measured by HPLC analysis by comparison with a
solution containing an authentic sample.
.90 (m, 4 H, aromatics), 6.93-6.99 (m, 1 H, aromatic), 7.03-
.14 (m, 3 H, aromatics). 13C NMR δ 35.12 (C-3), 55.38 (C-2),
9.41 (C-11), 115.45 (CH), 115.85 (CH), 120.76 (CH), 126.55
(
(
C), 126.83 (C), 127.67 (CH), 127.83 (CH), 128.35 (CH), 130.52
CH), 131.09 (CH), 152.71 (C-7), 153.65 (C-13), 171.48 (C-1).
Anal. Calcd for C16
0.95. Found: C, 58.20; H, 5.67; N, 4.25; Cl, 10.86.
Tyrosine alkylation products on N and O atoms are distin-
4
H18NO Cl: C, 58.35; H, 5.60; N, 4.33; Cl,
1
P h otoch em ica l Sta bility of th e Ad d u cts a t p H 12.
Gen er a l P r oced u r e. ꢀ-N-(2-hydroxybenzyl)-L-lysine hydro-
1
3
-2
guishable through the C chemical shift and the multiplicity
chloride (12) (3.2 mg, 1 × 10 mmol) and cysteine (1.3 mg,
1
-2
of the H signals of the methylene group, i.e., CH
2
at position
1.1 × 10 mmol) were dissolved in 10 mL of water in order to
-
3
1
1, deriving from the former o-quinone methide moiety. C-11
obtain a 10 M solution in both reagents. The pH of the
solution was then adjusted to 12 by NaOH addition (0.5 M).
The sample was then irradiated at 254 nm for 10 min at room
temperature. After irradiation, alkylation adduct 11 was
in 14 is more shielded (46.35 ppm) than the corresponding
carbon atom in 15 (69.41 ppm). H-11 protons resonate as an
AB system in 14 and as a singlet in 15, indicating that the

5
2
J . Org. Chem., Vol. 66, No. 1, 2001
Modica et al.
measured by HPLC analysis by comparison with a solution
containing an authentic sample.
Com p etition Exp er im en ts. Gen er a l P r oced u r e. An
aqueous solution (5 mL) of (2-hydroxybenzyl)trimethylammo-
were obtained from the fit of the absorbance data to a single-
exponential function and were reproducible to ( 2%. The
-
1
-1
second-order rate constants kNu (M
s ) for the reaction of
nucleophiles with 1 were determined as the least-squares
slopes of linear plots of kobsd against the total concentration of
nium iodide (2, 10 mg, 0.034 mmol), t-BuNH
mmol), PrNH (40 mg, 0.68 mmol), and Na CO
2
(50 mg, 0.68
(144 mg, 1.35
-
2
-4
2
2
3
rt ha ne gn eu.cleophile, used in 5 × 10 to 2 × 10 M concentration
mmol) was kept for 5 min at 80 °C in order to achieve a
conversion yield of the substrate 2 lower than 30%. The ratio
Kin etic Stu d ies. Hyd r a tion . Kinetic studies of acid- and
base catalyzed o-QM hydration were similarly carried out. No
buffers were used to change the pH, which was adjusted in
the range 1 < pH < 3.4 or 11 < pH < 13 by HCl or NaOH
addition. In a fashion similar to nucleophile kinetic studies
ionic strength was kept constant at I ) 0.1 M by NaCl.
La ser F la sh P h otolysis. The laser pulse photolysis ap-
paratus consisted of a Nd:YAG laser used at the fourth
harmonic of its fundamental wavelength. It delivered a
maximum power of 10 mJ at 266 nm with 10 ns pulse duration.
The monitor system, arranged in a cross-beam configuration,
consisted of a 275 W Xe arc lamp, an F/3.4 monochromator,
and a five-stage photomultiplier supplied by Applied Photo-
physics. The signals were captured by means of a Hewlett-
Packard 54510A digitizing oscilloscope, and the data was
processed on a 486-based computer system using software
developed in-house. Solutions for analysis were placed in a
fluorescence cuvette (d ) 10 mm). The absorbance of each
solution was adjusted to 1.5.
3
/4 ) 4.9 of the resulting alkylated products has been
measured by HPLC.
A water solution (15 mL) obtained by mixing a t-BuNH
2
solution (5 mL, 0.18 M), a n-prNH
2-hydroxybenzyl)trimethylammonium iodide solution (5 mL,
.018 M), and Na CO (360 mg, 3.40 mmol) in a quartz tube
20 mL capacity) was flushed with nitrogen and then irradi-
ated at 254 nm for 45 min at room temperature. The ratio
2
solution (5 mL, 0.18 M), a
(
0
2
3
(
3
/4 ) 4.6 of the resulting alkylated products has been
measured by HPLC.
The other competition experiments with morpholine/piperi-
-
dine, and t-BuNH
2
/OH , were similarly carried out.
Kin etic Stu d ies. Nu cleop h ile Rea ctivity. Kinetic studies
of nucleophiles addition to o-QM (with the exception of
hydration), were carried out at 25 °C, keeping constant at 7.0
or 12.0 the pH (according to Table 4). Phosphate buffer or a
NaOH solution were used to adjust the pH. The pH of each
solution was measured using an Orion SA520 pH meter with
a 8102 Ross electrode. Ionic strength was also kept constant
at I ) 0.1 M by NaCl, since Cl- is unreactive with o-QM in
water. 1 was produced by LFP, flashing a dilute solution of 2
Ack n ow led gm en t. Financial support from Pavia
University (“Fondo Giovani Ricercatori 1999”) is grate-
fully acknowledged.
-
4
(
5 × 10 M). The disappearance of 1 was followed, under
pseudo first-order conditions, by monitoring the absorbance
decrease at 400 nm. Pseudo-first-order rate constant (kobsd
)
J O0006627