Glutathione peroxidase (GPx)-like antioxidant activity of the organoselenium drug ebselen: Unexpected complications with thiol exchange reactions

Article abstract of DOI:10.1021/ja052794t

The factors that are responsible for the relatively low glutathione peroxidase (GPx)-like antioxidant activity of organoselenium compounds such as ebselen (1, 2-phenyl-1,2-benzisoselenazol-3(2H)-one) in the reduction of hydroperoxides with aromatic thiols

Full text of DOI:10.1021/ja052794t

Published on Web 07/23/2005
Glutathione Peroxidase (GPx)-like Antioxidant Activity of the
Organoselenium Drug Ebselen: Unexpected Complications
with Thiol Exchange Reactions
Bani Kanta Sarma and G. Mugesh*
Contribution from the Department of Inorganic & Physical Chemistry,
Indian Institute of Science, Bangalore 560 012, India
Received April 29, 2005; E-mail: mugesh@ipc.iisc.ernet.in
Abstract: The factors that are responsible for the relatively low glutathione peroxidase (GPx)-like antioxidant
activity of organoselenium compounds such as ebselen (1, 2-phenyl-1,2-benzisoselenazol-3(2H)-one) in
the reduction of hydroperoxides with aromatic thiols such as benzenethiol and 4-methylbenzenethiol as
cosubstrates are described. Experimental and theoretical investigations reveal that the relatively poor GPx-
like catalytic activity of organoselenium compounds is due to the undesired thiol exchange reactions that
take place at the selenium center in the selenenyl sulfide intermediate. This study suggests that any
substituent that is capable of enhancing the nucleophilic attack of thiol at sulfur in the selenenyl sulfide
state would enhance the antioxidant potency of organoselenium compounds such as ebselen. It is proved
that the use of thiol having an intramolecularly coordinating group would enhance the biological activity of
ebselen and other organoselenium compounds. The presence of strong S‚‚‚N or S‚‚‚O interactions in the
selenenyl sulfide state can modulate the attack of an incoming nucleophile (thiol) at the sulfur atom of the
-Se-S- bridge and enhance the GPx activity by reducing the barrier for the formation of the active species
selenol.
Scheme 1. GPx Cycle of Ebselena
Introduction
Ebselen (1, 2-phenyl-1,2-benzisoselenazol-3(2H)-one), a cy-
clic organoselenium compound with low toxicity, exhibits
interesting therapeutic properties for a number of disease states
including antiinflammatory activities.1 These properties are
apparently due to the ability of ebselen to catalyze the reduction
of hydroperoxides by glutathione (GSH) or other thiols,
mimicking the catalytic activity of one of the selenoenzymes,
glutathione peroxidase (GPx).2 Because of these properties,
ebselen is being used as a standard for comparing the GPx
activity of selenium compounds. Despite its importance in
biology and medicine, the catalytic GPx cycle of ebselen is
controversial probably due to major differences in working
conditions such as solvents, pH, and the nature of the hydro-
peroxides that are used by many research groups to characterize
the potential catalytic intermediates.3 Although several mech-
anisms have been proposed to explain the observed GPx activity
of ebselen, the available information reveals a hypothetical
a GSH has been replaced by PhSH.
catalytic cycle as shown in Scheme 1. According to this model,
the Se-N bond in ebselen is readily cleaved by thiols to produce
the corresponding selenenyl sulfides, which upon reduction by
excess thiols produce selenol 3. The catalytically active selenol
then reduces hydroperoxides to form a selenenic acid, which
further reacts with thiols to regenerate the selenenyl sulfides.
However, recent studies have shown that ebselen, irrespective
of the reaction pathway, is a relatively inefficient catalyst in
the reduction of hydroperoxides with aryl and benzylic thiols
such as PhSH and BnSH as cosubstrates,4 and the reason for
the relatively low activity is still not clear. Back et al. have
noted a similar lack of reactivity with certain other selenenyl
(1) (a) Mu¨ller, A.; Cadenas, E.; Graf, P.; Sies, H. Biochem. Pharmacol. 1984,
33, 3235 (b) Wendel, A.; Fausel, M.; Safayhi, H.; Tiegs, G.; Otter, R.
Biochem. Pharmacol. 1984, 33, 3241. (c) Sies, H. Angew. Chem., Int. Ed.
1986, 25, 1058. (d) Sies, H. Free Radical Biol. Med. 1993, 14, 313. (e)
Sies, H.; Masumoto, H. AdV. Pharmacol. 1997, 38, 22229.
(2) (a) Flohe´, L.; Gu¨nzler, E. A.; Schock, H. H. FEBS Lett. 1973, 32, 132. (b)
Rotruck, J. T.; Pope, A. L.; Ganther, H. E.; Swanson, A. B.; Hafeman, D.
G.; Hoekstra, W. G. Science 1973, 179, 588. (c) Bock, A. Selenium Proteins
Containing Selenocysteines. In Encyclopedia of Inorganic Chemistry; King,
R. B., Ed.; John Wiley & Sons: Chichester, England, 1994; Vol. 8. p 3700.
(d) Birringer, M.; Pilawa, S.; Flohe´, L. Nat. Prod. Rep. 2002, 19, 693. (e)
Jacob, C.; Giles, G. I.; Giles, N. M.; Sies, H. Angew. Chem., Int. Ed. 2003,
42, 4742.
(3) (a) Fisher, H.; Dereu, N. Bull. Soc. Chim. Belg. 1987, 96, 757. (b) Kice, J.
L.; Purkiss, D. W. J. Org. Chem. 1987, 52, 3448. (c) Engman, L.; Stern,
D.; Cotgreave, I. A.; Andersson, C. M. J. Am. Chem. Soc. 1992, 114, 9737.
(d) Mugesh, G.; Singh, H. B. Chem. Soc. ReV. 2000, 29, 347.
9
10.1021/ja052794t CCC: $30.25 © 2005 American Chemical Society
J. AM. CHEM. SOC. 2005, 127, 11477-11485
11477

A R T I C L E S
Sarma and Mugesh
Table 1. Initial Rates (v0) for the Reduction of H2O2 (7 mM) by
PhSH (5 mM), 4-MeC6H4SH (5 mM), or GSH in the presence of
Ebselen (0.47 mM) at 23 °C
1
thiol
method
vo
(
µM
‚
min-
)
control
ebselen
GSH
PhSH
4-MeC6H4SH
GSH
PhSH
GSH-GSSGa
HPLC
HPLC
GSH-GSSG
HPLC
15.11 ( 0.69
2.46 × 10-3
1.85 × 10-3
91.68 ( 2.37
0.38 × 10-3
0.56 × 10-3
4-MeC6H4SH
HPLC
Figure 1. Nucleophilic attack of thiol at selenium or sulfur.
a Reactions were carried out in 0.1 M phosphate buffer, pH 7.24, with
EDTA (1 mM), GSH (1 mM), NADPH (0.2 mM), GSSG reductase (0.6
unit/mL), ebselen (0.025 mM), and H2O2 (2 mM).
sulfides and have shown that these compounds undergo a
deactivation pathway that competes with the main catalytic
cycle.4b The use of 3-carboxy-4-nitrobenzenethiol, which has
been shown to enhance the GPx activity of the semisynthetic
enzyme selenosubtilisin,5 does not increase the catalytic ef-
ficiency of ebselen.4e Because the therapeutic effect of ebselen
has been linked to its peroxidase activity and the peroxidase
activity of this compound depends on the reduction of the
selenenic acid 4 to the corresponding selenol 3 by thiols, it is
important to study the effect of the nature of the thiols on GPx-
like catalytic activity. Although the effect of various substituents
attached to the selenium moiety on the GPx cycle of selenium
compounds has been extensively studied,3d,6 the effect of such
substituents attached to the thiol cosubstrates is very poorly
understood. It has been reported that dithiols such as dihydro-
lipoate may serve better as cofactors than GSH in the peroxidase
activity of ebselen.7,8
Because the selenium center in GPx readily changes its
oxidation states during the catalytic cycle, several weak interac-
tions between selenium and nearby amino acid moieties have
been shown to modulate such changes. The objective of this
work is not to study the nature of Se‚‚‚X (X ) heteroatom)
interactions in selenium compounds10,13,14 but to describe, with
the help of 77Se NMR spectroscopic studies, HPLC methods,
and theoretical calculations, the role of heteroatoms in the GPx
activity of ebselen and related derivatives. In this paper, we
also show that the low activity of ebselen and related derivatives
is mainly due to the thiol exchange reactions and a novel
approach to overcome these problems.
Results and Discussion
The GPx-like activity of ebselen was studied with H2O2 as a
substrate and GSH, PhSH, and 4-MeC6H4SH as thiol cosub-
strates. The catalytic activity of ebselen with GSH was studied
by using the classical GSH-GSSG coupled assay,9 and with
aromatic thiols it was studied by using reversed-phase HPLC
methods. The initial rates (V0) for the reduction of H2O2 by thiols
in the presence and absence of ebselen were calculated from
the first 5-10% of the reaction by choosing a linear fit (Table
1). The catalytic rates were corrected for the background reaction
between H2O2 and thiols.
Interestingly, the rates for reduction of H2O2 in the presence
of aromatic thiols were found to be much lower than those
observed with GSH and also significantly lower than those of
uncatalyzed reductions. In this regard, not only is ebselen an
inactive catalyst in the PhSH and 4-MeC6H4SH thiol systems,
but it also acts as an inhibitor of its own redox reactions. To
probe this unexpected behavior, the active intermediates of
ebselen in the GPx cycle (2-4) were studied using 77Se NMR
spectroscopy, HPLC methods, and quantum chemical calcula-
tions to identify the role of carbonyl oxygen during the catalytic
cycle. Because the intramolecular Se‚‚‚O/Se‚‚‚N nonbonded
interactions result in an apparent downfield shift of the 77Se
NMR chemical shift and an increase in the electropositive
character of selenium,13,15 the calculated Se‚‚‚O/Se‚‚‚N distances
and the theoretical and, in some cases, experimental 77Se NMR
It is known that basic amino groups in the active site of GPx
or similar groups in model compounds interact with selenium
to modulate its reactivity.9,10 Interestingly, although such
interactions have been shown to increase the electrophilic
reactivity of selenium, these interactions in some of the
intermediates become detrimental to the biological activity of
synthetic selenium compounds. For example, strong Se‚‚‚N
interactions in the selenenyl sulfide intermediate in the GPx
cycle enhance a nucleophilic attack of thiol at selenium instead
of the desired attack at sulfur, leading to a thiol exchange
reaction that would hamper the regeneration of the active species
selenol. Recent studies on model compounds show that the
reduction of selenenyl sulfides to selenol requires surmounting
a substantial barrier (∼50 kcal/mol),11 and the nuleophilic attack
of thiol at selenium is both kinetically and thermodynamically
more favorable than at sulfur (Figure 1).12
(4) (a) Back, T. G.; Moussa, Z. J. Am. Chem. Soc. 2002, 124, 12104. (b) Back,
T. G.; Moussa, Z. J. Am. Chem. Soc. 2003, 125, 13455. (c) Back, T. G.;
Moussa, Z.; Parvez, M. Angew. Chem., Int. Ed. 2004, 43, 1268. (d) Zhang,
X.; Xu, H.; Dong, Z.; Wang, Y.; Liu, J.; Shen, J. J. Am. Chem. Soc. 2004,
126, 10556. (e) Dong, Z.; Liu, J.; Mao, S.; Huang, X.; Yang, B.; Ren, X.;
Luo, G.; Shen, J. J. Am. Chem. Soc. 2004, 126, 16395.
(5) (a) Wu, Z.-P.; Hilvert, D. J. Am. Chem. Soc. 1989, 111, 4513. (b) Wu,
Z.-P.; Hilvert, D. J. Am. Chem. Soc. 1990, 112, 5647. (c) House, K. L.;
Dunlap, R. B.; Odom, J. D.; Wu, Z.-P.; Hilvert, D. J. Am. Chem. Soc.
1992, 114, 8573. (d) Syed, R.; Wu, Z.-P.; Hogle, J. M.; Hilvert, D.
Biochemistry 1993, 32, 6157. (e) House, K. L.; Garber, A. R.; Dunlap, R.
B.; Odom, J. D.; Hilvert, D. Biochemistry 1993, 32, 3468.
(6) Mugesh, G.; du Mont, W.-W.; Sies, H. Chem. ReV. 2001, 101, 2125 and
references therein.
(7) Haenen, G. R. M. M.; De Rooij, B. M.; Vermeulen, N. P. E.; Bast, A.
Mol. Pharmacol. 1990, 37, 412.
(8) Biewenga, G. Ph.; Bast, A. Methods Enzymol. 1995, 251, 303.
(9) Wilson, S. R.; Zucker, P. A.; Huang, R.-R. C.; Spector, A. J. Am. Chem.
Soc. 1989, 111, 5936.
(10) Iwaoka, M.; Tomoda, S. J. Am. Chem. Soc. 1994, 116, 2557.
(11) Benkova, S.; Ko´n˜a, J.; Gann, G.; Fabian, W. M. F. Int. J. Quantum Chem.
2002, 90, 555.
(12) Bachrach, S.; Demoin, D. W.; Luk, M.; Miller, J. V., Jr. J. Phys. Chem.
2004, 108, 4040.
(13) For excellent articles on Se‚‚‚X (X ) N, O, Cl, Br, F, etc.) interactions,
see: (a) Iwaoka, M.; Tomoda, S. Phosphorus, Sulfur Silicon Relat. Elem.
1992, 67, 125. (b) Iwaoka, M.; Tomoda, S. J. Am. Chem. Soc. 1996, 118,
8077. (c) Komatsu, H.; Iwaoka, M.; Tomoda, S. Chem. Commun. 1999,
205. (d) Iwaoka, M.; Komatsu, H.; Tomoda, S. Chem. Lett. 1998, 969. (e)
Iwaoka, M.; Komatsu, H.; Katsuda, T.; Tomoda, S. J. Am. Chem. Soc.
2002, 124, 1902. (f) Iwaoka, M.; Katsuda, T.; Tomoda, S.; Harada, J.;
Ogawa, K. Chem. Lett. 2002, 518. (g) Iwaoka, M.; Komatsu, H.; Katsuda,
T.; Tomoda, S. J. Am. Chem. Soc. 2004, 126, 5309. (h) Iwaoka, M.;
Katsuda, T.; Komatsu, H.; Tomoda, S. J. Org. Chem. 2005, 70, 321.
(14) For interesting papers on Se‚‚‚O interactions, see: (a) Spichty, M.; Fragale,
G.; Wirth, T. J. Am. Chem. Soc. 2000, 122, 10914. (b) Wirth, T.; Fragale,
G.; Spichty, M. J. Am. Chem. Soc. 1998, 120, 3376.
9
11478 J. AM. CHEM. SOC. VOL. 127, NO. 32, 2005

GPx-like Antioxidant Activity of Ebselen
A R T I C L E S
Table 2. Theoretical Data for 1-4 Obtained by DFT Calculations
at the B3LYP/6-31G(d)//B3LYP/6-311+G(d,p) Levels along with
the GIAO 77Se NMR Chemical Shifts
with a weak Se‚‚‚O interaction even after inclusion of the solvent
effect. This suggests that selenol 3, in contrast to 5, may exist
in the neutral form, favored by about 6-7 pKa units over the
corresponding zwitterion. Therefore, the existence of 5 as the
selenolate accounts for the higher reactivity of 5 as compared
with the undissociated selenol 3.
rSe
θO
δ(
77Se NMR)a
δ(
77Se NMR)
ESe
‚‚‚O
‚‚‚O
/
N
-Se-X
compd
(Å)
(deg)
qSe
(calcd) (ppm)
(exptl) (ppm)
(kcal/mol)
1
2
3
4
1.893
86.6b 0.622
941
604
203
959
588
232
2.470 177.3
2.655 166.0
2.375 172.6
0.377
0.227
0.732
19.01
7.29
28.05
1110
1143
a Referenced to the peak for Me2Se. b N-Se-C (C atom of the benzene
ring fused to the five-membered ring).
chemical shifts were used to probe the reactivity of various
selenium species. The Se‚‚‚O/Se‚‚‚N distances and the 77Se
NMR chemical shifts along with some important theoretical data
for compounds 1-4 are summarized in Table 2.
The stability of selenenic acids against further oxidation and
their fast reaction with thiols are the two major factors in the
second step of the catalytic cycle. The secondary amino group in
4 may not play a crucial role in the activation of selenenic acid.
Instead, the carbonyl oxygen is expected to interact with the
selenium as shown by the crystal structure of 7 in which the
oxygen atom interacts with selenium.18 In agreement with this,
the B3LYP/6-31G(d)-level-optimized geometry of 4 shows the
presence of strong Se‚‚‚O interactions (ESe‚‚‚O ) 28.05 kcal/mol).
The experimental 77Se NMR chemical shift (1143 ppm) for this
compound is in close agreement with the calculated value (1110
ppm), indicating the presence of a strong Se‚‚‚O interaction in
solution. The experimental 77Se NMR chemical shift value is sig-
nificantly downfield shifted compared with that of o-nitrobenzene-
selenenic acid (1066 ppm), but this value is close to that reported
for the selenenic acid derived from 6 (1173 ppm).10,19 In the
natural GPx cycle, such interactions have been shown to facili-
tate the attack of nucleophilic sulfur at the electrophilic selenium
atom, thereby producing the selenenyl sulfide. The crystal struc-
ture of human plasma GPx shows that the selenium atom is in-
volved in weak interactions with Gln79 and Trp153 in the selen-
inic acid state (Se‚‚‚N distances of 3.5 and 3.6 Å, respectively)
in addition to the normal hydrogen-bonding interactions of other
amino acid residues which are located in close proximity to
selenium.20
The first step of the catalytic cycle of ebselen after the
reduction of selenenyl sulfide 2 by thiols is believed to be the
oxidation of selenol 3 by hydrogen peroxide to produce the
corresponding selenenic acid 4. It has been proposed that the
introduction of amino/imino/ether groups in close proximity to
selenium could enhance the GPx activity of organoselenium
model compounds.13b The heteroatoms present in these mol-
ecules have been proposed to (i) deprotonate the selenol moiety
to increase its nucleophilic character and (ii) interact with
selenium in the selenenic acid state to increase the electrophilic
reactivity of selenium and hence to enhance the reaction with
thiols before the possible conversion to “overoxidized” selenium
species.9,10,13b,16 Therefore, theoretical investigations have been
carried out to determine the role of the amide moiety in the
catalysis. The B3LYP/6-31G(d)-level-optimized geometry and
the calculated 77Se NMR chemical shift for the selenol 3 show
that the carbonyl oxygen or the -NH- group does not facilitate
the deprotonation of the selenol moiety. The experimental 77Se
NMR chemical shift of 232 ppm for this compound is in close
agreement with the calculated value (203 ppm), but the signal
is very downfield shifted compared with that of 5 having a basic
amino group (6 ppm). Therefore, the GPx-like activity of 3 is
expected to be lower than that of 5, and this is indeed the case
as shown in this study with aromatic thiols. Unexpectedly, the
calculations on selenol 5 show that the basic amino group does
not help in deprotonating the selenol, but it helps in stabilizing
the selenol moiety by weak Se‚‚‚N interaction with a stabiliza-
tion energy (ESe‚‚‚N) of 5.19 kcal/mol. Because the pKa of a
quarternary trialkylammonium species is much higher (∼10-
11) than that of an aromatic selenol (∼6), a selenol such as 5
having a tertiary amine side chain would be driven toward the
zwitterion by about 4-5 pKa units. To understand the nature
of the selenol moiety in aqueous solution, we included the
solvent effects in the calculations by using the isodensity
polarized continuum model (IPCM).17 The single-point energy
calculations in water medium at the B3LYP/6-31G(d)-level-
optimized gas-phase geometries show that the zwitterion form
of 5 is at least 2.37 kcal/mol more stable in water than that
with a weak Se‚‚‚N interaction. On the other hand, the zwitterion
form of 3 was found to be 5.45 kcal/mol less stable than that
As selenenyl sulfides are the crucial intermediates in the GPx
cycle, we next carried out a detailed theoretical investigation
on some selenenyl sulfides derived from compounds having
heteroatoms in close proximity to selenium. In general, the
reduction of selenenic acids by thiols to the corresponding
selenenyl sulfides increases the shielding of selenium nuclei.
As expected, the B3LYP/6-31G(d)-level-optimized geometry of
the selenenyl sulfide 2 shows a strong Se‚‚‚O interaction (rSe‚
‚‚ ) 2.470 Å; ESe‚‚‚O ) 19.01 kcal/mol), and the calculated
O
77Se NMR chemical shift for this compound (604 ppm) is very
downfield shifted compared to that of PhSeSPh (437 ppm),
suggesting that the electron density around the selenium atom
in 2 is considerably reduced by the coordinating group. It should
be noted that the mesomeric effect of an aromatic ring having
a delocalizing/electron-withdrawing carbonyl group in the ortho
position may also reduce the electron density around selenium.
However, the cross conjugation of the amide nitrogen with the
π-electron system of the benzene ring may allow only a weak
(15) (a) Nakanishi, W.; Hayashi, S.; Toyota, S. Chem. Commun. 1996, 371. (b)
Nakanishi, W.; Hayashi, S.; Sakaue, A.; Ono, G.; Kawada, Y. J. Am. Chem.
Soc. 1998, 120, 3635.
(16) Spector, A.; Wilson, S. R.; Zucker, P. A. U.S. Patent, 321,138 (C1.546-
224, C07C37/02), 1994; Chem. Abstr. 1994, 121, P256039r.
(17) Foresman, J. B.; Keith, T. A.; Wiberg, K. B.; Snoonian, J.; Frisch, M. J. J.
Phys. Chem. 1996, 100, 16098.
(18) Fong, M. C.; Gable, R. W.; Schiesser, C. H. Acta Crystallogr., C 1996,
52, 1886.
(19) Reich, H. J.; Willis, W. W.; Wollowitz, S. Tetrahedron Lett. 1982, 23,
3319.
(20) Ren, B.; Huang, W.; Åkesson, B.; Ladenstein, R. J. Mol. Biol. 1997, 268,
869.
9
J. AM. CHEM. SOC. VOL. 127, NO. 32, 2005 11479

A R T I C L E S
Sarma and Mugesh
Table 3. A Comparison of the Calculated 77Se NMR Chemical
Shift Values of a Few Selenenyl Sulfides Having S‚‚‚N/O and
Se‚‚‚N/O Interactions
mesomeric effect. This can be rationalized by comparing the
electron density on the sulfur atom in 2 with that of PhSSePh.
In the case of the mesomeric effect, compound 2 should have
an electron-deficient sulfur atom covalently bonded to the
selenium. In contrast to this assumption, the calculations show
that the sulfur atom in 2 carries less positive charge (0.029)
than the sulfur in PhSSePh (0.089), indicating that the weak
Se‚‚‚O interactions play a major role in decreasing the electron
density around the selenium atom. Although Se‚‚‚O interactions
favor the nucleophilic attack of thiol at selenium in the selenenic
acid state, these interactions are detrimental to the reactivity of
selenenyl sufides as such interactions would lead to thiol
exchange reactions rather than attack at sulfur to regenerate the
selenol. The disadvantage of having a coordinating group near
the selenium in the selenenyl sulfide state was also confirmed
by experimental 77Se NMR spectroscopy. The addition of an
equimolar amount of PhSH to ebselen readily produced the
expected selenenyl sulfide 2 (588 ppm), but the addition of an
excess of PhSH did not produce the expected selenol 3.
Similarly, the addition of N-acetyl-L-cysteine to ebselen did not
produce any signal for the selenol, but it produced the
corresponding selenenyl sulfide 8 (565 ppm) as a stable product.
r(Se
δ(
77Se NMR)a
r(Se
‚‚‚O)/(S/Se‚‚‚N)
δ(
77Se NMR)a
‚‚‚O)
/
(S
/
Se‚‚‚N)
compd
(Å)
(calcd) (ppm)
compd
(Å)
(calcd) (ppm)
2
2.470
2.608
2.576
2.686
604
547
525
405
19
20
21
2.751
2.630
2.850
405
500, 331
460
13
17
18
a Referenced to the peak for Me2Se.
To support the observations that a strong Se‚‚‚O/N interaction
increases the electrophilic reactivity of selenium, we carried out
experiments on diselenide 11, which has been shown to be an
inactive catalyst in the reduction of H2O2 by PhSH.21 When
diselenide 11 was treated with an excess amount of N-acetyl-
L-cysteine in CDCl3, the 77Se NMR signal due to the diselenide
disappeared completely to produce a new signal at 564 ppm
for the corresponding selenenyl sulfide 12. As in the case of
ebselen, no signal was detected for the corresponding selenol
even with a large excess of N-acetyl-L-cysteine. On the other
hand, the addition of PhSH to the reaction mixture produced a
signal at 574 ppm for another selenenyl sulfide (13) as a result
of a thiol exchange reaction. The calculations on selenol 14
and selenenic acid 15 are also in close agreement with those of
selenol 3 and selenenic acid 4 derived from ebselen (Table 4).
Interestingly, the 77Se NMR chemical shift (-48 ppm) calculated
for selenolate 5 having a basic amino group is in close agreement
with the experimental value (6 ppm),21 but it is very upfield
shifted as compared with those of 3 and 14. This indicates that
selenol 5 can react with H2O2 much faster than 3 and 14, which
is consistent with the experimental observations. Tomoda et al.
have shown, in a number of studies, that NBO second-order
perturbation analysis can be conveniently used for studying
hypervalent Se‚‚‚N/O/X (X ) F, Cl, Br) interactions.10,13
Therefore, we carried out NBO analyses22 on a number of
selenols, which show that the selenols can also be stabilized by
hypervalent Se‚‚‚N/O interactions in the gas phase, although
the stabilization energies (ESe‚‚‚N) due to these interactions are
relatively small (Tables 2 and 4). However, the calculations
performed by including the effect of water show that selenol 5
exists in a completely dissociated form, whereas the selenols
having amide and oxazoline substituents (3 and 14) exist as
undissociated selenols. As can be seen from Table 3, the 77Se
NMR chemical shift for compound 17 (525 ppm) is very upfield
shifted compared with that of 2 (602 ppm), suggesting that a
nucleophilic attack of thiol at the selenium center in 17 is less
favored as compared with the corresponding attack in 2. This
can also be rationalized by comparing the positive charge on
the selenium atom in different selenenyl sulfides. According to
these calculations, the selenenyl sulfides 2 (0.377) and 13 (0.352)
carry a much higher positive charge on the selenium atom than
the selenenyl sulfide 17 (0.289). This also suggests that a
nucleophilic attack of thiol at the selenium center in 17 is less
favored as compared to a similar attack at the selenium center
in 2 or 13. Therefore, the higher GPx activity of 5 as compared
with that of 3 or 14 must be ascribed not only to the presence
The results obtained by HPLC experiments are in agreement
with those of 77Se NMR experiments and theoretical calcula-
tions. Addition of an excess amount of PhSH to ebselen pro-
duced only the selenenyl sulfide as a stable product with a reten-
tion time of 6.9 min. Similarly, the reaction of ebselen with
N-acetyl-L-cysteine also afforded the corresponding selenenyl
sulfide as the only product (retention time 2.0 min) (see Figure
S11 in the Supporting Information). These observations indicate
that the selenenyl sulfides formed during the reactions are very
stable and inactive toward further reaction with thiol or the ex-
cess thiol attacks at the selenium atom, leading to thiol exchange
reactions. A crossover experiment was carried out using ebselen
to confirm the possibility of a thiol exchange reaction. When
ebselen was treated with p-thiocresol, it produced the expected
selenenyl sulfide 9 (600 ppm); however, the addition of an
excess amount of ethanethiol failed to produce the selenol even
strictly under an inert atmosphere. On the other hand, it produced
a signal at 532 ppm for a new selenenyl sulfide (10). These
observations strongly suggest that the strong Se‚‚‚O interactions
in the selenenyl sulfide state are detrimental to the biological
activity of ebselen and related derivatives. The HPLC experi-
ments also show that the addition of PhSH to selenenyl sulfide
8 produces compound 2 and the addition of N-acetyl-L-cysteine
to 2 generates compound 8 by a thiol exchange reaction.
(21) Mugesh, G.; Panda, A.; Singh, H. B.; Punekar, N. S.; Butcher, R. J. J. Am.
Chem. Soc. 2001, 123, 839.
(22) (a) Reed, A. E.; Curtiss, L. A.; Weinhold, F. Chem. ReV. 1988, 88, 899.
(b) Glendening, E. D.; Reed, J. E.; Carpenter, J. E.; Weinhold, F. Natural
Bond Orbital (NBO) Version 3.1.
9
11480 J. AM. CHEM. SOC. VOL. 127, NO. 32, 2005

GPx-like Antioxidant Activity of Ebselen
Table 4. Summary of Quantum Chemical Calculations, 77Se NMR Chemical Shifts, and NBO Analysis on 5, 7, 13-17, 19, and 26
θN/O ESe/S
‚‚‚N/O
A R T I C L E S
rSe/S
δ(
77Se NMR)a
δ(
77Se NMR)
‚‚‚N/O
-Se/S
-X
compd
(Å)
(deg)
qSe
(calcd) (ppm)
(exptl) (ppm)
(kcal/mol)
5
7
2.829
2.725
2.608
2.759
2.471
2.485
2.576
2.751
2.569
165.9
171.8
177.5
167.0
172.3
172.7
175.2
176.9
177.1
0.144
0.446
0.352
0.219
0.694
0.617
0.289
0.215
0.374
81
253
547
6b
253c
574
5.19
6.13
13.45
6.35
23.01
21.64
15.49
5.46
13
14
15
16
17
19
26
197
270b
1206b
1164b
562b
455
1021
1052
525
405
632
622d
14.35
a Referenced to the peak for Me2Se. b Reference 21. c Reference 27. d Reference 13g.
of 5 in its zwitterion form but also to the ability of selenenyl
sulfide 17 to regenerate the selenol 5 in the catalytic cycle.
It has been shown that the conversion of selenenyl sulfide to
selenol could be achieved by increasing the thiol concentration.21
To gain further insight into the effect of thiol concentration in
the GPx-like reaction of ebselen, a detailed kinetic study was
undertaken by using an HPLC assay. The initial rates derived
for various concentrations of PhSH were plotted against the
concentration of thiol. In contrast to the GPx catalytic cycle,
saturation kinetics was not observed even at a very high
concentration of PhSH and also the expected linear increase in
the rate was not observed at these concentrations. Alternatively,
a polynomial increase in the rate was observed with an increase
in the concentration of PhSH. The rates were calculated to be
lower than the control rates at lower concentrations of thiol as
shown in Table 1, but the rate of the reaction gradually increased
over the control rates at higher concentrations (Figure 2). This
is in agreement with the assumption that the thiol exchange
reactions could be overcome, at least partially, by increasing
the thiol concentration.
However, a closer look at the chromatograms of the reaction
of ebselen with various concentrations of PhSH shows that the
concentration of selenenyl sulfide 2 (retention time 6.3 min)
with 9 mM PhSH (Figure 3A) is almost identical to that with
24 mM PhSH (Figure 3B). In fact, the selenenyl sulfide was
found to be the major species during the entire concentration
range, and the peak due to this compound was unaffected even
after 300 min. These observations indicate that only a minor
amount of selenenyl sulfide 2 could be converted to the
corresponding selenol by increasing the concentration of thiol.
Although the poor catalytic activity of ebselen can be
attributed both to the low reactivity of the selenol toward
oxidation and to the thiol exchange reactions at selenium, further
studies are required to give a firm conclusion regarding the rate-
determining step. For this purpose, we prepared the selenol and
selenenyl sulfide independently and compared their reactivity
with peroxide and thiol to understand the origin of the poor
catalytic activity. The reactions were followed by 77Se NMR
spectroscopy. The signal due to the selenenyl sulfide 2 was
unaffected by the addition of PhSH, whereas the signal due to
the selenol 3 completely disappeared upon the addition of
peroxide to produce a new signal at 1143 ppm. This strongly
suggests that the poor catalytic activity of ebselen is due to the
thiol exchange reaction at selenium as opposed to sulfur in the
selenenyl sulfide. Is there any other way to avoid the thiol
exchange reactions and increase the GPx activity of organose-
lenium compounds? As already mentioned, the reduction of a
selenenyl sulfide to selenol in the GPx cycle requires surmount-
Figure 2. Initial rate (V0) plotted against the substrate concentration. The
initial rates were calculated from at least six HPLC experiments for each
concentration of PhSH. The concentrations of ebselen and H2O2 were fixed
at 0.47 and 9.8 mM, respectively.
Figure 3. (A) Reversed-phase HPLC chromatogram (305 nm) of ebselen (0.5 mM) with PhSH (9.0 mM) and H2O2 (9.8 mM). The chromatographic runs
were carried out at (a) 300 min, (b) 240 min, (c) 200 min, (d) 160 min, (e) 120 min, and (f) 0 min. (B) Reversed-phase HPLC chromatogram of ebselen (0.5
mM) with PhSH (24.0 mM) and H2O2 (9.8 mM). The chromatographic runs were carried out at (a) 0 min, (b) 120 min, (c) 160 min, (d) 200 min, (e) 240
min, and (f) 300 min.
9
J. AM. CHEM. SOC. VOL. 127, NO. 32, 2005 11481

A R T I C L E S
Sarma and Mugesh
ing a substantial barrier (∼50 kcal/mol).11 A careful analysis
of the active site features of GPx reveals that the sulfur atom in
the selenenyl sulfide intermediate is involved in a weak inter-
action with the amido nitrogen of the Thr54 residue, lowering
the energy barrier to increase the possibility of a nucleophilic
attack of negatively charged thiolate at the sulfur atom in the
-Se-S- bridge.23,24 Another selenoenzyme, thioredoxin re-
ductase (TrxR), having proximal histidines25 may also use such
a strategy to overcome the thiol exchange reactions. To probe
the role of Thr54 in the GPx cycle, we carried out experimental
as well as theoretical investigations on some model compounds
which have internal chelating groups near the sulfur. We carried
out investigations on three different types of selenenyl sulfides
having (i) only a Se‚‚‚O/N interaction (selenenyl sulfide 2), (ii)
only a S‚‚‚O/N (selenenyl sulfide 19), and (iii) both Se‚‚‚O/N
and S‚‚‚O/N interactions (selenenyl sulfide 24).
formation of the expected selenide. The HPLC experiments also
show that the selenenyl sulfide 19 is stable enough to react with
23, producing the disulfide and PhSeSePh. These studies clearly
show that the thiol readily attacks at the sulfur center in 19 to
produce the corresponding disulfide and benzeneselenol and the
instability of the resulting selenol leads to the oxidation of this
compound to the corresponding diselenide.
The experimental 77Se NMR chemical shift (455 ppm) for
selenenyl sulfide 19 is very downfield shifted compared to those
for 13 (575 ppm) and PhSeSPh (526 ppm). This clearly indicates
a possibility that the S‚‚‚N interaction decreases the electro-
positive character of the selenium atom in 19. The calculations
show that the nitrogen atom in 19 interacts with the sulfur (S‚
‚‚N ) 2.751 Å), leading to an elongation of the -S-Se- bond
(Se-S ) 2.238 Å). The NBO second-order perturbation energy
for the S‚‚‚N interaction (ES‚‚‚N) in 19 is calculated to be 5.46
kcal/mol. The calculated 77Se chemical shift for 19 (405 ppm)
also shows a large upfield shift as compared with that of 13
(547.0 ppm), confirming the expected decrease in the electro-
philic character of selenium. The experimental 77Se NMR results
are in agreement with the theoretical results, and the upfield
signal observed for an authentic sample of 19 (455 ppm) also
suggests that the selenium nucleus in 19 is considerably shielded
due to the S‚‚‚N interactions even in solution. The NBO analysis
shows a decrease in the positive charge on selenium and an
increase in the positive charge on sulfur on moving from 13
[0.352 (Se), 0.033 (S)] to 19 [0.215 (Se), 0.152 (S)]. This indi-
cates that the nonbonded interactions with sulfur in the selenenyl
sulfide intermediate not only increase the electropositive
character of sulfur but also decrease the electrophilic character
of selenium. In other words, the S‚‚‚N interactions would
certainly enhance the possibility of the thiol attack at sulfur
rather than at selenium. The increase in the electron density
around selenium or the decrease in the electron density around
sulfur in 19 by S‚‚‚N interactions was further verified by
replacing the sulfur atom in 19 with selenium (compound 20).
In this case, the calculated 77Se chemical shifts show that the
selenium atom located near the amino group is more deshielded
(500 ppm) as compared with the other selenium atom (331 ppm).
As previously shown, the Se‚‚‚O interaction in 2 does not
favor a nucleophilic attack at sulfur, but it favors a nucleophilic
attack at selenium. The reaction of diphenyl diselenide with thiol
23 having a heteroatom in close proximity to sulfur produced
selenenyl sulfide 19. Interestingly, addition of an excess amount
of thiol completely reduced the selenenyl sulfide to the
corresponding selenol, which was unstable and oxidized in
solution to produce diphenyl diselenide. Similarly, when pure
selenenyl sulfide 19 was treated with 1 equiv of thiol 23, it
produced a signal at 460 ppm for diphenyl diselenide. To
confirm that the formation of diphenyl diselenide is not due to
the disproportionation reactions, which are normally observed
with selenenyl sulfides, the reaction of 19 with thiol 23 was
monitored by 77Se NMR spectroscopy and reversed-phase
HPLC. When the reaction of 19 with 23 was carried out strictly
under an inert atmosphere, it produced the expected selenol,
which was trapped by treating the reaction mixture with
iodoacetic acid. The 77Se NMR spectrum recorded in CDCl3/
MeOH showed a sharp signal at 203 ppm, indicating the
Finally, we focused our attention on selenenyl sulfide 24,
which has both Se‚‚‚O/N and S‚‚‚O/N interactions. As expected,
the reaction of ebselen with thiol 23 produced the corresponding
selenenyl sulfide 24. This compound exhibited a 77Se NMR
signal at 547 ppm, which is significantly upfield shifted
compared with that of 2 (588 ppm). Crucially, the addition of
an excess amount of 23 to selenenyl sulfide 24 readily produced
the corresponding selenol, which is unstable and upon oxidation
afforded the diselenide 25 (440 ppm) in nearly quantitative yield.
These observations strongly suggest that the introduction of
coordinating amino or other groups in the thiols would enhance
the GPx-like catalytic activity of ebselen and other related
organoselenium compounds (Figure 4). Although ebselen ef-
fectively reduces H2O2 in the presence of thiol 23, all attempts
to obtain the reaction rate were unsuccessful. The reaction of
24 with thiol 23 produced the desired selenol 3 and disulfide
22, but unfortunately, the disulfide, in contrast to PhSSPh,
reacted with H2O2 to produce the corresponding sulfenic acid
or other oxidized products. The cleavage of the -S-S- bond
in 22 by H2O2 may be due to the presence of strong S‚‚‚N
interactions (S‚‚‚N ) 2.815 Å, S‚‚‚N ) 2.787 Å),26 which
(23) Mugesh, G.; du Mont, W.-W. Chem.sEur. J. 2001, 7, 1365.
(24) Aumann, A. D.; B rf, N.; Brigelius-Flohe´, R.; Schomburg, D.; Flohe´, L.
Biomed. EnViron. Sci. 1997, 10, 136.
(25) Sandalova, T.; Zhong, L.; Lindqvist, Y.; Holmgren, A.; Schneider, G. Proc.
Natl. Acad. Sci. U.S.A. 2001, 98, 9533.
9
11482 J. AM. CHEM. SOC. VOL. 127, NO. 32, 2005

GPx-like Antioxidant Activity of Ebselen
A R T I C L E S
Figure 4. S‚‚‚N interactions modulate the attack of an incoming thiol at
the sulfur center in ebselen selenenyl sulfide 24.
activates the -S-S- bond. Because of this unexpected behavior
of 22, the initial rate could not be determined by HPLC or UV
spectroscopic methods.
Having established the effect of novel S‚‚‚N/O interactions
on the reactivity of selenenyl sulfides, we next set out to
investigate the correlation between S‚‚‚N/O distances and the
strength of Se‚‚‚N/O interactions. For this purpose, we carried
out DFT calculations on 26 and 27, which are simplified models
for 2 and 24, respectively. The Se‚‚‚O distance in 27 (2.631 Å)
is found to be longer than that in 26 (2.569 Å). NBO analysis
shows that the orbital interaction energy ESe‚‚‚O in 27 (11.43
kcal/mol) is lower than that in 26 (14.35 kcal/mol) due to the
presence of S‚‚‚N interactions (ES‚‚‚N ) 4.78 kcal/mol). As the
decrease in interaction energy would lead to a decrease in the
electrophilic character of selenium, the calculated 77Se NMR
chemical shift for 27 is very upfield shifted (586 ppm) as
compared with that of 26 (632 ppm). A comparison of the
experimental 77Se NMR chemical shift between 2 (588 ppm)
and 24 (547 ppm) also supports that S‚‚‚N interactions are
indeed present in 24, which increases the possibility of a
nucleophilic attack at the sulfur center of the -Se-S- bridge.
The energy profile obtained for 27 by varying the dihedral angle
γ(1,2,3,4) from -180° to +180° stepwise by 10° is shown in
Figure 5. In the rotation profile, the energy minima were found
at -160° (A), 0° (C), and +160° (E) and the maxima were
found at -90° (B) and +80° (D). The S‚‚‚N distance was found
to be the shortest (2.786 Å) at the point where the most stable
minimum (C) occurs, and this distance is in good agreement
with the optimized structure (Figure 5). The other two minima
at -160° (A) and +160° (E) correspond to conformations where
the nitrogen atom of the oxazoline ring is completely away from
the sulfur atom. At this stage, the structures are stabilized by
interaction between sulfur and the oxygen atom present in the
five-membered ring. The S‚‚‚O distances (2.623 Å at γ ) +160°
and 2.623 Å at γ ) -160°) indicate that the S‚‚‚O interactions
also lead to some stabilization of the molecule although these
conformations are less stable than those obtained by S‚‚‚N
interactions (relative energy for 27, 0.0, 2.77, and 2.58 kcal/
mol at γ ) 0°, +160°, and -160°, respectively). The conforma-
tions become unstable when the oxazoline ring lies perpendicular
to the benzene ring. In these conformations, there is no S‚‚‚N
or S‚‚‚O interaction as the nitrogen and oxygen atoms lie either
above or below the plane of the benzene ring.
Figure 5. Variation of the relative stabilization energy of 27 for the rotation
of the oxazoline ring around the C-C bond attached to the benzene ring.
The calculations were performed at the B3LYP/6-31G* level of theory.
Figure 6. Theoretical versus experimental 77Se NMR chemical shifts (ppm).
The data were taken from Tables 2 and 4.
X-ray analysis.29 The calculated GIAO 77Se NMR chemical
shifts for most of the compounds in the present study also
correlate well with the experimental data. A plot of theoretical
77Se NMR chemical shifts versus the experimental shifts shows
a linear correlation as shown by Bayse,30 demonstrating the
accuracy of the calculations (Figure 6).
Conclusions
The experimental and theoretical observations on ebselen and
related derivatives suggest that strong Se‚‚‚O/N interactions in
the selenenyl sulfide states are detrimental to the biological
activity of synthetic selenium compounds due to thiol exchange
The theoretical data28 obtained under the present study are
in good agreement with the experimental results. For example,
bond lengths and angles obtained by theoretical calculations for
13 are consistent with the data obtained from single-crystal
(27) Fischer, H.; Terlinden, R.; Lo¨hr, J. P.; Romer, A. Xenobiotica 1988, 18,
1347.
(28) For details, see the Supporting Information.
(29) Mugesh, G.; du Mont, W.-W.; Wismach, C.; Jones, P. G. ChemBioChem
2002, 3, 440.
(26) Mugesh, G.; Singh, H. B.; Butcher, R. J. Eur. J. Inorg. Chem. 1999, 1229.
(30) Bayse, C. A. Inorg. Chem. 2004, 43, 1208.
9
J. AM. CHEM. SOC. VOL. 127, NO. 32, 2005 11483

A R T I C L E S
Sarma and Mugesh
precipitate was dissolved in dry ether (30 mL), the solution was cooled
to 0 °C, and sulfur powder (0.32 g, 10 mmol) was added. After 1 h all
sulfur was consumed to give a yellow solution. This solution was poured
into a beaker containing K3Fe(CN)6 and stirred for 2 days by adding
CH2Cl2 from time to time. The resulting solution was extracted with
CH2Cl2, dried over anhydrous sodium sulfate, and concentrated to give
a yellow oil. Recrystallization of the product in chloroform/methanol
reactions. The relatively poor GPx-like catalytic activity of
ebselen is, therefore, due to the undesired thiol exchange
reactions that take place at the selenium center in the selenenyl
sulfide intermediate. This study provides the first experimental
evidence that any substituent that is capable of enhancing the
nucleophilic attack of thiol at sulfur in the selenenyl sulfide
state would enhance the antioxidant potency of organoselenium
compounds. These results lead to an assumption that some of
the glutathione peroxidases utilize non-GSH cosubstrates prob-
ably to overcome thiol exchange reactions.
1
(1:1) afforded the disulfide as white needles: yield 0.8 g, (50%); H
NMR (CDCl3) δ 7.78 (d, 1H), 7.72 (d, 1H), 7.32 (t, 1H), 7.20 (t, 1H),
4.12 (s, 2H), 1.44 (s, 6H); 13C NMR (CDCl3) δ 160.7, 138.6, 131.2,
129.8, 126.2, 125.9, 125.4, 78.9, 68.8, 28.6; MS (TOF MS ES+) m/z
435 [M + Na]+.
Experimental Section
Synthesis of 23. The thiol 23 was prepared by reducing the
corresponding disulfide with NaBH4 at 0 °C under a N2 atmosphere.
To a cooled solution (0 °C) of the disulfide 22 in MeOH was added
NaBH4 (10 equiv). After the resulting solution was stirred for 1 h at 0
°C, the organic layer was extracted with CH2Cl2, dried over anhydrous
sodium sulfate, and concentrated to give a green-yellow oil. This
compound was essentially pure and was used without further purifica-
tion. This compound is stable under an inert atmosphere, but oxidizes
slowly under air to produce the corresponding disulfide.
Synthesis of 24. To a stirred solution of 1 (21 mg, 0.076 mmol) in
CH2Cl2 was added the thiol 23 (15.75 mg, 0.076 mmol). After resulting
solution was stirred for 1 h at room temperature, the solvent was
evaporated under reduced pressure, and the product obtained was
washed with petroleum ether to remove any unreacted thiol or disulfide
formed during the reaction. The yellow oil of 24 was obtained in good
yield (90%): 1H NMR (CDCl3) δ 8.05 (d, 1H), 7.99 (s, 1H), 7.72 (d,
2H), 7.65 (t, 4H), 7.41 (q, 4H), 7.16-7.24 (m, 2H), 4.16 (s, 2H), 1.47
(s, 6H); 13C (CDCl3) δ 166.2, 161.4, 137.4, 137.3, 137.1, 132.5, 131.7,
131.1, 129.9, 129.4, 129.2, 127.0, 126.9, 126.7, 126.2, 125.5, 125.1,
120.7, 78.9, 68.7, 28.6; 77Se (CDCl3) δ 547; MS (TOF MS ES+) m/z
477 [M - 5]+.
HPLC Assay. In this assay, we employed a mixture containing a
2:1 molar ratio of PhSH or 4-MeC6H4SH and H2O2 in dichloromethane/
methanol (95:5) at room temperature as our model system. Runs with
and without 10 mol % added ebselen were carried out under the same
conditions. Periodically, aliquots were removed, and the concentrations
of the product diphenyl disulfide (PhSSPh) were determined by
reversed-phase HPLC, using pure PhSSPh as an external standard. The
catalytic activity of ebselen with 4-MeC6H4SH was studied by following
a similar method using the corresponding disulfide as an external
standard. The amount of disulfide formed during the course of the
reaction was calculated from the calibration plot for each standard.
All reactions were carried out under nitrogen or argon using standard
vacuum-line techniques. Solvents were purified by standard procedures
1
and were freshly distilled prior to use. H (400 MHz), 13C (100.56
MHz), and 77Se (76.29 MHz) NMR spectra were obtained on a Bruker
400 MHz NMR spectrometer. Chemical shifts are cited with respect
to SiMe4 as internal (1H and 13C) and Me2Se as external (77Se) standards.
Mass spectral studies were carried out on a Q-TOF micro mass
spectrometer with ESI MS mode analysis. In the case of isotopic
patterns, the value given is for the most intense peak. Ebselen was
synthesized by following the literature method.31
Synthesis of 2. To a stirred solution of ebselen (1) (21 mg, 0.076
mmol) in CH2Cl2 was added benzenethiol (8 µL, 0.076 mmol). After
the resulting solution was stirred for 1 h at room temperature, the solvent
was evaporated under reduced pressure, and the product obtained was
washed with petroleum ether to remove any unreacted thiol or disulfide
formed during the reaction. The selenenyl sulfide 2 was obtained as a
light yellow oil in quantitative yield: 1H NMR (CDCl3) δ 8.23 (d,
1H), 7.94 (s, 1H), 7.71 (d, 1H) 7.62 (d, 2H) 7.51 (d, 2H) 7.32-7.41
(m, 3H) 7.13-7.28 (m, 5H); 13C (CDCl3) δ 166.1, 137.7, 137.2, 136.5,
132.5, 131.2, 129.4, 129.3, 129.1, 128.9, 126.5, 126.2, 125.7, 125.2,
120.7; 77Se (CDCl3) δ 588; HRMS (TOF MS ES+) m/z 407.9937 (M
+ Na)+.
Synthesis of 9. To a stirred solution of 1 (21 mg, 0.076 mmol) in
CH2Cl2 was added 4-methylbenzenethiol (9.42 mg, 0.076 mmol). After
the resulting solution was stirred for 1 h at room temperature, the solvent
was evaporated under reduced pressure, and the product obtained was
washed with petroleum ether to remove any unreacted thiol or disulfide
formed during the reaction. The selenenyl sulfide 9 was obtained as a
light yellow oil in quantitative yield: 1H NMR (CDCl3) δ 8.27 (d,
1H), 7.93 (s, 1H), 7.69 (d, 1H) 7.61 (d, 2H) 7.50 (t, 1H) 7.30-7.42
(m, 5H) 7.18 (t, 1H) 7.04 (d, 2H) 2.29 (s, 3H); 13C (CDCl3) δ 166.1,
137.8, 137.2, 136.8, 133.1, 132.4, 131.3, 129.7, 129.5, 129.3, 129.0,
126.6, 126.2, 125.2, 120.6, 20.2; 77Se (CDCl3) δ 600; HRMS (TOF
MS ES+) m/z 422.0094 (M + Na)+.
Synthesis of 19. To a stirred solution of diphenyl diselenide (150.72
mg, 0.48 mmol) in CH2Cl2 was added the thiol 23 (100 mg, 0.48 mmol).
After the resulting solution was stirred for 1 day at room temperature,
the solvent was evaporated under reduced pressure to give a yellow
residue. This was purified by column chromatography using petroleum
ether/ethyl acetate (5:1) as eluent to give compound 19 as a yellow oil
in 30% yield: 1H NMR (CDCl3) δ 7.84 (d, 1H), 7.74 (d, 1H), 7.56 (d,
2H) 7.34 (t,1H) 7.20-7.29 (m, 4H), 4.11 (s, 2H), 1.44(s, 6H); 13C
(CDCl3) δ 160.9, 137.8, 132.0, 131.1, 129.8, 129.5, 129.3, 129.2, 127.4,
126.7, 125.8, 78.9, 68.7, 28.6; 77Se (CDCl3) δ 455; HRMS (TOF MS
ES+) m/z 364.0274 (M + H)+.
Computational Methods
All calculations were performed using the Gaussian98 suite of
quantum chemical programs.32 The hybrid Becke 3-Lee-Yang-Parr
(B3LYP) exchange correlation functional was applied for DFT calcula-
tions.33 Geometries were fully optimized at the B3LYP level of theory
using the 6-31G(d) basis sets. All stationary points were characterized
as minima by corresponding Hessian indices. The NMR calculations
were done at the B3LYP/6-311+G(d,p) level on B3LYP/6-31G(d)-
level-optimized geometries using the GIAO method.34 Orbital interac-
tions were analyzed using the natural bond orbital (NBO) method at
the B3LYP/6-31G(d) level, and charges were calculated from natural
population analysis (NPA).22
Synthesis of 22. The disulfide 22 was synthesized by following the
literature method with some modifications.26 To a solution of 4,4-
dimethyl-2-phenyloxazoline (1.75 mL, 1.79 g, 10 mmol) in dry hexane
(50 mL) was added a 1.6 M solution of n-BuLi in hexane (6.8 mL, 11
mmol). After the resulting solution was stirred for 1 h at room
temperature, a white precipitate of lithiated compound was obtained.
The supernatant solvent was removed with a syringe. The white
Acknowledgment. This study was supported by the Depart-
ment of Science and Technology (DST), New Delhi, India. We
express appreciation to Prof. Michio Iwaoka, Prof. Waroˆ
(32) Gaussian98; Gaussian, Inc.: Pittsburgh, PA, 1998. The full reference is
given in the Supporting Information.
(33) (a) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785. (b) Becke,
A. D. J. Chem. Phys. 1993, 98, 5648.
(34) Nakanishi, W.; Hayashi, S. J. Phys. Chem. A 1999, 103, 6074 and references
therein.
(31) Engman, L.; Hallberg, A. J. Org. Chem. 1989, 54, 2964.
9
11484 J. AM. CHEM. SOC. VOL. 127, NO. 32, 2005

GPx-like Antioxidant Activity of Ebselen
A R T I C L E S
Nakanishi, Dr. Mao Minoura, Dr. Craig A. Bayse, and Dr. R.
B. Sunoj for helpful discussions.
and calculated GIAO chemical shifts (PDF). This material is
available free of charge via the Internet at http://pubs.acs.org.
Supporting Information Available: Experimental data (77Se
NMR spectra) and archive entries for the optimized geometries
JA052794T
9
J. AM. CHEM. SOC. VOL. 127, NO. 32, 2005 11485