Structure-reactivity relationship of piperidine nitroxide: Electrochemical, ESR and computational studies

Article abstract of DOI:10.1021/jo101961m

We have synthesized several nitroxides with different substituents which vary the steric and electronic environment around the N-O moiety and have systematically investigated the role of substituents on the stability of the radicals. Our results demonstrated the reactivity toward ascorbate correlates with the redox potential of the derivatives. Furthermore, ab initio calculations also indicated a correlation between the reduction rate and the computed singly occupied molecular orbital-lowest unoccupied molecular orbital energy gap, but not with solvent accessible surface area of the N-O moiety, supporting the experimental results and suggesting that the electronic factors largely determine the radicals' stability. Hence, it is possible to perform virtual screening of nitroxides to optimize their stability, which can help to rationally design novel nitroxides for their potential use in vivo.

Full text of DOI:10.1021/jo101961m

pubs.acs.org/joc
Structure-Reactivity Relationship of Piperidine Nitroxide:
Electrochemical, ESR and Computational Studies
Toshihide Yamasaki,† Fumiya Mito,† Yuko Ito,† Sokkar Pandian,‡ Yuichi Kinoshita,†
Koji Nakano,§ Ramachandran Murugesan,‡,^ Kiyoshi Sakai,z Hideo Utsumi,z and
Ken-ichi Yamada*,†
†Department of Bio-functional Science, Faculty of Pharmaceutical Sciences, Kyushu University, 3-1-1
Higashi-ku, Maidashi, Fukuoka 812-8582, Japan, ‡School of Chemistry, Madurai Kamaraj University,
Madurai-625021, India, §Department of Applied Chemistry, Graduate School of Engineering, Kyushu
University, 7-4-4 Moto-oka, Nishi-ku, Fukuoka 819-0395, Japan, ^Networking Resource Center in
Biological Sciences, Madurai Kamaraj University, Madurai-625021, India, and zInnovation Center for
Medical Redox Navigation, Kyushu University, 3-1-1 Higashi-ku, Maidashi, Fukuoka 812-8582, Japan
yamada@pch.phar.kyushu-u.ac.jp
Received October 5, 2010
We have synthesized several nitroxides with different substituents which vary the steric and electronic
environment around the N-O moiety and have systematically investigated the role of substituents on
the stability of the radicals. Our results demonstrated the reactivity toward ascorbate correlates with
the redox potential of the derivatives. Furthermore, ab initio calculations also indicated a correlation
between the reduction rate and the computed singly occupied molecular orbital-lowest unoccupied
molecular orbital energy gap, but not with solvent accessible surface area of the N-O moiety,
supporting the experimental results and suggesting that the electronic factors largely determine the
radicals’ stability. Hence, it is possible to perform virtual screening of nitroxides to optimize their
stability, which can help to rationally design novel nitroxides for their potential use in vivo.
Introduction
agents,7,8 owing to their interesting redox behavior. The
nitroxides are readily oxidized to oxoammonium cations or
reduced to hydroxylamines by various in vivo oxidants or
reductants.9 Upon reaction with superoxide, the nitroxides
undergo one-electron oxidation and subsequent two-electron
reduction, which is driven forward by the redox potential of
Nitroxides are stable free radicals and are widely used as
spin probes,1-3 antioxidants,4-6 and radiation protective
*To whom correspondence should be addressed at Kyushu University.
Phone: þ81-92-642-6624. Fax: þ81-92-642-6626.
(1) Kuppusamy, P.; Li, H.; Ilangovan, G.; Cardounel, A. J.; Zweier, J. L.;
Yamada, K.; Krishna, M. C.; Mitchell, J. B. Cancer Res. 2002, 62, 307.
(2) Matsumoto, K. I.; Hyodo, F.; Matsumoto, A.; Koretsky, A. P.;
Sowers, A. L.; Mitchell, J. B.; Krishna, M. C. Clin. Cancer Res. 2006, 12,
2455.
(6) Xavier, S.; Yamada, K.; Samuni, A. M.; Samuni, A.; DeGraff, W.;
Krishna, M. C.; Mitchell, J. B. Biochim. Biophys. Acta 2002, 1573, 109.
(7) Cotrim, A. P.; Hyodo, F.; Matsumoto, K.; Sowers, A. L.; Cook, J. A.;
Baum, B. J.; Krishna, M. C.; Mitchell, J. B. Clin. Cancer Res. 2007, 13, 4928.
(8) Metz, J. M.; Smith, D.; Mick, R.; Lustig, R.; Mitchell, J.; Cherakuri,
M.; Glatstein, E.; Hahn, S. M. Clin. Cancer Res. 2004, 10, 6411.
(9) Kocherginsky, N.; Swartz, H. M. In Nitroxide Spin Labels: Reactions
in Biology and Chemistry; Kocherginsky, N., Swartz, H. M., Eds.; CRC
Press: Boca Raton, FL, 1995; p 27.
(3) Yamada, K.; Yamamiya, I.; Utsumi, H. Free Radical Biol. Med. 2006,
40, 2040.
(4) Wilcox, C. S. Pharmacol. Ther. 2010, 126, 119.
(5) Utsumi, H.; Yamada, K.; Ichikawa, K.; Sakai, K.; Kinoshita, Y.;
Matsumoto, S.; Nagai, M. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 1463.
DOI: 10.1021/jo101961m
r
Published on Web 12/29/2010
J. Org. Chem. 2011, 76, 435–440 435
2010 American Chemical Society

Yamasaki et al.
SCHEME 1. Synthetic Scheme of Piperidine Nitroxidesa
JOCArticle
FIGURE 1. Nitroxides used in this study.
the redox couple.10 The nitroxides are readily reduced by
ascorbic acid,11 posing a limitation to a probe of their in vivo
redox status due to their shortened lifetimes. Highly stable
nitroxides would be less sensitive to the redox status of their
microenvironment. Both the stability and reactivity of ni-
troxides are equally important for their potential in vivo
application. It has been reported that the reduction rate of
cyclic nitroxides can be altered by the substituents on the
ring.9,12,13 Recently imidazole and isoindoline nitroxides
with ethyl groups proximal to the N-O moiety were re-
ported to be relatively more stable to reduction by
ascorbate.14,15 Although the exact mechanism for this stabil-
ity is not clearly understood, it could be explained by the
steric hindrance the ethyl groups offer the N-O moiety. The
electronic environment around the N-O moiety may also
influence its stability. A detailed study on the structure-
reactivity relationship of the nitroxides is important to
understand the underlying mechanism for stability as well
as to augment their potential use in vivo.
We previously reported that the 4-oxo-2,2,6,6-tetraethyl-
piperidine-N-oxyl radical showed higher resistance to reduc-
tion by ascorbate compared with the methyl substituted
analogue,16 demonstrating the influence of neighboring
substituents. Recently, we developed a synthetic strategy to
derivatize piperidine-N-oxyl radicals at 2 and 6 positions
with a wide variety of substituents including spiro rings.17
Using this method, we have synthesized piperidine-N-oxyl
derivatives with various steric and electronic environments
around the N-O moiety. Herein we report the electrochem-
ical behavior of 2,6-substituted piperidine-N-oxyl radicals
and their reactivity toward ascorbate. We also report our
investigation on the relationship between electrochemical
properties and computed structural properties of nitroxides.
aReagents and conditions: (a) tetrahydro-4H-thiopyran-4-one, NH4Cl,
DMSO; (b) ethylene glycol, PTSA, benzene; (c) Raney-Ni, MeOH;
(d) HCl, acetone; (e) H2O2, Na2WO4 2H2O, MeOH; (f) 1-acetyl-
4-piperidone, NH4Cl, DMSO; (g) tetrahydro-4H-thiopyran-4-one
1,1-dioxide, NH4Cl, DMSO.
3
Results and Discussion
Synthesis of Substituted Piperidine-N-oxyl Radicals. Ni-
troxides used in this study are shown in Figure 1. Nitroxides
3-6,7a, and 8a were synthesized as previously reported.17
Nitroxides 2, 7b, 8b, and 8c were synthesized in a similar
manner as outlined in Scheme 1. Compound 10 was synthe-
sized from 1,2,2,6,6-pentamethylpiperidin-4-one and tetra-
hydro-4H-thiopyran-4-one. The carbonyl group of 10 was
protected with a ketal group and desulfurization was carried
out with Raney-Ni16 to give 12. Subsequent elimination of
the ketal group and the oxidation of 13 yielded 2,2-diethyl-
6,6-dimethylpiperidin-4-one-N-oxyl 2. Compounds 14 and
15 were synthesized with 1-acetylpiperidin-4-one and tetra-
hydro-4H-thiopyrane-4-one 1,1-dioxide, respectively, as
(10) Krishna, M. C.; Grahame, D. A.; Samuni, A.; Mitchell, J. B.; Russo,
A. Proc. Natl. Acad. Sci. U.S.A. 1992, 89, 5537.
(11) Saphier, O.; Silberstein, T.; Shames, A. I.; Likhtenshtein, G. I.;
Maimon, E.; Mankuta, D.; Mazor, M.; Katz, M.; Meyerstein, D.; Meyerstein,
N. Free Radical Res. 2003, 37, 301.
(12) Morris, S.; Sosnovsky, G.; Hui, B.; Huber, C. O.; Rao, N. U.; Swartz,
H. M. J. Pharm. Sci. 1991, 80, 149.
(13) Kirilyuk, I. A.; Bobko, A. A.; Grigor’ev, I. A.; Khramtsov, V. V.
Org. Biomol. Chem. 2004, 2, 1025.
(14) Marx, L.; Chiarelli, R.; Guiberteau, T.; Rassat, A. J. Chem. Soc.,
Perkin Trans. 1 2000, 1181.
(15) Bobko, A. A.; Kirilyuk, I. A.; Grigor’ev, I. A.; Zweier, J. L.;
Khramtsov, V. V. Free Radical Biol. Med. 2007, 42, 404.
(16) Kinoshita, Y.; Yamada, K.; Yamasaki, T.; Sadasue, H.; Sakai, K.;
Utsumi, H. Free Radical Res. 2009, 43, 565.
(17) Sakai, K.; Yamada, K. i.; Yamasaki, T.; Kinoshita, Y.; Mito, F.;
Utsumi, H. Tetrahedron 2010, 66, 2311.
436 J. Org. Chem. Vol. 76, No. 2, 2011

Yamasaki et al.
JOCArticle
TABLE 1. Experimental Redox Potentials of Oxoammonium Cation/
Nitroxide Redox Couplea
TABLE 2. Experimental Redox Potentials of Nitroxide/Deprotonated
Hydroxylamine Redox Couplea
b
c
f
nitroxide
Epa
Epc
E1/2d
ΔEe
ipa/ipc
b
c
f
nitroxide
Epa
Epc
E1/2d
ΔEe
ipa/ipc
1
2
3
4
5
6
7a
7b
8a
8b
8c
0.939
0.926
0.806
0.878
1.013
1.053
1.113
1.019
1.303
1.122
1.116
0.869
0.860
0.747
0.811
0.874
0.940
0.904
0.893
0.776
0.845
0.944
0.997
0.070
0.065
0.058
0.067
0.139
0.113
2.30
1.56
2.64
2.97
3.29
3.06
1
2
3
4
5
6
7a
7b
8a
8b
8c
0.524
0.453
0.226
0.244
0.515
0.644
0.613
0.585
0.653
0.694
0.416
-0.371
-0.401
-0.298
-0.141
-0.364
-0.386
-0.247
-0.381
-0.357
-0.494
-0.427
0.076
0.026
-0.036
0.051
0.076
0.129
0.183
0.102
0.148
0.100
-0.006
0.895
0.854
0.524
0.385
0.879
1.029
0.860
0.967
1.010
1.189
0.844
0.20
0.20
2.12
0.99
0.09
0.19
0.27
0.31
2.18
0.16
0.22
g
-
0.902
0.961
0.117
3.34
g
-
1.012
1.020
1.067
1.068
0.110
0.096
3.51
3.65
aGlassy carbon electrode, Ag/AgCl, Pt, sweep rate: 0.1 V s-1
.
Potentials were shown as vs SHE. bAnodic peak potential. Cathodic
c
aGlassy carbon electrode, Ag/AgCl, Pt, sweep rate: 0.1 V s-1
.
peak potential. dE1/2 = (Epa þ Epc)/2. eΔE = Epa -Epc
.
fThe peak
c
Potentials were shown as vs SHE. bAnodic peak potential. Cathodic
peak potential. dE1/2 = (Epa þ Epc)/2. eΔE = Epa -Epc. The peak
currents (ipa and ipc) were measured from the respective baseline
currents. gThe current of the cathodic peak was too low to determine
the potential value.
f
currents (ipa and ipc) were measured from the respective baseline
currents.
potentials (E1/2) of compounds 1-3 were found to decrease
with the number of ethyl groups (Figure 2), suggesting the
electron-donating substituents stabilize oxoammonium cat-
ions. Compounds 5, 6, 7b, 8b, and 8c showed relatively
larger ΔE and higher peak current ratio, which correspond
to the nitroxides with heteroatoms in the spiro ring. More-
over, the anodic and/or cathodic potentials of compounds 5,
6, 7a, and 8a were higher than those of the parent compound
(4). These results indicate that substituents with heteroatoms
(compounds 5-8) destabilizeoxoammonium cations because
of the electron-withdrawing inductive effect.
A broad and irreversible wave was observed for one-
electron reduction of nitroxide because of the hydroxyl-
amine. The redox potentials for this redox-couple are listed
in Table 2. The E1/2 values of compounds 1-3 decrease pro-
portionally with the number of the ethyl group at the 2 and 6
positions, and the ethyl group had a greater impact on the
anodic potential than the cathodic one. On the other hand,
the E1/2 values of compounds 5, 6, 7a, and 8a were more
positive than that of compound 4. These results suggested
that the electron-withdrawing groups at the 2 and 6 positions
of the piperidine ring destabilize the nitroxides, while elec-
tron-donating substituents stabilize them. In addition, the com-
parison of compounds 7a and 7b (or 8a, 8b, and 8c) showed that
E1/2 values decreased when electron-withdrawing groups were
replaced by the electron-donating groups methyl or ethyl. These
results are also in agreement with the results of compounds 1-3.
Reactivity of 2,6-Substituted Piperidine Nitroxides with
Ascorbate. Nitroxides (25 μM) were allowed to react with
an excess of ascorbate (1 mM) and the rate of radical decay
was monitored by using an X-band ESR spectrometer. The
ESR signal decay rates of the nitroxides are shown in Figure 3.
The nitroxide decay rate follows the order 1>2>3, which is
inversely proportional to the number of ethyl groups. Com-
pound 3, with four ethyl groups, was more esistant to the
reduction than compound 2, with two ethyl groups.
FIGURE 2. Cyclic voltammogram of nitroxides 1-3. Nitroxides
(2 mM) were dissolved in PBS (pH 7.4). Sweep rate: 0.1 V s-1
.
reactants with 9. The products were oxidized with hydrogen
peroxide to obtain nitroxides 7b and 8b, respectively. Com-
pound 17 was synthesized from the desulfurization of com-
pound 16 and oxidized to afford the nitroxide 8c. Compound 16
was refluxed in methanol in the presence of Raney-Ni to
yield mainly compound 17 by partial desulfurization. When
the same reaction was carried out in ethanol, it resulted in
compound 3 by complete desulfurization.
Cyclic Voltammetry. The redox potentials of the ni-
troxides were determined by cyclic voltammetry in phosphate
buffer saline (PBS, pH 7.4), using a glassy carbon electrode.
Compounds 4 and 5 were measured in PBS containing 5%
acetonitrile because of the solubility issues. Nevertheless, the
redox potential of compound 1 did not vary under these two
different conditions. Nitroxides undergo one-electron oxida-
tion corresponding to the formation of N-oxoammonium
cation. The redox potentials of this one-electron redox
couple are listed in Table 1. The redox potentials of com-
pounds 7a and 8a were not estimated because the cathodic
peaks could not be measured. The peak separation, ΔE =
Epa - Epc, of compounds 1-4 was 58-70 mV, close to the
theoretical Nernstian value of 59 mV. However, the intensi-
ties of the anodic and cathodic currents were not equal,
indicating a slight loss of the kinetic reversibility. The redox
The reduction rates of compounds 5, 6, 7a, and 8a with
ascorbate were found to be higher than that of their parent com-
pound 4. This can be explained by the presence of heteroatoms
J. Org. Chem. Vol. 76, No. 2, 2011 437

Yamasaki et al.
JOCArticle
FIGURE 3. ESR signal decay rates of the reaction with ascorbate.
Nitroxides (25 μM) were allowed to react with an excess of ascorbate
(1 mM) and the radical decay rate was monitored with an X-band
ESR spectrometer. The values are the average of three experiments.
Error bars shown are standard deviations.
FIGURE 5. Plot of the SOMO-LUMO energy gap against ESR
signal decay rate of nitroxides in ascorbic acid solution.
is shown in Figure 4a. The ΔEN-A value is the subtraction of
the redox potential of the ascorbic acid (0.058 V vs standard
hydrogen electrode, SHE)18 from that of the nitroxide, which
corresponds to the electromotive force. ESR signal decay
rate increases with an increase in ΔEN-A and it approaches
zero in the negative region of ΔEN-A. Substituents at the
4-position in a six-membered-ring nitroxide or at the 3-posi-
tion in a five-membered-ring nitroxide were reported to
influence the N-O reduction rate by ascorbate.12 However,
there is no report identifying the effect of substituents at 2
and 6 positions of the piperidine ring. Our results demon-
strate a good correlation between the ΔEN-A value and the
ESR signal decay rate of nitroxides differing in substituents
proximate to th N-O moiety at 2 or 6 positions. To further
understand the role of the substituents, the Gibbs free energy
changes (ΔG, calculated from the ΔEN-A by using the
standard expression: ΔG = -nFΔEN-A) were plotted against
the ESR signal decay rate (Figure 4b). The ESR signal decay rate
shows a good correlation with ΔG in the negative ΔG region
(r2 = 0.988). The results indicate that reduction of the nitroxide
by ascorbate occurred spontaneously when the ΔG value was
negative and that the reduction was not spontaneous when the
ΔG value was positive.
Computational Studies. The effect of substituents on the
stability of nitroxide was further validated by computational
methods. The nitroxides were modeled by using ab initio
methods and various descriptors (SOMO and LUMO en-
ergies, geometry energy, and dipole moment) were calcu-
lated. These descriptors were analyzed to correlate with the
stability of the nitroxides to reduction with ascorbate. The
SOMO-LUMO energy gap of nitroxides showed a good
correlation with their stability (Figure 5), the reduction rate
of nitroxides increasing with an increase in the energy gap.
This correlation can be explained by the hard and soft
(Lewis) acids and bases theory;soft acids preferably react
with soft bases and hard acids preferably react with hard
bases.19 The SOMO-LUMO energy gap of the radical cor-
responds to its absolute hardness and the higher the hard-
ness, the higher is its reactivity with ascorbate. Hence, the
nitroxides and ascorbate can be viewed as hard acids and
hard bases, respectively. Moreover, the energy gap of ni-
troxide SOMO-ascorbate HOMO correlates with the ESR
signal decay rate (Figure S15, Supporting Information). The
SOMO energy levels of nitroxides were found to be higher
FIGURE 4. Relationship between the ESR signal decay rates and
ΔEN-A (a) or ΔG (b). (a) ΔEN-A: subtraction of the redox potential of
ascorbate from that of nitroxides. (b) ΔG was calculated from ΔEN-A
with use of standard expression, ΔG = -nFΔEN-A• where n is the
number of electrons per mole of product and F is the Faraday
constant. The correlation coefficient when ΔG was negative was
0.988.
in the spiro ring which act as an electron-withdrawing group and
decrease the electron density around the N-O moiety, thereby
favoring the reduction reaction. The trend of redox potentials for
nitroxide reduction from electrochemical experiments is exactly
the same as that of the nitroxide reduction rate by ascorbate,
which supports our conclusions on the stabilization of nitroxides
by substituents at the 2 and 6 positions.
Relationship between the Redox Potential and the Reactiv-
ity with Ascorbate. The relationship between the redox
potential of the nitroxide and the reactivity with ascorbate
(18) Loach, P. A. In Handbook of Biochemistry and Molecular Biology,
3rd ed.; Fasman, G. D., Ed.; CRC Press: Cleveland, OH, 1976; p 123.
(19) Pearson, R. G. J. Am. Chem. Soc. 1985, 107, 6801.
438 J. Org. Chem. Vol. 76, No. 2, 2011

Yamasaki et al.
JOCArticle
than the HOMO energy level of ascorbate. Ascorbate-
mediated reduction is driven by the energy difference be-
tween nitroxide SOMO and ascorbate HOMO. Lower en-
ergy gaps facilitated the reaction and higher energy gaps
retarded it. Electron-donating (ED) substituents raised the
SOMO energy and made the radical nucleophilic, whereas
electron-withdrawing (EW) substituents decreased the
SOMO energy and increased the electrophilic nature
(Table S2, Supporting Information). Hence ED groups
decrease the reduction rate of nitroxides and EW groups
increase the reduction rate. Nevertheless, it is to be pointed
out that the reaction was found to correlate more closely with
the SOMO-LUMO energy gap of nitroxides than that of
nitroxide SOMO and ascorbate HOMO. This suggests that
there is a possible involvement of nitroxide LUMO in the
reduction reaction. However, the LUMO energy of nitrox-
ides did not directly correlate with the reduction rate. Hence,
it requires further investigation to understand the role of
nitroxide LUMO in the reaction. To further investigate the
role of steric factors in the stability of nitroxides, solvent
accessible surface areas (ASA) of N-O atoms were calculated.
ASA of N-O did not correlate with the stability to reduction,
mixture was adjusted to pH 9 with 10% aq K2CO3, and extracted
with AcOEt (4ꢀ). The AcOEt extract was washed with brine, dried
over anhydrous sodium sulfate, and concentrated in vacuo. The
residue was separated by column chromatography (CHCl3) to
afford 10 as a brown oil (0.85 g, 30%). MS (FABþ) 214.24(Mþ þ 1);
1
ν
max/cm-1 1695 (CdO); H NMR (400 MHz; CDCl3) δ (ppm)
1.228 (s, 6H), 1.767 (ddd, J1 = 12.2 Hz, J2 = 3.2 Hz, 2H), 1.907-
1.950 (m, 2H), 2.229 (s, 2H), 2.281 (s, 2H), 2.394-2.430 (m, 2H),
3.003 (ddd, br, 2H); 13C NMR (75 MHz; CDCl3) δ(ppm) 24.0, 32.5,
40.9, 53.7, 54.8, 55.3, 56.2, 210.2. Anal. Calcd for C11H19NOS: C,
61.93; H, 8.98; N, 6.57. Found: C, 61.56; H, 8.93; N, 6.45.
1,4-Dioxa-10-thia-14,14-dimethyl-13-azaspiro[4.1.5.3]penta-
decane (11). Compound 10 (426 mg, 2 mmol), PTSA (p-toluene-
sulfonic acid) (760 mg, 4 mmol), and ethylene glycol (0.22 mL,
4 mmol) in benzene (6 mL) were added to a round-bottomed
flask equipped with a Dean-Stark apparatus to remove the
water by azeotropic distillation. This solution was heated to
100 ꢀC for 3 h. The resulting mixture was cooled to rt and adjusted
to pH 10 with 10% aq K2CO3. This solution was extracted
with CHCl3, dried with Na2SO4, and evaporated in vacuo. The
residue was purified by column chromatography with CHCl3
as the eluent to give 11 (346 mg, 67%). Mp 86.1-88.6 ꢀC;
MS (FABþ) 258.3 (Mþ þ 1); νmax/cm-1 1056 (CdO); 1H NMR
(400 MHz; CDCl3) δ (ppm) 1.202 (s, 6H), 1.507 (s, 2H), 1.554
(s, 2H), 1.809 (t, br, 2H), 1.99 (m, 2H), 2.44 (m, 2H), 2.928 (t, br,
2H), 3.934 (s, 4H); 13C NMR (75 MHz; CDCl3) δ (ppm) 24.0,
32.5, 40.7, 44.4, 45.9, 51.2, 52.0, 64.0, 109.0. Anal. Calcd for
C13H23NO2S: C, 60.66; H, 9.01; N, 5.44. Found: C, 60.26; H,
8.91; N, 5.34.
2
˚
although the most stable compound 3 has the lowest (8.6 A )
accessible area. The calculations suggest that although steric
hindrance is a significant factor for nitroxides’ stability, electro-
nic factors play a dominant role in the stabilization.
Conclusions
7,7-Diethyl-9,9-dimethyl-1,4-dioxa-8-azaspiro[4.5]decane (12).
A slurry of freshly prepared Raney-Ni in methanol (3 mL) was
added to a well-stirred solution of compound 11 (150 mg, 0.584
mmol) in methanol (5 mL) and the mixture was refluxed at 60 ꢀC.
After 1.5 h, the mixture was filtered through Celite and the filtrates
concentrated in vacuo. The residue was separated by column chro-
matography to give the orange oil 12 (77 mg, 58%). MS (FABþ)
The redox potentials and the reduction rates of piperidine
nitroxides were influenced by the electronic effect of substit-
uents at 2 and 6 positions of the piperidine ring. A good
correlation between the redox potential of the nitroxide and
its reactivity toward ascorbate was demonstrated in this
study. Molecular modeling studies indicated a good correla-
tion between the SOMO-LUMO energy gap of nitroxides
and the reduction rates. Stability and reactivity of a nitroxide
go hand in hand with each other when such a compound is
considered as a potential in vivo spin probe. Different
stabilities and reactivities are desired for specific applications
and hence it is important to optimize these properties for a
given objective. Our electrochemical and molecular model-
ing studies consistently explained the effect of the substituent
on the stability of nitroxides studied. Since there is a correla-
tion between theoretically computed SOMO-LUMO energy gap
and reduction rate, it is now possible to perform virtual screening
of nitroxides, which rationally limits the need to synthesize all of
the possible compounds, thereby increasing the chance of getting
potential hits for better applications in vivo.
1
228.31 (Mþ þ 1); νmax/cm-1 1091 (ketal); H NMR (400 MHz;
CDCl3) δ (ppm) 0.797 (t, J = 7.4 Hz, 6H), 1.202 (s, 6H), 1.329 (m,
2H), 1.502 (s, 2H), 1.551 (s, 2H), 1.703 (m, 2H), 3.931 (s, 4H); 13
C
NMR (75 MHz; CDCl3) δ (ppm) 7.9, 31.3, 32.3, 42.0, 45.9, 51.1,
56.0, 63.9, 109.4; HRMS (FABþ) calcd for C13H26NO2 [M þ H]þ
228.1964, found 228.1999.
2,2-Diethyl-6,6-dimethylpiperidin-4-one (13). HCl aq (1.5 mL,
17.5-18.5%) was added dropwise to a well-stirred solution of
compound 12 (77 mg, 0.339 mmol) in acetone (1 mL) and the
mixture was stirred at rt for 6 h. The reaction mixture was
diluted with H2O, adjusted to pH 10 with K2CO3 aq (10%), and
extracted with CHCl3. The extract was dried over Na2SO4 and
concentrated in vacuo. The residue was separated by column
chromatography with hexane and ethyl acetate as the eluent to
afford compound 13 (23 mg, 37%). MS (FABþ) 184.24 (Mþ þ 1);
1
ν
max/cm-1 1708 (CdO); H NMR (400 MHz; CDCl3) δ (ppm)
0.836 (t, J=7.6Hz,6H),1.224(s,6H),1.384(m,2H),1.511(m,2H),
2.228 (s, 2H), 2.266 (s, 2H); 13C NMR (75 MHz; CDCl3) δ (ppm)
8.0, 32.1, 32.2, 50.6, 54.5, 54.7, 59.9, 211.6. Anal. Calcd for C11H21-
NO: C, 72.08; H, 11.55; N, 7.64. Found: C, 71.86; H, 11.36; N, 7.54.
2,2-Diethyl-6,6-dimethylpiperidin-4-one-1-oxyl (2). Com-
Experimental Section
Reagents and solvents from commercial suppliers were used
without further purification. Silica gel (100-200 mesh) was used
for column chromatography. Nitroxide 1 was purchased from a
commercial supplier and used as received. Nitroxides 3-6,7a,
8a, and 16 were synthesized as previously reported.17
2,2-Dimethyl-9-thia-1-azaspiro[5.5]undecan-4-one (10). NH4Cl-
(4.8 g, 89.7 mmol) was added portionwise to a stirred solution of
1,2,2,6,6-pentamethylpiperidin-4-one 9 (2.25 g, 13.3 mmol) and
tetrahydro-4H-thiopyrane-4-one (5.0 g, 43.1 mmol) in DMSO
(30 mL) at room temperature, and the mixture was heated 14 h
at 60 ꢀC. The reaction mixture was diluted with H2O (40 mL),
followed by acidification with 7% aq HCl (10 mL), and then
extracted with ether (3ꢀ) to remove a neutral fraction. The reaction
pound 13 (20 mg, 0.109 mmol) and Na2WO4 2H2O (18 mg,
3
0.055 mmol) were added to methanol (0.5 mL) and H2O2 (30%,
0.2 mL) was slowly added to the solution with stirring for 24 h at
rt. After being stirred, the solution was saturated with K2CO3
and extracted with chloroform. The chloroform extracts were
dried and evaporated. The residue was separated by column
chromatography (hexane/AcOEt: 1/1) to afford yellow oil 2
(20 mg, 92%). MS (FABþ) 198.26 (Mþ); νmax/cm-1 1713 (CdO).
Anal. Calcd for C11H20NO2: C, 66.63; H, 10.17; N, 7.06. Found:
C, 66.58; H, 10.10; N, 6.98. AN = 1.54 mT.
J. Org. Chem. Vol. 76, No. 2, 2011 439

Yamasaki et al.
JOCArticle
9-Acetyl-2,2-dimethyl-1,9-diazaspiro[5.5]undecan-4-one (14).
NH4Cl (2.68 g, 50.0 mmol) was added portionwise to a stirred
solution of 9 (1.69 g, 10.0 mmol) and 1-acetyl-4-piperidone (1.69 g,
12.0 mmol) in DMSO (8 mL) at room temperature, then the
mixture was heated 4 h at 45 ꢀC. The reaction mixture was
diluted with H2O, followed by acidification with 7% aq HCl,
and then washed with ether (3ꢀ) to remove a neutral fraction.
The reaction mixture was adjusted to pH 9 with 10% aq K2CO3,
then extracted with CHCl3 (4ꢀ). The CHCl3 extract was washed
with brine, dried over anhydrous sodium sulfate, and concen-
trated in vacuo. The residue was separated by column chroma-
tography (CHCl3/MeOH: 99/1) to afford compound 14 as a
product was recrystallized from hexane/AcOEt to give orange
crystals of 8c. Yield 260 mg (77%); mp 116.1-121.2 ꢀC; MS
(FABþ) 289.3 (Mþ þ 1); νmax/cm-1 1715 (CdO), 1292 (SO2).
Anal. Calcd for C13H22NO4S: C, 54.14; H, 7.69; N, 4.86. Found: C,
54.17; H, 7.66; N, 4.80. AN = 1.49 mT.
Cyclic Voltammetry. The electrochemical experiments were
conducted in a sealed, three-electrode glass cell. A platinum wire
and Ag/AgCl electrode were used as counter and reference
electrodes, respectively. A glassy carbon electrode with a disk
diameter of 1 mm was used as the working electrode. The
working electrode was polished with polishing alumina suspen-
sion and a polishing cloth, and rinsed with deionized distilled
water. All electrochemical measurements were performed in
PBS (pH 7.4). Before collecting cyclic voltammograms, the
solution was purged with argon gas for 10 min to minimize
the effect of molecular oxygen and the inert atmosphere was
maintained by a continuous argon flow over the solution during
the experiment. The cyclic voltammograms were collected at
brown oil (238 mg, 10%). MS (FABþ) 239.2 (Mþ þ 1); νmax
/
cm-1 1629 (-N-CO-), 1708 (CdO); 1H NMR (400 MHz;
CDCl3) δ (ppm) 1.188 (s, 3H), 1.222 (s, 3H), 1.496-1.691 (m,
4H), 2.048 (s, 3H), 2.271 (s, 2H), 2.294 (s, 2H), 3.343-3.781 (m,
4H); 13C NMR (75 MHz; CDCl3) δ (ppm) 21.5, 37.6, 40.0, 42.6,
52.4, 55.9, 168.9, 208.7; HRMS (ESIþ) calcd for C13H23N2O2
[M þ H]þ 239.1760, found 239.1741.
ambient temperature at potential sweep rates of 0.1 V s-1
.
9-Acetyl-2,2-dimethyl-1,9-diazaspiro[5.5]undecan-4-one-9-oxyl
(7b). Compound 14 (0.28 g, 1.17 mmol) was oxidized accord-
ing to the method described for compound 2. Yield 40%
(0.12 g); mp 162-163 ꢀC; MS (FABþ) 337.3 (Mþ þ 1);
Rate Constant of Reaction with Ascorbate. A solution of
nitroxides (50 μM) in PBS (pH 7.4) containing 100 μM diethy-
lenetriaminepentaacetic acid (DTPA) was prepared. Sodium
ascorbate solution (2 mM) was prepared in PBS (pH 7.4) with
100 μM DTPA and was mixed in an equal amount of nitroxide
solution. The resulting mixture of nitroxides (25 μM) and ascor-
bate (1 mM) was immediately taken up into a quartz capillary and
then introduced into an ESR tube and transferred into the cavity of
an ESR spectrometer. The ESR spectra were recorded at room
temperature as a function of time. Each experiment was repeated
three times. The loss of signal intensity of the low magnetic field ESR
signal was measured and fitted to exponential decay to compute the
rate constants of each nitroxide.
ν
max/cm-1 1720 (CdO), 1637 (-N-CO-CH3); HRMS (ESIþ)
found 276.1467, calcd for C13H21N2O3 ([M þ Na]þ) 276.1450;
AN = 1.58 mT.
2,2-Dimethyl-9-thia-4,9,9-trioxo-1-azaspiro[5.5]undecane (15).
Tetrahydro-4H-thiopyran-4-one 1,1-dioxide (2.0 g, 13.5 mmol)
was used in place of tetrahydro-4H-thiopyran-4-one according to
the method described for compound 10. Yield 13% (145 mg); mp
182.9 ꢀC (hexane-AcOEt); MS (FABþ) 246.1 (Mþ þ 1); νmax
/
1
cm-1 1698 (CdO), 1291 (SO2); H NMR (300 MHz; CDCl3)
δ (ppm) 1.253 (s, 6H), 2.044-2.229 (m, 4H), 2.282 (s, 2H), 2.326 (s,
2H), 2.811-2.872 (m, 2H), 3.491 (ddd, J1 = 13.2 Hz, J2 = 3.8 Hz,
2H); 13C NMR (75 MHz; CDCl3) δ (ppm) 31.5, 36.9, 47.4, 53.1,
54.9, 55.6, 56.0, 208.6. Anal. Calcd for C11H19NO3S: C, 53.85; H,
7.81; N, 5.71. Found: C, 53.84; H, 7.86; N, 5.75.
ESR Spectra Measurement. ESR experimental conditions
were the following: frequency, 9.4 GHz; power, 10 mW; mag-
netic field, 334 mT; modulation amplitude, one-third of the line
width of each nitroxides; and time constant, 0.03 s.
Molecular Modeling. The General Atomic and Molecular
Electronic Structure System (GAMESS)20 in conjunction with
the Ghemical21 program was used for geometry optimization
and other calculations of all nitroxides. The Ghemical program
was used to generate input structures for GAMESS. Input
structures were energy minimized with AMBER force field22
prior to ab initio calculations. SCF wave function, restricted
open shell Hartree-Fock (ROHF)23 was used with Pople’s
STO-nG Gaussian minimal basis set24 for calculating equilibrium
geometry. Orbital symmetry was not forced during the calculation.
All the calculations were done until the SCF converged. Energy
valuesof SOMO, LUMO, total geometry energy, solvent-accessible
2,2-Dimethyl-9-thia-4,9,9-trioxo-1-azaspiro[5.5]undecane-9-
oxyl (8b). Compound 15 (145 mg, 0.59 mmol) was oxidized
according to the method described for compound 2. The crude
product was recrystallized from AcOEt to give 8b. Yield 153 mg
(99%); mp 157.5-160.8 ꢀC; MS (FABþ) 261.2 (Mþ þ 1); νmax
/
cm-1 1714 (CdO), 1296 (SO2). Anal. Calcd for C11H18NO4S: C,
50.75; H, 6.97 N, 5.38. Found: C, 50.85; H, 7.00; N, 5.26. AN =
1.54 mT.
2,2-Diethyl-9-thia-1-azaspiro[5.5]undecan-4-one (17). A slurry
of freshly prepared Raney-Ni in methanol (6 mL) was added to a
well-stirred solution of compound 1617 (1.1 g, 4.24 mmol) in
methanol (20 mL) and the mixture was refluxed at 75 ꢀC. After
3 h, the mixture was filtered through Celite and the filtrates were
concentrated in vacuo. The residue was separated by column
chromatography (hexane/AcOEt = 1/1) to give the light yellow
oil 17 (284 mg, 32%). MS (FABþ) 242.3 (Mþ þ 1); νmax/cm-1
1708 (CdO); 1H NMR (300 MHz; CDCl3) δ (ppm) 0.853 (t, J=
7.2 Hz, 6H), 1.430 (q, J = 7.2 Hz, 2H), 1.442 (q, J = 7.5 Hz, 2H),
1.840 (ddd, J1 = 7.9 Hz, J2 = 3.5 Hz, 4H), 2.265 (s, 2H), 2.267 (s,
2H), 2.439-2.517 (m, 2H), 2.867-2.954 (m, 2H); 13C NMR (75
MHz; CDCl3) δ (ppm) 8.1, 24.3, 33.0, 41.7, 50.4, 52.6, 54.9, 59.3,
210.7. Anal. Calcd for C13H23NOS: C, 64.68; H, 9.60; N, 5.80.
Found: C, 64.66; H, 9.70; N, 5.76.
˚
surface area (with 1.4 A radius probe), and dipole moment were
calculated for the equilibrium structure.
Acknowledgment. This study was partially supported by
the Development of Systems and Technology for Advanced
Measurement and Analysis of the Japan Science and Tech-
nology Agency, and a Grant-in-Aid for Young Scientists
from the Japan Society for the Promotion of Science.
Supporting Information Available: 1H and 13C NMR spec-
tra of nonradical intermediates in Scheme 1, ESR parameters,
calculated descriptors, supplemental figure, STO-3G/ROHF
optimized coordinates of nitroxides, and corresponding total
energy. This material is available free of charge via the Internet
at http://pubs.acs.org.
2,2-Diethyl-9-thia-4,9,9-trioxo-1-azaspiro[5.5]undecane-9-
oxyl (8c). Compound 17 (284 mg, 1.18 mmol) was oxidized
according to the method described for compound 2. The crude
(20) Michael, W. S.; Kim, K. B.; Jerry, A. B.; Steven, T. E.; Mark, S. G.;
Jan, H. J.; Shiro, K.; Nikita, M.; Kiet, A. N.; Shujun, S.; Theresa, L. W.;
Michel, D.; John, A. M., Jr. J. Comput. Chem. 1993, 14, 1347.
(21) Tommi, H.; Mikael, P. J. Comput. Chem. 2001, 22, 1229.
(22) Wang, J.; Wolf, R. M.; Caldwell, J. W.; Kollman, P. A.; Case, D. A.
J. Comput. Chem. 2004, 25, 1157.
(23) Guesta, M. F.; Saunders, V. R. Mol. Phys. 1974, 28, 819.
(24) Davidson, E. R.; Feller, D. Chem. Rev. 1986, 86, 681.
440 J. Org. Chem. Vol. 76, No. 2, 2011