Synthesis and reduction kinetics of sterically shielded pyrrolidine nitroxides

Article abstract of DOI:10.1021/ol302506f

A series of sterically shielded pyrrolidine nitroxides were synthesized, and their reduction by ascorbate (vitamin C) indicate that nitroxide 3, a tetraethyl derivative of 3-carboxy-PROXYL, is reduced at the slowest rate among known nitroxides, i.e., at a 60-fold slower rate than that for 3-carboxy-PROXYL.

Full text of DOI:10.1021/ol302506f

ORGANIC
LETTERS
XXXX
Vol. XX, No. XX
000–000
Synthesis and Reduction Kinetics of
Sterically Shielded Pyrrolidine Nitroxides
Joseph T. Paletta,† Maren Pink,‡ Bridget Foley,† Suchada Rajca,† and Andrzej Rajca*,†
Department of Chemistry, University of Nebraska, Lincoln, Nebraska 68588-0304,
United States, and IUMSC, Department of Chemistry, Indiana University,
Bloomington, Indiana 47405-7102, United States
arajca1@unl.edu
Received September 11, 2012
ABSTRACT
A series of sterically shielded pyrrolidine nitroxides were synthesized, and their reduction by ascorbate (vitamin C) indicate that nitroxide 3, a tetraethyl
derivative of 3-carboxy-PROXYL, is reduced at the slowest rate among known nitroxides, i.e., at a 60-fold slower rate than that for 3-carboxy-PROXYL.
Nitroxide radicals have a wide range of applications
in organic synthesis,1 materials chemistry,2,3 biophysics,4
and biomedicine.5 These applications rely on the stability,
paramagnetic properties,6 and/or redox properties7 of the
radicals at ambient conditions. While fast redox reactions
of nitroxide radicals are favorable for many applications,
fast in vivo reduction of paramagnetic nitroxide radicals
to diamagnetic hydroxylamines by antioxidants, such as
ascorbate (vitamin C) or enzymatic systems, hinders their
applications in biomedicine.
3-Carboxy-PROXYL (1) (Figure 1) and its carboxylic acid
derivatives are among the most commonly used nitroxides
in vivo because of their resistance to reduction.9 However, their
half-life is only ∼2 min in the bloodstream and major organs,
as determined by magnetic resonance imaging (MRI) studies
in mice.9 Therefore, their applications as paramagnetic con-
trast agents in MRI or in vivo spin labels are rather limited.10
† University of Nebraska.
‡ Indiana University.
(1) Tebben, L.; Studer, A. Angew. Chem., Int. Ed. 2011, 50, 5034–5068.
(2) (a) Ratera, I.; Veciana J. Chem. Soc. Rev. 2012, 41, 303–349.
(b) Lahti, P. M. Adv. Phys. Org. Chem. 2011, 45, 93–169.
(3) Nishide, H.; Oyaizu, K. Science 2008, 319, 737–738.
(4) Reginsson, G. W.; Schiemann, O. Biochem. J. 2011, 434, 353–363.
(5) Burks, S. R.; Makowsky, M. A.; Yaffe, Z. A.; Hoggle, C.; Tsai, P.;
Muralidharan, S.; Bowman, M. K.; Kao, J. P. Y.; Rosen, G. M. J. Org.
Chem. 2010, 75, 4737–4741.
(6) (a) Dane, E. L.; Griffin, R. G.; Swager, T. M. Org. Lett. 2009, 11,
1871–1874. (b) Rajca, A.; Mukherjee, S.; Pink, M.; Rajca, S. J. Am.
Chem. Soc. 2006, 128, 13497–13507. (c) Olankitwanit, A.; Kathirvelu,
V.; Rajca, S.; Eaton, G. R.; Eaton, S. S.; Rajca, A. Chem. Commun. 2011,
47, 6443–6445. (d) Zagdoun, A.; Casano, G.; Ouari, O.; Lapadula, G.;
Rossini, A. J.; Lelli, M.; Baffert, M.; Gajan, D.; Veyre, L.; Maas, W. E.;
Rosay, M.; Weber, R. T.; Thieuleux, C.; Coperet, C.; Lesage, A.; Tordo,
P.; Emsley, L. J. Am. Chem. Soc. 2012, 134, 2284–2291.
(7) (a) Oyaizu, K.; Nishide, H. Adv. Mater. 2009, 21, 2339–2344.
(b) Gryn’ova1, G.; Barakat, J. M.; Blinco, J. P.; Bottle, S. E.; Coote,
M. L. Chem.;Eur. J. 2012, 18, 7582–7593.
Figure 1. Pyrrolidine nitroxides.
(9) Hyodo, F.; Matsumoto, K.; Matsumoto, A.; Mitchell, J. B.;
Krishna, M. C. Cancer Res. 2006, 66, 9921–9928.
(10) (a) Davis, R. M.; Sowers, A. L.; DeGraff, W.; Bernardo, M.;
Thet-ford, A.; Krishna, M. C.; Mitchell, J. B. Free Radical Biol. Med.
2011, 51, 780–790. (b) Khan, N.; Blinco, J. P.; Bottle, S. E.; Hosokawa,
K.; Swartz, H. M.; Micallef, A. S. J. Magn. Reson. 2011, 211, 170–177.
(8) Keana, J. F. W.; Pou, S.; Rosen, G. M. Magn. Res. Med. 1987, 5,
525–536.
r
10.1021/ol302506f
XXXX American Chemical Society

Recently, we reported the design and synthesis of an MRI
contrast agent based upon macromolecular polyradicals
derived from spirocyclohexyl nitroxide 2 (Figure 1).11,12
The agent possesses a comparatively long in vivo lifetime
2,2,6,6-tetraethylpiperidine-4-one in 15% yield.15 How-
ever, we found that the major product was alcohol 8,16
which could be obtained in about 50% yield using a larger
excess of Raney Ni. We expected that, under similar condi-
tions for the preparation of Tempone (16),17 oxidation of 8
using m-CPBAwouldprovide 10. Instead, we obtained9,16
which was then oxidized further with DessꢀMartin peri-
odinane (DMP) to give ketone-nitroxide 10. By adapting
the previously reported procedure for bromination of
Tempone,18 11 was obtained by protonation of nitroxide
10, followed by bromination, and recovery of nitroxide
with sodium nitrite. Nitroxide 11 in potassium hydroxide
solution at 20 °C undergoes Favorskii rearrangement18 to
give the target nitroxide 3. Treatment of 3 with N-hydro-
xysuccinimide (NHS) and N,N0-dicyclohexylcarbodiimide
(DCC) in THF provides spin label 3-NHS in good yield.
The synthesis of bicyclic pyrrolidine nitroxides is based
on partial Favorskii rerrangement of 12 leading to 4.19
Sodium diethyl malonate is added to a solution of a
bromoketone in anhydrous THF, and then the reaction
mixture is stirred at 50°C for 1.5ꢀ2.5 h, toproduce bicyclic
nitroxides 4ꢀ6 in about 50% isolated yields as mixtures of
diastereomers (Scheme 2).
1
and high H water relaxivity (r1) and provides contrast-
enhanced high resolution MRI in mice for over 1 h.11
To further improve the development of organic based
MRI contrast agents, we explore nitroxides that are more
resistant to reduction than 2. It is well-established that the
slower rate of reduction of nitroxides is associated with
ring size, i.e. five-membered (pyrrolidine) vs six-membered
(piperidine) ring structures,9 as well as steric shielding of
the nitroxide moiety.13,14 To search for nitroxides with
increased resistance to reduction, we focus on a series of
sterically shielded pyrrolidine nitroxides and their rate of
reduction with ascorbate. Typically, the rate of reduction
of a nitroxide by ascorbate in vitro provides a qualitative
guideline to its expected half-life in vivo.11,14
Herein we report the synthesis of sterically shielded
pyrrolidine nitroxides 3ꢀ6 (Figure 1) and reduction ki-
netics of nitroxides 1ꢀ6 by ascorbate.
Scheme 1. Synthesis of Nitroxides 3 and 3-NHSa
Scheme 2. Synthesis of Nitroxides 4ꢀ6
Structures of nitroxides 2 and 3 are confirmed by single-
crystal X-ray analysis (Figure 2).20
a Isolated yields.
Synthesis of nitroxide 3 is outlined in Scheme 1. Starting
from 7, we followed the desulfurization procedure using
Raney Ni, which was previously reported to provide
Figure 2. X-ray structures for nitroxide radicals 2 (left) and 3
(right). Carbon, nitrogen, and oxygen atoms are depicted with
thermal ellipsoids set at the 50% probability level.
(11) Rajca, A.; Wang, Y.; Boska, M.; Paletta, J. T.; Olankitwanit, A.;
Swanson, M. A.; Mitchell, D. G.; Eaton, S. S.; Eaton, G. R.; Rajca, S.
J. Am. Chem. Soc. 2012, 134, 15724–15727.
(12) Kirilyuk, I. A.; Polienko, Y. F.; Krumkacheva, O. A.; Strizhakov,
R. K.; Gatilov, Y. V.; Grigor’ev, I. A.; Bagryanskaya, E. G. J. Org. Chem.
2012, 77, 8016–8027.
(13) (a) Marxa, L.; Chiarellia, R.; Guiberteaub, T.; Rassat, A. J. Chem.
Soc., Perkin Trans. 1 2000, 1181–1182. (b) Bobko, A. A.; Kirilyuk, I. A.;
Grigor’ev, I. A.; Zweier, J. L.; Khramtsov, V. V. Free Radical Biol. Med.
2007, 42, 404–412.
(15) Sakai, K.; Yamada, K.; Yamasaki, T.; Kinoshita, Y.; Mito, F.;
Utsumi, H. Tetrahedron 2010, 66, 2311–2315.
(16) Wetter, C.; Gierlich, J.; Knoop, C. A.; Muller, C.; Schulte, T.;
Studer, A. Chem.;Eur. J. 2004, 10, 1156–1166.
(17) Cella, J. A.; Kelley, J. A.; Kenehan, E. F. J. Chem. Soc., Chem.
Commun. 1974, 943–943.
€
(14) Emoto, M.; Mito, F.; Yamasaki, T.; Yamada, K.; Sato-Akaba,
H.; Hirata, H.; Fujii, H. Free Radical Res. 2011, 45, 1325–1332.
(18) Sosnovsky, G.; Cai, Z. J. Org. Chem. 1995, 60, 3414–3418.
B
Org. Lett., Vol. XX, No. XX, XXXX

Table 1. Second-Order Rate Constants, k (Mꢀ1sꢀ1), for Initial
Rates of Reduction of Nitroxides (0.04ꢀ0.2 mM) with 20-Fold
Excess Ascorbate in 25 mM PBS pH 7.4 at 295 Ka
pyrrolidine nitroxides
piperidine nitroxides
k
k
1
2
3
4
5
6
0.063 ( 0.002
0.031 ( 0.003
e0.001b
14
15
9
5.6 ( 0.2
2.57 ( 0.03
0.039 ( 0.003
6.32 ( 0.01
3.2 ( 0.2
0.354 ( 0.006
0.170 ( 0.004
0.01
16
17
10
0.044 ( 0.002
a Mean ( two standard deviations from at least three measurements;
see: Tables S5ꢀS6 in the SI. b 1 mM 3 with 20 or 100 mM ascorbate in
125 mM PBS.
The high purity of the bulk samples of nitroxides 3ꢀ6, as
well as all other nitroxides used for kinetic studies, is deter-
mined by paramagnetic 1H NMR spectra and EPR spec-
troscopic spin concentrations.
Figure 3. Reduction profiles of 0.2 mM nitroxides with 4 mM
ascorbate in 25 mM PBS pH 7.4 at 295 K (5, 6, and 17 have
insufficient solubility to be included in the plot).
EPR spectra of nitroxides 3ꢀ6 in chloroform show triplet
patterns due to 14N hyperfine splitting, aN ≈ 14ꢀ16 G, and
g-values of about 2.006, similar to those for 1 and 2. For
bicylic nitroxide 4, each of the triplet peaks is resolved into
nonets, corresponding to the coupling of eight protons with
1H hyperfine splitting, aH ≈ 0.5 G. This EPR splitting pattern
is reproduced by DFT calculations,21 using simplified struc-
tures for cis and trans diastereomers, in which the diethylma-
lonate moiety is replaced with a methyl group. Specifically,
each diastereomer has eight protons with computed aH ≈ 0.5
G, i.e., two protons at the bridghead and six protons in the
methyls that are syn to the cyclopropane moiety.22
Rates of reduction for pyrrolidine nitroxides 1ꢀ6 are
studied under pseudo-first-order conditions using a 20-
fold excess of ascorbate in pH 7.4 PBS buffer. For com-
parison, the rates of reduction of the gem-dimethyl, spir-
ocyclohexyl, and gem-diethyl piperidine nitroxides 9, 10,
14ꢀ179,14,23 are measured under identical conditions.
Second-order rate constants, k, are obtained by monitor-
ing the decay of the low-field EPR peak height of nitr-
oxides at 295 K (Figure 3 and Table 1). Also, decays of
EPR single integrated peak height are examined and found
to produce similar values of k for most nitroxides (Tables
S5 and S6, Supporting Information (SI)).
nitroxide 4 (k = 0.354 Mꢀ1 sꢀ1) is reducedata significantly
faster rate than 1 (Figure 3). The spirocyclohexyl nitrox-
ides 2, 5, 15, and 17 are reduced at rates that are about two
times slower than those for the corresponding gem-di-
methyl nitroxides. The rates of reduction for gem-diethyl
nitroxides 3, 6, 9, and 10 are decreased by another factor of
20ꢀ70, compared to the spirocyclohexyl nitroxides.
gem-Diethyl pyrrolidine nitroxides 3 (k e 0.001
Mꢀ1
s
ꢀ1) and 6 (k ≈ 0.01 Mꢀ1
s
ꢀ1) are reduced at sig-
nificantly slower rates, compared to previously reported
nitroxides, including five-membered ring nitroxides shielded
with gem-diethyl groups, such as 3,4-dimethyl-2,2,5,5-tetra-
s
ethylperhydroimidazol-1-yloxy (k = 0.02 Mꢀ1 ꢀ1).13b As
illustrated in Figure 3, 0.2 mM nitroxide 3 in the presence of
a 20-fold excess of ascorbate shows no detectable decay
over 3 h. Only when much higher concentrations of 3 and
ascorbate are used, the reduction becomes detectable
(Figure 4); for 1 mM 3 treated with 20 and 100 mM
ascorbate, the EPR signal of nitroxide 3 decreases by only
5 and 15%, respectively, after 3 h.
The reduction profiles (Figure 4) suggest that the slow
,
rate at the initial stage of the reaction (k ≈ 0.001 Mꢀ1 sꢀ1
Table 1) is followed by slower decay of the EPR signal in
the later stage. This observation may be interpreted as the
initial slow reaction of ascorbate with nitroxide, which
produces hydroxylamine and the ascorbate radical, reaching
an “equilibrium” (Figure 4, eq 1). The equilibrium is shifted
toward hydroxylamine by slow decay of the ascorbate radical
via a known multistep mechanism.13b,24 Addition of reduced
glutathione (GSH), which is known to scavenge the ascor-
bate radical with k ≈ 10 Mꢀ1 sꢀ1 or dehydroascorbate (dis-
proportionation product of ascorbate radical),13b,24,25 leads to
The values of k for gem-dimethyl nitroxides 1, 14, and 16
in Table 1, which may be compared to the corresponding
k = 0.06, 5.46, 5.78 Mꢀ1 sꢀ1 at 293 K reported by Rigo,23
reflect a well-known finding that five-membered ring nitr-
oxides are reduced at relatively slower rates, compared to
six-membered ring nitroxides. The gem-dimethyl bicyclic
ꢀ
ꢀ
(19) Babic, A.; Pecar, S. Synlett 2008, 1155–1158.
(20) X-ray structure of 17 is described in the SI.
(21) Frisch, M. J. et al. Gaussian 09, revision A.01; Gaussian: Wallingford,
CT, 2009.
(22) Computed 1H hyperfine splitting within each syn methyl
averages to |aH| ≈ 0.5 G. All other gem-dimethyls have significantly
smaller values of average |aH| (Table S9, SI).
(23) Vianello, F.; Momo, F.; Scarpa, M.; Rigo, A. Magn. Reson.
Imaging 1995, 13, 219–226.
(24) Bielski, B. H. J.; Allen, A. O.; Schwarz, H. A. J. Am. Chem. Soc.
1981, 103, 3516–3518.
(25) Winkler, B. S.; Orselli, S. M.; Rex, T. S. Free Radical Biol. Med.
1994, 17, 333–349.
Org. Lett., Vol. XX, No. XX, XXXX
C

Figure 5. Stick and space-filling plots for X-structures of 2 (left)
and 3 (right). Carbon, nitrogen, and oxygen atoms are colored in
black, blue, and red, respectively.
in which the diethylmalonate moiety is replaced with a
methyl group, are computed at the same level of theory.
Space-filling plots for these DFT-computed nitroxides suggest
that steric shielding of the nitroxide moiety increases from
gem-dimethyl to spirocyclohexyl to gem-diethyl (Figure S6,
SI). Also, steric shielding is more effective in 1ꢀ3, compared
to the simplified structures of 4ꢀ6 (Figure S7, SI).26 Thus, the
experimentally observed rate constants for reduction of 1ꢀ6
by ascorbate may be qualitatively correlated with steric shield-
ing of the NO moiety in pyrrolidine nitroxides.
Figure 4. Reduction profiles of 1.0 mM nitroxides in 125 mM
PBS pH 7.4 at 295 K. Concentrations of ascorbate (20 or 100 mM)
and GSH (0, 5, or 25 mM) are indicated for each experiment.
higher conversion of nitroxide to hydroxylamine. This effect
is more pronounced at higher concentrations of ascorbate
(Figure 4). In the absence of ascorbate, GSH does not reduce
stable nitroxides (Figure S5, SI), and the initial rates of the
reduction of studied nitroxides with ascorbate are only
slightly increased by the addition of GSH (Table S6, SI).
The slow rate of reduction of 3 compared to that of 2
may be related to increased steric shielding of the nitroxide
moiety by gem-diethyl vs spirocyclohexyl groups, asshown
by the space-filling plots of the X-ray structures (Figure 5).
For further insight into the structure of five-membered
ring nitroxides, geometries of 1ꢀ3 were optimized at the
UB3LYP/6-311þG(d,p)þZPVE level of theory;21 for
4ꢀ6, simplified structures for cis and trans diastereomers,
Acknowledgment. We thank the National Science
Foundation and National Institutes of Health for support
of this work through Grants CHE-1012578 and NIBIB
EB008484. We thank Molly Sowers for her help with the
synthesis of 3.
Supporting Information Available. Complete acknowl-
edgment, ref 21, experimental and computational details,
and CIF files. This material is available free of charge via
the Internet at http://pubs.acs.org.
(26) Space-filling plots suggest that R-alkylation of the malonate
moiety in the cis isomer of 4ꢀ6 may significantly increase steric shielding
of nitroxide moiety.
The authors declare no competing financial interest.
D
Org. Lett., Vol. XX, No. XX, XXXX