Synthesis, structure, and X-band (9.5 GHz) EPR characterization of the new series of pH-sensitive spin probes: N,N-disubstituted 4-amino-2,2,5,5- tetramethyl-3-imidazoline 1-oxyls

Article abstract of DOI:10.1021/jo0510890

An approach to the synthesis of new imidazoline nitroxides bearing an N′,N′-disubstituted amidine group is reported. The approach is based on the alkylation of diamagnetic 4-R-amino-1,2,2,5,5-pentamethyl-3-imidazolines with bromoacetic acid ethyl ester; t

Full text of DOI:10.1021/jo0510890

Synthesis, Structure, and X-Band (9.5 GHz) EPR Characterization
of the New Series of pH-Sensitive Spin Probes: N,N-Disubstituted
4-Amino-2,2,5,5-tetramethyl-3-imidazoline 1-Oxyls
Maxim A. Voinov,*,† Julya F. Polienko,‡,§ Thomas Schanding,| Andrey A. Bobko,
,#
Valery V. Khramtsov, Yury V. Gatilov,‡ Tatyana V. Rybalova,‡ Alex I. Smirnov,† and
Igor A. Grigor’ev‡,§
Novosibirsk Institute of Organic Chemistry, Akad. Lavrent’ev Ave. 9, 630090 Novosibirsk, Russia,
Department of Chemistry, North Carolina State University, Raleigh, North Carolina 27695-8204,
Novosibirsk State University, Pyrogova St. 2, 630090 Novosibirsk, Russia, Fachbereich Chemie der
Universita¨t Kaiserslautern, Erwin-Schro¨dinger-Strasse, D-67663 Kaiserslautern, Germany, Institute of
Chemical Kinetics & Combustion, Novosibirsk 630090, Russia, and Dorothy M. Davis Heart & Lung
Research Institute, The Ohio State University, Columbus, Ohio 43210
mvoynov@ncsu.edu
Received May 31, 2005
An approach to the synthesis of new imidazoline nitroxides bearing an N′,N′-disubstituted amidine
group is reported. The approach is based on the alkylation of diamagnetic 4-R-amino-1,2,2,5,5-
pentamethyl-3-imidazolines with bromoacetic acid ethyl ester; the products of alkylation are further
oxidized to the corresponding nitroxides. The approach allows a variety of functional groups to be
introduced into the nitroxide molecule structure. Alkylation with bromoacetic acid ethyl ester was
found to proceed with high regioselectivity and afford the products of exo-alkylation. The
regiochemical assignment is made on the basis of 13C NMR spectra and confirmed by X-ray
diffraction study. All of the nitroxides synthesized here were shown to have pH-dependent EPR
spectra with pKa ranging from 3.5 to 6.2. For nitroxides 13 bearing the carboxylic group remote to
the nitroxide moiety, the changes in isotropic magnetic parameters of EPR spectra due to reversible
deprotonation of the carboxylic group were found to be small. For these nitroxides, we demonstrate
an alternative approach for pKa determination that is based on measuring the peak-to-peak line
width of the EPR spectrum in the presence of the paramagnetic broadening agent potassium
ferricyanide. The partition coefficients of nitroxides in octanol/H2O and octanol/phosphate buffer
solution mixtures were measured to reveal a range of their lipophilicities.
Introduction
adjacent to the nitroxide moiety.1a,d,2 Typically, pH-
induced changes in EPR spectra are assessed from
measurements of the isotropic hyperfine coupling con-
stant for nitrogen, aN, or, sometimes, for phosphorus, aP,1d
although the electronic g-factor is also affected by proton
An EPR method for pH value determination is based
on exquisite dependence of EPR spectra of some nitrox-
ides on the reversible protonation of functional groups
* To whom correspondence should be addressed.
† North Carolina State University.
‡ Novosibirsk Institute of Organic Chemistry.
§ Novosibirsk State University.
(1) (a) Khramtsov, V. V.; Weiner, L. M.; Grigor’ev, I. A.; Volodarsky,
L. B. Chem. Phys. Lett. 1982, 91 (1), 69-72. (b) Khramtsov, V. V.;
Weiner, L. M.; Eremenko, S. I.; Belchenko, O. I.; Schastnev, P. V.;
Grigor’ev, I. A.; Reznikov, V. A. J. Magn. Reson. 1985, 61, 397-408.
(c) Khramtsov, V. V.; Vainer, L. M. Russ. Chem. Rev. 1988, 57 (9),
824-839. (d) Cle´ment, J.-L.; Barbati, S.; Fre´javille, C.; Rockenbauer,
A.; Tordo, P. J. Chem. Soc., Perkin Trans. 2 2001, 1471-1475.
| Universita¨t Kaiserslautern.
Institute of Chemical Kinetics & Combustion.
# The Ohio State University.
10.1021/jo0510890 CCC: $30.25 © 2005 American Chemical Society
Published on Web 10/21/2005
9702
J. Org. Chem. 2005, 70, 9702-9711

EPR Characterization of pH-Sensitive Spin Probes
SCHEME 1
of the compounds. Specifically, only N′-monosubstituted
derivatives, such as compound 1, can be synthesized by
this method. Meanwhile, an introduction of a second
substituent, e.g., bearing a functional group, to the exo-
cyclic nitrogen of the amidine moiety would allow one to
manipulate the properties of the spin probe (∆aN, pKa,
lipophilicity, chemical binding capabilities) to a greater
extent, thus facilitating the further progress of the
method. Recently, an elegant approach to the synthesis
of N,N-disubstituted 4-amino-3-imidazoline-1-oxyls via
substitution of the cyano group in 4H-imidazole-5-car-
bonitrile 3-oxides with amines has been suggested.9
However, it appears that the necessity of using organo-
metallic reagents in the final step of the synthesis
imposes certain limitations on the structure of the amine
to be used. Specifically, this synthesis might not be
suitable for the amines bearing functional groups sensi-
tive toward Grignard reagents. We speculate that modi-
fication of the amidine function of the nitroxides 1 by
means of an alkylation reaction appears to be one of the
most attractive solutions. Though the example of alkyl-
ation of N′-unsubstituted amidine nitroxide 1 (R ) H)
with a highly reactive enone Mannich base methiodide
has been reported,10 the scope of application of this
particular approach is apparently yet restricted by the
synthesis of aryl derivatives. Our previous attempts to
directly alkylate both N′-unsubstituted (R ) H) and N′-
monosubstituted amidines 1 (R ) Ph, CH3) with methyl
iodide, acrylonitrile, bromoacetic acid ethyl ester, and
dimethyl sulfate failed. An apparent reason for decreased
reactivity of the amidine function is a strong electron-
withdrawing influence of the nitroxide moiety. As shown
earlier, the treatment of the diamagnetic amidine 2 with
dimethyl sulfate readily afforded the alkylation products
3 and 4 with subsequent oxidation yielding the new
nitroxides 5 and 6 with pH-dependent EPR spectra.11
Here we report an approach to the synthesis of N,N-
disubstituted 4-amino-2,2,5,5-tetramethyl-3-imidazoline
1-oxyls via alkylation of the amidine function in the
diamagnetic 4-R-amino-1,2,2,5,5-pentamethyl-3-imid-
azolines 2.
exchange reactions. The method was initially reported
in the early eighties.1 Since then, pH-sensitive spin
probes have found numerous applications in studying
proton-related phenomena, such as transmembrane pro-
ton transport in model systems,3 surface potential and
polarity of membranes and proteins,4 and acidity in the
interior ionite grains and at the solid-liquid interface.5
Recently, with development of low-frequency EPR spec-
troscopy and imaging, the method was extended to
noninvasive pH measurements in living organisms.6
Other experiments, such as site-directed pH determina-
tion of proteins and membrane-protein complexes using
high-field EPR, have been recently described.7 Further
advances in these applications call for a series of new
spin probes that not only have EPR spectra sensitive to
pH changes over a wide range but also have functional
groups suitable for residue-specific modification of com-
plex biological molecules.
Among all the pH-sensitive spin probes the 3-imid-
azoline 1-oxyl series of nitroxides 1 (Scheme 1) containing
protonatable amidine functionalities as a part of the
heterocycle are, perhaps, the most widely used in EPR
studies. These probes exhibit an exceptional sensitivity
of EPR spectra to protonation (e.g., ∆aN ≈ 0.7-0.9 G)
and a biologically important range of pK values from ca.
4.0 to 7.2 units.8a,b 4-R-amino-2,2,5,5-tetramethyl-3-
imidazoline 1-oxyls 1 are readily obtained through 1,3-
dipolar cycloaddition reaction of isocyanates to aldo-
nitrone 2,2,5,5-tetramethyl-3-imidazoline 1-oxyl.8a,b How-
ever, current synthetic procedures described for such
nitroxides impose certain restrictions on the structure
All the nitroxides synthesized in the present work were
characterized by X-band (9-10 GHz) EPR and were
shown to have pH-dependent EPR spectra with pKa
ranging from 3.5 to 6.2. For nitroxides with the carboxylic
group remote to the nitroxide moiety, we demonstrate
an alternative approach for the carboxylic group pKa
determination that is based on measurements of EPR line
width as a function of pH in the presence of the
paramagnetic broadening agent potassium ferricyanide.
We also report the partition coefficients of nitroxides in
octanol/H2O and octanol/phosphate buffer solution mix-
tures.
(2) Keana, J. F. W.; Acarregui, M. J.; Boyle, S. L. M. J. Am. Chem.
Soc. 1982, 104, 827-830.
(3) (a) Balakirev, M.; Khramtsov, V. V. J. Chem. Soc., Perkin Trans.
2 1993, 2157-2160. (b) Khramtsov, V. V.; Panteleev, M. V.; Weiner,
L. M. J. Biochem. Biophys. Meth. 1989, 18, 237-246.
(4) Khramtsov, V. V.; Marsh, D.; Weiner, L. M.; Reznikov, V. A.
Biochim. Biophys. Acta 1992, 1104, 317-324.
(5) (a) Molochnikov, L. S.; Kovalyova, E. G.; Grigor’ev, I. A.;
Reznikov, V. A. Metal-Containing Polymeric Mater. 1996, 395-401.
(b) Molochnikov, L. S.; Kovalyova, E. G.; Lipunov, I. N.; Grigor’ev, I.
A. Proc. Am. Chem. Soc. 1997, 77, 612-613. (c) Molochnikov, L. S.;
Kovalyova, E. G.; Grigor’ev, I. A.; Zagorodni, A. A. J. Phys. Chem. B
2004, 108, 1302-1313.
Results and Discussion
We chose the readily available bromoacetic acid ethyl
ester as an alkylation reagent in the synthesis shown in
Scheme 2.
(6) Khramtsov, V. V.; Grigor’ev, I. A.; Foster, M. A.; Lurie, D. J.;
Nicholson, I. Cell. Mol. Biol. 2000, 46(8), 1361-1374.
(7) Smirnov, A. I.; Ruuge, A.; Reznikov, V. A.; Voinov, M. A.;
Grigor’ev, I. A. J. Am. Chem. Soc. 2004, 126, 8872-8873.
(8) (a) Berezina, T. A.; Martin, V. V.; Volodarsky, L. B.; Khramtsov,
V. V.; Weiner, L. M. Bioorg. Khim. 1990, 16, 262-269. (b) Balakirev,
M.; Khramtsov, V. V.; Berezina, T. A.; Martin, V. V.; Volodarsky, L.
B. Synthesis 1992, 12, 1223-1225. (c) Khramtsov, V. V.; Volodarsky,
L. B. In Biological Magnetic Resonance, Vol. 14, Spin Labeling: The
Next Millennium; Berliner, L. J., Ed.; Plenum Press: New York, 1998;
pp 109-180.
(9) Kirilyuk, I. A.; Shevelev, T. G.; Morozov, D. A.; Khromovskih,
E. L.; Skuridin, N. G.; Khramtsov, V. V.; Grigor’ev, I. A. Synthesis 2003,
6, 871-878.
(10) Khlestkin, V. K.; Tikhonov, A. Y. Heterocycl. Commun. 2002, 8
(3), 249-254.
(11) Voinov, M. A.; Martin, V. V.; Volodarskii, L. B. Bull. Russ. Akad.
Sci., Div. Chem. Sci. (Engl. Transl.) 1992, 41, 2091-2095.
J. Org. Chem, Vol. 70, No. 24, 2005 9703

Voinov et al.
SCHEME 2
SCHEME 3
FIGURE 1. Feasible structures for compound 8b and selected
heteronuclear coupling constants for exo- and endo-forms. See
the text for further details.
in addition to the signals of methyl groups of imidazoline
ring, a signal at δ 3.29 ppm, assigned to the exo-N-CH3
group, ethyl group multiplets at δ 1.22 and 4.12 ppm and
the signal of the N-CH2 moiety at δ 4.52 ppm. We failed
to obtain this compound in an analytically pure form. The
hydrochloride and p-toluenesulfonate salts were also
prepared but found to be highly hygroscopic and useless
for characterization of the compound as well. Finally, the
ester obtained was treated with excess of methylamine
solution in ethanol, resulting in readily crystallizable
N-(1,2,2,5,5-pentamethyl-3-imidazoline-4-yl)-N-methyl-
amino acetic acid N-methyl amide 8b. The latter com-
pound was characterized and used for further regiochem-
ical assignment. The assignment was made on the basis
of a splitting pattern present in the proton-coupled13C
NMR spectrum of 8b. In the spectrum recorded in
acetone-d6 CH2 group reveals a triplet of quartets at δ
56.0, with 1JC-H ) 90.2 Hz and 3JC-H ) 2.4 Hz; CH3 group
reveals a quartet of triplets at δ 37.8, with 1JC-H ) 109.6
Refluxing a mixture of 4-phenylamino-1,2,2,5,5-penta-
methyl-3-imidazoline 2a, BrCH2COOEt, and K2CO3 in
acetonitrile afforded a compound whose IR spectrum
revealed an intense absorption at 1750 cm-1 that is
1
characteristic for ester carbonyl. H NMR spectrum of
the resulting compound showed, in addition to resonances
of methyl groups in the 1, 2, and 5 positions of the
imidazoline ring, ethyl group multiplets at δ 1.30 and
4.30 ppm, and a N-CH2 moiety signal at δ 5.00 ppm. To
elucidate the position of the new alkyl group, 13C NMR
and UV spectra of the compound synthesized were
compared with those of endo- and exo-derivatives 3 and
4 previously described11,12 (Scheme 3).
3
Hz and JC-H ) 3.7 Hz (see Figure 1).
The chemical shift of a phenyl group ipso-carbon atom
in N-phenyl-substituted amidines has been previously
shown to be a reliable criterion for regiochemical assign-
ment in a series of isomeric N-alkylated amidine deriva-
tives.12 If the CdN bond is in a direct conjugation with
the phenyl group, as in compound 4 (endo-alkylation),
the resonance of the ipso-carbon atom of the phenyl group
will be observed at a lower field (δ 150.31 ppm) than for
a compound with no conjugation, such as compound 3 (δ
145.72 ppm, exo-alkylation). The 13C NMR spectrum of
the compound synthesized showed the ipso-carbon atom
resonance at δ 139.93 ppm, which allowed us to assign
to the compound the structure of exo-isomeric N-(1,2,2,5,5-
pentamethyl-3-imidazoline-4-yl)-N-phenylamino acetic
acid ethyl ester 7a. The close resemblance of an UV
spectrum of compound 7a to that of 4-methylphenyl-
amino-1,2,2,5,5-pentamethyl-3-imidazoline 311 as com-
pared with the spectrum of endo-compound 4 (Scheme
3) is an additional evidence in favor of the assignment
made. None of the products of endo-alkylation were
detected in the reaction mixture. Thus, the alkylation of
4-phenylamino-1,2,2,5,5-pentamethyl-3-imidazoline 2a
with bromoacetic acid ethyl ester proceeds on the exo-
cyclic nitrogen of the amidine function with high regio-
selectivity.
Assuming that the product of endo-alkylation should
show negligibly small, if any, JC-H coupling constants
5
(Figure 1), the resulting compound was assigned similar
to the previous case, as the exo-alkylation product 7b.
Basically, examples of endo-alkylation of amidines have
been described earlier.13 Obviously, in the case of com-
pounds 2a,b the reason for regioselective exo-alkylation
is the steric hindrance of the endo-cyclic nitrogen atom
with bulky methyl groups in the position 2 of the
imidazoline ring.
The ester group in the compounds 7a,b allows one to
obtain various functional derivatives. For example, treat-
ment of 7a with ammonia afforded the amide 9 (Scheme
4). Alkaline hydrolysis of the latter as well as of ester 7a
led to the corresponding N-phenylamino acetic acid
derivative 10. Because of the high hygroscopicity of the
latter, its hydrochloric salt was obtained and used for
characterization.
Attempts to prepare the corresponding nitroxides by
oxidation of 9, 7a, and 10 in H2O2/Na2WO4/CH3OH have
been made, but did not yield satisfactory results. Ester
7a and amino acid 10 undergo destruction under the
oxidation conditions, and in the case of amide 9 the
only product isolated was the nitroxide amino acid 13a
formed in a negligible (6%) yield (Scheme 4). As we
discovered, the secondary amide function is much more
In the same manner, the alkylation of 4-methylamino-
1,2,2,5,5-pentamethyl-3-imidazoline 2b was performed.
1H NMR spectrum of the obtained compound displayed,
(13) (a) Goerdeler, J.; Roth, W. Chem. Ber. 1963, 96 6(1-2), 534-
549. (b) Skipper, P. L.; Tannenbaum, S. R.; Baldwin, J. E.; Scott, A.
Tetrahedron Lett. 1977, 49, 4269-4271. (c) Fossey, J.; Loupy, A.;
Strzelecka, E. Tetrahedron 1981, 37 (10-11), 1935-1941.
(12) Berezina, T. A.; Reznikov, V. A.; Volodarsky, L. B. Tetrahedron
1993, 49 (46), 10693-10704.
9704 J. Org. Chem., Vol. 70, No. 24, 2005

EPR Characterization of pH-Sensitive Spin Probes
SCHEME 4
SCHEME 5
FIGURE 2. Representative room temperature 9.5 GHz (X-
band) EPR spectra of the carboxylic acid nitroxide 13a taken
in 50 mM buffer solutions at various pH (concentration of
nitroxide ca. 0.1 mM). Approximate positions of the peak
height of the high-field nitrogen hyperfine components for the
protonated (pH ) 3.0) and deprotonated (pH ) 9.0) forms of
the nitroxide are shown by dashed lines.
unambiguity of regiochemical assignment made on the
basis of 13C NMR spectra.
The X-band (9.5 GHz) EPR spectra of the nitroxides
synthesized were studied as a function of pH. All new
nitroxides were found to show reversible effects of pH
on EPR spectra. Figure 2 shows a series of EPR spectra
of the carboxylic derivative 13a in aqueous buffers with
pH ranging from 3.0 to 9.0 units. It was found that each
individual EPR spectrum in such a series could be
approximated as a superposition of spectra from pro-
tonated, R•H+, and nonprotonated, R•, forms of the
nitroxide probe. These forms differ in their isotropic
nitrogen hyperfine coupling constants aN and g-factors.1b
The two-component spectra shown in Figure 2 are
indicative of a slow, on the EPR time scale, R• a R•H+
chemical exchange1b and were observed for all nitroxides
studied in this work.
We have utilized two approaches to derive pKa’s from
the nitroxide EPR spectra. The first approach is based
on calibrating experimentally measured splitting be-
tween the low field and the central nitrogen hyperfine
components of the first-derivative experimental EPR
spectra as a function of pH (see the Experimental Section
and Table 1, method A). If nitroxide EPR spectra fall into
a regime of a fast tumbling and the chemical exchange
R• a R•H+ is fast on the EPR time scale, then all
anisotropies and the contributions from the two nitroxide
forms will be averaged out. As result the experimental
splitting between the nitroxide hyperfine components will
be approximately equal to a weighted average of the
isotropic nitrogen hyperfine coupling constant of the two
forms, R• and R•H+, with the weights proportional to the
fraction of the radical in each of the forms. For nitroxides
in a regime of intermediate and slow exchange this condi-
tion is not satisfied. However, at X-band (9.5 GHz) no
splitting of low field and central nitrogen hyperfine
components is observed unless the nitroxide is per-
deuterated. That allowed many researchers to use ex-
stable under the oxidation conditions as compared to the
primary amide. Thus, the oxidation of N-methylamide
derivative 8a in the system H2O2/Na2WO4/CH3OH at a
slightly elevated temperature led to the paramagnetic
N-(2,2,5,5-tetramethyl-3-imidazoline-1-oxyl-4-yl)-N-phenyl-
amino acetic acid N-methyl amide 14a in a good yield
(Scheme 5). Using similar conditions, N-(2,2,5,5-tetra-
methyl-3-imidazoline-1-oxyl-4-yl)-N-methylamino acetic
acid N-methyl amide 14b was synthesized in a satisfac-
tory yield (see the Experimental Section). Paramagnetic
amino acids 13a,b were obtained by alkaline hydrolysis
of N-methyl amides 14a,b. Esterification of amino acid
13a was performed using the standard CH3OH/HCl
procedure. Along with methyl ester of N-(2,2,5,5-tetra-
methyl-3-imidazoline-1-oxyl-4-yl)-N-phenylamino acetic
acid 15a, the formation of another product with a lower
Rf value was observed by TLC. This product was not
isolated in a pure form and characterized. It was found
that it could be easily oxidized to ester 15a with PbO2 in
chloroform solution or by air on aluminum oxide TLC
plate. This low Rf product was assumed to be the
hydroxylamino derivative 16a, which results from the
disproportionation of the nitroxide 15a in an acidic
medium.14
The crystal structure of methyl ester 15a was deter-
mined by a single-crystal X-ray diffraction analysis. The
X-ray data (see the Supporting Information) confirm the
(14) Rozantsev, E. G. Free Nitroxyl Radicals; Plenum: New York,
1970.
J. Org. Chem, Vol. 70, No. 24, 2005 9705

Voinov et al.
TABLE 1. ∆aN Values, pKa’s, and Partition Coefficients
(P) for Nitroxides 13-15
aN, G
R•H+
Pa
∆aN,
G
no.
R•
pKa
A
B
C
13a
13b
14a
15.09 15.82 0.73 5.5b
5.94cb
5.97d
1.1
0.12
14.95 15.83 0.88 6.2b
0.015
6.63c
6.60d
14.95 15.83 0.88 4.2b
5.0
0.4
5.1
0.6
4.33c
14.91 15.80 0.89 5.2c
15.00 15.85 0.85 3.5b
14.89 15.81 0.92 5.0c
14b
15a
15b
27.7
30.6
1 (R ) Ph)
1 (R ) R-C10H7)
1 (R ) CH3)
ATI
4.958 d 8.8
118.0
197.0
2.24
1.8
6.5
5.08 d 11.5
6.211
0.79 6.2c
1.27 4.7c
0.57
0.14
2.0
2.8
13.4
HMI
a A: measured in octanol-H2O mixture. B: measured in
octanol-NaOH (0.1 M) mixture, ref 9. C: measured in octanol-
phosphate buffer solution (0.1 M) mixture, pH ) 7.5. b Method A
(see the Experimental Section). c Method B, 50 mM buffer solu-
tions. d Method B, 5 mM buffer solutions.
FIGURE 3. (a) Experimental room-temperature 9.5 GHz EPR
spectrum of the nitroxide 13a taken in 50 mM phthalate buffer
at pH ) 6.0; least-squares simulated spectra of the protonated
(b) and nonprotonated (c) forms of the nitroxide; (d) least-
squares simulated spectrum of the nitroxide 13a at pH ) 6.0;
(e) residual of the fitsthe difference between experimental and
simulated spectra.
perimentally observed splittings as an approximation of
the weighted average of the isotropic nitrogen hyperfine
coupling constants even when the fast-exchange condition
is not satisfied.8c
In the second approach (see the Experimental Section
and Table 1, method B), we have modeled experimental
spectra as a superposition of two EPR spectra from
nitroxides in a fast motion limit. The line shape of each
of the nitrogen hyperfine components is approximated
by a Voigt function, which is a convolution of Gaussian
and Lorentzian shapes.15 Parameters of each of the two
nitroxide spectra as well as the weight of the components
were adjusted during Levenberg-Marquardt optimiza-
tion. Figure 3 shows a representative spectrum of the
nitroxide 13a taken at pH ) 6.0 together with results of
least-squares simulation with the use of software de-
scribed earlier.15,16 A residual of the fitsthe difference
between experimental and best-fit spectrum (bottom line
in the Figure 3)sdemonstrates that two nitroxide model
fits the experimental spectrum rather well.
Figure 4 shows weighted, according to the double-
integrated intensities of the species present, aN for
nitroxides 13a,b as a function of pH as determined from
least-squares simulation of slow-exchange X-band EPR
spectra. The entire aN titration curve is fitted exception-
ally well to the Henderson-Hasselbalch equation17
FIGURE 4. Weighted average aN from least-squares simula-
tion of the two-component X-band EPR spectra as a function
of pH. Spectra were taken at room temperature in 50 mM
VWR buffer solutions. Corresponding least-squares Hender-
son-Hasselbalch titration curves are shown as solid lines:
filled circles, compound 13a; filled triangles, 13b.
ab10(pH - pK ) + aa
a
a )
(1)
to 9.0 the hyperfine splitting increased from 15.09 to
15.82 G for 13a and from 14.94 to 15.83 G for 13b. pKa
values for nitroxides 13-15 determined using eq 1 are
summarized in Table 1.
Table 1 shows that pKa values derived using these two
methods are somewhat different (see, for instance, 13a
and 13b). It is worthwhile to note here that these
deviations arise not from the difference in experimental
conditions such buffer type and/or buffer concentration
(see the Experimental Section, X-Band EPR Titration of
Nitroxides 13-15) but rather from approximations dur-
ing the analysis. It is likely that differential line width
1 + 10(pH - pK )
a
where aa and ab represent nitrogen hyperfine coupling
constants for the acidic and the basic form of the
nitroxide, respectively. Upon increasing the pH from 3.0
(15) Smirnov, A. I.; Belford, R. L. J. Magn. Reson. A 1995, 98, 65-
73.
(16) Smirnova, T. I.; Smirnov, A. I.; Clarkson, R. B.; Belford, R. L.
J. Phys. Chem. 1995, 99, 9008-9016.
(17) Khramtsov, V. V.; Weiner, L. M. In Imidazoline Nitroxides; Vol.
II, Applications; Volodarsky, L. B., Ed.; CRC Press: Boca Raton, FL,
1988; Chapter 2, p 42.
9706 J. Org. Chem., Vol. 70, No. 24, 2005

EPR Characterization of pH-Sensitive Spin Probes
broadening effects are responsible for some systematic
errors in the first method, which is based on measure-
ments of experimental splitting between the low field and
the central nitrogen hyperfine components.
For all nitroxides studied in this work the differences
between aa and ab were found to be ∆aN ≈ 0.9 G, which
is similar to ∆aN observed for other spin probes bearing
the amidine function.8b There are appreciable differences
in the aN values of radicals 13a and 13b (∼0.14 G, see
Figure 4 and Table 1) and radicals 15a and 15b (∼0.11
G, see Table 1) at pH ) 3.0. This indicates that the spin
density on the nitrogen atom of the nitroxyl group for
these nitroxides is somewhat different for protonated
forms but is essentially the same for nonprotonated
molecules. We speculate that one of the likely reasons
for the difference in hfs’s of protonated forms of nitroxides
is the interaction of the lone pair of electrons of exo-cyclic
nitrogen atom of amidine function with delocalized π
orbitals of the aromatic ring. For the protonated form of
the nitroxide this interaction could result in a partial
delocalization of the positive charge over the aromatic π
system. This would decrease the electron-withdrawing
effect of N3 atom of imidazoline heterocycle and, thus,
increase nitrogen hyperfine coupling constant. Other
effects such as a conformation of phenylamino moiety,
steric perturbation and solvation may contribute to the
differences in hfs’s observed for the protonated form of
nitroxides. We believe that detailed consideration and
modeling of these fine effects lay outside the scope of this
study.
pKa values of the nitroxides listed in the Table 1 show
a clear trend with the character of the substituents at
the exo-cyclic nitrogen of the amidine moiety. pKa values
of compounds 13a,b are noticeably higher than those of
14a,b and 15a,b because of a strong electron-donating
effect of COO- group that increases the amidine function
basicity. Relatively high ∆aN (∼ 0.9 G) values for the
compounds 13-15 demonstrate the high sensitivity of
their EPR spectra toward the changes in acidity of the
medium. The pKa values of the compounds 13-15 spread
from 3.5 to 6.2 pH units, making them attractive pH-
sensitive spin probes for EPR application in vivo, includ-
ing imaging. pH-sensitive unnatural amino acid deriva-
tives 13a,b could be covalently attached to either terminal
or side chain amino groups of peptides and, hence, are
potential candidates for investigating peptide structure
and polarity using SDSL (site-directed spin labeling) EPR
methods.4,18,19
FIGURE 5. Experimental pH dependence of the line width
of the EPR spectrum central nitrogen hyperfine component of
the nitroxide 13b in 1 mM Na-phosphate buffer solution in
the presence of 32 mM K3[Fe(CN)6] (2) and corresponding
calculated values of the term (-[R-]/([R] + [R-]) + [R+]/([R] +
[R+])) (b), where R- and R+ represent deprotonated and
protonated forms of the nitroxide, respectively, and R repre-
sents uncharged zwitterion form. At pH below 4.0, the above
term is equal to the fraction f of the radical in protonated form
(f(R+) ) [R+]/([R] + [R+]) and [R-] ≈ 0). At pH above 4.0, the
term is equal to the fraction of the radical in deprotonated
form (f(R-) ) [R-]/([R] + [R-]) and [R+] ≈ 0) taken with sign
minus, -f(R-), to keep similar graphic representation for the
experimental (2) and calculated (b) dependencies. The solid
line represents the nonlinear least-squares fit of the data to
the Henderson-Hasselbalch equation, yielding pK1 ) 6.07 and
pK2 ) 2.44.
group. However, the pKa of the carboxylic group still could
be determined from the EPR spectra if the nitroxide line
width measured as a function of pH in the presence of
K3[Fe(CN)6]. Paramagnetic K3[Fe(CN)6] is an effective
broadening agent for nitroxide radicals in solution.20 The
leading mechanism of magnetic interactions of paramag-
netic anions with nitroxides is a spin exchange during
bimolecular collisions. Such collisions are affected by a
charge on the nitroxide and this leads to different
broadening of the EPR spectra from nitroxides in acidic,
neutral, and basic forms. Thus, one would expect to
observe a progressive change in the line width of EPR
spectra when a nitroxide containing both the carboxylic
and the amidine groups, such as 13a or b, is titrated in
the presence of K3[Fe(CN)6]. The maximum spectral line
width should correspond to a nitroxide form with both
carboxylic and amidine groups protonated (low pH, a
positively charged spin probe). Figure 5 (filled triangles)
shows the peak-to-peak line width of the central EPR
spectral component of the nitroxide 13b measured in the
presence of 32 mM of K3[Fe(CN)6], as a function of pH.
As seen in Figure 5, the line width titration curve reveals
a plateau in the pH range from 3 to 4 pH units that is
characteristic for zwitterions. Because the experimental
EPR spectra line width arises from a superposition of the
two nitroxide forms, the pK value cannot be determined
from the line width plot directly. Instead, the fraction f
It is worthwhile to note here that although nitroxides
13a,b are ampholytes, like amino acids, their EPR
titration curves do not display the plateau characteristic
for zwitterion region of the latter. The EPR titration
curves of these nitroxides have only one transition
corresponding to the pKa value of the amidine function
(Figure 4, Table 1). The ionization of the carboxylic group
does not affect the hfs because this group is located
further away from the nitroxide moiety than the amidine
(18) (a) Monaco, V.; Formaggio, F.; Crisma, M.; Toniolo, C.; Hanson,
P.; Millhauser, G.; George, C.; Deschamps, J. R.; Flippen-Anderson,
J. L. Bioorg. Med. Chem. 1999, 7, 119-131. (b) Tominaga, M.; Barbosa,
S. R.; Poletti, E. F.; Zukerman-Schpector, J.; Marchetto, R.; Schreier,
S.; Paiva, A. C.; Nakaie, C. R. Chem. Pharm. Bull. 2001, 49 (8), 1027-
1029.
(19) Griffith, O. H.; Waggoner, A. S. Acc. Chem. Res. 1969, 2, 17-
24.
(20) Keith, A. D.; Snipes, W.; Mehlhorn, R. J.; Gunter, T. Biophys.
J. 1977, 19, 205-218.
J. Org. Chem, Vol. 70, No. 24, 2005 9707

Voinov et al.
of the nitroxide in the certain ionization state has to be
calculated and plotted as a function of pH.
carboxylic groups of most amino acids (typically, pKa )
1.77-2.58 units).cf.21
To derive a fraction f of a radical in protonated (f )
[R+]/([R] + [R+])) and deprotonated (f ) [R-]/([R] + [R-]))
states from the peak-to-peak EPR line width of the
central nitrogen hyperfine component measured in the
presence of K3[Fe(CN)6], we assumed that: (1) the proton
exchange between all the radical forms is slow, (2) the
line shape of individual components of nitroxide spectra
is well approximated by a Lorentzian function, and (3)
the changes in the magnetic field position of the central
nitrogen hyperfine component due to different g-factors
for the nitroxide forms are negligible compared with the
line width. Then, at any pH value the first-derivative
shape I′(H) of nitroxide central component can be mod-
eled as a superposition of two Lorentzian lines with two
different peak-to-peak widths Wa and Wb
Partition coefficients (P) are important characteristics
of a compound lipophilicity.22 Partition coefficients of new
nitroxides were determined and are listed in the Table 1
along with the corresponding data for widely used pH-
sensitive spin probes 1 (R ) Ph, R-naphthyl, CH3), ATI
(4-amino-2,2,5,5-tetramethyl-3-imidazoline-1-oxyl), and
HMI (2,2,3,4,5,5-hexamethylimidazolidine 1-oxyl). The
compounds synthesized in this work reveal a wide range
of lipophilicities (viz. from 0.015 to 27) that makes them
suitable pH-probes in varieties of heterogeneous environ-
ments, particularly in biological systems (e.g., blood, cell
membranes).
Conclusion
We have synthesized a series of new nitroxides using
the approach based on the alkylation of diamagnetic 4-R-
amino-1,2,2,5,5-pentamethyl-3-imidazolines. The ap-
proach allows one to overcome difficulties resulting from
the reduced reactivity of the amidine function in nitrox-
ides toward alkylation. In contrast to nitroxide amidine
1, diamagnetic amidines 2 are available using both
isocyanates and isothiocyanates as addends in 1,3-dipolar
cycloaddition reactions.11 All of the nitroxides synthesized
were characterized with X-band EPR spectroscopy and
were shown to have pH-dependent slow-exchange two-
component EPR spectra. The pKa values of the nitroxides
determined from EPR spectra were found to lie in the
range from 3.5 to 6.6 pH units. This pKa range is
particularly suitable for studies of biological systems. The
approach for the determination of pKa’s of groups remote
from the nitroxide moiety based on the nitroxide spectral
line width broadening by paramagnetic K3[Fe(CN)6],
measured as a function of pH, is proposed. Keeping in
mind the possible areas of application, the partition
coefficients of new nitroxides were determined as a
measure of their lipophilicity. The synthetic approach
described here noticeably extends the conceivable set of
nitroxides with pH-dependent EPR spectra and tunable
variable properties (∆aN, pKa, lipophilicity, binding ca-
pabilities), thus providing the basis for further advances
in application of pH-sensitive spin probes.
I'(H) )
fWa(H0 - H)
(1 - f)Wb(H0 - H)
A
+
(2)
2
2
3
4
3
4
Wa2 + (H0 - H)2
Wb2 + (H0 - H)2
[
]
(
)
(
)
where A is a constant proportional to the total quantity
of the nitroxide and H0 is the common center of the lines.
The peak-to-peak width ∆Hp-p of such a superposition
spectrum could be used to derive the fraction f of the
radical in a particular form.
Indeed, when H) H0 ( ∆Hp-p/2 the function I′(H) is
either maximal or minimal and
∂I′(H)
)
|
∂H H)H0(∆Hp-p/2
fWa(∆Hp-p2 - Wa ) (1 - f)Wb(∆Hp-p2 - Wb )
2
2
48A
+
) 0
2 3
2 3
(3Wa2 + ∆Hp-p
)
(3Wb2 + ∆Hp-p
)
[
]
(3)
From this equation f can be easily determined. The
experimentally measured line widths of the particular
forms R•H+, R• and R•- were used in the calculations for
the individual Lorentzian lines. We assumed that the line
width observed at pH ) 1.0 (3.1 G) corresponds to the
R•H+ form with both carboxylic and amidine functions
protonated, while the line width at pH ) 8.0 (1.16 G)
corresponds to the deprotonated form, R•-. The peak-to-
peak line width at pH ) 4.0 (2.64 G) was assigned to the
zwitterion form, R• (pH region from 3 to 4 pH units).
Using these values for Wa and Wb, the fraction f was
calculated from the line width data using eq 3 and plotted
in Figure 5 (filled circles) as a function of pH. Titration
plots, depicted in Figure 5, similar to conventional amino
acids titration curves, show the regions corresponding
ionization of carboxylic and amidine groups as well as
the plateau of the zwitterion form. pKa values of car-
boxylic and amidine groups were determined from non-
linear least-squares fit of the data to the Henderson-
Hasselbalch equation. From this data, the pKa value for
the amidine function (pKa ) 6.07) is rather close to pKa
) 6.60 obtained from aN titration plot (see Table 1). The
difference is likely due to approximation of the EPR line
shape by a Lorentzian function. The pKa value for
carboxylic group determined using this approach (pKa )
2.4) agrees well with the literature data for terminal
Experimental Section
IR spectra were recorded in KBr pellets (the concentration
was 0.25%, the pellet sickness was 1 mm) and CCl4 solutions.
UV spectra were recorded in EtOH solutions. 1H NMR spectra
were recorded at 400.136 and 200.132 MHz. 13C NMR spectra
were recorded at 100.614 and 50.323 MHz. Solvent signal was
used as internal standard. CDCl3 (δH ) 7.24 ppm, δC ) 76.69
ppm) was used as a solvent unless otherwise noted. The
proton-coupled 13C NMR spectrum of 8b in acetone-d6 was
acquired at 100.614 MHz. Single crystals of compound 15a
suitable for diffraction study were obtained from slow recrys-
tallization from pentane at ambient temperature over a period
of several days. X-band (9.5 GHz) EPR spectra were recorded
with samples placed into 100 µL quartz capillaries and in
PTFE capillaries (0.81 × 1.12 mm). The PTFE capillaries were
(21) Handbook of Chemistry and Physics; Lide, D. R., Ed.; CRC
Press: Boca Raton, 1995; p 7-1.
(22) Lipinski, Ch. A.; Lombardo, F.; Dominy, B. W.; Feeney, P. J.
Adv. Drug Del. Rev. 2001, 46, 3-26.
9708 J. Org. Chem., Vol. 70, No. 24, 2005

EPR Characterization of pH-Sensitive Spin Probes
folded twice and inserted into standard 3 by 4 mm quartz EPR
tube. Compounds 2a,b were synthesized according to pub-
lished procedures.11 For calibration of pH-sensitive nitroxides
the following standard buffer solutions were used: (a) potas-
sium tetraoxalate at pH ) 1.68; (b) hydrochloric acid/glycine
at pH ) 2.0; (c) potassium hydrogen phthalate/hydrochloric
acid at pH ) 3.0; (d) potassium hydrogen phthalate at pH 4.0;
(e) acetic acid/sodium acetate at pH )4 .63; (f) potassium
hydrogen phthalate/sodium hydroxide at pH ) 5.0; (g) potas-
sium hydrogen phthalate/sodium hydroxide at pH ) 6.0; (h)
sodium phosphate/potassium phosphate at pH ) 6.86, 7.0,
7.38, and 8.0; (i) boric acid/potassium chloride/sodium hydrox-
ide at pH ) 9.0; (j) sodium borate at pH ) 9.18; (k) sodium
bicarbonate/ sodium carbonate at pH ) 10. All standard buffer
solutions were colorless, contained 0.5% of a biocide Dowicide
A (sodium o-phenylphenate tetrahydrate), and at 50 mM
concentration except the pH 6.86 phthalate buffer which was
25 mM. These reference buffer solutions are specified to be
accurate to at least (0.02 pH at 25 °C and deviate from the
specified values by not more than (0.05 pH at 40 °C. Some
measurements were carried out with freshly prepared potas-
sium hydrogen phthalate/sodium hydroxide (pH ) 4.3; 4.74;
5.31) and sodium phosphate/sodium hydroxide (pH ) 5.0; 6.3)
buffer solutions. pH measurements were performed with a
digital pH meter.
Alkylation of 4-R-amino-1,2,2,5,5-pentamethyl-3-imid-
azolines 2a,b with BrCH2COOEt (General Procedure).
The mixture of 2 (0.03 mol), BrCH2COOEt (0.1 mol), and
K2CO3 (0.75 mol) in 50 mL of CH3CN was placed into a round-
bottomed flask, fitted with CaCl2 tube and condenser, and
refluxed for 80 h in the case of 2a and 15 h in the case of 2b.
The reaction was monitored by chromatography on TLC plates
(2a: Al2O3, hexane/benzene/ THF, 2:2:1; 2b: Al2O3, hexane/
EtOAc, 1:1, treatment in iodine vapors). After the reaction was
completed, the inorganic precipitate was filtered off, solvent
was removed under reduced pressure, and the residue was
separated by column chromatography (Al2O3, hexane/ether, 1:1
for 2a; Al2O3, petroleum ether/EtOAc, 10:7 for 2b) to give 7a,b.
[(1,2,2,5,5-Pentamethyl-2,5-dihydro-1H-imidazol-4-yl)-
phenyl-amino]acetic acid ethyl ester 7a was isolated as a
white powder (6.9 g of 2a afforded 8.0 g of 7a, 84%): mp 176-
178 °C (hexane/EtOAc, 1:3); IR (KBr) 1750 (CdO), 1615 (CdN);
UV (C2H5OH), λmax (lg ꢀ) 210 (4.10), 247 (3.60); 1H NMR
(CDCl3), δ ) 1.00 (s, 6H, 2CH3), 1.29 (t, 3H, CH2CH3), 1.59 (s,
6H, 2CH3), 2.18 (s, 3H, N-CH3), 4.29 (q, 2H, CH2CH3), 5.01 (s,
2H, CH2), 7.24-7.45 (m, 5H, Ph); 13C NMR (CDCl3), δ ) 14.0,
23.6, 25.9, 25.4, 59.6, 62.7, 66.4, 80.4, 128.4, 129.8, 130.4, 139.9,
166.4, 168.3. Anal. Calcd for C18H27N3O2‚HBr: C, 54.27; H,
7.03; N, 10.55. Found: C, 54.37; H, 7.35; N, 10.37. [(1,2,2,5,5-
Pentamethyl-2,5-dihydro-1H-imidazol-4-yl)methylamino]-
acetic acid ethyl ester 7b was isolated as a pale brown
viscous oil (5.1 g of 2b afforded 5.3 g of 7b, 69%, crude
product): IR (CCl4) 1742 (CdO), 1658 (CdN); 1H NMR
(CDCl3), δ ) 1.25 (t, 3H, CH2CH3), 1.40, 1.48 (both s, 6H,
2CH3), 2.24 (s, 3H, N(1)-CH3), 3.29 (s, 3H, N-CH3), 4.22 (q,
2H, CH2CH3).
to give amides 8a,b. N-Methyl-2-[(1,2,2,5,5-pentamethyl-
2,5-dihydro-1H-imidazol-4-yl)phenylamino]acetamide 8a
was isolated as colorless crystals (6.3 g of 7a afforded 4.6 g of
8a, 76%): mp 81-82 °C (hexane); IR (KBr) 1688 (CdO), 1606
(CdN), 3210 (N-H); UV (C2H5OH), λmax (lg ꢀ) 205 (4.24), 246
(3.64); 1H NMR (CDCl3), δ ) 0.86, 1.28 (both s, 6H, 2CH3),
2.19 (s, 3H, N(1)-CH3), 2.79 (d, 3H, NHCH3), 4.24 (s, 2H, CH2),
7.18-7.33 (m, 5H, Ph), 8.10 (br s, 1H, NH); 13C NMR (DMSO-
d6), δ ) 24.1, 27.1, 25.6, 56.2, 65.7, 83.5. Anal. Calcd for
C17H26N4O: C, 67.55; H, 8.61; N, 18.54. Found: C, 67.43; H,
8.55; N, 18.44. N-Methyl-2-[(1,2,2,5,5-pentamethyl-2,5-di-
hydro-1H-imidazol-4-yl)methylamino]acetamide 8b was
isolated as colorless crystals (4.8 g of 7b afforded 4.4 g of 8b,
91%): mp 105-107 °C (hexane); IR (KBr) 1650 (CdO), 1601
(CdN), 3401 (N-H); UV (C2H5OH), λmax (lg ꢀ) 213 (3.86); 1H
NMR (CDCl3), δ ) 1.22, 1.34 (both s, 6H, 2CH3), 2.26 (s, 3H,
N(1)-CH3), 3.61 (s, 3H, N-CH3), 2.74 (d, 3H, NHCH3), 3.98 (s,
2H, CH2), 7.45 (br s, 1H, NH). Anal. Calcd for C12H24N4O‚
H2O: C, 55.81; H, 10.08; N, 21.70. Found: C, 55.97; H, 10.10;
N, 21.53.
2-[(1,2,2,5,5-Pentamethyl-2,5-dihydro-1H-imidazol-4-
yl)phenylamino]acetamide 9. NH4OH (30%, 25 mL) was
added to the solution of ester 7a (3.17 g, 0.01 mol) in CH3OH
(10 mL) and allowed to stay at ambient temperature for a few
days. The course of reaction was monitored by chromatography
on TLC plates (Al2O3, hexane/ether, 1:1). The further treat-
ment was performed as in the case of amide 8. The residue,
obtained after solvent evaporation, solidified after trituration
with hexane/ether (1:1) mixture to give product 9 as a white
powder (2.6 g, 92%): mp 138-140 °C (hexane/EtOAc, 1:1); IR
(KBr) 1700 (CdO), 1578 (CdN), 3320, 3125 (NH2); UV
1
(C2H5OH), λmax (lg ꢀ) 208 (4.22), 245 (3.79); H NMR (CDCl3),
δ ) 0.69, 1.29 (both s, 6H, 2CH3), 2.21 (s, 3H, N(1)-CH3), 4.28
(s, 2H, CH2), 7.25-7.32 (m, 5H, Ph), 5.42 (br s, 1H, NH,
CONH2), 8.00 (br s, 1H, NH, CONH2). Anal. Calcd for
C16H24N4O: C, 66.67; H, 8.33; N, 19.45. Found: C, 66.86; H,
8.42; N, 19.17.
[(1,2,2,5,5-Pentamethyl-2,5-dihydro-1H-imidazol-4-yl)-
phenylamino]acetic Acid 10 Hydrochloride. A solution of
ester 7a (0.29 g, 0.0009 mol) and NaOH (0.036 g, 0.0009 mol)
in 5 mL of 1/10 H2O/CH3OH mixture was heated on a water
bath at 45 °C. The course of the reaction was monitored by
chromatography on TLC plates (Al2O3, hexane/ether, 1:1). The
reaction mixture was diluted with H2O (10 mL) and extracted
with CHCl3 (1 × 5 mL, chloroform extract was discarded). The
water phase, neutralized with 1% H2SO4, was extracted with
CHCl3 (4 × 5 mL), the extract was dried over MgSO4, the
solvent was removed in a vacuum, and the residue solidified
to give chromatographically pure 10 in 70% yield. The stream
of dry HCl gas was passed through the solution of amino acetic
acid 10 (0,58 g, 0,002 mol) in the ether/CH2Cl2 (5:1) mixture.
The light yellow precipitate of the hydrochloric salt was
collected on the filter; the mother solution was evaporated
under reduced pressure to give, after trituration with ether,
an additional amount of hydrochloric salt. The crude product
was recrystallized from CH3OH/ether (1:4) to give 10 hydro-
chloride as white crystals: mp 133-134 °C; IR (KBr) 1747
Synthesis of N-(1,2,2,5,5-Pentamethyl-3-imidazolin-4-
yl)-N-R-amino Acetic Acid N-Methyl Amides 8 (General
Procedure). A solution of ester 7 (0.02 mol) in 2 M CH3NH2
ethanol solution (50 mL) was allowed to stay at ambient
temperature for a few days. After the reaction was complete
(control by TLC, 7a: Al2O3, petroleum ether/EtOAc, 1:1,
treatment in iodine vapors, Rf ) 0.79 for 7a, Rf ) 0.3 for 8a;
7b: SiO2, CHCl3 + 1% CH3OH), the solvent was removed
under reduced pressure, the residue was dissolved in H2O (25
mL) and extracted with CHCl3 (5 × 20 mL), and the chloroform
solution was extracted with 3% HCl (5 × 15 mL). The acidic
solution was extracted with CHCl3 (2 × 15 mL, chloroform
extract was discarded), neutralized with NaHCO3, and then
extracted with CHCl3 (2 × 15 mL), the extract was dried over
MgSO4, the solvent was removed under reduced pressure, and
the chromatographically pure residue solidified upon standing
1
(CdO), 1658 (CdN); UV (C2H5OH), λmax (lg ꢀ) 207 (4.17); H
NMR (DMSO-d6), δ ) 1.06 (br s, 6H, 2CH3), 1.41 (s, 6H, 2CH3),
2.24 (s, 3H, N-CH3), 4.60 (s, 2H, CH2), 7.53 (s, 5H, Ph); 13C
NMR (DMSO-d6), δ ) 23.1, 25.5, 25.8, 57.5, 66.7, 80.4, 128.7,
129.6, 130.0, 139.9, 167.1, 167.8. Anal. Calcd for C16H23N3O2‚
2HCl‚2H2O: C, 48.25; H, 7.28; N, 10.55; Cl, 17.81. Found: C,
48.32; H, 7.22; N, 10.60; Cl, 17.60.
Synthesis of Nitroxides 14a,b (General Procedure).
Na2WO4 (0.1 g, 0.00034 mol) and H2O2 (3 mL) were added to
the solution of N-methylamide 8 (0.002 mol) in CH3OH (8 mL).
The resulting mixture was heated on a water bath at temper-
atures not exceeding 50 °C. The course of reaction was
monitored by chromatography on TLC plates (14a: SiO2,
CHCl3 + 1% CH3OH; 14b: Al2O3, petroleum ether/EtOAc, 1:1).
The reaction mixture was diluted with H2O (15 mL) and
J. Org. Chem, Vol. 70, No. 24, 2005 9709

Voinov et al.
extracted with CHCl3 (3 × 10 mL). The chloroform extract was
dried over MgSO4, the solvent was removed under reduced
pressure, and the chromatographically pure 14 solidified after
evaporation. 2-[(1-Oxyl-2,2,5,5-tetramethyl-2,5-dihydro-
1H-imidazol-4-yl)phenylamino]-N-methylacetamide 14a
was isolated as canary crystals (0.6 g of 8a affords 0.5 g of
14a, 85%): mp 148-150 °C (EtOAc/hexane, 1:15); IR (KBr)
1680 (CdO), 1580 (CdN), 3245 (NH); UV (C2H5OH), λmax (lg
ꢀ) 209 (4.21), 243 (3.88). Anal. Calcd for C16H23N4O2: C, 63.36;
H, 7.59; N, 18.48. Found: C, 62.96; H, 7.99; N, 18.35. 2-[(1-
Oxyl-2,2,5,5-tetramethyl-2,5-dihydro-1H-imidazol-4-yl)-
methylamino]-N-methylacetamide 14b was isolated as
yellow plates (0.48 g of 8b affords 0.24 g of 14b, 50%): mp
59-61 °C (hexane/EtOAc, 1:2); IR (KBr) 1671 (CdO), 1595
(CdN), 3451 (NH); UV (C2H5OH), λmax (lg ꢀ) 217 (4.09). Anal.
Calcd for C11H21N4O2‚H2O: C, 50.96; H, 8.88; N, 21.62.
Found: C, 50.88; H, 8.90; N, 21.81.
N-(2,2,5,5-Tetramethyl-3-imidazolin-1-oxyl-4-yl)-N-R-
amino Acetic Acids 13a,b (General Procedure). KOH
(0.06 g) was added to the solution of N-methylamide 14 (0.005
mol) in 50% aqueous methanol (6 mL), and the resulting
mixture was heated on the oil bath at 80 °C. The reaction was
monitored by chromatography on TLC plates (13a: SiO2,
CHCl3 + 5% CH3OH; 13b: Al2O3, petroleum ether/EtOAc, 1:1).
CH3OH was removed under reduced pressure, and the residue
was neutralized with 1% H2SO4 and extracted with CHCl3 (7
× 10 mL). The chloroform extract was dried over MgSO4, and
amino acetic acids 13a,b were obtained after the solvent
evaporation as the solid products. [(1-Oxyl-2,2,5,5-tetra-
methyl-2,5-dihydro-1H-imidazol-4-yl)phenylamino]acetic
acid 13a was isolated as pale yellow crystals (1.52 g of 14a
afforded 0.9 g of 13a, 68%): mp 148-150 °C dec (hexane/
EtOAc, 4:3); IR (KBr) 1740 (CdO), 1591 (COO-), 1649 (CdN);
UV (C2H5OH), λmax (lg ꢀ) 206 (4.22), 243 (3.81); Anal. Calcd
for C15H20N3O3: C, 62.07; H, 6.90; N, 14.48. Found: C, 62.20;
H, 7.11; N, 14.27. [(1-Oxyl-2,2,5,5-tetramethyl-2,5-dihydro-
1H-imidazol-4-yl)-methyl-amino]-acetic acid 13b is iso-
lated as pale yellow crystals (1.2 g of 14b affords 0.74 g of
13b, 65%). mp 136-137 °C (hexane/EtOAc, 1:5); IR (KBr):
1598 (COO-), 1667 (CdN); UV (C2H5OH), λmax (lg ꢀ): 215
(4.06); Anal. Calcd for C10H18N3O3: C, 52.63; H, 7.89; N, 18.42.
Found: C, 52.77; H, 7.55; N, 18.32.
N-(2,2,5,5-Tetramethyl-3-imidazolin-1-oxyl-4-yl)-N-R-
amino Acetic Acid Methyl Ester 15a,b (General Proce-
dure). A stream of the dry HCl gas was passed into the
methanol solution of amino acetic acid 13 (0.0034 mol) for 15-
20 min. The course of the reaction was monitored by chroma-
tography on TLC plates (Al2O3, petroleum ether/EtOAc, 1:1).
After the reaction was completed, the solvent was removed
under vacuum, the solid residue, containing, along with
nitroxide 15, the corresponding hydroxylamine derivative 16a
was dissolved in CHCl3, and PbO2 (0.007 mol) was added with
vigorous stirring. After the hydroxylamine derivative vanished
(control by TLC analysis), the lead oxides were filtered off, the
solvent was removed under reduced pressure, and the residue
was separated on preparative TLC plate (Al2O3, petroleum
ether/EtOAc, 1:1). [(1-Oxyl-2,2,5,5-tetramethyl-2,5-dihydro-
1H-imidazol-4-yl)phenylamino]acetic acid methyl ester
15a was isolated as yellow crystals (0.98 g of 13a afforded 0.9
g of 15a, 92%): mp 106-107 °C (pentane); IR (KBr) 1750
(CdO), 1588 (CdN); UV (C2H5OH), λmax (lg ꢀ) 203 (4.18), 242
(3.78). Anal. Calcd for C16H22N3O3‚1.5H2O: C, 58.01; H, 7.55;
N, 12.68. Found: C, 58.28; H, 7.78; N, 12.60. [(1-Oxyl-2,2,5,5-
tetramethyl-2,5-dihydro-1H-imidazol-4-yl)-methyl-amino]-
acetic acid methyl ester 15b is isolated as light yellow
crystals (0.77 g of 13b affords 0.74 g of 15b, 90%). mp 89-91
°C (hexane/EtOAc, 5:1); IR (KBr): 1750 (CdO), 1588 (CdN);
Anal. Calcd for C11H20N3O3: C, 54.54; H, 8.26; N, 17.35.
Found: C, 54.28; H, 7.97; N, 17.24.
5 mM sodium phosphate buffer and 0.1 mM DTPA with the
final concentration of the nitroxides of ca. 0.1 mM were
prepared. The resulting solutions were titrated with HCl or
NaOH solutions to the required pH using a digital pH-meter
equipped with a glass electrode. The accuracy of pH value
measurements was (0.05 pH units. The EPR spectra of the
samples were recorded in a 200 µL quartz EPR flat cell.
Spectrometer settings were as follows: modulation amplitude,
0.8 G; scan width, 50 G; sweep time, 200 s; time constant, 200
ms; incident microwave power, 10 mW. Hyperfine splitting
constants were measured as a distance between the low field
and the central nitrogen hyperfine components of the EPR
spectra and are accurate to 0.02 G. To determine the pKa
values of the nitroxides, the experimental curves of the hfs
were fitted to the Henderson-Hasselbalch eq 1.
Method B. A series of the nitroxide solutions in standard
50 mM buffer solutions in a pH range from 2.0 to 10.0 with
the final concentration of nitroxide ca. 0.1 mM were prepared.
The room-temperature X-band EPR spectra were measured.
Hfs constants of the nitroxides (aN) were determined from
least-squares simulation of slow-exchange EPR spectra with
the software described.15,16 To determine the pKa values the
experimental aN were fitted to the Henderson-Hasselbalch
eq 1. Typical spectrometer settings were as follows: modula-
tion amplitude, 1 G; time constant, 32 ms; microwave power,
3.1 mW; sweep time, 60 s; scan width, 80 G. To be assured
that the spectrometer settings did not affect the results of
titration, the experiments (for nitroxide 14a) were repeated
with the following combinations of modulation amplitude and
microwave power (pKa values obtained are given in parenthe-
ses): 1 G, 3.1 mW (4.33); 0.2 G, 3.1 mW (4.36); 1 G, 0.4 mW
(4.37); 0.8 G, 0.4 mW (4.37); 0.5 G, 0.4 mW (4.34); 0.2 G, 0.4
mW (4.37).
To elucidate the role of molarity of the buffer solutions, the
titrations of nitroxides 13a and 13b were repeated in 5 mM
buffer solutions. Deionized water (pH)4.66) was used as a
dilutant. The final pH values of the samples were measured
with three-point calibrated digital pH-meter and corrected
values were used in an aN-pH plot (see Table 1).
X-Band EPR Titration of the Nitroxide 13b in the
Presence of K3[Fe(CN)6]. K3[Fe(CN)6] (32 mM) was added
to 0.5 mM solution of the nitroxide 13b in 1 mM phosphate
buffer. The resulting solution was titrated with solutions of
HCl or KOH to the required pH value. The peak-to-peak line
width of the central nitrogen hyperfine component of the
spectrum was measured as a function of pH. Spectrometer
settings were the same as in X-band EPR titration experi-
ments, method A.
Partition Coefficients. The partition coefficients were
determined with a “shake flask” method. To ensure accuracy,
the measurements were repeated twice with two different
nitroxide concentrations, 0.5 and 0.7 mM. Preliminary tests
showed that shaking the solution for more than 5 min did not
influence the reproducibility of the results. Typically, a solid
sample of a nitroxide was placed in a vial containing 1 mL of
H2O (or phosphate buffer solution at pH ) 7.5) and octanol (1
mL). The mixture was shaken vigorously for 5 min and allowed
to stand for 4 h until separation of the phases was completed.
The partition coefficients were determined from EPR spectra
of the nitroxides as octanol/water phase integral intensities
ratio. The spectrometer settings were as follows: time con-
stant, 100 ms; sweep time, 160 s; modulation amplitude, 0.8
G; microwave power, 12.5 mW. The solutions of 2,2,6,6-
tetramethylpiperidine 1-oxyl (Tempo) and N-(2,2,5,5-tetra-
methyl-3-imidazoline-1-oxyl-4-yl)-N-phenylamino acetic acid
13a of known concentration in phosphate buffer solution,
water, and octanol, were used as the references. Error in the
partition coefficients was estimated up to 5%.
Acknowledgment. We thank N. G. Skuridin
(Novosibirsk) for technical assistance in some EPR
experiments. The comments of Dr. R. McArthur
X-Band EPR Titration of Nitroxides 13-15. Method
A. Titration of nitroxides was performed similarly to the
published procedure.8b Solutions of nitroxides in a mixture of
9710 J. Org. Chem., Vol. 70, No. 24, 2005

EPR Characterization of pH-Sensitive Spin Probes
(Raleigh) are gratefully acknowledged. The National
Science Foundation through the Fellowship made to Dr.
Voinov (Grant No. DGE-0312165) provided support for
this work. The Russian Foundation for Basic Research
(01-03-32452a) and INTAS (INTAS-99-01086) also pro-
vided financial support of this work. The project was
also partially supported by the National Institutes of
Health (KO1 EB03519) and CRDF (RUC1-2635-NO-05).
Acknowledgment is made to the donors of the American
Chemical Society Petroleum Research Fund for partial
support of this research through ACS PRF No. 40232-
G7 (A.I.S.).
Supporting Information Available: Crystallographic
information file (CIF) and ORTEP for compound 15a. This
material is available free of charge via the Internet at
http://pubs.acs.org.
JO0510890
J. Org. Chem, Vol. 70, No. 24, 2005 9711