Real-time monitoring of drug-induced changes in the stomach acidity of living rats using improved pH-sensitive nitroxides and low-field EPR techniques

Article abstract of DOI:10.1016/j.jmr.2006.06.002

New improved pH-sensitive nitroxides were applied for in vivo studies. An increased stability of the probes towards reduction was achieved by the introduction of the bulky ethyl groups in the vicinity of the paramagnetic N{single bond}O fragment. In addition, the range of pH sensitivity of the approach was extended by the synthesis of probes with two ionizable groups, and, therefore, with two pK_a values. Stability towards reduction and spectral characteristics of the three new probes were determined in vitro using 290 MHz radiofrequency (RF)- and X-band electron paramagnetic resonance (EPR), longitudinally detected EPR (LODEPR), and field-cycled dynamic nuclear polarization (FC-DNP) techniques. The newly synthesized probe, 4-[bis(2-hydroxyethyl)amino]-2-pyridine-4-yl-2,5,5-triethyl-2,5-dihydro-1H-imidazol-oxyl, was found to be the most appropriate for the application in the stomach due to both higher stability and convenient pH sensitivity range from pH 1.8 to 6. LODEPR, FC-DNP and proton-electron double resonance imaging (PEDRI) techniques were used to detect the nitroxide localization and acidity in the rat stomach. Improved probe characteristics allowed us to follow in vivo the drug-induced perturbation in the stomach acidity and its normalization afterwards during 1 h or longer period of time. The results show the applicability of the techniques for monitoring drug pharmacology and disease in the living animals.

Full text of DOI:10.1016/j.jmr.2006.06.002

Journal of Magnetic Resonance 182 (2006) 1–11
www.elsevier.com/locate/jmr
Real-time monitoring of drug-induced changes in the stomach
acidity of living rats using improved pH-sensitive nitroxides
and low-field EPR techniques
a
b
b
c
Dmitrii I. Potapenko , Margaret A. Foster , David J. Lurie , Igor A. Kirilyuk ,
b
c
a
James M.S. Hutchison , Igor A. Grigor’ev , Elena G. Bagryanskaya ,
d,e,
*
Valery V. Khramtsov
a
International Tomography Center, Novosibirsk, Russia
Biomedical Physics, School of Medical Sciences, University of Aberdeen, Aberdeen AB25 2ZD, UK
b
c
N.N.Vorozhtsov Novosibirsk Institute of Organic Chemistry, Novosibirsk, Russia
d
Institute of Chemical Kinetics and Combustion, Novosibirsk, Russia
Dorothy M. Davis Heart and Lung Research Institute, Division of Pulmonary, Critical Care and Sleep Medicine, the Ohio State University,
Columbus, OH 43210, USA
e
Received 6 March 2006; revised 18 April 2006
Available online 22 June 2006
Abstract
New improved pH-sensitive nitroxides were applied for in vivo studies. An increased stability of the probes towards reduction was
achieved by the introduction of the bulky ethyl groups in the vicinity of the paramagnetic NAO fragment. In addition, the range of
pH sensitivity of the approach was extended by the synthesis of probes with two ionizable groups, and, therefore, with two pKa values.
Stability towards reduction and spectral characteristics of the three new probes were determined in vitro using 290 MHz radiofrequency
(RF)- and X-band electron paramagnetic resonance (EPR), longitudinally detected EPR (LODEPR), and field-cycled dynamic nuclear
polarization (FC-DNP) techniques. The newly synthesized probe, 4-[bis(2-hydroxyethyl)amino]-2-pyridine-4-yl-2,5,5-triethyl-2,5-dihy-
dro-1H-imidazol-oxyl, was found to be the most appropriate for the application in the stomach due to both higher stability and conve-
nient pH sensitivity range from pH 1.8 to 6. LODEPR, FC-DNP and proton–electron double resonance imaging (PEDRI) techniques
were used to detect the nitroxide localization and acidity in the rat stomach. Improved probe characteristics allowed us to follow in vivo
the drug-induced perturbation in the stomach acidity and its normalization afterwards during 1 h or longer period of time. The results
show the applicability of the techniques for monitoring drug pharmacology and disease in the living animals.
ꢀ 2006 Elsevier Inc. All rights reserved.
Keywords: Nitroxide; In vivo EPR techniques; In vivo pH measurement; Stomach acidity
1. Introduction
lular acidosis in tumors [5], depth-specific tissue pH varia-
tions during skin treatments [6] or wound healing [7], etc.
Disruption of normal functioning of the gut, such as asso-
ciated with ulcer formation, or therapeutic intervention of
drugs can significantly alter gastric pH [8,9].
Non-invasive pH detection mostly relies on endogenous
or/and exogenous molecular probes that are weak acids or
bases with pH-dependent spectral properties. Absorption
and fluorescent techniques are particularly effective for
Spatially and temporally addressed in vivo pH measure-
ments are of considerable clinical relevance. Aberrations of
the normal pH homeostasis includes local acidosis induced
by ischemia [1,2], infection [3] or inflammation [4], extracel-
*
Corresponding author. Fax: +1 614 293 4799.
E-mail address: valery.khramtsov@osumc.edu (V.V. Khramtsov).
1090-7807/$ - see front matter ꢀ 2006 Elsevier Inc. All rights reserved.
doi:10.1016/j.jmr.2006.06.002

2
D.I. Potapenko et al. / Journal of Magnetic Resonance 182 (2006) 1–11
pH studies on cellular levels [10,11], while nuclear magnetic
resonance, NMR, and electron spin resonance, EPR, have
the advantage for in vivo applications in animals and
humans [12,13]. 31P NMR has been most often used for
pH measurement in living tissues based on the pH sensitiv-
ity of the chemical shift of endogenous inorganic phos-
phate, Pi. However pH assessment using 31P NMR and
Pi has its own limitations, including lack of resolution
(about 0.2–0.3 units and even lower resolution at lower
pH), the fact that Pi concentrations vary with metabolism
and ischemia, chemical shift dependence on ionic strength,
and narrow range of pH measurement near its pKa value,
6.75 [14–16]. From the above considerations, a number
of endogenous and exogenous analogues of phosphates
and phosphonates have been tested as alternatives to Pi,
among them phosphorus-containing metabolites, such as
ATP, 2,3-phosphoglycerate, and glucose 6-phosphate
[17–19], and exogenous phosphorus-containing pH probes
based on alkyl- and aminophosphonates [20,21]. To
increase intrinsic sensitivity of the pH detection by
NMR, several exogenous pH probes for 19F NMR
measuring pH in vivo in rats [27,32]. To extend the accessi-
ble range of pH the nitroxide 2 with the pKa = 3.0 was also
tested [27]. However hydrophobic character of the probe 2
favoured its membrane localization proved by PEDRI
imaging data, and, therefore, compromised its utility to
report stomach acidity. The short life time of the probes
in vivo is another important factor limiting their applica-
tion. Thus nitroxide 1 showed exponential LODEPR signal
decay with characteristic time about 16 min [32]. In this
work three improved pH-sensitive nitroxides, namely
3,4-dimethyl-2,2,5,5-tetraethylimidazolidine–1-oxyl, (4),
4-dimethylamino-5,5-dimethyl-2-ethyl-2-pyridine-4-yl-2,5-
dihydro-1H-imidazol-oxyl (5) and newly synthesized
4-[bis(2-hydroxyethyl)amino]-2-pyridine-4-yl-2,5,5-triethyl-
2,5-dihydro-1H-imidazol-oxyl (6) were studied (see Scheme
1). To increase stability of the probes towards reduction the
bulky ethyl groups in the vicinity of the NAO fragment
were introduced. In addition to protonatable nitrogen
N-3, a second ionizable group was incorporated into the
structure of the probes 5 and 6, therefore, extending their
range of pH sensitivity and the probe hydrophilicity. One
of the probes, the nitroxide 6, was found to be the most
appropriate for the pH monitoring in the stomach due to
its higher stability, hydrophilic character and convenient
pH sensitivity range from pH 1.8 to pH 6. The kinetics
of in vivo drug-induced perturbation in the stomach acidity
in living rats and its normalization afterwards were mea-
sured in the physiologically relevant pH range by low-field
EPR-based techniques for the first time.
1
[22–24] and H NMR [25,26] have been also developed.
A crucial advantage of exogenous EPR probes over
NMR is more than three orders of magnitude in sensitivity.
However, EPR applications in biological systems are limit-
ed by low depth of microwave penetration into aqueous
samples. Despite these formidable problems, recently
developed low-field EPR-based techniques, such as low-
field RF-EPR, LODEPR, FC-DNP offer unique opportu-
nity for non-invasive measurements of the functional
EPR probes in vivo (for recent reviews see [27,28]). In com-
bination with the steady progress in the design of pH-sen-
sitive nitroxides [29,30] this makes in vivo pH detection and
imaging possible.
2. Materials and methods
2.1. Chemicals
Gallez et al. [31] applied pH-sensitive nitroxide 1 of imi-
dazolidine type and nitroxide 3 of imidazoline type (see
Scheme 1) for non-invasive measurement of pH in the
stomach of living mice using L-band EPR spectroscopy.
Authors demonstrated an ability to detect drug-induced
Diethylenetriaminopentacetic acid (DTPA), sodium and
potassium phosphate, NaOH, citric acid, and sodium borate
were purchased from Sigma. For measurements of calibra-
tion curves buffer containing 50 mM citrate, phosphate,
and borate. The synthetic procedures for the nitroxides
3,4-dimethyl-2,2,5,5-tetraethylperhydroimidazol-1-yloxy,
4, and 5,5-dimethyl-4-(dimethylamino)-2-ethyl-2-pyridine-
4-yl-2,5-dihydro-1H-imidazol-1-oxyl, 5 (Scheme 1), are
described earlier in the references [30,29], respectively. The
stomach acidity changes. However, the probes
1
(pKa = 4.6) and 3 (pKa = 1) were practically pH-insensitive
to the stomach acidity in the most physiologically relevant
range from pH 2 to 3.6 due to their inappropriate pKa val-
ues. In our previous work, we demonstrated suitability of
LODEPR and FC-DNP for monitoring nitroxide 1 and
nitroxide
4-[bis(2-hydroxyethyl)amino]-2-pyridine-4-yl-
2,5,5-trimethyl-2,5-dihydro-1H-imidazol-oxyl, 6 (Scheme 1),
HO(H2C)2
(CH2)2OH
N
X
N
N
N
N
N
Et
Et
Et
Et
(CH2)2COOH
N
Et
Et
Et
Et
N
O
N
O
N
O
N
O
N
O
N
N
1 (X=Me)
2 (X=Ph)
3
4
5
6
Scheme 1. Structures of the pH-sensitive spin probes 1–6.

D.I. Potapenko et al. / Journal of Magnetic Resonance 182 (2006) 1–11
3
was prepared by analogy with the nitroxide 5 [29] as
described in Appendix.
the right part of Eq. (1) were obtained by the substitution
of the numerical values for the Planck constant, h, Bohr
magneton, lB, and nitroxide g-factor, g = 2.0 (providing
A and H in Gauss, and x in MHz). Hyperfine splitting
values, A, were determined by fitting calculated spectra to
the experimental ones using Eq. (1) for the line positions
and Lorentzian function for the lineshape.
2.2. Animals
Thirteen locally bred, male Sprague Dawley rats with
weight (BW) 180–200 g were used for in vivo studies. The
rats were fasted prior to oral intubation (gavage) of 3 ml
of 5 mM or 7.5 mM solution of a pH nitroxide label, with
or without bicarbonate. Animals were anesthetized by
intraperitoneal injection of 41 mg/kg BW ketamine (Veta-
lar; Parke-Davis, Pontypool, UK) and 20.5 mg/kg BW
xylazine (Rompun; Bayer, Bury St. Edmunds, UK), given
1 min after pH label dose. For examination the anesthe-
tized rats were placed in the prone position to aid their res-
piration and were secured to a plastic support couch that
was fitted into the coil assembly of both the LODEPR
and the FC-DNP/PEDRI systems. After an examination
the animals were killed by pentabarbitone overdose while
under anesthesia. All animal procedures were carried out
in accordance with local guidelines and under British
Home Office project licence no. PPL 60/2309 (M.A.F.).
ꢀ
ꢁ
2
h
glB
h
glB
2
3A xꢀ2
x
3Aðx=2:79Þꢀ2ðx=2:79Þ
Hlow-field
¼
¼
h
B
Aꢀ2gl
x
Aꢀ2ðx=2:79Þ
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2
4ðgl xÞ ꢀ9A2
h
2
4ðx=2:79Þ ꢀ9A2
hx
glB
B
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Hcenter-field
¼
¼ ðx=2:79Þ
2
4ðgl xÞ ꢀA2
h
2
4ðx=2:79Þ ꢀA2
B
ꢀ
ꢁ
2
h
glB
h
glB
2
3A xþ2
x
3Aðx=2:79Þþ2ðx=2:79Þ
Hhigh-field
¼
¼
h
B
Aþ2gl
x
Aþ2ðx=2:79Þ
ð1Þ
2.5. LODEPR
For this study an in-house-built imager/spectrometer
was used in a spectrometer mode [35]. The 50 ml glass or
plastic bottle with titrated buffer solution containing
0.5 mM of the radical was placed into magnet bore and
spectra were registered. For both in vitro and in vivo mea-
surements spectra were obtained using an excitation fre-
2.3. X-band EPR measurements
Bruker EMX and Bruker ER-200D-SRC spectrometers
were used. The measurements were done in 100 ll quartz
tubes. The samples containing buffer solutions of 50 lM
nitroxide were titrated to the required pH, placed in the
EPR cavity and the spectra were acquired afterwards. Stan-
dard spectrometer settings were as follows: microwave
power from 1 to 5 mW; number of points, 1024, over sweep
width, 80 G (resolution 0.08 G); modulation amplitude,
1 G, for the compounds 5 and 6, and 3 G for the com-
pound 1 with the broad linewidth; sweep time, 41 s; time
constant, 40.96 ms. Hyperfine splitting constant values
were determined by fitting of the experimental spectra with
WinSim-2000 program package (EPR Spectral Simulation
for MS-Windows 9x, NT Version 0.98; Public EPR
Software tools; David A. O’Brien, David R. Durling and
Dr. Yang C. Fann; National Institute of Environmental
Health Sciences; National Institute of Health, USA).
quency of 304 MHz and
a detection frequency of
425 kHz. The field was swept over 45 G centered on
102.5 G with a sweep time of 51.2s or 128 s depending on
signal intensity. A time constant of 50 ms was used and
normally four averages were taken. The incident RF power
´
level was 5W continuous throughout each acquisition
period.
2.6. FC-DNP/FC-PEDRI
An in-house-built imager/spectrometer used to collect
both FC-DNP spectra and FC-PEDRI images as previous-
ly described [36,37]. Images and spectra were collected
using field-cycling techniques with an EPR irradiation
frequency of 121 MHz applied for 500 ms before each col-
lection of a proton signal, and an NMR frequency of
2.5 MHz. The repetition time (TR) of the pulse sequence
was 1200 ms. The average incident power during an acqui-
sition was 8.3 or 16.6 W depending on the study. Spectra
were obtained by means of a field-cycled DNP pulse
sequence in which the evolution field strength was stepped.
The EPR irradiation frequency was maintained constant
(as was the NMR frequency), so each step of the evolution
field was equivalent to the sampling of an EPR signal at a
different magnetic field value [37]. The number of steps and
their separation defined the overall width of the observed
spectrum and its resolution. In this study, spectra of 60
points over a field range of 55 G centered on 37.5 G (reso-
lution 0.5 G) provided the full three-line spectrum. Spectra
2.4. RF-EPR measurements
A home-made RF-EPR spectrometer was used, operat-
ing at 300 MHz [33]. Measurements were carried out on
aqueous samples of 1 ml of 0.5–2 mM solution of the rad-
ical in glass sample tubes of 10 mm external diameter.
Spectra were collected as eight point-by-point averages of
320 point sweeps of a total field 45 G centered on
102.5 G. Sweep time was 32 s or 80.6 s depending on signal
intensity. Incident power of 10 mW was applied. Modula-
tion amplitude was 0.5 G at a modulation frequency of
25 kHz. Resonance line positions for nitroxide radical in
low magnetic field are given by the Eq. (1) which take into
account Breit–Rabi effect [34]. The simplified expressions in

4
D.I. Potapenko et al. / Journal of Magnetic Resonance 182 (2006) 1–11
were obtained from living rats or from 150 ml plastic
bottles of solutions. The half of the distance between field
positions of low and high spectral lines has been used as
pH-sensitive spectral parameter. Calculation of the abso-
lute HFS values from FC-DNP spectra was not performed
due to low spectral resolution.
Field-cycling PEDRI images were collected as 15x15 cm
coronal projective images of the chest and abdomen of the
rat with irradiation of the high-field EPR spectral line of
the nitroxide at 121 MHz. An interleaved saturation recov-
ery pulse sequence (TR = 1200 ms) with alternate EPR-ON
and EPR-OFF was used to collect Overhauser-enhanced
and standard NMR images. The difference image obtained
by subtraction of the data shows the location of the spin
probe in the rat.
observed HFS values were calculated as half the distance
between the positions of the high-field and low-field spec-
tral lines. Radical solutions, 50 lM for X-band EPR and
0.5 mM for low-field techniques were prepared in phos-
phate–citrate buffer (10 mM each). pH was serially adjust-
ed by titration with sodium hydroxide solution or
hydrochloric acid to the desired pH with subsequent acqui-
sition of EPR spectra.
3. Results and discussion
3.1. In vitro studies
Scheme 1 shows the structures of the compounds 4–6
used in these studies. The compound 4 is tetraethyl-substi-
tuted analog of the nitroxide 1 previously used for in vivo
studies of the stomach acidity [31,32]. Two other nitroxides
are three-ethyl substituted nitroxides with an additional
bulky group, pyridine, in their structure bound to atom
C2 of the radical heterocycle. The introduction of bulky
ethyl substitutes in the vicinity of the NAO fragment signif-
icantly increases the stability of the nitroxides, as well as
prolonging their lifetime in the blood (for the radical 4
and 1 the bimolecular rate constants of the reduction by
ascorbate are 0.04 and 0.85 Mꢀ1sꢀ1, respectively [30]). Typ-
ical RF-EPR, LODEPR and FC-DNP spectra of the radi-
cals 4–6 are shown in Figs. 1 and 2. Poor solubility of the
nitroxide 4 in aqueous solutions and significant broadening
2.7. pH titration
Studies of variation in the observed HFS constant, Aobs
,
with pH were made with the use of three low-frequency
techniques and by X-band EPR spectroscopy. At X-band
three spectral lines are equally spaced and the Aobs can
be calculated as the distance between neighboring peaks.
At low detection frequencies, however, the three lines are
unequally spaced due to the Breit–Rabi equation (34).
The Aobs values were obtained from LODEPR and RF-
EPR spectra by fitting the experimental spectra using
Lorentzian lineshapes and Eq. (1). For FC-DNP the
a
b
100
110
120
100
110
120
80
90
80
90
c
30
40
60
30
40
60
10
20
50
10
20
50
Field, G
Fig. 1. Low-field spectra of solutions of the nitroxides 4 (left) and 6 (right) in phosphate–citrate buffer (10 mM each) at pH 5.0: (a) RF-EPR spectrum of
the samples of 2 ml volume, radical concentration 2 mM, sweep time being equal to 80.6 s; (b) LODEPR spectrum of the samples of 50 ml volume, radical
concentration 0.5 mM, sweep time 128 s, RF power 5 W, RF frequency 296 MHz, modulation frequency 193 kHz; (c) FC-DNP spectrum of the sample of
150 ml volume, radical concentration 0.8 mM, sweep time 80.6 s; frequency of EPR irradiation 121.3 MHz. The other spectrometer settings were as
described in Section 2.

D.I. Potapenko et al. / Journal of Magnetic Resonance 182 (2006) 1–11
5
a
b
10
20
30
40
50
60
Field, G
Fig. 2. FC-DNP spectra of the nitroxide 5 obtained in phosphate–citrate
buffer (10 mM each) at pH 7 (a) and pH 2 (b). Sample volume was 150 ml.
Frequency of EPR irradiation 121.4 MHz. All other settings were as
described in Section 2. A dotted line is extended from each peak of the
spectra (a) to aid the eye.
of its signals compared with the nitroxides 5 and 6 (see
Figs. 1 and 2) apparently limit further applications of this
probe, particularly in vivo, where its membrane localization
is expected. Fig. 3 shows the dependencies of the HFS con-
stants of the nitroxides 5 and 6 on pH obtained by X-band
and all three low-field EPR techniques. Decrease in pH val-
ue of the solution results in reversible decrease of the HFS
value.
The titration curves clearly show pH sensitivity of the
HFS within extended pH range apparently due to the pres-
ence of two ionizable groups with different pKa values,
pKa1 and pKa2, in the structure of the nitroxides. The solid
lines, shown in Fig. 3, were calculated according to the fol-
lowing equation for the standard titration curve for the
compound with two ionizable groups:
Fig. 3. pH Dependencies of the observed HFS constant of the nitroxides 5
(a) and 6 (b). Data were obtained from analysis of FC-DNP (n),
LODEPR (,), RF-EPR (h) and X-band EPR (s) spectra as described in
Section 2. Solution of the nitroxides in phosphate–citrate buffer (10 mM
each) was titrated to necessary pH value and spectra were registered.
Sample volumes were 150 ml (nitroxide concentration 0.5 mM), 50 ml
(0.5 mM), 2 ml (1 mM) and 150 ll (50 lM) for FC-DNP, LODEPR, RF-
and X-band EPR measurements. Solid lines correspond to best fit of
obtained data to Eq. (2) yielding parameters Ai and pKai listed in Table 1.
a1ꢀpKa2
Þ
A1 þ A2 ꢁ 10ðpHꢀpK Þ þ A3 ꢁ 10ð2pHꢀpK
a1
Aobs
¼
ð2Þ
Deviation from high magnetic field approximation most
significantly affected the FC-DNP spectra acquired at low-
est magnetic field compared with other techniques (cf. field
scales for the spectra in Fig. 1). First, small satellite peaks
at magnetic field around 29 and 44 G for the nitroxide 5
(Fig. 2) and nitroxide 6 (Fig. 1c, right) were observed.
The origin of these peaks is due to the appearance of tran-
sitions between electron-nuclear spin levels which are not
allowed at higher magnetic fields. Second, the Breit–Rabi
effect results in significantly unequal distances between
the components of the triplet spectrum (see Eq. (1)). For
example, according to (1) for the nitroxide with hypefine
splitting, A = 15 G, and spectrometer frequency,
x = 300 MHz, the corresponding distances between high-
field and central lines and central and low-field lines are
16.15 G and 14.0 G, correspondingly. Table 1 summarizes
pKa and HFS values for the nitroxides 4–6 obtained from
the fitting of the experimental data to Eq. (2), and pH
range accessible to measurement. The nitroxides 5 and 6
a1ꢀpKa2
Þ
1 þ 10ðpHꢀpK Þ þ 10ð2pHꢀpK
a1
where Aobs is the experimentally measured HFS (see Mate-
rial and methods); pH = ꢀlg[H+]; Ai (i = 1, 2, 3) is the
HFS for the nitroxide in the different ionizable forms,
namely, neutral, protonated at nitrogen atom N-3, and
protonated both at N-3 and pyridine nitrogen atom. Calcu-
lated curves are in good agreement with the experimental
data with correlation coefficients more than 0.999 for the
nitroxides 5 and 6. Note that the HFS values from
RF-EPR and LODEPR spectra, Aobs, were calculated tak-
ing into account the Breit–Rabi effect (see Section 2) and
therefore demonstrates a good agreement with the HFS
values obtained from X-band EPR spectra (Fig. 3). Alter-
natively, a half of the distance between high- and low-field
spectral lines can be used to estimate Aobs which demon-
strates similar pH dependence but yields slightly different
absolute HFS values from X-band data particularly for
FC-DNP method (see Fig. 3).

6
D.I. Potapenko et al. / Journal of Magnetic Resonance 182 (2006) 1–11
Table 1
pKa and HFS values for the nitroxides 4–6 obtained from fitting the experimental data to Eq. (2), and pH range accessible to measurement by the probe
Nitroxide
X-band EPR
RF-EPR
LODEPR
FC-DNP
4a
A1, G
A2, G
pKa
15.28 0.02
13.92 0.02
4.96 0.03
4–6
—
—
—
15.37 0.07
13.79 0.01
5.19 0.04
16.62 0.14
14.97 0.12
5.10 0.16
pH range
5
A1, G
A2, G
A3, G
pKa1
pKa2
pH range
15.28 0.01
14.39 0.02
13.95 0.01
5.25 0.04
3.09 0.08
2.3–6.0
15.28 0.01
14.41 0.02
13.96 0.01
5.23 0.03
3.27 0.07
15.22 0.01
14.41 0.01
13.93 0.01
5.37 0.02
3.26 0.04
16.36 0.04
15.40 0.05
14.89 0.05
5.04 0.05
2.69 0.09
6
A1, G
A2, G
A3, G
pKa1
pKa2
pH range
14.93 0.01
14.40 0.01
13.54 0.01
4.91 0.03
2.82 0.02
1.7–5.8
14.96 0.01
14.42 0.02
13.53 0.01
4.95 0.06
2.73 0.04
14.90 0.01
14.39 0.01
13.48 0.01
4.87 0.04
2.70 0.02
16.03 0.04
15.67 0.05
14.50 0.05
4.91 0.05
2.83 0.05
a
An Eq. (2) was applied for fitting HFS dependencies for the nitroxide 4 with a single ionizable group supposing pKa1 = pKa and pKa2 = +1.
have much broader ranges of pH sensitivity compared with
the nitroxide 4, while total pH effects on the HFS value are
similar. Normal stomach acidity in rodents varies signifi-
cantly in the range from pH 2 to 4 depending on the par-
ticular animal and the diet [38]. The probe 6 is the most
sensitive to pH in this acidity range showing 0.38 G change
in the HFS per one pH unit compared with 0.21 G/pH unit
for the nitroxide 5 and 0.05 G/pH unit for the nitroxide 3.
Note also that previously used nitroxides 1 and 3 have very
low sensitivity to pH in the range of pH from 2 to 4 (0.13
and 0.04 G/pH unit, respectively) [31,32].
a
80
90
100
110
120
3.2. In vivo studies
Field, G
The nitroxides 5 and 6 with extended spectral pH sensi-
tivity, that covers the stomach acidity, were selected for
in vivo measurements. Fig. 4 shows LODEPR and
FC-DNP spectra from a living rat, which received a 3 ml
dose of 5 mM solution of the nitroxide 6 in deionised water
by oral intubation. The spectra were registered 11 and
20 min after administration of the dose, respectively, and
show good signal to noise ratio (SNR are equal to 20
and 35 for the spectra shown in Figs. 4a and b, correspond-
ingly). The baseline of LODEPR spectra is slightly distort-
ed apparently due to the manganese presented in the rat
diet [32]. The FC-DNP spectrum is superimposed on a
ramped baseline due to the increase in equilibrium proton
magnetization with evolution magnetic field strength. Both
techniques provide adequate resolution for accurate deter-
mination of line splitting, and hence, of the pH of stomach
media.
b
10
20
30
40
Field, G
50
60
Fig. 4. LODEPR (a) and FC-DNP (b) spectra from a living rat after
receiving a 3 ml dose of 5 mM solution of the nitroxide 6. Spectra were
registered 11 min (a) and 20 min (b) after dosing, respectively. Spectrom-
eter settings were as described in Section 2.
Typical time dependencies of pH value of stomach
media for two rats measured by LODEPR and FC-DNP
techniques are shown in Fig. 5. Every rat was administered
with 3 ml of gavage contained 5 mM nitroxide 6 alone or
together with bicarbonate in deionized water.
Administration of the nitroxide 6 decreases the acidity in
the stomach followed by its relaxation to physiological
value of pH 2–3 within 20 min. This is not surprising and

D.I. Potapenko et al. / Journal of Magnetic Resonance 182 (2006) 1–11
7
1.0
0.8
0.6
0.4
0.2
0.0
6
5
4
3
2
0
20
40
60
80
100
120
0
10
20
30
40
50
Time, min
Time, min
Fig. 5. Time dependencies of pH changes in the stomach measured by
LODEPR after giving 3 ml of gavage containing 5 mM nitroxide 6 alone
(h) or with 50 mM bicarbonate (d). After dosing the rats were placed in
the LODEPR coil assembly and a series of spectra were obtained. In the
first case (h) after 15 min the rat was pulled out and placed into the FC-
DNP/PEDRI system for subsequent measurements at times marked by
arrows. After that the rat was returned to LODEPR coil assembly and one
more spectrum was recorded. The sweep time of the LODEPR measure-
ments was 52 s. The frequency of EPR irradiation during FC-DNP
spectral acquisition was 121.3 MHz. All other settings were as described in
2. HFS values from each spectrum were determined as described in Section
2 and pH values were calculated from them using data in Fig. 3 as
calibration curves. The solid line corresponds to the best fit of the
experimental data to exponential function with characteristic time
(8.9 1.1) min.
Fig. 6. Time dependencies of signal intensities of the nitroxides 5 (s) and 6
( ) in the rat stomach. The rats received a 3 ml doses containing 5 mM of
n
the nitroxides in deionized water. After administration of a dose animals
were placed in the LODEPR coil assembly and the spectra were obtained.
Spectrometer settings were as described in Section 2. Signal intensities were
measured as the height of the middle line of the recorded spectra.
tration. Note that the rapid in vivo decay of the signal
intensity of the previously used pH sensitive nitroxide,
1, with characteristic time of about 16 min [32] was a signif-
icant limiting factor in its application (in addition to its low
spectral sensitivity to the acidity of solution at pH < 3.6).
Fig. 7 shows a set of FC-PEDRI images of a rat mea-
sured 33 min after administration of 3 ml of 5 mM nitrox-
ide 6 in deionized water. It is clearly seen that spin label is
localized in the stomach during the time of the experiment,
supporting assignment of the previously measured pH
values (see Fig. 5) to acidity in the stomach media. The
FC-PEDRI images of the distribution of a 5 mM solution
of probe 5 measured 23 min after gavage show similar
localization of the probe 5 in the animal’s stomach (data
not shown).
may be explained by the high concentration of the probe,
5 mM, and its weak basic properties with pKa1 = 4.9 well
above of stomach acidity. The absolute values of the
changes in pH of the stomach media and characteristic
times of its relaxation varied between individual rats, and
in most cases allowed fitting to exponential pH relaxation
with characteristic time from 5 to 10 min (see Fig. 5 cap-
tion). An addition of antacid, bicarbonate, results in a big-
ger decrease of the initial stomach acidity compared to
nitroxide alone, and in much longer relaxation to the phys-
iological value as shown in Fig. 5. Similar in vivo kinetics of
pH changes were observed upon administration of the nitr-
oxide 5 in the presence or absence of bicarbonate (data not
shown). However the data were less reproducible probably
due to higher toxicity of the nitroxide 5 (two of six animals
given this compound died under anesthesia within half an
hour of receiving the gavage).
The observed data demonstrate the ability of the tech-
nique to monitor stomach acidity in living animals using
nitroxide 6 during the periods of up to 1–2 h over a wide
range of pH (Fig. 5). Note that while probe itself perturbs
pH in the stomach, its long-life time in vivo allows monitor-
ing stomach acidity after pH normalization. In other words
the probe really reports the news rather making the news.
Fig. 6 shows the decay of intensity of the central line of
the LODEPR spectrum for the 5 mM nitroxides 5 and 6
measured in vivo. The rate of the decay allows linear inter-
polation, namely about 50% of the radical intensities are
decayed during the first 60 min (nitroxide 6) or 30 min
(nitroxide 5), still leaving signal intensity enough for regis-
4. Conclusions
Newly synthesized pH-sensitive nitroxides were found to
be more stable towards reduction in vivo and have extended
ranges of pH sensitivity compared with previously avail-
able probes. These two factors make these pH probes very
useful for in vivo real-time monitoring of the stomach acid-
ity. In particular, the nitroxide 6 with two ionizable groups
has convenient pH sensitivity range from pH 1.8 to pH 6.
The hydrophilic structure of the probe limits its diffusion
from the stomach into the membrane compartments while
bulky groups in the vicinity to the nitroxyl fragment pre-
vent the radical from reduction. The localization of the
nitroxides in the rat stomach was proved by its imaging
using PEDRI. The use of the new probes and low-field
EPR techniques, LODEPR and FC-DNP, allowed us to
follow, for the first time, in vivo drug-induced perturbation
in the stomach acidity and its subsequent normalization in
physiologically relevant pH range over periods of 1 h or
longer. The results show the applicability of the approaches

8
D.I. Potapenko et al. / Journal of Magnetic Resonance 182 (2006) 1–11
Fig. 7. FC-PEDRI images of a rat gavaged with 3 ml of a 5 mM solutions of nitroxide 6. Images were taken 33 min after administration of the nitroxide
with EPR irradiation off (a), or on (b) and difference between them (c). The diagram (d) demonstrates the anatomy (RL and LL, right and left lungs; H,
heart; S, stomach; RK and LK, right and left kidney).
for monitoring drug pharmacology and disease in living
animals.
stirred for 35 h at 25 ꢁC and allowed to stand at ꢀ5 ꢁC
overnight. The crystalline precipitate was filtered off,
washed with cold 50% EtOH and with cold water, to give
dihydroimidazole 8 (3.6 g, 70%) as colorless crystals, mp
132–134 ꢁC (ethyl acetate) (Found: C, 66.85; H, 8.30; N,
18.03. Calcd for C13H19N3O: C, 66.92; H, 8.21; N,
18.01); mmax(KBr)/cmꢀ1 3161, 2966, 2915, 2877, 1643,
1606, 1566, 1482, 1421, 1381, 1312, 1235, 1029, 1003,
893, 872, 830, 786 and 769; kmax (EtOH)/nm 257 (lge
3.46); dH (400 MHz; CDCl3) 0.77 (3 H, t, J 7.2, CH3, Et),
0.92 (3H, t, J 7.2, CH3, Et), 1.39 and 2.06 (2H each, quartet
AB, J1 7.2, J2 14.4, 2· CH2, 4-Et), 1.49 (2 H, quartet, J 7.2,
CH2, Et), 1.84 (3 H, d, J2, CH3C@N), 5.43 (1H, quartet, J
2.0, CH), 7.19 and 8.20 (4 H, AA0BB0, J 6.0, Py), 8.35 (1 H,
c, OH); dC (100 MHz; CDCl3) 8.44, 9.40 (CH3, Et), 25.74,
29.10 (CH2, Et), 16.96 (CH3AC@N), 78.82 (C5), 92.23
(C2), 122.46 (C3, Py), 148.68 (C2, Py), 150.80 (Ci, Py) and
178.35 (C@N).
Acknowledgments
This work was partly supported by grants from NIH
(KO1 EB03519), CRDF RUC1-2635-NO-05, RFBR
04-03-32299, RFBR-05-04-48632 and from INTAS
(99-1086).
Appendix A. Synthesis of the nitroxide 6
The nitroxide 4-[bis(2-hydroxyethyl)amino]-2-pyridine-
4-yl-2,5,5-trimethyl-2,5-dihydro- 1H-imidazol-oxyl, 6, was
prepared by analogy with the nitroxide 5 [29] using
3-hydroxyamino-3-ethylpentan-2-one as a starting com-
pound as shown in Scheme 2.
A.1. 5,5-Diethyl-4-methyl-2-pyridine-4-yl-2,5-dihydro-1H-
imidazole-1-ol (8)
A.2. 4,4-Diethyl-5-methyl-2-pyridine-4-yl-4H-imidazole
3-oxide (9)
Pyridine-4-carbaldehyde (2.4 g, 22 mmol) was added to
a stirred solution of 3-hydroxyamino-3-ethylpentan-2-one
hydrochloride, 7 (4 g, 22 mmol), in ethanol (8 ml) and
25% aqueous ammonia (9 ml). The reaction mixture was
A suspension of 8 (2.3 g, 10 mmol), and PbO2 (4.78 g,
20 mmol) in CH2Cl2 (30 ml) was stirred for 1–3 h. After
the reaction was complete (control by TLC analysis, Silu-

D.I. Potapenko et al. / Journal of Magnetic Resonance 182 (2006) 1–11
9
N
N
O
4-PyCHO, NH3
EtOH, H2O
PbO2
N
NH
N
O
CH2Cl2
N
N
OH
7
OH
8
9
i-PrONO
i-PrONa
OH
OH
HO
N
NC
N
N
N
N
TsCl
Et3N
HN(CH2CH2OH)2
N
N
O
N
N
N
N
O
O
11
10
12
1. EtMgBr, THF
2. H2O
3. MnO2
OH
OH
N
N
N
N
O
6
Scheme 2. Synthetic route for the synthesis of the probe 6.
fol, eluent Et2O–methanol 25:1, development with I2
vapour) the lead oxides were filtered off and the solvent
was removed in vacuum to give 9 (2 g 90%) as colorless
crystals, mp 81–84 ꢁC (EtOAc–hexane 1:1), (Found: C,
67.29; H, 7.40; N, 18.19; Calcd for C13H17N3O: C, 67.51;
H, 7.41; N, 18.17); mmax (KBr)/cmꢀ1 2973, 2932, 2879,
1595, 1544, 1522, 1477, 1460, 1428, 1403, 1311, 1202,
1060, 991 and 837; kmax (EtOH)/nm 335 (lge 4.04), 278
(3.81); dH (200 MHz; CDCl3) 0.42 (6 H, t, J 7.2, 2 · CH3,
4-Et), 1.69 and 1.94 (2 H each, quartet AB, J1 7.2, J2
14.4, 2 · CH2, 4-Et), 2.13 (3H, s, 5-Me), 8.33, 8.61 (2H
each, AA0BB0, J 4.8 Hz, Py); dC (50 MHz; CDCl3) 6.96
(CH3, Et), 28.53 (CH2, Et), 17.14 (5-Me), 91.28 (C4),
145.76 (C2), 178.56 (C5), Py: 132.97 (Ci), 120.40 (C3),
150.18 (C2).
solution of NaCl (20 ml) was added to the residue and
the precipitate was filtered off and recrystallized from
EtOAc to give 10, (3.3 g, 80%), mp 214–217 (EtOAc),
(Found: C, 59.84; H, 6.16; N, 21.60; Calcd for
C13H16N4O2: C, 59.99; H, 6.20; N 21.52); mmax (KBr)/
cmꢀ1 1603, 1556, 1523, 1470, 1440, 1409, 1380, 1322,
1206, 1026, 1002, 837, 727 and 706; kmax (EtOH)/nm 359
(lge 4.02), 256 (4.47); dH (400 MHz; ; CDCl3-CD3OD
1:1) 0.57 (6 H, t, J 7.2, 2 · CH3, 4-Et), 2.11 and 2.29 (2
H each, quartet AB, J1 7.2, J2 14.4, 2 · CH2, 4-Et), 8.05
(1H, s, HC@NAO), 8.52, 8.71 (2H each, AA0BB0 J 6 Hz,
4-Py); dC (50 MHz; CDCl3-CD3OD 1:1) 7.56 (CH3, Et),
31.35 (CH2, Et), 92.90 (C4), 147.87 (C2), 171.75 (C5),
144.29 (HC@NO), Py: 134.12 (Ci), 121.28 (C3), 150.03 (C2).
A.4. 4,4-Diethyl-2-pyridine-4-yl-4H-imidazole-5-carbonitrile
3-oxide (11)
A.3. 4,4-Diethyl-5-methyl-2-pyridine-4-yl-4H-imidazole-5-
carbaldehyde oxime (10)
TsCl (0.95 g, 5 mmol) was added portionwise to a stirred
solution of oxime 10 (1.3 g, 5 mmol) of in a mixture of
CHCl3 (7.5 ml) and triethylamine (16 ml, 11 mmol). The
resulting solution was stirred for 1 h, washed with water
and dried over MgSO4. The CHCl3 was removed in vacu-
um and the residue was separated by column chromatogra-
phy (Kieselgel 60, Merck, eluent chloroform) to give 11
(0.97 g, 80%), yellow crystals, mp 89–91 (hexane) (Found:
C, 64.16; H, 5.76; N, 22.95; Calcd for C13H14N4O: C,
64.45; H, 5.82; N, 23.13); mmax (KBr)/cmꢀ1 3043, 2974,
2937, 2882, 2220, 1597, 1554, 1508, 1470, 1406, 1409,
1329, 1065, 991, 833, 770 and 710; kmax (EtOH)/nm 372
Na (1 g, 41 mmol) was dissolved in isopropanol (30 ml);
after the reaction became slow the mixture was heated to
60 degrC until Na was completely dissolved. The solution
was allowed to cool to room temperature to form a suspen-
sion of i-PrONa. Isopropyl nitrite (3.5 ml, 39 mmol) and a
solution of 9 (3.7 g, 16 mmol) in 20 ml of isopropanol were
added subsequently to the stirred suspension of i-PrONa in
isopropanol and the mixture was allowed to stand for 1 h.
After the reaction was complete (TLC, Silufol UV-254, elu-
ent EtOAc) the mixture was acidified with AcOH to pH 6
and isopropanol was removed in vacuum. A saturated

10
D.I. Potapenko et al. / Journal of Magnetic Resonance 182 (2006) 1–11
(lge 3.74), 300 (3.95), 239 (4.17), 229 (4.17); dH (400 MHz;
CDCl3) 0.68 (6 H, t, J 7.6, 2 · CH3, 4-Et), 2.09 and 2.10 (2
H each, quartet AB, J1 7.6, J2 14.4, 2 · CH2, 4-Et), 8.31,
8.77 (2H each, AA0BB0 J 6 Hz, 4-Py); dC (100 MHz;
CDCl3) 7.03 (CH3, Et), 29.17 (CH2, Et), 93.92 (C4),
111.27 (C „ N), 147.66 (C2), 148.36 (C5), Py: 131.90 (Ci),
119.59 (C3), 150.61 (C2).
2971, 2937, 2876, 1589, 1552, 1466, 1414, 1368, 1327,
1298, 1199, 1138, 1184, 1167, 1066, 1053, 1007, 979, 903,
852, 836 and 810; kmax(EtOH)/nm 222 (lge 4.06).
References
[1] J.L. Zweier, P. Wang, A. Samouilov, P. Kuppusamy, Enzyme-
independent formation of nitric oxide in biological tissues, Nat. Med.
1 (1995) 804–809.
[2] J.M. Bond, E. Chacon, B. Herman, J.J. Lemasters, Intracellular pH
and Ca2+ homeostasis in the pH paradox of reperfusion injury to
neonatal rat cardiac myocytes, Am. J. Physiol. 265 (1993)
C129–C137.
A.5. 2-[(2-Hydroxyethyl)(4,4-diethyl-2-(4-pyridyl)-3-
oxido-4H-imidazol-5-yl)amino]ethanol (12)
2-[(Hydroxyethyl)amino]ethanol (6.5 ml) was triturated
with 11 (2.5 g, 10 mmol) and the resulting homogeneous
solution was allowed to stand overnight. The reaction mix-
ture was then poured into saturated solution of NaCl
(15 ml) and extracted with CHCl3 (3 · 5 ml). The extract
was dried over K2CO3, the CHCl3 was removed in vacuum,
and the residue was triturated with Et2O, the precipitate
was filtered off to give 12 (1.7 g, 53%), yellow crystals,
mp 148–150 ꢁC (THF) (Found: C, 59.71; H, 7.63; N,
17.49. Calcd for C16H24N4O3: C, 59.98; H, 7.55; N,
17.49); mmax (KBr)/cmꢀ1 2976, 2935, 2861, 1603, 1586,
1558, 1520, 1447, 1428, 1418, 1391, 1312, 1268, 1211,
1176, 1129, 1066, 1052, 1002, 969, 952, 836, 792 and 708;
kmax (EtOH)/nm 392 (lge 3.69), 267 (4.19); dH (200 MHz;
CDCl3) 0.68 (6 H, t, J 7.4, 2 · CH3, 4-Et), 1.95 and 2.22
(2 H each, quartet AB, J1 7.4, J2 14.4, 2 · CH2, 4-Et),
3.80 (8H, br m, CH2), 4.60 (2H, br s, OH), 8.42, 8.64
(2H each, AA0BB0 J 6 Hz, 4-Py); dC (50 MHz; CDCl3)
7.72 (CH3, Et), 28.79 (CH2, Et), 37.99 (N-Me), 52.00 (br,
NACH2), 59.56 (OACH2), 86.11 (C4), 146.92 (C2), 170.14
(C5), Py: 133.93 (Ci), 121.49 (C3), 149.81 (C2).
[3] L.A. Ferrari, L. Giannuzzi, Clinical parameters postmortem analysis
and estimation of lethal dose in victims of a massive intoxication with
diethylene glycol, Forensic Sci. Int. 153 (2005) 45–51.
[4] U. Issberner, P.W. Reeh, K.H. Steen, Pain due to tissue acidosis: a
mechanism for inflammatory and ischemic myalgia? Neurosci. Lett.
208 (1996) 191–194.
[5] N. Raghunand, R.J. Gillies, pH and drug resistance in tumors, Drug.
Resist. Updat. 3 (2000) 39–47.
[6] C. Kroll, W. Hermann, R. Stosser, H.H. Borchert, K. Ma¨der,
Influence of drug treatment on the microacidity in rat and human
skin—an in vitro electron spin resonance imaging study, Pharm. Res.
18 (2001) 525–530.
[7] B. Greener, A.A. Hughes, N.P. Bannister, J. Douglass, Proteases and
pH in chronic wounds, J. Wound Care 14 (2005) 59–61.
[8] T. Fukuchi, K. Ashida, H. Yamashita, N. Kiyota, R. Tsukamoto, H.
Takahashi, D. Ito, R. Nagamatsu, Influence of cure of Helicobacter
pylori infection on gastric acidity and gastroesophageal reflux: study
by 24-h pH monitoring in patients with gastric or duodenal ulcer, J.
Gastroenterol. 40 (2005) 350–360.
[9] R. Zhou, P. Moench, C. Heran, X. Lu, N. Mathias, T.N. Faria, D.A.
Wall, M.A. Hussain, R.L. Smith, D. Sun, pH-dependent dissolution
in vitro and absorption in vivo of weakly basic drugs: development of a
canine model, Pharm. Res. 22 (2005) 188–192.
[10] R.G. Zhang, S.G. Kelsen, J.C. LaManna, Measurement of intracel-
lular pH in hamster diaphragm by absorption spectrophotometry, J.
Appl. Physiol. 68 (1990) 1101–1106.
[11] A. Molich, N. Heisler, Determination of pH by microfluorometry:
intracellular and interstitial pH regulation in developing early-stage
fish embryos (Danio rerio), J. Exp. Biol. 208 (2005) 4137–4149.
[12] R.J. Gillies, N. Raghunand, M.L. Garcia-Martin, R.A. Gatenby, pH
imaging. A review of pH measurement methods and applications in
cancers, IEEE Eng. Med. Biol. Mag. 23 (2004) 57–64.
A.6. 4-[Bis(2-hydroxyethyl)amino]-2-pyridine-4-yl-2,5,5-
triethyl-2,5-dihydro- 1H-imidazol-oxyl (6)
One molar of EtMgBr in THF was added dropwise to a
stirred solution of 12 (1.6 g, 5 mmol) in THF (10 ml). The
reaction was controlled by TLC (Al2O3 Polygram Alox N/
UV 254, Macherey-Nagel, eluent CHCl3–methanol
50:1-2). Usually 3–5 ml of the organometallic reagent solu-
tion were sufficient for the reaction to be complete. The reac-
tion mixture was allowed to stand for 0.5 h. The reaction
mixture was quenched with water (3 ml). The brown resin
precipitate formed was separated from the solution, water
(3 ml), Et2O (20 ml), Na2CO3 (1 g) and MnO2 (3 g,
34.5 mmol) were added and the mixture was stirred vigor-
ously for 2 h, the oxidant was filtered off and thoroughly
washed with CHCl3–ethanol mixture 1:1. The filtrate was
evaporated under reduced pressure, diluted with 1 N NaOH
(20 ml) and extracted with t-BuOMe. The extract was dried
over Na2CO3. The solvent was removed in vacuum and the
nitroxide 6 was isolated from the residue by column chroma-
tography on Al2O3, eluent—t-BuOMe–ethanol 30:1, yield
0.35 g, 20%, yellow crystals, mp 103–105 ꢁC (hexane–
t-BuOMe 1:1) (Found: C, 61.87; H, 8.65; N, 15.68. Calcd for
C18H29N4O3: C, 61.87; H, 8.36; N, 16.03); mmax (KBr)/cmꢀ1
[13] V.V. Khramtsov, Biological imaging and spectroscopy of pH, Curr.
Org. Chem. 9 (2005) 909–923.
[14] R.J. Gillies, J.R. Alger, J.A. den Hollander, R.G. Shulman, Intracel-
lular pH measured by NMR: methods and results, in: R. Nuccitelli,
D.W. Deamer (Eds.), Intracellular pH: Its Measurement, Regulation
and Utilization in Cellular Functions, Alan R. Liss, New York, 1992,
pp. 79–104.
[15] S. Martel, J.L. Clement, A. Muller, M. Culcasi, S. Pietri, Synthesis
and 31P NMR characterization of new low toxic highly sensitive pH
probes designed for in vivo acidic pH studies, Bioorg. Med. Chem. 10
(2002) 1451–1458.
[16] A.S. Ojugo, P.M. McSheehy, D.J. McIntyre, C. McCoy, M. Stubbs,
M.O. Leach, I.R. Judson, J.R. Griffiths, Measurement of the
extracellular pH of solid tumours in mice by magnetic resonance
spectroscopy: a comparison of exogenous (19)F and (31)P probes,
NMR Biomed. 12 (1999) 495–504.
[17] P.M.L. Robitaille, P.A. Robitaille, G.G. Brown, An analysis of the
pH-dependent chemical-shift behavior of phosphorus-containing
metabolites, J. Magn. Reson. 92 (1991) 73–84.
[18] G.D. Williams, T.J. Mosher, M.B. Smith, Simultaneous determina-
tion of intracellular magnesium and pH from the three 31P NMR
chemical shifts of ATP, Anal. Biochem. 214 (1993) 458–467.

D.I. Potapenko et al. / Journal of Magnetic Resonance 182 (2006) 1–11
11
[19] L.H. Schliselfeld, C.T. Burt, R.J. Labotka, 31P nuclear magnetic
resonance of phosphonic acid analogues of adenosine nucleotides as
functions of pH and magnesium ion concentration, Biochemistry 21
(1982) 317–320.
[20] K. Bruynseels, N. Gillis, P. Van Hecke, F. Vanstapel, Phosphonates
as 31P-NMR markers of extra- and intracellular space and pH in
perfused rat liver, NMR Biomed. 10 (1997) 263–270.
[21] S. Pietri, M. Miollan, S. Martel, F. Le Moigne, B. Blaive, M. Culcasi,
alpha- and beta-phosphorylated amines and pyrrolidines, a new class
of low toxic highly sensitive 31P NMR pH indicators. Modeling of
pKa and chemical shift values as a function of substituents, J. Biol.
Chem. 275 (2000) 19505–19512.
[22] C. Deutsch, J.S. Taylor, M. Price, pH homeostasis in human
lymphocytes: modulation by ions and mitogen, J. Cell. Biol. 98
(1984) 885–893.
[23] S. He, R.P. Mason, S. Hunjan, V.D. Mehta, V. Arora, R. Katipally,
P.V. Kulkarni, P.P. Antich, Development of novel 19F NMR pH
indicators: synthesis and evaluation of a series of fluorinated vitamin
B6 analogues, Bioorg. Med. Chem. 6 (1998) 1631–1639.
[24] Y. Aoki, K. Akagi, Y. Tanaka, J. Kawai, M. Takahashi, Measure-
ment of intratumor pH by pH indicator used in 19F-magnetic
resonance spectroscopy. Measurement of extracellular pH decrease
caused by hyperthermia combined with hydralazine, Invest. Radiol.
31 (1996) 680–689.
[25] M.S. Gil, Imidazol-1-ylalkanoate esters and their corresponding
acids. A novel series of extrinsic 1H NMR probes for intracellular pH,
Bioorg. Med. Chem. Lett. 2 (1992) 1717–1722.
[26] R. van Sluis, Z.M. Bhujwalla, N. Raghunand, P. Ballesteros, J.
Alvarez, S. Cerdan, J.P. Galons, R.J. Gillies, In vivo imaging of
extracellular pH using 1H MRSI, Magn. Reson. Med. 41 (1999)
743–750.
[28] V.V. Khramtsov, I.A. Grigor’ev, M.A. Foster, D.J. Lurie, In vitro
and in vivo measurement of pH and thiols by EPR-based techniques,
Antioxid. Redox Signal. 6 (2004) 667–676.
[29] I.A. Kirilyuk, A.A. Bobko, V.V. Khramtsov, I.A. Grigor’ev, Nitrox-
ides with two pK values—useful spin probes for pH monitoring
within a broad range, Org. Biomol. Chem. 3 (2005) 1269–1274.
[30] I.A. Kirilyuk, A.A. Bobko, I.A. Grigor’ev, V.V. Khramtsov,
Synthesis of the tetraethyl substituted pH-sensitive nitroxides of
imidazole series with enhanced stability towards reduction, Org.
Biomol. Chem. 2 (2004) 1025–1030.
[31] B. Gallez, K. Ma¨der, H.M. Swartz, Noninvasive measurement of the
pH inside the gut by using pH-sensitive nitroxides. An in vivo EPR
study, Magn. Reson. Med. 36 (1996) 694–697.
[32] M.A. Foster, I.A. Grigor’ev, D.J. Lurie, V.V. Khramtsov, S. McCal-
lum, I. Panagiotelis, J.M. Hutchison, A. Koptioug, I. Nicholson, In
vivo detection of a pH-sensitive nitroxide in the rat stomach by low-
field ESR-based techniques, Magn. Reson. Med. 49 (2003) 558–567.
[33] J.M.S. Hutchison, A simple low-field ESR spectrometer, in: Interna-
tional Workshop on In Vivo EPR and Related Studies, Dartmouth
Medical School, Hanover, 1998, p. 35.
[34] G. Breit, I.I. Rabi, Measurement of nuclear spin, Phys. Rev. 38 (1931)
2082–2083.
[35] I. Nicholson, F.J. Robb, D.J. Lurie, Imaging paramagnetic species
using radiofrequency longitudinally detected ESR (LODESR imag-
ing), J. Magn. Reson. Ser. B 104 (1994) 284–288.
[36] D.J. Lurie, M.A. Foster, D. Yeung, J.M. Hutchison, Design,
construction and use of a large-sample field-cycled PEDRI imager,
Phys. Med. Biol. 43 (1998) 1877–1886.
[37] D.J. Lurie, I. Nicholson, J.R. Mallard, Low-field EPR measurements
by field-cycled dynamic nuclear polarisation, J. Magn. Res. 95 (1991)
405–409.
[27] V.V. Khramtsov, I.A. Grigor’ev, M.A. Foster, D.J. Lurie, I.
Nicholson, Biological applications of spin pH probes, Cell. Mol.
Biol. 46 (2000) 1361–1374.
[38] R.J. Simpson, T.J. Peters, Forms of soluble iron in mouse stomach
and duodenal lumen: significance for mucosal uptake, Br. J. Nutr. 63
(1990) 79–89.