Dendrimeric-containing nitronyl nitroxides as spin traps for nitric oxide: Synthesis, kinetic, and stability studies

Article abstract of DOI:10.1021/ma021292m

The in-vivo in-situ spin-trapping of NO^. with iron chelates is limited in its ability to localize NO^. at defined cellular compartments. Nitronyl nitroxides, such as 2-(4-carboxyphenyl)-4,4,5,5-tetramethylimidazoline 3-oxide 1-oxyl, have been found to react with NO^.. The resulting nitroxide, 2-(4-carboxyphenyl)-4,4,5,5-tetramethylimidazolidine 1-oxyl, exhibits an EPR spectrum that is characteristically distinct from the parent nitroxide. However, nitronyl nitroxides are unstable in biologic milieu. This limits their ability to in vivo identify NO^. at specific tissue compartments. Herein, we report on the preparation of a family of dendrimer-containing nitronyl nitroxides, which were developed to spin-trap NO^.. Once synthesized, we determined their rate of reaction with NO^. and their capacity to spin-trap NO^. and estimated the stability of dendrimer-linked nitronyl nitroxides under various experimental paradigms. For the largest of the dendrimeric nitronyl nitroxides prepared, the reaction with NO^. changed the EPR spectrum from one that was broad to that with a major component with narrow and defined spectral lines. These compounds promise to be excellent biological spin-trapping agents for NO^..

Full text of DOI:10.1021/ma021292m

Macromolecules 2003, 36, 1021-1027
1021
Dendrimeric-Containing Nitronyl Nitroxides as Spin Traps for Nitric
Oxide: Synthesis, Kinetic, and Stability Studies
Ger a ld M. Rosen ,*,†,‡,§ Su p a tr a P or a su p h a ta n a ,† P ei Tsa i,† Nich ola s P . Am bu los,
Va ler i E. Ga ltsev,§,# Ka zu h ir o Ich ik a w a ,§,#,$ a n d How a r d J . Ha lp er n §,#
Department of Pharmaceutical Sciences, University of Maryland School of Pharmacy, Baltimore,
Maryland 21201; Medical Biotechnology Center, University of Maryland Biotechnology Institute,
Baltimore, Maryland 21201; Center for Very Low-Frequency EPR Imaging for In Vivo Physiology,
University of Maryland School of Pharmacy, Baltimore, Maryland 21201, and The University of
Chicago, Chicago, Illinois 60637; Department of Microbiology and Immunology, University of
Maryland School of Medicine, Baltimore, Maryland 21201; and Department of Radiation and
Cellular Oncology, University of Chicago, Chicago, Illinois 60637
Received August 9, 2002; Revised Manuscript Received November 26, 2002
ABSTRACT: The in-vivo in-situ spin-trapping of NO• with iron chelates is limited in its ability to localize
NO• at defined cellular compartments. Nitronyl nitroxides, such as 2-(4-carboxyphenyl)-4,4,5,5-tetra-
methylimidazoline 3-oxide 1-oxyl, have been found to react with NO•. The resulting nitroxide, 2-(4-
carboxyphenyl)-4,4,5,5-tetramethylimidazolidine 1-oxyl, exhibits an EPR spectrum that is characteristi-
cally distinct from the parent nitroxide. However, nitronyl nitroxides are unstable in biologic milieu.
This limits their ability to in vivo identify NO• at specific tissue compartments. Herein, we report on the
preparation of a family of dendrimer-containing nitronyl nitroxides, which were developed to spin-trap
NO•. Once synthesized, we determined their rate of reaction with NO• and their capacity to spin-trap
NO• and estimated the stability of dendrimer-linked nitronyl nitroxides under various experimental
paradigms. For the largest of the dendrimeric nitronyl nitroxides prepared, the reaction with NO• changed
the EPR spectrum from one that was broad to that with a major component with narrow and defined
spectral lines. These compounds promise to be excellent biological spin-trapping agents for NO•.
In tr od u ction
resulting imino nitroxide, 2-(4-carboxyphenyl)-4,4,5,5-
tetramethylimidazolidine 1-oxyl [6], exhibits an EPR
spectrum characteristically unique to this latter nitrox-
ide.3,4 The chemical flexibility, offered by this class of
spin traps for NO•, allows detection of this free radical
at specific tissue sites. Excitement for these nitronyl
nitroxides must, however, be tempered by the rapidity
by which these spin-labels are bioreduced.5
Dendrimers with terminal functional groups have
attracted much interest as models to incorporate a
localized high concentration of a specific guest mol-
ecule.6 On the basis of previous reported syntheses of
nitroxide-linked dendrimers,7 we have prepared a family
of dendrimer-containing nitronyl nitroxides. With these
compounds, we measured the rate of spin-trapping NO•
and the capacity of this reaction. Further, we estimated
the stability of dendrimer-linked nitronyl nitroxides
under various experimental paradigms.
Nitric oxide (NO•) is a free radical with a remarkably
long lifetime in biological milieu.1 Although this dimin-
ished rate of reaction toward many targets leads to a
high degree of specificity, an essential prerequisite for
a biologic second messenger, this low reactivity presents
problems when attempting to identify this free radical
in animal models in real time. In fact, only in the past
several years has evidence surfaced, documenting the
in-vivo generation of this free radical.2 These in-vivo in-
situ spin-trapping/EPR spectroscopic experiments en-
tailed the reaction of NO•, secreted by activated murine
macrophages, with iron chelates,2 like the highly charged
ferro-di(N-methyl-D-glucamine dithiocarbamate) [1] or
the more hydrophilic ferro-di(N-(dithiocarboxy)sar-
cosine) [2] (Scheme 1). Despite the fact that these spin
traps have successfully detected this free radical in mice
in real time,2 we continue to search for spin traps that
may allow localization of NO• at defined sites. Further,
in-vivo toxicity of these iron chelates remains a serious
problem.
Exp er im en ta l Section
Reagen ts. Hypoxanthine was purchased from Sigma Chemi-
cal Co. (St. Louis, MO). Superoxide dismutase, catalase, and
xanthine oxidase were obtained from Boehringer Mannheim
(Indianapolis, IN). All other reagents were purchased from
commercial venders and were used without further purifica-
tion.
Recent studies, however, have suggested that nitronyl
nitroxides, such as 2-(4-carboxyphenyl)-4,4,5,5-tetra-
methylimidizoline 3-oxide 1-oxyl [5], may be a viable
alternative to iron chelates (Scheme 1).3,4 Nitronyl
nitroxide [5] reacts with NO• at reasonable rates.3a The
DAB-Am -4-Den dr im er -(NH-2-(4-car boxyph en yl)-4,4,5,5-
tetr a m eth ylim id a zolin e-1-oxyl 3-Oxid e [8]. To a solution
of 2-(4-carboxyphenyl)-4,4,5,5-tetramethylimidazoline-1-oxyl
3-oxide [5] (prepared as reported,3a 0.5 g, 1.8 mmol) and
N-hydroxysuccinimide (Aldrich Chemical Co., Milwaukee, WI,
0.21 g, 1.8 mmol) in CHCl3 (50 mL) was added 1,3-dicyclo-
hexylcarbodiimide (Aldrich, 0.4 g, 1.98 mmol). The reaction
mixture was stirred for 1 day at room temperature, filtered,
and evaporated in vacuo. The remaining residue was taken
up in CH2Cl2 (10 mL), filtered, and evaporated in vacuo to
† University of Maryland School of Pharmacy.
‡ University of Maryland Biotechnology Institute.
§ University of Maryland School of Pharmacy and The Univer-
sity of Chicago.
University of Maryland School of Medicine.
University of Chicago.
On leave from Kyushu University, Fukuoka, J apan.
#
$
* Corresponding author: Tel 410-706-0514; FAX 410-706-8184;
e-mail grosen@umaryland.edu.
10.1021/ma021292m CCC: $25.00 © 2003 American Chemical Society
Published on Web 01/18/2003

1022 Rosen et al.
Macromolecules, Vol. 36, No. 4, 2003
Sch em e 1
dryness. The residual oil was passed through a chromato-
graphic column containing silica gel (Aldrich, mesh 230-400)
and eluted with chloroform:acetone (94:6). 2-{4-[(2,5-Dioxo-1-
pyrrolidinyl)oxy]carboxyphenyl}-4,4,5,5-tetramethylimidazo-
line-1-oxyl 3-oxide [7] was isolated as a deep blue solid (0.48
g, 72%); mp 156-160 °C (with decomposition). IR (CHCl3):
NO• (fixed at 10-28 µM, final concentration). These solutions
were prepared by introducing NO• gas (99%, Matheson Gas
Products, Baltimore, MD) into deionized H2O, which had
previously been passed through a glass column containing
Chelex-100 ion-exchange resin (BioRad, Richmond, CA). The
NO• solution was mixed with oxyhemoglobin (30 µM, final
concentration) in the absence and presence of freshly prepared
nitronyl nitroxide [5], dinitroxide [9], and dendrimer-contain-
ing nitronyl nitroxides [8, 10, 11] (varied from 0 to 200 µM).
This reaction was conducted in sodium phosphate buffer (30
mM, pH 7.4). The concentration of NO• was estimated by its
reaction with oxyhemoglobin using an extinction coefficient
of 12 mM-1 cm-1 at 576 nm.8
1,3
1772, 1742, and 1698 (CdO), 1646 cm-1 (CdN-O). EPR: AN
) 8.2 G.
A solution of DAB-Am-4-N,N,N′,N′-tetrakis(3-aminopropyl)-
1,4-butanediamine (Aldrich, 0.22 g, 0.69 mmol) dissolved in
CH2Cl2 (5 mL) was added to a solution of [7] (1.26 g, 3.3 mmol)
dissolved in CH2Cl2 (50 mL). This reaction mixture was stirred
at room temperature for 1 day. After evaporation to dryness,
in vacuo, the remaining oil was purified by column chroma-
tography using silica gel (Aldrich, mesh 230-400) and chlo-
roform:acetone (94:6), which separated excess [7] from the
reaction mixture. Changing the effluent to methanol:ammo-
nium hydroxide (30%) (9:1), the desired dendrimer-containing
nitroxide [8] was afforded as an oil, which solidified upon
Sta bility of Nitr on yl Nitr oxid es [5], Din itr oxid e [9],
a n d Den d r im er -Con ta in in g Nitr on yl Nitr oxid es [8, 10,
11]. The stability of nitronyl nitroxide [5], dinitroxide [9], and
dendrimer-containing nitronyl nitroxides [8, 10, 11] in the
presence of superoxide (O2•-) (10 µM/min) and glutathione (100
µM) was evaluated and presented in Table 3. For a typical
experiment, dendrimer-linked nitronyl nitroxide [11] (5 µM,
placement under high vacuum (0.45 g, 48%). IR (CHCl3): 1647
cm-1 (CdN). AN
) 8.2 G with broadening (Figure 1b).
1,3
•-
final concentration) was added to a O2 generating system
Gen er a l P r oced u r e for th e P r ep a r a tion of Den d r im er -
(NH-2-(4-ca r boxyp h en yl)-4,4,5,5-tetr a m eth ylim ida zolin e-
1-oxyl 3-Oxid e)n . Following the general method described
above, other dendrimer-containing 2-(4-carboxyphenyl)-4,4,5,5-
tetramethylimidazoline-1-oxyl 3-oxides were prepared and
purified by flash column chromatography using silica gel (mesh
230-400) and different effluent systems: (a) nitronyl nitroxide
[9] from 1,4-diaminopropane (Aldrich), chloroform:methanol
(9:1); (b) dendrimer-linked nitroxide [10] from TREN (tris(2-
aminoethyl)amine) (Pressure Chemical Co., Pittsburgh, PA),
chloroform:methanol (8:2); (c) dendrimer-linked nitroxide [11]
from DAB-Am-8-(4,17-bis(3-aminopropyl)-8,13-bis[3-[bis(3-
aminopropyl)amino]propyl]-4,8,13,17-tetraazaeicosane-1,20-di-
amine, Aldrich), methanol:ammonium hydroxide (30%) (98:
2).
consisting of xanthine/xanthine oxidase, pH 7.4. Addition of
SOD (30 U/mL) confirmed that the loss of nitroxide was O2
•-
dependent. This reaction mixture was immediately added to
an EPR flat quartz cell and introduced into the cavity of an
EPR spectrometer (Varian Associates, model E-109, Palo Alto,
CA). EPR spectra were recorded at room temperature. Identi-
cal studies were conducted with nitronyl nitroxide [5], dini-
troxide [9], and dendrimer-containing nitronyl nitroxides [8,
10].
Su p er oxid e P r od u ction . Xanthine oxidase was added to
a sodium phosphate buffer, pH. 7.4, containing xanthine such
•-
that the initial rate of O2 generation was 10 µM/min. The
superoxide flux was estimated as the SOD-inhibitable reduc-
tion of ferricytochrome c, monitored at 550 nm spectrophoto-
metrically as previously described.9 Initial rate of O2 pro-
•-
Kin etics of Sp in -Tr a p p in g Nitr ic Oxid e by Nitr on yl
Nitr oxid es [5, 8-11]. The apparent rate constant for the spin-
trapping of NO• was estimated using anaerobic solutions of
duction was calculated using an extinction coefficient of 21
mM-1 cm-1
.

Macromolecules, Vol. 36, No. 4, 2003
Dendrimeric-Containing Nitronyl Nitroxides 1023
Ta ble 1. MALDI-TOF Ma ss Sp ectr om etr y of Nitr on yl
Nitr oxid e-Lin k ed Den d r im er s
average mass [g/mol]
nitronyl nitroxides molecular formula
calcd
measd
9
10
8
C
C
C
C
31H40N6O6
592.7
924.0
1353.6
2847.5
588.4
922.8
1352.3
2844.2
48H63N10O9
72H100N14O12
152H216N30O24
11
Ta ble 2. Ra te Con sta n ts for th e Sp in -Tr a p p in g of Nitr ic
Oxid e
rate constant,
spin trap
× 103 M-1 s-1
161 ( 48
experimental conditions
[5]a
anaerobic solution of NO• gas
in H2O added to sodium
phosphate buffer, pH 7.4
anaerobic solution of NO• gas
in H2O added to sodium
phosphate buffer, pH 7.4
anaerobic solution of NO• gas
in H2O added to sodium
phosphate buffer, pH 7.4
anaerobic solution of NO• gas
in H2O added to sodium
phosphate buffer, pH 7.4
anaerobic solution of NO• gas
in H2O/DMSO (0.1%) added to
sodium phosphate buffer, pH 7.4
sodium phosphate buffered
solutions of NO•, pH 7.4
[9]a
191 ( 48
179 ( 42
242 ( 16
40.2 ( 5.2
10.1
[10]a
[8]a
[11]a
[5]b
a
[5, 9, 8, 10, 11] ) 2-(4-carboxyphenyl)-4,4,5,5-tetramethylimi-
dazoline-1-oxyl 3-oxide, dinitroxide, and dendrimer-linked nitronyl
nitroxides; apparent rate constants using nitronyl nitroxides as
b
competitive inhibitors for oxyhemoglobin.12 [5] ) 2-(4-carboxy-
phenyl)-4,4,5,5-tetramethylimidazoline-1-oxyl 3-oxide; absolute
rate constants (data taken from ref 3a).
Ta ble 3. Sta bility of Den d r im er -Con ta in in g Nitr on yl
Nitr oxid es in th e P r esen ce of Red u ced Glu ta th ion e a n d
Su p er oxid ea
spin trap
half-life (min)a
half-life (min)b
[5]
[9]
[10]
[8]
[11]
20 ( 2.0
13 ( 2.5
stablec
9 ( 1.4
15 ( 1.8
10 ( 0.7
19 ( 1.1
37 ( 2.8
17 ( 2.3
182 ( 6.6
a
The half-life of dendrimer-containing nitronyl nitroxides (5
F igu r e 1. (a) Representative EPR spectra of nitronyl nitroxide
[5] (5 µM), dinitroxide [9] (5 µM), and dendrimer-linked
nitronyl nitroxides [10] (5 µM) in sodium phosphate buffer (30
mM, pH 7.4). After reaction with NO• (at 2 µM/min, from the
hydrolysis of SPER-NO), the EPR spectra of imidazolidinoxyls
[6, 12, 13] were obtained. EPR spectra were continually
recorded. EPR spectra shown were at 40 min, independent of
whether SPER-NO was added. Receiver gain for nitroxides [5,
6, 9, 10, 12, 13] was 4 × 103 using a Varian Associates model
E-109 spectrometer. (b) Representative EPR spectra of den-
drimer-linked nitroxides [8] (5 µM) in sodium phosphate buffer
(30 mM, pH 7.4). After reaction with NO• (at 2 µM/min, from
the hydrolysis of SPER-NO), the EPR spectrum of imidazoli-
dinoxyls [14] was recorded. Dendrimer-linked nitronyl nitrox-
ide [11] was dissolved in sodium phosphate buffer (30 mM,
pH 7.4 containing DMSO, 3.6% final concentration) prior to
reaction with NO•, at a flux of 2 µM/min from the hydrolysis
of SPER-NO. EPR spectra were continually recorded. EPR
spectra shown were 40 min, independent of whether SPER-
NO was added. Receiver gain for nitroxides [8, 14] was 4 ×
103, and for nitroxide [11, 15] it was 5 × 104. EPR spectra
were obtained using s Varian Associates model E-109 spec-
trometer.
µM) in the presence of reduced glutathione (100 µM), potassium
phosphate buffer, (30 mM, pH 7.4) was determined. In the absence
of glutathione, each of the nitronyl nitroxides were stable for at
least 80 min. Data presented was the average of three independent
b
experiments. The half-life of nitronyl nitroxides (5 µM) exposed
•-
to a O2 generating system consisting of xanthine and xanthine
•-
oxidase. The rate of O2 production was 10 µM/min. Data
presented was the average of three independent experiments.
c The term stable refers to the fact that the EPR spectrum for
dendrimer-linked nitronyl nitroxide [10] was unchanged for over
30 min.
zyl) was used as a marker of g value ) 2.0036. The sample
volume was adjusted to make the EPR signal height of DPPH
the same as a 5 µM aqueous solution of each of the nitronyl
nitroxides tested. DPPH was sandwiched between 5 mm
square pieces of Scotch tape, which in turn was taped to the
EPR quart flat cell. This cell was then fitted into the cavity of
the EPR spectrometer. Spectrometer parameters were micro-
wave power, 10 mW; modulation frequency, 100 kHz; modula-
tion amplitude, 1.0 G; response time, 0.3 s; sweep width, 100
G; and number of points, 512.
g Va lu e Deter m in a tion for Nitr on yl Nitr oxid es. All
EPR spectra (Varian Associates, model E-12, Palo Alto, CA)
were recorded at room temperature using a frequency counter
(Hewlett-Packard 5342A). DPPH (1,1′-diphenyl-2-picrylhydra-
Resu lts a n d Discu ssion
Syn th esis. In the preparation of dendrimer-contain-
ing nitronyl nitroxides, a substituent on the aromatic

1024 Rosen et al.
Macromolecules, Vol. 36, No. 4, 2003
Sch em e 2
using matrix-assisted laser desorption and ionization-
time-of-flight (MALDI-TOF) mass spectrometry using
2,4,6-trihydroxyacetophenone and R-cyano-4-hydroxy-
cinnamic acid as matrices (Table 1). In each case a small
deviation from calculated mass (m/z) values has been
found. Further, the MALDI-TOF mass spectra showed
fragmentation due to repetitive loss of oxygen.
EP R Sp ectr oscop y. The EPR spectrum of nitronyl
nitroxide [5] is composed of five lines in the ratio of
1:2:3:2:1, from the coupling of two identical nitrogen
atoms (nuclear spin I ) 1) to the electron spin (Figure
1a). The nitrogen isotropic hyperfine coupling constant
1,3
AN
) 8.2 G is about half of that found with simple
dialkyl nitroxides such as 3-carboxy-2,2,5,5-tetramethyl-
1-pyrrolinidinoxyl.3a,4 In the presence of NO•, nitronyl
nitroxide [5] was converted to the corresponding imi-
dazolidine 1-oxyl [6], as demonstrated by its character-
1
3
istic seven-line EPR spectrum, AN ) 9.8 G and AN
)
4.9 G (Figure 1a). The identity of the nitrogen atoms
has been lost by the elimination of the oxygen at one of
the nitrogen atoms.3a As the number of nitroxides linked
to dendrimers increased, from n ) 3 to n ) 4 and to n
) 8 active site attachments, there appeared a broadened
peak that overlaps the very visible five-line EPR spec-
trum of typical nitronyl nitroxides (Figure 1b). However,
only a small percent of the doubly integrated EPR
spectrum derives from the narrow component. In con-
trast, for dendrimer-linked pyrrolinidinoxyls, the nar-
row nitroxide triplet of typical pyrrolinidinoxyls with
an AN ) 15 G was lost by attachment of as few as four
of these nitroxides to the dendrimer.7 The g values for
dendrimer-linked nitroxides [10, 11] were 2.0066 and
2.0065, respectively.
ring of 2-phenyl-4,4,5,5-tetramethylimidazolidine 1-oxyl
was used to link the imidazolidine to various dendrim-
ers. For these initial studies, we activated 2-(4-carboxy-
phenyl)-4,4,5,5-tetramethylimidazoline-1-oxyl 3-oxide
[5] through its reaction with N-hydroxysuccinimide in
the presence of 1,3-dicyclohexylcarbodiimide (Scheme
2).7b The resulting activated ester [7] was combined with
poly(propyleneimine tetraamine) dendrimers (DAB-
Am)n where, for instance, n ) 4, yielding the corre-
sponding dendrimer-linked nitronyl nitroxide [8] (Scheme
2). In a similar fashion other dendrimer-containing
nitronyl nitroxides were prepared using tris(2-amino-
ethyl)amine and poly(propyleneimine octaamine) den-
drimer (DAB-Am)n where n ) 8 as well as 1,3-
diaminopropane (Scheme 3). The resulting dinitroxide
[9] and dendrimer-linked 2-(4-carboxyphenyl)-4,4,5,5-
tetramethylimidazoline-1-oxyl 3-oxides [8, 10, 11] were
purified by flash column chromatography using silica
gel and various combinations of polar solvents as
effluents. For the dinitroxide [9] and dendrimer-
containing nitronyl nitroxides [8, 10], the solvent
mixture of chloroform:methanol (9:1) and chloro-
form:methanol (8:2), respectively, was found to be ideal.
In the case of dendrimer-containing nitronyl nitroxide
[11], a more polar solvent mixture was required, metha-
nol:ammonium hydroxide (30%) (98:2) to isolate the
desired product, nitroxide [11]. The identification of the
dendrimer-containing nitronyl nitroxides was confirmed
Reaction of dinitroxide [9] and dendrimer-containing
nitronyl nitroxides [8, 10, 11] with NO• leads to the
corresponding imidazolidinoxyls [12-15] (Figure 1). Of
particular interest is the finding of a shift in the EPR
spectrum from a near exclusive broad-lined of that found
for dendrimer-containing nitroxide [11] to the seven-
lined imidiazolidininoxyl [15] with a substantial narrow
component. The narrow component seen in [15] repre-
sents as much as 25% of the EPR spectral intensity.
This increase in the narrow component can be qualita-
tively understood by noting that this change in the EPR
spectrum results from the loss of an oxygen atom from
the nitronyl nitroxide moiety of dendrimer-containing
nitroxide [11]. This should reduce the steric hindrance
of the imidiazolidininoxyls in [15] relative to the nitronyl
nitroxides in [11] and allow rapid spinning of the former,
thereby narrowing the EPR spectral lines. Work is
underway to better understand this phenomenon. This
narrowing of the EPR spectrum of imidazolidinoxyls will
make it easier to identify NO• at cellular fluxes, as has
been observed with LPS-stimulated RAW macrophages
Sch em e 3

Macromolecules, Vol. 36, No. 4, 2003
Dendrimeric-Containing Nitronyl Nitroxides 1025
using 2-(4-carboxyphenyl)-4,4,5,5-tetramethylimidizo-
line 3-oxide 1-oxyl [5].5b Further, we suggest that this
spectral property will be important in the in-vivo in-
situ measurement of NO• using low-frequency EPR
spectroscopy.
by fixing the concentration of nitronyl nitroxide [5],
while the concentration of [Hb(Fe2+)O2] was varied.
Here, the ratio of the peak height of the first line in the
EPR spectra of nitronyl nitroxide [5] and imino nitroxide
[6] was calculated. For dendrimer-linked nitroxides,
however, only spectrophotometric assays were used, as
the broadened EPR spectra of these compounds made
accurate estimates of peak heights difficult. From these
experimental designs, the apparent rate constants for
nitronyl nitroxide [5], dinitroxide [9], and dendrimer-
linked nitronyl nitroxides [8, 10, 11] were obtained and
are presented in Table 2. We were puzzled as to why
the rate constant for the spin-trapping of NO• by
dendrimer-containing nitronyl nitroxide [11] was sig-
nificantly different than those rate constants attained
for nitronyl nitroxide [5], dinitroxide [9], and dendrimer-
linked nitronyl nitroxides [8, 10]. All control experi-
ments, including those that examined the reaction of
NO• with DMSO, did not result in any insightful theory
as to why differing rate constants were found for the
reaction of NO• with dendrimer-linked nitronyl nitroxide
[11]. Nevertheless, we are currently exploring possible
explanations.
Kin etics. The rate constant for the reaction of NO•
with 2-(4-carboxyphenyl)-4,4,5,5-tetramethylimidizoline-
1-oxyl 3-oxide [5] is given in Table 2. In aerobic aqueous
solutions, NO• combines with O2 according to the
following:
NO• + 1/2O2 f NO2
NO2 + NO• f N2O3
(1)
(2)
(3)
N2O3 + H2O f 2H+ + 2NO2
-
in which -d[NO•]/dt ) k[O2] [NO•]2. Thus, the reaction
is second order in NO• and first order in O2.1,11 To
simplify the kinetic model, we used low concentrations
of NO• added to an anaerobic solution of nitronyl
nitroxide [5]. The concentration of nitronyl nitroxide [5]
was set sufficiently high to trap essentially all of the
NO• in the absence of oxyhemoglobin.8 In the presence
of oxyhemoglobin Hb(Fe2+)O2 there are two competing
reactions:
Sta bility. Nitronyl nitroxide [5], dinitroxide [9], and
dendrimer-linked nitronyl nitroxides [8, 10, 11] were
stable for at least 80 min in chelexed potassium
phosphate buffer (30 mM, pH 7.4) (Table 3). However,
after 24 h in the buffer, dendrimer-linked nitronyl
nitroxides [8, 10, 11] began to lose nitroxide groups from
the dendrimer, leading to an EPR spectrum character-
istic of nitronyl nitroxide [5]. This loss of nitroxides from
the dendrimer was greatest with dendrimer-linked
nitronyl nitroxide [11] at 43%, for dendrimer-linked
nitronyl nitroxide [8] at 22% and dendrimer-linked
nitronyl nitroxide [10] at 5%. Interestingly, the EPR
spectrum of dinitroxide [9] did not change over this 24
h period, suggesting, as in the case of nitronyl nitroxide
[5], that dinitroxide [9] is stable in phosphate buffer at
pH 7.4. While the mechanism remains undefined,
perhaps crowding of the nitroxides in the dendrimeric
molecule, which cannot occur with either nitronyl ni-
troxide [5] or dinitroxide [9], was responsible for the
degradation of dendrimer-linked nitronyl nitroxides [8,
10, 11]. Of importance from an experimental perspective
is the requirement that only freshly prepared aqueous
solutions of dendrimer-linked nitronyl nitroxides should
be used in studies measuring NO• production.
NO• + nitroxide [5] f nitroxide [6]
NO• + Hb(Fe2+)O2 f Hb(Fe3+) + NO3
(4)
(5)
-
Thus, from eqs 4 and 5, the rate of NO• elimination can
be expressed as
V ) -d[NO•]/dt ) kapp[5][NO•] +
kHb[Hb(Fe2+)O2][NO•] (6)
In the absence of nitronyl nitroxide [5], eq 6 can be
described as
v ) -d[NO•]/dt ) d[Hb(Fe3+)]/dt )
kHb[Hb(Fe2+)O2][NO•] (7)
By dividing eq 7 into eq 6 and rearranging the terms,
competing reactions can be represented as
Except for nitroxide [10], each of the nitroxides [5, 9,
8, 11] was unstable in the presence of reduced glu-
tathione (100 µM) (Table 3). Thus, in intracellular
environments where levels of glutathione are high (in
the millimolar range, depending on cell type) nitroxide
[10] would be the best of these dendrimer-linked nitro-
nyl nitroxides at detecting nitric oxide synthase pro-
duced NO•.
V/v ) 1 + kapp{nitroxide [5]}/kHb[Hb(Fe+2)O2] (8)
As the initial rate of spin-trapping is first-order with
respect to [NO•], kapp/kHb[Hb(Fe2+)O2] becomes the
constant, k′ at fixed [Hb(Fe2+)O2]. Then, eq 8 can be
simplified to
V/v ) 1 + k′{nitroxide[5]}
(9)
Purified nitric oxide synthase, independent of the
isozyme, has been reported to generate O2•-, even in the
presence of L-arginine.13 In light of this, it was impor-
tant to determine the stability of the dendrimer-linked
By plotting V/v vs {nitroxide[5]}, k′ can be determined.12b
From this, the kapp for the spin-trapping of NO• by
nitronyl nitroxide [5] was calculated (Table 2).
Several different experimental designs were used to
determine the apparent rate constant for spin-trapping
NO• by nitronyl nitroxide [5]. As described above, the
concentration of [Hb(Fe2+)O2] was held constant, and
the concentration of nitronyl nitroxide [5] was varied.
In this case, spectrophotometry was used to obtain
measurements by taking advantage of the absorbance
of [Hb(Fe2+)O2].8 We also used an alternative approach
•-
nitroxides in the presence of a continued flux of O2
(Table 3). For these experiments, the metabolism of
xanthine by xanthine oxidase was used as a source of
•-
O2 (generated at a rate of 10 µM/min) to which
nitroxides [5, 8-11] were subjected. Inclusion of SOD
(30 U/mL) confirmed that the loss of EPR spectra was
•-
attributed to the effects of O2 on each nitroxide. Not
unexpectedly, the stability of the dendrimer-linked

1026 Rosen et al.
Macromolecules, Vol. 36, No. 4, 2003
of dendrimer-containing nitronyl nitroxide [11] with
NO•, affording imidazolidinoxyl [15], resulted in the
EPR spectrum shown in Figure 1b with a much larger
fraction of the spectrum contained in the doubly inte-
grated intensity of the narrow lines, approximately 25%.
Initially, we suspected that this larger fraction might
be due to monomeric nitroxide that detached upon
treatment of dendrimer-linked nitronyl [11] with NO•.
To test this possibility, nitroxide [11] was incubated
with sufficient NO• so that the resultant EPR spectrum
corresponded to that of nitroxide [15]. Then, mass
spectral analysis of the reaction mixture was deter-
mined. To our surprise, we observed a mass spectral
line corresponding to imidazolidinoxyl [15]. This finding
suggests that the seven-line EPR spectrum typical of
these imidazolidinoxyls4 does not arise from hydrolysis
of dendrimer-attached nitronyl nitroxides. But rather,
we suspect that elimination of oxygen atoms from
dendrimer-linked nitronyl nitroxide [11], for instance,
after attack by NO• that resulted in NO2 , may free the
•
imidazolidinoxyl [15] to rotationally average and narrow
the EPR spectrum (Figure 1b, Scheme 3). This large
fraction of the spectral intensity due to a narrow
component is very favorable for spin-trapping NO• under
a variety of experimental paradigms, particularly for the
in-vivo detection of NO•. A large number of nitronyl
nitroxides can be delivered via a dendrimer to the in-
vivo site of NO• formation with very low spectral
amplitude per nitronyl nitroxide due to the dipolar
broadening. Spin-trapping of NO• then generates a
narrow and highly measurable seven-line EPR spectrum
corresponding to imidazolidinoxyl.
F igu r e 2. Plot of the EPR spectral peak height ratio vs
concentration of NO•. Nitric oxide was generated from SPER-
NO at a NO• flux of 1 µM/min, added to either nitronyl
nitroxide [5] (5 µM, open circles), dendrimer-linked nitronyl
nitroxide [8] (5 µM, solid circles), or nitronyl nitroxide [5] (20
µM, triangles).
nitroxides enhanced as the number of nitroxide groups
attached to the dendrimer was increased. Of interest
was the finding that dendrimer-linked nitronyl nitroxide
[10] was considerably more resistant to O2•--mediated
reduction than would be predicted based on dinitroxide
[9] and dendrimer-linked nitroxide [8]. Further, den-
drimer-linked nitroxide [11] exhibited a half-life of over
3 h. This finding was rather surprising and suggests
that nitroxide [11] may be completely resistant to
bioreduction in the harsh environment of activated
phagocytes, where NO• and O2•- are likewise produced.
NO• Sp in -Tr a p p in g Ca p a city. The ratio of the
conversion to dendrimer-containing nitronyl nitroxides
to imidazolidinoxyls was estimated as a function of NO•
concentration. For these experiments, the concentration
of nitroxides [5, 8-11] was fixed at 5 µM, and the
capacity to spin-trap NO• was estimated as a function
of NO• concentration. As the rate constants for the spin-
trapping of NO• are similar (Table 2), one would expect
differences in the capacity to reflect the number of
nitroxides attached to a specific dendrimer. To test this
theory, we compared the capacity of nitronyl nitroxide
[5] to dendrimer-linked nitronyl nitroxide [8]. At equal
molar concentrations, there was, as expected, approxi-
mately a 4-fold increase in NO• spin-trapping capacity
for dendrimer-linked nitroxide [8] as compared to nit-
ronyl nitroxide [5] (Figure 2). When the concentration
of nitroxide [5] was increased 4-fold, there was no
difference in the capacity to spin-trap NO• between
nitroxide [5] and dendrimer-linked nitroxide [8] (Figure
2).
Con clu sion
Herein, we report on the first synthesis of dendrimer-
linked nitronyl nitroxides as spin-traps for NO•. Unex-
pectedly, the unique EPR spectrum of nitronyl nitrox-
ides was still observable, although small, within the
broaden spectrum of these nitroxide-attached dendrim-
ers. As the number of nitronyl nitroxides attached to
the dendrimer increased, however, the singular five-line
EPR spectrum was less apparent, ∼ 1% spectral inten-
sity for dendrimer-linked nitronyl nitroxide [11]. Sur-
prisingly, after reaction with NO•, the seven-line EPR
spectrum of imidazolinoxyls was clearly visible, up to
25% spectral intensity for imidazolinoxyl [15]. The rate
constants for the spin-trapping of NO• by the dendrimer-
linked nitronyl nitroxides were similar to nitronyl
nitroxide [5], except for dendrimer-linked nitronyl ni-
troxide [11]. One possible explanation for the smaller
rate constant may be that as we increase the size of the
dendrimer, packing of the nitronyl nitroxide groups
retarded the rate of their reaction with NO•. We are
currently investigating this possibility. Finally, the
increased capacity of these dendrimer-linked nitronyl
nitroxides may make it possible to effectively detect NO•
in vivo and in real time. These studies are currently
ongoing.
EP R Sp ectr a of Den d r im er -Con ta in in g Nitr on yl
Nitr oxid es a n d Im id a zolid in oxyls. The EPR spec-
trum of dendrimer-containing nitronyl nitroxide [11]
shown in Figure 1b is consistent with less than 1%
spectral intensity contained in the narrow component.
While mass spectral data of [11] (Table 2) are supportive
of one compound, it is conceivable that the 1% may be
monomer, detached from the dendrimer during dissolu-
tion in sodium phosphate buffer (30 mM at pH 7.4,
containing DMSO 3.6%, final concentration). Reaction
Ack n ow led gm en t. This research was supported in
part by a grant from the National Institutes of Health,
RR12257, as part of the Center for EPR Imaging In Vivo
Physiology.
Refer en ces a n d Notes
(1) Kharitonov, V. G.; Sundquist, A. R.; Sharma, V. S. J . Biol.
Chem. 1994, 269, 5881.

Macromolecules, Vol. 36, No. 4, 2003
Dendrimeric-Containing Nitronyl Nitroxides 1027
(2) (a) Komarov, A.; Mattson, D.; J ones, M. M.; Singh, P. K.; Lai,
C.-S. Biochem. Biophys. Res. Commun. 1993, 195, 1191. (b)
Lai, C.-S.; Komoarov, A. M. FEBS Lett. 1994, 345, 120. (c)
Yoshimura, T.; Yokoyama, H.; Fujii, S.; Takayama, F.;
Oikawa, K.; Kamada, H. Nat. Biotechnol. 1996, 14, 992. (d)
Fujii, H.; Koscielniak, J .; Berliner, L. J . Magn. Reson. Med.
1997, 38, 565.
(3) (a) Akaike, T.; Yoshida, M.; Miyamoto, Y.; Sato, K.; Kohno,
M.; Sasamoto, K.; Miyazaki, K.; Ueda, S.; Maeda, H. Bio-
chemistry 1993, 32, 827. (b) J oseph, J .; Kalyanaraman, B.;
Hyde, J . S. Biochem. Biophys. Res. Commun. 1993, 192, 926.
(c) Woldman, Y. Y.; Khramtsov, V. V.; Grigor’ev, I. A.;
Kiriljuk, I. A.; Utepbergenov, D. I. Biochem. Biophys. Res.
Commun. 1994, 202, 195. (d) Konorev, E. A.; Tarpey, M. M.;
J oseph, J .; Baker, J . E.; Kalyanaraman, B. Free Radical Biol.
Med. 1995, 18, 169. (e) Pieper, G. M.; Siebeneich, W. J .
Pharmacol. Exp. Ther. 1997, 283, 138.
(4) Ullman, E. F.; Osiecki, J . H.; Boocock, D. G. B.; Darcy, R. J .
Am. Chem. Soc. 1972, 94, 7049.
(5) (a) Griffeth, L. K.; Rosen, G. M.; Rauckman. E. J .; Drayer,
D. P. Invest. Radiol. 1984, 19, 553. (b) Akaike, T.; Maeda, H.
Methods Enzymol. 1996, 268, 211.
(6) (a) Tomalia, D. A.; Naylor, M. A.; Goddard, W. A., III Angew.
Chem., Int. Ed. Engl. 1990, 29, 138. (b) Wo¨rner, C.; Mu¨lhaupt,
R. Angew. Chem., Int. Ed. Engl. 1993, 32, 1306. (c) J ansen,
J . F. G. A.; de Brabander-van den Berg, E. M. M.; Meijer, E.
W. Science 1994, 266, 1226. (d) Scherrenberg, R.; Coussens,
B.; van Vliet, P.; Edouard, G.; Brackman, J .; de Bradander,
E.; Mortensen, K. Macromolecules 1998, 31, 456. (e) Rietveld,
I. B.; Smit, J . A. M. Macromolecules 1999, 32, 4608.
(7) (a) Matyjaszewski, K.; Shigemoto, T.; Fre´chet, J . M.; Leduc,
M. Macromolecules 1996, 29, 4167. (b) Bosman, A. W.;
J anssen, R. A. J .; Meijer, E. W. Macromolecules 1997, 30,
3606.
(10) (a) A. Abragam. The Principles of Nuclear Magnetism;
Clarendon Press: Oxford, 1961; p 300. (b) Wyard, S. J . Proc.
Phys. Soc. 1965, 86, 587. (c) Frank, N. L.; Clerac, R.; Sutter,
J .-P.; Daro, N.; Kahn, O.; Coulon, C.; Green, M. T.; Golhen,
S.; Ouahab, L. J . Am. Chem. Soc. 2000, 122, 2053.
(11) (a) Fontecave, M.; Pierre, J .-P. Bull. Soc. Chim. Fr. 1994, 131,
620. (b) Goldstein, S.; Czapski, G. J . Am. Chem. Soc. 1995,
117, 12078.
(12) (a) Sawada, Y.; Yamazaki, I. Biochim. Biophys. Acta 1973,
327, 257. (b) Asada, K.; Takahashi, M.-A.; Nagate, M. Agric.
Biol. Chem. 1974, 38, 471. (c) Finkelstein, E.; Rosen, G. M.;
Rauckman, E. J . J . Am. Chem. Soc. 1980, 102, 4994. (d) Pou,
S.; Tsai, P.; Porasuphatana, S.; Halpern, H. J .; Chandramouli,
G. V. R.; Barth, E. D.; Rosen, G. M. Biochim. Biophys. Acta
1999, 1427, 216.
(13) (a) Pou, S.; Pou, W. S.; Bredt, D. S.; Snyder, S. H.; Rosen, G.
M. J . Biol. Chem. 1992, 267, 24173. (b) Wever, R. M. F.; van
Dam, T.; van Rijn, H. J . M.; de Groot, F.; Rabelink, T. J .
Biochem. Biophys. Res. Commun. 1997, 237, 340. (c) Xia, Y.;
Roman, L. J .; Masters, B. S. S.; Zweier, J . L. J . Biol. Chem.
1998, 273, 22635. (d) Va´squez-Vivar, J .; Kalyanaraman, B.;
Marta´sek, P.; Hogg, N.; Masters, B. S. S.; Karoui, H.; Tordo,
P.; Pritchard, K. A., J r. Proc. Natl. Acad. Sci. U.S.A. 1998,
95, 9220. (e) Xia, Y.; Tsai, A.-L.; Berka, V.; Zweier, J . L. J .
Biol. Chem. 1998, 273, 25804. (f) Stroes, E.; Hijmering, M.;
van Zandvoort, M.; Wever, R.; Rabelink, T. J .; van Faasen,
E. E. FEBS Lett. 1998, 438, 161. (g) Pou, S.; Keaton, L.;
Surichamorn, W.; Rosen, G. M. J . Biol. Chem. 1999, 274,
9573. (h) Va´squez-Vivar, J .; Hogg, N.; Marta´sek, P.; Karoui,
H.; Pritchard, K. A., J r.; Kalyanaraman, B. J . Biol. Chem.
1999, 274, 26736. (i) Yoneyama, H.; Yamamoto, A.; Kosaka,
H. Biochem. J . 2001, 360, 247. (j) Rosen, G. M.; Tsai, P.;
Weaver, J .; Porasuphatana, S.; Roman, L. J .; Starkov, A. A.;
Fiskum, G.; Pou, S. J . Biol. Chem. 2002, 277, 40275.
(8) Doyle; M. P.; Hoekstra, J . W. J . Inorg. Biochem. 1981, 14,
351.
(9) Kuthan, H.; Ullrich, V.; Estabrook, R. Biochem. J . 1982, 203,
551.
MA021292M