Synthesis and characterization of 5-hydroxymethyl-5-methyl-pyrroline N-oxide and its derivatives

Article abstract of DOI:10.1016/j.bmc.2010.11.052

Synthesis and spin trapping behavior of three novel DMPO derivatives, namely 5-hydroxymethyl-5-methyl-pyrroline N-oxide (HMMPO), 5-(2-furanyl)- oxymethyl-5-methyl-pyrroline N-oxide (FMMPO), and 5-(2-pyranyl)-oxymethyl-5- methyl-pyrroline N-oxide (PMMPO) towards different oxygen- and carbon-centered radicals are described. The stabilizing effect of a series of cyclodextrins on the superoxide adducts was tested.

Full text of DOI:10.1016/j.bmc.2010.11.052

Bioorganic & Medicinal Chemistry 19 (2011) 985–993
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Synthesis and characterization of 5-hydroxymethyl-5-methyl-pyrroline
N-oxide and its derivatives
Klaus Stolze a, , Natascha Rohr-Udilova b, Anjan Patel c, Thomas Rosenau c
⇑
a Molecular Pharmacology and Toxicology Unit, Department of Biomedical Sciences, University of Veterinary Medicine Vienna, Veterinärplatz 1, A-1210 Vienna, Austria
b Division of Gastroenterology and Hepatology, Department of Internal Medicine III, Medical University of Vienna, Währinger Gürtel 18-20, A-1090 Vienna, Austria
c Department of Chemistry, University of Natural Resources and Life Sciences (BOKU), Muthgasse 18, A-1190 Vienna, Austria
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 6 June 2010
Revised 17 November 2010
Accepted 20 November 2010
Available online 26 November 2010
Synthesis and spin trapping behavior of three novel DMPO derivatives, namely 5-hydroxymethyl-
5-methyl-pyrroline N-oxide (HMMPO), 5-(2-furanyl)-oxymethyl-5-methyl-pyrroline N-oxide (FMMPO),
and 5-(2-pyranyl)-oxymethyl-5-methyl-pyrroline N-oxide (PMMPO) towards different oxygen- and car-
bon-centered radicals are described. The stabilizing effect of a series of cyclodextrins on the superoxide
adducts was tested.
Ó 2010 Elsevier Ltd. All rights reserved.
Keywords:
EPR
Spin trapping
DMPO derivatives
Cyclodextrin
Superoxide
1. Introduction
Although the spin trapping properties of the final product,
HMMPO, must be expected to be not optimal for superoxide radi-
Previous studies about DEPMPO and EMPO derivatives1–7 were
focused on the stability of superoxide adducts as well as investiga-
tions of mitochondria targeted spin traps8–10 bearing positively
charged phosphonium groups. Based on these studies we were
aiming at synthesizing EMPO derivatives with a function to which
a mitochondria targeting group could be readily attached in a sin-
gle step. Unfortunately, the synthesis of 5-ethoxycarbonyl-
4-hydroxymethyl-5-methyl-pyrroline N-oxide (the analogue of
the previously reported 5-diethoxyphosphoryl-4-hydroxymethyl-
5-methyl-1-pyrroline N-oxide, 4-HMDEPMPO8,9) is not straightfor-
ward, and several attempts to synthesize this compound in pure
form have failed in our hands.11 We therefore decided to first de-
velop strategies involving different protective groups for the hy-
droxyl groups starting from simple related model compounds. In
this study we report the synthesis of 5-hydroxymethyl-5-methyl-
pyrroline N-oxide (HMMPO), a DMPO derivative having an extra
hydroxyl group as an anchor group where specific functionalities
can be attached thereby imposing a site-specific reactivity of the
trap. We decided to investigate the effects of tetrahydrofuranyl
(THF)12 or tetrahydropyranyl (THP)13 protecting groups on the syn-
thetic yields of the subsequent steps of the synthetic sequence.
cals, utilization of those protecting groups in combination with the
pyrrolidine N-oxide moiety seems to offer a valuable and versatile
building block for later modification with a series of different tar-
geting groups, for example, towards cationic spin traps which have
a specific selectivity for mitochondria.
2. Results
2.1. Structure of the spin traps
The spin traps reported within this study are derived from
the commercially available parent compound 5,5-dimethyl-
1-pyrroline-N-oxide (DMPO), to which an additional hydroxyl
substituent (unprotected or protected by a THF or THP group)
had been introduced to one of the methyl groups in position
5. This led to 5-hydroxymethyl-5-methyl-1-pyrroline-N-oxide
(HMMPO), 5-(2-furanyloxymethyl)-5-methyl-1-pyrroline-N-oxide
(FMMPO), and 5-(2-pyranyloxymethyl)-5-methyl-1-pyrroline-
N-oxide (PMMPO). The structures are shown in Figure 1a, the
synthetic steps in Figure 1b, and the spectroscopic data are
summarized in Tables 1–6 (NMR, MS, IR, UV–vis and EPR
parameters).
The structural identity of the novel spin traps was confirmed by
⇑
Corresponding author. Tel.: +43 1 25077 4406; fax: +43 1 25077 4490.
E-mail address: klaus.stolze@vetmeduni.ac.at (K. Stolze).
13C NMR (Table 1), 1H NMR (Table 2), MS and microanalysis
0968-0896/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2010.11.052

986
K. Stolze et al. / Bioorg. Med. Chem. 19 (2011) 985–993
a
b
HMMPO:
R
R
R
=
=
=
H
ROCH
2
+
N
O-
THF-HMMPO:
THP-HMMPO:
CH
3
O
O
O
n
n
i
ii
HO
O
NO2
O
O
O
NO2
NO2
n
iii
iv
O
HO
N+
O
N+
O
O
n=1: FMMPO
n=2: PMMPO
HMMPO
i: 2,3-dihydrofurane (or 3,4-dihydro-(2H)-pyrane), H+; ii: H2C=CH-CHO, MeCN, NEt3;
iii: NH4Cl / Zn, H2O/MeOH (v/v=6:4); iv: HCl
Figure 1. (a) General structure of the spin traps and (b) general synthetic scheme.
Table 1
13C NMR data [ppm] of the spin traps
4´
4´
3´
5´
6´
3´
2´
4
3
4
4
3
3
2´
5b
5a
5b
5a
5b
5a
5´
2
2
2
HO
1
N
O
1´
O
1
N
O
1´
O
1
N
5
5
5
O
O
O
0
0
0
0
0
2C
3C
4C
5C
5aC
5bC
2 C
3 C
4 C
5 C
6 C
EMPO3
HMMPO 137.5
FMMPO
134.9
25.4
25.3
31.9
28.3
78.5
76.5
20.3
20.5
169.3 (C@O) 61.7 (CH2)
13.4 (CH3)
—
—
—
—
—
67.1
—
—
—
66.0
—
134.7 & 135.2 25.3 & 25.4
29.2 & 29.6 76.1 & 76.4 21.5 &21.6
69.7 & 69.9
103.9 &104.1 23.3 & 23.5 32.4
PMMPO 134.1 & 134.7 25.55 & 25.6 29.6 & 29.9 76.2 & 76.5 21.7 & 21.8 70.2 & 70.9
98.2 & 99.7
30.6 & 30.8 19.1 & 20.0 25.62 61.6 & 63.0
Table 2
1H NMR data [ppm] of the spin traps
4´
3´
4´
3´
5´
6´
4
3
4
4
3
3
2´
2´
5b
5a
5b
5a
5b
5a
5´
2
2
2
HO
1
N
O
1´
O
1
N
O
1´
O
1
N
5
5
5
O
O
O
0
0
2CH
3CH2
4CH2
5aCH3
1.72 s
5BCH3
—
2 CH
3 CH2
4’CH2
—
5’CH2
—
6’CH2
—
OH
—
EMPO3
6.97 t 2.75 m
2.16 m
2.60 m
4.26 m (CH2) 1.31 t (CH3)
HMMPO 7.04 s 2.55–2.72 m 1.92–2.02 m 1.38 s
2.40–2.49 m
3.43 d
3.93 d
—
—
—
—
—
—
4.50 s, br
—
FMMPO
PMMPO
6.92 s 2.51–2.60 m 1.95–2.03 m 1.36 s & 1.38 s 3.23 d 5.13 m
1.75–1.87 m 1.83–1.88 m 3.83–3.92 m
1.90–1.96 m
6.95 s
2.40–2.48 m
3.49 d
3.87 d
4.04 d
6.90 s 2.57–2.70 m 1.96–2.06 m 1.38 s & 1.41 s 3.20 d 4.15 m
1.45–1.63 m 1.64–1.79 m 1.45–1.63 m 3.48–3.56 m
3.83–3.92 m 3.79–3.88 m
2.47–2.56 m
3.60 d
3.85 d
4.15 d

K. Stolze et al. / Bioorg. Med. Chem. 19 (2011) 985–993
987
Table 3
MS/microanalysis of the spin traps
[MH]+
Calcd [MH]+
C calcd
C found
H calcd
H found
N calcd
N found
HMMPO
FMMPO
PMMPO
130.17
200.26
214.28
130.17
200.26
214.29
55.80
60.28
61.95
55.72
60.19
61.88
8.58
8.60
8.98
8.72
8.73
9.17
10.84
7.03
6.57
10.88
6.98
6.39
Table 4
IR data (cmꢀ1) of the spin traps4
EMPOa
DMPO
2985 2940 2874
1464 1446 1377 1341 1288
—
—
—
1024
—
950 926 862 796
—
—
1741 1582
2933 2872 1662
1236 1182
1232 1144
1107
3086
1458 1367 1344 1271
1117 1026 935 824 779 714 690 635 584 507
2974
2980
1583
2902 2872 1670
HMMPO
1541 1456 1383 1346 1294 1124 1024 976 945 779 725 698 652 586 500
3307
2927
1597
1215 1196
1065
FMMPO 2980 2949 2879 1659 1589 1458 1375 1346 1296
1196
970 945 920 860 781 696 654 592 509
1225
1101 1047
1124 1070 1036
PMMPO
2872 2850 1651 1585 1456 1377 1348 1323 1228 1200
974 905 870 814 779 694 646 594 560 500
2947 2916
Intensities: strong (1741), medium (1464), weak (950).
a
Data from Stolze et al. (2003).4
Table 5
Half-life of the superoxide adducts, n-octanol/buffer partition coefficients, and UV–vis spectroscopic data of the spin traps
Compound
Apparent t1/2 (min)
partition coefficient n-octanol/
phosphate buffer (100 mM, pH 7.0)
UV data kmax and molar
absorptivity
e )
(l molꢀ1 cmꢀ1
DMPO
0.835 (—)1
0.13
—
0.91 (
1.53 (b-CD)
0.75 ( -CD)
a-CD)
c
1.65 (2,6-DM-b-CD)
EMPO
8.6 (—)3
0.15
231 nm, 7417 46
228 nm, 6123 54
228 nm, 7015 46
228 nm, 6666 26
17.8 (
53.3 (b-CD)
26.1 ( -CD)
a-CD)
c
66.1 (2,6-DM-b-CD)
3.53 (—)
HMMPO
FMMPO
PMMPO
0.041 0.001
0.347 0.006
0.845 0.002
4.71 (
7.82 (b-CD)
4.02 ( -CD)
a-CD)
c
4.42 (2,6-DM-b-CD)
3.43 (—)
3.72 (
4.70 (b-CD)
4.57 ( -CD)
a-CD)
c
5.76 (2,6-DM-b-CD)
4.76 (—)
4.42 (
6.03 (b-CD)
5.08 ( -CD)
6.73 (2,6-DM-b-CD)
a-CD)
c
(Table 3), FT-IR spectroscopy (Table 4) and UV–vis spectroscopy
(Table 5).
2-furanyl and 2-pyranyl protecting groups agree with the
expectations.
A complete set of 1H, H–H correlated, 13C, HMQC and HMBC
spectra was recorded for each compound, which allowed for a
complete signal assignment in both the 1H and 13C domains.
13C NMR confirmed the presence of a heterocyclic pyrroline
ring. The resonances for C-2 were found between 134.1 and
137.5 ppm, comparable to EMPO and its derivatives (around
134.9 ppm). The C-3 resonances (signal at about 25 ppm) were also
comparable to EMPO. A slight upfield shift by about 2 ppm was
seen for the neighboring C-4 (approx. 28–30 ppm, vs 31.9 ppm
for EMPO), and for C-5 (approx. 76.1–76.5 ppm, vs 78.5 ppm for
EMPO). The 5a-substituent, resonating at 20.3 ppm in EMPO, was
shifted downfield by about 1 ppm, being found around 22 ppm
for the derivatives described herein. The carboxylic carbon in
EMPO (resonating at 169.3 ppm) was replaced by a 5b-hydroxy-
methyl group in HMMPO (66 ppm), and 5b-alkoxymethyl groups
in FMMPO and PMMPO (approx. 70 ppm). The resonances of the
The proton spectra showed some interesting features. H-2
(around 7 ppm) appeared as a singlet, in contrast to EMPO and
most of its derivatives, for which the expected triplet was found
that indicated similar coupling constants to the geminal H-3 pro-
tons. The observed singlet indicates that the coupling constants
to both H-3 are close to zero.
The resonance of the 4-methylene and the 5b-oxymethylene
group show a strong enantiotopic/diastereotopic splitting of about
0.5 ppm.
The major absorption peaks in the respective FT-IR spectra were
seen around 1580 cmꢀ1 (C@N) and 1205 cmꢀ1 (N–O).
2.2. Superoxide radical adducts
Previously reported methods for the generation of superoxide
adducts were applied in our study.2–7,18,19 However, the low stabil-

988
K. Stolze et al. / Bioorg. Med. Chem. 19 (2011) 985–993
ity of superoxide adducts of HMMPO, FMMPO, and PMMPO ren-
dered the detection of the expected ESR signal rather difficult. In
addition, several secondary products were also formed, one of
them being the hydroxyl radical adduct also seen in the case of
commercially available DMPO. Half-lives of the superoxide spin
adducts showed no significant advantage over the spin trap DMPO.
In order to increase the stability we decided to employ the forma-
tion of stable cyclodextrin complexes which has recently been re-
ported by several groups.14–17 Since it was not known which of the
commercially available cyclodextrins and cyclodextrin derivatives
would be best suitable for spin adduct stabilization, a set of 4 com-
a
b
c
pounds, namely
a-, b-, c-cyclodextrin, and heptakis-(2,6-di-O-
methyl)-b-cyclodextrin, was tested. The stabilizing effects of the
different cyclodextrins are shown in Figs. 2–4. For comparison, also
cucurbit[6]uril, a cavitand build of methylene-bridged glycoluril
d
e
a
b
c
d
20 G
Figure 3. Superoxide radical spin adducts formed from FMMPO in a hypoxanthine/
xanthine oxidase system stabilized by different cyclodextrin derivatives. FMMPO
(8 mM) was dissolved in oxygenated phosphate buffer (100 mM, pH 7.4, containing
0.4 mM DTPA), catalase (250 U/mL), hypoxanthine (0.2 mM) and xanthine oxidase
(500 mU/mL). (a) Without cyclodextrin added (rel. int.: 21), (b)
(16 mM, rel. int.: 29), (c) b-cyclodextrin (16 mM, rel. int.: 70), (d)
a
c
-cyclodextrin
-cyclodextrin
(16 mM, rel. int.: 43), (e) heptakis-(di-O-methyl-)-b-cyclodextrin (16 mM, rel. int.:
55). After 15 min incubation maximum intensity was reached. SOD (150 units/mL)
was added and a series of consecutive ESR spectra (every 90 s for 30 min) was
recorded using the following EPR parameters: sweep width, 80 G; modulation
amplitude, 0.83 G; microwave power, 20 mW; time constant, 0.08 s; receiver gain,
2 ꢂ 105; scan rate, 57 G/min.
e
f
units, was tested, but did not lead to a stabilization of any of the
investigated superoxide adducts (spectra not shown).
Superoxide radicals were generated using a hypoxanthine
(0.2 mM)/xanthine oxidase (500 mU/ml) system in the presence
of the spin trap (8 mM) and a 2-fold excess of the respective cyclo-
dextrin (16 mM) in 10 mM oxygenated phosphate buffer pH 7.4,
containing 1 mM DTPA. When the maximum ESR intensity was ob-
tained (ranging from ca. 1 min for DMPO to 15 min for EMPO) SOD
(ca. 150 U/ml) and catalase (250 U/mL) were added to stop super-
oxide production. Decay kinetics of the superoxide adducts were
calculated as follows: A series of repetitive scans was recorded
for each incubation (every 45 s over 15 min for DMPO, every 90 s
over 30 min for the other compounds) using a first-order exponen-
tial decay approximation (Table 5, Pearson correlation coefficient
R2 > 0.99).
The resulting ESR spectra obtained with 8 mM HMMPO alone
(no cyclodextrin added) are shown in Figure 2a (five scans accumu-
lated from 0 to 6 min after mixing) and in Figure 2b (recorded from
7.5 to 15 min after mixing). Both ESR spectra show a mixture of
several paramagnetic species with variable intensities. The EPR
lines of the rapidly decaying HMMPO/ꢁOOH spin adduct became
visible only in the form of a difference spectrum (Fig. 2c, corre-
g
20 G
Figure 2. Superoxide radical spin adducts formed from HMMPO in a hypoxanthine/
xanthine oxidase system stabilized by different cyclodextrin derivatives. HMMPO
(8 mM) was dissolved in oxygenated phosphate buffer (100 mM, pH 7.4, containing
0.4 mM DTPA), catalase (250 U/mL), hypoxanthine (0.2 mM) and xanthine oxidase
(500 mU/mL). After 15 min incubation maximum intensity was reached, SOD
(150 units/mL) was added and a series of consecutive ESR spectra (every 90 s for
30 min) was recorded using the following EPR parameters: sweep width, 80 G;
modulation amplitude, 1.17 G; microwave power, 20 mW; time constant, 0.16 s;
receiver gain, 2 ꢂ 105; scan rate, 57 G/min. (a) Scans accumulated from 0 to 6 min
after mixing (relative intensity: 47). (b) Scans accumulated from 7.5 to 15 min (rel.
int.: 60). (c) Difference spectrum: Fig. 2a—0.9 ꢂ Fig. 2b (rel. int.: 18). Under
otherwise identical conditions as in (a), the following cyclodextrins were added: (d)
+
+
a
c
-Cyclodextrin (16 mM, rel. int.: 25). (e) + b-Cyclodextrin (16 mM, rel. int.: 32). (f)
-Cyclodextrin (16 mM, rel. int.: 27). (g) + Heptakis-(di-O-methyl-)-b-cyclodex-
trin (16 mM, rel. int.: 41).

K. Stolze et al. / Bioorg. Med. Chem. 19 (2011) 985–993
989
The superoxide adducts obtained with PMPPO (8 mM) are
shown in Figure 4a–e (without cyclodextrin, and with 16 mM of
a
-cyclodextrin, b-cyclodextrin, c-cyclodextrin, and heptakis-(2,6-
a
b
di-O-methyl)-b-cyclodextrin, respectively). The relative intensities
of the EPR spectra are listed in the figure legends.
2.3. Hydroxyl radical adducts
In order to form the respective hydroxyl radical adducts, the
spin traps (40 mM) were incubated with an aqueous Fenton sys-
tem20,21 (H2O2 (0.2%), EDTA (2 mM), and iron(II) sulfate (1 mM)).
After 10 s the reaction mixture was diluted 1:1 with phosphate
buffer (300 mM, pH 7.4; 20 mM DTPA), and the EPR spectra were
recorded immediately (Fig. 5). Only one major compound was
found for HMMPO/ÅOH (Fig. 5a), whereas two different species
were detected with FMMPO/ÅOH (ratio 58%/42%; Fig. 5b) and
PMMPO/ÅOH (ratio 64%/36%; Fig. 5c). The secondary species is a
carbon-centered radical of unknown structure, most probably de-
rived from degradation of the spin trap or spin adduct. The respec-
tive HFS data, obtained by computer simulation, are listed in
Table 6.
c
d
e
2.4. Methyl radical adducts
In analogy to our previous investigations with EMPO deriva-
tives,3–7 we synthesized the respective methyl radical adducts
using an aqueous Fenton system in the presence of 10% DMSO.20
Also here the reaction mixture was diluted 1:1 with phosphate
20 G
Figure 4. Superoxide radical spin adducts formed from PMMPO in a hypoxanthine/
xanthine oxidase system stabilized by different cyclodextrin derivatives. PMMPO
(8 mM) was dissolved in oxygenated phosphate buffer (100 mM, pH 7.4, containing
0.4 mM DTPA), catalase (250 U/mL), hypoxanthine (0.2 mM) and xanthine oxidase
a
(500 mU/mL). (a) Without cyclodextrin added (relative intensity: 29), (b)
cyclodextrin (16 mM, rel. int.: 35), (c) b-cyclodextrin (16 mM, rel. int.: 55), (d)
a
c
-
-
cyclodextrin (16 mM, rel. int.: 43), (e) heptakis-(di-O-methyl-)-b-cyclodextrin
(16 mM, rel. int.: 50). After 15 min incubation maximum intensity was reached.
SOD (150 units/mL) was added and a series of consecutive ESR spectra (every 90 s
for 30 min) was recorded using the following EPR parameters: sweep width, 80 G;
modulation amplitude, 0.83 G; microwave power, 20 mW; time constant, 0.08 s;
receiver gain, 2 ꢂ 105; scan rate, 57 G/min.
b
sponding to 2a—0.9 ꢂ 2b), indicating that the intensity of the sec-
ondary products changed only gradually during the first 15 min.
The respective EPR spectra recorded in the presence of 16 mM
a-cyclodextrin (Fig. 2d), 16 mM b-cyclodextrin (Fig. 2e), 16 mM
c-cyclodextrin (Fig. 2f), and 16 mM heptakis-(2,6-di-O-methyl)-b-
cyclodextrin (Fig. 2g) were obtained in the same way, as difference
spectra as discussed above (Fig. 2c). The shape of the ESR spectra
strongly varies with the complexing agent used. Due to the low
stability of the non-complexed superoxide adducts and the limited
solubility of b-cyclodextrin in buffer, concentration-dependent
studies of superoxide half lives and HFS parameters were not pos-
sible. Therefore, a constant 2:1 ratio between the spin trap and the
cyclodextrin was chosen, the concentration of cyclodextrin being
16 mM, which was the highest concentration of b-cyclodextrin in
aqueous buffer achievable.
c
20 G
The stability of the superoxide adducts of FMMPO and PMMPO
(Figs. 3 and 4) were significantly higher than those with HMMPO,
the EPR spectra could therefore be detected directly, without calcu-
lation of difference spectra. All experiments were performed with
8 mM of FMPPO with no cyclodextrin added (Fig. 3a), and with
16 mM of a-cyclodextrin (Fig 3b), b-cyclodextrin (Fig. 3c), c-cyclo-
dextrin (Fig. 3d), and heptakis-(2,6-di-O-methyl)-b-cyclodextrin
(Fig. 3e), respectively.
Figure 5. Iron-dependent formation of hydroxyl radical spin adducts from HMMPO,
FMMPO, and PMMPO. (a) HMMPO (40 mM, initial concentration) was incubated
with a Fenton system containing FeSO4 (1 mM), EDTA (2 mM), H2O2 (0.2%). The
reaction was stopped after 10 s by 1:1 dilution with phosphate buffer (300 mM, pH
7.4, containing 20 mM DTPA) and the spectrum was recorded using the following
spectrometer settings: sweep width, 80 G; modulation amplitude, 0.21 G; micro-
wave power, 20 mW; time constant, 0.08 s; receiver gain, 1 ꢂ 104; scan rate, 57 G/
min (rel. int.: 45), (b) same as in (a), except that FMMPO was used (rel. int.: 33), (c)
same as in (a), except that PMMPO was used (gain 2.5 ꢂ 104, rel. int.: 37).

990
K. Stolze et al. / Bioorg. Med. Chem. 19 (2011) 985–993
Table 6
Comparison of the EPR parameters of radical adducts of EMPO, DMPO, HMMPO, FMMPO, and PMMPO.
Radical
ÅOOH
HFS (G)
EMPOa
DMPO
HMMPO
FMMPO
PMMPO
transb
cisb
(6%)
13.25
10.15
1.50
(ꢀ)
aN
(94%)c
(100%)
14.30
11.70
1.25
(62%)
(38%)
13.70
10.23
1.00
(35%)
(65%)
13.68
10.15
(0.80)
(45%)
(55%)
13.75
13.75
0.80
13.27 13.25
12.37 9.30
13.55
12.17
0.55
13.69
12.60
—
13.73
12.73
0.55
aH
aH
(55.5%/44.5% of of transc)
ÅOOH
OOH
.OOH
ÅOOH
(
a
-CD)
(38%)
13.22
11.90
—
(62%)
13.22
9.66
—
(54%)
14.15
12.55
—
(46%)
14.15
10.05
—
(34%)
15.15
12.31
0.55
(66%)
14.49
10.20
1.00
(35%)
13.70
12.60
—
(65%)
13.70
10.18
0.80
(49%)
13.73
12.40
0.55
(51%)
13.73
9.89
aN
aH
aH
0.70
(b-CD)
aN
(61%)
12.73
12.00
—
(39%)
12.73
9.82
—
(50%)
13.57
12.27
—
(50%)
13.58
9.84
—
(52%)
13.65
12.00
0.55
(48%)
13.60
9.70
(53%)
13.35
12.10
(0.55)
(47%)
13.30
9.55
(59%)
13.10
11.80
0.55
(41%)
13.10
9.05
aH
aH
0.70
(0.70)
0.70
(2,6-DMBCD)
(50%)
12.56
11.95
—
(50%)
12.56
9.75
—
(48%)
13.50
12.40
—
(52%)
13.52
9.95
—
(44%)
13.30
12.27
0.55
(56%)
13.26
9.74
(27%)
13.00
12.35
0.85
(73%)
13.10
9.50
(42%)
13.05
11.95
0.55
(58%)
12.90
9.65
aN
aH
aH
0.95
0.90
0.90
(
c
-CD)
(55%)
13.16
11.81
—
(45%)
13.17
9.51
—
(53%)
14.06
12.48
—
(47%)
14.02
10.10
—
(40%)
13.54
12.25
0.55
(60%)
13.55
9.95
(32%)
13.50
12.60
—
(68%)
13.55
10.04
1.00
(32%)
13.20
12.60
0.55
(68%)
13.57
10.35
1.15
aN
aH
aH
0.93
EMPO
DMPO
HMMPO
FMMPO
PMMPO
transb
cisb
ÅOH
(76%)
14.11
12.80
0.63
(24%)
14.18
15.27
0.62
(100%)
14.80
14.80
(100%)
14.44
14.98
(58%)
14.60
14.67
(42%)f
15.46
22.67
(64%)
14.62
14.73
(36%)f
15.49
22.85
aN
aH
aH
aH
aH
aH
0.43
0.50
0.21(3)
0.13(2)
0.29(3)
0.07(2)
ÅH
aN
aN
aH
aH
(100%)
15.52
22.21
20.82
(100%)
16.00
21.50
21.50
(100%)
16.00
21.20
22.85
(100%)
16.00
21.40
22.70
(100%)
16.00
21.31
22.80
ÅCH3
ÅOCH3
(100%)
15.42
22.30
—
(100%)
16.56
23.68
(100%)
15.79
23.52
(100%)
15.83
23.30
(100%)
15.76
23.46
aN
aH
aH
—
(50%)
13.74
10.87
—
(50%))
13.74
7.81
—
(100%)
13.58
7.61
—
—
—
—
(100%)e
14.1
9.9
(100%)e
14.0
10.1
aN
aH
aH
1.85 (organic solvent)
1.4
1.4
ÅCH2OH
(100%)
14.95
21.25
—
(67%)
14.94
20.82
(100%)
16.20
23.00
—
(100%)
16.00
23.00
(100%)
15.36
22.40
—
(100%)
15.33
22.87
(100%)
15.38
22.00
—
(100%)
15.39
22.57
(100%)
15.39
22.07
—
(100%)
15.40
22.55
aN
aH
aH
ÅCH(OH)CH3
(33%)
15.00
22.40
aN
aH
ꢀ
ꢁCO2
(100%)
14.74
17.16
—
(100%)
15.80
18.80
(70%)
15.15
19.02
—
(30%)
14.49
14.94
0.50
(58%)
15.18
18.70
—
(42%)
14.61
14.61
0.60
(90%)
15.22
18.78
—
(10%)
14.64
14.64
0.60
aN
aH
—
a
b
c
Data from Stolze et al.3,7, except for b and c.
Data from Culcasi et al.24
Mixture of rotamers.
Approximate values, obtained from difference spectra.
Carbon-centered radical adduct.
d
e
buffer (300 mM, pH 7.4; 20 mM DTPA) after brief incubation. Due
to the high line width the respective EPR spectra showed the pres-
ence of only one major component of HMMPO/ÅCH3 (Fig. 6a),
PMMPO/ÅCH3 (Fig. 6b) and PMMPO/ÅCH3 (Fig. 6c).
The EPR spectra of the ÅCH2OH radical adducts of HMMPO,
FMMPO, and PMMPO can be seen in Figure 7a–c, respectively.
EPR parameters for the CH(CH3)OH radical adducts are listed in
Table 6. Due to the presence of more than one chiral center, dif-
ferent diastereomeric forms of the spin adducts can be expected.
Because of the high line width these diasteromeric forms can,
however, not be resolved. The lines of low intensity in between
most probably are traces of the respective hydroxyl and methoxyl
radical adducts.
Å
2.5. Methanol and ethanol-derived radical adducts
The above-mentioned Fenton-type incubation system20,21 was
used, except that methanol or ethanol was used instead of DMSO.

K. Stolze et al. / Bioorg. Med. Chem. 19 (2011) 985–993
991
a
b
c
a
b
c
20 G
Figure 7. Iron-dependent formation of hydroxymethyl radical spin adducts from
HMMPO, FMMPO, and PMMPO in the presence of methanol. (a) HMMPO (40 mM,
initial concentration) was incubated with
a Fenton system containing FeSO4
Figure 6. Iron-dependent formation of methyl radical spin adducts from HMMPO,
FMMPO, and PMMPO in the presence of DMSO. (a) HMMPO (40 mM, initial
concentration) was incubated with a Fenton system containing FeSO4 (1 mM),
EDTA (2 mM), H2O2 (0.2%) in 20% DMSO. The reaction was stopped after 10 s by 1:1
dilution with phosphate buffer (300 mM, pH 7.4, containing 20 mM DTPA) and the
spectrum was recorded using the following spectrometer settings: sweep width, 80
G; modulation amplitude, 0.21 G; microwave power, 20 mW; time constant, 0.08 s;
receiver gain, 1 ꢂ 104; scan rate, 57 G/min (rel. int.: 11), (b) same as in (a), except
that FMMPO was used (rel. int.: 12), (c) same as in (a), except that PMMPO was used
(gain 2.5 ꢂ 104, rel. int.: 9).
(1 mM), EDTA (2 mM), H2O2 (0.2%) in 20% methanol. The reaction was stopped after
10 s by 1:1 dilution with phosphate buffer (300 mM, pH 7.4, containing 20 mM
DTPA) and the spectrum was recorded using the following spectrometer settings:
sweep width, 80 G; modulation amplitude, 0.21 G; microwave power, 20 mW; time
constant, 0.08 s; receiver gain, 1 ꢂ 104; scan rate, 57 G/min (rel. int.: 19), (b) same
as in (a), except that FMMPO was used (rel. int.: 22), (c) same as in (a), except that
PMMPO was used (gain 2.5 ꢂ 104, rel. int.: 52).
observation of the respective Hꢁ adducts with all three compounds
(spectra not shown, EPR parameters listed in Table 6).
2.6. Other spin adducts formed
3. Discussion
Similarly, the use of an aqueous sodium formate solution
(200 mM) under otherwise identical conditions lead to the forma-
tion of the respective carbon dioxide anion radical adducts.20,21 Tꢀhe
characteristic ESR spectra are shown in Figure 8a (HMMPO/ꢁCO2 ),
Figure 8b (FMMPO/ꢁCO2ꢀ), and Fig. 8c (PMMPO/ꢁCOꢀ2 ). Two clearly
different diastereomeric forms could be distinguished with all spin
traps. The HFS data are listed in Table 6. We also tried to synthesize
the respective methoxyl radical adducts according to a procedure
previously published by Dikalov et al.22 In our previous studies
using a series of EMPO derivatives,3–7 transient ꢁOCH3 adducts
could be detected from most of the spin traps investigated. With
HMMPO only the hydroxymethyl adduct (see above, Fig. 7) was
seen, FMMPO and PMMPO formed the respective hydroxymethyl
adduct as the major species with traces of a very short lived meth-
oxyl radical adduct, detectable only as poorly resolved difference
spectra (not shown). Approximate HFS constants have been calcu-
lated and added to Table 6.
Three novel spin traps that can be regarded as DMPO deriva-
tives with a hydroxymethyl or alkoxymethyl group in position 5
of the pyrroline ring were synthesized in this study and were com-
pared with the parent compound DMPO as well as with similar
spin traps, such as EMPO and its derivatives.3–7 The chemical struc-
ture and purity of all compounds was comprehensively supported
by full NMR assignment (1H and 13C), MS, microanalysis, FT-IR, and
UV–vis spectroscopy (Tables 1–5).
The presence of functional groups in spin traps, such as hydro-
xyl or amino functionalities, allows for structural modification un-
der mild conditions so as to direct the spin trap to specific areas or
organells. The presence of two chiral carbon centers leads to the
formation of two different diastereomers of the investigated spin
traps. Addition of radicals from two different sides eventually leads
to four different stereoisomers, which have similar EPR parameters
and can, therefore, not always be clearly distinguished. In addition,
formation of secondary products due to side reactions and degra-
dation of spin adducts complicates the situation even further.
Unfortunately, the stabilities of the respective superoxide
adducts are rather low, that is, similar to the parent compound
DMPO or slightly higher, but significantly lower than the structur-
ally related EMPO derivatives. On the other hand, previous studies
with hydroxyl-substituted DEPMPO compounds revealed higher
Reduction of spin traps with KBH4 usually leads to the forma-
tion of pseudo-Hꢁ adducts. This was also confirmed for the three
new spin traps in the present study: Addition of a small amount
of KBH4 (ca. 0.5 mg/500 ll) to the respective spin trap (40 mM),
followed by readjusting of the pH to 7.4 by 1:1 dilution with phos-
phate buffer (300 mM, pH 7.4, containing 20 mM DTPA) led to the

992
K. Stolze et al. / Bioorg. Med. Chem. 19 (2011) 985–993
adducts in all cases. In contrast to this, stable carbon-centered
radical adducts are formed from all investigated spin traps. The
prospect of facile modification of the novel compounds under mild
conditions justifies further research with respect to better yields
and higher adduct stabilities.
a
5. Experimental
5.1. Chemicals
Acrolein and triethylamine were commercially available from
Fluka, 2-nitropropanol, 2,3-dihydrofurane, and 3,4-dihydro-(2H)-
pyrane were from Sigma–Aldrich, all other chemicals were ob-
tained from VWR.
b
5.2. Syntheses
Synthesis and characterization of the compounds were per-
formed in analogy to those reported previously for the synthesis
of derivatives of EMPO3–7,18,23 with modifications as given below.
5.2.1. 2-Nitropropanol tetrahydrofuranyl and tetrahydropyranyl
ether
c
To 7 ml of 2-nitropropanol (8.3 g; 78.9 mmol) either 1 ml of 2,3-
dihydrofurane (0.927 g; 13.07 mmol for the tetrahydrofuranyl
ether) or 1 ml of 3,4-dihydro(2H) pyrane (0.93 g; 11.05 mmol, for
the tetrahydropyranyl ether) was added. HPLC (75% methanol in
water as the eluent on a 250 ꢂ 4.6 mm C-18 column) was run as
a control. Toluenesulfonic (10 mg) acid was added as a catalyst
and the reaction was monitored every 15 min by HPLC. One milli-
liter of portions of either 2,3-dihydrofurane (THF ether) or 3,4-
dihydro-(2H)-pyrane (THP ether) were added at 10 min intervals,
until the HPLC runs confirmed that the reaction was complete
(after ca. 1 h and a total amount of 7 ml of 2,3-dihydrofurane or
8 ml of 3,4-dihydro-(2H)-pyrane added). Average yield: 94% of
crude product, which was used without further purification.
20 G
Figure 8. Iron-dependent formation of carbon-centered spin adducts from HMMPO,
FMMPO, and PMMPO in the presence of formate (a) HMMPO (40 mM, initial
concentration) was incubated with a Fenton system containing FeSO4 (1 mM),
EDTA (2 mM), H2O2 (0.2%) and sodium formate (200 mM). The reaction was stopped
after 10 s by 1:1 dilution with phosphate buffer (300 mM, pH 7.4, containing 20 mM
DTPA) and the spectrum was recorded using the following spectrometer settings:
sweep width, 80 G; modulation amplitude, 0.21 G; microwave power, 20 mW; time
constant, 0.08 s; receiver gain, 1 ꢂ 104; scan rate, 57 G/min (rel. int.: 15), (b) same
as in (a), except that FMMPO was used (rel. int.: 13), (c) same as in (a), except that
PMMPO was used (gain 2.5 ꢂ 104, rel. int.: 65).
5.2.2. 2-(2-Furanyl)-oxy-2-nitropentanal and 2-(2-pyranyl)-oxy-
2-nitropentanal
superoxide adduct stabilities,8–10 although the reported synthetic
procedure is rather complex. In view of the low stabilities of the
superoxide adducts we investigated the stabilizing effect of
b-cyclodextrin and compared it to the values obtained with DMPO
The crude product of the respective tetrahydrofuranyl ether
(73.6 mmol; 13 g) or tetrahydropyranyl ether (73.6 mmol; 14 g)
was dissolved in 50 ml acetonitrile at 0 °C and 2 ml of triethyl-
amine was added. Acrolein (6 ml; 88.8 mmol) was slowly added
in 1 ml portions while the reaction was monitored by HPLC
(75% methanol; 250 ꢂ 4.6 mm C-18). After 30 min the ice bath
was removed und the reaction mixture was stirred at room
temperature until the reaction was complete (after ca. 2 h).
The solvent and the triethylamine were removed at reduced
pressure and the purity of the crude product (ca. 95% yield)
was assessed by HPLC. The product was used without further
purification.
and EMPO. Replacement of b-cyclodextrin by
a-, c-, or heptakis-
(2,6-di-O-methyl)-b-cyclodextrin gave considerably different re-
sults, the use of cucurbit[6]uril (spectra not shown) did not show
any stabilizing effect at all.
All spin traps are forming characteristic hydroxyl radical ad-
ducts: Methoxyl radical adducts were rather unstable and could
only be detected from FMMPO and PMMPO as minor components
in mixtures with the predominant hydroxymethyl radical adduct.
From two EPR spectra recorded at different time intervals poorly
resolved spectra of the FMMPO and PMMPO methoxyl radical ad-
ducts could be calculated, whereas from HMMPO no methoxyl rad-
ical adduct was detectable under these conditions. The most stable
spin adducts were formed from carbon-centered radicals, for
example, from DMSO (ꢁCH3), methanol (ꢁCH2OH), ethanol
5.2.3. Synthesis of the nitrones
Synthesis of the nitrones was performed in analogy to previ-
ously published EMPO derivatives3–7 with the following
modifications:
Ca. 70 mmol of the crude 2-(2-furanyl)-oxy-2-nitropentanal (or
2-(2-pyranyl)-oxy-2-nitropentanal) was dissolved in 40 ml of
methanol and a solution of 3 g NH4Cl in 60 ml water was added
which causes a partial precipitation of the solutes. Zn (11 g) pow-
der (168 mmol) was gradually added in steps of 1 g under HPLC
(ꢁCH(CH3)OH), and formate (ꢁCOꢀ).
2
4. Conclusion
control (dissolution of 20
in an Eppendorf tube and centrifugation for 1 min). After 1 h, acetic
acid was carefully added in small amounts (ca. 200 l) until the
product peak in the HPLC control had passed its maximum. The
ll of the mixture in 1 ml of methanol
In conclusion, all three novel DMPO-derived spin traps formed
rather unstable superoxide adducts (t1/2 ca. 3–5 min) and can
therefore not be recommended for superoxide trapping, although
cyclodextrins showed a significant stabilizing effect on the radical
l

K. Stolze et al. / Bioorg. Med. Chem. 19 (2011) 985–993
993
solution was filtered and washed three times with 30 ml methanol.
The filtrate was extracted 5 times with 100 ml of organic solvent
(75% petroleum ether/25% methylene chloride) in order to remove
unreacted starting material (the respective pentanal). The organic
phase was extracted 5 times with 30 ml water and was discarded
afterwards. The original filtrate (containing the Zn salts) was ex-
tracted five times with 50 ml of methylene chloride. The aqueous
phase containing Zn was discarded, the methylene chloride phase
was re-extracted five times with 30 ml of water. After removal of
the solvent (mostly water) about 6–8 g slightly orange to light
brown material was obtained (average yield ca. 35%). Column chro-
matography on silica gel with a methylene chloride/ethanol gradi-
ent provided an almost colorless product.
100 mM phosphate buffer, pH 7.4. The mixture was vortexed for
2 min at room temperature. If necessary, the procedure was re-
peated several times, until an equilibrium between the two phases
was achieved. After careful separation of the phases, the absor-
bance was read at the maximum around 235 nm after dilution
with methanol. For EPR experiments, Bruker spectrometers
(ESP300E and EMX) were used, operating at 9.7 GHz with
100 kHz modulation frequency, equipped with
a rectangular
TE102 or a TM110 microwave cavity. All calculations for spectral
simulation were done using the SimFonia Program by Bruker
(Table 6).
Acknowledgments
Additional purification was done on a 5 ml solid phase extrac-
tion column (Chromabond C-18 100 mg column obtained from
Macherey-Nagel, Düren, Germany) using a water/methanol gradi-
ent as the eluant. Colored solutions were subjected to the same
procedure until the solution containing the N-oxides remained
colorless.
The authors wish to thank P. Jodl for skilful technical assistance
in synthesis, purification, and characterization of the spin traps.
The financial support of the Christian Doppler laboratory for ad-
vanced cellulose chemistry and analytics is gratefully
acknowledged.
The products were extracted three times with CH2Cl2, the sol-
vent was removed and the identity and purity was checked, aver-
age yield: 25% (purified product).
Supplementary data
The purity of the obtained products was assessed by TLC, HPLC,
and UV–vis spectroscopy. Final identification of the purified prod-
ucts was performed by 1H NMR, 13C NMR, and IR spectroscopy.
Purity was confirmed by microanalysis (see Table 3).
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmc.2010.11.052. These data
include MOL files and InChiKeys of the most important compounds
described in this article.
5.2.4. 5-Hydroxymethyl-5-methyl-pyrroline N-oxide
References and notes
7 g of PMMPO (32.8 mmol) were dissolved in 10 ml water at
1. Fréjaville, C.; Karoui, H.; Tuccio, B.; Le Moigne, F.; Culcasi, M.; Pietri, S.;
Lauricella, R.; Tordo, P. J. Med. Chem. 1995, 38, 258.
2. Stolze, K.; Udilova, N.; Nohl, H. Free Radical Biol. Med. 2000, 29, 1005.
3. Stolze, K.; Udilova, N.; Nohl, H. Biol. Chem. 2002, 383, 813.
4. Stolze, K.; Udilova, N.; Rosenau, T.; Hofinger, A.; Nohl, H. Biol. Chem. 2003, 384,
493.
5. Stolze, K.; Rohr-Udilova, N.; Rosenau, T.; Stadtmüller, R.; Nohl, H. Biochem.
Pharmacol. 2005, 69, 1351.
6. Stolze, K.; Rohr-Udilova, N.; Rosenau, T.; Hofinger, A.; Kolarich, D.; Nohl, H.
Bioorg. Med. Chem. 2006, 14, 3368.
7. Stolze, K.; Rohr-Udilova, N.; Rosenau, T.; Hofinger, A.; Nohl, H. Bioorg. Med.
Chem. 2007, 15, 2827.
8. Hardy, M.; Chalier, F.; Ouari, O.; Finet, J.-P.; Rockenbauer, A.; Tordo, P. Chem.
Commun. 2007, 1083.
9. Chalier, F.; Hardy, M.; Ouari, O.; Rockenbauer, A.; Tordo, P. J. Org. Chem. 2007,
72, 7886.
10. Hardy, M.; Rockenbauer, A.; Vasquez-Vivar, J.; Felix, C.; Lopez, M.; Srinivasan,
S.; Avadhani, N.; Tordo, Paul.; Kalyanaraman, B. Chem. Res. Toxicol. 2007, 20,
1053.
0 °C and 250 ll concentrated hydrochloric was added. The cleaving
of the protective THP group was monitored every 10 min by HPLC.
After completion, the solution was neutralized with stoichiometric
amounts of solid sodium hydrogen carbonate and the organic
impurities were extracted with 75% petroleum ether/25% CH2Cl2
(five times 15 ml, both phases checked by HPLC). The aqueous
phase was evaporated carefully in a rotary evaporator at 35–
40 °C and the product was dissolved in methylene chloride in order
to separate the precipitated sodium chloride. Solid phase extrac-
tion on mini columns (5 g of C-18 material) yielded a purified
material.
5.3. Instruments
NMR spectra were recorded on a Bruker Avance at 400 MHz for
1H, and 100 MHz for 13C at 22 °C. CDCl3 was used as the solvent
throughout, TMS (tetramethylsilane) as the internal standard.
Concentrations were set to 10 mg sample/0.6 ml solvent, tempera-
ture to 293 K. 13C peaks were assigned by means of APT (attached
proton test), HMQC (1H-detected heteronuclear multiple-quantum
coherence) and HMBC (heteronuclear multiple bond connectivity)
spectra. All chemical shift data are given in ppm units. Mass
spectra were obtained on a ESI Q-TOF MS on a Waters Micromass
Q-TOF Ultima Global in 70% aqueous methanol containing 0.1% for-
11. Patel, A.; Rosenau, T.; unpublished results.
12. Baati, R.; Valleix, A.; Mioskowski, C.; Barma, D. K.; Falck, J. R. Org. Lett. 2000, 2,
485; cit. lit. Greene, T. W.; Wuts, P. G. M. Protective Groups in Organic Synthesis;
Wiley: New York, 1999. Chapter 2, p. 57.
13. Mohammadpoor-Baltork, I.; Kharamesh, B. J. Chem. Res. (S) 1998, 146; cit. lit.:
Greene, T. W.; Wuts, P. G. M. Protective Groups in Organic Synthesis, 2nd ed.;
Wiley: New York, 1991.
14. Karoui, H.; Rockenbauer, A.; Pietri, S.; Tordo, P. Chem. Commun. 2002, 3030.
15. Bardelang, D.; Rockenbauer, A.; Karoui, H.; Finet, J.; Biskupska, I.; Banaszak, K.;
Tordo, P. Org. Biomol. Chem. 2006, 4, 2874.
16. Han, Y.; Tuccio, B.; Lauricella, R.; Villamena, F. A. J. Org. Chem. 2008, 73, 7108.
17. Šnyrychová, I. Free Radical Biol. Med. 2010, 48, 264.
18. Zhang, H.; Joseph, J.; Vasquez-Vivar, J.; Karoui, H.; Nsanzumuhire, C.; Martásek,
P.; Tordo, P.; Kalyanaraman, B. FEBS Lett. 2000, 473, 58.
19. Dikalov, S.; Jiang, J.; Mason, R. P. Free Radical Res. 2005, 39, 825.
20. Roubaud, V.; Lauricella, R.; Bouteiller, J. C.; Tuccio, B. Arch. Biochem. Biophys.
2002, 397, 51.
mic acid at a flow rate of 5 ll/min. IR spectra were recorded as film
on an ATI Mattson Genesis Series FT-IR spectrometer (see also
Table 4).UV–vis spectra were recorded on Hitachi 150-20 and
U-3300 spectrophotometers in double-beam mode against a blank
of the respective solvent (Table 5). Determination of the concentra-
tions was done measuring the absorption maxima in the range be-
tween 200 and 300 nm. For measurements of the partition
21. Stolze, K.; Rohr-Udilova, N.; Hofinger, A.; Rosenau, T. Bioorg. Med. Chem. 2008,
16, 8082.
22. Dikalov, S. I.; Mason, R. P. Free Radical Biol. Med. 2001, 30, 187.
23. Olive, G.; Mercier, A.; LeMoigne, F.; Rockenbauer, A.; Tordo, P. Free Radical Biol.
Med. 2000, 28, 403.
coefficients, 500
ll of n-octanol was added to 500 ll of a solution
24. Culcasi, M.; Rockenbauer, A.; Mercier, A.; Clément, J.-L.; Pietri, S. Free Radical
Biol. Med. 2006, 40, 1524.
of the respective spin trap (100 mM or 5–10 mg, respectively) in