Simplified Synthesis of Isotopically Labeled 5,5-Dimethyl-pyrroline N-Oxide

5,5-Dimethylpyrroline N-oxide (15N) and 5,5-di(trideuteromethyl)pyrroline N-oxide were synthesized from the respective isotopically labeled 2-nitropropane analogs obtained from the reaction of sodium nitrate with 2-halopropanes. This facile, straightforward process allows synthesizing isotopically labeled DMPO analogs in a 4-step reaction without special equipment.


Introduction
Nitrone spin traps such as the most commonly used 5,5-dimethylpyrroline N-oxide (DMPO) are important reagents for the detection of free radicals by means of ESR spin trapping [1]. For certain mass spectrometry experiments or to investigate the fidelity of spin trapping, it is helpful to use isotopically labeled spin traps.
The di(trideuteromethyl) analog of DMPO with its mass shift of +6 compared to unlabeled DMPO can be used for dual spin-trapping mass spectrometry experiments. Here, spin traps labeled with stable isotopes ( 2 H, 15 N or 13 C) are utilized to simplify the interpretation of mass spectrometry experiments [2,3]. With equal amounts of the labeled and unlabeled spin trap present, adducts of the OPEN ACCESS trapped radicals will appear as ion pairs in the mass spectrum (with the mass differences depending on the isotopes incorporated in the spin trap and on the charge state of the ion). This facilitates and clarifies the identification of radical-derived analytes.
The 15 N analog can be used to unambiguously determine the susceptibility of a particular spintrapping experiment to the Forrester-Hepburn artifact in an ESR experiment [4]. A Forrester-Hepburn artifact is the result of nucleophilic attack of the radical precursor on the spin trap with subsequent oxidation to the identical nitroxide radical as derived from genuine spin trapping. It is difficult to distinguish between nucleophilic attack and free-radical trapping, with normal chemical and biological control experiments being of no use. Timmins et al. reported a method based on spin traps with different isotopes at the α-or β-positions to the nitrogen of the spin trap [4]. The substrate is preincubated with a spin trap (first isotope), and then the spin trap labeled with the second isotope is added simultaneously with the initiation of free radical formation. Because the ESR signals are different, the origin of the signal can be determined as artifact or genuine signal. Employing that technique, we were able to identify the DMPO/ • CN radical, supposedly generated by horseradish peroxidase and hydrogen peroxide, as an artifact (unpublished data).
The classical synthesis of DMPO, as reported by Janzen et al. [5], is based on the synthesis of pyrrolines [6]. Later, Le et al. [7] published a synthesis of 2-14 C-DMPO that avoided the direct Michael reaction of nitropropane with methyl acrylate, thereby improving the low yield and eliminating the difficult purification of this step. For the synthesis of isotopically labeled DMPO, Pou et al. [8] published an effective method starting from 15 N-hydroxylamine, which involved the use of hydrogen gas in an autoclave as well as a reaction with ozone derived from an ozone generator. We have developed a more facile synthetic pathway for the synthesis of DMPO based on 15 N-sodium nitrite or 2-bromopropane (D7) as the isotopically labeled starting material. The intermediate 2-nitropropane was prepared in a one-step reaction according to the reaction principle described by Kornblum et al. [9]. This principle has been used for nitrone spin trap synthesis [10]. Nitropropane was then used for a DMPO synthesis similar to that of Le et al. [7] (Scheme 1).

2-Nitropropane
The synthesis of 2-nitropropane can be accomplished directly by nitration of 2-halopropanes with sodium nitrite [9]. With 2-iodopropane, the reaction is carried out in dry DMF in the presence of urea. For the slower reacting 2-bromopropane, a longer reaction time and the presence of phloroglucinol as a nitrite ester scavenger is required.
To synthesize 15 N-DMPO, 2-iodopropane was used in slight excess in the absence of phloroglucinol because phloroglucinol did not increase the yields with respect to the 15 N-nitrite. The nitrite forms the nitro compound and, as a byproduct in a slow process, the nitrite ester. In an undesired reaction, the latter can react with already formed nitropropane, but the nitrite ester can be removed with phloroglucinol [9]. However, under the chosen conditions, the formation of the nitrite ester is not the limiting factor for the yield. After purification by vacuum distillation, the yield of 15 N-2-nitropropane was 35%. Perdeuterated 2-nitropropane (D7) was synthesized from 2-bromopropane (D7) in the presence of phloroglucinol since this was more economical than employing the corresponding deuterated iodo-compound. The reaction gave perdeuteronitropropane, formed with 27% yield after purification. The yields are not very high for one-step reactions, but still result in a higher overall yield of 2-nitropropane than a multistep process (e.g., Gabriel-synthesis of the corresponding phthalate, cleavage of the 2-propylamine [11] and subsequent oxidation by ozone, which can induce isomerization).

Methylnitrovaleric Acid Methyl Ester
In order to avoid the reported problems of the direct aldehyde synthesis via the Michael reaction [5], we chose, instead, the conditions described by Moffett [12] to optimize the yield with respect to the nitropropane. The yield of 15 N-nitropropane was 51% and the yield of the 1,1,1,3,3,3-hexadeutero-2nitropropane was 80%, which was comparable to that reported in the literature [12].

Methylnitropentanal
The corresponding pentanal was formed by reduction with diisobutylaluminum hydride, as described by Le et al. [7]. Some experimental details of the protocol were modified according to [13]. In order to achieve a selective reduction to the aldehyde with minimal further reduction to the alcohol, which is difficult to remove once formed, the bath temperature should be maintained at −90 °C during the reaction. The use of wet silica gel allows the quenching procedure to be carried out at the reaction temperature [13]. The reaction gave yields of 58% for the hexadeutero-and 74% for the 15 N-compound, which is in the expected range for diisobutylaluminum hydride reductions [13].

DMPO
The final step was carried out as described in the literature [7]. The yield of DMPO was 15% with both analogs. In preliminary experiments, the activation of zinc or variations of the zinc equivalents did not improve the yield to 60%, as previously reported [14]. In our hands, we obtained ca. 15-20% (as also reported by Le et al. [7]). In retrospect, for future syntheses it may be beneficial to protect the aldehyde as dioxolane before the reduction step [5] because Rosen et al. also reported a low yield for the direct reaction [15].
The synthesized DMPO analogs and 14 N-DMPO were analyzed using high mass resolution mass spectrometry (Rs > 10,000). The protonated molecular ion of each analog was observed and the resulting exact mass measurement was determined. The resulting M+H + ions observed for each DMPO analog were as follows: DMPO (M+H) + ion of m/z 114.0915 (theoretical M+H + = 114.0919), 15 N-DMPO (M+H) + ion of m/z 115.0882 (theoretical M+H + = 115.0889), and 6 D-DMPO (M+H) + ion of m/z 120.1282 (theoretical M+H + = 120.1295). According to their elemental compositions, these exact mass measurements correspond to mass accuracies of 3.5 ppm, 6.1 ppm, and 10.8 ppm, respectively. In addition, similar fragmentations of each analog were observed.
With a spin-trapping experiment, the coupling constants of the respective hydroxyl radical adducts were determined. A strong signal was detected with both analogs (Figures 1a,d). With the di(trideuteromethyl) analog, a signal with similar coupling constants and the same intensity as with 14 [16,17]. In the absence of hydrogen peroxide (b and e) or iron (c and f), no significant signal was detected with either analog.

N-5,5-Dimethyl-1-pyrroline N-oxide.
In a 500 mL three-neck flask with an addition funnel, ethanol (95%, 60 mL) was cooled to 2 °C. Then 15 N-4-nitro-4-methyl-1-pentanal (2.85 g, 19.2 mmol) and zinc dust (2.68 g, 0.41 mmol, 2.1 equivalents) were added. Under brisk magnetic stirring, glacial acetic acid (4.2 mL) was added dropwise while the temperature was kept below 10 °C. After stirring 1 h, the apparatus was put in a refrigerator suitable for flammable materials. Any hydrogen gas which may have formed was allowed to dissipate. After an additional 30 h of stirring at 10 °C, the solution was filtered to remove zinc and any undissolved zinc acetate. The filtrate was extracted with cold ethanol (3 × 100 mL). From the combined solutions, the ethanol was removed with a rotational evaporator. The residue was diluted with dichloromethane (200 mL) and washed twice with saturated sodium bicarbonate solution and water (50 mL each). The organic layer was dried with magnesium sulfate, filtered and the dichloromethane removed with a rotary evaporator to give a reddish-brown crude product. The product was purified by two consecutive sublimations (0.1 torr) from room temperature to 0 °C, and 305 mg (2.7 mmol, 15%) of 15  Mass Spectrometry. Samples of the DMPO analogs were initially prepared at 100 mM in acetonitrile, diluted 1000-fold just prior to analysis with a solution of water with 0.1% formic acid and infused into the mass spectrometer at 500 nL/min using a syringe pump. The instrumental parameters for the MS analyses were as follows: capillary voltage, 3.5 kV; cone voltage, 30 V; collision energy, 4 eV and source temperature, 80 ºC. MS/MS data of the DMPO analogs were acquired using collision energies of 15-25 eV. For calibration, a solution of glu-fibrinopeptide B (500 fmol/μL) in water/acetonitrile 80:20 (v/v) with 0.1% formic acid and a mass of 785.8496 (2+) was used. Data analysis was accomplished using MassLynx software supplied by the manufacturer.

Conclusions
Here we describe a synthesis pathway for the isotopically labeled spin traps 15 N-5,5-dimethyl-1-pyrroline N-oxide and 5,5-di(trideuteromethyl)-1-pyrroline N-oxide. Though syntheses for isotopically labeled DMPO have already been reported [8], the elimination of hydrogen gas and ozone as reactants and the respective special equipment facilitates the synthesis and gives reasonable yields. As reported previously [7], Michael reaction to the methyl ester and subsequent reduction to the aldehyde allows for better yields and easier purification compared to direct aldehyde formation. For reduction to DMPO, it may be beneficial to protect the aldehyde as the corresponding dioxolane first since the yield of the direct reaction was below the yield reported in the original paper. This issue has also been reported elsewhere [15]. Our approach is well suited for the synthesis of isotopically labeled DMPO analogs.