Synthesis of two novel pyridine annulated pyrrolidine nitroxides -Five-step Synthetic Route of Pyridine-Annulated Pyrrolidine Nitroxides-N

: Two novel, unsubstituted, pyridine-annulated pyrrolidine nitroxides; 1,1,3,3-tetramethyl-2,3-dihydro-1H-pyrrolo[3,4-c]pyridin-2-yloxyl and 1,1,3,3-tetramethyl-2,3-dihydro-1H-pyrrolo[3,4-c]pyridin-5-oxide-2-yloxyl were synthesized in overall yield of ~15% (5 steps starting from pyridine-3,4-dicarboxyic acid) via a Grignard approach. Grignard methylation of N-benzylcinchomeronic imide that was derived from pyridine-3,4-dicarboxylic anhydride, furnished the tetramethylpyrrolidine precursor in an isolated yield of 18%. Hydrogenation of the tetramethyl pyrrolidine precursor followed by oxidation afforded, depending on the oxidation conditions, each of the two nitroxides.

• Side product formation was the main reason for the limited yield at Grignard step.
• Two novel heterocyclic nitroxides were produced via a five-step pathway in ~15% yield. Step (c): Yield limiting step due to the formation of numerous, unavoidable side products Therefore, the overall yield of the final product has been reduced.

INTRODUCTION
Isoindoline nitroxides, a commercially valuable class of nitroxides have gained widespread attention mainly due to some advantages they possess over the commercially available other classes of nitroxides. If the 6-membered aromatic ring of the isoindoline skeleton is replaced by a heteroaromatic ring, the nitroxide skeleton would have been modified further with some novel vital properties. Established literature supports that monocyclic nitroxides containing heteroatoms have been synthesized in early 1980s (Keana et al., 1982). Some of the uses of monocyclic heteroaromatic nitroxides such as EPR probes in biomedicine and related fields to monitor oxidative stress and reactive radical species in biological systems (Bobko et al., 2012;Zhelev et al, 2019), pH sensitive spin probes (Khlestkin et al., 2005), contrast enhancing agents (Keana et al., 1987) for magnetic resonance imaging (MRI) applications and molecular units (Vaz et al., 1999;Laget et al., 1998) in the synthesis of molecular magnetic materials are of notable importance. However, there are some disadvantages associated with monocyclic nitroxides containing heteroatoms as well. For instance, the electron-withdrawing effect of the heteroatom in the ring can destabilize the positive charge of the resulting oxo-ammonium cation (Hicks, 2010). Furthermore, the heteroatom can also facilitate ring opening reactions (Keana et al., 1982;Gryn'ova et al., 2012) in monocyclic nitroxides and thereby promote degradation.
These issues can be evaded if the heteroaromatic ring is fused to the monocyclic nitroxide skeleton. This fused heteroaromatic ring could confer some advantages to the skeleton such as rigidity and resistance towards ring opening reactions (Gryn'ova et al., 2012;Hansen and Blinco, 2018) and resistance towards the alteration of functions of biomolecules when they act as spin labels (Kalai et al., 2000). If the fused heteroaromatic ring is pyridine, it would further impart good σ donor capabilities as a monodentate ligand (Budzelaar, 2012). Therefore, this paper focuses on developing a novel unsubstituted heteroaromatic nitroxide with a fused pyridine ring.

Chemicals & Apparatus
All chemicals used were of analytical reagent grade purchased from chemical suppliers such as Sigma-Aldrich. Dichloromethane (DCM) was freshly distilled from calcium hydride and tetrahydrofuran (THF) from sodium benzophenone ketal prior to use. Both toluene and diethyl ether were dried over sodium wire and triethylamine was dried over potassium hydroxide. All air-sensitive reactions were performed under an ultra-high purity argon atmosphere. All other reagents were purchased from commercial suppliers and used without further purification. 1H and 13C NMR spectra were recorded on a 400 MHz spectrometer and referenced to the relevant solvent peak ( Scheme 2: Pyridine-fused pyrroline nitroxide synthesis using spontaneous cyclization reactions. Scheme 3: Preparation of precursor for pyridine annulated pyrroline nitroxide using cobalt (0) catalyzed cycloaddition reactions. mass spectra were obtained using a QTOF LC mass spectrometer which utilized electrospray ionisation (recorded in the positive mode) with a methanol mobile phase. Melting point values were collected on a Variable Temperature Apparatus using the capillary method and were uncorrected. Analytical HPLC was carried out on a HPLC system using an Agilent 1100 Prep-C18 scalar column (4.6 × 150 mm, 10 μm) with a flow rate of 1 mL/ min in the stated mixtures of methanol and water with detection at 254 nm. In all HPLC analyses, the solvent system used was MeOH : H 2 O, 65: 35 except for compound 21a and 21b. For compounds 21a and 21b, MeOH: H 2 O, 70:30 was used as the solvent system. Merck Silica Gel 60 F254 TLC plates were used for analytical Thin-Layer Chromatography (TLC) while Silica Gel 60 (230-400 mesh) was used for preparative column chromatography.

RESULTS & DISCUSSION
Synthesis of the starting imide, N-benzylcinchomeronic imide (13) was achieved starting from commercially available pyridine-3,4-dicarboxylic acid (11) via two steps (Scheme 4) with an overall yield of 89%. Next, Grignard methylation was performed on imide 13 to synthesize 1,1,3,3-tetramethyl adduct. Following Griffiths standard procedure (Griffiths et al., 1983) of synthesizing 2-benzyl-1,1,3,3-tetramethylisoindoline adduct, imide 13 was treated with 6.0 equivalents of MeMgI in refluxing toluene for 3 h. Examining the reaction mixture (after the aqueous NH 4 Cl work-up) by TLC and isolating the components by column chromatography showed two relatively polar products, 16 and 17, as being formed in the mixture (Scheme 5). When the methylation reaction was undertaken on imide 13 with excess MeMgI for an extended reflux time (5 h), two relatively non-polar components appeared (Scheme 5) in the reaction mixture (by TLC and HPLC).
The desired tetramethyl adduct 14 was identified (by NMR and Mass) as being formed in the reaction mixture along with ethyltrimethyl adduct 15 (Scheme 5).
Optimizing the experimental conditions of the Grignard methylation of imide 13 to improve the proportion of 14 in the reaction mixture was conducted by carrying out the Grignard methylation reactions at 110 °C under different reaction times and analyzing them by HPLC in order to calculate the relative HPLC product ratios of 14, 15, 16 and 17 appearing in the chromatogram (Table 1). Based on these calculations, the relative HPLC proportion of 14 was improved to 23% (Entry 4, Table 1) by increasing the amount of MeMgI from 6.0 to 8.0 equiv. and refluxing the reaction mixture for 72 h in toluene. The compounds 14, 15, 16 and 17 were isolated from the reaction mixture by column chromatography (hexane:ethyl acetate, 1:3, 4:1), and their yields were 18%, 5%, 25% and 27%, respectively.
Based on the relative HPLC product ratios shown in Table 1, some mechanistic insights of the Grignard  Each entry (1-4) corresponds to a Grignard methylation performed on 13 following the general reaction conditions described for the preparation of 14 (Experimental section).

Scheme 6:
A proposed mechanism for the formation of 14, 15, 16 and 17, starting from 13 during the Grignard methylation.
methylation of imide 13 were revealed (Scheme 6). Previous literature (Jayawardena et al., 2013) supports that products such as 17 could possibly arise from an iminium intermediate such as 19, during the aqueous work-up. This intermediate 19, sighted as the initial precursor intermediate of forming 1,1,3,3-tetraalkyl adduct during the Grignard alkylation of imides could also undergo ring methylation (at fourth position of the ring) leading to a structure like 19a which is inert to further methylation (Colwell et al., 2011). Absence of any methyl groups at the fourth position of the pyridine ring of two tetraalkyl adducts (14 or 15) exhibits that intermediate 19a would remain as an inert structure for further Grignard methylation.
The variation of the HPLC product ratios in entries 1 and 2 (Table 1) Table 1) has surprisingly led to a significant decrease in the relative amount of 14 along with an increase in the relative amount of 15. This suggests that iminium ion 19 could possibly be converted to another intermediate like 19c efficiently (over the long reflux time), due to the deprotonation driven by the strongly basic Grignard environment. This intermediate 19c would eventually give rise to a higher amount of 15 by reacting with some unreacted MeI in the Grignard reagent, followed by three Grignard methylations. Previous findings suggest that exocyclic amides such as 19c could be the possible precursors for the formation of unusual mixed alkylated adducts like 15 (Jayawardena et al. unpublished).
Tetramethyl adduct 14 could be derived in the reaction mixture via two possible pathways. One mechanism (Braslau and Chaplinski, 1998) is three alternative additions of MeMgI at C 1 and C 3 of intermediate 19.
The other possibility of deriving 14 is via 1,1-dimethyl intermediate (19b) by two methyl additions at C 3 . It was earlier observed that 1,1-dimethyl amides are convertible to the 1,1,3,3-tetramethyl adduct during the Grignard methylation reactions (Jayawardena et al. unpublished). Based on these comparisons, it is reasonable to hypothesize that tetramethyl adduct 14 could be derived via a 1,1-dialkylamide intermediate like 19b through 1,1-addition. This suggestion also explains why tetramethyl adduct 14 was formed during the Grignard methylation of 13.  The next step, hydrogenation of 14 was successfully achieved with stirring in MeOH (as solvent) in the presence of H 2 /Pd/C at room temperature (3-4 h) in a yield of 95%. Oxidation of the secondary amine 20 resulted from the hydrogenation of 14 was approached using H 2 O 2 in the presence of Na 2 WO 4 .2H 2 O/NaHCO 3 in a mixture of MeOH/ MeCN (7:1) resulting the target nitroxide 21a in a yield of 80% (48 h). Another novel nitroxide 21b was also isolated from this reaction in a yield of 17%. However, an extended reaction time (120 h) generated the pyridine oxide-typenitroxide 21b as the major product (85%) along with some small amount of 21a (6%). This protocol involving aqueous H 2 O 2 provides milder yet higher yielding synthesis for the two novel nitroxides, 21a and 21b. When the oxidation of 20 was undertaken with mCPBA, the reaction specifically led to nitroxide 21b within 2 h (81%). Notably, the conversion of nitroxide 21a to 21b was also achieved with both H 2 O 2 and mCPBA.
This methodology provides a short and convenient pathway (5 steps) to synthesize the novel heterocyclic nitroxides in practically preparative scales.

CONCLUSIONS
The main scope of this study was to introduce the synthesis of tetraalkylated pyridine-annulated heterocyclic nitroxides via a short and convenient pathway. Finally, this approach furnished, in good overall synthetic yields (up to ~15%), two novel pyridine-annulated tetramethyl nitroxides (21a and 21b) starting from pyridine-3,4-dicarboxylic acid 11, via a short (5 steps) and convenient pathway.

Data Availability
Datasets used to characterize the compounds of this study are available upon request.