o -Nitrophenylacetonitrile Michael additions and cyclocondensations: a novel quinoline synthesis

Michael additions of o -nitrophenylacetonitrile have been investigated. The adducts are formed diastereoselectively under mild conditions. Higher temperatures result in an unusual annulation involving intramolecular nucleophilic attack of an enolate on the nitro group, giving rise to 2,3,4-trisubstituted quinolines. In contrast, under similar conditions ethyl o -nitrophenylacetic acid reacts with N -methylmaleimide to give an N -hydroxyindole. A possible explanation for the divergent chemistry, and mechanisms for these reactions are proposed.


Figure 1. o-Nitrophenylacetonitriles as synthons for benzo-fused heterocycles.
As part of efforts to achieve the total synthesis of the marine pyrroloacridine natural product alpkinidine, [32][33][34] we investigated Michael addition of 1 to quinones. Somewhat surprisingly, these reactions were unsuccessful. 33 For the purposes of calibration, we explored the reactions of 1 with other Michael acceptors. In the course of these studies, we encountered some interesting and unexpected reactions affording quinolines and Nhydroxyindoles, which are reported here.

Results and Discussion
Our investigation into nucleophilic reactions of the anion derived from 1 began with α-methylation ( Table 1). The reaction failed to go to completion with K2CO3 in EtOH, and it was apparent that the purple carbanion had adsorbed to the poorly soluble base, effectively sequestering it from the electrophile. A switch to the more soluble base Cs2CO3, or a dipolar aprotic solvent, 9 improved the yield of 8 considerably. Michael additions were then explored ( Table 2). In these reactions, the combination of K2CO3/EtOH worked well, as expected given the requirement for only catalytic base and the stabilization of transition states imparted by the protic solvent. Reactions in MeCN were much slower and gave side-products and inferior yields. The Michael addition of 1 to methyl acrylate, using potassium t-butoxide in THF, has been reported to provide the corresponding adduct in 76% yield. 35 In our hands, milder conditions with ethyl acrylate (9) led to a higher yield of 14.
Michael additions to N-methylmaleimide (10) or cyclohexenone (11), gave 10:1 and 5:1 mixtures of diastereomeric adducts 15 and 16, respectively, as determined by 1 H NMR spectroscopy of the crude products. Attempts to separate the stereoisomers by chromatography resulted in significant mass losses; however, in both cases the major diastereomer was easily isolated by washing the crude solid product with Et2O. The relative configurations of the major isomers were determined by X-ray crystallography ( Figure 1).

Figure 1.
Representations of the X-ray crystal structures of 15 (left) and 16 (right). Crystals of 16 contained two independent molecules in the asymmetric unit. In the molecule not shown, the nitro group is slightly more twisted out of the plane of the benzene. Displacement envelopes are at 50% probability amplitude, with hydrogen atoms assigned arbitrary radii. Michael addition to diethyl maleate (12) proceeded as expected, but in this case afforded a 11:14 mixture of diastereomers 17 that were neither separated nor assigned to spectroscopic features. The reaction with dimethyl acetylenedicarboxylate (DMAD, 13) was undertaken in MeOH to exclude the possibility of transesterification. In contrast to the other examples, this reaction produced a complex array of products and attempted chromatographic separation failed to yield any identifiable compounds.
In 1960, Loudon and Wellings reported the synthesis of N-hydroxyindole 19 during an attempted recrystallisation of 18 from weakly basic EtOH/H2O 11 (Scheme 1). Given the structural similarities, we subjected Michael adduct 15 to treatment with K2CO3 in refluxing EtOH. A new major product was isolated following acidic work-up; however, it was clearly not the expected N-hydroxyindole 20. One and 2D NMR spectroscopy experiments suggested the unexpected product was a quinoline, and X-ray crystallography confirmed the structure 22 (Figure 2). It was also noted from TLC that the Rf of 22 was different from that of the major product observed in the reaction mixture, suggesting that a further transformation had occurred during acidic work-up. Thus, the reaction was repeated with a basic aqueous work-up, leading to the isolation of the imino congener 21, again confirmed by X-ray crystallography ( Figure 2). The identification of 21 shed light on a possible mechanism for the unusual reaction (Scheme 2).  The proposed mechanism is supported by the reaction of the crude diastereomeric mixture of adducts 17 under the same conditions, which afforded quinoline 28 (Scheme 3). To the best of our knowledge, the closest precedent to this rare and unusual cyclocondensation, initiated by nucleophilic attack of an enolate on a nitro group, involves 2'-nitrobiphenyl-2-acetic acid derivatives 29, which cyclize under strongly basic conditions to give the phenanthridine-N-oxides 30 (Scheme 3). 36,37 Scheme 3. Base-catalyzed cyclocondensations involving a nitro group -current (top) and closest precedents 36,37 (bottom).
Many years later, and in a different lab, we investigated the one-pot Michael addition/cyclocondensation of o-nitrophenylacetonitrile (1) with N-methylmaleimide (10) (Scheme 4). Surprisingly, this reaction gave the quinaldic acid 31 in moderate yield, after acidic workup. It seems likely in this case that adventitious water either provided an alternative nucleophile for the ring opening of intermediate 26 (Scheme 2), or led to saponification of 21/27 under the reaction conditions. However, drying of the K2CO3 (overnight 400 °C) and ethanol (over freshly activated sieves for >24 h) had no bearing on the outcome of the reaction, which raises the possibility that carbonate is the nucleophile that ring-opens intermediate 26 in these instances. In contrast to the two-step synthesis of 28 (Table 2 and Scheme 3), heating 1, diethyl maleate (12) and K2CO3 from the outset gave a complex mixture of products, suggesting that, with some electrophiles, one-pot annulations of this type may best be carried out with a room-temperature Michael-addition step prior to heating to induce the cyclocondensation reaction.
An attempt to effect an analogous one-pot annulation reaction of 1 with ethyl acrylate (9) gave predominantly the Michael adduct 14 ( Table 2). The slow/lack of annulation of adduct 14 in refluxing ethanol suggests that: 1) intramolecular nucleophilic attack of enolates on the nitro group is reversible; and, 2) an adjacent electron-withdrawing group facilitates tautomerization, e.g., step 24 → 25 in scheme 2, enabling irreversible dehydration/aromatization. In line with these hypotheses, a switch to the higher boiling solvent 1-butanol gave 4-cyanoquinaldic acid (32) (Scheme 5); once again, ester saponification due to adventitious water seems likely. Difficult multistep purification exacerbated by poor solubility contributed to the low isolated yield of 32; nevertheless, further experimentation is required to make the annulation of 1 with Michael acceptors lacking a β-electron-withdrawing group synthetically viable. Scheme 4. Successful and attempted annulation reactions of 1 with Michael acceptors. Each reaction was followed by a weakly acidic aqueous work-up. A representation of the X-ray crystal structure of 31 is shown on the right. Displacement envelopes are at 50% probability amplitude, with hydrogen atoms assigned arbitrary radii.
Scheme 5. One-pot annulation reaction of 1 and 9 requires a higher boiling solvent.
We also briefly investigated reactions of the related enolate generated from ethyl o-nitrophenylacetate (33), with N-methylmaleimide (10) (Scheme 6). Interestingly, with K2CO3 in ethanol at room temperature, no color associated with the relevant carbanion was observed, and no Michael adduct 34 was detected. Upon heating under reflux, N-hydroxyindole 35 was the predominant product, as confirmed with an X-ray crystal structure (Scheme 6). The same transformation was achieved in higher yield at room temperature with stronger base.
Key intermediates in the proposed mechanism for the formation of 35 are shown in Scheme 7. Michael addition of ester 33 to 10 gives adduct 34. A second deprotonation generates enolate 36, which cyclizes onto the nitro group, affording spirocycle 37. This species is presumably in equilibrium with the oxoammonium ion 38. Reversible attack of liberated, or adventitious, hydroxide on the more sterically hindered succinimide carbonyl group of 38 gives 39, which is poised to undergo a retro-Claisen-like ring scission, affording nitronate 40. Decarboxylation and tautomerization/protonation then give the observed N-hydroxyindole 35. Scheme 6. Annulation reactions of 33 and N-methylmaleimide (10). The representation of the X-ray crystal structure of N-hydroxyindole 35 has displacement envelopes at 50% probability amplitude, with hydrogen atoms assigned arbitrary radii.
The two different annulation pathways -quinolines from 1 and an N-hydroxyindole from 33 -probably result from the different steric demands of the benzylic cyano and ethoxycarbonyl groups, which influence which nucleophilic center of the possible enolate intermediates can achieve the geometry required for nucleophilic attack on the nitro group.

Scheme 7.
Key intermediates in the proposed mechanism for the formation of N-hydroxyindole 35.

Conclusions
The carbanion mildly generated from o-nitrophenylacetonitrile is a good nucleophile for diastereoselective Michael additions. More forcing conditions result in an unusual cyclocondensation affording 2,(3),4trisubstituted quinolines in moderate to good yields. Analogous isoquinoline syntheses involving other o-nitrobenzylic carbanions are likely possible, provided the benzylic substituent is not too sterically demanding. In contrast, higher temperatures or stronger bases are required to elicit reaction of ethyl o-nitrophenylacetate with N-methylmaleimide (and presumably other Michael acceptors), and an N-hydroxyindole product predominates in this case. The reactions described herein provide rapid access to substituted benzo-fused heterocycles that would otherwise require multi-step synthetic routes, and expand the rich chemistry of the carboxylic acid derivatives of o-nitrophenylacetic acid.

Experimental Section
General. General experimental details are as reported previously. 33,38 Reactions were conducted under ambient conditions unless otherwise indicated. Stated temperatures refer to the temperature of the heating bath. Dry solvents were stored in sealed bottles over activated type 3A sieves for at least 24 h before use.

2-(2-Nitrophenyl)propanenitrile (8). Method 1.
K2CO3 (74 mg, 0.54 mmol) was added to a stirred solution of o-nitrophenylacetonitrile (1) (18 mg, 0.11 mmol) in dry EtOH (2 mL) under argon. MeI (0.03 mL, 0.5 mmol) was added to the purple suspension and stirring was continued at 40 °C for 3 h, over which time the solution turned pale yellow. The reaction mixture was acidified with 1 M HCl (5 mL) and diluted with water (10 mL), then extracted with EtOAc (4 × 5 mL). The extract was dried and evaporated to give a yellow oil, which was purified by flash chromatography. Elution (EtOAc/hexanes, 3:17) gave 8 as a yellow oil (6 mg, 35%), identical with the material described below. Method 2. K2CO3 (429 mg, 3.11 mmol) was added to a stirred solution of o-nitrophenylacetonitrile (1) (323 mg, 1.99 mmol) in dry DMF (15 mL) under argon. MeI (0.24 mL, 4.0 mmol) was added to the purple suspension and stirring was continued at 50 °C for 90 min, over which time the solution turned pale yellow. The mixture was acidified with 1 M HCl (10 mL) and diluted with water (20 mL), then extracted with EtOAc (3 × 15 mL). The extract was dried and evaporated to give a yellow oil, which was purified by flash chromatography. Elution (EtOAc/hexanes, 3:17) gave 8 as a yellow oil (288 mg, 80%), identical with the material described below.  (32). Ethyl acrylate (9) (68 μL, 0.64 mmol) was added to a stirred mixture of K2CO3 (125 mg, 0.902 mmol) and o-nitrophenylacetonitrile (1) (97 mg, 0.60 mmol) in dry n-butanol (20 mL). The reaction mixture was then heated under reflux under N2 for 24 h, then cooled and acidified with 1 M HCl (10 mL), diluted with water (30 mL) and extracted with EtOAc (3 × 20 mL). The extract was dried and evaporated to give a brown residue (92 mg), which was triturated with minimal CHCl3. The remaining brown solid was dissolved in EtOAc (20 mL) and extracted with half-saturated NaHCO3 (3 × 20 mL). The basic extract was acidified with 4 M HCl (20 mL) and extracted with EtOAc (3 × 20 mL). The organic extract was dried and evaporated to give a brown solid, which was subjected to preparative TLC. Development NaH 60% dispersion in mineral oil (24 mg, 0.61 mmol) was added to dry EtOH (6 mL) with stirring. Ester 33 (191 mg, 0.913 mmol) was added portion-wise. No color change was observed. After 5 min, N-methylmaleimide (10) (59 mg, 0.53 mmol) was added to the solution. After 24 h TLC indicated the reaction was incomplete and a second portion of Nmethylmaleimide (10) (51 mg, 0.46 mmol) was added and the reaction mixture was stirred for a further 24 h. The reaction mixture was acidified with 1 M HCl (10 mL), diluted with water (20 mL) and extracted with EtOAc (3 × 10 mL). The extract was dried and evaporated to yield a pale-brown oil, which was subjected to flash chromatography. Elution (EtOAc/hexanes, 1:4) and recrystallisation from DCM/hexanes gave the Nhydroxyindole 35 as a pale-yellow solid (114 mg, 45%), identical with the material described below. Method 2. N-Methylmaleimide (10) (138 mg, 1.24 mmol) was added to a stirred mixture of K2CO3 (127 mg, 0.919 mmol) and ester 33 (124 mg, 0.595 mmol) in dry EtOH (20 mL). The reaction mixture was then heated under reflux under N2 for 16 h, cooled to rt, and treated with additional N-methylmaleimide (10) (71 mg, 0.64 mmol) before being heated under reflux for a further 24 h. The reaction mixture was cooled, acidified with 1 M HCl (5 mL), diluted with water (30 mL) and extracted with EtOAc (3 × 20 mL). The extract was dried and evaporated to yield an orange-red oil, which was subjected to flash chromatography. Elution (EtOAc/hexanes, 25:75)