Enantioselective Synthesis of 4- and 6-Azaindolines by a Cation-Directed Cyclization.

Functionalized 4- and 6-azaindolines are accessible with high levels of enantioselectivity by the cation-directed cyclization of aminopyridine-derived imines via phase-transfer catalysis. The extension of this methodology to diastereoselective cyclizations is also described.

As a bioisostere of the n class="Chemical">indoline moiety, pan> class="Chemical">azaindolines represent attractive synthetic targets which are often incorporated into pharmaceutically active compounds.[1] Relatively few general procedures exist for their synthesis,[2] with the majority of methods directed toward the preparation of 5- or 7-azaindolines. Reports of 4-azaindolines are particularly scarce,[3] with existing synthetic routes involving base-mediated nucleophilic attack of nitrogen onto tosylates,[1a] acid catalyzed N-cyclization onto alkenyl intermediates,[4] intramolecular carbolithiation,[5] palladium-catalyzed C–H activation,[6] or radical cyclization.[2g,7] Moreover, to the best of our knowledge, only two enantioselective syntheses of azaindolines have been reported to date.[6b,8]

Cation-directed enantioselective cyclizations have previously been employed in the synthesis of n class="Chemical">indolines, pan> class="Chemical">indanes, and complex polycyclic systems,[9−11] and we envisaged that this approach could be extended to encompass the preparation of a series of enantioenriched azaindoline scaffolds. The feasibility of this transformation was investigated through the cyclization of imine 1, which was prepared in three simple steps.[12] An initial attempt at cyclization using aqueous potassium carbonate under phase-transfer conditions in the presence of N-benzylcinchonidinium chloride 3 led to a promising 90:10 er at room temperature (Table ). Attempts to lower the temperature of the reaction to improve enantioselectivity resulted in reduced reactivity and poor conversion of starting material. However, a switch to solid bases resulted in smooth conversion to the desired azaindoline 2 at lower temperatures, with solid KOH at −30 °C proving to be the optimum base (93:7 er). We subsequently evaluated several cinchonine-derived ammonium salts as catalysts, initially focusing on tuning the N-substituent. Catalyst 4, bearing a 9-anthracenylmethyl group, gave an increased er (95:5), and N-2-naphthylmethyl cinchonidinium bromide 5 afforded the desired azaindoline in 97:3 er. We also varied the electronic properties of the pendant N-benzylic group (as in 4-methoxy catalyst 6 and 4-trifluoromethyl catalyst 7), but this resulted in no increase in enantioselectivity over catalyst 5 (93:7 er vs 92:8 er respectively). With optimized conditions for the 6-azaindoline cyclization in hand, the scope of this transformation was then investigated (Figure ). The reaction is compatible with a range of substituted aryl imines, with high enantioselectivities observed with ortho- and para-substituted aromatic substrates.[13] For instance, ortho-substituted substrates 9, 10, and 11 cyclize in high er (98:2, 93:7, and 97:3 er respectively). Para-substituted substrates cyclize equally effectively (p-chloro 12 94:6 er, p-bromo 13 94:6 er, p-nitro 14 95:5 er). A substrate bearing an electron-withdrawing group in the meta- position affords product 15 with only moderate er; however, an electron-donating substituent in the same position is well tolerated in the case of 16 (95:5 er).

Table 1

Reaction Optimizationa

base	catalyst	temp, °C	er
K2CO3 (aq.)	3	rt	(+) 90:10
K2CO3 (aq.)	3	–15	nr
CsOH·H2O (s)	3	–15	(+) 73:27
KOH (s)	3	–15	(+) 88:12
KOH (s)	3	–30	(+) 93:7
KOH (s)	4	–30	(+) 95:5
KOH (s)	5	–30	(+) 97:3
KOH (s)	6	–30	(+) 93:7
KOH (s)	7	–30	(+) 92:8
KOH (s)	8	–30	(+) 92:8

Conditions: 1.0 equiv of base, 10 mol % catalyst, 3:1 toluene/CH2Cl2. Er determined by chiral stationary phase HPLC; (+) refers to sign of the optical rotation; nr = no reaction.

Figure 1

Scope of enantioselective 4-azaindoline formation. Yields are for isolated materials; er determined by chiral stationary phase HPLC.  Yield calculated over two steps from aniline.

Conditions: 1.0 equiv of base, 10 mol % catalyst, 3:1 n class="Chemical">toluene/n class="Chemical">CH2Cl2. Er determined by chiral stationary phase HPLC; (+) refers to sign of the optical rotation; nr = no reaction.

Scope of enantioselective n class="Chemical">4-azaindoline formation. Yields are for isolated materials; er determinpan>ed by chiral stationary phase HPLC.  Yield calculated over two steps from n class="Chemical">aniline.

Bulky aromatic groups (as in 17) and heterocycles (as in 18) are also effective substituents in the cyclization (95:5 er and 97:3 er respectively). n class="Chemical">Imines derived from pan> class="Chemical">alkyl aldehydes were found to be particularly reactive; consequently, reactions involving these aldehydes were performed without purification of the intermediate imine. This approach afforded 4-azaindolines bearing branched and cyclic alkyl substituents with high enantioselectivity (19; 93:7 er and 20; 98:2 er). We were also able to incorporate straight-chain alkyl groups (as in 21, 88:12 er) with moderate enantioselectivity. Single crystal X-ray diffraction experiments on the major enantiomer of 10 allowed the absolute configuration of the cyclization product to be determined, and it is assumed by analogy that the stereochemical outcome of the reaction is the same in all other cases.[14] Having demonstrated the scope of the reaction for the synthesis of 4-azaindolines, we examined the extension of this approach to the synthesis of isomeric 6-azaindolines.[15] Precursors for cyclization to the 6-azaindoline core were prepared by a similar reaction sequence to that applied in the 4-azaindoline case.[16] In general, the synthesis of imine intermediates in this system was significantly more challenging than in the 4-azaindoline case.[17] However, the efficiency and enantioselectivity of the cyclization of these materials as demonstrated by 22 (90:10 er) and 23 (84:16 er) were similar to those observed in the 4-azaindoline system (Figure ).

Figure 2

Cyclizations of 6-azaindolines. Yields are for isolated materials; er determined by chiral stationary phase HPLC.

Cyclizations of n class="Chemical">6-azaindolines. Yields are for isolated materials; er determinpan>ed by chiral stationary phase HPLC.

It was possible to isolate an n class="Chemical">anisaldehyde-derived imine, albeit inpan> low yield. The observed er for cyclization of this substrate to generate 24 was slightly lower (er 91:9) than the correspondinpan>g n class="Chemical">4-azaindoline substrate 15 (er 97:3).

We have previously demonstrated that substrates which contain distinct electron-withdrawing groups rather than two n class="Chemical">isopropyl esters canpan> cyclize to form pan> class="Chemical">indoline products containing two contiguous stereocenters, one of which is an all-carbon quaternary center.[9,18] These cyclizations can proceed with excellent diastereoselectivity, and we rationalized that this would form an attractive route to stereochemically complex azaindolines (Scheme ).

Scheme 1

Diastereoselective Cyclization

Dr determined by 1H NMR spectroscopy of the crude reaction mixture.

Dr determined by n class="Chemical">1H n class="Chemical">NMR spectroscopy of the crude reaction mixture.

We synthesized a suitable substrate 25 to probe the potential for diastereo- and enantioselective cyclization with a chiral but racemic α-aryl acetonitrile substituent, rather than the n class="Chemical">malonate derivative previously employed. Treatmenpan>t of pan> class="Chemical">imine 25 with CsOH·H2O in the presence of tetra-n-butylammonium hydrogensulfate (TBAHS) resulted in complete conversion of starting material and the formation of a mixture of diastereomers, which were separated by chromatography and analyzed by single crystal X-ray diffraction to determine their relative stereochemistry. Under these conditions the reaction favored formation of the diastereomer 26 with the phenyl groups in an anti- relationship in 2.7:1 dr. The reaction was then attempted utilizing the previously optimized conditions in the presence of a chiral phase-transfer catalyst (Table ). Cyclization of 25 with previously identified cinchonidinium catalyst 4 led to 26 and 27 in significantly lower enantioselectivity than previously observed in the cyclization of malonate-derived substrates such as 1.[19]

Table 2

Attempted Diastereo- and Enantioselective Cyclizationsa

catalyst	temp, °C	dr (26:27)	er (26)	er (27)
4	rt	2.0:1.0	racemic	52:48
4	–10	1.0:3.0	59:41	56:44
4	–30	1.0:4.5	64:36	58:42
5	–30	1.0:1.5	77:23	62:38

Dr determined by 1H NMR spectroscopy of the crude reaction mixture; er determined by chiral stationary phase HPLC.

Dr determined by n class="Chemical">1H n class="Chemical">NMR spectroscopy of the crude reaction mixture; er determined by chiral stationary phase HPLC.

Further variation of the reaction conditions demonstrated that diastereoselectivity is dependent on the temperature. This led us to suspect that that the cyclization may be reversible under the reaction conditions, with the thermodynamically favored anti-phenyl isomer 26 predominating at higher temperatures. To test this hypothesis, n class="Chemical">imine 25 anpan>d pure samples of both product diastereoisomers 26 anpan>d 27 were treated with solid pan> class="Chemical">CsOH·H2O and TBAHS at different temperatures (Scheme ). 1H NMR spectroscopic analysis of crude reaction mixtures revealed that imine 25 afforded a diastereoisomeric ratio of 2.7:1.0, 26:27 under these conditions. This ratio of products is close to that observed upon treatment of either of the pure diastereoisomers 26 or 27 under these conditions (2.6:1.0 and 2.1:1.0 26:27 respectively). This is consistent with the cyclization reaction being reversible and the anti-phenyl diastereomer 26 being the thermodynamic product of the reaction. The diastereoisomeric ratios observed during the equilibration experiments suggest that the ground state energy difference between the products is relatively small. Monitoring the equilibration reaction with TBAHS at −30 °C using 1H NMR spectroscopy showed that interconversion between isomers 26 and 27 still occurred at this temperature, but at a much slower rate than at rt. This rate of equilibration is slowed further when chiral phase transfer catalysts are employed, suggesting that equilibration has a small role in the enantioselective cyclization but that it is not solely responsible for the poor selectivity observed. Unfortunately, attempts to lower the reaction temperature further to minimize equilibration resulted in incomplete cyclization.

Scheme 2

Equilibration Experiments

Dr determined by 1H NMR spectroscopy of the crude reaction mixture.

Dr determined by n class="Chemical">1H n class="Chemical">NMR spectroscopy of the crude reaction mixture.

In conclusion we have developed an efficient route to enantiomerically enriched n class="Chemical">4-azaindolines bearinpan>g aryl or alkyl substituenpan>ts via a cationpan>-directed asymmetric cyclizationpan>, anpan>d this route was successfully extenpan>ded to the synpan>thesis of pan> class="Chemical">6-azaindolines. From a mechanistic perspective this reaction is likely to proceed via a kinetically controlled 5-endo-trig mechanism rather than the alternative electrocyclization manifold. This approach is complementary to methods reported for azaindoline synthesis to date and allows access to a diverse range of useful synthetic intermediates.