Asymmetric deprotonation and dearomatising cyclisation of N-benzyl benzamides using chiral lithium amides: formal synthesis of (–)-kainic acid

Abstract

Chiral lithium amides deprotonate N-benzyl benzamides enantioselectively, initiating an asymmetric dearomatising cyclisation to enantiomerically enriched isoindolinones.

---

We have previously reported¹ that lithiated N-benzyl amides cyclise at temperatures approaching 0 °C (−78 °C in the presence of HMPA^2,3) to yield dearomatised products which take the form of partially saturated isoindolinones (Scheme 1). These heterocycles, with their three new stereogenic centres and versatile functional groups, are versatile synthetic intermediates, and we have published a synthesis of (±)-kainic acid using the cyclisation of 1b as a key step.⁴ In this paper we report an asymmetric cyclisation achieved† by means of chiral lithium amide bases which proceeds via an asymmetric benzylic deprotonation.

Scheme 1 Dearomatising cyclisation of lithiated N-benzyl benzamides. (i) [for (±)-5] LDA, −78–+25 °C, 1.5 h; (ii) [for (−)-5] chiral lithium amide, −78–+25 °C, 1.5 h; (iii) NH₄Cl; (iv) HCl, H₂O.

α-Amido benzyllithiums related to 2 may be generated enantioselectively with s-BuLi–(−)-sparteine and are configurationally stable over a period of 10 h at −78 °C.^5,6 However, while treatment of 1a with s-BuLi–(−)-sparteine⁷ in THF gave the cyclised product 5a, it turned to be racemic. Insufficient solubility prevented the use of Et₂O or t-BuOMe; toluene turned out to be too acidic, becoming lithiated itself and leading to benzyl ketones; no reactions took place in cumene.

As we knew that LDA deprotonates 1a at temperatures approaching −40 °C,¹ we turned to chiral versions of LDA, namely the lithium amides 6·Li, 7·Li, 8·Li and 9·Li₂.⁸ Each base was used with the amide 1a, and the crude product was hydrolysed to the enone with aqueous hydrochloric acid. The products 5a–d were purified by flash chromatography, and the yields and enantiomeric excesses obtained are shown in Table 1. The best results were with 7·Li: not only did it give the highest enantiomeric excesses, but it was also the easiest to remove from the product mixture by a simple acid wash.‡ A single recrystallisation was enough to return the products 1a and 1b in >99% ee. Attempts to improve enantioselectivity by changing the temperature régime (entries 2, 10) were unsuccessful. The absolute stereochemistry of the products (which were all laevorotatory with 6·Li–8·Li and dextrorotatory with 9·Li₂) was determined by X-ray crystallography of the enone (−)-10 obtained by bromination (Br₂, Et₃N, CH₂Cl₂: 80%) of enantiomerically pure (−)-5a.

Table 1 Asymmetric cyclisation of N-benzyl benzamides

Entry	Starting material	Base	(−)-5 yield (%)	(−)-5 ee^a (%)	(−)-5 ee^b (%)
a Ee after flash chromatography. b Ee after recrystallisation. c Reaction carried out −78–−60 °C, 16 h. d (+)-5 is the major enantiomer. e Reaction carried out −50 °C–+25 °C.
1	1a	6 ·Li	73	80	>99
2	1a	6 ·Li^c	65	50	—
3	1a	7 ·Li	72	80	>99
4	1a	8 ·Li	72	17	—
5	1a	9 ·Li2	38	−75^d	—
6	1b	6 ·Li	64	75	>99
7	1b	7 ·Li	72	81	>99
8	1b	9 ·Li2	23	−30^d	—
9	1c	6 ·Li	59	73	–
10	1c	6 ·Li^e	67	60	—
11	1d	7 ·Li	87	84	—
12	11	7 ·Li	70	60	0

---

Enantioselective cyclisation of the para-substituted amide 11 was also successful under these conditions, though unfortunately it was impossible to improve the enantiomeric excess of the product by recrystallisation. However, amides bearing ortho substituents such as 12 or 13 gave either low yields or low enantioselectivities, probably because such ortho-substituted tertiary aromatic amides are chiral compounds at low temperatures.§

As for the origin of the enantioselectivity, two limiting possibilities present themselves. The first is that the reaction proceeds via an asymmetric deprotonation of the achiral starting material 1, generating an enantiomerically enriched organolithium which is configurationally stable on the timescale of the reaction and which cyclises stereospecifically. The second is that the reaction is a stereoselective cyclisation of an intermediate organolithium (which may or may not be configurationally stable) under the asymmetric influence of the amine such as 7·H regenerated after deprotonation. The known configurational stability of lithio amides similar to 2^5,6 at least does not rule out the first mechanism. To distinguish between these two possibilities, we carried out two experiments.⁹

In the first experiment, we treated 1a with t-BuLi to generate organolithium (±)-2a, and then added chiral amine 7·H, recreating the conditions of the asymmetric cyclisation but with a starting organolithium that must, at least initially, be racemic (Scheme 2). After an aqueous quench and acidic work-up, the product 5a of this reaction was nearly racemic, proving that the enantioselectivity in the asymmetric reaction is determined before the cyclisation step.

Scheme 2 Mechanism of the cyclisation (i) t-BuLi, THF, −78 °C; (ii) 6·H, −78 °C–+25 °C; (iii) NH₄Cl then HCl, H₂O; (iv) CD₃OD; (v) 6.Li, THF, −78–+25 °C.

In the second experiment, we used racemic, deuterium-labelled (±)-d-1a (obtained by deuteration of (±)-2a, Scheme 2) as the starting material for the asymmetric cyclisation. Cyclisation of (±)-d-1a with 6·Li gave racemic material 5a which was >99% deuterated — a result consistent only with a primary kinetic isotope effect k_H/k_D of at least 100 which has completely overturned the enantioselectivity of the chiral base.^10,11 This in turn implies the enantioselective step must be a step subject to a primary kinetic isotope effect, i.e. the deprotonation.

While the s-BuLi–(−)-sparteine has commonly been used for the enantioselective removal of one of a pair of enantiotopic geminal H atoms to give a chiral organolithium,⁷ the less basic chiral lithium amide bases are much more frequently employed to generate chiral planar anions, such as enolates, by selecting between enantiotopic enolisation sites of prochiral carbonyl compounds.⁸ By carrying out the asymmetric deprotonation at temperatures too low for cyclisation to occur (<−40 °C), we hoped to observe the asymmetric formation and external quench of a chiral, non-planar organolithium generated with a chiral lithium amide, a reaction observed before only with Cr(CO)₃-complexed benzyl groups.^12,13 When we treated the amide 1a with 7·Li at −78 °C, warmed to −40 °C over 1 h to ensure deprotonation, and then added an electrophile, the ee of the product was dependent on the electrophile used (Scheme 3). Presumably the rate of racemisation of the organolithium 2a is similar to the rate of reaction with an electrophile, and racemisation outpaces attack on MeI but not on CO₂.¶ We conclude that asymmetric deprotonation of one of the pair of enantiotopic protons of 1 gives a benzylic organolithium with sufficient configurational stability to cyclise stereospecifically but insufficient configurational stability to prevent partial or complete racemisation on the timescale of its addition to an external electrophile.

Scheme 3 External electrophiles (i) 7·Li, −78–−40 °C, 1 h; (ii) E⁺.

Our published route to (±)-kainic acid⁴ uses 18 as an intermediate. As a demonstration of the utility of the asymmetric cyclisation, we therefore converted 5d (which was produced in 84% ee by the most enantioselective cyclisation we could find, Table 1, entry 11) to 18 in four steps (Scheme 4). 5d could not be recrystallised to enantiomeric purity, but after addition of Me₂CuLi and deprotection/hydrolysis of the enol ether 15, recrystallisation of pyrrolidinone 16 gave enantiomerically pure material in an overall yield of 48% from 5d. Boc protection of 16 and oxidation of the m-MeOC₆H₄ ring of 18 gave the enantiomerically pure ester 18, which we have converted to kainic acid in five steps.⁴

Scheme 4 Formal synthesis of (−)-kainic acid (i) Me₂CuLi, Me₃SiCl; (ii) CF₃CO₂H; (iii) recrystallise; (iv) Boc₂O, DMAP, Et₃N, CH₂Cl₂; (v) RuCl₃, NaIO₄, H₂O, MeCN then CH₂N₂.

We are grateful to the EPSRC for a studentship (to CJM), to Dr C. S. Frampton for determining the absolute stereochemistry of 10, and to Miss K. Hebditch for studying the cyclisation of 12.

Notes and references

1. J. Clayden, C. J. Menet and D. Mansfield, Org. Lett., 2000, 2, 4229.
2. A. Ahmed, J. Clayden and S. A. Yasin, Chem. Commun., 1999, 231.
3. J. Clayden, K. Tchabanenko, S. A. Yasin and M. D. Turnbull, Synlett, 2001, 302.
4. J. Clayden and K. Tchabanenko, Chem. Commun., 2000, 317.
5. S. Wu, S. Lee and P. Beak, J. Am. Chem. Soc., 1996, 118, 715; Y. S. Park, M. L. Boys and P. Beak, J. Am. Chem. Soc., 1996, 118, 3757.
6. M. Schlosser and D. Limat, J. Am. Chem. Soc., 1995, 117, 12342; C. Barberis, N. Voyer, J. Roby, S. Chénard, M. Tremblay and P. Labrie, Tetrahedron, 2001, 57, 2965.
7. D. Hoppe and T. Hense, Angew. Chem., Int. Ed., 1997, 36, 2282.
8. P. O′Brien, J. Chem. Soc., Perkin Trans. 1, 1998, 1439.
9. For a similar discussion, see: P. Beak, A. Basu, D. J. Gallagher, Y. S. Park and S. Thayumanavan, Acc. Chem. Res., 1996, 29, 552.
10. D. Hoppe, M. Paetow and F. Hintze, Angew. Chem., Int. Ed., 1993, 32, 394.
11. D. R. Anderson, N. C. Faibish and P. Beak, J. Am. Chem. Soc., 1999, 121, 7553.
12. E. L. M. Cowton, S. E. Gibson (née Thomas), M. J. Schneider and M. H. Smith, Chem. Commun., 1996, 839.
13. R. A. Ewin, A. M. MacLeod, D. A. Price, N. S. Simpkins and A. P. Watt, J. Chem. Soc., Perkin Trans. 1, 1997, 401.
14. A. Ahmed, J. Clayden and M. Rowley, Chem. Commun., 1998, 297.
15. A. Ahmed, R. A. Bragg, J. Clayden and K. Tchabanenko, Tetrahedron Lett., 2001, 42, 3407.
16. R. A. Bragg, J. Clayden, M. Bladon and O. Ichihara, Tetrahedron Lett., 2001, 42, 3411.
17. R. A. Bragg and J. Clayden, Tetrahedron Lett., 1999, 40, 8323.
18. F. Hammerschmidt and A. Hanninger, Chem. Ber., 1995, 128, 823.
19. A. Ahmed, R. A. Bragg, J. Clayden, L. W. Lai, C. McCarthy, J. H. Pink, N. Westlund and S. A. Yasin, Tetrahedron, 1998, 54, 13277; J. Clayden and J. H. Pink, Tetrahedron Lett., 1997, 38, 2561.
20. R. A. Bragg and J. Clayden, Tetrahedron Lett., 1999, 40, 8327.

---

Footnotes

† Related cyclisations of N-benzyl-1-naphthamides^14,15 have been made asymmetric by the use of a chiral auxiliary¹⁶ or chiral starting material.¹⁷

‡ 7·H was made in 61% yield by stirring (S)-α-methylbenzylamine with isopropyl bromide.¹⁸

§ Tertiary 1-naphthamides and ortho-substituted benzamides exhibit atropisomerism at low temperature,¹⁹ and hence 12 and 13 must be viewed as racemic compounds under the conditions of the deprotonation. This feature has a governing role in the cyclisation of chiral 1-naphthamides.²⁰

¶ The absolute stereochemistry of (+)-14 (E = CO₂H) is unknown.

---