Origin of regioselectivity in electrophilic reaction of ambident enaldimines

0040-4039(94}EO20l-8

Kiyoshi Tornioka,* Tetsuji 0kamot0~ Motomu Kanai, and Hiroshi Yamataka 77reIns&u& of SeienZ@ andIru&st?MResearcli, Osaka U&e&&

ibiuak~ Osaka 567, Japan

ne o&wed substituent-dependent regioseiectiviLy f&Z- vs I,l-additionl iir the &cmphil& ma&~~of antbiaknt imines1-3 B mtioMl& undemood &Ythe relative m@fnitude

Abstract:

of the LUMO coelpriena of the Mnes. We have recently reported the succemful chiral ~g~-m~a~ ~~ 1,2- and l,~a~tion reactions of organolithiums to various imines of aryl and a#-unsaturated aldehydes.1 Thus, organolithiums add to the cyclohexylimmes 1.2 (R=c-c6H11) in 1.4fashion to give, after hydrolysis, the oorresponding aldehydes in up to 99% ee and in high to good yield& On the other hand, the reactions with arylimines 1, 2 (R--Q-MeOPh) exclusively afford 1,2-adducts in up to 99% ee and in nearly quantitative yields, giving us a good methcd for the preparationof chiral atnines.3.4 It is also interesting in that the reactions with both cyclohexyl- and arylimines of 1-fluo~naph~alene-2-c~bal~hy~ 3 exclusively proceed in 1,4-fashion to provide nucleophik aromatic substitution products in high yields,5 Since deep understanding of the factors governing regioselectivity is a long-standing problem6.7 and essential for further development of the much more effective catalytic asymmetric reactions of the eneimines, we carried out molecular orbital calculations of l-3.8

AR s-cis and s-tram &rwtum 1,2 were fully optimized with the ab initio (HWSTO-3G) method under the Cs const&nt. PM3 o@nization (precise mode} for l-3 (s-c&) and HFf3-21G for 2b (R=ph, s&) were also carried out for oomparison. The optimized geometries of s-crk conformation are shown in the Figure I together with the absolute values of the LUMO coefficients and the total oharges (in parentbe& at the each reaction site (2- and 4-positions). Comparison between la (R=Me, R’=H) and lb (R=Ph, R’=H) revealed that the experimefitaly obseroed regioseleotivity is rationalized by the relative magnitu& of the LUMO coefficients, The arylimine lb has a much larger coefficient at the 2-position than the 4-position, con&tent with the exciuaive 1.2.addition for lb,3 while the dcient is larger at the 4-position for the alkylii la. for which the l,&dduct is the major product.2 Similarly, the 4-position has a much larger coefficient than the 2position for the alkylimine 2a (R=Me), which gave only the 1,4-adduct, the product thermodynamically less stable than the 1.2~adductdue to the breakdown of the naphthakne aromaticity. In contmst, the size of the LUMO coefficient is similar at the 2- and 4-positions fa the arylimine 2b (R=Ph), for which the 1,2addition is the preferredreaction,3 probably due to the aromaticity. 1891

1892

Results calculation

for the s-trams (HF/3-21G,

m/STO-3G.

for the substituent-dependent electron withdrawing

isomers

the same as s-cis.

and PM3) has little influence

reversal

and donating

are nearly

of the LUMO coefficient

and the difference on a qualitative

magnitude

in the method

conclusion.

is probably

prom

the substituent-dependent

efforts

toward

regioselectivity

much more effective

of the LUMO coefficient of the ambident

asymmetric

reactions

to

is one of the major factors

ene-imines.

by varying

attached

Based on the result,

imine-substihrents

are in

in our laboratories.9

la ObMe, R’4-i)

lb FM’h,

Figure

charges

::

to the

nature of the aryl and alkyl groups, respectively.

Thus, it became clear that the relative magnitude further

The reason

ascribable

PM3 calculations for 3a,b (R=Me, Ph) revealed the larger coefficient at the carbon fluorine than that at the imine carbon, indicating the experimentally observed SNAr reaction.5 governing

of

R’=H)

2a (&Me)

2b (RPh)

Optimized geometries (Chem3D presentation), (in parentheses) at the reaction sites of 1-3.

I.

- 3b (R=Ph)

LUMO

coefficients

and total

References and Notes Tomioka. K.&J&~& 1990,541. Tomioka. K.; Shindo, M.; Koga. K. J. Am. Chem. Sot. 1989, Ill, 8266; Rawson. D. J.; Meyers, A. I. J. Org. Chem. 1991.56.2292.

3. Tomioka. K.; Inoue, I.; Shindo. M.: Koga. K. Te~&e&on Lett. 1990,31, 6681; Idem. ibid. 1991 32 3095houe, I.; Shindo. M.; Koga. K.; Tomioka, K. Te@&edron: Asymmehy 1993,4, 1603; Idem. T&&,&on & 4. litzmi, S.; Yanaka, I-I.; Ha&is&a. C.; Ito. K.J. Chem 5+x., PerkzhDtzns.11991.1341; Katr&ky, A. R.: Harris P. A. Tefmkedtwz: Asymmetry 1992,3, 437; Soai, K.; Hatanaka. T.; Miyazawa. T. J. Chem. Sot.. C&k Commun. 1992.1097. Shindo, M.; Koga, K.; Tanioka, K.J. Am. Chem. Sot. 1992.114.8732. 2: Similar regio++vity has been reported in the reaction of ketoitnines without clear explanation. Keuk, B. P.; Mmze, B.; L. $vnthesS 1977,638. 7. For theoreh -Y treatment for regioselectivily of elqc addition to an ambident ketone or aldehyde, see Lefour, +-M.; Laupy, A. Tebzzheahn 1978.34,2597, and &xence~ ’ 8. ckkdatmns were carried out withIm! R!.%O~ using the GAUSS%9=&: Frisch M J - Trucks G W.; Head-. M.; Gill, P. M. W.; Wang. M. byw .; Faresman. J. B.; Johnson, B. G.; Schlegei, ti. ii.: Robb,‘M: , E S.; Gomperrs. R.; An&es. J. L; Raghava&ri, K.; B . J. S.; Gonzakz C.: M&in, R. L: Fox, $\T ezza , D. J.; Baker, J.; Stewart, J. J.; Pople, J. A. GAU.!WAh 7 9 , RevMt.v~8, Gaussianinc., F’&+&&, USA. PlbOcaluualationwasalsocarriedoutwith le Macintosh (MOPAC Ver. 6.00 (QCPE No. 4451, Stewart, J. J. QCPE Bull. 1990,10(4), 86; Himno, T. % J E Newsletter, 1991.2(4). 39; Revised as Ver. 6.02 by Toyoda, J. for Apple Mac&to&). 9. We are gmteful to Gras-in-Aid for Scientific Research, Ministry of Education, Science and Culture, Japanend ‘IYEFugakuTrustfor Medicinal Research, for jxutial support of this work.

(Received in Jquan 29 Ocmber 1993; accepted 10 January MM)