100445

Gilman et al.

3286 J. Org. Chem., Vol. 58, No. 12, 1993

Table I. Blological Data for Imidazo! 1,5-aJi 1,4]ben zodiaupinee

compd	(*H]diaztpam binding; IC*o‘nM	antipentyleneteuszole; EDjcr mg kg
2	5	1
3	4	0.1
10	88	inactive
11	>1000	inactive
12	24	5
12a	19.5	1.9
12b	250	131
18	>1000	inactive
19	>1000	&5
19m	>1000	inactive
19b	>1000	1.6
25	13	3.4
Um	7	0.18
25b	220	6.6
30	>1000	21
31	54	138
3ls	26.5	55
3lb	540	78

* The method deacribed in ref 3 w aa u«»d for this aaaay. ‘ A modiflcation of the Evaract and Richarda¹'' matbod wat usad for this aaaay. Reaulta are reported aa 95% fiducial limits.

i*. X ■ F, R ■ H

ognition by the benzodiazepine receptor in the imida-zobenzodiazepine senes of compounds.

This report describes the synthesis and separation of the diaatereomeric imidazo (1,5-a) [ 1 ,4] benzodiazepine* 10 and 11, and also, the synthesis of the enantiomere of the imidazo[l,5-a][l,4]benzodiazepine8 12,25, and 31, all of which are structurally related to midazolam (3) or the mized agonia ta/antagonista 6 and 6a. The biological testing data for these compounds (Table I) indicated that 12 and 31 are agoniata, while 25 ia a znized agoniat' antagonist. To datę, no reporta have appeared deacribing the separation of enantiomere of imidazo [l,5-a][ 1,4]-benzodiazpinea whoae asymmetry ia due only to a chiral piane. The interaction of theae compounds with the benzodiazepine receptor will be discussed.

Reaulta and Diaousaion

Syntheaia and Separation of the Diaatereomeric Benzodiazepineb 10 and 11. Midazolam (3) ezists as the racemate 3a, 3b, and cannot be resolved at room temperaturę due to the rapid interconveraion 3a «■* 3b (Chart I). A high temperaturę NMR study (200 °C) of 3 showed a collapse of the AB paltem for the 4-methylene protons, indicating a barrier to interconvereion of 3a and 3b of less than 24 kcal/mol. We had previoualy shown⁶ that subetitution of a ferf-butyl group for the methyl group in diazepam (2) to give 4 was sufficient to allow for the separation of the enantiomere 4a and 4b. In this case there was no collapse of the 3-methylene protone in the NMR spectrum at 200 *C, indicating a barrier of invereion of >24 kcal/mol.¹⁰ From an ezamination of the structure of midazolam (3), it was evident that the fused benzene and imidazo ringB form a biphenyl type system in which the introduction of a bulky substituent (such as ferf-butyl) at the 1-poeition might provide enough steric hindrance between the 1 and 10 positions during the interconvsrsion process (see structure 3 for numbering) to allow for the separation of the atropisomere." Molscular mechanics calculations have now been performed to investigate the structural and energetic aspecta of the ring inversion for compound 4 as we U as compounds 2, 3, and 31. The calculations were carried out using the standard Tripoe force field in the SYBYL program.¹²⁴ The pathways for ring invereion were determined by computing the potential energy surface as a function of the C2-C3, C9a-Nl, and C5-C6a torsional angles for compounds 2 and 4 and as a function of the C3a-C4, CiOa-N, and C6-C6a torsional angles for 3 and 31 (see structures 2 and 3 for numbering). For 4, the potential energy surface ehows a saddle point

(10)    Shanaa-AUdi. H. Bar-Eli, K. H. J. Phyt. Chom. 1970,74,Ml. We carried out the calculation* by Mfuming that en equal popuiation of the two enanltomen were preeen'..

(11)    EUel. E. L. Stertochtmutry of Carbon Compoundr, McGraw-Hill: New York. 1962. p 177.

(12)    SYBYL. Ver»ion 5.41: Tripoe Aeeociatat. 1699 Henley Road. St. Louie, MO 63144.