What is the Bioactive Conformation of Aminoalkylindenes at the CB1 Receptor?
Insights Gained from E and Z Napthylidene Indenes

P.H. Reggio, S. Basu-Dutt, D.P. Hurst, M.J. Patel,
H.H. Seltzman and B. F. Thomas

 Introduction
 Methods and Results
 Conclusions

Introduction

We’ve spent the past year studying the aminoalkylindoles. Above, we presented the pharmacophore we’ve developed for the AAIs including hypothesized binding sites at the CB1 receptor. One always begins pharmacophore development by performing ligand conformational analyses. We discovered early that the AAIs have two distinctive families of conformers. These families are characterized by the relative orientation of the aroyl moiety relative to the indole system. In the S-cis conformation, the aromatic ring(s) of the aroyl moiety are stacked (in a tilted-T stack) over the indole ring system. The carbonyl oxygen is then near the C-2 substituent on the indole ring (about 20° - 40° out of the indole plane). In the S-trans conformation the opposite occurs. Here the aromatic ring(s) of the aroyl system are over the C-2 position while the carbonyl oxygen is over the indole system (the oxygen is 20° - 40° out of the indole plane). A change in the C-2 substituent alters the preference for S-cis or S-trans. When C-2 is a hydrogen the S-trans conformation is favored. When C-2 is methyl, however, the S-cis conformation is favored. The energy differences between the two forms can be anywhere between 0.5 and 2.0 kcal/mole depending on the AAI being studied. The energy difference for most AAIs is small enough that both conformers would be expected to exist at ordinary temperatures.

Now if one is trying to model the interaction of AAIs at CB1 this presents a problem: which conformer is the bio-active conformer? Prompted by a discussion in the first Sterling Winthrop AAI paper (M.R. Bell et al J.Med.Chem. 34, 1099-1110, 1991), we designed rigid analogs of the AAIs that locked the molecule into either an S-cis (Z) or S-trans (E) type geometry. In order to rigidify the structure, the carbonyl oxygen had to be removed. The carbonyl oxygen could potentially be a key interaction site itself. So, removal of the carbonyl could result in lowered affinity for both geometric isomers. However, this was not the case. While the syntheses were in progress, another Sterling Winthrop paper was published. (V. Kumar et al Bio. Org. Med. Chem. Lett., 5, 381-386, 1995). In this paper, binding and MVD results for the compounds we were in the process of making were reported. However, the pharmacological data in each case was for a mixture of both geometric (E and Z) isomers. The data showed that both the C-2 Hydrogen and C-2 methyl indenes had very good affinity for CB1 and good efficacy in the MVD assay. These results indicated that the indenes retained affinity although the carbonyl oxygen was no longer present. However, because the geometric isomers had not been separated, the Kumar paper did not shed light on which conformation was the bioactive conformation. We therefore synthesized the E and the Z isomers of the C-2 Hydrogen (protio) and the C-2 methyl compounds. The table shows that in both cases, we found that the E (S-trans) isomer had the better affinity and activity (as measured by the twitch response of the Guinea Pig Ileum assay) at CB1 (and also at CB2) and had the higher activity. The compounds possessed CB2 affinity similar to their CB1 affinities, as well.

Methods and Results

A. Conformational Analysis

We began by performing complete conformational analyses of the E and Z isomers of the protio and methyl indenes using the semi-empirical AM1 method. For the C-2 Hydrogen (protio) geometric isomers, the global minimum energy of the E isomer was lower in energy. For the C-2 Methyl geometric isomers, the global minimum conformer of the Z isomer was lower in energy.

B. REV Screens

Conformers were screened using our Receptor Essential Volume Maps (REV 1 and REV 2) which delineate regions of steric interference at the AAI CB1 binding site. The remaining conformers were considered accessible. As an example, the conformers found for the protio E Indene and the Z Indene are illustrated in the figures. We used our REV 1 and 2 maps to screen the calculated conformers. After using the REV maps, the following conformers of the E and Z isomers remained.

C. Receptor Docking

The global minimum energy conformer of each indene was used for docking studies within the CB1 model. Both the E and Z isomers were found capable of binding in the same general region of CB1, Hxs 3,4 and 5. The figure shows the methyl E indene in CB1. As indicated in the figure, there are numerous interactions available to the methyl E indene. The primary ligand/receptor interaction appears to be aromatic stacking. Aromatic stacks commonly are "tilted-T" type stacks, in which the hydrogens on the edge of one aromatic ring are pointing toward the pi electron density of another aromatic system. Energetically, the tilted T has been estimated to be equivalent to a Hydrogen bonding interaction between uncharged residues (~ 1.5kcal/mol) [S.K. Burley and G.A. Petsko, Science 229,23-28, 1985.]. The methyl E has stacking interactions with the aromatic residues shown in yellow. In this position, W5.43 (in white) provides steric hindrance postulated above. The next figure shows the methyl Z indene at CB1. The Z indene has fewer aromatic stacking interactions (shown in yellow). The C-2 protio indene isomers dock as did their methyl counterparts.

Conclusion

Hxs 3,4 and 5 appear to be involved in the binding of the napthylidene indenes at CB1. The E(trans) geometric isomer has many more aromatic interactions than does the Z (cis) isomer. This may explain the greater CB1 affinity of the E indenes.