Design and synthesis of phosphorus-based inhibitors for the HIV-1 proteinase

Abstract

A number of peptide fragments including (2S)-Pro-(2S)-Ile-NHiBu (72) and (2S)-Phe-(2S)-Ile-NHiBu (73) and l-(benzyloxycarbonyl)-aminophosphinic acid methyl esters (analogues of Phe (76), Cha (77) and Leu (78)) have been synthesised and coupled to give a series of phosphonamidate methyl ester-based peptide inhibitors of HIV-1 proteinase. These compounds were tested against the proteinase enzyme in vitro using a spectrophotometric assay and displayed activities in the 1-100 muM range. Remarkably, comparison of these data with data obtained for activities against HIV-1 in cultured human lymphocytes showed an in vitro:in vivo ratio of approximately 1:1. These results are indicative of a highly efficient cell uptake mechanism and exceed the reported in vitro:in vivo ratios for HIV proteinase inhibitors by a factor of 103. It is also apparent from the results that the compounds are not rapidly degraded in human lymphocytes. The activities of these compounds, however, was rather insensitive to small structural changes, undermining our attempts to optimise them. A phosphonamidate ethyl ester (Cbz-Phe-PO(OCH2CH3)NH-(2S)-Phe-(2S)-Ile- NHiBu (94)) was prepared and this possessed an IC50 of 4.5 muM, indicating that it might be possible to incorporate larger ligands onto the ester portion of the molecule. Due to the problems associated with peptidic compounds as therapeutic agents, a number of non-peptidic targets were designed. 3,3'-Di- (benzyloxycarbonyl)-aminobenzoin (103) was synthesised in three steps from m- nitrobenzaldehyde and this displayed an IC50 of 8 muM. Molecular modelling of a second target, a bicyclic phosphorodiamidate (106), showed that this compound should adopt a skewed position within the enzyme's active site, with the OH group between the catalytic aspartates (Asp 25 and Asp 25') and the P=0 strongly hydrogen-bonded to one of the flap Ile's but relatively distant from the other. A synthesis of this molecule from tris (hydroxymethyl)amino-methane was undertaken, but problems with this led to the design of the less-substituted phosphorodiamidate (116). An alternative route from diethyl malonate was begun, although to date this has not been completed. This work is now under investigation by others in our group.