2,5-Anhydro sugar diacid and 2,5-anhydro sugar diamine based C₂ symmetric peptidomimetics as potential HIV-1 protease inhibitors

Abstract

Conformationally constrained molecular frameworks of the 2,5-anhydro sugar diacid (9) and 2,5-anhydro sugar diamines (10, 11) were used to construct architecturally beautiful novel C₂ symmetric peptidomimetics 1–8. Although none of these compounds showed any significant HIV-1 protease inhibitory activity, further refinements in design may lead to protease inhibitors based on these rigid carbohydrate-derived scaffolds.

Among the several specific targets in the cell cycle of the human immunodeficiency virus (HIV), the virally encoded homodimeric HIV-1 protease, which is an aspartyl protease required for maturation of the infectious virion, has emerged as a promising target leading to the design and evaluation of a vast array of compounds with diverse structural motifs as possible inhibitors, some of which have already been approved for the treatment of AIDS.^1–6

We describe herein the development of a new class of compounds 1–8 as potential HIV-1 protease inhibitors that are based on carbohydrate–peptide hybrid structures. In this approach, identical peptide chains are anchored on both sides of a core carbohydrate motif—a C₂ symmetric 2,5-anhydro sugar diacid or sugar diamine—leading to the formation of C₂ symmetric peptidomimetics.⁷ Carbohydrate-based molecular designs are increasingly drawing chemists’ attention.^8,9 Detailed studies on the development of new HIV-1 protease inhibitors based on acyclic carbohydrates have recently been reported.^10–12 It is also being increasingly felt that small molecule protease inhibitors need to have restricted degrees of freedom. The success of cyclic urea based inhibitors supports this feeling.^13,14 This has prompted us to look for cyclic carbohydrate based core foundations as conformationally rigid scaffolds to build a new class of molecular frameworks as potential protease inhibitors.

Two different types of cyclic carbohydrate framework—sugar diacid and sugar diamine—are employed in our designs. These molecules are 2,5-anhydro-D-idaric acid (9), 1,6-diamino-2,5-anhydro-1,6-dideoxy-D-iditol (10) and 1,6-diamino-2,5-anhydro-1,6-dideoxy-D-mannitol (11). The sugar diacid 9, to the best of our knowledge, was reported for the first time by us along with the synthesis of the first sugar diacid based peptidomimetic 1.¹⁵ The twofold symmetry of the molecule was evident in its ¹H and ¹³C NMR spectra which showed peaks for only half of the molecule. Two-dimensional NMR studies in combination with constrained molecular dynamics (MD) simulations revealed a very ordered C₂ symmetric structure for 1, consisting of identical intramolecular H bonds at the two ends between the LeuNH→sugar-OH. A very small temperature coefficient for the LeuNH chemical shift (Δδ/ΔT=−1.2 ppb K⁻¹) was observed in DMSO-d₆. The presence of two ‘cis-β-hydroxycarboxyl’ moieties, the core structural motif believed to be responsible for such intramolecular H bonds,¹⁶ on two sides of the tetrahydrofuran ring nucleated identical β-turn-like structures in 1 at both ends.

To further these studies, we decided to prepare a few more C₂ symmetric peptidomimetics and examine their structures and properties, especially HIV-1 protease inhibition activities, as these compounds structurally resemble many well-established C₂ symmetric HIV-1 protease inhibitors. In this paper we describe the results of that investigation.

Compounds 2 and 4 were prepared in 60 and 70% yields, respectively, from 1 in two steps: saponification followed by the reaction with p-bromoaniline (for 2) or H₂N–Ala–OMe (for 4¹⁷) under standard solution phase peptide synthesis conditions using 1-hydroxybenzotriazole (HOBt) and 1-ethyl-3-(3-(dimethylamino)propyl)carbodiimide hydrochloride (EDCI) as coupling agents and amine-free dry DMF and/or dry CH₂Cl₂ as solvents. The synthesis of 3 was carried out following the same procedure described earlier for the synthesis of 1.¹⁵ The di-O-benzyl derivative of 2,5-anhydro-D-idaric acid¹⁵ (9) was treated with an excess of amino-terminal free peptide molecule to furnish 3 in 55% yield.

Structural studies of these molecules (2–4) did not reveal significant turn structures in solution as seen in 1. While compounds 3 and 4 still maintained C₂ symmetric structures as shown by their NMR studies in DMSO-d₆, the p-bromoanilide derivative 2 lost its symmetry with each of its amide protons showing two peaks in the same solvent. Its LeuNH chemical shifts, however, exhibited low temperature coefficients of −2.1 and −2.4 ppb K⁻¹ maintaining their participation in intramolecular H bonding, similar to the one observed in 1, albeit weaker. Hydrogenation of 3 gave a complex mixture of products and its structure and properties were studied in the Bn-protected form, which showed high temperature coefficients for the amide proton chemical shifts. The temperature coefficient of the LeuNH peak in 4 was −2.8 ppb K⁻¹, indicating gradual loss in its ability to nucleate the desired secondary structure, probably due to the increased steric hindrance.

Next, we turned our attention to the remaining peptidomimetics, 5–8, based on sugar diamines 10 and 11. Syntheses of 5 and 6 were accomplished according to Scheme 1. The starting material 12, prepared from 1,3:4,6-di-O-benzylidene-D-mannitol,¹⁵ was converted to 14 in four steps—ditosylation, azidination and selective azide reduction, followed by protection in situ using Boc₂O for purification purposes giving the di-N-Boc-protected intermediate 14 in 50% overall yield from 12. Purified 14 was treated with TFA in CH₂Cl₂ and peptide chains were attached at both ends using EDCI–HOBt as coupling reagents. Finally, debenzylation using H₂–Pd–C (10%) in MeOH furnished the final product 5 (and 6) in ~70% overall yield.¹⁸

Scheme 1.

Synthesis of 5 and 6

Syntheses of 7 and 8 from 3,4-di-O-benzyl-1,2:5,6-diisopropylidene-D-mannitol (15) are depicted in Scheme 2. Monodeprotection gave a diol intermediate whose primary hydroxyl was selectively acylated to get a 55% yield of acetate 16. Conversion of 16 to the terminal epoxide 17 was achieved in two steps in 86% yield. Acid-catalyzed cycloetherification of 17 went smoothly to furnish the 2,5-anhydro-D-mannitol framework 18, which was converted to the diamine and finally to the desired products 7 and 8,¹⁹ following the procedure described above in Scheme 1.

Scheme 2.

Synthesis of 7 and 8

All the sugar diamine based mimetics, 5–8, had C₂ symmetric structures in solution, as determined by their ¹H NMR studies in DMSO-d₆, which showed peaks for only half of the molecule. However, very high temperature coefficients for all of their amide proton chemical shifts indicated the absence of any intramolecular H bonding in these molecules.

The potential of the peptidomimetics 1–8 as HIV protease inhibitors were evaluated using an assay protocol developed by Toth and Marshall.²⁰ Unfortunately, none of these compounds showed any enzyme inhibitory activities up to 2 μM inhibitor concentration. The large distances between the phenyl rings of the P1/P1′ positions are probably responsible for their low activities. Efforts are now underway to improve our designs and to prepare active HIV-1 protease inhibitors from conformationally constrained 2,5-anhydro sugar-based scaffolds.