AmericanChemicalSociety.com

Synthesis. This grant supports the synthesis of chiral, Lewis-basic macrocycles, and the evaluation of their use as enantioselective host molecules, both in free-base form and as labile metal complexes. Over the past year, we have synthesized the trans-1,2-diaminocyclohexane based macrocycles (R)-1 through (R)-7, and have made significant progress towards the isolation of diastereomeric macrocycles (R,R)-8 and (R,S)-8 (see Figure 1). Each of the isolated compounds has been prepared in large quantities (~ 500 mg – several grams), and has been characterized by ¹H, ¹³C and COSY NMR spectroscopy.

Figure 1.

X-ray crystallography and molecular modeling. Compounds (R)-4 and (R)-7 were characterized by X-ray crystallography (Figures 2 and 3, respectively). Both structures show the same (R,S) configuration of the two stereogenic N atoms to facilitate an intramolecular hydrogen bond. The arene rings of both structures are also directed away from the cyclohexyl groups, fashioning a relatively rigid cleft in which the vicinal diamino moieties reside. The meta linkages of (R)-4 appear to enforce a more extended macrocyclic structure, whereas the 20-membered ring of (R)-7, which contains ortho linked spacers, exhibits a pronounced kink.

Molecular modeling studies (Monte-Carlo searching using MM94 force field followed by single point calculations of the 100 lowest E structures at the HF/3-21G level) confirm that the macrocycles become less planar as the arene linkage geometry is changed from para to meta to ortho. The optimized structure of (R)-4 is virtually superimposable on the crystal structure shown in Figure 2, whereas the crystal structure (R)-7 is very similar to one of the lowest energy computed structures.

Figure 2.

Figure 3.

Enantioselective binding of mandelic acid. Several of the macrocycles prepared were observed by ¹H NMR to bind mandelic acid (MA) enantioselectively in chloroform. The addition of a macrocycle to a racemic solution of MA causes upfield shifting and splitting of the benzylic C-H signal of MA into two equal intensity singlets that is greatest when the molar ratio of macrocycle to MA is near 0.3. Macrocycles (R)-3, (R)-4, and (R)-7 provide the greatest levels of enantiodiscrimination and are significantly more effective than the acyclic receptor (R)-9. Of the three best macrocycles, (R)-4 is the most effective, with maximal peak separations > 0.1 ppm achievable. On the other hand, receptors (R)-1, (R)-2, (R)-5, and (R)-6 were less effective than the acyclic control (R)-9in discriminating MA enantiomers.

            Various MA derivatives bearing halogen and methoxy substituents on the phenyl ring were also tested. In general the results confirmed that (R)-3, (R)-4, and (R)-7 are most effective and are capable of providing baseline resolved separation between enantiomers. No binding by any of the macrocycles to the ester methyl mandalte is observed, which implies that the carboxylic acid group is involved in the primary associative interaction. The utility of the present macrocyclic receptors as chiral NMR shift reagents was demonstrated in the case of (R)-3, which was used to determine the enantiomeric excess of five MA samples with less than 2% error by integration of the resolved signals. In all cases, the magnitude of the enantiodiscrimination was found to increase with total concentration of MA and macrocycle. The poor solubility of MA in chloroform is not a limiting factor however, because in the presence of a macrocycle, the solubility of MA is dramatically enhanced.

            The specific binding processes were investigated by ¹H NMR titrations in which single MA enantiomers were used as guests. Both continuous variation and mole ratio studies support the existence of macrocycle:MA complexes of 1:1 and 2:1 stoichiometry, likely resulting from proton transfer from the carboxylic acid group of MA to the amine group of the receptor, with subsequent ion pairing. Comparison of analyses of MA enantiomers suggests that for (R)-3 and (R)-4, it is the 2:1 and not the 1:1 complexes, which provide enantioselectivity, whereas for (R)-7, both 1:1 and 2:1 complexes exhibit enantioselectivity.

            Future studies will be aimed at quantifying the 1:1 and 2:1 apparent association constants, and characterizing the Zn(II) and Cu(II) complexes of the macrocycles. Preliminary work along these lines suggest that the metal linked macrocycles (R)-3 and (R)-4 from well defined Zn(II) complexes, whereas the   

Collaboration: The X-ray crystal structures were determined in collaboration with Professor Joseph Tanski of Vassar College. Much of the macrocycle synthesis was performed by Philip Atwood (Chemistry, '12), Tyler Moore (Chemistry, '10), and Shaun Ben-Ari (Chemistry, '12). The majority of the NMR binding analyses were performed by Thomas Quinn (Biology, '11). Philip Atwwod and Tyler Moore were directly supported by the present award.