Contemporary macrocyclic synthetic receptors include cryptophanes, carcerands, cyclophanes, spherands, hemispherands, crown ethers, and calixarenes. Of these, the water soluble cyclophanes stand alone in their ability to bind to lipophilic substrates in aqueous media through predominantly solvent driven forces. Cyclophanes, like many natural receptors, contain a lipophilic pocket, can be more or less preorganized, and can be made water-soluble at physiological pH. These characteristics validate the use of this class of molecule as valuable model systems for the study of hydrophobic binding.
The use of cyclophanes for studies in molecular recognition began in earnest in the seventies with Whitlock, Tabushi, Murakami, and Koga. Odashima and Koga reported the first direct evidence for the stoichiometric inclusion of a nonpolar guest, 2,7-naphthalenediol, with their cyclophane host by a crystal structure of the complex. Characteristic of these early host molecules, including those introduced by Diederich and Wilcox, was the use of substituted 4,4'-methylenedianiline compounds as achiral host building blocks. Dougherty made an important contribution by introducing the substituted ethenoanthracene structural unit for chiral host manufacture. Reports from Wilcox introduced several analogs of Tröger's base, which can function as a chiral molecular hinge with a resting angle of 92-98o, an angle more acute than that provided by ethenoanthracene (120o) and more readily incorporated into rectangular structural motifs.
From the beginning, the special interest in this area has been in controlling the flexibility of water soluble cyclophanes. It was reasoned that a less flexible molecule might show higher selectivity and lead to less ambiguity in interpretation of binding data. A new cyclophane molecule recently synthesized in the Wilcox lab incorporates all three of the building blocks described above. This host molecule (above) binds to neutral cyclohexanoid guests with high affinity and is diastereoselective and enantioselective.
The diastereoselectivity of this molecule is a direct consequence of its shape selectivity. The distance across the shorter dimension of the cavity is perfect for sandwich-type binding cyclohexane rings, but the distance is too close to allow axial substituents to enter. The result is that diastereomers with axial groups are less effectively bound than diastereomers with only equatorial groups. The observed enantioselectivity, however, was very small. A crystal structure of the host molecule confirmed the results of computer aided molecular modeling that guided the redesign effort. The toroidal shape of the host is formally chiral, but the interior space created by the macrocycle has a rather smooth and featureless character. This is somewhat like a lock with few tumblers into which many keys would fit. In this context, the macrocycle can be said to have "fuzzy" chirality, and changes are required to make its chirality less "fuzzy".
The goal is to make a host molecule (16, below) that shows greater discrimination among enantiomers than any proposed macrocyclic structure to date. It was reasoned that rearranging the structure so that bulkier groups would rise above and below the rim of the host would increase host-guest contacts and enhance the receptor's ability to differentiate enantiomers. Additionally, incorporating more rigid building blocks would enhance the ability of the host molecule to bind to potential guests. This project focuses on the development of a new chiral column using this synthetic receptor tethered to a solid support. The receptor is designed to bind small alicyclic substrates that are typically difficult to separate chemically. The acid functionalities on the periphery of the molecule provide the means for attachment to the solid support.
Because the host molecule is chiral, the several building blocks will have to be synthesized and resolved separately, then combined to form the macrocyclic structure. The first building block to be made will be the ethenoanthracene derivative, the synthesis of which is outlined in Scheme 1. A simple Diels-Alder reaction between anthracene and dimethyl acetylenedicarboxylate will produce ethenoanthracene (1). Aromatic nitration will result in the production an enantiomeric mixture of 2-nitroethenoanthracenes (2a,b). Saponification of the methyl esters to yield the dicarboxylic acid (3a,b) followed by treatment with (-)-menthol and dicyclohexylcarbodiimide (DCC) will produce the diastereomeric menthyl esters (4a,b). Separation of the diastereomers via differential recrystallization from a hexane and ethyl acetate solvent system (75%/25%) will yield the enriched isomer 2a after removal of the menthyl esters and replacement with methyl esters. Aromatic nitration will produce diastereomeric dinitro compounds (5a,b), which, when monoreduced, can be separated to yield the enantiomerically pure 2-amino-6-nitroethenoanthracene biulding block (6a). The methodology of producing pure diastereomer 2a is compatible with fairly large-scale work - 90 grams of racemic diester (2a,b) can be converted to 10 grams of resolved methyl ester (2a) in about 36 hours.
Scheme 1. 
The spribifluorene unit is synthesized according to Scheme 2. A Grignard reaction involving 2-iodobiphenyl and 9-fluorenone followed by refluxing in acetic acid and aqueous hydrochloric acid produces spirobifluorene (7). Friedel-Crafts acylation with acetyl chloride produces a racemic mixture of 2,2’-diacetylspirobifluorene (8a,b). Oxidation with bromine in base will yield the racemic diacids (9a,b), which are then reduced to the diol with Red-Al (10a,b). Resolution of the racemic diols is accomplished with (-)-camphanic chloride, which is subsequently removed with base to form the enantiomerically pure diol (10a). Oxidation to the diacid, and conversion to the acid chloride (12a) completes the synthesis of this building block.
Scheme 2. 
The macrocycle synthesis is accomplished by the combination of the two building blocks (6a and 12a) as described in Scheme 3. A 2:1 molar ratio of 2-amino-6-nitroethenoanthracene (6a) and the acid chloride derived from spirobifluorene (12a) produces the dinitro diamide (13). Reduction of the dinitro compound to the diamine is effected by hydrogenation yielding 14. Final macrocyclization is effected by Tröger’s base formation to give 15. This is a critical step. A diamino compound similar to 14 had been cyclized in 48% yield through the action of hexamethylenetetramine (HMTA) in trifluoroacetic acid. A similar strategy will be used here. The macrocyclic tetramethyl esters are hydrolyzed with LiOH in aqueous methanol and, after acidification, the resulting tetraacid 16 will be formed in quantitative yield completing the synthesis.
Scheme 3. 
The final macrocyclic structure 16 will then be tethered to a suitable solid support, such as Restek’s Ultra Amino Column, by recycling a solution containing 16 through an amine column, thereby covalently attaching the macrocycle to the column packing. The column is aminopropylsilane chemically bonded to silica with a 5mm particle size (1.0 mm ID, 30 mm in length). The effectiveness of the column toward the separation of racemic mixtures of small alicyclic compounds will then be determined. For instance, the separation of racemic mixtures of 1,2-cyclohexandiol, menthol, camphor, and other cyclohexane derivatives will be attempted. In addition, aromatic substrates, such as benzoin and deoxybenzoin will be separated.
| |
| |