This application is a continuation-in-part of U.S. Serial No. 08/357,663, filed December 16, 1994, a continuation- in-part of 08/188,146, filed January 27, 1994, U.S. Serial No. 07/901,401, the contents of which are hereby incorporated by reference into this application.

This invention was made with government support under grant #GM-44525 from the National Institutes of Health and grants #CHE89-11008 and #CHE92-08254 from the National Science Foundation. Accordingly, the U.S. Government has certain rights in the invention.

Background of the Invention

This invention relates to the field of molecular recognition of small ligands. More particularly, the invention relates to compositions useful for the purification of enantiomers of amino acid derivatives and for the purification of certain compounds able to form hydrogen bonds, methods for preparing these compositions, and methods for using them.

Standard approaches to the optical resolution and purification of organic and biological molecules include crystallization, distillation, extraction, and chro - atography (Eliel, Stereochemistry of Carbon Compounds, New York: McGraw-Hill, 1962). Each methodology is based on a physical or chemical interaction of a molecule with an element of its environment, and may involve molecular sizing, electrostatics, hydrophobicity, sterics, or polarity. The efficiency of purification increases as

the differences in interaction energy for all the species present in the mixture increase. The relevant interactions for cystallization are crystal lattice forces and solvation of the molecule; for distillation, the interaction is a liquid-gas phase transition; while for extraction and chromatography, the interaction is exchange between non-miscible phases. Common to all these classic methods is the limitation that as molecular structures become increasingly similar, the energy differentials for the relevant interaction diminish to the extent that high resolution is no longer feasible. A general approach to purification necessitates an enhanced capability for transcending this natural tendency toward shrinking energy differences. The ability to purify very similar or chiral molecules is of economic and practical importance to the developing fields of biotechnology, and should greatly accelerate the development of new pharmaceuticals and bioactive and other useful compounds.

The ability to distinguish similar molecules is an important goal of research in the field of molecular recognition. Early efforts to bind molecules selectively involved naturally occurring host molecules, such as clathrates, cholic acid, and cyclodextrinε (Diederich, Angew. Chem. Int. Ed. Enq.. 27, 362 (1988); Breslow, Science (Washington. D.C.) , 218, 532 (1982)) . The first example of a synthetic system specifically designed to undergo inclusion complexation was a cyclophane (Stetter & Roos, Chem. Ber.. 88, 1390 (1955)). Synthetic crown ethers and cyclic polyamines were designed to complex metal ions selectively by adjusting ring size and number of heteroato s (Pederson, Angew. Chem. Int. Ed. Eng. , 16, 16 (1972)). Macrobicyclic compounds have been prepared which show selectivity for trihydrobenzenes with certain substitution patterns (Eb eyer and Vogtle, Angew. Che . Int. Ed. Engl.. 28, 79 (1989)).

The use of chiral components in constructing host compounds has led to the development of molecules which are, in principle, capable of diasteroselective complexation with chiral guests. While several systems have exhibited some diastereoselectivity, numerous attempts to produce chiral hosts have not produced any known compounds of practical utility prior to the present invention. The earliest preparations of chiral crown amino ethers were applied to cation complexation, and not to chiral discrimination by diastereoselective complexation (Wudl & Gaeta, J. Chem. Soc. , Chem. Commun. , 107 (1972)). Chiral hosts based on biphenyl-macrocycles have shown promise (Kyba, et al., J. Amer. Chem. Soc.. 100, 4555 (1978)). A recent example intended to dis- tinguish enantiomers of amino acids and arylpropionic acids however appears from binding studies not to function as a host for nonpolar molecules (Rubin, et al., J. Org. Chem.. 51, 3270 (1986)).

High enantioselectivity has thus largely eluded prior workers in the field. A bilaterally symmetric host con¬ taining two diiodotyrosine moieties was one of the first to exhibit a measurable difference in binding energy with mirror image guest molecules (Sanderson, et al., J. Amer. Chem. Soc.. Ill, 8314 (1989)); free energy differences ranged from -0.15 to 0.48 kcal/mole, with binding site saturations up to 67%. More recently, a related chiral host was made with pyridyl moieties replacing benzene rings in the macrocycle, which showed free energy differences up to about 1 kcal/mole and a range of binding saturations of approximately 40-80% (Liu, et al., J. Org. Chem.. 55, 5184 (1990)). Chiral hosts in which the enantioselection energies exceed 1 kcal/mole have been virtually nonexistant prior to the present invention.

Progress toward a completely chemoselective or enant- ioselective host has been limited, proceeding roughly in parallel with growing understanding of intermolecular interactions controlling binding affinity in natural rec- eptors like enzymes and hormone receptors. The subject invention provides a composition of matter which possesses enzyme-like enantioselectivity which is sufficiently high to offer practical utility in optical resolution and chemical purification of organic compounds.

Summary of the Invention

The subject invention relates to a composition of matter having the structure:

wherein each of A, B, C, X, Y, and Z is independently 0, NH, N(CH ₂ ) _m CH ₃ , N(C=0) (CH ₂ ) _m CH ₃ , CH ₂ , S, or Se; each of R,, R ₂ , and R ₃ is independently phenyl, 4-hydroxyphenyl, pyridyl, pyrrolyl, indolyl, naphthyl, thiophenyl, (C=0) (CH ₂ ) _p CH ₃ , NH(C=0) (CH ₂ ) _p CH ₃ , OH, COOH, NH.,, or SH; and m, n, and p are integers between 0 and 5.

The invention provides a process of obtaining a purified enantiomeric isomer of a compound of interest from a mixture of isomers of such compounds which comprises contacting the mixture of isomers with the composition under conditions such that the enantiomeric isomer binds

to the composition to form a complex, separating the resulting complex from the mixture, treating the complex so as to separate the enantiomeric isomer from the composition, and recovering the purified enantiomeric isomer.

The invention also provides a process of obtaining a purified organic compound of interest able to form hydrogen bonds from a mixture of organic compounds which comprises contacting the mixture with the composition under conditions such that the organic compound binds to the composition to form a complex, separating the re¬ sulting complex from the mixture, treating the complex so as to separate compound from the composition, and recovering the purified compound.

The invention further provides a process of preparing the composition which comprises: (a) reacting a chiral multifunctional reagent containing at least one protecting group with a compound having the structure:

wherein each of A, B, C, X, Y, and Z is independently O, NH, N(CH ₂ ) _m CH ₃ , N(C=0) (CH ₂ ) _m CH ₃ , CH ₂ , S, or Se, under conditions permitting formation of a compound having the structure:

wherein each of A, B, C, X, Y, and Z is independently O, NH, N(CH ₂ ) _m CH ₃ , N(C=0) (CH ₂ ) _m CH ₃ , CH ₂ , S, or Se; each of R ₁ , R ₂ , and R ₃ is independently phenyl, 4-hydroxyphenyl, pyridyl, pyrrolyl, indolyl, naphthyl, thiophenyl, (C=0) (CH ₂ )pCH ₃ , NH(C=0) (CH ₂ ) _p CH ₃ , OH, COOH, NH ₂ , or SH; and m, n, and p are integers between 0 and 5;

(b) treating the compound formed in step (a) under suitable conditions so as to cleave one protecting group and form a compound having the structure:

wherein each of A, B, C, X, Y, and Z is independently O, NH, N(CH ₂ ) _m CH ₃ , N(C=0) (CH^CI^, CH ₂ , S, or Se; each of R,, R ₂ , and R ₃ is independently phenyl, 4-hydroxyphenyl, pyridyl, pyrrolyl, indolyl, naphthyl, thiophenyl, (C=0) (CH ₂ ) _p CH ₃ , NH(C=0) (CH ₂ ) _p CH ₃ , OH, COOH, NH ₂ , or SH; and m, n, and p are integers between 0 and 5;

(c) treating the compound formed in step (b) with a condensing agent under conditions permitting multiple macrolactamization so as to thereby form the composition.

The subject invention further provides a composition of matter having the structure:

wherein A has the structure:

and R ₁ and R ₂ are independently the same or different and are H, F, a linear or branched chain alkyl, arylalkyl, alkoxyalkyl, aminoalkyl, alkylaminoalkyl, hydroxyalkyl, (cycloalkyl) alkyl, or acylalkyl group, or an aryl group, a linear or branched chain alkylaryl, pyridyl, thiophene, pyrrolyl, indolyl or naphthyl group; X is CH ₂ or NH; Y is C=0 or SO,; and n is 0 to about 3.

The subject invention also provides a composition of matter having the structure:

wherein A has the structure :

and R,, is H, a linear or branched chain alkyl, arylalkyl, alkoxyalkyl, aminoalkyl, alkylaminoalkyl, hydroxyalkyl, (cycloalkyl)alkyl, or acylalkyl group, or an aryl group, a linear or branched chain alkylaryl, pyridyl, thiophene, pyrrolyl, indolyl or naphthyl group.

The subject invention also provides a composition of matter having the structure:

wherein R ₁ , R ₂ and R ₃ are C ₆ H ₄ (OCH ₂ CH=CH ₂ ) ; A, B and C are CH ₂ ; X, Y and Z are S; and n is 1.

The subject invention provides a composition of matter having the structure:

wherein each of A, B, C, X, Y and Z is independently O, NH, N(CH ₂ ) _m H, N(C=0) (CH ₂ ) _m H, CH ₂ , S or Se; each of R, , R- and R ₃ is independently linear or branched alkyl, aryl, (C=0) (CH ₂ ) _p H, NH(C=0) (CH ₂ ) _p H, OH, COOH, NH., or SH; and m, n, and p are integers from 0 to about 6.

The invention provides a composition of matter having the structure :

wherein A has the structure:

and R ₁ and R ₂ are each independently H, F, a linear or branched chain alkyl, arylalkyl, alkoxyalkyl, aminoalkyl, alkylaminoalkyl, hydroxyalkyl, (cycloalkyl)alkyl, or acylalkyl group, or an aryl group, a linear or branched chain alkylaryl, pyridyl, thiophene, pyrrolyl, indolyl or naphthyl group; wherein X is NH, O, S, S=0, S(=0) ₂ or NR where R is linear or branched alkyl, acyl, aryl, or 0; wherein Q is selected from the group consisting of H,

linear or branched chain alkyl, acyl and aryl; and m and n are integers from 0 to about 3.

The subject invention provides a composition of matter having the structure:

wherein each of A, B and C is independently 0, NH, N(CH _? ) _m H, N(C=0) (CH,) _m H, CH ^~ Z ' or Se; wherein Q is selected from a group consisting of H, linear or branched chain alkyl, acyl and aryl; and m and n are integers from 0 to about 5.

The subject invention also provides a composition of matter having the structure:

wherein R ₁ and R ₂ are each independently H, F, a linear or branched chain alkyl, arylalkyl, alkoxyalkyl, aminoalkyl, alkyla inoalkyl, hydroxyalkyl, (cycloalkyl) alkyl, acyl- alkyl, aryl, a linear or branched chain alkylaryl, pyridyl, thiophene, pyrrolyl, indolyl or naphthyl group; wherein Q is selected from a group consisting of H, linear or branched chain alkyl, acyl or aryl; and n is an integer from 0 to about 3.

Brief Description of the Figures

Figure 1 shows a scheme for the practical synthesis host molecule 2: (a) Methanol/ammonia 4:1, rt> 2 days, 97%; (b) Boc ₂ 0, i-Pr ₂ NEt, 4-DMAP (10 ol %) , CH ₂ C1 ₂ , 1 h, 90%; (c) NaN(TMS) ₂ , THF, -78°C, 3 in; add tetra-n- butylammoniu iodide and methyl 3,5-bis(bromomethyl)benz¬ oate; warm to 10°C, 2 h, 82%; (d) benzene-1,3, 5-trithiol, i-Pr ₂ NEt, THF, 8 h, 78%; (e) TFA, anisole, CH ₂ C1 ₂ , rt, 16 h, quant; (f) Boc ₂ 0, i-Pr ₂ NEt, K ₂ C0 ₃ , CH ₂ C1 ₂ , rt, 24 h, 86%; (g) THF, EtOH, H ₂ 0, LiOH, 6 h, quant; (h) F ₅ -phenol, EDC, THF, rt, 4 h, 68%; (i) TFA, anisole, CH ₂ C1 ₂ , rt, 4 h, quant; (j) TFA salt in DMA dropwise to i-Pr ₂ NEt, THF, rt, 40 h, 78%.

Figure 2 shows a diagram of a model for receptor- substrate binding.

Figure 3 shows composition 13a*.

Figure 4 shows a synthesis of 13a*.

Figure 5 shows compositions 11a* and 12*.

Figure 6(a) shows a synthesis of composition 11a*.

Figure 6(b) shows a synthesis of composition 12*.

Figure 7(a) shows compositions laa, lbb, lcc and 2aa.

Figure 7(b) shows a synthesis of composition 2aa.

Figure 8(a) shows compositions A ₄ B ₆ , precursors thereof (A, B) , and precursors of A ₄ B1 ₆ and A ₄ B2 ₆ (Bl and B2, respectively) .

Figure 8(b) shows a synthesis of labeled composition 2'.

Figure 9(a) shows compositions 1 ' ' and 2 ' ' .

Figure 9(b) shows a comparison of models for interaction between compositions 1 and 1 ' ' with a peptide substrate.

Figure 9(c) shows key intermediates 3'' and A ' ' in the synthesis of composition 1".

Figure 10 shows gray-scale histograms of substrate sequence binding for 2".

Detailed Description of the Invention

The subject invention provides a composition of matter having the structure:

wherein each of A, B, C, X, Y, and Z is independently O, NH, N(CH ₂ ) _m CH ₃ , N(C=0) (CH ₂ ) _m CH ₃ , CH.,, S, or Se; each of R,,

R. 2' and R ₃ is independently phenyl, 4-hydroxyphenyl, pyridyl, pyrrolyl, indolyl, naphthyl, thiophenyl, (C=0) (CH ₂ ) _p CH ₃ , NH(C=0) (CH ₂ ) _p CH ₃ , OH, COOH, NH.,, or SH; and m, n, and p are integers between 0 and 5. In one embodiment of the invention, X, Y, and Z are each O; in another embodiment, they are each S. In certain other embodiments, R.,, R ₂ , and R ₃ are each phenyl, or they are each 4-hydroxyphenyl. Additionally, in certain embod¬ iments, n is desirably 1.

The invention further provides a process of obtaining a purified enantiomeric isomer of a compound of interest

from a mixture of isomers of such compounds which comprises contacting the mixture of isomers with the chiral host composition defined hereinabove under conditions such that the enantiomeric isomer binds to the composition to form a complex, separating the resulting complex from the mixture, treating the complex so as to separate the enantiomeric isomer from the composition, and recovering the purified enantiomeric isomer. In one embodiment, the process is used to purify enantiomers of amino acid derivatives, of which diamides are particularly effective.

The invention also provides a process of obtaining a purified organic compound of interest from a mixture of organic compounds able to form hydrogen bonds, which com¬ prises contacting the mixture with the chiral host com¬ position defined hereinabove under conditions such that the organic compound binds to the composition to form a complex, separating the resulting complex from the mixture, treating the complex so as to separate the compound from the composition, and recovering the purified compound. In one embodiment, the process is used to purify derivatives of amino acids differing in side-chains. The process is particularly well suited to purify diamide derivatives of amino acids.

One application of the composition is to bind it to a solid support such that a chromatographic adsorbent results which is specific for enantiomeric isomers of compounds of interest and other organic compounds of interest which differ only in side-chain substitution. Effective use of the composition bound to a solid support is made to obtain the enantiomeric isomers of an amino acid derivative in a purified form and to obtain a purified organic compound of interest able to form hydrogen bonds from a mixture of compounds. The compound

to be purified by the composition is preferably a diamide.

The invention further provides a process of preparing the composition, which comprises:

(a) reacting a chiral multifunctional reagent containing at least one protecting group with a compound having the structure:

wherein each of A, B, C, X, Y, and Z is independently 0, NH, N(CH ₂ ) _m CH ₃ , N(C=0) (CH ₂ ) _m CH ₃ , CH ₂ , S, or Se, under

conditions permitting formation of a compound having the structure:

wherein each of A, B, C, X, Y, and Z is independently O, NH, N(CH ₂ ) _m CH ₃ , N(C=0) (CH ₂ ) _m CH ₃ , CH.,, S, or Se; each of R,, R ₂ , and R ₃ is independently phenyl, 4-hydroxyphenyl, pyridyl, pyrrolyl, indolyl, naphthyl, thiophenyl, (C=0) (CH ₂ )pCH ₃ , NH(C=0) (CH ₂ ) _p CH ₃ , OH, COOH, NH-,, or SH; and m, n, and p are integers between 0 and 5;

(b) treating the compound formed in step (a) under suitable conditions to cleave one protecting group to form a compound having the structure:

wherein each of A, B, C, X, Y, and Z is independently O, NH, N(CH ₂ ) _m CH ₃ , N(C=0) (CH ₂ ) _m CH ₃ , CH ₂ , S, or Se; each of R , R ₂ , and R ₃ is independently phenyl, 4-hydroxyphenyl, pyridyl, pyrrolyl, indolyl, naphthyl, thiophenyl, (C=0) (CH ₂ ) _p CH ₃ , NH(C=0) (CH ₂ ) _p CH ₃ , OH, COOH, NH ₂ , or SH; and m, n, and p are integers between 0 and 5; (c) treating the compound formed in step (b) with a con- densing agent under conditions permitting multiple macro- lactamization, thereby forming the desired composition.

The preparation of the composition strategically exploits its C ₃ symmetry. The synthesis of the composition could proceed in a manner analogous to the detailed exper¬ imental examples given hereinbelow for embodiments in which X, Y, and Z are S, and R _1# R ₂ , R ₃ are 4-hydroxy-

phenyl, except that if there is only one protecting group in the chiral multifunctional reagent of step (a) , then none of the side-group protection reactions would pertain.

The coupling of step (a) above can be carried out by several alternative methods of forming amide bonds. One approach is to contact the achiral tetraaromatic triamino triester above shown with the p-nitrophenyl active ester of the chiral multifunctional reagent, made from p- nitrophenol, N-hydroxybenzotriazole, and N,N- dicyclohexylcarbodiimide. The reaction may be performed in the presence of aprotic dipolar solvents, such as N,N- di ethylformamide, tetrahydrofuran, or dimethylsulfoxide, diluted with a miscible cosolvent, such as dichloromethane, to the extent required to achieve solubility of all reactants, at temperatures from about 0 to 100°C, preferably from 0 to 30°C. The preparation of the starting material for step (a) can be obtained by trialkylation of 1,3,5-trimercaptobenzene or phloroglucinol with N-protected methyl 3-(aminomethyl)-5- (bromomethyl)benzoate, followed by cleavage of the N- protecting group. In one embodiment of the invention, the chiral multifunctional reagent containing at least one protecting group in step (a) is an amino acid containing an N-protecting group. In certain embodiments, the amino acid is L-phenylalanine or L- tyrosine. The N-protecting group is preferably chosen such that it may be removed in process step (b) by an acid, for example, trifluoroacetic acid.

In general, process step (b) involves the removal of three protecting groups on the tetraaromatic intermediate. This reaction could be effected by any method corresponding to the lability of the protecting group. A large variety of protecting groups are available for the purpose, including t-butyloxycarbonyl

(BOC) , benzyloxycarbonyl, 2-bromobenzyloxycarbonyl, and p-toluensulfonyl. While a preferred method is to use acid-sensitive BOC groups, other effective protecting groups also removable by acid include biphenylisopropyl- oxycarbonyl (Bpoc) and adamantyloxycarbonyl (Adoc) . Still other protecting groups may be selected such that alternative methods of removal are feasible according to the invention, including photolytic, reductive, electro¬ chemical, and mild base conditions. This flexibility allows a wide range of chiral multifunctional reagents to be used to prepare the composition.

Prior to condensation process (c) , the protecting ester group (for example, methyl) on each of the three aromatic moieties could be cleaved to give the carboxylic acid by (i) transesterification with trimethylsilylethanol, followed by (ii) fluoride-induced silane elimination. The condensing agent in step (c) could comprise a reagent generated (i) from an agent selected from a group comprising pentafluorophenol, hydroxybenzotriazole, 4- nitrophenol, 2-nitrophenol, pentachlorophenol, hydroxy- succinimide, and hydroxypiperidine and (ii) from an agent selected from a group consisting of N,N-dicyclohexyl- diimide, d i i sopropy 1 carbod i im i de , and carbonyldiimidazole. Other condensing methods may also serve the purpose, including Woodward's reagent K, mixed anhydrides, triphenylphosphine/2,2'-dipyridyl sulfide, ketenimines, and acyloxyphosphonium salts. In a pre¬ ferred embodiment, the condensing agent is the combin- ation of N,N-dicyclohexylcarbodiimide and penta- fluorophenol. If the multifunctional chiral reagent of step (a) contains an alcohol function, the process of steps (b) and (c) could be simply adapted to generate three ester linkages after multiple macrolactonization. Other modifications in the multifunctional chiral reagent of step (a) could be readily envisioned to form such

alternative linkages as thioesterε, thionoesters, and phosphoramides.

The protecting groups which may be present on side-group functionalities could be cleaved by a method corres¬ ponding to their lability. In one embodiment, R , R ₂ , and R ₃ are 4-hydroxyphenyl which should be made by coupling with the suitably protected multifunctional chiral reagent Boc-L-tyrosine (Tyr) . The protecting group on the Tyr is preferably an allyl ether. The processes described provided the embodiments of the composition, wherein R _{l r} R ₂ , and R ₃ are phenyl, in 30% overall yield for the trithia receptor and 7% yield for the trioxa receptor, respectively referred to hereinafter as 1 and 2. Preparation of the tyrosine trithia macrocycle is described in Examples 1 to 16, which serve as an ennabling model illustrative for all embodiments of the composition.

The subject invention also provides a composition of matter hereinafter denoted 9 having the structure:

wherein A has the structure:

and R- and R ₂ are independently the same or different and are H, F, a linear or branched chain alkyl, arylalkyl, alkoxyalkyl, aminoalkyl, alkylaminoalkyl, hydroxyalkyl, (cycloalkyl)alkyl, or acylalkyl group, or an aryl group, a linear or branched chain alkylaryl, pyridyl, thiophene, pyrrolyl, indolyl or naphthyl group; X is CH ₂ or NH; Y is

C=0 or S0 ₂ ; and n is 0 to about 3. In one embodiment, the subject invention provides a composition wherein X is NH. In another embodiment, the invention provides a composition wherein X is CH ₂ , Y is C=0 and n is 1. In another embodiment, the invention provides a composition wherein R ₁ and R ₂ are H.

The subject invention also provides a composition of matter (hereinafter referred to as 10) having the structure:

wherein A has the structure:

and R ₁ is H, a linear or branched chain alkyl, arylalkyl, alkoxyalkyl, a inoalkyl, alkylaminoalkyl, hydroxyalkyl, (cycloalkyl)alkyl, or acylalkyl group, or an aryl group, a linear or branched chain alkylaryl, pyridyl, thiophene, pyrrolyl, indolyl or naphthyl group. In one embodiment, the invention provides a composition wherein R ₁ is a phenyl group. In another embodiment, the invention provides a composition wherein R ₁ is a benzyloxymethyl group.

The subject invention further provides a composition of matter (hereinafter referred to as 2A) having the structure:

wherein R _u R ₂ and R ₃ are C ₆ H ₄ (OCH ₂ CH=CH ₂ ) ; A , B and C are CH ₂ ; X , Y and Z are S ; and n is 1 .

The subject invention also provides a compound which comprises the compositions of matter 9, 10, or 2A, bound to a solid support.

The subject invention further provides a complex which comprises the compositions 9, 10, or 2A, bound to a derivative of an amino acid. In one embodiment, the invention provides a composition wherein the derivative is an amide.

The subject invention provides a process of obtaining a purified enantiomeric isomer of a compound of interest from a mixture of isomers of such compounds which comprises contacting the mixture of isomers with the compositions 9, 10, or 2A, under conditions such that the enantiomeric isomer binds to the compositon to form a complex, separating the resulting complex from the mixture, treating the complex so as to separate the enantiomeric isomer from the composition, and recovering the purified enantiomeric isomer.

The subject invention further provides a process of obtaining a purified organic compound of interest from a mixture of organic compounds able to form hydrogen bonds, which comprises contacting the mixture with the compositions 9, 10, or 2A, under conditions such that the organic compound binds to the composition to form a complex, separating the resulting complex from the mixture, treating the complex so as to separate the organic compound from the composition, and recovering the purified compound. In one embodiment, the invention provides a process wherein the purified organic compound is an amino acid derivative.

The subject invention also provides a process of preparing the composition having the structure:

wherein A has the structure:

wherein R ₁ and R ₂ are H and n is 1 which comprises:

(a) condensing a compound having the structure:

with a compound having the structure:

NHBoc

under suitable conditio ^Υ ns to produce a compound having the structure:

(b) hydrolyzing the compound formed by step (a) under suitable conditions to form an acid compound having the structure:

(c) treating the compound formed in step (b) under suitable conditions so as to activate the acid compound to form a compound having the structure:

(d) reacting the compound formed in step (c) under suitable conditions with a compound having the structure:

NH.

to form a compound having the structure:

C0 ₂ Me

(e) saponifying the compound formed by step (d) under suitable conditions to form a diacid having the structure:

C0 ₂ H

(f) activating the diacid formed in step (e) under suitable conditions to form a compound having the structure:

C0 ₂ C ₆ F ₅

(g) deprotecting the compound formed in step (f) under suitable conditions to form a diamino diacid having the structure:

C0 ₂ H

(h) di erizing the diamino diacid formed in step (g) under suitable conditions to form the composition having the structure:

wherein A has the structure:

and R and R ₂ are H and n is l.

In condensing step (a) it is to be understood that esters other than methyl esters may be used in an equivalent manner for the purposes of the process. Other useful esters include ethyl, propyl, phenol, and benzyl esters.

The condensing agent in step (a) could comprise a reagent generated (i) from an agent selected from a group comprising pentafluorophenol, hydroxybenzotriazole, 4- nitrophenol, 2-nitrophenol, pentachlorophenol, hydroxy- succinimide, and hydroxypiperidine and (ii) from an agent selected from a group consisting of N,N-dicyclohexyl- diimide, diisopropylcarbodiimide, and carbonyldiimid- azole. Other condensing methods may also serve the purpose, including Woodward's reagent K, mixed anhyd- rides, triphenylphosphine/2,2'-dipyridyl sulfide, keten- imines, and acyloxyphosphonium salts. Hydrolyzing step (b) may be performed using base or acid catalysis, though preferably base catalysis. Favorable results obtain using sodium hydroxide. Treating step (c) is effected by a wide variety of procedures, including reaction of pentafluorophenol with DCC or l-(3-(dimethylamino)prop- yl)-3-ethylcarbodiimide hydrochloride (EDC) . Reacting step (d) is performed in the presence of a nonnucleophil- ic base such as triethylamine. Good solvents for the purpose include dimethyl acetamide or dimethyl formamide. Saponifying step (e) is carried out using a base, such as sodium hydroxide. Other bases which effect the step include lithium hydroxide, potassium hydroxide, and tetramethylam onium hydroxide. Activating step (f) is effectively performed using an activating agent such as

pentafluorophenol. Other agents include hydroxy- benzotriazole, 4-nitrophenol, 2-nitrophenol, pentachloro- phenol, hydroxysuccinimide, and hydroxypiperidine. Deprotecting step (g) is carried out preferrably under mildly acidic conditions. Useful acids include trifluoroacetic, trichloroacetic acid and hydrochloric acid in dioxane solution. Scavengers such as anisole help prevent untoward alkylation reactions. Dimerizing step (h) may be effectively performed in the presence of a mild nonnucleophilic base, such as diisopropyl- ethylamine or triethylamine in a dipolar nonaqueous solvent, such as tetrahydrofuran.

The subject invention also provides a process of preparing the composition having the structure:

wherein A is a 1,3,5-trisubstituted phenyl moiety and R ₁ and R ₂ are H and n is 1 which comprises: reacting a compound having the structure:

C0C1

with a compound having the structure:

NH ₂

under suitable conditions to form a compound:

wherein A has the structure:

and R _χ and R ₂ are H and n is 1.

The reacting step may be effectively performed as a one- pot procedure in the presence of a mild nonnucleophilic base such as diisopropylethylamine. Useful solvents include dipolar nonaqueous solvents such as dimethyl formamide and tetrahydrofuran. The reaction may be carried out over a range of temperatures from -25°C to 60°C, but preferably at 0-10°C.

The subject invention also provides a process of preparing the composition having the structure:

which comprises:

(a) reacting a compound having the structure:

with ammonia under suitable conditions to form a compound having the structure:

(b) reacting the compound formed by step (a) with an acylating agent under suitable conditions to form a plurally acylated compound having the structure:

(c) reacting the plurally acylated compound formed by step (b) with a compound having the structure:

C0 ₂ CH ₃

under suitable conditions to form an alkylated amide having the structure:

B

(d) reacting the alkylated amide formed by step (c) with benzene-l,3,5-trithiol under suitable conditions to form a sulfide having the structure:

Bo

(e) deprotecting the sulfide formed by step (d) under suitable conditions to form a free amine ester having the structure:

(f) re-acylating the free amine ester formed by step (e) under suitable conditions to form an acylamine ester having the structure:

B

(g) saponifying the acylamine ester formed by step (f) under suitable conditions to form an acylamine acid having the structure:

(h) activating the acylamine acid formed by step (g) under suitable conditions to form an acylamine activated ester having the structure:

(i) the acylamine activated ester formed by step (h) under suitable conditions to form a free amine activated ester having the structure:

and (j) cyclizing the free amine activated ester formed by step (i) under suitable conditions to form the composition.

Reacting step (a) may be carried out in the presence of a miscible co-solvent such as methanol, and occurs in high yield when performed at ambient temperatures. Reacting step (b) may be carried out using a variety of acylating agents in the presence of nonnucleophilic base and 4-dimethylaminopyridine catalyst. Common agents include t-Boc-Cl and Amyloxycarbonyl chloride. Reacting step (c) is efficiently performed using sodium hexamethyldisilylazide in tetrahydrofuran solution. Preferred temperatures range from -80°C to -70°C. A dry ice bath providing a temperature of -78°C is convenient for this purpose. Reacting step (d) is readily effected in the presence of a nonnucleophilic base such as diisopropylethylamine in a dipolar nonaqueous solvent such as tetrahydrofuran. Deprotecting step (e) occurs well by using a mild acid, such as trifluoacetic acid in the presence of a scavenger such as anisole. Re- acylating step (f) is carried out using a variety of acylating agents. t-Boc ₂ 0 is a preferred acylating agent for the purposes of the synthesis. Saponifying step (g) may be carried out using such bases as lithium hydroxide and sodium hydroxide. Lithium hydroxide is a preferred base. Activating step (h) is carried out using pentafluorophenol in the presence of various condensing agents, including DCC and EDC. De-protecting step (i) may be performed using a mild acid such as trifluoroacetic acid and a scavenger. Cyclizing step (j) is performed using a dropwise addition technique and a nonnucleophilic base such as diisopropylethylamine in a dipolar nonaqueous solvent such as dimethyl aceta ide or dimethyl formamide.

Receptors 1 and 2 are capable of high binding selectivity among simple amino acid derivatives (Table I) . With Boc- protected, W-methylamide acid derivatives, enantio- selectivity ranges from 1.7 to 3.0 kcal/mole with the L isomer always being bound preferentially (entries 1/2, 5/6, 7/8, 9/10, 12/13).

Side-chain functionality can also be distinguished by the chiral receptors (Table I; entries 1-8 vs 9, 10 and 12,13). The side-chain hydroxyls of serine and threonine contribute ^' about 2 kcal/mole to association energies and effectively distinguish these amino acids from Ala, Val, and Leu. Such hydroxylated L-amino acids bind better than O-benzyl-L-serine (entry 11) by about 3 kcal/mole. Nuclear magnetic resonance data suggest that the operative mode of complexation involves close proximity of the C-terminal group of the amino acid derivatives to all four aromatic rings in the host. Entries 14-17 (Table I) suggest that other binding modes may apply to amino acid derivatives having small N-terminal func¬ tionalities such as acetyl.

The chiral host compounds may be utilized in any manner suitable for the intended purpose. For example, the host may be covalently bound to a polymer by modification of the synthetic method described above by replacing phloroglucinol or a similar starting material with one which has the additional substitution of an alkyl, aryl, or aralkyl, linker containing a reactive moiety at its terminus, comprising a halide, amine, carboxylate, alcohol, or thiol, if necessary in suitably protected form.

The resulting chiral polymer may serve as an adsorbent for use as a convenient extractive reagent, in which the polymer may be combined with a mixture of racemic amino acid derivatives or a mixture of compounds related by

differing side-chain substitution in a range of polar or nonpolar solvents. After sufficient agitation at a temperature suitable for promoting binding of one component in the mixture, ranging from -90 to 180°C, preferably from 0 to 35°C, the polymeric complex is then separated by gravity or suction filtration, centrifugation, or sedimentation and decanting. The desired enantiomeric derivative or related compound may be obtained by washing the polymer with a suitable buffer, solvent, or mixture of solvents at a temperature suitable for releasing the derivative from the polymeric host. The chiral polymer may also serve as an adsorbent in a chromatographic column, in which the mixture of enantiomers or related compounds may bind with different affinities, and then be eluted after washing with a suitable buffer, solvent, or mixture of solvents. The adsorbent is preferably prepared using finer meshes (>400 U.S. mesh) of chloromethylated 0.5-2.0% divinyl-benzene cross-linked polystyrene and either aminoethyl, hydroxyethyl, or carboxyethyl derivatives of phloroglucinol or benzenetrithiol, according to the described procedure. Any polymeric resin selected from the group consisting of polyacrylamide, phenol- formaldehyde polymers, polymethacrylate, carbohydrates, aluminates, and silicates may serve as the solid medium.

While not wishing to be bound by a particular theory of action, the high selectivity in the binding of various substrates to a host molecule as observed while practicing the subject invention could result from high conformational homogeneity and substantial host/guest contact. Monte Carlo conformational searching using the MacroModel/AMBER force field, in which Phe is modeled by Ala, predicts that the chiral receptors have similar conformations with C3 symmetry. The Phe's are folded into turns around the periphery of a large binding cavity with dimensions (~βA diameter) similar to those of -

cyclodextrin. Some variability remains in the central ring Ar-X-CH ₂ -Ar' torsion angles, with little effect on the shape and nature of the binding cavity. Experimental evidence supporting the predicted structure includes NH- CH _a coupling -constants (J l = 8.1 Hz; J 2 = 8.0 Hz) and N-H infrared bands (free and hydrogen-bonded: 3434,3321) cm ^-1 ) in dilute CDC1 ₃ solution. The chiral host on forming a bound complex undergoes only minor conformational change, according to simulated annealing calculations. Specific contacts which may be responsible for the selective binding interactions in the complex could include three N-H/0=C hydrogen bonds, according to molecular mechanics modelling.

The chiral hosts of the subject invention bind diamides of certain amino acids with high selectivity which is dependent upon the nature of the amino acid side chain (2-kcal/mol for serine vs alanine) and the identity of the N-alkyl substituent (>3 kcal/ ol for methyl vs tert- butyl) . These synthetic hosts are among the most enantioselective known, and bind certain derivatives of L-amino acids with selectivities as high as 3 kcal/mol. No other composition has been available to the art which achieves binding energy differentials of the magnitude herein disclosed for diastereoselective complexation of amino acid derivatives.

The subject invention provides a composition of matter having the structure:

wherein each of A, B, C, X, Y and Z is independently O, NH, N(CH ₂ ) _m H, N(C=0) (CH ₂ ) _m H, CH ₂ , S or Se; each of R, , R- and R ₃ is independently linear or branched alkyl, aryl, (C=0) (CH ₂ ) _p H, NH(C=0) (CH ₂ ) _p H, OH, COOH, NH ₂ or SH; and m, n, and p are integers from 0 to about 6. In one embodiment, the invention provides the composition wherein A, B and C are O. In another embodiment, the invention provides the composition wherein A, B and C are S. The invention also provides the composition wherein R.,, R ₂ , and R ₃ are each independently phenyl, 4-hydroxy- phenyl, pyridyl, pyrrolyl, indolyl or naphthyl. The invention further provides the composition wherein R.*, R ₂ , and R ₃ are 4-hydroxymethylphenyl. The invention also provides the composition wherein R.,, R ₂ , and R ₃ are each independently 4-allyloxyphenyl, 4-alkoxyphenyl, 4-

acyloxyphenyl or 4-(dye-substituted-acyloxy)phenyl. In one embodiment, the invention provides the composition wherein R ₁ , R ₂ , and R ₃ are 4-(dye-substituted-alkoxy)phenyl and wherein the dye-substituted-alkoxy moiety is OCH ₂ CH ₂ N(Et)C ₆ H ₄ -N=N-C ₆ H ₄ N0 ₂ -para,trans,para. In another embodiment, the invention provides the composition 43, wherein n is 1.

The invention provides a compound which comprises the composition having the structure:

wherein each of A, B, C, X, Y and Z is independently O, NH, N(CH ₂ ) _m H, N(C=0) (CH ₂ ) _m H, CH ₂ , S or Se; each of R,, R ₂ and R ₃ is independently linear or branched alkyl, aryl, (C=0) (CH ₂ ) _p H, NH(C=0) (CH ₂ ) _p H, OH, COOH, NH ₂ or SH; and m, n, and p are integers from 0 to about 6; wherein the composition is bound to a solid support. In one embodiment, the invention provides a compound which

comprises the above composition bound to a derivative of an amino acid. In a preferred embodiment, the invention provides the compound above wherein the derivative is an oligopeptide.

The subject invention provides a process of obtaining a purified enantiomeric isomer of a compound of interest from a mixture of optical isomers of such compound which comprises contacting the mixture of isomers with the composition having the structure:

wherein each of A, B, C, X, Y and Z is independently O, NH, N(CH ₂ ) _m H, N(C=0) (CH ₂ ) _m H, CH ₂ , S or Se; each of R,, R ₂ and R ₃ is independently linear or branched alkyl, aryl, (C=0) (CH ₂ ) _p H, NH(C=0) (CH ₂ ) _p H, OH, COOH, NH., or SH; and , n, and p are integers from 0 to about 6; under conditions such that the enantiomeric isomer binds to the composition to form a complex, separating the resulting

complex from the mixture, treating the complex so as to separate the enantiomeric isomer from the composition, and recovering the purified enantiomeric isomer.

The invention also provides a process of obtaining a purified organic compound of interest from a mixture of organic compounds able to form hydrogen bonds, which comprises contacting the mixture with the composition shown above under conditions such that the organic compound binds to the composition to form a complex, separating the resulting complex from the mixture, treating the complex so as to separate the organic compound from the composition, and recovering the purif: i compound. In one embodiment, the invention provides the above process wherein the purified organic compound is an amino acid derivative. In a preferred embodiment, the invention provides the above process wherein the amino acid derivative is an oligopeptide. In another embodiment, the invention provides the above process wherein the purified organic compound is a biopolymer. In a particular embodiment, the invention provides the above process wherein the biopolymer is an enzyme. In another embodiment, the invention provides the above process wherein the purified organic compound is a monosaccharide or a polysaccharide. In additional embodiments, the invention provides the above processed directed to optical isomers and organic compounds wherein the composition shown above is bound to a permeable membrane.

The invention provides a composition of matter having the structure:

wherein A has the structure:

and R ₂ and R ₂ are each independently H, F, a linear or branched chain alkyl, arylalkyl, alkoxyalkyl, aminoalkyl, alkylaminoalkyl, hydroxyalkyl, (cycloalkyl) alkyl, or acylalkyl group, or an aryl group, a linear or branched chain alkylaryl, pyridyl, thiophene, pyrrolyl, indolyl or naphthyl group; wherein X is NH, O, S, S=0, S(=0) ₂ or NR where R is linear or branched alkyl, acyl, aryl, or O; wherein Q is selected from the group consisting of H,

linear or branched chain alkyl, acyl and aryl; and m and n are integers from 0 to about 3. In one embodiment, the invention provides the above composition wherein R and R ₂ are H. In another embodiment, the invention provides the composition wherein Q is an acyl group. In a certain embodiment, the invention provides the above composition wherein Q is an acyl moiety sustituted by a dye molecule. In a particular embodiment, the invention provides the above composition wherein the acyl moiety is: (C=0)CH ₂ CH ₂ (C=0)OCH ₂ CH ₂ N(Et)C ₆ H ₄ -N=N-C ₆ H ₄ N0 ₂ - para,trans,para. The invention also provides a compound which comprises the above shown composition of matter bound to a solid support. The invention also provides a compound which comprises the above shown composition bound to a derivative of an amino acid. In a preferred embodiment, the invention provides the above compound wherein the derivative is an oligopeptide.

The subject invention provides a process of obtaining a purified enantiomeric isomer of a compound of interest from a mixture of optical isomers of such compound which

comprises contacting the mixture of isomers with the composition having the structure:

wherein A has the structure:

and R ₁ and R ₂ are each independently H, F, a linear or branched chain alkyl, arylalkyl, alkoxyalkyl, aminoalkyl, alkylaminoalkyl, hydroxyalkyl, (cycloalkyl)alkyl, or acylalkyl group, or an aryl group, a linear or branched chain alkylaryl, pyridyl, thiophene, pyrrolyl, indolyl or naphthyl group; wherein X is NH, O, S, S=0, S(=0) ₂ or NR where R is linear or branched alkyl, acyl, aryl, or 0; wherein Q is selected from the group consisting of H,

linear or branched chain alkyl, acyl and aryl; and m and n are integers from 0 to about 3; under conditions such that the enantiomeric isomer binds to the compositon to form a complex, separating the resulting complex from the mixture, treating the complex so as to separate the enantiomeric isomer from the composition, and recovering the purified enantiomeric isomer.

The invention also provides a process of obtaining a purified organic compound of interest from a mixture of organic compounds able to form hydrogen bonds, which comprises contacting the mixture with the above shown composition under conditions such that the organic compound binds to the composition to form a complex, separating the resulting complex from the mixture, treating the complex so as to separate the organic compound from the composition, and recovering the purified compound. In one embodiment, the invention provides the above process wherein the purified organic compound is an amino acid derivative. In a preferred embodiment, the invention provides the above process wherein the amino acid derivative is an oligopeptide. The invention also provides the above process wherein the purified organic compound is a biopolymer. In a particular embodiment, the invention provides the above process wherein the biopolymer is an enzyme. In another embodiment, the invention provides the above process wherein the purified organic compound is a monosaccharide or a polysaccharide. In additional embodiments, the invention provides the above processes directed to optical isomers and organic compounds wherein the composition shown above is bound to a permeable membrane.

The subject invention provides a composition of matter having the structure:

wherein each of A, B and C is independently O, NH, N(CH ₂ ) _m H, N(C=0) (CH ₂ ) _m H, CH ₂ , S or Se; wherein Q is selected from a group consisting of H, linear or branched chain alkyl, acyl and aryl; and m and n are integers from 0 to about 5. In one embodiment, the invention provides the above composition wherein A, B and C are 0. In another embodiment, the invention provides the above composition wherein A, B and C are NH. In a certain embodiment, the invention provides the above composition wherein Q is an acyl group. In a particular embodiment, the invention provides the above composition wherein Q is an acyl moiety sustituted by a dye molecule. In a more particular embodiment, the invention provides the above

composition wherein the acyl moiety sustituted by a dye molecule is:

(C=0)CH ₂ CH ₂ (C=0)OCH ₂ CH ₂ N(Et)C ₆ H ₄ -N=N-C ₆ H ₄ N0 ₂ - para,trans,para. In another embodiment, the invention provides the above composition wherein n is 1 or 2. In another embodiment, the invention provides a compound which comprises the above shown composition of matter bound to a solid support. In one embodiment, the invention provides a compound which comprises the above shown composition bound to a derivative of an amino acid. In a particular embodiment, the invention provides the compound above wherein the derivative is an oligopeptide.

The invention also provides a process of obtaining a purified enantiomeric isomer of a compound of interest from a mixture of optical isomers of such compound which

comprises contacting the mixture of isomers with the composition having the structure:

wherein each of A, B and C is independently O, NH, N(CH ₂ ) _m H, N(C=0) (CH ₂ ) _m H, CH ₂ , S or Se; wherein Q is selected from a group consisting of H, linear or branched chain alkyl, acyl and aryl; and m and n are integers from 0 to about 5; under conditions such that the enantiomeric isomer binds to the compositon to form a complex, separating the resulting complex from the mixture, treating the complex so as to separate the enantiomeric isomer from the composition, and recovering the purified enantiomeric isomer.

The invention also provides a process of obtaining a purified organic compound of interest from a mixture of organic compounds able to form hydrogen bonds, which comprises contacting the mixture with the above shown

composition under conditions such that the organic compound binds to the composition to form a complex, separating the resulting complex from the mixture, treating the complex so as to separate the organic compound from the composition, and recovering the purified compound. In one embodiment, the invention provides the above process wherein the purified organic compound is an amino acid derivative. In another embodiment, the invention provides the above process wherein the amino acid derivative is an oligopeptide. In yet another embodiment, the invention provides the above process wherein the purified organic compound is a biopolymer. In a particular embodiment, the invention provides the above process wherein the biopolymer is an enzyme. In another particular embodiment, the invention provides the above process wherein the purified organic compound is a monosaccharide or a polysaccharide. In additional embodiments, the invention provides the above processes directed to optical isomers and organic compounds wherein the composition shown above is bound to a permeable membrane. As used herein, the permeable membrane is one of the type used to filter and separate molecules according to selected physical parameters, such as size, molecular weight, etc. A variety of synthetic and natural memranes are suitable for the purpose, including polyamide, nylon, perfluoroethylene, cellulose acetate, etc.

The subject invention also provides a composition of matter having the structure:

wherein R ₂ and R ₂ are each independently H, F, a linear or branched chain alkyl, arylalkyl, alkoxyalkyl, aminoalkyl, alkylaminoalkyl, hydroxyalkyl, (cycloalkyl)alkyl, acyl¬ alkyl, aryl, a linear or branched chain alkylaryl, pyridyl, thiophene, pyrrolyl, indolyl or naphthyl group; wherein Q is selected from a group consisting of H, linear or branched chain alkyl, acyl or aryl; and n is an integer from 0 to about 3. In one embodiment, the invention provides the above composition wherein R _λ and R ₂ are H. In another embodiment, the invention provides the above composition wherein Q is an acyl group. In a particular embodiment, the invention provides the above composition wherein Q is an acyl moiety sustituted by a dye molecule. In another particular embodiment, the

invention provides the above composition wherein the acyl moiety sustituted by a dye molecule is:

(C=0)CH ₂ CH ₂ (C=0)OCH ₂ CH ₂ N(Et)C ₆ H ₄ -N=N-C ₆ H ₄ N0 ₂ - para,trans,para.

The invention also provides a compound which comprises the composition of matter having the above structure bound to a solid support. The invention also provides a compound which comprises the above shown composition bound to a derivative of an amino acid. In a preferred embodiment, the invention provides the above compound wherein the derivative is an oligopeptide.

The invention also provides a process of obtaining a purified enantiomeric isomer of a compound of interest from a mixture of optical isomers of such compound which

comprises contacting the mixture of isomers with the composition having the structure:

wherein R _λ and R ₂ are each independently H, F, a linear or branched chain alkyl, arylalkyl, alkoxyalkyl, aminoalkyl, alkylaminoalkyl, hydroxyalkyl, (cycloalkyl)alkyl, acyl¬ alkyl, aryl, a linear or branched chain alkylaryl, pyridyl, thiophene, pyrrolyl, indolyl or naphthyl group; wherein Q is selected from a group consisting of H, linear or branched chain alkyl, acyl or aryl; and n is an integer from 0 to about 3; under conditions such that the enantiomeric isomer binds to the compositon to form a complex, separating the resulting complex from the mixture, treating the complex so as to separate the enantiomeric isomer from the composition, and recovering the purified enantiomeric isomer.

The invention also provides a process of obtaining a purified organic compound of interest from a mixture of organic compounds able to form hydrogen bonds, which comprises contacting the mixture with the above shown composition under conditions such that the organic compound binds to the composition to form a complex, separating the resulting complex from the mixture, treating the complex so as to separate the organic compound from the composition, and recovering the purified compound. In one embodiment, the invention provides the above process wherein the purified organic compound is an amino acid derivative. In a preferred embodiment, the invention provides the above process wherein the amino acid derivative is an oligopeptide. In another embodiment, the invention provides the above process wherein the purified organic compound is a biopolymer. In a particular embodiment, the invention provides the above process wherein the biopolymer is an enzyme. In another embodiment, the invention provides the above process wherein the purified organic compound is a monosaccharide or a polysaccharide. In additional embodiments, the invention provides the above processes directed to optical isomers and organic compounds wherein the composition shown above is bound to a permeable membrane.

Table I. ΔG's of Association (kcal/mole) of 1 and 2 with Amino Acid Derivatives

-ΔG ^a Saturation % ΔΔG ^C Entry Peptide substrate 1 2 1 2 1 2

1 N-Boc-D-Ala-NHMe 1.7 2.1 53 70

2 N-Boc-L-Ala-NHMe 3.9 3.8 93 90 2.2 1.7

3 N-Boc-L-Ala-NHBn 1.4 51

4 N-Boc-L-Ala-NHtBu nc ^d

5 N-Boc-D-Val-NHMe 1.5 1.5 51 54

6 N-Boc-L-Val-NHMe 4.4 4.0 79 74 2.9 2.5

7 N-Boc-D-Leu-NHMe 1.5 1.6 64 60

8 N-Boc-L-Leu-NHMe 4.1 3.8 88 78 2.6 2.2

9 N-Boc-D-Ser-NHMe 3.8 4.4 86 94

10 N-Boc-L-Ser-NHMe >6.1 >6.2 95 96 >2.3 >1.8

11 N-Boc-L-Ser(OBn)-NHMe 3.1 83

12 N-Boc-D-Thr-NHMe 3.2 3.6 84 90

13 N-Boc-L-Thr-NHMe >6.2 lg ^e >95 >3.0

14 N-Ac-D-Ala-NHMe 2.7 90

15 N-Ac-L-Ala-NHMe 3.9 94 1.2

16 N-Ac-D-Ala-NHtBu 2.0 59

17 N-Ac-L-Ala-NHtBu 3.0 85 1.0

a Measured by NMR titration at 25C with 1 or 2 at 0.5 mM ϊntration in CDC1 ₃ . ^b Extent of extrapolated saturation at end of titration. ^c Enantioselectivity, ΔG(D) - ΔG(L). ^d No complexation detected. ^e Too large to measure accurately.

The following Experimental Details are set forth to aid in an understanding of the invention, and are not intended, and should not be construed, to limit in any way the invention set forth in the claims which follow thereafter.

Experimental Details

EXAMPLE 1

Preparation of methyl-3.5-dimethyl-benzoate: A solution of 3,5-dimethyl benzoic acid (25 g, 0.17 mol) in methanol (250 ml) was treated with sulfuric acid (1 ml, cat. amount) and heated to reflux. After 10 hours, the solution was cooled to room temperature, concentrated to approximately 1/2 volume and poured into 200 ml of crushed ice. The mixture was extracted twice with 200 ml portions of diethyl ether. The organic phase was extracted with saturated aqueous sodium carbonate, dried over magnesium sulfate, and concentrated under reduced pressure. The resulting solid was recrystallized from hexanes to yield the product (24.5 g, 90% yield) as volatile white plates. mp=32-35°C (lit. mp=35-36°C)

EXAMPLE 2

Preparation of methyl-3.5-bis (bromomethyl) -benzoate:

A solution of methyl-3,5-dimethyl-benzoate (16.8 g, 0.10 mol) in carbon tetrachloride (150 ml) was treated with N- bromosuccinimide (35.6 g, 0.20 mmol) , and benzoyl peroxide (500 mg, cat. amount) , and heated to reflux. After 3 hours, the mixture was cooled to room temperature and filtered through a sintered glass funnel. The filtrate was concentrated under reduced pressure and recrystallized from diethyl ether/hexanes (1:1) to yield the product (18.0 g, 55% yield) as a granular white solid. mp 64-70°C (lit. mp 65-69°C) ; TLC (20% ethyl acetate/hexanes) : R _f =0.65 (UV active, CAM stain)

EXAMPLE 3

Preparation of methv 1 - 3 -bromo ethy 1 - 5 - bis(Boc)aminomethyl-benzoate:

A mixture of sodium hydride (3.6 g, 90 mmol) and N,N- dimethylformamide (150 ml) was cooled to 0°C and treated with solid di-tert-butyliminodicarboxylate (17.4 g, 76.0 mmol) with vigorous stirring. The mixture was stirred at 0°C for 15 minutes, the ice bath was removed and a solution of methyl-3,5-bis(bromomethyl)-benzoate (23.6 g, 73.2 mmol) in N,N-dimethylformamide was added dropwise over 30 minutes. After 12 hours the mixture was poured into 100 ml 1/4 saturated aqueous ammonium chloride and extracted with the three 100 ml portion of hexanes. The organic layer was dried over magnesium sulfate, concentrated under reduced pressure and chromatographed using a gradient of 10-20% ethyl acetate/hexanes to yield the product (22.0 g, 66% yield). TLC (20% ethyl acetate/hexanes) : R _f =0.55 (UV active, CAM stain)

EXAMPLE 4

Preparation of benzene-1.3.5-tris)methyl-3'-thiomethyl- 5 ^/ -bis(Boc)aminomethyl-benzoatel :

A solution of benzene-l,3,5-trithiol (900 mg, 5.16 mmol) in tetrahydrofuran 0 ml) was treated with N,N-diiso- propylethylamine (3.1 ml, 17.67 mmol) with vigorous stirring. The mixture was allowed to stir until all solids had dissolved and was treated with solution of methyl-3-bromomethyl-5-bis(Boc)aminomethyl-benzoate (8.1 g, 17.67 mmol) in tetrahydrofuran. After 24 hours the mixture was poured into 100 ml saturated aqueous ammonium chloride and extracted with three 100 ml portions of diethyl ether. The organic phase was dried over magnesium sulfate, concentrated under reduced pressure, and the resulting oil was chromatographed using a gradient of 20-40% ethyl acetate/hexanes to yield the product (5.10 g, 76% yield) as a colorless oil. TLC (40% ethyl acetate/hexanes) : R _f =0.65 (UV active, C1 ₂ /TDM stain)

-72 -

EXAMPLE 5

5

Preparation of benzene-1.3.5-trisfmethyl-3'-thiomethyl- 5 '-aminomethyl-benzoate hydrochloride salt]: A solution of benzene-1,2,3,5-tris[methyl-3'-thiomethyl- 5'-bis(Boc)aminomethyl-benzoate hydrochloride] (4.8 g, 0 3.67 mmol) in absolute methanol (25 ml) was treated with 25 ml "10% methanolic HCl" (a mixture of 2.5 ml acetyl chloride and 22.5 ml absolute methanol) and allowed to stir at room temperature for 3 hours. All volatiles were removed under reduced pressure and the resulting white 5 powder was dried under high vacuum. The product (3.00 g, quant, yield) was used without additional purification.

EXAMPLE 6

0 Preparation of N-α-Boc-L-tyrosine methyl ester:

A solution of L-tyrosine methyl ester hydrochloride (10.0 g, 43.2 mmol) in N,N-dimethylformamide (100 ml) was cooled to 0°C and treated with solid di-tert-butyl dicarbonate (13.0 g, 65 mmol) and triethylamine (6.6 ml, 5 47.5 mmol). After one hour the ice bath was remove and the solution was allowed to warm to room temperature. After 4 hours the solution was poured into 200 ml ethyl acetate, extracted with 100 ml 1.0 M aqueous hydrochloric acid, 200 ml saturated aqueous sodium bicarbonate, 200 ml 0 saturated aqueous sodium chloride, dried over magnesium sulfate, and concentrated under reduced pressure. The crude product (12.56 g, 98% yield) was used without additional purification.

5 EXAMPLE 7

Preparation of N-Q-BOC-L-tyrosine-O-allyl ether methyl ester:

A solution of N-α-BOC-L-tyrosine methyl ester (12.5 g, 42.3 mmol) in N,N-dimethylformamide (100 ml) was treated with allyl bromide (4.5 ml, 51.8 mmol), tetra-n- butylammonium iodide (1.5 g, 4.3 mmol) and potassium carbonate (12 g, 86.4 mmol) and allowed to stir overnight. After 14 hours the mixture was poured into 200 ml ethyl acetate and extracted with 100 ml 1.0 M aqueous citric acid, 100 ml saturated aqueous sodium bicarbonate, saturated aqueous sodium chloride, dried over magnesium sulfate and concentrated under reduced pressure. The resulting oil was chromatographed using a gradient of 20-40% ethyl acetate/hexanes to yield the product (14.0 g, 99% yield) a colorless oil. TLC (40% ethyl acetate/hexanes) : R _f =0.40 (UV active, Cl ₂ /TDM stain)

EXAMPLE 8

Preparation of N- -BOC-L-tyrosine-O-allyl ether:

A solution of N-α-BOC-L-tyrosine-O-allyl ether methyl ester (14.0 g, 41.7 mmol) in a mixture of tetrahydrofuran (100 ml) and water (10 ml) was treated with lithium hydroxide (10.0 g, 260 mmol) and allowed to stir at room temperature. After 6 hours all of the starting material had been consumed, as determined by thin layer chromatography, and the reaction mixture was diluted with 200 ml ethyl acetate and acidified to pH 2 with 1.0 M aqueous potassium hydrogen sulfate. The organic phase was dried over magnesium sulfate and concentrated under reduced pressure to yield the product (13.4 g, quant, yield) . The product was used without additional purification.

EXAMPLE 9

Preparation of N-α-BOC-L-tyrosine-O-allyl ether-p- nitrophenyl ester:

A solution of N-α-BOC-L-tyrosine-O-allyl ether (13.4 g, 41.7 mmol) in chloroform (100 ml) was cooled to 0°C and was treated with p-nitrophenol (17 g, 128 mmol) , N- hydroxybenzotriazole (3.0 g, 21.3 mmol) and N, N'- dicyclohexylcarbodiimide (10.5 g, 51.2 mmol). The mixture was allowed to stir overnight at room temperature. After 15 hours the solution was filtered to remove N,N'-dicyclohexylurea, concentrated under reduced pressure and chromatographed using 100% chloroform to yield the product (13.7 g, 74% yield) as a yellow oil which solidified upon standing. TLC (100% chloroform) : R _f =0.30 (UV active, ninhydrin stain)

EXAMPLE 10

Preparation of benzene-1,3.5-tris[methyl-3'-thiomethyl- 5 '-aminomethyl-(N-α-BOC-L-tyrosine-amide-O-allyl-ether)- benzoate1 :

A solution of benzene-1,3,5-tris[methyl-3'-thiomethyl-5'- aminomethyl-benzoate hydrochloride salt] (2.33 g, 2.86 mmol) in N,N-dimethylformamide (30 ml) was treated with N,N-diisopropylethylamine (1.9 ml, 10.87 mmol) with vigorous stirring until all solids had dissolved. The solution was cooled to 0°C and treated with solid N-α-BOC- L-tyrosine-O-allyl-ether-p-nitrophenyl ester (2(0 (4.3 g, 9.72 mmol). After one hour the ice bath was removed and the mixture was mixed with silica gel (13 g) and all volatiles were removed under reduced pressure. The pre- absorbed reaction mixture was placed directly onto a chromatography column containing silica gel equilibrated with 40% ethyl acetate/hexanes. The column was eluted with 40% ethyl acetate/hexanes to remove unreacted p- nitrophenyl ester and most of the p-nitrophenol. The product was then eluted with 10% methanol/chloroform to yield a fine yellow powder slightly contaminated with p-

nitrophenol. The mixture was redissolved in methylene chloride and extracted with two 200 ml portions of 0.5 M aqueous sodium hydroxide. The organic phase was dried over magnesium sulfate and concentrated under reduced pressure to yield to product (4.41 g, 96% yield) as a pale yellow powder. TLC (8% acetone/methylene chloride) : R _f =0.45 (UV active, C1 ₂ /TDM stain)

EXAMPLE 11

Preparation of benzene-1, 3 , 5-tris f 2"- (trimethyl)silylethyl-3'-thiomethyl-5 '-aminomethyl-(N-α- Boc-L-tyrosine-amide-O-allyl-ether)-benzoate] : A suspension of benzene-1,3,5-tris[methyl-3'-thiomethyl- 5'-aminomethyl-(N-cκ-BOC-L-tyrosine-amide-0-allyl-ether)- benzoate] (4.3 g, 2.66 mmol) in 2-(-trimethyl)silyl- ethanol (10 ml, 70 mmol) and toluene (10 ml) was thoroughly purged with argon and treated with titanium ethoxide (0.050 ml, catalytic amount) and heated to reflux. After 6 hours the mixture was cooled to room temperature, filtered through a pad of Celite (diatomaceous earth) and concentrated under reduced pressure. The resulting oil was chromatographed using a gradient of 100% methylene chloride-5% methanol/methylene chloride to yield the product (4.2 g, 84% yield) as a pale yellow powder. TLC (8% acetone/methylene chloride) : R _f =0.85 (UV active, C1 ₂ /TDM stain).

EXAMPLE 12

Preparation of benzene-1. 3 . 5-tris r 3 ' -thiomethyl-5 ' -amino¬ methyl- (N-α-BOC-L-tyrosine-amide-O-allyl-ether) -benzoic acid] :

A s o l u t i o n o f b e n z e n e - 1 . 3 . 5 - 1 r i s [ 2 ' - (trimethyl) silylethyl-3 ' -thiomethyl-5 ' -aminomethyl-) N-α-

BOC-L-tyrosine-amide-O-allyl-ether) -benzoate] (4.2 g,

2.24 mmol) in tetrahydrofuran (75 ml) was treated with

tetra-n-butylammonium fluoride (1.0 M solution in tetrahydrofuran) (10.08 ml, 10.08 mmol). After 4 hours the solution was diluted with 100 ml ethyl acetate and acidified to pH 2 with 1.0 M aqueous potassium hydrogen sulfate. The organic layer was dried over magnesium sulfate and concentrated under reduced pressure to yield the product (3.50 g, quant, yield). The product was used without additional purification. TLC (10% methanol/- chloroform) : R _f =0.15 (UV active, C1 ₂ /TDM stain).

EXAMPLE 13

Preparation of benzene-1,3,5-tris[pentafluorophenyl-3 ' - thiomethyl-5 '-aminomethyl-(N-α-BOC-L-tyrosine-amide-O- allyl-ether)-benzoatel :

A solution of benzene-1,3,5-tris[3'-thiomethyl-5'-amino- methyl-(N-α-BOC-L-tyrosine-amide-O-allyl-ether)-benzoic acid] (3.50 g, 2,24 mmol) in tetrahydrofuran (50 ml) was treated with pentafluorophenol (3.7 g, 20.16 mmol), and was allowed to stir at room temperature. After 3 hours the reaction mixture was concentrated under reduced pressure and the resulting oil was chromatographed using a gradient of 100% methylene chloride-10% acetone/methylene chloride to yield the product (2.60 g, 56 % yield) as a white powder. TLC (5% acetone/methylene chloride): R _f =0.45 (UV active, CAM stain).

EXAMPLE 14

Preparation of benzene-1.3.5-tris[-fpentafluorophenyl-3'- thiomethyl-5 '-aminomethyl-(L-tyrosine-amide-O-allyl- ether)-benzoate -trifluoroacetate salt! : A solution of benzene-1,3,5-tris[pentafluorophenyl-3'- thiomethyl-5 '-aminomethyl-(N-α-BOC-L-tyrosine-amide-O- allyl-ether)-benzoate] (2.5 g, 1.21 mmol) in methylene

chloride (100 ml) was treated with anisole (10.0 ml, 93.0 mmol) and trifluoroacetic acid (50.0 ml). After 3 hours the mixture was concentrated under reduced pressure, resuspended in toluene and concentrated again. Finally, the product was triturated three times in diethyl ether to yield the product (1.90 g, quant, yield) as a white powder. The product was used without additional purification.

EXAMPLE 15

Preparation of L-tyrosine macrocycle tris-allyl ether: A solution of benzene-1,3,5-[{pentafluorophenyl-3'-thio- methyl-5 '-aminomethyl-(L-tyrosine-amide-O-allyl-ether)- benzoate}-trifluoroacetate salt] (1.47 g, 1.21 mmol) in N,N-dimethylacetamide (25 ml) was added via syringe pump (33 hours) to a stirring solution of N,N- diisopropylethylamine (30.0 ml, 180 mmol) in tetrahydrofuran (500 ml) . Twelve hours after the addition had been completed, the reaction mixture was diluted with an equal volume of ethyl acetate, extracted with two 200 ml portions of 5% aqueous hydrochloric acid, two 200 ml portions of saturated aqueous sodium bicarb¬ onate, and 100 ml saturated aqueous sodium chloride. The organic phase was dried over magnesium sulfate, con¬ centrated under reduced pressure and chromatographed using a gradient of 100% chloroform - 5% methanol/chlor¬ oform to yield the product (892 mg, 65% yield) as a pale yellow powder. TLC (25% acetone methylene chloride) : R _f =0.45 (UV active, CAM stain).

EXAMPLE 16

Preparation of the L-tyrosine macrocycle 1: A solution of the L-tyrosine macrocycle tris-allyl ether (50 mg, 0.041 mmol) in tetrahydrofuran (15 ml) was treated with 5,5-dimethyl-cyclohexan-l,3-dione (100 mg,

0.71 mmol) and tetrakis-(triphenylphosphine)palladium (10.0 mg, cat. amount) and allowed to stir at room temperature. After 4 hours the solution was diluted with 50 ml ethyl acetate and extracted with three 20 ml portions of saturated aqueous sodium bicarbonate and 20 ml saturated aqueous sodium chloride. The organic phase was dried over magnesium sulfate and concentrated under pressure. The resulting solid was chromatographed using 10% methanol/chloroform to yield the product (1; 40.0 mg, 89% yield) as a pale yellow powder. TLC (10% methanol/chloroform): R _f =0.20 (UV active, CAM stain).

EXAMPLE 17

Preparation of a receptor bound to a solid support:

A solid phase peptide reaction vessel was charged with Merrifield resin (chloromethylated polystyrene cross- linked with 2% divinylbenzene; 100 mg, 0.100 meq) , the macrocyclic tris-phenol made according to Example 16 (110.0 mg, 0.100 mmol), potassium carbonate (14 mg, 0.100 mmol) , and N,N-dimethylformamide (2 ml) . The mixture was placed on a rotary agitator for four days. The reaction mixture was washed successively with 5x5 ml portions of methylene chloride, methanol, deionized water, methanol, and methylene chloride. The resulting solid was dried under high vacuum and weighed to dermine the amount of alkylation. The coupled resin weighed 116.3 mg (approximately 15% based on chloromethyl groups) . The organic washes were diluted with 100 ml ethyl acetate and extracted with 50 ml portions of 1 M aqueous potassium hydrogen sulfate, saturated aqueous sodium bicarbonate, and saturated aqueous sodium chloride. The organic phase was dried over magnesium sulfate, concentrated under reduced pressure, and chromatographed using 10% methanol/chloroform to recover unreacted tris-phenol (40.1 mg) . The infrared spectrum shows type I, II, and III amide bands (1650, 1510, and 1230 cm ^"1 ) .

EXAMPLE 18

Method of resolution of N-α-Boc-DL-valine methylamide: The resin-bound tyrosine receptor (50 mg) prepared in Example 17 was placed in a solid-phase peptide synthesis reaction vessel (a cylinderical glass container with a ground glass joint (standard taper 14/20) on top, a coarse glass frit, and a stopcock at the bottom; treated with dichlorodimethylsilane to reduce adhesion to the glass surface) was pre-swelled by washing 5 times with 50 ml portions of chloroform and forcing excess solvent out with a stream of argon. A solution of 10 mM N-α-BOC-DL- valine methylamide (57.6 mg) was dissolved in perdeuterobenzene, and incubated with the resin-bound host for five minutes. The resin was washed with acetone (5 times 50. ml) . The collected washings were concentrated under reduced pressure to afford 14.5 mg of resolved N-α-BOC-valine methylamide. The extent of enantiomeric enrichment was determined as follows. The BOC group was removed by treatment with a large excess of anhydrous methanolic HCl. On neutralization with tri¬ ethylamine, the resulting amine was reacted with N-α-BOC- L-alanine p-nitrophenyl ester to give N-α-BOC-L- alanylvaline methylamide (19.0 mg, 97.2%) after chromatography. NMR integration and comparison with authentic DL diastereomeric compounds revealed an 85:15 mixture of diastereomers, i.e., 70% enantiomeric enrichment. The resin could be regenerated by washing five times with 50 ml portions of methanol, dried under a stream of argon, and re-swelled with chloroform.

EXAMPLE 19

Synthesis of an 0-allγl tyrosyl C ₃ -svmmetric receptor: N-Boc-O-Allyl-L-tyrosine amide 3. Di-tert-butyl dicarbonate (13.0 g, 59.6 mmol) was added to a solution of L-tyrosine methyl ester hydrochloride (10.0 g, 43.2

mmol) and i-Pr ₂ NEt (6.6 mL, 38.0 mmol) in DMF (100 mL) . The reaction mixture was poured into 1 M aq KHS0 ₄ after 8 h and extracted with ethyl acetate (3X) . The combined extracts were washed with aq NaHC0 ₃ and brine. Drying and evaporation afforded a yellow oil which was dissolved in DMF (100 mL) . Potassium carbonate (12.0 g, 86 mmol), allyl bromide (4.5 mL, 51.8 mmol), and n-Bu ₄ -NI (1.5 g, 4.3 mmol) were added. The reaction mixture was stirred for 16 h, poured into 1 M aq KHS0 ₄ , and extracted with ethyl acetate (3X) . The combined organic layers were washed with aq NaHC0 ₃ and brine. Drying and solvent removal afforded N-Boc-O-allyl-L-tyrosine methyl ester as a yellow oil.

Ammonia (20 L) was condensed into a solution of N-Boc-O- allyl-L-tyrosine methyl ester in CH ₃ OH (60 mL) at -78 °C in a high pressure glass reaction vessel. The vessel was sealed and slowly warmed to rt. After 2 days, the vessel was cooled to -78 °C and opened. Argon was bubbled through the solution while it was allowed to warm slowly to rt. After 1 h, the solution was transferred to a round-bottom flask and all volatiles were removed. The light brown solid residue was washed with hexane/ethyl acetate (2:1) to yield the N-Boc-O-allyl-L-tyrosine amide 3 (13.0 g, 94%) as a white solid: mp 145 °C; R _f 0.28 (diethyl ether); ² H NMR (CDC1 ₃ ) δ 1.40 (9 H, s) , 2.97 (1 H, dd, J=7.2, 13.6 Hz), 3.05 (1 H, dd, J = 6.4, 13.6 Hz), 4.30 (1 H, m) , 4.51 (d, J = 5.2 Hz), 5.06 (1 H, m) , 5.29 (1 H, dd, J = 1.4, 10.8 Hz), 5.38 (1 H, dd, J = 1.2, 19.2 Hz), 5.40 (1 H, bs) , 5.78 (1 H, bs) , 6.04 (1 H, m) , 6.86 (2 H, d, J = 8.4 HZ), 7.13 (2 H, d, J = 8.4 Hz); ¹³ C NMR (CDC1 ₃ ) δ 27.9, 37.3, 55.2, 68.5, 77.2, 114.6, 117.3, 128.4, 130.0, 132.9, 148.8, 157.3, 173.7; IR (KBr) 3677, 3390, 3195, 1678, 1661, 1515, 1248, 1168 cm ^-1 ; HRMS calcd for C ₁₇ H ₂₄ N ₂ 0 ₄ 320.1736; found 320.1741.

N-Boc-O-Allyl-L-tyrosine N,N-di-Boc-amide 4. To a solution of 3 (3.0 g, 9.38 mmol) in CH ₂ C1 ₂ at rt was added i-Pr ₂ -NEt (6.52 mL, 37.5 mmol) , DMAP (192 mg, 1.56 mmol), and di-tert-butyl dicarbonate (5.12 g, 23.5 mmol). After 2 h the reaction mixture was washed with 1 M aq KHS0 ₄ and 1 M aq NaHC0 ₃ . Drying, concentration, filtration through a pad (10 g) of silica gel with 30% ether in pentane and evaporation afforded crude 4 (4.39 g, 90%) as a pale yellow oil. Trituration with hexane gave 4 as a white amorphous solid: mp 87°C; R _f 0.55 (50% ether/hexane) ; ^X H NMR (CDC1 ₃ ) cS-1.36 (9 H, s) , 1.53 (18 H, s) , 2.76 (1 H, dd, J = 6.8, 14.0 Hz), 3.13 (1 H, dd, J = 4.8, 14.0 Hz) , 4.51 (d, J = 5.2 Hz), 5.05 (1 H, d, J = 9.6 Hz), 5.26 (1 H, dd, J = 1.6, 10.0 Hz), 5.41 (1 H, dd, J = 1.6, 18.0 Hz), 5.57 (1 H, dd, J=4.8, 6.8 Hz) , 6.05 (1 H, m) , 6.84 (2 H, d, J = 8.4 Hz), 7.12 (2 H, d, J = 8.4 Hz); ¹³ C NMR (CDC1 ₃ ) δ 27.6, 28.3, 38.3, 54.8, 68.8, 79.6, 85.2, 114.6, 117.5, 128.1, 130.7, 133.4, 149.2, 155.0, 157.8, 174.7; IR (KBr) 2979, 2361, 1788, 1728, 1609, 1511, 1458, 1368, 1316, 1224, 1144, 1011 cm ^"1 ; HRMS (M + 1) calcd for C ₂₇ H ₄₃ N ₂ 0 ₈ 521.2863, found 521.2854. Anal. Calcd for C ₂₇ H ₄₂ N ₂ 0 ₈ : C, 62.05; H,8.10; N,5.36. Found: C,62.29; H,7.70; N,5.38.

Boc-Amidomethyl methyl (bromomethyl)benzoate 5.

NaN(TMS) ₂ (l M THF; 4.75 mL, 4.75 mmol) was added dropwise to a solution of 4 (2.50 g, 4.81 mmol) in THF (40 mL) at -78 °C. Methyl 3,5-bis(bromomethyl)benzoate (1.86 g, 5.77 mmol) and n-Bu ₄ NI (431 mg, 1.17 mmol) were added after 5 min, and the reaction mixture was warmed to 10-15 °C. After 45 min the reaction mixture was diluted with ether (40 mL) and washed with aq NH ₄ C1. The aqueous phase was extracted with ether (2X) and the extracts were washed with brine. Drying, concentration, and flash chromatography (10-20% ethyl acetate/hexane) afforded 5 (2.96 g, 82%) as a white foam; R _f 0.45 (50% ether/hexane); ^λ NMR (CDC1 ₃ ) 1.33 (9 H, s) , 1.42 (18 H, s) , 3.09 (1 H,

dd, J = 8.4, 14.0 Hz), 3.44 (1 H, dd, J = 6.4, 14.0 Hz), 3.89 (3 H, s) , 4.45 (2 H, s) , 4.51 (d, J = 5.2 Hz) , 4.56 (1 H, J = 15.2 Hz), 4.95 (1 H, dd, J = 15.2 Hz) , 5.29 (1 H, d, J = 10.8 Hz), 5.41 (1 H, d, J = 17.6 Hz), 5.72 (1 H, dd, J = 6.0, 8.4 Hz), 6.05 (1 H, m) , 6.81 (2 H, d, J

= 8.4 Hz), 7.18 (2 H, d, J = 8.4 Hz), 7.50 (1 H, s) , 7.90 (1 H, s) , 7.92 (1 H, s) ; ¹³ C NMR (d ₆ -DMSO) 527.2, 27.3, 33.2, 35.0, 48.0, 52.2, 61.5, 68.1, 82.0, 83.1, 114.3, 117.1, 120.9, 128.0, 128.7, 130.1, 130.5, 133.8, 133.8, 139.0, 139,3. 151.9, 151.9, 156.8, 165.6, 173.6; IR (film) 3286, 2979, 2933, 1787, 1752, 1710, 1611, 1511, 1481, 1458, 1368, 1302, 1238, 1142, 1027, 999, 923 cm ^"1 ; HRMS (M + 1) calcd for C ₃₇ H _5Q O ₁₀ N ₂ Br 763.2637, found 763.2755.

Nona-Boc Trisulfide 6. Compound 5 (2.0 g, 2.63 mmol) was added to a suspension of benzene-1,3,5-trithiol ⁶ (140 mg, 0.80 mmol) and i-Pr ₂ NEt (610 μL, 35.1 mmol) in THF (20 mL) at rt. The reaction mixture was quenched with aq NH ₄ C1 after 6 h and extracted with ether (2X) . After a brine wash and concentration, flash chromatography (5-3:1:1 pentane:benzene:diethyl ether) gave 6 (1.38 g, 78%) as a solid white foam: mp 80 °C; R _f 0.35 (33% hexane/diethyl ether); ¹ H NMR (CDC1 ₃ ) δ 1.31 (27 H, s) , 1.40 (54 H, s) , 3.13 (3 H, dd, J = 8.4, 13.2 Hz), 3.44 (3 H, dd, J = 6.0, 13.2 Hz), 3.86 (9 H, s) , 4.08 (6 H, s) , 4.50 (6 H, d, J = 5.2 Hz), 4.54 (3 H, d, J = 15.2), 4.98 (3 H, J = 15.2 HZ), 5.27 (3 H, J = 9.6 Hz), 5.40 (3 H, d, J = 15.2 Hz) , 5.71 (3 H, dd, J = 6.0, 8.4 Hz), 6.03 (3 H, m) , 6.82 (6 H, d, J = 8.4 HZ), 7.02 (3 H, S) , 7.19 (6 H, d, J = 8.4 HZ), 7.51 (3 H, S) , 7.86 (6 H, s) ; ¹³ C NMR (d ₆ -DMSO) δ 27.2, 27.3, 35.0, 35.7, 48.0, 52.0, 61.4, 81.9, 83.0, 114.3, 117.1, 120.9, 124.2, 127.1, 128.5, 129.8, 130.5, 132.6, 133.8, 137.6, 138.0, 138.9, 151.3, 151.8, 156.8, 165.6, 173.6; IR (film) 3420, 2979, 1791, 1733, 1653, 1636, 1609, 1558, 1511, 1474, 1457, 1436, 1368, 1314, 1219, 1145, 1011, 960, 926, 850, 772, cm ^"1 . Anal. Calcd

for C ₁₁₇ H ₁₅₀ N ₆ O _3O S ₃ : C,63.40; H,6.82; N,3.79. Found: C,62.87; H,6.80; N,3.68.

Tri-Boc Trisulfide 7. Trifluoroacetic acid (75 mL) and anisole (19 mL) were added via syringe to a solution of 6 (11.4 g, 5.15 mmol) in CH ₂ C1 ₂ (150 mL) at rt. After 18 h, concentration gave a light pink residue which was triturated with ether to yield a white powder (low resolution mass spectrophotometric data; m/z = 1316) . That material was dissolved in DMF (80 mL) containing K ₂ C0 ₃ (3.78 g, 31 mmol), i-Pr ₂ NEt (5.4 mL, 31 mmol), and di-tert-butyl dicarbonate (5.61 g, 25.75 mmol) . After 17 h, the reaction mixture was poured into ethyl acetate (1000 mL) and washed with 1 M aq KHS0 ₄ , NaHC0 ₃ , and brine. Drying, concentration, and trituration with ether afforded 7 (7.07 g, 85%) as a white powder: mp 138°C; R _f 0.39 (10% acetone/CH ₂ Cl ₂ ) ; ² H NMR (d ₆ -DMSO) δ 1.29 (27 H, s) , 2.67 (3 H, dd, J = 10.4, 14.0 Hz), 2.86 (3 H, dd, J = 4.4, 14.0 Hz), 3.76 (9 H, s) , 4.10 (3 H, m) , 4.27 (12 H, m) , 4.45 (6 H, d, J = 5.2 Hz) , 5.20 (3 H, d, J = 10.4 Hz), 5.33 (3 H, J = 17.6 Hz) , 6.04 (3 H, m) , 6.77 (6 H, d, J = 8.4 Hz), 7.10 (3 H, s) , 7.11 (6 H, d, J = 8.4 Hz), 7.51 (3 H, s) , 7.73 (3 H, s) , 7.81 (3 H, s) , 8.50 (3 H, m) ; ¹³ C NMR (d ₆ -DMSO) 527.2, 30.8, 35.8, 36.6, 41.7, 52.1, 56.2, 68.0, 78.0, 114.2, 117.2, 123.7, 127.0, 128.1, 129.8, 130.1, 132.6, 133.8, 137.7, 140.6, 155.3, 156.6, 162.3, 165.9, 172.0,; IR (KBr) 3310, 2926, 1720, 1511, 1437, 1367, 1310, 1242, 1167, 1024 cm ^-1 . Anal. Calcd for ^C ₈₇ ^H ₁ 0 ₂ ^N ₆ ° ₁₈ ^S ₃ ^: C,64.66; H,6.36; N,5.20. Found: C,64.07; H,6.50; N,5.02.

Pentafluorophenyl Ester of 7. A solution of 1 M aq LiOH (15 mL, 15 mmol) was added to 7 (500 mg, 0.309 mmol) in THF/EtOH/H ₂ 0 (6:3:2, 100 mL) . The reaction mixture was poured into 1. M aq KHS0 ₄ after 8 h and extracted with ethyl acetate (3X) . After the extracts were washed with

brine and dried, solvent removal afforded the crude acid as a light brown powder which was washed with ether.

Pentafluorophenol (600 mg, 3.26 mmol) and l-(3-(dimethyl- amino)propyl)-3-ethylcarbodiimide hydrochloride (320 mg, 1.69 mmol) were added to a stirred solution of the crude acid (470 mg) in THF (7.0 mL) . After 4 h of stirring, concentration gave a brown residue from which the tris(pentafluorophneyl ester) (435 mg, 68%) was isolated by flash chromatography (0-10% acetone/CH ₂ C1 ₂ ) as an amorphous white solid: mp 158 °C; R _f 0.24 (5% acetone/CH ₂ Cl ₂ ) ; ² H NMR (CDC1 ₃ ) δ 1.35 (27 H, s) , 2.98 (6 H, d, J = 7.2 Hz), 4.02 (6 H, s) , 4.35 (6 H, d, J = 4.4 Hz), 4.44 (6 H, ddd, J = 1.6, 1.8, 5.6 Hz), 5.25 (3 H, dd, J = 1.6, 10.4 Hz), 5.36 (3 H, dd, J = 1.6, 17.2 Hz) , 5.99 (3 H, dddd, J = 1.6, 1.8, 10.4, 17.2 Hz) , 6.76 (6 H, d, J = 8.4 Hz), 6.98 (3 H, s) , 7.05 (6 H, d, J = 8.4 Hz) , 7.34 (3 H, s) , 7.90 (3 H, s) , 7.99 (3 H, s) ; ¹³ C NMR (CDC1 ₃ ) δ 29.5, 32.3, 39.0, 39.3, 44.1, 57.5, 70.1, 81.7, 116.1, 119.0, 128.9, 129.9, 130.1, 131.4, 131.6, 134.5, 135.7, 138.7, 140.0, 140.9, 141.0, 157.0, 158.9, 163.3, 173.5; IR (KBr) 3371, 2979, 1686, 1615, 1517, 1444, 1368, 1224, 1166 cm ^-1 . Anal. Calcd for C ₁₀₂ H ₉₃ F ₁₅ 0 ₁₈ S ₃ : C, 59.13; H,4.52; N,4.06. Found: C,58.53; H,4.52; N,4.06.

Tyrosine Macrocycle 2A. Anisole (12 mL) and trifluoroacetic acid (60 mL) were added via syringe to a stirring solution of the above tris(pentafluorophenyl ester) (3.16 g, 1.52 mmol) in CH ₂ -C1 ₂ (125 mL) . After 6 h, the reaction mixture was concentrated. The resulting pink oil was triturated with ether to yield the tris-TFA amine salt as a white powder (3.20 g) .

A solution of the above tris-TFA amine salt (1.50 g, 0.710 mmol) in N,N-dimethylacetamide (25 mL) was added dropwise over 36 h to a rapidly stirred solution of i-

Pr ₂ NEt (30 mL, 172 mmol) in THF (1200 mL) at rt. After

the solution was stirred for an additional 12 h, 800 mL of THF was removed and the remaining solution was diluted with ethyl acetate (400 mL) . The solution was then washed with 0.5 M aq HCl (2X) , aq NaHC0 ₃ (2X) , and brine. Drying, concentration, and flash chromatography (10-50% acetone in CH ₂ C1 ₂ ) afforded 2A as an amorphous white solid (680 mg, 78%): mp 200 °C; R _f 0.38 (20% acetone/CH ₂ C1 ₂ ) ; ^~ Ε NMR (CDC1 ₃ ) δ 3.09 (3 H, dd, J = 7.2, 14.0 Hz) , 3.24 (3 H, dd, J = 7.2, 14.0 Hz), 3.78 (3 H, d, J = 15.2 Hz), 4.03 (3 H, dd, J = 4.2, 14.2 Hz), 4.11 (3 H, d, J = 15.2 Hz), 4.38 (3 H, dd, J = 6.8, 14.2 Hz) , 4.53 (6 H, d, J = 5.2 Hz), 4.80 (3 H, dd, J = 7.2, 15.6 Hz), 5.28 (3 H, dd, J = 1.2, 9.6 Hz), 5.41 (3 H, dd, J =• 1.2, 17.2 Hz), 6.04 (3 H, m) , 6.65 (3 H, d, J = 8.0 Hz) , 6.68 (3 H, s) , 6.83 (3 H, bs) , 6.a2 (6 H, d, J = 8.4 Hz) , 7.08 (3 H, s) , 7.21 (6 H, d, J = 8.4 HZ), 7.44 (3 H, s) , 7.55 (3 H, s) ; ¹³ C NMR (CDC1 ₃ ) δ 32.3, 35.8, 37.6, 45.0, 56.1, 70.2, 116.3, 119.1, 127.3, 129.7, 130.3, 131.5, 131.6, 131.8, 134.2, 134.5, 134.7, 137.9, 140.0, 141.0, 158.9, 168.6, 172.4; IR (KBr) 3310, 2926, 1654, 1510, 1457, 1242, 1178, 1113, 1019, 926, 824 cm ^"1 ; HRMS calcd for C ₆₉ H ₆₆ N ₆ 0 ₉ S ₃ 1219.4130, found 1219.4093.

EXAMPLE 20

Determination of Optical Purity of 5. K ₂ C0 ₃ (20 mg, 0.145 mmol) was added to a stirred solution of 3 (100 mg, 0.131 mmol) in CH ₃ OH (2 mL) . After 30 min, the reaction mixture was filtered, diluted with ether (10 mL) , and washed with aq NH ₄ C1 and • brine. Concentration followed by flash chromatography (20% diethyl ether/pentane) afforded N- Boc-O-allyltyrosine methyl ester (40.0 mg, 92%).

The above methyl ester was dissolved in CH ₃ OH (5.0 mL) and acetyl chloride (1.0 mL, 13.5 mmol) was carefully added by pipette. After 3 h, all volatiles were removed and the resulting white solid was taken up in diethyl ether

which was washed with 0.5 M aq LiOH and brine. Drying and solvent removal afforded O-allyl-tyrosine methyl ester as a waxy solid (26.8 mg, 95%) .

O-Allyltyrosine methyl ester (20 mg, 0.084 mmol) was added to a stirred solution of (S)-(-) -methoxy- (trifluoromethyl) phenyl-acetic acid (28.0 mg, 0.120 mmol) and DCC (40 mg, 0.20 mmol) in CH ₂ C1 ₂ (0.50 mL) . After 3 h the reaction mixture was diluted with CH ₂ C1 ₂ (10.0 mL) , filtered, and washed with 0.5 M aq NaOH. Drying and solvent removal afforded the crude (S)-(-)- methoxy-(trifluoromethyl)phenyl-acetamide as a waxy oil containing DCC and N,N-dicyclohexylurea: R _f 0.55 (50% ether/hexane); ^~ -E NMR (CDC1 ₃ ) δ 3.05 (1 H, dd, J = 6.4, 14.4 Hz), 3.13 (1 H, dd, J = 5.4, 14.4 Hz), 3.24 (3 H, s) , 3.74 (3 H, s) , 4.52 (2 H, d, J = 5.2 Hz), 4.87 (1 H, ddd, J = 5.4, 6.0, 6.4 Hz), 5.29 (1 H, d, J = 10.4 Hz), 5.41 (1 H, d, J = 17.6 Hz), 6.06 (1 H, m) , 6.83 (2 H, d, J = 8.4 Hz), 7.05 (2 H, d, J = 8.4 Hz), 7.28 (1 H, d, J = 6.0 Hz), 7.39 (3 H, m) , 7.52 (2 H, m) ; ¹³ C NMR (CDC1 ₃ ) δ 28.9, 37.0, 52.5, 53.4, 53.7, 55.0, 69.0, 77.9, 115.2, 117.8, 127.8, 128.1, 128.7, 129.8, 130.4, 133.4, 158.0, 166.2, 171.9,; IR (film) 3411, 3140, 2953, 2851, 1745, 1696, 1511, 1244, 1233, 1224, 1178 cm ^"1 ; HRMS (M + 1) calcd for C ₂₃ H ₂₅ F ₃ N0 ₅ 452.1685, found 452.1685.

O-Allyltyrosine methyl ester (20 mg, 0.084 mmol) was added to a stirring solution of (RS) -(±)-methox (tri- fluoromethyl) -phenylacetic acid (28.0 mg, 0.120 mmol) and DCC as described in the preceding paragraph to yield an authentic mixture of diastereomeric MTPA amides: R _f 0.55 and 0.57 (50% ether/hexane); ^~ -H NMR (CDC1 ₃ ) δ 2.95-3.15 (2 H, m) , 3.24/3.46 (3 H, s) , 3.74/3.76 (3 H, s) , 4.50 (2 H, m) , 4.80-5.0 (1 H, m) , 5.26 (1 H, d, J = 10.4 Hz) , 5.38 (1 H, d, J = 17.4 HZ), 6.03 (1 H, m) , 6.71/6.83 (2 H, d, J = 8.4 HZ), 6.77/7.05 (2 H, d, J = 8.4 Hz) , 7.28 (1 H, m) , 7.39 (3 H, m) , 7.50 (2 H, m) ; ¹³ C NMR (CDC1 ₃ ) δ 28.9,

36.62, 37.0, 52.2, 52.5, 53.4, 53.7, 54.6, 69.0, 77.9, 114.6, 114.7, 117.4, 117.5, 127.1, 127.7, 128.2, 128.3, 129.1, 129.3, 129.9, 130.0, 133.0, 158.0, 166.2, 171.9; IR (film) 3411, 3140, 2953, 2851, 1745, 1696, 1511, 1244, 1233, 1224, 1178 cm ^"1 ; HRMS (M + 1) calcd for C ₂₃ H ₂₅ F ₃ N0 ₅ 452.1685, found 452.1695.

EXAMPLE 21

One-step synthesis of 9. To an ice cold solution of (-)- (IR, 2R) -dia inocyclohexane (24 mg, 0.211 mmol) and iPrNEt ₂ (0.11 mL, 0.417 mmol) in THF (100 mL) and dimethylacetamide (10 mL) was added 1,3,5- benzenetricarbonyl trichloride (36 mg, 0.139 mmol) as a single portion with stirring. After 2 hours at 0 °C, the mixture was allowed to warm to room temperature and stirred for an additional 12 hours. All volatiles were then removed at reduced pressure and the residue was purified by flash chromatography on silica gel using methylene chloride:methanol = 97:3 to give, as the most mobile compound, an amorphous white solid (9; 5.8 mg, 13%) ^~~ K NMR (CDC1 ₃ ) 58.58 (s, IH) , 8.47 (s, IH) , 8.19 (s, IH) , 7.80 (d, IH, J=6.84 Hz), 7.51 ( , IH) , 7.11 (m, IH) , 4.31 (m, IH) , 4.09 (m, IH) , 3.71 (m, IH) , 2.3 (m, IH) , 2.17 (m, 1H), ^° 2.08 (m, IH) , 1.94 (m, IH) , 1.70-1.03 (m, 8H) ; ¹³ C NMR (CDC1 ₃ ) 5 169.2, 166.4, 165,3, 135.1, 134.8, 133.8, 130.2, 129.6, 128.1, 57.5, 54.7, 52.4, 32.6, 31.7, 31.2, 25.0, 24.7, 24.1; IR (neat) 3336, 2937, 1644, 1537, 1322 cm ^"1 ; MS (FAB) m/z 1309 (M+) . HRMS (FAB) calcd for ^C ₇₂ ^H ₈₅ ° ₁₂ ^N ₁₂ 1309.6410, found 1309.6311.

EXAMPLE 22 Compound 13a*-1

To a solution of 1.03 g of trimesic acid (A) dimethyl ester (4.33 mmol; S. Kasina, J. Nematollahi, Tetrahedron Lett . , 1978, 19, 1403) , 0.93 g of monoBoc-(IR,2R) -

diaminocyclohexane (B, 4.33 mmol; J.B. Hansen, M.C. Nielsen, U. Ehrbar, O. Buchardt, Synthesis , 1982, 404) and 0.59 g of HOBT (4.33 mmol) in 50 ml of CH ₂ C1 ₂ was added 1.0 g of EDC (5.20 mmol) at 0°C. After stirring for 2 hr at 0°C and additional 8 hr at r.t., all volatiles were removed at reduced pressure. The residue was extracted with EtOAc and washed with IN HCl solution (3x50 ml) , saturated NaHC0 ₃ solution (3x50 ml) and saturated NaCl solution (1x50 ml) . After drying over MgS0 ₄ , all volatiles were removed at reduced pressure. The residue was purified by flash chromatography on silica gel using l/l=hexane/EtOAc to give an amorphous white solid (1.49 g, 79.3%).

R _f =0.62 (l/l=hexane/EtOAc)

¹ H NMR (CDC1 ₃ ) 5 1.23 (m, 13H) 1.69 (m, IH) 2.19 (m, IH)

3.45 (m, IH) 3.71 (m, IH) 3.88 (s, 6H) 4.81 (d, IH,

J=8.25 Hz) 7.45 (d, IH, J=7.20 Hz) 8.61 (d, 2H, J=1.53

Hz) 8.69 (t, IH, J=1.56 Hz) ¹³ C NMR (CDC1 ₃ ) 5 165.41, 164.97, 157.11, 135.23, 133.00,

132.03, 130.90, 79.69, 56.63, 53.40, 52.34, 32.36, 32.11,

28.05, 24.99, 24.32

IR (neat) 3320, 2872, 1770, 1700, 1665, 1576 cm ^"1

MS (FAB) m/z 435 (M+l) , 335 (M-Boc)

Compound 13-la* (Hexamethyl ester)

To a solution of 0.15 g of 13a*-l Boc-B-A(C0 ₂ Me) ₂ (0.437 mmol) and 0.10 ml of anisole in 10 ml of CH ₂ C1 ₂ was added 3 ml of TFA. After stirring for 2 hr at r.t., all volatiles were removed at reduced pressure. The residue was washed with ether and dissolved in 10 ml of DMA. To above solution were added 0.13 ml of TEA (0.961 mmol) and 35 mg of 1,3,5-benzenetricarbonyl chloride (0.131 mmol). After stirring overnight at r.t., all volatiles were removed at reduced pressure. The residue was purified by

flash chromatography on silica gel using 5% MeOH in CH ₂ C1 ₂ to give an amorphous white solid (0.15 g, 98.5%).

R _f =0.45 (5% MeOH in CH ₂ C1 ₂ )

¹ H NMR (CDC1 ₃ ) 5 166.82, 166.21, 165.00, 136.22, 134.64,

132.39, 131.94, 130.92, 129.44, 59.82, 54.88, 52.10,

32.10, 31.98, 25.42, 24.98

IR (neat) 3326, 2821, 1730, 1682, 1574 cm ^"1

MS (FAB) m/z 1160 (M+l)

Compound 13-la** (Hexakis(pentafluorophenyl)ester

To a solution of 0.14 g of hexamethyl ester (13-la*, 0.120 mmol) in 5 ml o^ THF, 3 ml of MeOH and 1 ml of water was added 0.84 ml of IN NaOH solution. After stirring for 5 hr at r.t., the reaction mixture was acidified with IN HCl solution and extracted with EtOAc (3x50 ml) . The crude hexacarboxyl acid was dissolved in 3 ml of THF and 10 ml of CH ₂ C1 ₂ , and 0.15 g of penta- fluorophenol 0.840 mmol) and 0.16 g of EDC (0.840 mmol) were added. After stirring for 5 hr at r.t., all volatiles were removed at reduced pressure. The residue was purified by flash chromatography on silica gel using 20% acetone in CH ₂ C1 ₂ to give an amorphous white solid (86 mg, 31.9%) .

R _f =0.58 (10% acetone in CH ₂ C1 ₂ )

¹ H NMR (CDC1 ₃ ) 5 1.40 (m, 4H) 1.83 (m, 2H) 2.11 (m, IH 2.25 (m, 1J) 3.97 (m, 2H) 7.09 (m, IH) 7.32 (m, IH) 8.30 (s, IH) 8.75 (s, 2H) 8.97 (s, IH)

IR (neat) 3320, 2825, 1715, 1675, 1573 cm ^"1

Compound 13* (Tetrahedral symmetric macrocyclic receptor

A solution of 46 mg of hexakis(pentafluorophenyl) ester 13-la** (0.0205 mmol) and 7 mg of (1R,2R) diamino- cyclohexane diTFA salt (B, 0.615 mmol) in 10 ml of DMA

was added to a solution of 0.10 ml of DIPEA (0.589 mmol) in 150 ml of THF at r.t. for 20 hr by syringe pump. After stirring for additional 8 hr, all volatiles were removed at reduced pressure. The residue was purified by flash chromatography on silica gel using 10% MeOH in CH ₂ C1 to give an amorphous white solid (13 mg, 50.0%)

R _f =0.45 (5% MeOH in CH ₂ Cl ₂ ) ^λ E NMR (DMSO-d ₆ ) 5 1.28 (m, IH) 1.72 (m, IH) 1.90 (m, IH) 3.96 (m, IH) 8.16 (s, IH) 8.24 (m, IH)

¹³ C NMR (DMS0-d ₆ ) 5 166.32, 134.70, 129.50, 54.81, 30.50,

24.68

IR (neat) 3321, 2882, 1680, 1573 cm ^"1

MS (FAB) m/z 1309 (M+l)

Compound 13a* (Dye-labelled receptor)

A solution of 86 mg of hexakis (pentafluorophenyl) ester 13* (0.0383 mmol) and 96 mg of N-succinyl dye-3R,4R- pyrrolidine diamine diTFA salt (8-2, 0.126 mmol) in 10 ml of DMA was added to a solution of 0.15 ml of DIPEA (0.884 mmol) in 150 ml of THF at r.t. for 20 hr by syringe pump. After stirring for additional 8 hr, all volatiles were removed at reduced pressure. The residue was purified by flash chromatography on silica gel using 10% MeOH in CH ₂ C1 ₂ to give an amorphous red solid (48 mg, 51.0%).

R _f =0.31 (10% MeOH in CH ₂ C1 ₂ )

^~~ H NMR (1/1=CD ₃ 0D/CDC1 ₃ ) 5 1.12 (m, 3H) 1.32 (m, 4H) 1.70 (m, 2H) 1.95 (m, 2H) 2.52 (m, 4H) 3.42 (m, 2H) 3.62 (m, 2H) 3.95 (m, 4H) 4.20 (m, 2H) 4.68 (m, 2H) 6.71 (d, 2H, J=9.0 Hz) 7.74 (m, 4H) 8.00 (s, IH) 8.18 (m, 4H)

¹³ C NMR (1/1=CD ₃ 0D/CDC1 ₃ ) 5 172.83, 170.12, 166.37 (m) , 156.54, 151.23, 147.11, 143.51, 134.37, 134.33, 133.70,

128.27 (m) , 126.04, 122.31, 111.19, 61.39, 61.29, 53.18

(m) , 46.85, 46.81, 45.43, 32.04, 31.76, 31.64, 28.46, 28.30, 28.18, 24.56, 24.40, 11.89 IR (neat) 3321, 2874, 1722, 1674, 1573 cm ^"1 MS (FAB m/z 2459 (M+l) EXAMPLE 23

Synthesis of A ₄ B1 ₆ : To an ice cold solution of (1R,2R)- diphenylethylenediamine (66 mg, 0.31 mmol) and iPr ₂ NEt (0.16 mL, 0.63 mmol) in THF (lOOmL) and dimethylacetamide (10 mL) was added 1,3,5-benzenetricarbonyl trichloride (55 mg, 0.21 mmol) as a single protion with stirring. After 2 hours at 0°C, the mixture was allowed to warm to room temperature and then to stand for an additional 12 hours. Volatile materials were removed at reduced pressure and the crude product was pruified by flash chromatography on silica gel (3% methanol in methylene chloride) . ^A ₄ ^B1 ₆ ^{was the} mos mobile compound chromatographically and was isolated as an amorphous white solid (9.8 mg, 10%): TLC (5% MeOH in CH ₂ C1 ₂ ) R _f =0.71; ^λ NMR (CDC1 ₃ ) 59.03 (d, IH, J=5.2 Hz), 8.57 (d, IH, J=9.2 Hz), 8.47 (s, IH) , 8.27 (s, IH) , 7.70 (s, IH) , 7.46-7.12 (m, 15H) , 6.81 (m, 1H),5.60 (m, 2H) , 5.38 (dd, IH, J=10.7, 6.9 Hz); ¹³ C NMR (CDC1 ₃ ) 5 168.7, 165.8, 164.7, 141.8-128.5 (m) , 64.7, 62.8, 60.0; IR (neat) 3345, 2933, 1652, 1538, 1321 cm ^'1 ; MS(FAB) m/z 1899 (M+l) .

EXAMPLE 24

3aa: To a solution of trimesic acid pentafluorophenyl dimethyl ester (0.42 g, 1.04 mmol) and tri(aminomethyl)benzene triHCl salt (86 mg, 0.31 mmol) in 10 mL of dry N,N-dimethylacetamide (DMA) was added 0.36 mL of iPr ₂ NEt. After stirring 8 hr, the mixture was concentrated at reduced pressure and purified by flash chromatography (silica gel, 5% MeOH in CH ₂ C1 ₂ ) to give 3aa as an amorphous white solid (0.20 g, 78%) . ^• ' ^■ H NMR (CDC1 ₃ ) 5 3.90 (s, 3H) , 4.38 (m, 2H) , 7.10 (s, IH) , 7.30 (m, IH) ,

8.61 (s, 2H) , 8.70 (s, IH) ; ¹³ C NMR (CDC1 ₃ ) 5 165.74,

165.36, 139.11, 134.99, 132.84, 132.18, 130.77, 126.35, 52.44, 43.71; IR (neat) 3324, 2815, 1730, 1674, 1573 cm ^"1 ; HRMS (FAB) cal-cd. for C ₄₂ H ₄₀ N ₃ O ₁₅ 827.2537 found 827.2539. 2aa: To 3aa (0.1 g, 0.121 mmol) in 5 mL THF, 3 L MeOH and 1 ml water was added 0.85 mL IN NaOH solution. After stirring for 5 hr, the mixture was acidified with IN HCl solution and extracted with EtOAc (3x50 mL) . The extracted hexaacid was dissolved in 3 mL THF and 10 L CH ₂ C1 ₂ , and C ₆ F ₅ OH (0.15 g, 0.84 mmol) and l-(3- dimethyla inopropyl) -3-ethylcarbodiimide (0.16 g, 0.84 mmol) were added. After stirring for 5 hr, concentration and flash chromatography (silica gel, 20% acetone in CH ₂ C1 ₂ ) gave 3bb as an amorphous white solid (63 mg, 30%) .

A solution of 3bb (44 mg, 0.025 mmol) and N-succinyl dye- 3R,4R-pyrrolidine diamine diTFA salt (63 g, 0.084 mmol) in 10 mL DMA was added with stirring to a solution of iPr ₂ NEt (0.19 mL, 1.09 mmol) in 200 mL dry THF at rt over 20 hr by syringe pump. After stirring for an additional 8 hr, concentration and flash chromatography (silica gel, 10% MeOH in CH ₂ C1 ₂ ) gave 2aa as an amorphous red solid (15mg, 26%). ^λ E NMR (1/1 CD ₃ 0D/CDC1 ₃ ) 5 0.70 (t, 3H, J=7.0 Hz), 2.45 (m, 4H) , 3.01 (q, 2H, J=7.0 Hz) , 3.36 (s, br, 2H) , 3.50 (s, br, 2H) , 3.74 (m, IH) , 3.92 (m, IH) , 4.12 (m, 4H) , 4.58 (m, IH) , 4.68 (m, IH) , 6.64 (d, 2H, J=9.1 HZ), 7.72 (m, 5H) , 7.98 (m, IH) , 8.10 (d, 2H, J=9.0 HZ), 8.23 (m, IH) ; ¹³ C NMR (1/1 = CD ₃ 0D/CDC1 ₃ ) 5 172.82, 170.20, 169.72, 169.61, 167.70, 138.51, 136.06, 133.53, 131.27, 130.54, 130.46, 129.54, 126.72, 125.92, 125.22,

124.37, 124.24, 122.31, 118.85, 112.22, 61.34, 55.74, 50.54, 45.70, 44.42, 38.28, 28.76, 28.27, 27.52, 27.22, 11.94; IR (neat) 3324, 2815, 1720, 1675, 1573 cm' ¹ ; MS (FAB) m/z 2127 (M+l) .

EXAMPLE 25

A ₂ B ₂ dimethyl ester (3'). A solution of 5.2 g of 1,3,5- benzenetricarboxyl acid -bis-pentafluorophenyl monomethyl ester (9.34 mmol) and 1.1 g of IR,2R-diaminocyclohexane Jbistrifluoroacetic acid (TFA ₂ ) salt (9.34 mmol) in 50 mL of dimethylacetamide (DMA) was added to a solution of 7.2 mL of iPr ₂ NEt (41.1 mmol, 4.4 equiv.) in 200 mL THF at 25°C over 20 hrs by syringe pump. After stirring for an additional 8 hrs, all volatiles were removed at reduced pressure. The residue was purified by flash chromatography on silica gel using 5% methanol in CH ₂ C1 ₂ to give 3' as an amorphous white solid (1.52 g, 54%): R _f =0.65 (silica gel, 5% MeOH in CH ₂ C1 ₂ ) ; IR (neat) 3440, 2987, 1720, 1672 cm ^"1 ; NMR (CDC1 ₃ , ppm) δ 8.54 (s, IH) , ¹³ C NMR (CDC1 ₃ , ppm) 5 24.78, 31.88, 52.60, 55.59, 129.58, 130.53, 130.91, 135.39, 165.27, 168.16; MS (FAB) m/z 605 (M ⁺ +l) ; HRMS Calc for C ₃₂ H ₃₇ N ₄ 0 ₈ 605.2612, found 605.2607.

A ₂ B ₂ bis(pentafluorophenyl)ester (4'). To a solution of 1 g of dimethyl ester (1.65 mmol) in 30 mL of THF and 15 mL of MeOH was added 3.3 mL of IN NaOH solution. After stirring for 5 hrs at 25°C, the reaction mixture was acidified with IN HCl solution and extracted with EtOAc (3x50 mL) . After solvent removal, the crude dicarboxylic acid was dissolved in 30 mL of THF and 30 mL of CH ₂ C1 ₂ , and 0.61 g of pentafluorophenol (3.3 mmol) and 0.63 g of EDC (3.3 mmol) were added. After stirring for 5 hrs at 25°C, all volatiles were removed at reduced pressure and the residue was purified by flash chromatography on silica gel using 20% acetone in CH ₂ C1 ₂ yielding 4' as an amorphous white solid (0.77 g, 51%): R _f =0.87 (silica gel, 10% acetone in CH ₂ C1 ₂ ) ; E NMR (CDC1 ₃ , ppm 5 9.14 (s, 2H) , 8.87 (s, IH) , 7.12 (m, 2H) , 4.01 (m, 2H) , 2.29 (m, 2H) , 1.90 (m, 2H) , 1.50 (m, 4H) ; HRMS Calc for C ₄₂ H ₃₁ N ₄ O ₈ F ₁₀ 909.6772, found 909.6767.

N-Succinyl dye-3R,4R-pyrrolidine diamine diTFA salt (5'). To a solution of 0.8 g of Disperse Red 1 (Aldrich, 2.09

mmol) , 0.35 L of Et ₃ N (2.75 mmol) and 0.035 g of DMAP (0.25 mmol) in 30 ml of CH ₂ C1 ₂ was added 0.25 g of succinic anhydride (2.50 mmol). After stirring overnight at 25°C, all volatiles were removed at reduced pressure. The residue was purified by flash chromatography on silica gel using 5% MeOH in CH ₂ C1 ₂ to give the mono-dye ester of succinic acid as an amorphous red solid (0.95 g, 96%): R _f =0.31 (silica gel, 5% MeOH in CH ₂ C1 ₂ ) ; ^X H NMR (CDC1 ₃ , ppm) 5 8,34 (d, J=8.9 Hz, 2H) , 3.44 (q, J=7.2 Hz, 2H) , 2.64 (t, J=5.9 Hz, 2H) , 2.49 (t, J=6.0 Hz, 2H) , 1.16 (t, J=7.2 HZ, 3H)

To a solution of 0.095 g of above mono-dye ester (0.201 mmol), 0.030 g of HOBT (0.221 mmol) and 0.087 g of 3R,4R- di(Boc-amino)pyrrolidine (0.221 mmol, prepared by Boc protection (Boc ₂ 0, NEt ₃ ) and debenzylation (ammonium formate, Pd/C) from the known N-benzyl 3R,4R- diaminopyrrolidine; D.R. Reddy and E.R. Thornton, J. Chem . Soc , Chem . Commun . , 172 (1992)) in 30 mL of CH ₂ Cl ₂ was added 0.042 g of EDC (0.221 mmol) at 0°C. After stirring for 2 hr at 0°C and additional 8 hr at 25 °C, all volatiles were removed at reduced pressure. The residue was purified by flash chromatography on silica gel using 5% MeOH in CH ₂ C1 ₂ to give Boc ₂ -5' as an amorphous red solid (0.12 g, 79%): R _f =0.62 (silica gel, 5% MeOH in CH ₂ C1 ₂ ) ; ^~ Ε NMR (CDC1 ₃ , ppm) 5 8.33 (d, J=8.9 Hz, 2H) , 7.92 (dd, J=12.0, 9.0 Hz, 4H) , 6.80 (d, J=8.9 Hz, 2H) , 3.54.(q, J=7.2 Hz, 2H) , 3.23-3.10 (m, 2H) , 2.65- 2.49 (m, 4H) , 1.44 (s, 18H) , 1.26 (t, J=7.2 Hz, 3H) ; ¹³ C NMR (CDC1 ₃ , ppm) 5172.83, 169.98, 156.73, 151.25, 126.66, 126.22, 125.37, 124.64, 122.61, 119.58, 111.42, 61.43, 54.86, 50.34, 48.70, 45.62, 28.76, 28.52, 28.28, 28.22, 12.24; MS (FAB) m/z 698 (M ⁺ +l) ; HRMS Calc for C ₃₄ H ₄₈ N ₇ 0 ₉ 698.3514, found 698.3512.

To a solution of 0.12 g of Boc ₂ -5' (0.159 mmol) and 0.15 ml of anisole in 15 ml of CH ₂ C1 ₂ was added 5 ml of TFA at

25°C. After stirring for 2 hr, all volatiles were removed at reduced pressure and the residue (crude 5' TFA ₂ ) was washed with ether and used for next step without further purification.

Dyed receptor (2'). A solution of 0.12 g of (pentafluorophenyl) ester 4' (0.131 mmol) and 0.11 g of 5'. TFA ₂ (0.146 mmol, l.l eq.) in 10 mL of DMA was added via syringe pump over 20 h to a solution o: 0.23 mL of iPr ₂ NEt (1.31 mmol) in 200 ml of THF at 2 °C. After stirring for additional 8 hr, all volatiles were removed at reduced pressure. The residue was then purified by flash chromatography on silica gel using 3% MeOH in CH ₂ C1 ₂ to give receptor 2 ' as an amorphous red solid (31 mg, 20.4%): R _f =0.41 (silica gel, 5% MeOH in CH ₂ C1 ₂ ) ; IR (neat) 3391, 2893, 1721, 1673, 1530 cm ^"1 ; ^X H NMR (1:1=CDC1 ₃ :CD ₃ 0D, ppm) 5 8.49 (s, 4H) , 8.27 (s, 4H) , 8.01 (s, 4H) , 7.98.(d, 4H, J=9.0 Hz), 7.64 (dd, 8H, J=8.8, 1.4 HZ), 6.55 (d, 4H, J=8.9 Hz) , 4.56 (m, 4H) , 4.08 (m, 8H) , 1.61 (m, 8H) , 1.21 (m, 16H) , 1.01 (t, 6H, J=7.1 Hz) ; ¹³ C NMR (CDC1 ₃ :CD ₃ 0D=1:1, ppm) 5 172.83, 170.45, 167.97, 165.32, 156.55, 151.20, 146.99, 134.17, 134.05, 133.67, 129.85, 128.81, 128.68, 128.51, 128.32, 125.97, 124.32, 122.22, 111.20, 61.34, 57.07, 54.49, 52.00, 45.28, 32.18, 31.47, 30.79, 28.41, 28.07, 24.81, 24.42, 24.29, 11.73; MS(FAB) m/z 2077 (M ⁺ +l) ; HRMS Calc for C ₁₀₈ H ₁₂₀ N ₂₂ O ₂₂ 2076.8850, found 2076.8776.

Solid Phase Binding Assay. The solid phase substrate library was prepared by the encoded split synthesis as described previously and included 50,625 different acylated tripeptide sequences corresponding to all possible combinations of the 15 acylating agents and 15 amino acids (used three times) listed in the text. (A. Borchardt and .C. Still, J. Am . Chem . Soc , 116, 373 (1991); M.H.J. Ohlmeyer, R.N. Swanson, L.W. Dillard, J.C.

Reader, G. Asouline, R. Kobayashi, M. Wigler and W.C. Still, Proc Natl . Acad . Sci . USA, 90, 10922 (1993))

To screen the substrate library for binding, a 10 mg sample of the library (-10 ⁵ beads) was mixed in a 1.5 mL Eppendorf tube with 0.3 mL of -50 μM 2 ' in CHC1 ₃ . After agitation on a wrist-action shaker for 48 h, ~1% of the beads were found to be stained deep red. Fifty-five of these deep red beads were picked by hand under a 4X wide- field microscope and photolyzed (350 nm, 4 hrs) in 1-2 μL of DMF to release the tag molecules. After silylation (CH ₃ C(OTMS)NTMS, -0.1 μL) , electron capture GC was used to analyze the tag complement of each picked bead.

Results and Discussion

Unlike most synthetic host molecules (Notable exceptions: Petti, M.A. ; Shepodd, T.J. ; Barrans, R.E.; Dougherty, D.A. J. Am. Chem. Soc. 1988 110. 6825. Mock, W.L.; Shih, N.-Y. J. Am. Chem. Soc. 1989 111. 2697. Sherman, J.C.; Cram, D.J. J. Am. Chem. Soc. 1989, 111. 4527. Jeong, K.- S.; Muehldorf, A.V. ; Rebek, J. J. Am. Chem. Soc. 1990, 112, 6144. Hong, J.-I.; Namgoong, S.K. ; Bernardi, A.; Still, W.C. J. Am. Chem. Soc. 1991, 113, 5111. Webb, T.H.; Suh, H. ; Wilcox, C.S. J. Am. Chem. Soc. 1991, 113. 5111. Webb, T.H. ; Suh, H. ; Wilcox, C.S. J. Am. Chem. Soc. 1991, 113. 8554. Tanner, M.E.; Knobler, C.B. ; Cram, D.J. J. Org. Chem. 1992, .57., 40) biological receptors are conformationally well defined and large enough to almost fully encapsulate the substrates which they often bind with exquisite selectivity. Constructing analogous synthetic receptors is challenging because such structures seem to require complex atomic networks to form large binding sites and position binding functionality.

A practical synthesis of a highly enantioselective, C ₃ - symmetric host molecule 2A has been developed. The basic strategy is a significant improvement over the relatively lengthy previous synthesis and involves direct addition of a Boc-tyrosine amide anion derivative 4 to methyl 3,5- bis(bromomethyl)-benzoate to give an advanced intermediate 5. The final step, a triple macrolactamization, closes three 19-membered rings simultaneously to produce the bridged macrotricyclic receptor in 70-80% yield.

The C ₃ -symmetric receptor 1A (Hong, J.-LI; Namgoong, S.K.; Bernardi, A.; Still, W.C. J. Am. Chem. Soc. 1991, 113.

5111) described hereinabove is one of the most enantioselective synthetic receptors yet reported and binds N-Boc-N"-methylamide derivatives of simple amino acids with enantioselectivity ranging from 2 to 3 kcal/mol (90-99% ee) (Other enantioselective hosts for neutral molecules: Canceill, J. ; Lacombe, L. ; Collet, A.; J. Am. Chem. Soc. 1985, 107. 6993. Pirkle, W.H. ; Pochapsky, T.C. J. Am. Chem. Soc. 1987, 109. 5957. Sanderson, P.E.J.; Kilburn, J.D.; Still, W.C. J. Am. Chem. Soc. 1989, 111. 8314. Castro, P.P.; Georgiadis, T.M. ; Diederich, F. J. Orσ. Chem. 1989, 54. _/ 5384. Liu, R. ; Sanderson, P.E.J.; Still, W.C. J. Org. Chem. 1990, 55, 5184. Jeong, K.-S.; Muehldorf, A.V. ; Rebek, J. J. Am. Chem. Soc. 1980, 112. 6144. Webb, T.H. ; Suh, H. ; Wilcox, C.S. J. Am. Chem. Soc. 1991, 113. 8554). Such highly enantioselective receptors could have practical applications as resolving agents. A practical synthesis is provided hereinabove of O-allyl tyrosyl receptor 2A, a derivative of 1A which could be covalently bound to a solid support.

For a derivative of 1A which could be bound to a solid support, the O-allyl derivative 2A is appropriate. Such otherwise stable ethers can be deprotected (Kunz, H. ; Unverzagt, C. Angew. Chem.. Int. Ed. Engl. 1984, £3_, 436) with transition metals to free phenols or attached (Tambute, A.; Begos, A.; Lienne, M. ; Macaudiere, P.; Caude, M. ; Rosset, R. ; New J. Chem. 1989, .13., 625) directly to a support using free radical chemistry. The present synthesis avoids the problematic di-tert-butyl iminodicarboxylate anion coupling and addition of nitrogen and amino acid in separate steps. Instead, a more convergent route is provided in which an N-anionic amino acid fragment would be added to bis(bromomethyl)benzoate in a single step. Use of a Boc- stabilized amide ion made from N-Boc-O-allyltyrosine amide is summarized in Figure 1, and proved more reactive

to aσylation than was the primary amide. Thus, the major product with 1 equiv of Boc ₂ 0/DMAP, the tri-Boc material could be isolated in 95% yield.

As shown in Figure 1, the desired Boc-stabilized amide anion could nevertheless be obtained and the planned coupling achieved. Thus starting with commercially available O-ailyl-N-Boc-tyrosine methyl ester 3, NH ₃ was used to prepare the corresponding primary amide which then formed the tri-Boc derivative 4.

On treatment of 4 with sodium hexamethyldisilylazide in THF at -78°C, a rapid deprotonation and Boc-migration occurred, leading to the Boc-stabilized amide anion shown below. While this anion was stable enough to be alkylated with benzylic bromides at low temperatures, warming it to 15°C caused elimination of tert-butoxide leading to 8. For preparation of 2A, 1.2 equiv of 3,5- bis-(bromomethyl)benzoate were used with Bu NI catalysis and obtained 5 in 82% yield.

Although the alkylation proceeded smoothly, 5 might be acidic enough to have racemized under the basic conditions of the alkylation. To test for such racemization, a sample of 5 was treated with K ₂ C0 ₃ in methanol and then HCl in methanol. The first treatment converted (Flynn, D.A. ; Zelle, R.E.; Grieco, P.A. J. Or . Chem. 1983, 4_8, 2824) the C-terminal Boc-amide to methyl ester while the second removed the two N-terminal Boc groups, yielding O-allyltyrosine methyl ester. This material was then coupled using DCC to (S)-α-methoxy-α-

(trifluoromethyl)phenylacetic acid (Mosher's acid) to provide the corresponding amide. ¹ H and ¹³ C NMR comparison of this material with corresponding amides from authentic D- and L-O-allyltyrosine methyl ester showed that very little (<5%) racemization had occurred. Under the conditions of an - ^~ K NMR experiment, as little

as 2% of the epimerized D-tyrosine derivative could have been detected.

The benzylic bromide 5 was then used to triply alkylate sym-trimercaptobenzene (Bellavita, V. Chim. Ital. 1932, 62. 655) using Hunig's base (i-Pr ₂ NEt) providing C ₃ - symmetric 6 in 78% yield. The remainder of the synthesis involved a triple macrolactamization via an activated benzoic acid ester. However, the Boc-substituted amide was quite labile toward acid and base, and conversion of the methyl ester to acid was difficult in its presence. Furthermore, the problematic Boc could not be removed from the C-terminal amide without simultaneously deprotecting the tyrosyl amine. An effective solution to the problem was to remove all Boc protecting groups with TFA and then restore Boc protection of the free amines with Boc ₂ 0 to obtain 7 in 86% yield over both steps.

The three methyl esters of 7 were hydrolyzed using aqueous lithium hydroxide and then the resulting acids were esterfied to pentafluorophenol using l-(3-

(dimethylamino)propyl)-3-ethylcarbodiimide in THF. Flash chromatography (Still, W.C; Kahn, M. ; Mitra, A. J. Org.

Chem. , 1978, 43_, 2923) provided the activated tris(pen- tafluorophenyl ester) in 68% yield. After removing the remaining Boc protecting groups using trifluoroacetic acid, the crude trifluoroacetate salt in N,N-dimethyl- acetamide was added via syringe pump to a large volume of dry THF containing excess Hunig's base. The addition was carried out at room temperature over 36 h using a syringe pump and the final concentration of reactant was -0.5 mM.

This triple macrolactamization was suitable for reactions of this type and provided 2A in 78% yield after silica gel chromatography.

II

A practical synthesis of the C ₃ -symmetric receptor 2A is provided which proceeds in an overall vield of 27% and requires no high resolution chromatographic separations. Solution-phase binding experiments in CDC1 ₃ showed that 2A bound N-Boc-N'-methylamides of amino acids with the same high enantioselectivities as found with 1A. Receptor 2A is useful in solid-phase resolution of protected amino acids.

Also described herein is an example of a large synthetic receptor which has only minimal structural complexity, but has binding selectivities approaching those of biological receptors. This receptor 9 is an A ₄ B ₆ cyclooligomer of trimesic acid (A) and (R,R)- diaminocyclohexane (B) . It binds amino acid residues in peptide chains with very high selectivities for chirality (up to 99+% ee) and side-chain identity (up to 3+ kcal/mol) .

In designing 9, minimal receptor flexibility was achieved by using fragments having few opportunities for conformational isomerism and by joining them with planar amide bonds. Because one of the fragments (A) has three joining points, the neutral, nonpolymeric condensation products of A and B are bridged polycyclics. Among the ways in which A and B can be combined, structure 9 is appealing because of its well-defined binding cavity and appropriately positioned hydrogen-bonding groups.

9 was made by first preparing an amide-linked Boc-B-A-B- A-B-Boc oligomer having the two internal carboxylates activated as pentafluorophenyl esters. When this material was deprotected (TFA, anisole) and slowly added to iPr ₂ NEt/THF, it dimerized to 9 in 39% yield. Alternatively, 9 could be prepared in a single step (13% yield) by simply mixing commercially available A acid

trichloride and B at 3 mM concentration with iPr ₂ NEt in dry THF.

¹ H NMR titrations in CDC1 ₃ showed that 9 formed 1:1 complexes with certain peptides and that W-Ac-L-Val-NHtBu was particularly well bound. To predict the structure of the most stable complex, a 5000-step MacroModel/SUMM conformational search (Goodman, J.M. ; Still, W.C. J. Comput. Chem. 1991, .12., 1110) was carried out using AMBER ^*3 and GB/SA chloroform (Still, W.C; Tempczyk, A. ; Hawley, R.C ; Hendrickson, T. J. Am. Chem. Soc. 1990 112. 6127. CHC1 ₃ parameter set: Hollinger, F. ; Still, W.C, unpublished results). The most stable structure found is a complex held together by four intermolecular hydrogen bonds forming a structure resembling a peptidic three-strand ?-sheet.

A related pair of intramolecular hydrogen bonds (between B's) closes the unbound end of 9 to produce a deep cavity which fully encapsulates the side chain (R) of a bound - peptide. With L-valine, this structure places the side- chain isopropyl near the face of the four aromatic rings (A) of 9. It is incompatible with the ¹ H NMR of the corresponding L-valine methylamide complex, which shows a 2.5 ppm upfield shift for the side-chain methyls and an ~lppm downfield shift of only one of the three different types of host NH's.

The picture which emerges from association energy measurements (Table II) is also in accord with the binding mode described supra which projects L-amino acid side chains into the central cavity of the receptor.

Table II. Binding Energies (Kcal/mol) of Receptor 9 and Peptides ³

By NMR trat on at 25°C o 0.5 mM 9 in CDC1 ₃

(each binding energy is the average of two to five independent measurements on different protons, and the average of two to five independent measurements on different protons, and the largest deviation from the average is <0.2 kcal/mol) . ^b Enantioselectivity favoring L. ^c PGly, phenylglycine. ^d NC, no complex observed. ^e Oc, octanoyl. ^f HSer, homoserine.

Thus peptide derivatives are bound with high selectivity for the L-configuration except when side chains are large (entries 7 and 8) . Valine and phenylglycine side chains appear to fit the binding cavity quite well, but substantial reductions in binding occur when even single methylenes are added (entries 3 vs 4 and 5 and 6 vs 7 and 8) . Removal of side-chain bulk from a near-optimal side chain (iPR) also diminishes binding. Thus stepwise truncation of side-chain iPR to Me to H costs 1.5 kcal/mol per step with L-amino acids. The effect is less significant with D-amino acids, which the model suggests to have side chains projecting away from the binding site and into solvent. Finally, the large binding energies in entries 14 and 15 suggest that 9 can interact associatively with as many as three residues, a feat that appears unique among synthetic receptors. Presumably, the terminal residues of such peptides are able to form additional hydrogen bonds to the outlying amides of the host (NHCO and CONH in the schematic) .

Thus it is possible to assemble a large, conformationally well-defined receptor with remarkable binding properties starting from a few conformationally restricted subunits and well-known synthetic operations. There are doubtless many other such readily accessible heterooligomeric assemblies which have structural and binding properties analogous to those associated with macromolecύlar receptors.

III

The A ₄ B ₆ macrotricycle described herein is remarkable for several reasons. First, it self-assembles in a single step from two commercially available materials, benzene- 1,3,5-tricarbonyl trichloride and the diamine (1R2R)- diaminocyclohexane. Though the yield of this extraordinary reaction is only 13%, the receptor is

readily isolated because it is the most chromatographically mobile of the product formed. Second, A ₄ B ₆ is a highly selective receptor for neutral peptides. For example, it binds derivatives of L amino acids with enantioselectivities as high as 99% ee and can also distinguish between peptides based on the steric requirements of their sidechains. In some cases, this sidechain selectivity can be quite large and exceed 3 kcal/mol even when the peptides being compared differ only by a single methylene (e.g. phenylglycine vs phenylalaninei. In the A B ₆ receptor, the conformationally rigid building blocks used minimize its flexibility. The synthesis and properties of two related A ₄ B ₆ cyclooligomers which are constructed from more conformationally flexible acyclic diamines (1R,2R)-1,2- diphenylethylenediamine (hereinafter Bl) and (2R,3R)-2,3- diaminobutane-l,4-diol (hereinafter B2) . The binding properties in this series of receptors are sensitive to the structure of the components used to assemble them, but rigid cyclic building blocks need not be used to obtain high binding selectivity.

To prepare the receptors, a simple one-step coupling was performed on the amines and the triacid chloride as described for A ₄ B ₆ . With Bl, the A ₄ B1 ₆ receptor was obtained in 10% yield when the coupling was carried out at a concentration corresponding to 6 nM in receptor.

Synthesis of A ₄ B1 ₆ : To an ice cold solution of (1R,2R)- diphenylethylenediamine (66 mg, 0.31 mmol) and IPr ₂ NEt (0.16 mL, 0.63 mmol) in THF (100 mL) and dimethylacetamide (10 mL) was added 1,3,5- benzenetricarbonyl trichloride (55 mg, 0.21 mmol) as a single portion with stirring. After 2 hours at 0°C, the mixture was allowed to warm to room temperature and then to stand for an additional 12 hours. Volatile materials were removed at reduced pressure and the crude product

was purified by flash chromatography on silica gel (3% methanol in methylene chloride) . ^A ₄ ^B1 ₆ ^was ^^ ^{e most} mobile compound chromatographically and was isolated as an amorphous white solid (9.8 mg, 10%); TLC (5% MeOH in CH ₂ C1 ₂ ) R _f = 0.71; ² H NMR (CDC1 ₃ ) 5 9.03 (d, IH, J=5.2 Hz), 8.57 (d, IH, J=9.2 Hz), 8.47 (s, IH) , 8.27 (s, IH) , 7.70 (s, IH) , 7.46-7.12 (m, 15H) , 6.81 (m, IH) , 5.60 (m,

1899 (M+l) .

With B2, a more dilute 1 mM concentration was used to prepare A ₄ B2 ₆ in 7% yield. Both products were readily isolated as the most mobile reaction product on silica gel and were identified by mass spectroscopy and by their symmetry as revealed by ¹³ C and ^λ E NMR.

Binding energies were measured by titrating 0.5 mM solutions of receptor in CDC1 ₃ with various N-acetyl amino acid methylamides and monitoring the receptor protons by 400 MHz NMR. In general, signals which showed the largest shifts upon binding were certain aromatic (H-C) and amide (H-N) protons. The binding energies found are given in Table III and all represent averages of at least two different binding measurements. Scatchard treatment of binding data indicated 1:1 complexes in all cases.

a. All peptides are N-acetyl, methylamides; b. binding energy (kcal/mol); c. enantioselectivity (kcal/mol); NC = no complexation observed.

The binding results obtained with all three receptors support the general picture of the complex shown in Figure 2. In the diagram, '+' and ' - ' represent receptor hydrogen bond donors (H-N) and acceptors (0=C) , respectively. These functionalities presumably not only bind the peptidic substrate by hydrogen bonds but also associate to close the other end of the receptor to create a conical binding cavity which can encapsulate the sidechain (R) of a bound L amino acid.

The binding data in Table III reveals a number of notable trends. First, all receptors bind all D amino acid substrates with roughly the same binding energy (2.0-2.5 kcal.mol) . Thus the high enantioselectivities observed originate from especially favorable binding to L peptide substrates, not by destabilization of binding to D substrates. Second, both the A ₄ B ₆ receptor and the A ₄ B1 ₆ analog have similar binding selectivities despite the construction of the latter from an acyclic diamine. Indeed, A ₄ B1 ₆ binds six of the eight substrates studied with higher enantioselectivity than does A ₄ B ₆ .

Both A ₄ B ₆ and A ₄ B1 ₆ show surprisingly high selectivity among L amino acids which are distinguished only by the size and shape of their unfunctionalized, hydrocarbon sidechains. Amino acids having branched sidechains bind well only when the branch occurs at the substrate β- carbon. Thus valine and isoleucine (R=i-Pr, s-Bu) bind well but leucine (R=i-Bu) does not. The receptors also distinguish substrates by sidechain length. Thus while alanine and butylglycine (R=Me, n-Bu) are rather poorly bound, ethylglycine and propylglycine (R=Et, n-Pr) are among the best substrates. All three receptors distinguish phenylglycine and phenylalanine by >3 kcal/mol. These observations are compatible with the conical-cavity model which favors substrates having more steric bulk near the enlarged, open end of the binding

cavity. Substrates with sidechains that are either too small to fill the cavity or too long to be accommodated are poorly bound. While binding selectivity based on steric effects is known (For example see: F. Diederich, K. Dick and D. Griebel, J. Am. Chem. Soc.. 108, 2273 (1986); W.L. Mock and N.-Y. Shih, J. Am. Chem. Soc.. 110, 4706 (1988); M.A. Petti, T.J. Shepodd, R.E. Barrans and D.A. Dougherty, J. Am. Chem. Soc.. 110, 6825 (1988); D.J. Cram, M.E. Tanner, S.J. Kelpert and C.B. Knobler, J. Am. Chem. Soc.. 113, 8909 (1991); K. Nae ura, K. Ueno, S. Takeuchi, Y. Tobe, T. Kaneda and Y. Sakata, J. Am. Chem. Soc.. 115, 8475 (1993); L. Garle, B. Lozach, J.-P. Dutasta and A. Collet, J. Am. Chem. Soc.. 115, 11652 (1993)), the subtle differences in sidechain bulk which these receptors are able to distinguish energetically by 1-2 kcal/mol is unusual with synthetic receptors. The key to such high steric selectivity appears to coincide with the receptor's ability to fully encapsulate the chemical substructure being distinguished.

Like A ₄ B ₆ and A ₄ B1 ₆ which bind L-peptides based on the steric reguirements of their sidechains, receptor A ₄ B2 ₆ also distinguishes peptide sidechains sterically but with different selectivity. In particular, A ₄ B2 ₆ selects for L-peptides whose sidechains are small and compact; thus alanine, valine and ethylglycine are well-bound while isoleucine, leucine, phenylglycine, propylglycine and butylglycine are more weakly bound relative to the other receptors. Thus A ₄ B2 ₆ appears to have a smaller binding cavity, a property which may follow from cavity occupancy by benzyloxymethyl substituents or from partial cavity collapse due to the flexible nature of the B2 fragment.

These findings suggest that the highly selective binding found with the A ₄ B ₆ receptor may be general to cyclo- oligomeric molecules of this class and that binding selectivity can be altered by starting with different

amine and acid chloride fragments. It may be noted that these receptors incorporate diamine fragments in two different structural environments: the upper and lower macrocycles include four equivalent B amines while two other B's link those macrocycles together. By varying these distinct B fragments independently, even more receptor diversity can be generated.

IV

A new tetrahedrally symmetric cagelike receptor 13*(A _A B ₆ ) , an isomer of 7* (A ₄ B ₆ ) where A and B's are combined in a different way, is described. The global minimum conformation of 13 by molecular mechanics calculation reveals several interesting structural features. First, it is conformationally homogeneous. Within 3 kcal/mol of the global minimum conformation, 13 exits in a single family of closely related conformations. Second, it has a well-defined cavity with hydrogen bond donor/acceptors on symmetrically positioned about the periphery. Each hydrogen bond donor/acceptors may interact with a peptide substrate bound in the central cavity.

The synthesis of 13a*, a dye-labeled derivative of 13* for a solid phase color assay with an encoded substrate library, followed a tridirectional strategy and began with acylation of B-A mono TFA salt having two methyl esters with trimesic acid trichloride as shown in Figure

4. Ester hydrolysis and EDC coupling led to the cyclization precursor. The final step was an intermolecular macrolactamization which used a hexakis

(pentafluorophenyl) ester and three dye-labeled pyrrolidine diamine diTFA salts to close the three 27- membered rings. This cyclization provided the intensely red receptor derivative 13a* in a 50 % yield. The synthesis of 13* followed that of 13a* except using diaminocyclohexanes instead of dye-labeled pyrrolidine

diamines in the final macrocyclization step (50% yield) . However, 13* was highly insoluble in CHC1 ₃ and it was not possible to study its binding properties by the ¹ H NMR titration method in free solution.

To survey the binding properties of the new receptor, a solid phase color assay was employed with an encoded combinatorial library of 50624 acylated tripeptide substrates. The side-chain protected substrate library was screened for binding by treatment with 50 _/ xM solution of the red receptor 13a* in CHC1 ₃ . After 24 hr of equilibration with the library, ca. 1 % of beads had become deep red colored. The deeply stained beads were picked and decoded using gas chromatography to yield the sequences of the most tightly binding substrates. The results are summarized in Table 12.

able 12. Sequences found in sidechain protected substrates bound by 13a* in CHC1 ₃ .

The binding data in Table 12 reveal a number of notable trends. First, sequence selectivity in the AA ₂ -AA- _L position was observed. For example, in 32 of the 55 deep red beads decoded, the residues in the AA _j AA*- _^ were L-Gln- L-Pro. Second, high selectivity was observed in the terminal R and AA ₃ position. In 19 of the 55 deep red beads decoded, the residues in the R-AA ₃ were Me-D-Gln. It is interesting that a simple receptor 13a can interact with substrate in two different ways with high selectivities. To confirm the findings and to estimate the binding energies of binding for some of the sequences, several peptides were resynthesized and their association with 13a* measured in CHC1 ₃ . The results are summarized below in Table 13. Inversion of the stereochemistry of the AA ₃ residue and changing terminal R group from Me to bulky cycPen residue reduces the binding energy by l.l and 1.2 kcal/mol, respectively (peptide substrate l, 2 and 3) . Changing AA ₂ -AA- _L from (L)Ala-Gly to (L)Gin-(L)Pro increases the individual binding energy by 1.1 kcal/mol (peptide substrate 3 and 4). Inversions of the stereochemistry of (L)Gln and (L)Pro weakens binding by 1.3 and 1.5 kcal/mol, respectively and inversion of both weakens binding by 2.1 kcal/mol (peptide substrate 4, 5, 6 and 7).

Table 13. Binding of 13a* with peptides in CHC1 ₃ .

Peptide substrate 1. Me-D-Gln-L-Ala-Gly-polymer 2.Me-L-Gln-L-Ala-Gly-polymer

3.cycPen-D-Gln-L-Ala-Gly-polymer 4.cycPen-D-Gln-L-Gln-L-Pro-polymer 5.cycPen-D-Gln-D-Gln-L-Pro-polymer 6. cycPen-D-Gln-L-Gln-D-Pro-polymer 7.cycPen-D-Gln-D-Gln-D-Pro-polymer

To get an idea of the origin of 13a*'s (L)Gin-(L)Pro selectivity, a 5000-step conformational search was conducted on 13* bound Me-Gly-L-Gln(Tr)-L-Pro-NHTr using MacroModel/AMBER* force field. It was found that within 3 kcal/mol above global minimum conformation the complex exists in a single family of closely related conformations. In the global minimum conformation of complex, shown below diagram, the peptide substrate is held in a rigid conformation by four hydrogen bonds with properly positioned hydrogen bond donor/acceptors of the receptor and ir-it interactions between the aromatic rings of 13* and the Tr protecting group of the substrate.

The origin of 13a*'s Me-(D)Gin selectivity is a matter of conjecture, but it is likely that 13a* interacts with the peptide substrate in a similar way to the C ₃ symmetric receptor described hereinabove. This proposal is based on the similar binding properties from solid phase color assay and binding measurements.

Selective complexation was also found by related binding studies using the analogous side-chain deprotected library. In that case, 13a* showed different and lower substrate selectivity as shown in Table 14.

Table 1 . Sequences found in sidechain deprotected substrates bound by 13a* in CHC1 ₃ .

The most commonly found side-chain deprotected seguences differed from that of the protected library. While in all sequences decoded, at least one Pro or Ser was found, there are no clear trends. The binding mode of 13a* with peptide substrate seems to be that a peptide substrate interacts with more than two hydrogen bond donor/acceptors through the central cavity simultaneously. This proposal is based on the results of the molecular mechanics calculation. When a conformational search was conducted on 13* bound to me- Gly-L-Gln(Tr)-L-Pro-NHTr, changing the Tr groups of the Gin side-chain and carboxyl amide to Ac or tBu allow the peptide substrate to pass through the central cavity of the receptor and interact with the other side of the receptor by hydrogen bonds. Whatever the binding mode of 13a* is with the peptide substrate, it is likely that receptor 13a* has a large enough cavity to accommodate peptide substrates in different ways.

Previously, the present inventors showed a remarkable effect of solvent size on the stability of a molecular complex. The binding energy between a C ₂ symmetric receptor and imidazole can be increased up to 3.3 kcal/mol by changing the solvent from CHC1 ₃ (3.7 kcal/mol) to CHC1 ₂ CHC1 ₂ (7.0 kcal/mol) . ³⁶ It was proposed that receptors with preorganized three-dimensional binding cavities should exhibit increases in binding energies in media composed of increasingly bulky solvent molecules.

Molecular mechanics calculations and CPK modeling studies show that receptor 13a* has a big cavity enough to accommodate a CHC1 ₃ solvent molecule, but too small to accommodate a CHC1 ₂ CHC1 ₂ . It is postulated that changing the solid phase assay solvent from CHC1 ₃ to CHC1 ₂ CHC1 ₂ would increase binding energies and lead to different selectivities. To verify this postulate, a solid phase color assay of 13a* was employed with the side-chain

protected substrate library in CHC1 ₂ CHC1 ₂ . The results are summarized in Table 15.

In the larger solvent, 13a* showed different and higher selectivity. The most commonly found sequences differed from Me-D-Gln-AA _j -AA, and R-D-AA ₃ -L-Gln-L-Pro in CHC1 ₃ and had Me ₂ N-AA ₃ -D-Ala-Gly. There are large differences in molecular dielectric constants as well as surface areas between CHC1 ₃ (e=4.81) and CHC1 ₂ CHC1 ₂ (e=8.20). To get some idea of the origin of the observed differences, the same solid phase color assay was employed in CH ₂ C1 ₂ e=9.08) which has surface area similar to that of CHC1 ₃ and molecular dielectric constant similar to that of CHC1 ₂ CHC1 ₂ , and in CHC1 ₂ CH ₂ C1 (e=8.78) which has surface area and molecular dielectric constant similar to those of CHC1 ₂ CHC1 ₂ . The results are summarized in Table 16 and Table 17.

Table 15. Sequences found in side-chain protected substrates bound by 13a* in CHC1 ₂ CHC1 ₂ .

1. 2. 3, 4, 5, 6, 7. 8, 9,

10.Me ₂ N ll.Me ₂ N 12.Me ₂ N 13.Me ₂ N 14.Me ₂ N 15.Me ₂ N 16.Me ₂ N 17.Me ₂ N 18.Me ₂ N 19.Me ₂ N 20.Me ₂ N 21.Me ₂ N 22.Me ₂ N 23.Me ₂ N 24.Me ₂ N 25.Me ₂ N 26.Me ₂ N 27.Me ₂ N 28.Me ₂ N 29.Me ₂ N 30.Me ₂ N 31.Me ₂ N 32.Me ₂ N 33.Me ₂ N

Table 16. Sequences found in side-chain protected substrates bound by 13a* in CH ₂ C1 ₂ .

Table 17. Sequences found in side-chain protected substrates bound by 13a* in CHC1 ₂ CH ₂ C1.

1. Me ₂ N A-a-K

2. Me ₂ N a-a-Q

3. Me ₂ N a-a-G

4. Me ₂ N a-a-G

5. Me ₂ N a-a-G

6. Me ₂ N V-n-P

7. Me ₂ N v-a-S

8. Me ₂ N v-a-G

9. Me ₂ N v-S-G 10.Me ₂ N 12.Me ₂ N 13.Me ₂ N 14.Me ₂ N 15.Me ₂ N 16.Me ₂ N 17.Me ₂ N 18.Me ₂ N 19.Me ₂ N

20.Me ₂ N 21. Me ₂ N 22.Me ₂ N 23.Me ₂ N 24.Me ₂ N 25.Me ₂ N 26.Me ₂ N 27.Me ₂ N 28.Me ₂ N 29.Me ₂ N

30. Me ₂ N 31.Me ₂ N 32.Me ₂ N 33.Me ₂ N 34.Me ₂ N

While in CHC1 ₂ CH ₂ C1 13a* showed binding selectivities similar to that found in CHC1 ₂ CHC1 ₂ with peptide substrates, in CH ₂ Cl ₂ 13a* showed binding selectivities similar to that found in CHC1 ₃ with peptide substrates. This suggests that the size of solvent molecules play an important role in the observed differences in binding selectivities of 13a* with peptide substrates.

This observation was the starting point to examine the binding properties of 13a* in a wide range of organic solvents. Ten organic solvents were chosen and checked 13a* ^, s binding properties with the peptide substrate library. In some solvents such as chlorobenzene, t-butyl methyl ether and water, the solid phase color assay was not possible due to insolubility of 13a*, receptor 13a* showed selective binding properties with peptide substrates in a number of organic solvents as summarized in Table 18.

Table 18. Binding property of 13a* with side-chain protected substrate in different solvents Solvent Selective staining

Acetone yes

THF yes Ethyl Acetate yes

Dimethoxyethane yes

MeOH yes (with PEG resin)

DMF no staining water/MeOH=l/l no

Although the exact nature of the sequences bound by 13a* in different solvents was not determined, it is interesting that 13a* shows some selective binding property with peptide substrates in a number of organic solvent including MeOH.

V

To discriminate between a vast array of different molecules, biological receptors use well-defined binding sites, with different sizes and arrays of functional groups, which are constructed by various combinations of structurally related building blocks such as α-amino acids. The synthesis and binding properties of new C ₃ symmetric receptors derived from A and B (Fig. 8(a)) are described herein.

Among the ways in which A and B can be combined, structure 11* (Fig. 5) is appealing because of its well- defined binding cavity and appropriately positioned hydrogen bonding groups. The central benzene ring presumably reduces conformational flexibility and provides a hydrophobic region for nonbonded interactions with peptides, and is likely crucial for the high observed binding energies and selectivities.

The synthesis of 11a*, a dye-labeled derivative of 11* for a solid phase color assay with a combinatorial substrate library, exploited its C ₃ symmetry ³⁵ and began with acylation of 1,3,5-tri-aminomethylbenzene with trimesic acid dimethyl mono(pentafluorophenyl) ester as shown in Figure 6(a). Ester hydrolysis and EDC coupling with pentafluorophenol led to the cyclization precursor. The final step was an intermolecular macrolactamization

which used a hexakis-(pentafluorophenyl) ester and three dye-labeled pyrrolidine diamine diTFA salts to close the three 21-membered rings. This cyclization provided the intensely red receptor derivative 11a* in 32% yield. The synthesis of 11* followed that of 11a* except using diaminocyclohexanes instead of dye-labeled pyrrolidine diamines in the final macrocyclization step (31% yield) . However, 11* was highly insoluble in CHC1 ₃ and it was not possible to study its binding properties in free solution. Receptor 12* (Fig. 5) was synthesized by first preparing macrolactam of B-A-B-A-B diTFA salts having two internal methyl carboxyl ester with trimesic acid methyl di(pentafluorophenyl) ester (slow addition to excess iPr ₂ NE _t solution in THF) as shown in Figure 6(b). Ester hydrolysis and EDC coupling with the dye-labeled amines provided 12* with intensely red color in 50% yield (two steps) .

To survey the binding properties of the new receptors, a solid phase color assay was employed with an encoded combinatorial library of 50,625 acylated tripeptide substrates as described hereinabove.

The protected substrate library was screened for binding by treatment with 50 μM solution of the red receptor 11a* in CHC1 ₃ . After 24 hr of equilibration with the library, ca. 10% of the beads had become colored with ca. 1% being very deep red. The most deeply stained beads were

selected and decoded using gas chromatography to yield the sequences of the most tightly binding substrates. The residues found at each position of these substrates are summarized in Table 7 along with the number of instances each residue is found. In the same assay receptor 12* does not stain any beads in the protected substrate library.

Table 7. Residues found in sidechain protected substrates bound by 11a* and Frequencies of occurrence of each residue (in brackets)

The binding data in Table 7 show that extraordinary selectivity was observed for the terminal acylating groups. For example, in 44 of the 50 deep red beads decoded, the terminal R was composed of three non- hydrogen atoms (31 for MeOCH ₂ and 13 Me ₂ N) . High selectivity was observed for the AA ₃ position. In 43 of the 50 deep red beads decoded, the residue in AA ₃ was composed of D-amino acids with an amide group in the side chain (31 for Gin, 7 for Asn and 5 for Lys) . Th-.--* data indicate that receptor 11,* _* * discriminates between substrates most effectively when structural differences occur near the free end of the substrate chain. Thus the number of accepted residues is minimal in the case of the terminal acylating group (R) and increases with distance from the terminus. Such high selectivity for the N- terminal group and the AA ₃ residue is similar to a known peptide receptor and suggest that the binding mode of 11a* is similar to that of the C ₃ symmetric receptors disclosed herein. ⁷¹ * It appears that the hydrogen bond donor/acceptors on the rim of 11a* bind the substrate by hydrogen bonds, and there are nonbonded interactions between the benzene-lined hydrophobic region of the receptor and the N-terminal group of the peptide substrate. Receptor 12* does not bind any peptide substrates. This suggests that the conformational homogeneity of 11a* is crucial for binding with peptide substrates.

A variety of peptides were synthesized and binding constants for association with 11a* measured in CHC1 ₃ using the solid phase binding methods as described supra. The results of these binding measurements are summarized below in Table 8. These measurements show that stereochemical inversion at Gin reduces binding energy by 1.3 kcal/mol. Changing the N-terminal group from Me ₂ N to iPr and cycPr reduce binding energy by 1.7 and 1.2 kcal/mol, respectively.

Table 8. Binding of 11a* with peptides in CHCl,

Peptide Substrate Me0CH ₂ -(D)Gln-Gly-(DPro-Polymer Me ₂ N-(D)Gln-Gly-( )Pro-Polymer iPr-(D)Gln-G_.y-(L.Pro-Polymer cycPr-(D)Gln-Gly-( )Pro-Polymer MeOCH ₂ -( )Gln-Gly-( )Pro-Polymer

While receptor 12* does not show any binding, highly selective complexation of 11a* was also found by a related binding assay using the side-chain deprotected

substrate library. In that case, lia* showed different and even higher substrate selectivity as shown in Table 9.

Table 9. Residues found in sidechain deprotected substrates bound by 11a* and Frequencies of occurrence of each residue (in brackets)

The most commonly found side-chain deprotected sequences differed from the MeOCH ₂ -(D)Gln-AA ₂ -AA ₁ found with the protected library and had R-(D)Gln-(D)AA ₂ -(L)Pro (40%) and R-(D)AA ₃ -(L)Gln-(D)Pro (40%). Although the exact binding mode is unknown, it is likely that the sidechain of Gin interacts with binding cavity of 11a* and Pro at AA, restricts conformational flexibility of peptide substrate to be preorganized for binding with the receptor. This suggests that 11a* recognizes the total structural features of the polypeptide substrates, not just partial structure. It is interesting that the stereochemistry and the site selectivity of Gin in the polypeptide substrates bound to 11a* depend on the stereochemistry of Pro at AA,, as shown in Table 11 (L-proline at AA, appeared with D-glutamine at AA ₃ and D-proline at AA, appeared with L-glutamine at AA ₂ ) .

11a* is a readily accessible heterooligomeric assembly from trimesic acid (A) and diamine (B) linked through a tris(aminomethyl)benzene. The results described here not

only show that 11a* is a highly selective receptor for peptide substrates but also demonstrate the power of directed screening of large chemical libraries as a method to find novel molecules having sought-after properties. The remarkable differences in peptide substrate binding properties between lla* and 12* suggest that conformational homogeneity may be a key to the designing the highly selective receptors .

Table 10.

1.M0M

2.MOM

3.MOM

4.MOM

5.MOM

6.MOM

7.MOM

8.MOM

9.MOM 10. OM 11.MOM 12.MOM 13.MOM 14.MOM 15.MOM 16.MOM 17. OM 18. OM 19.MOM 20.MOM 21.MOM 22.MOM 23.MOM 24.MOM 25.MOM 26.MOM 27. OM

28.MOM

29.MOM 30.MOM 31.MOM 32.Me ₂ N 33.Me ₂ N 34.Me ₂ N 35.Me ₂ N 36.Me ₂ N 37.Me,N

Table 11. Sequences found in sidechain deprotected substrates bound by lla*. tSlassay" ²⁵ - ³ t«M l .Me ₂ N D-Gln Gly L-Pro

2.Me N D-Gln Gly L-Pro 3.Me ₂ N D-Gln Gly L-Pro

4. MOM D-Lys Gly L-Pro

5.Me ₂ N D-Gln D-Lys L-Pro

6.Ph D-Gln D-Lys L-Pro

7.Ph D-Gln D-Lys L-Pro 8. Et D-Gln D-Lys L-Pro

9.Et D-Gln D-Lys L-Pro lO . cPr D-Gln D-Lys L-Pro

11. Mor D-Gln D-Lys L-Pro

12. cPen D-Gln D-Lys L-Pro 13 . MOM D-Gln L-Lys L-Pro

14 .AOM D-Gln D-Gln L-Pro

15. Me D-Gln D-Gln L-Pro

16. OM D-Gln L-Ala L-Pro

17. Ph D-Gln D-Ala L-Pro 18. Mor D-Gln D-Ala L-Pro

19. MOM D-Gln D-Ser L-Pro

20. MOM D-Lys L-Lys L-Pro

21 . MOM D-Lys L-Ala L-Pro

22. MOM L-Asn D-Gln L-Pro 23. MOM D-Asn L-Ser L-Pro

24 .Me ₂ N D-Val L-Gln D-Pro

25.Me ₂ N D-Val L-Gln D-Pro

26.Me ₂ N D-Val L-Gln D-Pro

27. Mor D-Val L-Gln D-Pro 28. Mor D-Val L-Gln D-Pro

29. Et D-Val L-Gln D-Pro

30. Ph D-Pro L-Gln D-Pro

32. tBu D-Pro L-Gln D-Pro

33.Me ₂ N D-Ala L-Gln D-Pro 34 . Ph D-Ala L-Gln D-Pro

35. MOM D-Gln L-Gln D-Pro

36. Ph D-Gln L-Gln D-Pro

37. or D-Gln L-Gln D-Pro 38.cPr D-Ser L-Gln D-Pro 39.Me N D-Ser L-Gln D-Pro 40. or D-Lys L-Gln D-Pro 41. OM D-Asn L-Gln D-Pro 42. OM L-Ser D-Asn D-Pro 43.cPr D-Gln D-Asn D-Pro 44.Me ₂ N D-Val L-Lys D-Pro 45.MOM D-Gln L-Pro D-Gln

Hexamethyl ester (11-1*)

To a solution of 0.42 g of trimesic acid dimethyl- mono(pentafluorophenylJester (1.04 mmol) and 86 mg of triaminomethyl benzene triHCl salt (0.312 mmol; F.L. Seitl, K.N. Raymond, J. Am. Chem. Soc , 1979, 101, 2728) in 10 ml of DMA was added 0.36 ml of DIPEA. After stirring overnight at r.t., all volatiles were removed at reduced pressure. The residue was purified by flash chromatography on silica gel using 5% MeOH in CH ₂ C1 ₂ to give an amorphous white solid (0.20 g, 77.7%).

R _f =0.55 (5% MeOH in CH ₂ C1 ₂ )

¹ H NMR (CDC1 ₃ ) <S 3.90 (s, 3H) 4.38 (m, 2H) 7.10 (s, IH) 7.30 (m, IH) 8.61 (s, 2H) 8.70 (s, IH)

¹³ C NMR (CDC1 ₃ ) δ 165.74, 165.36, 139.11, 134.99, 132.84,

132.18, 130.77, 126.35, 52.44, 43.71

IR (neat) 3324, 2815, 1730, 1674, 1573 cm ^"1

C3 symmetric macrocyclic receptor (11*)

A solution of 63 mg of hexakis(pentafluorophenyl) ester (0.0362 mmol) and 13 mg of diaminocyclohexane diTFA salt (B, 0.109 mmol) in 10 ml of DMA was added to a solution of 0.19 ml of DIPEA (1.09 mmol) in 200 ml of THF at r.t. for 20 hr by syringe pump. After stirring for additional 8 hr, all volatiles were removed at reduced pressure. The residue was purified by flash chromatography on

silica gel using 10% MeOH in CH ₂ C1 ₂ to give an amorphous white solid (12 mg, 31.0%) .

R _f =0.45 (5% MeOH in CH ₂ C1 ₂ ) ¹ H NMR (1/1=CD ₃ 0D/CDC1 ₃ ) δ 1.50 (m, 2H) 2.02 (m, IH) 2.28

(m, IH) 3.72 (m, IH) 4.12 (m, IH) 4.28 (m, IH) 5.02 (m, IH) 7.46 (s, IH) 8.22 (s, IH) 8.72 (s, IH) ¹³ C NMR (1/1 = CD ₃ 0D/CDC1 ₃ ) δ 170.24, 169.91, 139.05, 134.83, 131.27, 130.75, 130.66, 130.19, 125.92, 56.21, 38.21, 35.44, 35.24, 29.08, 29.49

IR (neat) 3324, 2826, 1674, 1573 cm ^"1 MS(FAB) m/z 976 (M+l)

Dye-labelled macrocyclic receptor (lla*)

A solution of 44 mg of hexakis(pentafluorophenyl) ester (0.0253 mmol) and 63 mg of N-succinyl dye - 3R,4R - pyrrolidine diamine diTFA salt (8-2, 0.0835 mmol) in 10 ml of DMA was added to a solution of 0.19 ml of DIPEA (1.09 mmol) in 200 ml of THF at r.t. for 20 hr by syringe pump. After stirring for additional 8 hr, all volatiles were removed at reduced pressure. The residue was purified by flash chromatography on silica gel using 10% MeOH in CH ₂ C1 ₂ to give an amorphous red solid (15 mg, 26.0%).

R _f =0.45 (10% MeoH in CH ₂ C1 ₂ )

¹ H NMR (1/1 = CD ₃ 0D/CDC1 ₃ ) δ 0.70 (t, 3H, J=7.0 Hz) 2.45 (m, 4H) 3.01 (q, 2H, J=7.0 Hz) 3.36 (s, br, 2H) 3.50 (s, br, 2H) 3.74 (m, IH) 3.92 ( , IH) 4.12 (m, 4H) 4.58 (m, IH) 4.68 (m, IH) 6.64 (d, 2H, J=9.1 Hz) 7.72 (m, 5H) 7.98 (m, IH) 8.10 (d, 2H, J-9.0 Hz) 8.23 (m, IH)

¹³ C NMR (1/1=CD ₃ 0D/CDC1 ₃ ) δ 172.82, 170.20, 169.72, 169.61, 167.70, 138.51, 136.06, 133.53, 131.27, 130.54, 130.46, 129.54, 126.72, 125.92, 125.22, 124.37, 124.24, 122.31, 118.85, 112.22, 61.34, 55.74, 50.54, 45.70, 44.42, 38.28, 28.76, 28.27, 27.52, 27.22, 11.94

IR (neat) 3324, 2815, 1720, 1675, 1573 cm ^"1

43. M. Bodanszky, A. Bodanszky, The Practice of Peptxde Synthesis ; Springer-Verlag, Berlin, 1984.

44. J.-I. Hong, Ph.D. Thesis, Columbia University, 1990.

45. J.M. Stewart, J.D. Young, Solid Phase Peptide Synthesis Pierce Chemical Company, Rockford, 1984.

50. A. Ricci, R. Danieli, S. Rossini, J. Chem. Soc Perkin I, 1976, 1691; . Offermann, F. Vogtle, Synthesis , 1977, 272.

VI

While laa (H.-J. Schneider, Angew. Chem. Int . Ed. , 32 848 (1993); T.H. Webb and C.S. Wilcox, Chem. Soc. Rev. , 383 (1993)) is available by an optimized synthesis (J.-I. Hong, S.K. Namgoong, A. Barnardi and W.C. Still, J. Am. Chem. Soc , 113, 5111 (1991); S.D. Erickson, J.A. Simon and W.C. Still, J. Org. Chem. , 58, 1305 (1993); R. Liu and W.C. Still, Tetrahedron Lett. , 34, 2573 (1993); A. Borchardt and W.C. Still, J. Am. Chem. Soc , 116, 373 (1994)), related receptors were sought which can be prepared by simpler methods and thus might find practical application to problems in separation science. In this regard, receptor 2aa and its relatives are attractive candidates. laa and 2aa are closely related in that their cup-shaped binding cavities have both similar dimensions and analogous patterns of unassociated hydrogen bond donors and acceptors on their peripheries. Described herein is a simple synthesis of 2aa and its binding properties as revealed using an encoded combinatorial library of -50,000 acylated tripeptide substrates. 2aa is a selective receptor for peptides.

The synthesis of 2aa is summarized below and involves coupling three equivalents of pentafluorophenyl dimethyl trimesate to 1,3,5-tri(aminomethyl)benzene to provide 3aa (78% yield) . Saponification of all six esters (NaOH, H ₂ 0, MeOH) and activation as the hexakis(pentafluorophenyl) ester (C ₆ F ₅ OH, EDC) then gave 3bb (30% yield) . This material underwent triple coupling with the dye (Disperse Red 1)-acylated, (R,R)-diaminopyrrolidine ((R,R)- Dia inopyrrolidine preparation: D.R. Reddy and E.R. Thornton, J. Chem. Soc Chem. Common . , 172 (1992) ; S.S. Yoon and W.C. Still, Tetrahedron , 0000 (1994)) shown to simultaneously close three 21-membered rings and provide the brilliant red-colored receptor 2aa in 26% yield. A

variety of related receptors can be prepared with other 1,2-diamines. For example, commercially available (2J?,2J?)-l,2-diaminocyclohexane was used with 3bb to prepare an analogous receptor (31% yield for the final tricyclization step) having the pyrrolidine rings replaced by simple cyclohexanes.

Synthesis of Receptor 2aa.

To establish the binding properties of receptor 2aa, the method of treating a dye-labeled receptor with a large collection or library of potential substrates was used. (A. Borchardt and W.C. Still, J. Am. Chem. Soc , 116, 373 (1994)) The substrate library was synthesized on 50-80 μ Merrifield polystyrene beads in such a way that each bead carried a single library member. The binding experiment simply involved mixing red 2aa with the bead- supported library and selecting the beads which turned red. Thus binding of 2aa to the entire library (here -50,000 acylated tripeptides) could be surveyed in a single experiment.

The polymer-supported substrate library used has been described (A. Borchardt and W.C. Still, J. Am. Chem. Soc , 116, 373 (1994)) and has the general structure:

RCO-AA3-AA2-AA1-NH(CH ₂ ) ₅ CONHCH ₂ -polystyrene

where AAn represents any one of the following fifteen side-chain-protected [ (N-trityl)Asn, (N-trityl)Gln, (N- Boc)Lys, (O-tBu)Ser] amino acids:

Gly, D-Ala, L-Ala, D-Ser, L-Ser, D-Val, L-Val, D-Pro, L- Pro, D-Asn, D-Gln, L-Gln, D-Lys, L-Lys

and where R represents any one of the following fifteen groups:

methyl (Me), ethyl (Et) , i-propyl (iPr) , t-butyl (tBu) , i-butyl (iBu) , neopentyl (neoPe) , trifluoro-methyl (TFM) , methoxymethyl (MOM) , cyclopropyl (cPr) , cyclobutyl (cBu) , cyclopentyl (cPe) , acetoxymethyl (AcOM) , phenyl (Ph) , dimethylamino (Me ₂ N) , morpholino (Mor)

The library was prepared using split synthesis . (A. Furka, M. Sebestyen, M. Asgedom and G. Dibo, Abstr. 14th

Jnt. Congr. Biochem. , Prague, Czechoslovakia, 5, 47 (1988) ; A. Furka, M. Sebestyen, M. Asgedom and G. Dibo, Abstr. 10th Int. Symp. Med. Chem. , Budapest, Hungary, 288 (1988) ; A. Furka, M. Sebestyen, M. Asgedom and G. Dibo, Int. J. Pept. Protein Res. , 37, 487 (1991) ; K.S. Lam, S.E. Salmon, E.M. Hersh, V.J. Hruby, W.M. Kazmierski and R.J. Knapp, Nature , 354, 82 (1991)) Thus the total possible number of different substrates in the library was 15 ⁴ (50,625). To allow for determination of the structure of the substrate on a bead that bound 2aa, the binary encoding method was used which employs electron capture gas chromatography (ECGC) to analyze tag-encoded structural information. (M.H.J. Ohlmeyer, R.N. Swanson, L.W. Dillard, J.C. Reader, G. Asouline, R. Kobayashi, M. Wigler and W.C. Still, Proc Natl . Acad. Sex . USA, 90, 10922 (1993)).

The binding assay was carried out by equilibrating a 10 mg sample (~10 ⁵ beads) of the above peptide library with -0.3 mL of 50 μM 2aa in CHC1 ₃ . After 24 hrs of agitation, -1% of the beads had developed deep red-orange coloration. These beads carried peptides that bound red 2aa most tightly. (Previous studies have shown the color assay readily distinguishes substrates differing in binding energy by as little as 1.0 kcal/mol. Control

experiments established that bead coloration reflects binding of the peptide, not the dye: neither the dye nor the succinylated dye stained nay of the library beads under the assay conditions.) By measuring [2aa] in solution over the beads at equilibrium, we established that the assay was selecting beads having minimum association constants (K _a . _mι - _n ) of -3000. Fifty of these most deeply colored beads were picked under a 4X microscope and their tag-encoded structural information was decoded by ECGC. The residue types found in the peptides on these beads are listed in Table la.

Table la. Residues in acylated, protected peptide substrates selectively bound by 2aa (number of occurrences from 50 decoded beads found given in parentheses) .

In several ways, the binding selectivities of 2aa are similar to those reported for laa. Thus, both receptors show the highest selectivity for residues (R,AA3) at the terminus of the peptide chain. They also selectively bind certain R groups having three nonhydrogen atoms and AA3 = (N-trityl)Gln. On the other hand, laa binds AA3=L- (N-Trityl)Gln whereas 2aa binds D-(N-trityl)Gln. This difference in enantioselection is rationalized by the enantiomeric patterns of substrate-binding hydrogen bonding groups (+=donor, -=acceptor) found at the periphery of the binding cavity.

Other selectivity differences for laa and 2aa are found at AA2 (laa prefers L-Pro, 2aa prefers Gly) and AA1 (laa has virtually no selectivity, 2aa prefers D-Pro, L-Pro or D-Ala) residues. The high preference of laa for R=cyclopropyl is absent with 2aa. Neither laa nor 2aa bind peptides having R=isopropyl. These and other selectivities were verified by independent binding energy measurements with 2aa that showed, inter alia , that changing R=MOM to cyclopropyl or AA3=D-(N-trityl)Gln to L-(n-trityl)Gln diminished binding by laa.3 kcal/mol each

relative to the preferred substrate MOM-CO-D-(N- trityl)Gln-Gly-L-Pro (ΔG _binding =-4.3 kcal/mol).

An analogous survey of 2aa with a substrate library having deprotected peptide side-chains is shown in Table 2a. Measurement of [2aa] _u indicated that the color assay in that case allowed selection of substrates having K _a a- _m m,i- _n n~5000.

Table 2a. Residues in acylated, deprotected peptide substrates selectively bound by 2aa (number of occurrences from 45 decoded beads given in parentheses).

MOM (12) D-Gln (23 ) L-Gln (18)L-Pro (23 ) Me ₂ N (10) D-Val ( 7 ) D-Lys (8)D-Pro (21 ) Ph (6) D-Lys (4 ) L-Lys ( 3 )D-Gln (1 ) Mor (6) D-Pro (3 ) Gly (4 ) Et (4) D-Ala (2) D-Gln ( 3 ) cPr (3) D-Ser (2) D-Asn (2) Me (1) D-Asn (2) D-Ala (2) cPen (1) L-Asn (1 ) L-Ala (2) tBu (1) L-Ser ( 1 ) L-Pro (1 ) AcOM (1) D-Ser ( 1 )

L-Ser ( 1 )

Whereas previous studies found laa to have similar selectivities for protected and deprotected peptide substrates, receptor 2aa bound a different set of peptides with the deprotected library. It was most discriminating at the internal AA1 site where Pro was strongly preferred. In particular, 2aa preferentially bound the two partial sequences XCO-(D)Gln-(D)X-(L)Pro (40% of beads) and XCO-X-(L)Gln-(D)Pro (40% of beads) [X indicates no significant with both the configuration and position of the downstream Gin is particularly interesting. This novel Gin...Pro selectivity may reflect hydrogen boning between the Gin sidechain and amides in the bottom of 2aa's binding cavity - a possibility not available to laa. The ability of 2aa to selectively bind peptide spans having as many as three residues is remarkable for such a small host molecule.

These studies and others suggest that a variety of receptors similar in structure to laa or 2aa have significantly sequence-selective peptide binding properties and that the particular selectivity of such a receptor can be controlled by varying the details of its structure. While the full range of variations that can be tolerated without losing binding has yet to be established, the full macrotricyclic framework of 2aa is

necessary for peptide binding - deletion of the central aromatic linker obliterates tight peptide binding as detected by solid phase color assay. Given the simplicity of 2aa's synthesis, its highly selective peptide-binding properties, and the availability of many building blocks with which to construct analogs, it is likely that a variety of C ₃ -symmetric receptors having a range of peptide sequence selectivities can be prepared along the lines described herein.

VII

When designing the A ₄ B ₆ receptor (Fig. 8(a)), the conformationally rigid building blocks A and B were chosen to minimize its flexibility. The present inventors also describe herein the synthesis and properties of two related A ₄ B ₆ cyclooligomers which are constructed from more conformationally flexible acyclic diamines Bl and B2. Binding properties in this series of receptors are sensitive to structure of the components used to assemble them, but rigid cyclic building blocks need not be used to obtain high binding selectivity.

To prepare these receptors, the simple one-step coupling of the amines was carried out with the triacid chloride A as described for A ₄ B ₆ . (S.S. Yoon and W.C. Still, J. Am. Chem. Soc , 115, 823 (1993)).

With Bl, the A ₄ B1 ₆ receptor was obtained in 10% yield when the coupling was carried out at a concentration corresponding to 6 mM in receptor. (EXAMPLE 23) With B2, it was necessary to use a more dilute 1 mM concentration to prepare A ₄ B2 ₆ in 7% yield. Both products were readily isolated as the most mobile reaction product on silica gel and were identified by mass spectroscopy and by their symmetry as revealed by ¹³ C and ¹ H NMR.

Binding energies were measured by titrating 0.5 mM solutions of receptor in CDC1 ₃ with various N-acetyl amino acid methylamides and monitoring the receptor protons by 400 MHz (H-C) and amide (H-N) protons. The binding energies found are given in Table IA and all represent averages of at least two different binding measurements. Scatchard treatment of binding data indicated 1:1 complexes in all cases.

Table IA. Peptide-Binding Properties of Macrotricyclic Receptors in CDC1 ₃ .

Peptide Substrate* -AG ^D AAG" -AG ^C AAG ^C -AG ^b AAG ^C

GLY 1.9° 1.4 1.5 1.7(89Xee)

L-ALA 3.5° 1.3 (80Xee) 4.1 1.8(90Xee) 3.7 1.7(89Zee)

D-ALA 2.2° 2.3 2.0

L-VAL 5.0° 2.6 ^d (98Xee) 4.5 2.4(96Xee) 3.8 1.5(84Zee)

D-VAL 2.4° 2.1 2.3

L-ILE 4.3 ^d 1.9 ^d (92Xee) 4.2 2.2(95 ee) 2.6 0.4(32Xee) D-ILE 2.4 ^d 2.0 2.2

L-LEU 3.4° 1.0°(68%ee) 3.6 1.5(84%ee) 2.5 0.3(24%ee) D-LEU 2.4° 2.1 2.2

L-PHE NC° NC NC D-PHE 2.0° 1.5 1.4

L-Phenylglycine 5.9° 3.0 ^α (>99Zee) 5.7 3.9(>99Xee) 3.4 1.6(87Xee) D-Phenylglycine 2.9° 1.8 1.8

L-Ethylglycine 5.7 3.3(>99Xee) 5.5 3.4(>99Xee) 5.5 3.3(>99Xee) D-Ethylglycine 2.4 2.1 2.2

L-Propylglycine 6.0 3.5(>99Xee) 5.7 3.5(>99Xee) 4.8 2.5(97Xee) D-Propylglycine 2.5 2.2 2.3

L-Butylglycine 3.9 1.4(82Xee) 3.8 1.6(87Zee) 2.5 0.2(16JCee)

a. All peptides are N-acetyl, methylamides; n. binding energy (kcal/mol; c. enantioselectivity (kcal/mol); d. from reference 1. NC = no complexation observed.

The binding results obtained with all three receptors support the general picture of the complex suggested by Yoon and Still.

The binding data in Table IA reveal a number of trends. First, all receptors bind all D amino acid substrates with roughly the same binding energy (2.0-2.5 kcal/mol). Thus the high enantioselectivities observed likely originate from especially favorable binding to L peptide substrates, not by destabilization of binding to D substrates. Second, both the original A ₄ B ₆ receptor and the A ₄ B1 ₆ binds six of the eight substrates studied with higher enantioselectivity than does A ₄ B ₆ .

Both A ₄ B ₆ and A ₄ B1 ₆ show surprisingly high selectivity among L amino acids which are distinguished only by the size and shape of their unfunctionalized, hydrocarbon sidechains. Amino acids having branched sidechains bind well only when the branch occurs at the substrate β- carbon. Thus valine and isoleucine (R=i-Pr, s-Bu) bind well but leucine (R=Et, n-Pr) are among the best substrates. All three receptors distinguish phenylglycine and phenylalanine by >3 kcal/mol. These observations are compatible with a conical-cavity model which favors substrates having more steric bulk near the enlarged, open end of the binding cavity. Substrates with side-chains that are either too small to fill the cavity or too long to be accommodated are poorly bound. While binding selectivity based on steric effects is known (for example, see F. Diederich, K. Dick and D. Griebel, J. Am. Chem. Soc , 108, 2273 (1986); W.L. Mock and N.-Y. Shih, J. Am. Chem. Soc , 110, 4706 (1988); M.A. Petti, T.J. Shepodd, R.e. Barrans and D.A. Dougherty, J. Am. Chem. Soc , 110, 6825 (1988); D.J. Cram, M.E. Tanner, S.J. Keipert and C.B. Knobler, J. Am/ Chem. Soc , 113, 8909 (1991) ; K. Naemura, K. Ueno, S. Takeuchi, Y. Tobe, T. Kaneda and Y. Sakata, J. Am. Chem. Soc. , 115 8475

(1993); L. Garel, B. Lozach, J.-P. Dutasta and A. Collet, J. Am. Chem. Soc , 115, 11652 (1993)), the subtle differences in side-chain bulk which the disclosed receptors are able to distinguish energetically by 1-2 kcal/mol is unusual with synthetic receptors. The key to such high steric selectivity appears to coincide with the receptor's ability to fully encapsulate the chemical substructure being distinguished.

Like A ₄ B ₆ and A ₄ B1 ₆ which bind L-peptides based on the steric requirements of their sidechains, receptor A ₄ B2 ₆ also distinguishes peptide sidechains sterically but with different selectivity. In particular, A ₄ B2 ₆ selects for L-peptides whose sidechains are small and compact. Thus, alanine, valine and ethylglycine are well-bound while isoleucine, leucine, phenylglycine, propylglycine and butylglycine are more weakly bound relative to the other receptors. ^A ₄ B2 ₆ appears to have a smaller binding cavity, a property which may follow from cavity occupancy by benzyloxymethyl substituents or from partial cavity collapse due to the flexible nature of the B2 fragment.

These findings suggest that the highly selective binding found with the original A ₄ B ₆ receptor may be general to cyclooigomeric molecules of this class and that binding selectivity can be altered by starting with different A and B fragments. Thus it should now be possible to prepare a wide range of interesting receptors by similar routes. It may also be noted that these receptors incorporate B fragments in two different structural environments: the upper and lower macrocycles include four equivalent B's while two other B'ε link those macrocycles together. By varying these distinct B fragments independently, even more receptor diversity can be generated.

VIII

While biological receptors manage to distinguish substrates by differences in their amino acid sequences, analogous selectivities are difficult to attain with synthetic receptors. While significant strides have been made in selectively binding derivatives of single amino acids (S.C. Peacock, L.A. Domeier, F.C.A. Gaeta, R.C. Helgeson, J.M. Timko and D.J. Cram, J. Am. Chem. Soc , 100, 8190 (1978); J. Rebek, B. Askew, P. Ballester and M. Doa, J. Am. Chem. Soc , 109, 4119 (1987); J.-I. Hong, S.K. Namgoong, A. Bernardi and W.C. Still, J. Am. Chem. Soc , 113, 5111 (1991); Y. Murakami, T. Ohno, 0. Hayashida and Y. Hisaeda, J. Chem. Soc , Chem. Commun . , 950 (1991) : A. Galan, D. Andreu, A.M. Echavarren, P. Prados and J. de Mendoza, J. Am. Chem. Soc , 114, 1511 (1992); R. Liu and W.C. still, Tetrahedron Lett . , 34, 2573 (1993); S.S. Yoon and W.C. Still, J. Am. Chem. Soc , 115, 832 (1993); S.S. Yoon and W.C. Still, Tetrahedron Lett. , 0000 (1994)), synthetic receptors which bind particular multiple amino acid sequences are unknown.

One reason that sequence-selective binding of peptide substrates is underdeveloped in host-guest chemistry is that peptides are physically large molecules and the receptors which would bind them would need binding sites of comparable dimensions. Unfortunately, most contemporary host molecules have rather small substrate- binding regions and thus cannot easily distinguish substrates having differentiating structural features which are distributed over a large area. It is not easy to formulate molecular structures that are both readily synthesized and have large, well-structured binding sites that are complementary to a given substrate. (Notable successes include: W.L. Mock and N.-Y. Shih, J. Am. Chem. Soc , 110 4706 (1988); M.A. Petti, T.J. Shepodd, R.E. Barrans and D.A. Dougherty, J. Am. Chem. Soc , 110, 6825

(1988); J.K. Judice and D.J. Cram, J. Am. Chem. Soc , 113, 2790 (1991); J.-I. Hong, S.K. Namgoong, A. Bernardi and W.C. still, J. Am. Chem. Soc , 113, 2790 (1991); J.- I. Hong, S.K. Namgoong, A. Bernardi and W.C. Still, J. Am. Chem. Soc , 113, 5111 (1991); T.H. Webb, H. Suh and C.S. Wilcox, J. Am. Chem. Soc , 113, 8554 (1991); M.E. Tanner, C.B. Knobler and D.J. Cram, J. Org. Chem. , 57, 40 (1992) .)

While the general problem of receptor design remains unsolved, several receptors are known which bind certain biological molecules with high selectivity. (E. g. , K.-S. Jeong, A.V. Meehldorf and J.E. Rebek, J. Am. Chem. Soc , 112, 6144 (1990); R.P. Dixon, S.J. Geib, A.D. Hamilton, J. Am. Chem. Soc , 114, 365 (1992).) Among these is receptor 1' (Fig. 8(b)) (S.S. Yoon and W.C. Still, J. Am. Chem. Soc , 115, 832 (1993); S.S. Yoon and W.C. Still, Tetrahedron Lett . , 0000 (1994).) This molecule is a cyclooligomer of trimesic acid and ( 1R, 2R) ~ diaminocyclohexane, and can be synthesized in one step from commercially available materials. Its most interesting property, however, is that it binds certain α-amino acid derivatives with high selectivity. In particular, 1' was found to bind L-amino acids enantioselectively (70-99% ee) and to select for amino acid side-chains having a particular size (e.g. phenyl>>benzyl, ethyl»methyl) . These properties were rationalized by a three-strand, ^-sheet-like binding mode involving peptide association with the circled hydrogen bond donors/acceptors and having an L-peptide sidechain buried within the center of the receptor. The present inventors have observed very strong binding of 1' to a certain tripeptide (N-Boc-Gly-(L)Val-gly-NHBn) which presumably involved additional hydrogen bonding to the outlying amides which are part of the B fragments of 1'. While these findings suggested the possibility of sequence-selective peptide binding, studies using 1' and

variouε alanine- and valine-containing tripeptides failed to detect it. In conjunction with other receptor work, Borchardt and Still described a new method for studying the scope of receptor binding using an encoded combinatorial library of substrates. (A. Borchardt and W.C. Still, J. Am. Chem. Soc , 116, 373 (1991).) Because such a library-based assay enables many thousands of binding experiments to be performed simultaneously, it is ideal for discovering binding phenomena about which little is known. The method to search for sequence- selective peptide binding by 2', a dye-labeled analog of 1' is used herein.

The approach to evaluating the binding properties of a synthetic receptor is closely related to methods developed for finding good ligands to biological receptors such as antibodies. (H.M. Geysen and T.J. Mason, Bioorg. Med. Chem. Lett. , 3, 397 (1993)) The general scheme involves labeling a receptor (e. g. with a fluorescent dye or radioisotope) so that it may be sensitively detected and then treating the labeled receptor with a large collection of potential substrates. If these substrates are spatially separated (e.g. on different solid particles or in different location on a plate) , then those areas occupied by substrates which bind the receptor will themselves become labeled. The appeal of the approach is its simplicity and the fact that very large numbers of different substrates can be screened for binding simultaneously. Even such simple procedures as examining substrate-bearing particles (e.g. Merrifield synthesis beads) through a low power microscope for the presence of label allows millions of substrates to be screened for binding by one person in a few hours.

By varying the concentration of the labeled receptor used, it is possible to control the minimum association

constant (D _a . _min ) which can be detected by the binding assay. Assuming simple bimolecular receptor/substrate binding, K _a is traditionally defined:

„ [Complex] [Receptor] _free [Substrate] _free

When [Complex]=[Substrate] _free , K _a simplifies to 1/[Receptor] _free . In the context of a solid phase assay using labeled receptor in solution and substrates on solid particles, this situation obtains when the substrate on a particular particle is half bound by labeled receptor. Hence, one may estimate the minimum K _a (here termed _a . _mιn )of receptor for substrates on any fully labeled particles by measuring the equilibrium concentration of the labeled receptor in free solution over the particles and taking its reciprocal.

The particular solid phase assay used here generally involves labeling the receptor with a colored dye and mixing it in dilute solution with a library of peptide- like substrates attached to Merrifield synthesis beads. After 48 hours of agitation to equilibrate a dilute solution of the colored receptor and the initially colorless substrate bead library, a small percentage of the beads take on deep colorations. Then the con¬ centration of free receptor remaining in solution is measured to determine K _a . _mjn , and then those beads having the deepest coloration are picked. By determining the structures of the substrates on those beads, one learns which substrates in the library bind the receptor with association constants of at least K _a _ _m * _n assuming the substrates on the deeply colored beads are at least 50% saturated by labeled receptor.

While the receptor labeling and assay procedures are straightforward, creation of an appropriate substrate library is less obvious. A variety of methods for creating large collections of diverse molecules have been developed and could be used (Reviews: G. Jung and A.G. Beck-Sickinger, Angew. Chem . Int . Ed. Engl . , 31, 367 (1992) ; M.R. Pavia, T.K. Sawyer and W.H. Moos, Bioorg. Med. Chem. Lett . , 3, 387 (1993)), but one method, combinatorial split synthesis (A. Furka, M. Sebestyen, M. Asgedom and G. Kibo, AJbstr. 14th Int . Congr. Biochem. , Prague, Czechchoslovakia, 5, 47 (1988); A. Furka, M. Sebestyen, M. Asgedom and G. Dibo, Abεtr 10th Int. Symp. Med Chem. Budapest, Hungary, p 288 (1988) ; A. Furka, M. Sebestyen, M. Asgedom and G. Dibo, Int . J. Pept . Protein Res . , 37 487 (1991); K.S. Lam, S.E. Salmon, E.M. Hersh, V.J. Hruby, W.M. Kazmierski and R.J. Knapp, Nature , 354, 82 (1991)), is particularly relevant to substrate library preparation. Given that organic synthesis proceeds via a series of steps and that many of these steps can often be carried out using a range of alternative reagents (or synthons, residues, etc.) to yield different products, split synthesis prescribes a simple protocol for preparing the library of products resulting from all possible combinations of all alternative reagents used. Split synthesis is carried out on small solid support particles (e.g. Merrifield beads) and yields a particle- supported library in which any particular particle carries the product from one particular set of reagents.

The split synthesis method was developed originally for oligopeptide synthesis and can yield very large libraries. For example, if a pentapeptide library is prepared by split synthesis in which 20 different amino acids are used at each of the five residue-coupling steps, the final library will contain 3,200,000 (20 ⁵ ) different pentapeptides. Furthermore, any particular synthesis particle will bear only one type of

pentapeptide (or at least have been submitted to only one particular, well-defined series of chemical steps) .

Split synthesis has been effectively limited to libraries of sequencable biopolymers such as peptides and nucleotides. The limitation arose because individual members of the library are produced on single solid support particles an these particles carry only picomolar quantities of product. Such quantities are far too small for most classical structure elucidation techniques.

While mass spectroscopy offers a partial solution to the product analysis problem, it is limited by its inability to distinguish isomers or to deal with the mixtures of products and byproduct which commonly result from solid phase synthesis.

The best general solution to the structure analysis problem involves a recently developed technique known as encoding. (S. Brenner and R.A. Lerner, Proc Natl . Acad. Sci. USA, 89, 5381 (1992); J.M. Kerr, S.C. Banville and R.N. Zuckermann. J. Am. Chem. Soc , 115, 2529 (1993) ; J. Nielson, S. Brenner and K.D. Janda, J. Am. Chem. Soc , 115, 9812 (1993); M.C. Needels, D.G. Jones, E.H. Tate, G.L. Heinkel, L.M. Kochersperger, W.J. Dower, R.W. Barrett and M.A. Gallop, Proc. Natl . Acad. Sci . USA, 90, 10700 (1993) ; M.H.J. Ohlmeyer, R.N. Swanson, L.W. Dillard, J.C Reader, G. Asouline, R. Kobayashi, M. Wigler and W.C. Still, Proc Natl . Acad. Sci . USA, 90, 10922 (1993).) Encoding entails attaching arrays of molecular tags to the solid support particles during each synthetic step to create unique, tag-encoded records of the particular reagents used in the synthesis of each library member. By analyzing the tag complement of any particular solid support particle after the synthesis, one can determine the particular synthetic steps which were used in the synthesis of the library member on that particle.

In the study of labeled receptor 2 > , a substrate library of 50,625 (15 ⁴ ) terminally acylated tripeptides is used and is prepared by split synthesis on 50-80μ polystyrene (Merrifield) beads as described above. The library is encoded using a set of sixteen highly electrophoric tagging molecules which can be detached from single synthesis beads and analyzed using electron capture capillary gas chromatography (ECGC) .

A labeled variant of receptor l' (Figure 8(b)) was made that could be visua? y detected by simple inspection. Since 1' itself has nc appropriate label attachment site, a relative of 1' in which the two spanning tranε- diaminocyclohexaneε (B in 1') were replaced by stereochemically similar trans -3 ,4-diaminopyrrolidineε was made. The pyrrolidine ring nitrogen could then serve as the label attachment point. For the label, an intense red dye, Disperse Red 1, was selected, and the final structure of the labeled receptor thus became 2'.

The synthesis of red-labeled receptor 2' is straight¬ forward and is outlined in Figure 8(c) (EXAMPLE 25). Because 2' is not significantly strained and has few rotatable bonds, it could be constructed by a few simple reactions which both couple fragments and macrocyclize in a single step. Thus, the macrocyclic tetramide 3' was prepared in >50% yield by a single reaction which linked and cyclized two molecules of trans-1,2- diaminocyclohexane and two molecules of diactivated trimesic ester. The synthesis was completed via diactivation of 3' as a Jis-pentafluorophenyl ester (4') followed by another single step coupling/macrocyclization (-30% yield) using labeled diamine 5' to give the desired dye-labeled receptor 2' .

The polymer-supported, encoded substrate library has been described and has the general structure:

Polystyrene-CH ₂ NHCO(CH ₂ ) ₅ NH-AA1-AA2-AA3-C(=0)R

where AAn represents any one of the following fifteen sidechain-protected (side-chain protection: Asn (trityl) , Lys (Boc) , Ser (tBu) ) amino acids (standard single letter codes for amino acids in parentheses) :

Gly(G), D-Ala(a), L-Ala(A), D-Ser(S), D-Val(v), L-Vat(V), D-Pro(p), L-Pro(P), D-Asn(n), L- Asn(N), D-Gln(q), L-Gln(Q), D-Lys( ), L-Lys(K)

and where R represents any one of the following fifteen groups:

methyl(Me), ethyl(Et), i-propyl(iPr) , t-butyl (tBu), i- butyl(iBu), neopentyl (neoPe) , trifluoromethyl (TFM) , methoxymethyl (MOM) , cyclopropyl (cPr) , cyclobutyl (cBu) , cyclopentyl (cPe) , acetoxymethyl (AcOM) , phenyl (Ph)m dimethylamino (Me ₂ N) , morpholino (Mor)

Because the substrate library was made by split synthesis using fifteen different amino acids at each of the three AA sites and terminated the tripeptide chain with fifteen different acylating agents, the total number of different substrates in the library is 15 ⁴ or 50,625. To encode this library, sixteen GC-distinct electrophoric tagging molecules were used.

To carry out the binding assay, a 10 mg sample (-10 ⁵ beads) of the substrate library was suspended in an Eppendorf tube containing -0.3 L CHC1 ₃ to which was then added -25 μL of a 600 μM CHC1 ₃ solution of red 2 ' . Upon mixing, the bead library immediately extracted most of the colored receptor from solution, but examination of the mixture through a low power microscope showed that all beads looked essentially the same. However, after 30

min of agitation on a wrist-action shaker, -5% of the beads had turned light orange. After 48 hours of agitation, -1% of the beads were stained deep red-orange along with more having various lighter orange colorations. At this point, a UV/VIS measurement of receptor remaining in solution revealed [2'] = 21 μM and thus K ^8" " ¹ '" would be -48000 for receptor-saturated beads. To estimate the degree of saturation for our most deeply stained beads, their colors were compared with those of standardized beads having 1%, 10% and 100% dye loading. This comparison suggested the most deeply stained beads to be -5% saturated. It was difficult to estimate the extent of saturation to better than a factor of 2-4. Assuming 5% saturation, the actual κ ^a"miπ would be -2400 which corresponds to a minimum binding energy for our red beads of -4.5 kcal/mol. Fifty-five of these most deeply stained beads were picked manually under a 4X microscope for decoding.

After washing the selected beads extensively with DMF, they were placed individually in separate melting point capillary tubes containing 1-2 μL of pure DMF for tag detachment. Because the tags were attached by photolabile ortho-nitrobenzylic carbonate linkages, they could be released for ECGC analysis by long wavelength ultraviolet (350 nm) irradiation. After silylation (biε- trimethylsilylacetamide) to increase tag volatility, the solution over each bead was injected into an electron capture capillary gas chromatograph for tag analysis.

The resulting chromatograms showed which tags were present and which were absent and thus revealed ^10b the structure of each substrate which had been selected by the solid phase binding assay.

Using the bead-supported peptide library and dye-labeled receptor 2 ' , the above-described solid phase assay and

ECGC decoding of 55 of the most intensely stained beads revealed the substrates whose sequences are listed in Table IB. Among these sequences, 53 unique sequences were found indicating that the 55 beads we picked included only a small fraction of all tightly binding sequences. Nevertheless, the data in Table 1 shows a number of clear trends.

Table IB. Peptide substrate sequences found to bind 2 *

The most obvious trend is a strong preference of 2' for L-valine (v) . Thus, 46 of the 52 sequences found (entries 1-46) contain V somewhere in the sequence and most commonly at the central AA2 site. This finding is in line with previous NMR studies of 1' which revealed the tightest binding for peptides containing V (among those amino acids used in the library) . All non-valine- containing substrates (entries 47-52) incorporate at least one glycine (G) . By comparing the observed frequencies of the various amino acids at the three different sites with their statistical expectation frequencies and standard deviations (σ) , which residue- sit selectivities are statistically significant can be discerned. Summarized below are the significant selectivities (>90% confidence level) at all substrate sites along with their deviations (units of σ) and the

corresponding confidence levels (%) as derived from counting statistics.

While these results indicate that 2 ' has significant selectivity for certain residues at each amino acid position, 2' has even more selectivity for particular amino acids relative to the L-valine (V) position. In particular, in 43 (93%) of the 46 V-containing sequences, the amino acid either immediately preceding or immediately following the V is D-(trityl)asparagine (n, 65% of sequences) or D-(N-trityl)glutamine (q, 32% of sequences). These two residues (n,q) are closely related in that they are the only D amino acids in the library having (N-trityl)carboxamide-bearing sidechains. In 36 (84%) of these 43 n/q-V sequences, n or q precedes V. In such n/q-V sequences, the site following V (i.e. AA1) also shows significant selectivity and is most commonly L-(O-tBu)serine (S) or glycine (G) . There is essentially no selectivity among the various N-terminal R groups.

These results indicate that 2' preferentially binds the tripeptide consensus sequence n-V-S and its relatives q- V-S, n-V-G and q-V-G. The most selectivity is seen at the dipeptide level where 2 ' binds n-V and q-V in the both frame-shifted sites AA3-AA2 and AA2-AA1. Considering that q and n are similar D-amino acids with carboxamide sidechains, 2 ^,/ s dipeptide selectivity for n/q-V is remarkable.

To confirm these findings and to estimate the energetic extents of some of the selectivities observed, a series of diastereomeric acetoxymethyl-(N-trityl)Asp-Val-(0-

tBu)Ser-linker-bead substrates were synthesized. By treating known quantities of these supported substrates with 2 ' in CHC1 ₃ , UV/VIS spectrophotometry (λ=476 nm) could be used to measure [2'] both before and after treatment with supported substrate and thus evaluate binding energies. The results of these solid phase binding measurements are summarized below:

Peptide Substrate Binding energy (kcal/mol) Found in Assay?

AcOM-n-V-S-polymer -4.4 Yes

AcOM-n-v-S-polymer -0.8 No

AcOM-N-V-S-polymer -3.4 No

AcOM-n-V-s-polymer -2.8 No

AcOM-N-v-s-polymer -3.0 No

AcOM-N-v-S-polymer -3.1 No

These measurements show that 2 ' binds the consensus sequence n-V-S (found by the solid phase, color assay) more tightly than any of the diastereomers studied (not found by assay) . Stereochemical inversion at V is most damaging to binding which is thereby reduced by 3.6 kcal/mol. Inversion of n or S also weakens binding but only by 1.1 and 1.6 kcal/mol, respectively. The residue binding energies are not simply additive as the fully inverted (enantiomeric) sequence N-v-s binds less tightly by only 1.4 kcal/mol.

The binding energies measured by solid phase experiments are not directly comparable to those from solution phase experiments because the environment inside the bead is not quite the same as free solution. In particular, the concentration of supported peptide in the bead is -0.1M and this relatively high concentration favors peptide substrate aggregation which would diminish binding to receptor. Comparison measurements of binding of several

(L)Ala- and (L)Val- containing peptides in free solution and on polystyrene supports using receptor 1' showed that peptide binding is 2-3 kcal/mol stronger when binding is measured in dilute free solution.

A proposed mode of interaction is based on molecular mechanics studies and ¹ H NMR measurements on the N-acetyl- (L)Val-NHtBu complex with 1' which fix the valine side- hain within the center of the receptor. (A. Borchardt and W.C. Still, J. Am. Chem. Soc , 116, 373 (1991)) A conformational search on Ac-(D)Asn(tBu)-(L)Val-NHMe bound to 1' found a certain structure as the global minimum. (AMBER* force field, GB/SA CHC1 ₃ solvation treatment, Monte Carlo conformational search, MacroModel/BatchMin V4.5) The other low energy structures found in the search had similar geometries which differed only in the details of the hydrogen bonding between the substrate side-chain carboxamide and receptor amides.

Whatever the origin of the observed discrimination in binding, receptor 2 ' binds peptide substrates having particular amino acid sequences with remarkable selectivity. This selectivity includes selection based on side-chain stereochemistry (favoring D-Asn/Gln, L- Val) , size (Val»Ala) and functionality (favoring the carboxamides of Asn and Gin) . These results apparently provide the first example of a synthetic receptor which binds peptidic substrates with significant sequence- selectivity.

Receptor 2 ' binds certain side-chain-protected di- and tripeptides with remarkable selectivity and shows a large preference for n/q-V-containing substrates. Based on the number of replicate substrate sequences found in the binding assay, the total number of different substrate sequences which are bound by 2 ' at -4 kcal/mol or better

are 500-1000 out of the entire library of 50,625 sequences.

Highly selective complexation was also found by related binding studies using the analogous side-chain- deprotected library. In that case, 2 ' showed even higher substrate selectivity with a replicate analysis indicating that only 40-70 sequences out of 50,625 are bound at the -4 kcal/mol level of binding. The most commonly found side-chain-deprotected sequences differed from the ubiquitous n/q-V sequences found with the protected library and had AA2=glycine (86%) and AA1=L- serine (49%) or glycine (37%) .

IX

Highly stereoselective binding of a neutral substrate by a receptor in an organic solvent often depends on the substrate's ability to participate in an array of favorable intermolecular electrostatic interactions such as hydrogen bonds. (E.g. Rebek, J.; Askew, B. ; Ballester, P.; Doa, M. J. Am. Chem. Soc , 109, 4119 (1987); Jeong, K.-S; Muehldorf, A.V. ; Rebek, J. J. Am. Chem. Soc , 112, 6144 (1990); Hong, J.-I; Mangoong, S.K.; Bernardi, A. ; Still, W.C. J. Am. Chem. Soc , 113, 5111 (1991); Dixon, R.P. ; Geib, S.J.; Hamilton, A.D. J. Am. Chem. Soc , 114, 365 (1992); Yoon, S.S.; Still, W.C. J. Am. Chem. Soc , 115, 832 (1993); Borchardt, A.; Still, W.C. J. Am. Chem. Soc 1994, 116 , 373) Biological receptors use similar means but also distinguish substrates or fragments by size and seem to make such distinctions based in part upon a substrate's ability to precisely fill a binding

cavity. Here described is a synthetic receptor (1 / / Figure 9(a)) that shows an analogous capability to distinguish closely related peptide substrates in organic solvents based both on stereochemistry and on subtle differences in residue size. l" exhibits high selectivity for binding tripeptides containing an internal L-proline (>99% de for L-Pro vs. D-Pro) , and also binds L-Pro more tightly than cyclic analogs which are both smaller and larger than Pro itself. Furthermore, 1'' stereoselectively binds substrates having L-amino acids adjacent to L-Pro (90-99% de for L- Ala) and with binding constants (K _a = 2.5 x 10 ⁵ for iPrCO- (L)Pro-(L)Ala) that are among the largest reported for binding a neutral guest by a synthetic host.

Receptor 1 has a cup-like shape with a ca. 6A diameter binding cavity surrounded by six unassociated hydrogen bond donors (D) and acceptors (A) . According to molecular mechanics, the design of 1" has a very similar conformation except that the three naphthalenes substantially enlarged the binding cavity (ca. 8A diameter). (A 5,000-step Monte Carlo conformational search (Goodman, J.M. ; Still, W.C. J. Comput . Chem. 1991, 12 , 1110) for 1 (L-Tyr modeled by L-Ala) using the MacroModel/AMBER ^* force field (McDonald, D.Q. ; Still, W.C. Tetrahedron Lett. 1992, 33, 7743 and references therein) found a single conformation within the lowest 2 kcal/mol.) Whereas the original receptor interacted primarily with small terminal substituents of simple peptides (cartoon O; Figure 9(b)), it appeared that the enlarged design might bind internal amino acid residues (R ₂ ) with formation of as many as six hydrogen bonds (cartoon N) .

Synthesis of enantiomerically pure 1" began with naphthol construction via Friedel-Crafts cyclization of the Stobbe-derived half ester 3" (Figure 9(c)). (Cf.

Johnson, W. S.; Graber, R.R. J ^" . Am. Chem. Soc. 1950, 72 , 925.) That material was elaborated by straightforward chemistry to 4" which underwent a triple macrolactamization on treatment with iPr ₂ NEt to provide 1" in 50% yield. Phenol deprotection ( (Ph ₃ P) ₄ Pd/dimedone) and alkylation (Bu ₄ NF, THF) with the mesylate of the azo dye, Disperse Red 1, gave the intensely red receptor derivative 2".

To survey the binding properties of our new receptor, a solid phase color assay was used employing an encoded combinatorial library of -50,000 acylated tripeptide substrates as described above. This library had the general structure R(C=0)-AA3-AA2-AA1-NH(CH ₂ ) ₅ C0NH- polystyrene where R represents 15 different acyl substituents and AA1-AA3 each represent 15 different D and L amino acids. (R = methyl (Me) , ethyl (Et) , isopropyl (iPr) , t-butyl (tBu) , neopentyl (neoPe) , trifluoromethyl (CF ₃ ) , isobutyl (iBu) , methoxymethyl (MOM) , acetoxymethyl (AcOM) , cyclopropyl (cPr) , cyclobutyl (cBu) , cyclopentyl (cPe) , phenyl (Ph) , morpholino (Morph) , dimethylamino (Me ₂ N) . AA1-AA3 = Gly

(G) , D-Ala (a), L-Ala (A), D-Ser(OtBu) (s) , L-Ser(OtBu)

(S), D-Val (v), L-Val (V), D-Pro (p) , L-Pro (P) , D-Asn(N- trityl) (n) , L-Asn(N-trityl) (N) , D-Gin(N-trityl) (q) , L- Gin(N-trityl) (Q) , D-Lys(N-Boc) (k) , L-Lys(N-Boc) (K) .) The library was supported on 50-80 μ polystyrene beads and was prepared both with and without N-trityl/N-Boc sidechain protection. Each bead carried only one type of tripeptide substrate.

Both protected and deprotected substrate libraries were screened for binding by treatment with -50 μM solutions of the red receptor 2" in CHC1 ₃ . After 24hrs of equilibration with each library, -10% of the beads had become colored with -1% being very deep red. The most deeply colored beads were then picked and decoded using

gas chromatography to yield the sequences of the most tightly binding substrates. The binding results were extraordinary. With the protected library, 63 out of 68 tight-binding peptide sequences contained L-Pro and 71% of these placed L-Pro at the AA2 site. Furthermore, in all such sequences having AA2 = L-Pro, the residues at AA1 and AA3 were either L-amino acids or Gly. With the unprotected substrate library, 38 beads were picked and found that all contained L-Pro. As before, L-Pro appeared most commonly in the AA2 position (79% of the sequences) , and sites AA1 and AA3 were nearly always occupied by L-amino acids.

Solid phase binding data for both libraries are represented in Figure 10 as gray-scale histograms which give the fractional occupancy of each substrate site (R, AA1-AA3) by each possible residue (R = Me, Et, . . . ; AAn = L-Ala, L-Ser, . . .) as a gray-scale running between white (0% occupancy) and black (>50% occupancy) . Thus the very dark areas under AA1 and AA2 for L-Pro indicate that most peptides bound by 2" had L-Pro at these substrate sites. The selectivity for AA1-AA3 = L-amino acids can be seen as a generally dark region in the upper half of the histograms just below AA1-AA3 . The large number of gray areas below R indicate little receptor selectivity for the terminal acyl group. The general similarity of the protected and unprotected histograms indicates that 2" binds both libraries with similar selectivity.

The preference of 2" for L-Pro-containing substrates and absence of significant selectivity for the N-terminal substituent suggested that the receptor could be binding substrate approximately as shown in carron N (R ₂ = Pro ring). Indeed, ¹ H NMR of the complex of 1" with iPrCO- (L)Ala-(L)Pro-(L)Ala-NHC ₁₂ H ₂₅ in CDC1 ₃ supported such a binding mode: thus the L-Pro ring CH's shifted upfield

to the region between -0.4 and -2.4 ppm upon binding, a result consistent with the Pro ring penetrating deeply into l'"s aromatic-lined binding cavity. Furthermore, a 40,000-step Monte Carlo conformational search of the complex using AMBER ^* in GB/SA CHCI^ provided additional support by locating just such a structure as the global minimum for the complex.

Table IIA. Binding Energies (kcal/mol of 1 ' ' + Peptides in CHC1,.

entry peptide substrate -ΔG

1 iPrCO-(L)Ala-(L)Pro-(L)Ala-NHC ₁₂ H ₂₅ 7.4

2 iPrCO-(D)Ala-(L)Pro-(L)Ala-NHC ₁₂ H ₂₅ 5.5 3 iPrCO-(L)Ala-(D)Pro-(L)Ala-NHC ₁₂ H ₂₅ 4.3

4 iPrCO-(L)Ala-(L)Pro-(D)Ala-NHC ₁₂ H ₂₅ 4.4

5 iPrCO-(L)Pro-(L)Pro-(L)Ala-NHC ₁₂ H ₂₅ 6.0

6 iPrCO-(L)Pro-(L)Pro-(D)Ala-NHC ₁₂ H ₂₅ 5.4

7 iPrCO-(L)Ala-(L)Azetidinyl-(L)Ala-NHC ₁₂ H ₂₅ 5.3 8 iPrCO-(L)Ala-(L)Pipacolinyl-(L)Ala-NHC ₁₂ H ₂₅ 6.3

To quantify l'"s selectivity for an internal L-proline flanked by L-amino acids, N-dodecylamides of several tight-binding substrates and certain diastereomers were synthesized. Their binding energies were measured in CHC1 ₃ solution ([1] = 100 nM) by titration (Table IIA) with monitoring by CD (Kearney, P.C.; Mizoue, L.S.; Kumpf, R.A. ; Forman, J.E.; McCurdy, A.; Dougherty, D.A. J. Am. Chem. Soc. 1993, 115 , 9907; several of the sequences bound 1" too tightly for monitoring by NMR). 1" bound L-Pro in the AA2 position 3.1 kcal/mol more strongly than D-Pro (entry 1 vs. 3), L-Ala at AA2 3.0 kcal/mol more tightly than D-Ala (entry 1 vs. 4) , and L- Ala at AA3 1.9 kcal/mol more tightly than D-Ala (entry 1 vs. 2) . Taken together, these findings suggest a strong, highly ordered interaction between l'"s chiral binding site and the entire tripeptide chain (as suggested by N

and Fig. 2) . Binding decreased significantly when L-Pro was replaced by either its four- or six-membered ring analog (entry 1 vs. 7 and B) . These results underscore the importance of filling a receptor's binding cavity precisely - binding is disfavored not only when a substrate is too large but also when it is too small. (Cf. Yoon, S.S.; Still, W.C. Tetrahedron Lett . 1994, 35 , 2117.)

Thus, receptor 1" has an enlarged binding cavity but that is otherwise very similar to peptide receptor l. ¹⁰ Though the structural similarities of these two receptors far outweigh the differences, their binding properties are very different. Whereas the previous receptor selects for small terminal residue substituents (as in O) , 1" shows little selectivity for the terminal substituent but instead selects for internal L-Pro with high stereochemical and steric selectivity (as in N) . 1" also shows sequence-selective tripeptide binding in the form of a strong preference for the L-configuration of flanking amino acids. More generally, these results illustrate how known principles of receptor design and the massive data gathering capacity of combinatorial library screening can be combined to create a powerful approach to problems in molecular recognition.