Biomimetic macrocyclic receptors for carboxylate anion recognition based on C-linked peptidocalix[4]arenes

Abstract

Two neutral macrobicyclic anion receptors 4 and 6, containing a calix[4]arene in the cone conformation, two l-alanine units, and a 2,6-diacylpyridine or a phthaloyl bridge, are described. The x-ray crystal structure of the acetone complexes of the pyridine containing macrocycle 6 shows the four amide NH groups to be in close proximity to the chiral pocket delimited by the pyridine and one aromatic nucleus of the calix[4]arene. This conformation is also the most stable in acetone-d6 solution, as proven by one- and two-dimensional NMR spectral measurements. Electrospray ionization–MS and ¹H NMR experiments reveal that the two ligands strongly bind carboxylate anions in acetone solution. H-bonding interactions between the carboxylate anions and the amide NH groups, together with π/π stacking, are invoked to explain the efficiency and the selectivity of these anion receptors.

The selective complexation of carboxylate anions by natural and synthetic hosts is a topic of current interest in bioorganic and supramolecular chemistry (1–5), because these species are involved in several molecular recognition phenomena of biological interest. It is worth noting the special role that carboxylate recognition plays in determining the biological activity of the vancomycin family of antibiotics (6). The enantioselective recognition of chiral carboxylates is also an important goal, because several pharmaceutical compounds possess this functional group. Efficient synthetic receptors for carboxylate anion recognition have been obtained by incorporating in more complex structures charged guanidinium, amidinium, or ammonium groups (2) and neutral H-bonding donor groups such as (thio)ureas (7), pyrroles (8), activated amides (9, 10), and trifluoromethyl alcohols (11). Only a few examples are known where natural amino acids are used as binding units in biomimetic receptors for carboxylate recognition (12). In this case, the intrinsic lower H-bonding donor ability of peptide NH groups could be strengthened by the cooperative action of other weak noncovalent interactions (π/π, CH/π, etc.), and the resulting host–guest binding could be enhanced.

We have recently synthesized a series of upper rim N-linked (13–15) and C-linked (16) peptidocalix[4]arenes with the aim of obtaining novel chiral receptors for amino acids and small peptides (Scheme S1).

Scheme 1.

Interestingly, some of the upper rim bridged peptidocalix[4]arenes (e.g., I) behave as vancomycin mimics and show promising antibiotic activity toward Gram-positive bacterial strands as a consequence of their ability to bind the d-alanyl-d-alanine (d-ala-d-ala) terminal part of peptoglycans (13, 17). Both the N-linked (e.g., II) and C-linked (e.g., III) cleft-like peptidocalix[4]arenes are very poor receptors in the recognition of ions and polar organic molecules, mainly because of their residual conformational flexibility, which, in the case of C-linked compounds, leads to extensive intramolecular H bonding (see III). To reduce this conformational flexibility, we decided to cap the C-linked peptidocalix[4]arenes with suitable aromatic units and to explore the host properties of macrobicyclic receptors 4 and 6, in particular toward anionic guest species. The solid-state structure of receptor 6 was determined by x-ray crystallography, whereas two-dimensional NMR was used to clarify the solution structures of both 4 and 6. The results show that the new ligands form strong complexes in acetone-d6 with carboxylate anions and are selective for benzoate, which is able to interact with the hosts through H bonding and π/π stacking.

Materials and Methods

The general synthetic procedures, methods, and instrumentation for x-ray crystallography, NMR spectroscopy, and mass spectrometry are published as supporting information on PNAS web site (www.pnas.org).

Synthesis of Compounds.

5,11-Diamino-25,26,27,28-tetra-n-propoxycalix[4]arene (1).

This compound was synthesized according to a literature procedure (18).

5,17-Bis(N-carbobenzyloxy-l-alanylamido)-25,26,27,28-tetra-n-propoxycalix[4]arene (2).

To a solution of 5,11-diamino-25,26,27,28-tetra-n-propoxycalix[4]arene (18) 1 (1.00 g, 1.61 mmol) and Et₃N (0.65 ml, 4.82 mmol) in dry CH₂Cl₂ (100 ml), N-Cbz-l-alanine (1.01 g, 4.82 mmol) and O-benzotriazol-1-yl-N,N,N′,N′-tetramethyluronium hexafluorophosphate (1.83 g, 4.82 mmol) were added. The reaction was stirred at room temperature for 14 h, then quenched by adding 1 M HCl (50 ml). The organic layer was separated, washed with H₂O (50 ml), 5% NaHCO₃ aqueous solution (50 ml), and H₂O (50 ml), dried over MgSO₄, and evaporated to dryness under reduced pressure. The crude was purified by flash chromatography (hexane/AcOEt 2:1, v:v) to obtain 2 as a white solid (1.49 g, 90% yield): mp 116–118°C; [α]-7.1 (c = 1.8, acetone); ¹H NMR (300 MHz, CDCl₃): δ = 7.95 (bs, 2H, ArNHCO), 7.43–7.26 (m, 10H, Ar of Cbz), 6.84 (bs, 4H, Ar), 6.82–6.59 (m, 4H, Ar), 6.21 (bs, 2H, Ar), 5.63 (bs, 2H, NH ala), 5.14 and 5.04 (d, 2H each, J = 12.2 Hz, NCOCH₂Ph), 4.43 (d, 4H, J = 13.7 Hz, H_ax of ArCH₂Ar), 4.24–4.09 (m, 2H, CH ala), 3.93 (t, 4H, J = 7.2 Hz, OCH₂CH₂CH₃), 3.73 (t, 4H, J = 6.5 Hz, OCH₂CH₂CH₃), 3.16 and 3.11 (d, 2H each, J = 13.7 Hz, H_eq of ArCH₂Ar), 2.04–1.79 (m, 8H OCH₂CH₂CH₃), 1.32 (d, 6H, J = 6.8 Hz, CH₃ ala), 1.03 and 0.94 (t, 6H each, J = 7.3 Hz, OCH₂CH₂CH₃); ¹³C NMR (75 MHz, CDCl₃): δ = 170.4 (s, CO), 157.0 (s, Ar ipso), 156.5 (s, CO Cbz), 153.5 (s, Ar ipso), 136.1 (s, Ar Cbz), 135.5 and 135.3 (s, Ar ortho), 134.8 (s, Ar ortho), 130.8 (s, Ar para), 128.7 (d, Ar meta), 128.5 (d, Ar Cbz), 128.3 (d, Ar meta), 128.1 and 127.9 (d, Ar Cbz), 122.0 (2d, Ar para and Ar meta), 120.6 (d, Ar meta), 77.4 and 76.8 (t, OCH₂CH₂CH₃), 67.1 (t, NCOCH₂Ph), 50.8 (d, CH ala), 31.0 and 30.9 (t, ArCH₂Ar), 23.2 and 23.0 (t, OCH₂CH₂CH₃), 17.9 (q, CH₃ ala), 10.4 and 10.0 (q, OCH₂CH₂CH₃); CI-MS (+): m/z = 1032.9 (100%) [M⁺]. Analysis calculated for C₆₂H₇₂N₄O₁₀: C, 72.07; H, 7.02; N, 5.42. Found: C, 72.21; H, 7.13; N, 5.35.

5,17-Bis(l-alanylamido)-25,26,27,28-tetra-n-propoxycalix[4]arene (3).

A mixture of compound 2 (1.40 g, 1.36 mmol) and Pd/C (10%) in AcOEt (50 ml) was stirred under H₂ pressure (2 bar) at room temperature for 48 h. The reaction was quenched by removing the catalyst by filtration, which was washed with AcOEt. The combined filtrates were evaporated to dryness under reduced pressure to give 3 as a white solid (0.99 g, 95% yield), which was used without further purification. Physical and spectroscopic data are in full agreement with those previously reported (16).

Cyclization Reaction.

To a stirred 1 × 10⁻³ M solution of compound 3 and NEt₃ (3 equiv.) in dry CH₂Cl₂, maintained at 0°C, a 5 × 10⁻² M solution of the aromatic dicarbonyl dichloride (1.1 equiv.) in dry CH₂Cl₂ was added dropwise. The reaction mixture was stirred at room temperature overnight and quenched by adding a 5% Na₂CO₃ aqueous solution. The organic layer was separated, washed with H₂O, dried over MgSO₄, and evaporated to dryness under reduced pressure.

Macrocycle 4.

The crude product was purified by flash chromatography (CH₂Cl₂/MeOH 25:1, v:v). This compound was obtained in 40% yield: mp 254–256°C; [α] + 7.4 (c = 1.0, acetone); ¹H NMR (400 MHz, acetone-d6): δ = 8.90 (s, 2H, ArNHCO), 8.05 (d, 2H, J = 7.9 Hz, NH ala), 7.97 (dd, 2H, J = 1.7, 7.7 Hz, Ar isophthalic), 7.71 (d, 1H, J = 1.7 Hz, Ar isophthalic), 7.52 (t, 1H, J = 7.7 Hz, Ar isophthalic), 7.15 and 7.06 (dd, 2H each, J = 1.5, 7.5 Hz, Ar), 6.85 (t, 2H, J = 7.5 Hz, Ar), 6.67 and 6.62 (d, 2H each, J = 2.5 Hz, Ar), 4.68–4.53 (m, 2H, CH ala), 4.50 and 4.48 (d, 2H each, J = 12.5 Hz, H_ax of ArCH₂Ar), 4.21–4.10 (m, 4H, OCH₂CH₂CH₃), 3.66 (t, 4H, J = 7.2 Hz, OCH₂CH₂CH₃), 3.22 and 3.16 (d, 2H each, J = 12.5 Hz, H_eq of ArCH₂Ar), 2.19–2.09 and 2.01–1.84 (m, 4H each, OCH₂CH₂CH₃), 1.37 (d, 6H, J = 7.1 Hz, CH₃ ala), 1.06 and 0.94 (t, 6H each, J = 7.6 Hz, OCH₂CH₂CH₃); ¹³C NMR (75 MHz, acetone-d6): δ = 171.0 and 166.8 (s, CO), 158.3 and 152.8 (s, Ar ipso), 137.3, 133.8, 133.1, 130.5, 129.9, 129.8, 129.5, 127.5, 123.6, 122.6 and 122.2 (Ar and Ar isophthalic), 78.7 and 76.9 (t, OCH₂CH₂CH₃), 51.5 (d, CH ala), 32.0 (t, ArCH₂Ar), 24.1 and 23.5 (t, OCH₂CH₂CH₃), 17.2 (q, CH₃ ala), 11.0 and 10.1 (q, OCH₂CH₂CH₃); CI-MS (+): m/z = 895.3 (100%) [M⁺ + 1]. Analysis calculated for C₅₄H₆₂N₄O₈: C, 72.46; H, 6.98; N, 6.26. Found: C, 72.60; H, 7.05; N, 6.17.

Macrocycle 5.

This compound, which was the first eluted in flash chromatography, was obtained in 4% yield; selected data: ¹H NMR (300 MHz, CD₃OD): δ = 8.38 (d, 2H, J = 1.6 Hz, Ar isophthalic), 8.03 (dd, 4H, J = 1.6, 7.8 Hz, Ar isophthalic), 7.58 (t, 2H, J = 7.8 Hz, Ar isophthalic), 7.04 (d, 4H, J = 2.3 Hz, Ar), 6.80 and 6.75 (d, 4H each, J = 7.3 Hz, Ar), 6.60 (d, 4H, J = 2.3 Hz, Ar), 6.56 (t, 4H, J = 7.3 Hz, Ar), 4.60 (q, 4H, J = 7.1 Hz, CH ala), 4.48 and 4.46 (d, 4H each, J = 12.6 Hz, H_ax of ArCH₂Ar), 3.90 and 3.81 (t, 8H each, J = 7.4 Hz, OCH₂CH₂CH₃), 3.15 and 3.11 (d, 4H each, J = 12.6 Hz, H_eq of ArCH₂Ar), 2.14–1.87 (m, 16H, OCH₂CH₂CH₃), 1.47 (d, 12H, J = 7.1 Hz, CH₃ ala), 1.07 and 1.01 (t, 12H each, J = 7.6 Hz, OCH₂CH₂CH₃); CI-MS (+): m/z = 1790.6 (100%) [M⁺ + 1], 1789.4 (50%) [M⁺]. Analysis calculated for C₁₀₈H₁₂₄N₈O₁₆: C, 72.46; H, 6.98; N, 6.26. Found: C, 72.55; H, 7.09; N, 6.21.

Macrocycle 6.

The crude product was purified by flash chromatography (hexane/AcOEt 4:1, v:v). This compound was obtained in 30% yield: mp 234–236°C; [α] + 17.8 (c = 1.0, acetone); ¹H NMR (300 MHz, acetone-d6): δ = 8.69 (s, 2H, ArNHCO), 8.62 (d, 2H, J = 7.0 Hz, NH ala), 8.49–8.14 (m, 3H, Ar pyridine), 7.14 and 7.11 (dd, 2H each, J = 1.6, 7.4 Hz, Ar), 6.93 (d, 2H, J = 2.6 Hz, Ar), 6.88 (t, 2H, J = 7.4 Hz, Ar), 6.60 (d, 2H, J = 2.6 Hz, Ar), 4.50 (d, 4H, J = 12.8 Hz, H_ax of ArCH₂Ar), 4.39–4.27 (m, 2H, CH ala), 4.26–4.12 (m, 4H, OCH₂CH₂CH₃), 3.67 (t, 4H, J = 7.0 Hz, OCH₂CH₂CH₃), 3.19 and 3.17 (d, 2H each, J = 12.8 Hz, H_eq of ArCH₂Ar), 2.20–2.02 and 2.03–1.87 (m, 4H each, OCH₂CH₂CH₃), 1.48 (d, 6H, J = 7.0 Hz, CH₃ ala), 1.11 and 0.94 (t, 6H each, J = 7.6 Hz, OCH₂CH₂CH₃); ¹³C NMR (75 MHz, acetone-d6): δ = 170.3 and 163.9 (s, CO), 158.4 (s, Ar ipso pyridine), 152.5 and 150.6 (s, Ar ipso), 140.3 (s, Ar para), 137.4 (s, Ar ortho), 133.8 (s, Ar ortho), 129.7 (d, Ar pyridine), 128.9 (d, Ar meta), 125.2 (d, Ar pyridine), 123.4 (d, Ar para), 120.6 and 120.5 (d, Ar meta), 78.5 and 77.1 (t, OCH₂CH₂CH₃), 52.6 (d, CH ala), 32.1 and 31.9 (t, ArCH₂Ar), 24.2 and 23.5 (t, OCH₂CH₂CH₃), 17.2 (q, CH₃ ala), 11.1 and 10.1 (q, OCH₂CH₂CH₃); CI-MS (+): m/z = 895.6 (100%) [M⁺]. Analysis calculated for C₅₃H₆₁N₅O₈: C, 71.04; H, 6.86; N, 7.82. Found: C, 71.11; H, 6.89; N, 7.72.

Macrocycle 7.

This compound, which was the first eluted in flash chromatography, was obtained in 3% yield; selected data: ¹H NMR (300 MHz, CD₃OD): δ = 8.16–8.00 (m, 6H, Ar pyridine), 6.95–6.81 (m, 8H, Ar), 6.70–6.53 (m, 4H + 4H, Ar), 6.48 (d, 4H, J = 1.8 Hz, Ar), 4.60–4.49 (m, 4H + 8H, CH ala + H_ax of ArCH₂Ar), 3.99 and 3.76 (t, 8H each, J = 7.0 Hz OCH₂CH₂CH₃), 3.15 (d, 8H, J = 13.2 Hz, H_eq of ArCH₂Ar), 2.11–1.86 (m, 16H, OCH₂CH₂CH₃), 1.56 (d, 12H, J = 7.3 Hz, CH₃ ala), 1.11 and 0.98 (t, 12H each, J = 7.4 Hz, OCH₂CH₂CH₃); CI-MS (+): m/z = 1791.3 (100%) [M⁺]. Analysis calculated for C₁₀₆H₁₂₂N₁₀O₁₆: C, 71.04; H, 6.86; N, 7.82. Found: C, 71.20; H, 6.90; N, 7.69.

Crystal Data and Experimental Details.

Empirical formula C₅₃H₆₁N₅O₈⋅1.5CO(CH₃)₂⋅3H₂O; formula weight, 1,039.28 amu; crystal size, 0.4 × 0.3 × 0.4 mm; crystal system, Monoclinic; space group C 2; a = 48.107 (10), b = 12.347(3), c = 18.490(4) Å; β = 97.97(3)°; V = 10,877(3) ų; Z = 8; ρ (calcd.) = 1.269 g⋅cm⁻³; F(000) = 4,464; T = 295 K; index ranges, −58 ≤ h ≤ 58, 0 ≤ k ≤ 15, 0 ≤ l ≤ 22; reflections: 12,333 collected, 10,740 independent (R_int = 0.042), 3,988 observed [F₀ ≥ 4σ(F₀)]; data/restraints/parameters, 10,740/22/1,299; weighting scheme, w = [σ²(F) + (0.2P)²]⁻¹, where P = (F + 2F)/3; goodness-of-fit on F² = 1.027; final R indices (obs. data), R₁ = 0.135, wR₂ = 0.33; largest diff. peak and hole = 0.92, −0.56 eÅ⁻³. [R₁ = Σ || F₀|− | F_c||/Σ |F₀|, wR₂ = [Σ w(F − F)²/Σw F]^1/2. Goodness-of-fit = [Σw(F − F)²/(n–p)]^1/2, where n is the number of reflections and p the number of parameters].

The structure was solved by direct methods by using SIR92 (19) and refined on F² by using SHELXL-97 (20). The geometrical parameters were obtained from PARST97 (21).

Results and Discussion

Synthesis and Conformational Properties of the Receptors in Solution.

Receptors 4 and 6 were synthesized starting from the known (18) diamino-tetrapropoxycalix[4]arene 1 (Scheme S2). The condensation of 1 with N-Cbz-l-alanine in the presence of O-benzotriazol-1-yl-N,N,N′,N′-tetramethyluronium hexafluorophosphate, followed by deprotection, gave compound 3 in 85% overall yield. The reaction of 3 with aromatic diacyl dichlorides gave the macrobicyclic receptors 4 and 6 in 40 and 30% yield, respectively. In both cases, also the corresponding macrotricyclic 2 + 2 condensation products 5 and 7 were isolated, although in very low yields. All compounds were characterized by NMR, MS, and elemental analysis.

Scheme 2.

Receptor 4 shows in CDCl₃, at room temperature, a broad ¹H NMR spectrum that becomes sharp after a modest heating (323 K) (see Fig. 3, which is published as supporting information). Dilution experiments do not produce substantial changes in the spectra, thus ruling out the formation of intermolecular aggregates and suggesting that broadening is because of the presence of different conformations, which slowly exchange (on the NMR timescale) in CDCl₃ because of some intramolecular H bonding. We were not able to “freeze” the conformational interconversion even at 243 K. In acetone-d6, the spectrum is sharp at room temperature, suggesting either the presence of a symmetric C₂ structure or a fast exchange among different conformers. The presence of crosspeaks in the rotating-frame Overhauser effect spectroscopy spectra (see Fig. 4, which is published as supporting information) between the signals of protons 2, 4, and 6 of the isophthaloyl aromatic nucleus and the alanyl amide protons supports the second hypothesis, where the equilibrating conformations differ for the reciprocal orientation (Scheme S3) of the NH protons, which can be trans (a) or cis-outside (b) and cis-inside (c) (22–26).

Scheme 3.

Interestingly, compound 6 shows a slightly different conformational behavior that indicates a more rigid structure. The ¹H NMR spectrum in CDCl₃ (Fig. 1a) is rather sharp at room temperature and indicates the presence of three different conformers, one of which is significantly more abundant (80%). A similar situation is observed in C₂D₂Cl₄, but heating at 363 K causes an extended broadening of the signals that are close to coalescence. On the other hand, the spectrum at room temperature in acetone-d6 is very sharp (Fig. 1b). The two chiral alanine units cause a splitting of the signals of some aromatic protons of the calixarene and of those of the methylene bridge. Moreover, the doublet relative to the alanine NHs is significantly downfield shifted (δ = 8.62 ppm) if compared with the analogous signal observed in compound 4 (δ = 8.05 ppm), probably because they are close to the pyridine nitrogen atom. In addition, the absence of any crosspeak between the alanine NHs and the pyridine aromatic protons in the rotating-frame Overhauser effect spectroscopy spectra indicates that the only conformation present in acetone has the two alanine NH protons in a reciprocal cis orientation (see Scheme S3c) and directed toward the interior of the macrocycle defined by the pseudopeptide. In this structure, which is similar to that found in the solid state (see below), it may be that the repulsive interaction between the lone pairs of the pyridine nitrogen atom and the amide carbonyl groups is minimized and the conformation is stabilized by attractive electrostatic interactions between the amide NHs and the pyridine nitrogen atom, as observed in similar systems (24–27).

Figure 1.

¹H NMR spectra (9.0–5.5 ppm, 400 MHz, 300 K) of receptor 6 in (a) CDCl₃ and (b) acetone-d6.

Solid-State Structure of Compound 6 and Molecular Modeling.

Compound 6 crystallizes from acetone in a single phase containing two different types of complexes with the solvent. They show a 1:2 (A) and a 1:1 (B) ligand/acetone stoichiometry, and their molecular structures are reported in Fig. 2. In both complexes, the calix[4]arene is in a flattened cone conformation with the two rings bearing the pyridine bridge almost parallel [7(2)° and 10(1)° in A and B, respectively] and the other two almost orthogonal [100(2)° and 90(1)° in A and B, respectively].

Figure 2.

Ball-and-stick representation of the x-ray molecular structures of 1:2 (A) and 1:1 (B) ligand 6/acetone complexes. Hydrogen atoms are omitted for clarity.

The most relevant data for the calix[4]arene conformation in the two complexes are reported in Table 1, according to the standard rules (28). Apart from the different host–guest stoichiometry, the most significant difference in the two complexes is found in the orientation of the pyridine nucleus with respect to the calixarene basket. In Fig. 2A, the pyridine ring (Py) is almost parallel with the phenolic unit a [Py∧a = 10(3)°, and the distance between the barycenters of the two rings is 7.56(8) Å], whereas in Fig. 2B, because of the absence of one acetone molecule, the pyridine is closer to the corresponding phenolic nucleus e [Py∧e = 25(1)° and the distance between the barycenters of the two rings is 7.07(8) Å].

Table 1.

Dihedral angles (°) between the molecular reference plane (the least-squares plane through the bridging methylene carbon atoms) (R_A in A and R_B in B) and the least-squares planes through the phenolic units (a–h), and relevant intermolecular host–guest distances (Å)

RA∧a	145 (2)	RB∧e	133 (1)	N4*⋯O1r	3.22 (6)	N5*⋯O1r	2.94 (6)
RA∧b	96 (2)	RB∧f	94.0 (6)	N2*⋯O1s	3.06 (7)	N4*⋯O1s	3.06 (6)
RA∧c	135 (1)	RB∧g	137.0 (8)	N2′⋯O1t	2.98 (3)	N4′⋯O1t	3.04 (4)
RA∧d	91 (2)	RB∧h	95.8 (9)

We also examined the possibility that the observed conformations could be stabilized by intramolecular H bonding between the alanine amide NH groups (N2* and N4* in Fig. 2A and N2′ and N4′ in Fig. 2B) and the pyridine nitrogen atom (N3* in Fig. 2A and N3′ in Fig. 2B), as suggested in the case of other systems containing the pyridine acylamide unit (25–27). Although the structure determination did not allow the hydrogen atoms from the electron density maps to be located, theoretical calculations with semiempirical methods at the PM3 level (29) in the spartan suite (30) allowed the equilibrium geometry of the ligand (see Fig. 5, which is published as supporting information) to be reproduced, so that the reciprocal orientation of the pyridine and of the amide NH groups could be established. The structure shows a H⋅⋅⋅N_Py distance of 2.7 Å, but the very low values of the N—H⋅⋅⋅N_Py angle (from 90 to 100°) make unlikely the formation of intramolecular hydrogen bonding. The host–guest distances reported in Table 1 reveal that the acetone guest molecules s and t in the 1:2 and in the 1:1 complex, respectively, are linked to the host through a symmetrical bifurcated hydrogen bond involving the N2—H and N4—H groups, whereas the second acetone molecule r in the 1:2 complex (Fig. 2A) interacts only with the N5*—H group, which, as a consequence, is forced out of coplanarity with the aromatic ring. Because no crystal structure was available for receptor 4, molecular modeling calculations at the PM3 level were carried out to explore its conformational space. The calculations predict that several molecular conformations having different folding patterns of the bridging isophthalic group are possible in such a way that the benzene ring in the center of the bridge can span the entire arch at the top of the structure. At least one of the low-energy conformations (Fig. 5, which is published as supporting information) is characterized by a trans orientation of the phthaloyl NH proton amide groups, in good agreement with the experimental two-dimensional NMR data (see above).

Binding Properties.

A preliminary screening by electrospray ionization (ESI)-MS was performed to qualitatively evaluate the binding properties of receptors 4 and 6 toward several anionic guests. Competitive binding studies were carried out by adding a mixture of inorganic (Cl⁻, I⁻, NO, H₂PO), and carboxylate anions, each in 1:1 ratio with the host, to an acetone solution of the receptor. Interestingly, a good selectivity for carboxylate over inorganic anions was observed (see Fig. 6, which is published as supporting information), which prompted us to quantitatively evaluate the association constants by means of NMR methods. To overcome the conformational problems encountered in CDCl₃ (see above), the complexation properties of ligands 4 and 6 were studied in acetone-d6, which has been previously used in anion binding studies (9, 23). In each binding experiment, a 1:1 stoichiometry was assumed for the complexation, which was generally supported by the good fit of the experimental data to the theoretical model and by the results of ESI-MS, which show the presence of only this species in different conditions. Association constants were determined both by standard titration experiments (31) and by concentrating a 6.2 × 10⁻⁵ M solution of the 1:1 host–guest complex up to 2.5 × 10⁻³ M. The observed shifts of several signals, induced by complexation, were analyzed by nonlinear least-squares fitting procedures. Because of the strong binding displayed by 4 and 6 toward the investigated anions, the second method proved more reliable and allowed data to be collected in the ideal range of 20–80% complexed species (31). Whereas in the presence of carboxylate anions we observed shifts of most of the calixarene protons and, in particular, significant downfield shifts of the amide NH protons, no shift and consequently no binding were observed with the corresponding carboxylic acids, indicating that the host–guest interaction is mainly because of the pseudopeptide NH groups, which act as H-bonding donors.

Having verified with acetate and benzoate anions that the two ligands show a quite similar complexation behavior (Table 2), we focused our attention on the pyridine ligand 6, which has a more rigid conformation. Inspection of Table 2 reveals some interesting features. Receptor 6 shows selectivity toward Y-shaped carboxylate anions, whose binding is quite strong, also considering the high donicity of the solvent. Association constants are much higher than those found for the cleft-like peptidocalix[4]arene III having the same number of H-bonding donor groups and for other preorganized receptors containing the 2,6-pyridine diamide (32, 33) or calix[4]arene amide (9, 34) moieties. As observed by ESI-MS, inorganic anions of different shapes (Cl⁻, NO) are more weakly bound. The poor binding of NO anion by ligand 6 is particularly striking in view of the nitrate selectivity observed in macrocyclic cyclophanes containing the 2,6-pyridine amide binding groups (32). Among the carboxylate anions, selectivity is toward the aromatic ones and especially for benzoate [pKa in DMSO = 11.0 (35)], which is more strongly bound than acetate anion [pKa in DMSO = 12.3 (35)]. The available data seem to indicate that for host 6, binding is occurring inside the pocket formed by the calix[4]arene and the pseudopeptide loop, and that additional noncovalent interactions, besides the most important COO⁻⋅⋅⋅H—N hydrogen bonding, are playing a role in stabilizing the complexes. To explain the benzoate/acetate selectivity and the small preference for phenyl substituted α-amino acid derivatives, we have to assume a cooperative stabilization of the complexes via π/π stacking interaction between the aromatic nucleus of the guest and the pyridine moiety and/or a phenolic unit of the host, as observed in other systems (34).

Table 2.

Binding constants (K_ass, M⁻¹) and free energy of complexation (−ΔG°, KJ/mol) for the 1:1 complexes formed by macrocycles 4, 6, and open-chain peptidocalixarene III with anionic substrates (as tetrabutylammonium salts) in acetone-d6 (300 K)

Host	Guest	Kass	−ΔG°
III	Acetate	370	14.8
	Benzoate	680	16.3
4	Acetate	7100	22.1
	Benzoate	44000	26.7
6	Acetate	10500	23.1
	Benzoate	40100	26.4
	p-Methoxybenzoate	33300	26.0
	N-Ac-GlyCO	6200	21.8
	N-Ac-L-AlaCO	4900	21.2
	N-Ac-D-AlaCO	5700	21.6
	N-Ac-L-PheCO	7900	22.4
	N-Ac-D-PheCO	10500	23.1
	Cl−	2800	19.8
	NO	200	13.2

The association constants are affected by errors ≤10%. 

Interestingly, by measuring the ESI-MS spectra of an acetone solution of the ligand and a mixture of the tetrabutyl ammonium carboxylates, each in 1:1 host–guest ratio, we found, on the basis of the relative intensities of each complex molecular peak (Fig. 7, which is published as supporting information), a selectivity order (benzoate>p-methoxybenzoate>N-acetyl-phenylalaninate>N-acetyl-alaninate) qualitatively similar to that determined by NMR experiments, with the exception of acetate anion, whose complexation appears reduced in the MS experiments. Although the two ligands 4 and 6 are chiral (C₂ symmetry), they show only a very modest enantioselectivity for the d-enantiomers in the few cases investigated.

Conclusion

In summary, we have been able to show that efficient receptors for carboxylate anion recognition can be obtained by bridging the C-linked 1,3-dialanylcalix[4]arenes in the cone conformation with 2,6-diacylpyridine or isophthaloyl moieties. In this way, the residual conformational mobility of cone calix[4]arenes is blocked, and the resulting more preorganized receptors display enhanced efficiency. Selectivity toward aromatic carboxylates and especially toward benzoate is observed, thanks to π/π stacking interactions with the pyridine moiety and/or a calix[4]arene aromatic nucleus, which act in addition to hydrogen bonding with amide NH groups.

Abbreviation

ESI-MS

electrospray ionization–MS