Skip to main content
NCBI home page
JACS Au logoLink to JACS Au
. 2025 Jan 23;5(2):902–912. doi: 10.1021/jacsau.4c01118

Acceleration and Selectivity of 1,3-Dipolar Cycloaddition Reactions Included in a Polar [4 + 2] Octa-imine Bis-calix[4]pyrrole Cage

Yifan Li †,‡, Chiara F M Mirabella †, Gemma Aragay †, Pablo Ballester †,§,*
PMCID: PMC11862939  PMID: 40017753

Abstract

graphic file with name au4c01118_0009.jpg

We describe the quantitative self-assembly (>90%) of a [4 + 2] octa-imine cage (1) in a CDCl3:CD3CN 9:1 solvent mixture containing 0.5% of acetic acid. Cage 1 is based on two identical aryl-extended calix[4]pyrrole units linked through eight dynamically reversible imine bonds. Cage 1 forms thermodynamically and kinetically highly stable inclusion complexes featuring 1:1 and 2:1 stoichiometry with suitable para-substituted pyridine-N-oxides. The ability of 1 for the pairwise inclusion of two different pyridine-N-oxides led us to investigate its properties as a reactor vessel. The coinclusion of 4-azido pyridine-N-oxide and 4-ethynyl pyridine-N-oxide did not produce a detectable acceleration of their 1,3-dipolar cycloaddition reaction. Conversely, the coinclusion in cage 1 of the same alkyne dipolarophile with 4-azido(alkyl) pyridine-N-oxides (alkyl= methyl, ethyl) produced significant reaction acceleration. We quantified the reactions’ acceleration with an effective molarity (EM) of ∼103 M, corresponding to the more prominent reported value of a bimolecular 1,3-dipolar cycloaddition reaction in a molecular vessel by directly detecting the ternary Michaelis complex. The included reactions are quantitative and regioselective, yielding exclusively the 1,4-disubstituted triazole isomers. We propose that the selectivity of 1 in accelerating the included 1,3-dipolar cycloadditions is related to (a) the entropy gain provoked by the reaction’s inclusion, (b) the rigidity of the container, and (c) the spatial fixation of the polar knobs (pyridine-N-oxide) carrying the reacting groups in its two functionalized hemispheres. The two latter characteristics render the distance between the reacting groups (azido and ethynyl) almost fixed by design, thus allowing or not achieving the transition state’s geometry. We support our hypothesis with the help of DFT calculations of the inclusion complexes’ structures.

Keywords: molecular container, dynamic covalent cage, reactor vessel, click chemistry, calix[4]pyrrole

Introduction

The confinement of molecules in nanometric spaces can change their physical and chemical properties. Enzymes bind substrates in confined environments, a.k.a active sites, to catalyze their chemical transformations under mild conditions.1,2 Moreover, the reacting groups of the bound substrates are placed in close proximity and adopt a suitable orientation to react. This results in an effective increase in the concentration of the reactants and endows intermolecular reactions with an intramolecular character, resulting in a significant rate increase. Page and Jencks quantified the entropic contribution to rate acceleration for bimolecular enzymatic reactions up to −35 e.u. (corresponding to a rate acceleration EM = eΔS/R = 4 × 107 M) or more.3 It is also proposed that enzymes bind elusive transition states more tightly than either substrates or products, causing additional rate acceleration.4−6

Synthetic molecular containers stabilize highly reactive species and high-energy conformations by including substrates in their cavities. Likewise, the inclusion of reactions in molecular containers led to altered pathways, changes in selectivity, and noticeable rate increases when compared to those in bulk solution.7−9

In particular, the acceleration experienced by cycloaddition reactions upon inclusion in covalent, supramolecular, and biological synthetic molecular containers was used to assess their potential as synthetic enzymes.10 Several examples of Diels–Alder (DA) reactions accelerated inside the cavity of molecular containers were reported.11−17 The inclusion of reactants in molecular containers can also influence the diastereoselectivity of the DA reaction (i.e., “endo”/“exo” products ratio).10 In most of the reported examples of DA reactions inside molecular containers, either one or both reactants contained heteroatoms, and the maximum acceleration value measured as EM = kintra/kbulk was ≤103 M. Recently, Nau and co-workers estimated that performing the cyclopentadiene dimerization inside cucurbit[7]uril led to an acceleration factor of ca. 4 × 105 M, which is close to the predicted maximum.3,18

Notably, the acceleration of Huisgen cycloadditions by inclusion in molecular containers has been less explored. In bulk, the Huisgen cycloaddition requires elevated temperatures. It often produces a mixture of two 1,2,3-triazole regioisomers (1,4 and 1,5).19,20 Copper21 or ruthenium-22 catalyzed alternatives are well-known to selectively obtain the 1,4 or the 1,5-disubstituted 1,2,3-triazole isomers, respectively.

In the nineties, Mock et al. described the acceleration of intermolecular cycloadditions between azide and alkyne ammonium derivatives included in cucurbit[6]uril.23,24 The reaction of N-tert-butylated substrates produced a product with [2]rotaxane topology, simplifying the analysis of the kinetic data. From the reported data, it was derived that the reaction’s acceleration factor expressed as EM could be estimated as ca. 1.9 × 104 M.25,26 The lack of direct detection of the ternary complex (Michaelis) made the kinetic data analysis and the derivatization of the EM value not straightforward.

In 2002, Rebek and Jia disclosed that the cycloaddition between phenyl azide and phenyl acetylene was accelerated when included in a self-assembled resorcin[4]arene capsular dimer.27 Cacciapaglia and co-workers calculated the EM acceleration factor to be 120 M.25 In this case, the ternary complex was observed, simplifying the kinetic data analysis and the derivatization of an EM value.

It is worth noting that the Huisgen reactions proceeding in the interior of CB6 and the self-assembled capsular dimer were regioselective, yielding exclusively the 1,4-isomer of the triazole products. The results of theoretical calculations of the reactions taking place inside these containers assigned entropic effects and geometric strain between reactants as the most relevant factors for the observed reaction accelerations.28,29 The putative increase in pressure experienced by the included reactants might also accelerate the reactions owing to their known negative activation volumes.18

A few years ago, our group described the template-directed synthesis of octa-imine cages based on aryl-extended calix[4]pyrrole units.30 A unique feature of these reported cages was that their hemispheres were defined by two endohedral and converging polar binding sites. The 4,4′-bis-pyridine-N,N′-dioxide resulted in a perfect template for the quantitative assembly of the dynamically covalent cages but could not be removed afterward. More recently, we published the self-assembly of a related [1 + 1] tetra-imine calix[4]pyrrole cage in a 9:1 CDCl3:CD3CN solvent mixture.31 The dynamic tetra-imine covalent cage was assembled in a good yield. Using weakly bound CD3CN molecules as the template for the cage’s assembly allowed the study and characterization of its binding properties with pyridine-N-oxide guests.

Herein, we report the self-assembly of a much larger [4 + 2] octa-imine bis-calix[4]pyrrole cage 1 featuring two identical endohedral functionalized hemispheres (Scheme 1). We describe the binding properties of cage 1 with several pyridine-N-oxide guests functionalized at their para-position with acetylene and azido groups. We investigate the kinetics of the 1,3-dipolar cycloaddition reaction between azido- and alkyne-derived pyridine-N-oxides (Scheme 2) included in the cavity of 1 (intra) and compare them with those in bulk (bulk). Some reactions experienced acceleration factors, quantified as effective molarity, EM= kintra/kbulk, larger than 103 M. The systems do not exhibit turnover but feature the largest acceleration value described for bimolecular reactions included in molecular vessels, for which the direct detection of the ternary complex (Michaelis) was possible. Our results demonstrate that substrate binding and confinement in a synthetic molecular container produced significant accelerations of bimolecular Huisgen cycloaddition reactions.

Scheme 1. Equilibrium of the Self-Assembly Process of the [2 + 4] Octa-imine Cage 1.

Scheme 1

Open in a new tab

The proton assignment for cage 1 and its synthetic precursors, tetra-amine 2 and terephthalaldehyde 3, is also shown.

Scheme 2. (Top) Scheme of the 1,3-Dipolar Cycloaddition Reaction between Azido (4a, 4b or 4c) and Ethynyl (5) Derivatives Producing the Two Regioisomeric 1,4- (6) and 1,5- (7) Disubstituted 1,2,3-Triazoles (The Corresponding Proton Assignment for 4a–c, 5, and 6a–c Is Shown); (Bottom) Molecular Structure of 4,4-(Buta-1,3-diyne-1,4-diyl)bis-pyridine-N-Oxide 8 Used as a Template in the Self-Assembly of the Octa-imine Cage 1 Is Also Shown.

Scheme 2

Open in a new tab

Synthesis

Pyridine-N-oxide Derivatives

We uneventfully prepared pyridine-N-oxide derivatives 4a-d using slightly modified reported methodologies (Scheme 2, see SI for experimental details).32−34 We performed the Cu(I) catalyzed cycloaddition reactions of the N-oxides 4a, 4b, and 4c, with 5, by standard literature procedures.21 Using preparative TLC, we isolated the corresponding 1,4- and 1,5- triazole isomers, 6 and 7. We also synthesized the bis-N-oxide 8 by dimerization of 5 using Hay reaction conditions.35 The prepared N-oxides were used as guests in the binding studies and kinetic experiments of cage 1 (vide infra).

Synthetic Precursors for the Self-Assembly of Cage 1

Tetra-amine tetra-ester aryl-extended calix[4]pyrrole 2 (AE-C[4]P 2) was prepared using a procedure reported by us.36 Terephthalaldehyde 3 was commercially purchased.

Results and Discussion

Self-Assembly of [2 + 4] Octa-imine Cage 1

Octa-imine Cage 1

We dissolved freshly distilled terephthalaldehyde 3 (0.25 mL 8.8 mM, 2.2 equiv) and tetra-amine tetra-ester calix[4]pyrrole 2 (0.25 mL 4 mM, 1 equiv) in a CDCl3:CD3CN 9:1 solvent mixture (0.5 mL) containing 1,3,5-trimethoxybenzene as internal standard (i.s.). The resulting solution was placed in a J Young NMR tube and analyzed at different time intervals using 1H NMR spectroscopy (Figure 1). The initial 1H NMR spectrum of the mixture showed sharp and well-defined signals corresponding to the hydrogen atoms of the unreacted starting materials (Figure 1a). Following the addition of 0.5% acetic acid, the 1H NMR spectrum of the mixture evidenced the appearance of broad signals and the concomitant decrease in intensity of the signals corresponding to the tetra-amine AE-C[4]P 2 and the bis-aldehyde 3 (Figure 1b). This observation suggested the formation of oligomeric species linked by imine bonds of unknown stoichiometry and ill-defined structures. The solution mixture was allowed to stand at r.t. After 36 h, the 1H NMR spectrum of the mixture revealed the appearance of a new set of sharp and well-defined proton signals (Figure 1c). We did not detect any of the signals corresponding to the protons of the tetra-amine AE-C[4]P 2. In contrast, we observed residual signals of the protons of bis-aldehyde 3 (H7 and H8).37

Figure 1.

Figure 1

Open in a new tab

Selected regions of the 1H HMR spectra (400 MHz, 298 K, CDCl3:CD3CN 9:1 mixture) of solutions containing (a) tetra-amine AE-C[4]P 2 (2 mM) and terephthalaldehyde 3 (4.4 mM); (b) same mixture of (a), following the addition of 0.5% mol of acetic acid; (c) same mixture of (b) after 36 h. See Scheme 1 for the proton assignment. Hydrogen atoms in cage 1 are depicted in bold format for clarity. i.s. internal standard.

The new set of proton signals agreed with the quantitative self-assembly of the octa-imine cage 1, displaying an apparent D4h symmetry (Scheme 1). We attributed the singlet resonating at δ = 8.4 ppm to the imine protons (H5). The pyrrole NHs (H2) appeared as a broad singlet centered at δ = 7.6 ppm. The aromatic protons of the spacer panels (H6) produced a singlet at δ = 7.7 ppm, while those of the meso-aryl substituents (H3 and H4) gave two ortho-coupled doublets, at δ = 7.0 and 7.1 ppm. A 2D ROESY experiment evidenced the existence of several cross-peaks due to close intramolecular contact in agreement with the octa-imine cage 1 structure and the proton assignment (Figure S3). Octa-imine cage was characterized by a complete set of high-resolution spectra (NMR and HRMS, Figures S1–S5).

Using the integral values of the internal standard (i.s.) as a reference, we determined that the self-assembly of the octa-imine cage 1 proceeded in almost quantitative yield (>90%).38 Previously, we showed that in the solid-state, structurally related mononuclear Pd(II) and Pt(II) C[4]P cages self-assembled in the same solvent mixture by including acetonitrile molecules in their cavities.39 The cone conformation of a “four wall” AE-C[4]P represents a perfect fit for including an acetonitrile molecule. The nitrogen atom of the included acetonitrile forms four convergent hydrogen bonds with the pyrrole NHs of the C[4]P cone conformation. Molecular modeling studies showed that for the octa-imine cage 1, one CD3CN molecule could be included in each one of its two C[4]P hemispheres, resulting in the solvate cage complex (CD3CN)2⊂1 having its middle aromatic cavity collapsed.40

Next, we studied the self-assembly of the octa-imine cage 1 in CDCl3 solution using 0.5 equiv of 4,4-(buta-1,3-diyne-1,4-diyl)bis-pyridine-N-oxide 8 as a template. The 1H NMR spectrum of a solution containing a mixture of tetra-amine 2 (1 equiv), bis-aldehyde 3 (2.2 equiv), and template 8 (0.5 equiv) was acquired approximately 1 h following its preparation. The spectrum showed sharp and well-defined proton signals (Figure S8) diagnostic of the quantitative self-assembly of the 8⊂1 cage inclusion complex (>90%). We characterized the encapsulation complex by a complete set of high-resolution spectra (NMR and HRMS, see SI). A 1H DOSY NMR experiment in a 9:1 CDCl3:CD3CN solvent mixture assigned the same diffusion constant value to the protons of the cage and the included guest, D = 3.10 ± 0.06 × 10–10 m2 s–1, evidencing their involvement in the same encapsulation complex (Figure S9). Moreover, the diffusion constant value coincided with that of the free octa-imine cage 1 in the same solvent mixture (Figure S5).

Single crystals suitable for X-ray diffraction grew from the chloroform solution containing the 8⊂1 complex (Figures 2 and S10). In the solid state, all imine bonds of the 8⊂1 complex showed E-conformation. The imine protons were arranged in pairs facing one another in the two C[4]P hemispheres of 1. This arrangement defined two differently sized portals for cage 1. The aromatic spacers were involved in CH-π interactions, with the two adjacent panels displaying alternate edge-to-face orientations.

Figure 2.

Figure 2

Open in a new tab

Side views of the X-ray crystal structure of the octa-imine 8⊂1 complex showing the two differently sized portals. Cimine–Cimine distance is shown for each portal. CH−π interactions between the edge-to-face oriented central aromatic panels are depicted. Bound guest is omitted for clarity. Thermal ellipsoids for C, N, and O atoms are set at 50% probability. H atoms are shown as spheres of 0.30 Å in diameter. To see the structure with the included guest, see Figure S10 in the Supporting Information.

Binding Studies of the Octa-imine Cage 1 with Pyridine-N-oxide Derivatives

As exemplified by the solid-state structure of the 8⊂1 complex, receptors based on AE-C[4]P units bind pyridine-N-oxide derivatives with high affinity in chlorinated solvents.30,36

Molecular modeling studies (Scigress, MM3) showed that pyridine-N-oxide derivatives 4a and 5 were good fits for the polar aromatic cavity of the octa-imine cage 1. The structures of the 2:1 homocomplexes, (4a)2⊂1 and (5)2⊂1, and the heterocomplex (4a·5)⊂1 were energy-minimized using DFT calculations at the RI41−43-BP8641-D3BJ44,45/def-SV(P)46,47 level of theory as implemented in Turbomole 7.0.48,49 The results of the quantum calculations produced sensible structures for all the inclusion complexes (see Figures S68, S69, and S71). In the (4a·5)⊂1 heterocomplex, the para-substituted reacting groups (azido and ethynyl) of the pyridine-N-oxides were located at a reasonable distance to engage in a 1,3-dipolar cycloaddition reaction (Figure S72).

Azido Pyridine-N-oxide 4a and 4-Ethynyl pyridine-N-oxide 5

First, we probed the binding properties of the octa-imine cage 1 with azido pyridine-N-oxide guest, 4a, by means of 1H NMR spectroscopic titrations (Scheme 2). Adding 1 equiv of the azido N-oxide 4a to a 2 mM solution of octa-imine 1 in CDCl3:CD3CN 9:1 solvent mixture induced the appearance of two new sets of signals for the protons of cage 1. We attributed these new sets of signals to bound 1 in two inclusion complexes of different stoichiometry. The proton signals of the free octa-imine 1 were still detected in the 1H NMR spectrum of the mixture (Figure S29).

Adding more than 1.0 equiv of 4a provoked a noticeable increase in the intensity of one of the two new sets of signals and the disappearance of that assigned to free 1. When 2.0 equiv of 4a were added, the 1H NMR spectrum of the solution showed an exclusive set of sharp and well-defined proton signals for 1 (Figure 3a). The pyrrole NHs resonated at δ = 9.54 ppm as a broad singlet. The downfield shift experienced by the NHs was indicative of their involvement in hydrogen bonds. The spacer’s imine and aromatic protons appeared as two sharp singlets centered at δ = 8.06 and δ = 7.74 ppm, respectively. Moreover, the proton signals of 4a resonated as two ortho-coupled doublets at δ = 5.77 and 4.96 ppm, (Δδ = −3.14 and −1.13 ppm, respectively). The upfield shifts experienced by the pyridine-N-oxide protons confirmed its inclusion in the polar C[4]P aromatic cavity defining the hemispheres of cage 1. The chemical shift changes and the number of proton signals were consistent with the quantitative formation of the (4a)2⊂1 complex featuring D4h symmetry (see energy minimized structure in Figure S68). The exclusive observation of the proton signals of the (4a)2⊂1 complex after adding 2.0 equiv of 4a allowed us to estimate its binding constant as larger than 108 M–2.50

Figure 3.

Figure 3

Open in a new tab

Selected regions of the 1H NMR spectra of (a) 1:2 mixture of 1:4a; (b) 1:2 mixture of 1:5; and (c) 1:2:2 mixture of 1:4a:5. Proton assignments in blue and red correspond to 1:1 and 2:1 homocomplexes with 4a and 5, respectively. Primed protons in black correspond to ternary heterocomplex (4a·5)⊂1 and protons in black to the free guest 5. The disproportionation equilibrium of homocomplexes (4a)2⊂1 and (5)2⊂1 into the heterocounterpart (4a·5)⊂1 is shown on the top.

Taken together, the results described above indicated that cage 1 included the azido pyridine-N-oxide 4a in its polar cavity, producing two inclusion complexes of 1:1, 4a⊂1, and 2:1, (4a)2⊂1, stoichiometry. In the 1:1 complex, 4a is likely coincluded with one acetonitrile molecule: (4a·CH3CN)⊂1.

We used the integral values of selected proton signals of the three species involving cage 1 to determine their concentrations’ ratios throughout the 1H NMR titration. Combining the experimentally determined concentrations’ ratios with the theoretical speciation profiles produced by the HySS2009 software, we estimated the stepwise macroscopic binding constant values as K[(CH3CN)2⊂1 + 4a ⇌ (4a·CH3CN)⊂1] = 2.0 × 104 M–1; K[(4a·CH3CN)⊂1 + 4a ⇌ (4a)2⊂1] = 3.1 × 104 M–1. This result assigned a reduced positive cooperativity to the second binding process.

In the same line, the observation of a single binding isotherm in the isothermal titration calorimetry (ITC) experiments (Figure S30) confirmed that the binding process of including two copies of 4a in the polar cavity of 1 did not feature significant levels of cooperativity, as anticipated from the 1H NMR titration results (α = (2 × K1:1⇌2:1)/(K1:1/2) = 6).51,52

We performed analogous 1H NMR and ITC titration experiments with 4-ethynyl pyridine-N-oxide 5. As before, we estimated the values for the stepwise binding constants, K[(CH3CN)2⊂1 + 5 ⇌ (5·CH3CN)⊂1] = 3.2 × 104 M–1; K[(5·CH3CN)⊂1 + 5 ⇌ (5)2⊂1] = 1.2 × 103 M–1, by comparing experimental and theoretical speciation profiles of the 1H NMR titration experiments using HySS2009 software (vide supra). From them, we derived that the binding process showed a small negative cooperativity factor, α = (2 × K1:1⇌2:1)/(K1:1/2) = 0.15. The ITC experiments conducted to accurately characterize the binding process of 1 and 5 in a CHCl3:CH3CN 9:1 solvent mixture yielded a single sigmoidal binding isotherm (Figure S35). The inflection point was centered at a 5/1 molar ratio of 2. The experimental data fit provided an average value for the binding constant of 4.5 × 104 M–1 for the two binding events. Most likely, the negative cooperativity calculated from the NMR titration data is too subtle to be detected by ITC experiments.

Pairwise Inclusion of Pyridine-N-Oxides 4a and 5 in Cage 1

We explored the formation of the ternary heterocomplex (4a·5)⊂1 by adding 2 equiv of each pyridine-N-oxide, 4a and 5, to a 2 mM solution of the octa-imine cage 1 in a 9:1 CDCl3:CD3CN solvent mixture. The 1H NMR spectrum of the solution evidenced, in the downfield region, the diagnostic signals of the imine and pyrrolic NH protons of the (4a)2⊂1 complex. However, we did not detect the corresponding signals for the (5)2⊂1 complex and the 1:1 complexes (4a·CH3CN)⊂1 and (5·CH3CN)⊂1. We observed a new set of split NH and imine signals that we assigned to the protons of the (4a·5)⊂1 ternary heterocomplex featuring two chemically nonequivalent hemispheres (Figure 3). Using the integral values of the NH signals of the two detected cage complexes, we estimated their concentrations to be [(4a)2⊂1] ∼ 0.9 mM and [(4a·5)⊂1] ∼ 0.8 mM. As expected, we also observed signals corresponding to the protons of the excess free N-oxides 4a and 5. Using the previously estimated constant values for the 1:1 and 2:1 homoinclusion complexes of 1 with 4a and 5 and the HySS2009 software (vide supra), we simulated the theoretical speciation profile of the mixture. The theoretical simulation agreed with the experimental speciation profile of the observed inclusion complexes when a binding constant value of β[(CH3CN)2⊂1 + 4a + 5 ⇌ (4a·5)⊂1] = 2.8 × 108 M–2 was assigned to the ternary heteroinclusion complex (see Figures S36–S38).

In summary, the ternary complexes (4a)2⊂1 and (4a·5)⊂1 are almost 1 order of magnitude more stable than the (5)2⊂1 counterpart (Table 1). This produced the exclusive observation of the two former complexes and the lack of detection of the latter in the presence of an excess of the N-oxides. It also resulted in an equilibrium constant for the disproportionation process of homocomplexes into the heterocounterpart of Kdispro= 3. The kinetics of the formation of the inclusion complexes of the pyridine-N-oxides and cage 1 are fast on the human time scale (i.e., seconds/mins). However, their binding processes showed slow kinetic on the 1H chemical shift time scale (i.e., separate signals for free and bound binding partners, Figure 3).

Table 1. Stepwise and Overall Binding Constant Values (K1:1, M–1; K1:1⇌2:1,M–1; β2:1 = K1:1 × K1:1⇌2:1, M–2) Determined for the 1:1 and 2:1 Homo- and Heteroternary Complexes of the Octa-imine Cage 1 with the Pyridine-N-oxide Derivatives 4a, 4b, 4c, and 5 as Guestsa.
host 1st guest K1:1 (M–1)b 2nd guest K1:1⇌2:1 (M–1)c β2:1 (M–2)d
1 4a 2.0 × 104 4a 3.1 × 104 6.2 × 108
1 5 3.2 × 104 5 1.3 × 103 4.0 × 107
1 4b 4.0 × 105 4b 4.0 × 103 1.6 × 109
1 4c 4.0 × 105 4c 1.0 × 104 4.0 × 109
1 4a 2.0 × 104 5 1.4 × 104 2.8 × 108
1 5 3.2 × 104 4a 8.8 × 103 2.8 × 108
1 4b 4.0 × 105 5 5.0 × 102 2.0 × 108
1 5 3.2 × 104 4b 6.3 × 103 2.0 × 108
1 4c 4.0 × 105 5 3.3 × 102 1.3 × 108
1 5 3.2 × 104 4c 4.1 × 103 1.3 × 108
Open in a new tab
a

Binding constant values were determined by fine-tuning theoretical speciation profiles produced using HySS2009 software to the experimentally measured concentration ratio of the different species detected in the 1H NMR spectra registered during spectroscopic titrations. For the formation of 1:1 and 2:1 homocomplexes, we use a binding model considering two reagents, (CH3CN)2⊂1, and G = 1st Guest = 2nd Guest forming two species (G·CH3CN)⊂1, and (G·G)⊂1. For the formation of the 2:1 heterocomplexes, we use a binding model considering three reagents, (CH3CN)2⊂1, G1 = 1st Guest, and G2 = 2nd Guest forming five species: (G1·CH3CN)⊂1, (G2·CH3CN)⊂1, (G1·G1)⊂1, (G2·G2)⊂1, and (G1·G2)⊂1. In the latter binding model, the constant values of the 1:1 and 2:1 homocomplexes were fixed to those determined with the former.

b

The K1:1 constants correspond to the binding equilibrium producing the 1:1 complex from the host solvate: (CH3CN)2⊂1 + 1st Guest ⇌ (1st Guest ·CH3CN)⊂1.

c

The K1:1⇌2:1 constants refer to the stepwise equilibrium for the formation of the 2:1 complexes from the 1:1 counterparts: (1st Guest · CH3CN)⊂1 + 2nd Guest ⇌ (1st Guest · 2nd Guest)⊂1.

d

β2:1 = K1:1 × K1:1⇌2:1, stands for the overall binding constant of the equilibrium producing the 2:1 complexes from the host solvate: (CH3CN)2⊂1 + 1st Guest + 2nd Guest ⇌ (1st Guest · 2nd Guest)⊂1.

Monitoring the 1,3-Dipolar Cycloaddition Reaction of 4a with 5 Included in 1

We used 1H NMR spectroscopy to analyze, at different time intervals and during 2 weeks, the solution containing the almost equimolar mixture of cage complexes (4a·5)⊂1 and (4a)2⊂1. We did not detect noticeable changes in the acquired 1H NMR spectra during this time. This result indicated that the cycloaddition reaction between 4a and 5 was not accelerated to a detectable extent by inclusion in the octa-imine cage 1 for 2 weeks under these experimental conditions.

Intending to have available the 1H NMR spectrum of the 6a⊂1 complex, corresponding to the 1,4-disubstituted triazole isomer of the cycloaddition reaction of 4a with 5 included in cage 1, we performed a solid–liquid extraction experiment of 6a using a 2 mM solution cage 1 in a 9:1 CDCl3:CD3CN solvent mixture.53 After sonication and filtration, the 1H NMR spectrum of the obtained solution did not show any sharp signal that could be assigned to the protons of bound 6a. Moreover, some proton signals of cage 1 broaden beyond detection (Figure S39).

We hypothesized that the length of the bis-N-oxide 6a was too short to span the gap between the two C[4]P cage’s hemispheres of 1 and produce a ditopic interaction with optimal hydrogen-bonding distances. To support this hypothesis, we optimized the structure of the two possible isomers of the 6a⊂1 complex at the RI41−43-BP8641-D3BJ44,45/def-SV(P)46,47 level of theory using Turbomole v7.0 (6aC⊂1 and 6aN⊂1 in Figure S72).48,49 The computed energies were similar for both isomers (ΔE < 0.2 kcal·mol–1). For any of the two isomers, the energy-minimized structure showed that the N–O···N–H hydrogen bonding distances of its two hemispheres were significantly different (3.1 and 3.9 Å, Figure 4a).

Figure 4.

Figure 4

Open in a new tab

Energy minimized structure (DFT) of complexes (a) 6aN⊂1 and (b) 6b⊂1. The N–O···N–H hydrogen-bonding distances of its two hemispheres are depicted.

This result suggested that reaching an energetically favorable transition state (TS) for the 1,3-dipolar cycloaddition reaction between 4a and 5 included in 1 would require that at least one of the pyridine-N-oxides knobs, holding the reacting groups, moves toward the other. In doing so, the N-oxide knob will be displaced from its optimal binding geometry. The distance of the hydrogen bonding, π–π, and CH−π interactions of the N-oxide with the AE-C[4]P unit will be elongated, producing a concomitant increase in the energy of the complex that will be translated to that of the TS. We consider that these geometric requirements might serve to explain why the cycloaddition between 4a and 5 is not significantly accelerated when included in 1.

This reasoning also finds that the reduced flexibility of cage 1 cannot cope with the inadequate complementarity of sizes (i.e., length) between the cage’s cavity and guest 6a to produce an optimal ditopic interaction nor with approaching the included reactants 4a and 5 without disrupting their intermolecular interactions with the container.

Study of the 1,3-Dipolar Cycloaddition Reaction of 4b with 5 Included in 1

We considered that 4-azido(methyl) pyridine-N-oxide, 4b, featuring a methylene unit between the azido group and the para-carbon of the pyridine-N-oxide, should locate the two reacting groups in the heterocomplex (4b·5)⊂1 in closer spatial proximity compared to the (4a·5)⊂1 analog described in the previous section. For the same token, the bis-N-oxide 6b, the 1,4-disubstituted triazole isomer of the cycloaddition reaction of 4b and 5, should be a superior fit for the dimensions of the cavity of cage 1 than the homologous 6a (Figure 4b). First, we experimentally investigated the inclusion of 6b in cage 1. The 1H NMR spectrum of an equimolar mixture of 6b and 1 produced sharp and well-defined proton signals diagnostic of the quantitative formation of the 6b⊂1 complex, featuring C4 symmetry (Figures S45–S47).54

With this information, we evaluated the pairwise inclusion of N-oxides 4b and 5 in the octa-imine cage 1 and their putative cycloaddition reaction once included. We prepared a 2 mM solution of 1 and added 1 equiv of the new dipole 4b and 2 equiv of the previously used dipolarophile 5. Owing to the larger thermodynamic stability of the (4b)2⊂1 and aiming to increase the concentration of the heteroternary complex (4b·5)⊂1, we used a 1:1:2 molar ratio of 1, 4b, and 5, instead of the 1:2:2 used for the assembly of the (4a·5)⊂1.55

The 1H NMR spectrum of the solution acquired following the preparation of the mixture showed the homoternary complexes (5)2⊂1 and (4b)2⊂1 as the major species (Figures 5a and S42). We also detected, to a much-reduced extent, the proton signals of the 1:1 complexes: (5·CH3CN)⊂1 and (4b·CH3CN)⊂1. We attributed an additional set of signals of cage 1 to the protons in the heteroternary (4b·5)⊂1 complex. Using integral values of selected proton signals, we quantified its concentration as [(4b·5)⊂1] ∼ 0.54 mM. Finally, we also identified proton signals of low intensity for the free cage 1, and the free N-oxides 4b and 5. The theoretical simulation (HySS2009) of the experimentally observed speciation profile assigned a stability constant value of β[(CH3CN)2⊂1 + 4b + 5 ⇌ (4b·5)⊂1] = 2.0 × 108 M–2 to the heteroternary complex (see Figure S44).

Figure 5.

Figure 5

Open in a new tab

Selected regions of the 1H NMR spectra corresponding to the kinetic study of the formation of the 6b⊂1 complex starting from a 1:1:2 mixture of 1:4b:5. (a) 1:1:2 mixture of 1:4b:5 (t = 0 h); (b) 1:1:2 mixture of 1:4b:5 (t = 8 h); and (c) 1:1:2 mixture of 1:4b:5 (t = 48 h). The percentage of bound species at t = 0 h is indicated in spectrum (a). Proton assignments in blue and red in spectrum (a) correspond to complexes with 5 and 4b, respectively. Primed protons in black correspond to the 6b⊂1 complex. The scheme of the 1,3-dipolar cycloaddition reaction between 4b and 5 inside the cage yielded complex 6b⊂1 is shown on the top.

We analyzed the changes in the solution with time using 1H NMR spectroscopy. After 30 min, we observed the emergence of a new set of proton signals of cage 1. This was especially clear in the upfield spectrum region where two new pyrrole NHs signals centered at δ = 9.2 and 8.9 ppm appeared. The intensity of the latest set of signals increased with time at the expense of those of the homocomplexes (4b)2⊂1 and (5)2⊂1. Moreover, the intensity of the proton signals of the heterocomplex (4b·5)⊂1 remained almost constant in the initial phase of the monitoring process (∼0.54 mM). After 48 h, we exclusively observed the set of signals of the newly formed species (Figure 5c), which coincided with those of the previously prepared (6b)⊂1 complex. Taken in concert, these results indicated that the 1,3-dipolar cycloaddition between 4b and 5 occurred when included in cage 1. Furthermore, it only produced the 1,4-triazole isomer 6b.56 We performed analogous kinetic experiments using different molar ratios of 1:4b:5 (1:1:1 and 1:5:5), producing the ternary complex (4b·5)⊂1 in lower concentrations (0.32 mM and 0.1 mM, respectively). We used the changes in the integral values of selected proton signals of the 6b⊂1 complex to determine its changes in concentration with time. First, we used the initial rates method to unequivocally assign to the included 1,3-dipolar cycloaddition reaction between 4b and 4a a first-order rate law concerning exclusively the concentration of the heteroternary complex, υintra_ini = d[(6b⊂1]/dt=kintra [(4b·5)⊂1]. We derived an average value for the reaction’s rate constant, k(6b-intra) = 5 (±3) × 10–5 s–1, by linear fitting the kinetic data at different concentrations.57

Second, we constructed an elaborated kinetic model considering all the binding equilibria in the solution and the irreversible 1,3-dipolar cycloaddition reaction of 4b and 5 occurring in 1 (Figure 6). The contribution from the uncatalyzed reaction (bulk) was not considered in the model (i.e., negligible at this concentration). Neither did we consider the dissociation of the cycloaddition product complex to give free 6b in solution (K[(CH3CN)2⊂1 + 6b ⇌ (6b·CH3CN)⊂1] > 108 M–1). This model will help derive rate constant values following the included cycloaddition reaction to completion and by a best-fit computer simulation (COPASI) of the kinetics over the whole reaction curve. We assumed that the rate constants of the equilibrium steps were faster than that of the irreversible included reaction.58 We manually fixed the kon/koff ratios of the binding equilibria to the determined association constants values of the formed complexes, K = kon/koff.59

Figure 6.

Figure 6

Open in a new tab

(Top) Theoretical kinetic model used for the nonlinear mathematical analysis of the experimental data. (Bottom) Changes in the concentration of the 6b⊂1 complex (empty dots) with time starting from a 1:1:1 mixture of 1:4b:5 in CDCl3:CD3CN 9:1. The solid line represents the fit of the kinetic data to the theoretical model using the parameter estimation module of COPASI Software Version 4.25. All binding equilibria’s kon/koff ratios were manually fixed based on the determined binding constants. k(intra) was the only variable parameter used for the fit.

The experimental kinetic data obtained for the complete formation of the 6b⊂1 complex at different molar ratios of 1:4b:5 showed an excellent fit to the elaborated kinetic model, producing an average rate constant for the included cycloaddition reaction of k(6b-intra) = 7 (±5) × 10–5 s–1 (Figures S55–S57). This value coincided with the one previously derived using the initial rates method.

In assessing the acceleration factor caused by including the cycloaddition reaction between 4-ethynyl pyridine-N-oxide 5 and 4-azido(methyl) pyridine-N-oxide 4b in the octa-imine 1, we determined the reaction’s rate constant in the bulk solution. To this end, we reacted the substrates at 25 mM concentration in a 9:1 CDCl3:CD3CN solvent mixture at 298 K. The high concentration used to carry out the bulk reaction and its low reaction rate required HPLC to monitor its progress. We used 4-methyl pyridine-N-oxide as the internal standard (i.s.) (Figures S59–S65 and Table S1).60 Under these conditions, the cycloaddition reaction produced a mixture of the 1,4-, 6b, and 1,5-, 7b, triazoles in approximately 2:1 molar ratio (Scheme 2). We fit the experimental data to a theoretical kinetic model, which considered the irreversible bimolecular reaction between 5 and 4b to produce isomer 6b using COPASI. The fit returned a value for k(6b-bulk) of 5.6 × 10–8 M–1 s–1.61 Considering this result, at 2 mM concentration of reactants, the 1,3-dipolar cycloaddition reaction between 5 and 4b in the bulk solution will require ∼317 years to produce a 1 mM concentration of 6b. For this reason, we considered the amount of 6b produced by the background reaction negligible and neglected its inclusion in the elaborated kinetic binding model. The analysis of the kinetics of the cycloaddition reaction between 4a and 5 in the bulk solution yielded a rate constant k(6a-bulk) of similar magnitude (see Table 2).

Table 2. Rate Constant Values for the 1,3-Dipolar Cycloaddition Reactions in Bulk (kbulk, M–1 s–1) and the Octa-imine Cage Cavity (kintra, s–1)a.

guest pair product kbulk (M–1 s–1)b kintra (s–1)c EM (M)d
(4a·5) 6a 5.1 × 10–8 n.d. n.d.
(4b·5) 6b 5.6 × 10–8 5.0 × 10–5 ∼103
(4c·5) 6c 3.6 × 10–8 8.1 × 10–5 ∼103
Open in a new tab
a

Acceleration factors reported as effective molarities (EM, M) are also listed.

b

Determined by best-computer fit of the kinetic experimental data using COPASI and a kinetic theoretical model for an irreversible first-order bimolecular reaction. Experimental data were obtained from the changes in concentration of the product, 6a–c, with time for the reactions between 4a–c with 5 at 298 K. The initial concentration of reactants was 25 mM in a 9:1 CDCl3:CD3CN solvent mixture. We used HPLC and calibration curves to accurately determine the concentrations of 6a–c.

c

Determined by best-computer fit of the kinetic experimental data using COPASI and an elaborated kinetic model considering six binding equilibria in the solution, producing the 1:1 and 2:1 homocomplexes, as well as the ternary (Michaelis) complex, and the irreversible pseudounimolecular reaction of the ternary complex producing the product bound in the octa-imine cage. Experimental kinetic data were derived from the concentration changes with the time of the complex that resulted from including the reaction product in the container cage. We used 1H NMR spectroscopy to monitor the reaction starting from a mM equimolar mixture of 1, 4, and 5 dissolved in a CDCl3:CD3CN 9:1 solvent mixture.

d

EM = kintra/kbulk; n.d. not determined. We did not detect noticeable changes in the 1H NMR spectra acquired during 2 weeks for the 1:1:1 mixture of 1, 4a, and 5.

We assessed the rate acceleration factor provoked by the inclusion of the cycloaddition reaction between 5 and 4b in the cavity of cage 1, determining its effective molarity, EM = k(6b-intra)/k(6b-bulk) = 1.0 (±0.2) × 103 M. This EM value can be assigned to an entropy factor of −13.7 e.u. or an energy difference of 4.1 kcal·mol–1 in favor to the transition state (TS) of the cycloaddition reaction between 5 and 4b included in 1 compared to that in the bulk solution. This result agreed with the hypothesis of Page and Jenks, stating that enzymes might carry out a significant fraction of their extraordinary rate accelerations, compensating reactions’ entropy costs by substrates’ binding and confinement in the active site, without the need to invoke other concepts or the enthalpic stabilization of the TS.3 A reaction’s acceleration factor was also assessed from the ratio of initial rates, v0(6b-intra)/v0(6b-bulk), in the presence and absence of cage 1, respectively. We calculated the hypothetical initial reaction rate of the reaction in bulk, v0(6b-bulk), at 2 mM concentrations of 4b and 5 using the determined value of k(6b-bulk) (see Table S1). We took the value of v0(6b-intra) from the reaction performed using 2 mM concentrations of 5, 4b, and 1. The ratio of initial rate values returned an acceleration factor of 104, which cannot be converted into an energy difference of the TS but provided a physical magnitude of the effect caused by the reaction inclusion on its rate.

Unfortunately, as with other reactions accelerated by inclusion in molecular containers, the tight binding of the product, bis-N-oxide 6b, to cage 1 inhibited turnover.

We also performed a series of control experiments (see Supporting Information section 5) to verify that the coinclusion of 4b and 5 in the cavity of cage 1 unequivocally caused the acceleration of their dipolar cycloaddition. For example, the 1,3-cycloaddition reaction of azido/ethynyl guest molecules that do not simultaneously fit into the cage’s cavity do not form the ternary heterocomplex (e.g., azido(methyl) pyridine-N-oxide 4b and 4-(4-ethynylphenyl) pyridine-N-oxide 10), is not accelerated to a measurable extent (see Supporting Information, Table S2).

Study of the 1,3-Dipolar Cycloaddition Reaction of 4c with 5 Included in 1

We also studied the cycloaddition reaction between azido(ethyl) pyridine-N-oxide 4c, a dipole featuring two methylene units between the azido group and the para-carbon of the pyridine-N-oxide, and the same dipolarophile 5. The regioselective reaction produced exclusively the 1,4-triazole isomer, 6c, in the 6c⊂1 complex. We assessed the EM reaction’s acceleration factor as ∼103 M. We did not expect that the cycloaddition reactions of 4b with 5 and of 4c with 5, included in 1, experienced EM acceleration factors of the same magnitude. The TS of the reaction between 4c and 5, when included in 1, required locking an additional single bond rotation compared to that of 4b with 5, which should be associated with an additional entropy cost and the corresponding increase of the TS barrier. However, from the obtained result, we concluded that 4c did not experience low-energy conformational changes within the heteroternary (4c·5)⊂1 complex, contributing negatively to the reaction’s acceleration.

Comparison of the Obtained Results with Previous Examples of Accelerating 1,3-Dipolar Cycloadditions by Inclusion in Molecular Containers

What is unique about octa-imine 1, is its sizable polar cavity equipped with two convergent and endohedrally functionalized binding sites, AE-C[4]P. These characteristics allowed the tight-binding and pairwise inclusion of two pyridine-N-oxides in a well-defined and fixed orientation. The arrangement of some of the included N-oxides provoked a suitable alignment of their para-substituents in a chemical reaction. We selected the 1,3-dipolar Huisgen cycloaddition reaction as a benchmark to investigate the acceleration caused by including the reacting groups in the cavity of 1. As mentioned in the introduction, we were not the first to examine the acceleration effect resulting from including a 1,3-dipolar cycloaddition of azide and alkyne in a molecular container. For example, Rebek and Jia studied the inclusion of the cycloaddition reaction of phenyl azide with phenylacetylene in a resorcin[4]arene dimer.27 The reaction was regioselective, and an EM acceleration factor of 1.2 × 102 M was later determined.25 The dimer had an aromatic cavity suitable for coincluding the two reactants. However, it lacked inner functional groups to control the relative positioning of the substrates, the reacting groups’ orientation, and the substrates’ binding affinity. As is the case here, the direct observation of the ternary-hetero complex simplified the kinetic data analysis.

We assigned the increase of an order of magnitude of the EM (1.0 × 103 M) measured for related dipolar cycloaddition reaction included in cage 1 to the tighter binding of the pyridine-N-oxide knobs of the substrates to the polar AE-C[4]P defining the hemispheres of the container. The N-oxides are located in fixed positions in the container’s cavity, owing to the formation of four convergent hydrogen bonds between their oxygen atoms and the pyrrole NHs. Their relative motions within the heteroternary complex are reduced, and the reacting groups are adequately oriented. The fixed position of the N-oxide substrates is evidenced by the lack of reactivity in the cycloaddition reaction of 4a and 5 included in 1.

Mock and co-workers24 used CB[6] as a molecular container to accelerate the cycloaddition reaction of azido and alkynyl N-tert-butylated substrates in water solution. The reaction was regioselective, but the ternary complex was not experimentally observed. This raised the problem of “non-productive binding”, that is, the existence of a heteroternary complex in which the second substrate had the reactive group in the exterior of CB[6], evidencing that it was not suitable to unequivocally control the reacting groups’ orientation as cage 1 does. All these considerations complicated the kinetic analysis and led to some internal data inconsistency. For this reason, the authors referred to the reported “acceleration factor” as an approximation. As for the case of Rebek and Jia, an EM acceleration factor was also estimated a few years later by Mandolini and co-workers to be 1.6 × 104 M.25 We consider that the Mock example represents a particular case of reaction acceleration induced by inclusion in a container. The substrates are charged, not neutral. Hence, the container neutralizes the repulsive Coulombic interaction occurring in the bulk. Due to its limited size, the container does not include the reactants, only the reacting groups. In short, the results presented here constitute the most significant acceleration reported for a bimolecular reaction included in a molecular container by directly detecting the ternary complex.

Conclusions

We self-assembled and characterized a [4 + 2] octa-imine calix[4]pyrrole cage in a CDCl3:CD3CN 9:1 solvent mixture. Adding 0.5% of acetic acid or a bis-pyridine-N-oxide template molecule significantly improved the yield of the self-assembly reaction (>90%). Octa-imine cage 1 formed thermodynamically and kinetically stable 1:1 and 2:1 homo- and heterocomplexes with pyridine-N-oxide guests featuring an azido (4a, 4b, 4c) or ethynyl substituent (5) in the para- position. Octa-imine cage 1 was found to accelerate the included 1,3-dipolar cycloaddition reactions between 4-azido(alkyl) pyridine-N-oxides, 4b and 4c, with 4-ethynyl pyridine-N-oxide 5. The ternary Michaelis inclusion complexes, (4b/4c·5)⊂1, were detected in solution. The calculated EMs (∼103 M) of the reactions included in octa-imine 1 are one figure larger than the one determined for an analogous reaction in Rebek’s resorcinarene dimer. We attributed our results to the tighter binding and fixed (and well-oriented) position of the reacting substrates within the polar container, reducing the entropy cost needed to achieve the TS.

Surprisingly, the included 1,3-cycloaddition of 4-azido pyridine-N-oxide 4a and 4-ethynyl pyridine-N-oxide 5 was not accelerated to a measurable extent under analogous conditions. Nevertheless, the (4a·5)⊂1 inclusion complex was present in solution. We hypothesized that the guests’ sizes (i.e., length) are inadequate for the productive cycloaddition reaction inside octa-imine cage 1. The fixed position of the included guests and the reduced flexibility of container 1 may not be able to handle the geometric requirements needed to achieve the TS of the included reaction. Specifically, the necessary elongation of N–O···N–H hydrogen bonds for at least one guest increases the energy barrier for achieving the TS.

In brief, including bimolecular 1,3-cycloaddition Huisgen reactions in cage 1 imposes geometric and steric constraints on their transition state. This explains the observed accelerations and reactions’ regioselectivity and the measured lack of acceleration in one of the included reactions.62 Currently, we are investigating the properties of 1 and its more giant analogs for the acceleration of other included bimolecular reactions. We hope to communicate our results in due time.

Acknowledgments

This research was funded by Gobierno de España MICINN/AEI/FEDER (PID2020-114020GB-I00, CEX2019-000925-S, and PID2023-149233NB-I00), the CERCA Programme/Generalitat de Catalunya, AGAUR (2021 SGR 00851), and the ICIQ Foundation. We thank Dr. Eduardo Escudero for the X-ray crystallographic data. We also thank Dirk Husstege and Mario Carratú for their help in different phases of this project.

Glossary

Abbreviations

EM

effective molarity

AE-C[4]P

aryl-extended calix[4]pyrrole

DA

Diels–Alder

ITC

isothermal titration calorimetry

TS

transition state

Data Availability Statement

All dataset collection of computational results of this manuscript is available in the ioChem-BD repository and can be accessed through this link http://dx.doi.org/10.19061/iochem-bd-1-358 CCDC: 2393702.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacsau.4c01118.

  • General information and methods, synthetic procedures, characterization of new compounds, 1H NMR titrations, ITC experiments, fit of the experimental data to the kinetic model, and DFT calculations (PDF)

  • Structure of octa-imine cage 1 complex with bis-pyridine-N-oxide guest 8 (CIF)

Author Contributions

This manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. CRediT: Yifan Li formal analysis, methodology, writing - original draft; Chiara F.M. Mirabella methodology; Gemma Aragay data curation, formal analysis, investigation, supervision, writing - original draft, writing - review & editing. Pablo Ballester conceptualization, funding acquisition, formal analysis, supervision, writing - review & editing.

The authors declare no competing financial interest.

Supplementary Material

au4c01118_si_001.pdf (4.3MB, pdf)
au4c01118_si_002.cif (5.4MB, cif)

References

  1. Richard J. P. Enzymatic Rate Enhancements: A Review and Perspective. Biochemistry 2013, 52 (12), 2009–2011. 10.1021/bi3017119. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Pauling L. Chemical Achievement and hope for the Future. Am. Sci. 1948, 36 (1), 51–58. [PubMed] [Google Scholar]
  3. Page M. I.; Jencks W. P. Entropic contributions to rate accelerations in enzymic and intramolecular reactions and the chelate effect. Proc. Natl. Acad. Sci. U.S.A. 1971, 68 (8), 1678–1683. 10.1073/pnas.68.8.1678. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Pauling L. Nature of Forces between Large Molecules of Biological Interest. Nature 1948, 161 (4097), 707–709. 10.1038/161707a0. [DOI] [PubMed] [Google Scholar]
  5. Houk K. N.; Leach A. G.; Kim S. P.; Zhang X. Binding Affinities of Host–Guest, Protein–Ligand, and Protein–Transition-State Complexes. Angew. Chem., Int. Ed. 2003, 42 (40), 4872–4897. 10.1002/anie.200200565. [DOI] [PubMed] [Google Scholar]
  6. The term “binding” is used here to mean that the enzyme transition-state complex displays a lower energy than the sum of the energies of the enzyme and the transition-state. We are aware that the “transition-state binding energy” cannot be measured directly and that the transition state of the uncatalyzed reaction may differ from that of the enzyme i.e. in composition and geometry.
  7. Wang R.; Yu Y. Site-selective reactions mediated by molecular containers. Beilstein J. Org. Chem. 2022, 18, 309–324. 10.3762/bjoc.18.35. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Liu W.; Stoddart J. F. Emergent behavior in nanoconfined molecular containers. Chem. 2021, 7 (4), 919–947. 10.1016/j.chempr.2021.02.016. [DOI] [Google Scholar]
  9. Yu Y.; Yang J.-M.; Rebek J. Molecules in Confined Spaces: Reactivities and Possibilities in Cavitands. Chem. 2020, 6 (6), 1265–1274. 10.1016/j.chempr.2020.04.014. [DOI] [Google Scholar]
  10. Kim S. P.; Leach A. G.; Houk K. N. The Origins of Noncovalent Catalysis of Intermolecular Diels–Alder Reactions by Cyclodextrins, Self-Assembling Capsules, Antibodies, and RNAses. J. Org. Chem. 2002, 67 (12), 4250–4260. 10.1021/jo011180d. [DOI] [PubMed] [Google Scholar]
  11. For examples of cyclodextrin see:Rideout D. C.; Breslow R. Hydrophobic acceleration of Diels-Alder reactions. J. Am. Chem. Soc. 1980, 102 (26), 7816–7817. 10.1021/ja00546a048. [DOI] [Google Scholar]
  12. For examples of hydrogen bonded dimeric capsules see:; a Kang J.; Rebek J. Acceleration of a Diels–Alder reaction by a self-assembled molecular capsule. Nature 1997, 385 (6611), 50–52. 10.1038/385050a0. [DOI] [PubMed] [Google Scholar]; b Kang J.; Santamaría J.; Hilmersson G.; Rebek J. Self-Assembled Molecular Capsule Catalyzes a Diels–Alder Reaction. J. Am. Chem. Soc. 1998, 120 (29), 7389–7390. 10.1021/ja980927n. [DOI] [Google Scholar]
  13. For examples of self-assembled coordination cages see:Kusukawa T.; Nakai T.; Okano T.; Fujita M. Remarkable Acceleration of Diels–Alder Reactions in a Self-Assembled Coordination Cage. Chem. Lett. 2003, 32 (3), 284–285. 10.1246/cl.2003.284. [DOI] [Google Scholar]
  14. For examples of porphyrin macrocycle see:Sanders J. K. M.; Nakash M.; Marty M.; Clyde-Watson Z.; Twyman L. J. Acceleration of a hetero-Diels–Alder reaction by cyclic metalloporphyrin trimers. Chem. Commun. 1998, 20, 2265–2266. 10.1039/a806070c. [DOI] [Google Scholar]
  15. For examples of purely organic covalent cages see:Warmuth R. First innermolecular Diels–Alder reaction of o-benzyne inside a molecular container compound. Chem. Commun. 1998, 1, 59–60. 10.1039/a703099a. [DOI] [Google Scholar]
  16. For examples of dynamically covalent cages:; a Brisig B.; Sanders J. K. M.; Otto S. Selection and Amplification of a Catalyst from a Dynamic Combinatorial Library. Angew. Chem., Int. Ed. 2003, 42 (11), 1270–1273. 10.1002/anie.200390326. [DOI] [PubMed] [Google Scholar]; b Ono K.; Niibe M.; Iwasawa N. A K+-promoted Diels–Alder reaction by using a self-assembled macrocyclic boronic ester containing two crown ether moieties. Chem. Sci. 2019, 10 (32), 7627–7632. 10.1039/C9SC01597C. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Murase T.; Horiuchi S.; Fujita M. Naphthalene Diels–Alder in a Self-Assembled Molecular Flask. J. Am. Chem. Soc. 2010, 132 (9), 2866–2867. 10.1021/ja9107275. [DOI] [PubMed] [Google Scholar]
  18. Tehrani F. N.; Assaf K. I.; Hein R.; Jensen C. M. E.; Nugent T. C.; Nau W. M. Supramolecular Catalysis of a Catalysis-Resistant Diels–Alder Reaction: Almost Theoretical Acceleration of Cyclopentadiene Dimerization inside Cucurbit[7]uril. ACS Catal. 2022, 12 (4), 2261–2269. 10.1021/acscatal.1c05659. [DOI] [Google Scholar]
  19. Huisgen R. 1,3-Dipolar Cycloadditions. Proc. Chem. Soc. 1961, 357–396. [Google Scholar]
  20. Huisgen R. Kinetics and Mechanism of 1,3-Dipolar Cycloadditions. Angew. Chem., Int. Ed. 1963, 2 (11), 633–645. 10.1002/anie.196306331. [DOI] [Google Scholar]
  21. Hein J. E.; Fokin V. V. Copper-catalyzed azide–alkyne cycloaddition (CuAAC) and beyond: new reactivity of copper(i) acetylides. Chem. Soc. Rev. 2010, 39 (4), 1302–1315. 10.1039/b904091a. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Johansson J. R.; Beke-Somfai T.; Said Stålsmeden A.; Kann N. Ruthenium-Catalyzed Azide Alkyne Cycloaddition Reaction: Scope, Mechanism, and Applications. Chem. Rev. 2016, 116 (23), 14726–14768. 10.1021/acs.chemrev.6b00466. [DOI] [PubMed] [Google Scholar]
  23. Mock W. L.; Irra T. A.; Wepsiec J. P.; Manimaran T. L. Cycloaddition induced by cucurbituril. A case of Pauling principle catalysis. J. Org. Chem. 1983, 48 (20), 3619–3620. 10.1021/jo00168a070. [DOI] [Google Scholar]
  24. Mock W. L.; Irra T. A.; Wepsiec J. P.; Adhya M. Catalysis by cucurbituril. The significance of bound-substrate destabilization for induced triazole formation. J. Org. Chem. 1989, 54 (22), 5302–5308. 10.1021/jo00283a024. [DOI] [Google Scholar]
  25. Cacciapaglia R.; Di Stefano S.; Mandolini L. Effective Molarities in Supramolecular Catalysis of Two-Substrate Reactions. Acc. Chem. Res. 2004, 37 (2), 113–122. 10.1021/ar020076v. [DOI] [PubMed] [Google Scholar]
  26. k4 = 0.023 s–1 from kinetic analysis; kuncat = 0.7 M s–1 (k4/K7 estimated second order rate)/6 × 105 (reported acceleration) = 1.16 × 10–6 M s–1; EM = 0.023 s–1/1.16 × 10–6 M s–1 = 1.6 −1.9 × 104 M.
  27. Chen J.; Rebek J. Selectivity in an Encapsulated Cycloaddition Reaction. Org. Lett. 2002, 4 (3), 327–329. 10.1021/ol0168115. [DOI] [PubMed] [Google Scholar]
  28. Carlqvist P.; Maseras F. A theoretical analysis of a classic example of supramolecular catalysis. Chem. Commun. 2007, 7, 748–750. 10.1039/B613434C. [DOI] [PubMed] [Google Scholar]
  29. Daver H.; Harvey J. N.; Rebek J. Jr.; Himo F. Quantum Chemical Modeling of Cycloaddition Reaction in a Self-Assembled Capsule. J. Am. Chem. Soc. 2017, 139 (43), 15494–15503. 10.1021/jacs.7b09102. [DOI] [PubMed] [Google Scholar]
  30. Galán A.; Escudero-Adán E. C.; Ballester P. Template-directed self-assembly of dynamic covalent capsules with polar interiors. Chem. Sci. 2017, 8 (11), 7746–7750. 10.1039/C7SC03731G. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Mirabella C. F. M.; Aragay G.; Ballester P. Influence of the solvent in the self-assembly and binding properties of [1 + 1] tetra-imine bis-calix[4]pyrrole cages. Chem. Sci. 2022, 14 (1), 186–195. 10.1039/D2SC05311J. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Adriaenssens L.; Acero Sánchez J. L.; Barril X.; O’Sullivan C. K.; Ballester P. Binding of calix[4]pyrroles to pyridine N-oxides probed with surface plasmon resonance. Chem. Sci. 2014, 5 (11), 4210–4215. 10.1039/C4SC01745E. [DOI] [Google Scholar]
  33. Kwok S. W.; Fotsing J. R.; Fraser R. J.; Rodionov V. O.; Fokin V. V. Transition-metal-free catalytic synthesis of 1,5-diaryl-1,2,3-triazoles. Org. Lett. 2010, 12 (19), 4217–4219. 10.1021/ol101568d. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Piccinno M.; Aragay G.; Mihan F. Y.; Ballester P.; Dalla Cort A. Unexpected Emission Properties of a 1,8-Naphthalimide Unit Covalently Appended to a Zn–Salophen. Eur. J. Inorg. Chem. 2015, 2015 (16), 2664–2670. 10.1002/ejic.201500258. [DOI] [Google Scholar]
  35. Siemsen P.; Livingston R. C.; Diederich F. Acetylenic Coupling: A Powerful Tool in Molecular Construction. Angew. Chem., Int. Ed. 2000, 39 (15), 2632–2657. 10.1002/1521-3773(20000804)39:15<2632::AID-ANIE2632>3.0.CO;2-F. [DOI] [PubMed] [Google Scholar]
  36. Escobar L.; Arroyave F. A.; Ballester P. Synthesis and Binding Studies of a Tetra-α Aryl-Extended Photoresponsive Calix[4]pyrrole Receptor Bearing meso-Alkyl Substituents. Eur. J. Org. Chem. 2018, 2018 (9), 1097–1106. 10.1002/ejoc.201701602. [DOI] [Google Scholar]
  37. Without the addition of acetic acid the quantitative self-assembly of the octa-imine 1 was capricious and dependent on the batch of CDCl3 used. Typically, it required extensive times (>72 h) and higher temperatures (310 K) to produce the octa-imine 1 in variable yields. We surmise that the variable amounts of acid impurities present in the different CDCl3 batches were responsible of the erratic behavior of the self-assembly.
  38. We determined a 70% yield for the assembly of the octa-imine cage 1 after heating the mixture for 72 h at 310 K in previously neutralized CDCl3:CD3CN solvent mixture (absence of a Bro̷nsted acid).
  39. Escobar L.; Escudero-Adán E. C.; Ballester P. Guest Exchange Mechanisms in Mono-Metallic PdII/PtII-Cages Based on a Tetra-Pyridyl Calix[4]pyrrole Ligand. Angew. Chem., Int. Ed. 2019, 58 (45), 16105–16109. 10.1002/anie.201909685. [DOI] [PubMed] [Google Scholar]
  40. Alternatively, forming a (CD3CN)3⊂1 inclusion complex in solution was also plausible, featuring an additional CD3CN molecule in the middle aromatic cavity (see link computational results).
  41. Perdew J. P. Density-functional approximation for the correlation energy of the inhomogeneous electron gas. Phys. Rev. B 1986, 33 (12), 8822–8824. 10.1103/PhysRevB.33.8822. [DOI] [PubMed] [Google Scholar]
  42. Eichkorn K.; Weigend F.; Treutler O.; Ahlrichs R. Auxiliary basis sets for main row atoms and transition metals and their use to approximate Coulomb potentials. Theor. Chem. Acc. 1997, 97 (1), 119–124. 10.1007/s002140050244. [DOI] [Google Scholar]
  43. Sierka M.; Hogekamp A.; Ahlrichs R. Fast evaluation of the Coulomb potential for electron densities using multipole accelerated resolution of identity approximation. J. Chem. Phys. 2003, 118 (20), 9136–9148. 10.1063/1.1567253. [DOI] [Google Scholar]
  44. Grimme S.; Ehrlich S.; Goerigk L. Effect of the damping function in dispersion corrected density functional theory. J. Comput. Chem. 2011, 32 (7), 1456–1465. 10.1002/jcc.21759. [DOI] [PubMed] [Google Scholar]
  45. Grimme S.; Antony J.; Ehrlich S.; Krieg H. A consistent and accurate ab initio parametrization of density functional dispersion correction (DFT-D) for the 94 elements H-Pu. J. Chem. Phys. 2010, 132 (15), 154104. 10.1063/1.3382344. [DOI] [PubMed] [Google Scholar]
  46. Rappoport D.; Furche F. Property-optimized Gaussian basis sets for molecular response calculations. J. Chem. Phys. 2010, 133 (13), 134105. 10.1063/1.3484283. [DOI] [PubMed] [Google Scholar]
  47. Schäfer A.; Horn H.; Ahlrichs R. Fully optimized contracted Gaussian basis sets for atoms Li to Kr. J. Chem. Phys. 1992, 97 (4), 2571–2577. 10.1063/1.463096. [DOI] [Google Scholar]
  48. TURBOMOLE V7.0 2015, a Development of University of Karlsruhe and Forschungszentrum Karlsruhe GmbH, 1989–2007, TURBOMOLE GmbH, 2007. https://www.turbomole.com.
  49. Ahlrichs R.; Bär M.; Häser M.; Horn H.; Kölmel C. Electronic structure calculations on workstation computers: The program system turbomole. Chem. Phys. Lett. 1989, 162 (3), 165–169. 10.1016/0009-2614(89)85118-8. [DOI] [Google Scholar]
  50. The addition of more than 2.0 equiv. of 4a did not produce noticeable changes in the proton signals assigned to the 2:1 (4a)2⊂1 complex. However, we observed the emergence of the proton signals of free 4a. This observation indicated that the binding equilibrium was slow on the 1H NMR chemical shift. In addition, an EXSY experiment did not show chemical exchange cross-peaks between the proton signals of the free and bound 4a guest. This result evidenced that the binding equilibrium was also slow on the EXSY time scale and that the dissociation rate constant of the complex, koff, was smaller than 10–2 s–1 (kexchange = k’ + koff ∼ 10–2 s–1).
  51. Microcal Origin Data Analysis v7.21, Malvern Instruments Limited.
  52. The normalized integrated data produced a single sigmoidal binding isotherm indicating that the binding cooperativity was not large enough to be detected by ITC. We fit the data to a theoretical binding model considering one set of sites which returned an average value for the binding constant of the two sites.
  53. Compound 6a has low solubility in CDCl3:CD3CN 9:1 solvent mixture.
  54. The energy minimized structure of the 6b⊂1 complex revealed a superior fit of the bis N-oxide guest for the cavity of 1 compared to 6a⊂1 with optimal and almost identical N–O···N–H hydrogen bonding interactions for both C[4]P hemispheres (3.0 and 3.1 Å).
  55. A comparison of experimental and theoretical speciation profiles for a 1H NMR titration experiments of 1 with incremental amounts of 4b assigned: K[(CH3CN)2⊂1 + 4b ⇌ (4b·CH3CN)⊂1] = = 3.9 × 105 M–1; K[(4b·CH3CN)⊂1 + 4b ⇌ (4b)2⊂1] = 3.9 × 103 M–1. Similarly, we determined β[(CH3CN)2⊂1 + 4b + 5 ⇌ (4b·5)⊂1] = 2.0 × 108 M–2.
  56. Most likely the bad fit of the 1,5-regioisomeric product, 7b (Scheme 2), for the cavity of 1 precluded its formation when the cycloaddition substrates are included in it (Figure S74).
  57. Lowering the concentrations of the ternary heterocomplex in solution led to proportionally slower reaction’s rates.
  58. The use of different values for kon’s of 105 to 108 times faster than the included reaction step with the corresponding koff’s adjusted to provide the measured association constant had no effect on the reaction profile observed.
  59. We set the koff for all inclusion complexes to 0.01 s–1 in agreement with the slow exchange rate constant estimated from the lack of exchange peaks in the EXSY experiment. Then kon was determined as kon = Ka × koff.
  60. The chromatogram of the cycloaddition reaction shows two peaks at 11.3 and 12.5 min which increased in intensity with time compared to that of the i.s. These were assigned to the 1,4 and 1,5-isomers of the cycloaddition products.
  61. Using initial rates method, we determined the initial reaction rate producing 6b as v0(6b-bulk) = 3.1 (±0.3) × 10–11 M s–1 by linear regression. Considering that the cycloaddition reaction was first-order for the two reactants, we determined the rate constant value to be k(6b-bulk) = v0(6b-bulk)/(0.025)2 = 5 (±0.5) × 10–8 M–1 s–1. The reaction crude was analyzed by HPLC using XBridge Hilic column and a gradient of CH3CN/H2O as eluent (from 98:2 up to 60:40 in 15 min). Initial rate was calculated from the integration of chromatogram peaks of the newly formed species 6b and 7b at different times (see SI).
  62. We are currently investigating the outcome of analogous 1,3-dipolar cycloaddition reactions using the octa-amino cage derived from the hydride reduction of octa-imine 1. Preliminary results with two pairs of substrates indicate the lack of significant differences in the kinetics of the reactions performed in the presence of the two containers. We expect to report these and other additional findings of the acceleration of reactions using the octa-amine container in due course.

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

au4c01118_si_001.pdf (4.3MB, pdf)
au4c01118_si_002.cif (5.4MB, cif)

Data Availability Statement

All dataset collection of computational results of this manuscript is available in the ioChem-BD repository and can be accessed through this link http://dx.doi.org/10.19061/iochem-bd-1-358 CCDC: 2393702.

Articles from JACS Au are provided here courtesy of American Chemical Society

RESOURCES