Research Article Chemistry—A European Journal doi.org/10.1002/chem.202201792 www.chemeurj.org Catalysis by [Ga4L6]12 Metallocage on the Nazarov Cyclization: The Basicity of Complexed Alcohol is Key Gantulga Norjmaa,[a] Fahmi Himo,[b] Jean-Didier Maréchal,[a] and Gregori Ujaque*[a] Abstract: The Nazarov cyclization is investigated in solution the most critical one for the acceleration of the reaction and within K12[Ga4L6] supramolecular organometallic cage by inside the metallocage. In order to calculate the relative means of computational methods. The reaction needs acidic stability of protonated alcohol inside the cavity, we propose a condition in solution but works at neutral pH in the presence computational scheme for the calculation of basicity for of the metallocage. The reaction steps for the process are species inside cavities and can be of general use. These analogous in both media: (a) protonation of the alcohol results are in excellent agreement with the experiments, group, (b) water loss and (c) cyclization. The relative Gibbs identifying key steps of catalysis and providing an in-depth energies of all the steps are affected by changing the understanding of the impact of the metallocage on all the environment from solvent to the metallocage. The first step reaction steps. in the mechanism, the alcohol protonation, turns out to be Introduction ligands and the final structure are mainly driven by coordina- tion rules.[3] This strategy gives rise to a great diversity of well- Supramolecular chemistry is an interdisciplinary field with a defined topologies at the nanoscale that could not be reached spectrum of applications that ranges from molecular recogni- with pure organic scaffolds. Their internal cavities are attractive tion to catalysis.[1,2] This field shares numerous similarities with for hosting molecules with high selectivity and specificity, with processes observed in enzymes, the most striking one is the many applications,[4–6] including catalysis.[7–9] Many metallocages importance host-guest interactions in dictating both the bind- have been employed as supramolecular catalysts[9–13] including ing process of the guest and controlling the catalytic profiles.[1,2] Raymond’s[14] Ga4L6, Fujita’s[15] Pd6L4, Nitschke’s[16] Fe4L6, or This parallelism between host-guest and receptor-ligand of Lusby’s Pd2L4 among others.[17] biomacromolecules has been largely pursued by chemists to The tetrahedron host K12[Ga4L6] designed by Raymond and make efficient cages. coworkers has been demonstrated to accelerate several chem- A range of synthetic hosts has been designed and ical reactions, such as the hydrolysis of orthoformates,[18] optimized enriching the toolbox of available supramolecules. reductive elimination from Au(III) and Pt(IV) complexes,[19–23] Amongst them, the so-called metallocages, Supramolecular hydroalkylation,[24] allyl alcohol isomerization,[25] and aza-Cope Organometallic Complexes (SOCs) or Metal–Organic Cages rearrangement,[26] among others.[27,28] The Nazarov cyclization (MOCs) are among of the most exciting ones. These are (Scheme 1a), the subject of the present investigation, is a very obtained by self-assembly of metal ions or clusters with organic useful reaction for the formation of cyclic compounds by means of C C bond forming process and has also been demonstrated to be catalyzed by the K12[Ga4L6] metallocage.[29] [a] Dr. G. Norjmaa, Dr. J.-D. Maréchal, Dr. G. Ujaque Under common reaction conditions, the Narazov cyclization Departament de Química and Centro de Innovación en Química Avanzada requires acidic media to proceed. However, in the presence of (ORFEO-CINQA) metallocage [Ga4L6]12 , 3, the process can take place at pH 8 Universitat Autònoma de Barcelona 08193 Cerdanyola del Valles, Barcelona, Catalonia (Spain) (Scheme 1a). The rate acceleration is ca. 106 times (Scheme 1b), E-mail: jeandidier.marechal@uab.cat reaching a degree of rate acceleration comparable to those gregori.ujaque@uab.cat observed in several enzymes.[30] The increase in reaction rate [b] Prof. Dr. F. Himo was initially attributed to a combination of several factors, such Department of Organic Chemistry, Arrhenius Laboratory, Stockholm as (a) preorganization of the encapsulated substrate, (b) University 10691, Stockholm (Sweden) transition state stabilization of the cyclization by constrictive E-mail: fahmi.himo@su.se binding, and (c) increase of the basicity of the complexed Supporting information for this article is available on the WWW under alcohol functionality. A subsequent mechanistic study by https://doi.org/10.1002/chem.202201792 Raymond and coworkers showed that the electrocyclization is © 2022 The Authors. Chemistry - A European Journal published by Wiley-VCH the rate-determining step for the uncatalyzed reaction, while GmbH. This is an open access article under the terms of the Creative the electrocyclization and the water loss from the substrate are Commons Attribution Non-Commercial License, which permits use, dis- tribution and reproduction in any medium, provided the original work is both rate-determining for the cage-catalyzed reaction.[31] More- properly cited and is not used for commercial purposes. over, according to their mechanistic analysis, transition state Chem. Eur. J. 2022, e202201792 (1 of 9) © 2022 The Authors. Chemistry - A European Journal published by Wiley-VCH GmbH Research Article Chemistry—A European Journal doi.org/10.1002/chem.202201792 Results and Discussion The Nazarov cyclization was experimentally investigated for three related 1,4-pentadien-3-ols substrates (Scheme 1b). Be- cause the asymmetric compound 1 shows the highest rate constant for the catalyzed reaction and yields no unexpected product (dihydrofulvene isomer), it was selected as case system. A general schematic representation of the most important intermediate structures for the mechanism are illustrated in Scheme 2 (black arrows for the uncatalyzed reaction and orange arrows for the catalyzed one).[29,31] The main steps for the cyclization process can be described as: (i) protonation of the alcohol group of the substrate, 2, (ii) water loss from the substrate, forming a carbocationic species, 4, and (iii) formation of the five-membered ring, 5, by electrocyclization. In what follows, we will present the computational analysis of the reaction mechanism in solution first, followed by the reaction inside the metallocage, step by step, involving encapsulation, protonation, water loss and cyclization. Reaction in Solution The equilibrium between substrate 1, with an alcohol group, and its protonated form 2, depends on the pH of the solution. The reaction requires acidic pH in the absence of the metallo- cage 3, whereas in its presence it can be performed at pH 8. Scheme 1. (a) Nazarov cyclization in presence of metallocage 3. (b) The energy of the protonated reactant can be obtained from Experimental rate constants obtained for three different substrates; the the pKa of protonated species 2, which under the reaction highlighted one has been selected for this study. Rate constants from Ref. [29] and [31]. (c) Schematic representation of the [Ga4L6]12 metallocage, conditions in the presence of the cage (pH 8.0 and 45 °C) was 3. experimentally estimated to be 5.0.[31] Therefore, the energy of the protonated 1,3-pentadienol substrate 2 was estimated by Raymond and coworkers to be 18.9 kcal/mol higher than the neutral substrate 1.[31] stabilization of the cyclization was found to significantly Computing relative Gibbs energy values for protonated vs. contribute to the dramatic rate enhancement.[31] non-protonated species corresponds to the calculation of the Computational chemistry has become an indispensable tool pKa of the protonated alcohol.[46] Following a standard proce- in the study of reaction mechanisms. It provides detailed dure for the calculation of pKa values,[46,47] we determined the understanding of the processes that is not easily accessible by Gibbs energy difference between protonated and non-proto- experimental techniques. In the case of supramolecular catal- nated states to be 19.5 kcal/mol, in excellent agreement with ysis, computational methods have only been applied recently the experimentally estimated value. The procedure to compute mainly because of the large dimensions of the systems.[21–23,32–44] pKa is by calculating the standard Gibbs energy of deprotona- In the current contribution, we present a computational tion (ΔGs(deprot)) of the protonated alcohol in gas phase along study with the aim of identifying the factors causing the rate with the Gibbs energy of solvation of all the species involved in enhancement of the Nazarov reaction inside the metallocage as the thermodynamic cycle: ROH2 +, (H2O)n, ROH and H3O + (H2O)n-1 compared to solution. We use a combination of quantum (see Scheme S1 in the Supporting Information). It is important mechanical (QM) calculations and molecular dynamics (MD) to point out that in order to obtain this Gibbs energy we simulations, and the overall description of the reaction mecha- evaluated how the number of explicit water molecules consid- nism is obtained by connecting experimental and computa- ered in the model affects the results. Thus, the Gibbs energy tional results to describe the reaction steps for the process. difference (ΔG) between the neutral substrate, 1, and the Overall, computational results are in excellent agreement with protonated form, 2, were calculated including one (1-w1 vs. 2- experiment and provide a deeper understanding of the process w1), two (1-w2 vs. 2-w2), three (1-w3 vs. 2-w3), or four explicit identifying the key aspects of catalysis. While this manuscript water molecules (1-w4 vs.2-w4). The results clearly show that was under review appeared a publication by Tantillo, Toste, the energies converge at 3 explicit water molecules (Figure S1 Bergman, Raymond and coworkers on the same topic reaching and Table S2). Calculations inside the metallocage will be also essentially the same conclusions about the origin of catalysis.[45] done including 3 water molecules; MD simulations of the protonated substrate show that an average of 3 water Chem. Eur. J. 2022, e202201792 (2 of 9) © 2022 The Authors. Chemistry - A European Journal published by Wiley-VCH GmbH Research Article Chemistry—A European Journal doi.org/10.1002/chem.202201792 Scheme 2. General schematic representation of the Nazarov cyclization of 1,4-pentadien-3-ol, 1, in solution and in the metallocage. Rate constants are taken from Ref. [29]. molecules are inside the metallocage (see below). In both following intermediate, 4-w3, is 5.3 kcal/mol higher than the 2- systems, solution and inside the metallocage, 3 water molecules w3 (+ 24.8 kcal/mol relative to the zero). At the transition state, need to be explicitly considered in the model. TS2-w3, the dC-Ow distance is elongated to 2.37 Å (Figure 2). From We thus selected 2-w3 as starting model of the protonated 4-w3 the electrocyclization takes place through TS4-w3, calcu- substrate for the reaction in solution. A conformational analysis lated to have an energy of 29.9 kcal/mol (Figure 3). This is the of 2 shows that the most stable conformation is the one named rate-determining step of the reaction, in agreement with the “open” in Figure 1, whereas the reactive conformation, named experimental value of 30.0 kcal/mol, obtained from the meas- “closed”, is calculated to be only 0.5 kcal/mol higher in energy, ured rate constant of 3.3 · 10 8 s 1 at 45 °C.[29,31] Moreover, the showing that it is highly accessible. When three explicit solvents difference between two calculated barriers, 26.6 kcal/mol for molecules are included in the calculations, the difference water loss and 29.9 kcal/mol for electrocyclization, is also in between open (the most stable) and closed one is 0.3 kcal/mol. good agreement with experimental observations of the latter Energies of other conformations are given in the Supporting step as the rate determining step. Information (Table S1). After the protonation, the next step in the reaction is the water loss from the protonated substrate. This was calculated to Reaction in metallocage: Encapsulation and protonation of have a barrier of 7.1 kcal/mol relative to 2-w3, which is substrate 26.6 kcal/mol relative to the global zero (Figure 3), and the The Gibbs energy associated to the encapsulation of alcohol 1 was experimentally determined to be 0.3 kcal/mol (ΔG4,exp.). It is expected that under reaction conditions all the metallocages are occupied with the substrate (the concentration of the reactant is much larger than that of the metallocages). The encapsulation process was calculated using a protocol we recently established[48] for calculating binding Gibbs energies in supramolecular metallocages using the attach–pull–release (APR) method (see computational details). This protocol showed good agreement between calculated and experimental binding Figure 1. Open (2’) and closed (2) conformers of the protonated substrate. Gibbs energies for a set of cationic guests encapsulated in the Chem. Eur. J. 2022, e202201792 (3 of 9) © 2022 The Authors. Chemistry - A European Journal published by Wiley-VCH GmbH Research Article Chemistry—A European Journal doi.org/10.1002/chem.202201792 Once encapsulated, the next step in the reaction is the protonation of the substrate. Direct computation of Gibbs energy for this protonation step is not affordable: it implies calculating the pKa of the protonated substrate inside the metallocage. The general thermodynamic cycle employed to calculate pKa cannot be utilized here to accurately determine the Gibbs energy for the protonation of the substrate within the metallocage; such a cycle needs the value of the Gibbs energy for the solvation of the encapsulated proton (from the gas phase to solution). Nevertheless, the calculation of this magnitude is not reliable using continuum methods.[46] Instead, the Gibbs energy difference between protonated and non- protonated substrate inside the metallocage (ΔG3) can be obtained using an alternative thermodynamic cycle, as the one shown in Figure 4. According to this, the relationship between the Gibbs energies for protonating the neutral substrate (in both, solution, ΔG2, and inside metallocage, ΔG3), must be the same than the relationship between the Gibbs energies of encapsulating both, the neutral, ΔG4., and protonated, ΔG1, substrates. For two of the steps of the thermodynamic cycle the Figure 2. Geometries of intermediates and TSs for the reaction in solution. Distances are in Å; Gibbs energies in kcal/mol; imaginary frequencies in Gibbs energy can be estimated, ΔG2,est and experimentally cm 1. obtained, ΔG4,exp. ΔG1,calc and ΔG4,calc, corresponding to the binding of the neutral and protonated substrate in the metallo- cage, are computed using the APR method based on molecular dynamics simulations. ΔG2,calc, in turn, which corresponds to the [Ga4L6]12 metallocage. Its application to substrate 1 gives a protonation of neutral substrate in solution, is calculated at DFT Gibbs binding energy (ΔG4,calc.) of 3.2 kcal/mol, in fairly good level. In Figure 4 are shown all computed Gibbs energies (in agreement with experiment. blue) and experimental and estimated values (in green). The computed values are in excellent agreement with experiment: Figure 3. Calculated Gibbs energy reaction profile for the Nazarov cyclization in solution and in the presence of metallocage 3. Chem. Eur. J. 2022, e202201792 (4 of 9) © 2022 The Authors. Chemistry - A European Journal published by Wiley-VCH GmbH Research Article Chemistry—A European Journal doi.org/10.1002/chem.202201792 To analyze the effect of the metallocage on the conforma- tion, the Gibbs energy of encapsulation for the open and closed conformations of the protonated substrate were calculated employing the above-mentioned protocol.[48] The values ob- tained are 5.4 kcal/mol and 7.4 kcal/mol, for open and closed conformations, respectively, showing that the closed conformation is favored by 2.0 kcal/mol. The analysis was complemented by performing a classical molecular dynamic (MD) simulation of the encapsulated protonated substrate in a periodic box of explicit solvents (see computational details) and additional DFT calculations. The MD simulations show that there is an average of two to three solvent water molecules around the protonated substrate within the metallocage (see Figure S3). For computing the relative Gibbs energies at DFT level we considered a model Figure 4. Thermodynamic cycle to obtain the basicity (ΔG3) of the encapsu- with 3 solvent molecules encapsulated. To identify the most lated reactant 1. (*) “closed” conformation of the protonated substrate is stable situation of the encapsulated system, different arrange- 0.3 kcal/mol higher in Gibbs energy than the most stable one. Gibbs ments (molecular conformations, water disposition) selected energies in kcal/mol. from the most populated structures along the MD were fully optimized at DFT level. Interestingly, the closed conformation ΔG2,est = 19.2 vs. ΔG2,calc = 19.5 kcal/mol, and ΔG4,exp = 0.3 vs. (the reactive one) is calculated to be the most stable one, 2- ΔG4,calc = 3.2 kcal/mol, respectively. The encapsulation of the w3�3 by 4.2 kcal/mol compared to the lowest open conforma- protonated alcohol, 2, to the metallocage, ΔG1, cannot be tion arrangement, 2’-w3�3. This confirms that the energy determined experimentally. Nevertheless, the binding energy difference between encapsulated open and closed conforma- for this process can be obtained using the protocol previously tions favors the latter by ~ 2-4 kcal/mol (note that in solution commented.[48] The calculation gives a binding Gibbs energy for both conformations were very similar in energy). Hence, this process, ΔG1,calc, of 7.4 kcal/mol. Encapsulating the encapsulation induces the proper preorganization of the protonated substrate would be an exergonic process (Fig- substrate, making the closed conformation the most stable one ure S2). inside the metallocage thus facilitating the reaction. In the thermodynamic cycle (Figure 4), the values of the Gibbs energies are ΔG2,calc = 19.5 kcal/mol, ΔG4,calc = 3.2 kcal/mol and ΔG1,calc = 7.4 kcal/mol. Assuming that these ΔG values can Reaction in metallocage: Water loss be combined despite being obtained by means of different computational approaches, the calculated Gibbs energy for the To obtain an accurate reaction profile within the metallocage at protonation of the encapsulated neutral substrate, ΔG3 is DFT level one needs to properly describe the behavior of 8.9 kcal/mol. According to this, the encapsulated protonated encapsulated protonated substrate, thus including the locally substrate, (2-w3�3), has a relative Gibbs energy of 12.1 kcal/ surrounding solvent molecules. In the previous section we mol within the reaction profile (Figure 3). presented classical molecular dynamic (MD) simulations of the The Gibbs energies associated to these protonation steps in encapsulated protonated substrate in a periodic box of explicit solution and inside the metallocage are 19.5 kcal/mol for 2-w3 solvents. The simulations indicate that there are around three and 12.1 kcal/mol for 2-w3�3, respectively (Figure 3). The solvent water molecules along with the protonated substrate comparison of these values reveals that the protonation process within the metallocage (see Figure S3). Thus, for calculating the is much more accessible inside the cage than in solution by Gibbs energy profile at DFT level we selected the most stable 7.4 kcal/mol. In a related study, Warshel and coworkers[38] also structure, 2-w3�3, which includes the protonated substrate on found that the rate acceleration by the same [Ga4L6]12 metallo- the closed conformation and three water molecules altogether cage on a related process, hydrolysis of orthoformate, is mainly in the [Ga4L6]12 metallocage. due to large stabilization of H3O + inside the metallocage The optimized geometries of intermediates and transition (pointing out a very low “local pH”) and the electrostatic states for encapsulated substrate along with three explicit water stabilization of the positively charged transition state. Similarly, molecules are shown in Figure 5. The encapsulated protonated the dramatic change of basicity of the alcohol over encapsula- intermediate, 2-w3�3, as previously shown, has an estimated tion observed here is crucial for accelerating the reaction. In relative Gibbs energy of 12.1 kcal/mol within the reaction fact, accessing to the 2-w3�3 intermediate takes place at the profile (Figure 3). The transition state for the water loss from beginning of the process, thus all the reaction steps taking the protonated substrate inside the metallocage, TS2-w3�3, has a place afterwards are largely affected. The Gibbs energy differ- relative Gibbs energy barrier of 9.7 kcal/mol, with a global ence for this protonation step in solution and inside the barrier of 21.8 kcal/mol. The distance of the protonated alcohol metallocage has the largest impact on the overall reaction. C O bond (dC-Ow) is elongated to 2. 37 Å at the transition state (TS2-w3�3) from 1.51 Å at the intermediate (2-w3�3). Chem. Eur. J. 2022, e202201792 (5 of 9) © 2022 The Authors. Chemistry - A European Journal published by Wiley-VCH GmbH Research Article Chemistry—A European Journal doi.org/10.1002/chem.202201792 Figure 5. Optimized geometries and relative Gibbs energies for the intermediates and TSs of the Nazarov cyclization inside metallocage 3; imaginary frequencies in cm 1. Once the substrate is protonated, next step in the process corresponding barrier inside the metallocage (from 4-w3�3 to corresponds to water loss. The Gibbs energy barrier for this TS2-w3�3) is 4.2 kcal/mol. This is in very good agreement with process in solution (from 2-w3 to TS2-w3) is 7.1 kcal/mol, whereas experiment where it has been shown that the backward the analogous step inside the metallocage is 9.7 kcal/mol (from reaction step inside the metallocage is more energy demanding 2-w3�3 to TS2-w3�3). Therefore, performing the process inside than in solution.[29,31] Altogether, the effect of the metallocage the metallocage provokes an increase of the barrier of 2.7 kcal/ on the water loss step, despite non favoring the process, is mol for this step. Water loss inside the cavity leads to a penalty relatively moderate. in terms of Gibbs energy barrier, although that penalty is not large. Such an increase on the Gibbs energy barrier is somewhat relevant because involves that this step becomes rate limiting Reaction in metallocage: Cyclization along with the cyclization (see below). At this point it is worth commenting how is affected the Before analyzing the cyclization step, we investigated the Gibbs energy barrier for the reversal process of water loss in accessibility for a conformational change of 4 inside the solution (water addition from intermediate 4-w3) and inside the metallocage; a modification from closed to other open con- metallocage (4-w3�3). In solution, the relative Gibbs energy formation should be highly detrimental for the reaction. To that barrier (from 4-w3 to TS2-w3) is 1.8 kcal/mol, whereas the purpose, we performed an Adaptively Biased Molecular Dynam- Chem. Eur. J. 2022, e202201792 (6 of 9) © 2022 The Authors. Chemistry - A European Journal published by Wiley-VCH GmbH Research Article Chemistry—A European Journal doi.org/10.1002/chem.202201792 ic (ABMD) simulation with AMBER 16 program using two torsion Conclusions angles as collective variables (see Figure S5). The simulation shows that the conformational change within the metallocage The reaction mechanism for the Nazarov cyclization was is too energetically demanding to be affordable (more than analyzed in solution and inside the [Ga4L6]12 metallocage by 20 kcal/mol). Thus, the closed conformation is maintained once means of computational methods combining DFT and classical the intermediate after water loss is formed. molecular dynamics calculations. The encapsulated intermediate after the water loss, 4- The reaction mechanism in solution and inside metallocage w3�3, is calculated to be at 17.6 kcal/mol in Gibbs energy have the same reaction steps: (i) protonation of the alcohol relative to 1. For the next step, cyclization, the lowest energy group of the substrate (ii) water loss from the substrate, barrier step was found with a different arrangement of the forming a carbo-cationic species, and (iii) formation of the five- water molecules inside the cavity, 4-w3’�3. The calculations member ring, 5, by electrocyclization. Nevertheless, upon show that the relative Gibbs energy of this intermediate is not encapsulation the Gibbs energy profile is dramatically affected. significantly affected by the solvent rearrangement inside the The first step in the reaction is the protonation of the metallocage; the calculated relative Gibbs energies are 17.6 and substrate. The associated Gibbs energies in solution and inside 17.9 kcal/mol for 4-w3�3 and 4-w3’�3, respectively. These the metallocage are 19.5 kcal/mol for ΔG2 and 12.1 kcal/mol for rearrangements of water molecules in the metallocage are ΔG3, respectively (Figure 4). The comparison of these values schematically illustrated in Figure 3 and their optimized geo- reveals that the protonation process is much more accessible metries are shown in Figure 5. For the cyclization process itself, inside the cage than in solution by 7.4 kcal/mol. This becomes several transition states with different molecular arrangements the most important factor for accelerating the reaction rate. were optimized. The lowest energy one, TS4-w3’�3, is 3.9 kcal/mol Additionally, whereas in solution an open conformation is above 4-w3’�3, located at 21.8 kcal/mol from 1, (Figure 3). The favored (too far from the catalytically competent geometry for optimized geometries of the transition state with different cyclization), inside the metallocage a closed conformation with rearrangements of water solvent molecules in the metallocage a geometry close to the reactive form is preferred (by 2 to are shown in the Supporting Information (Figure S4b). Such 4 kcal/mol, depending on the method employed). Therefore, water arrangements have larger effect on the Gibbs energy of the metallocage promotes a preorganization of the substrate the transition states than on the intermediates (4-w3�3 and 4- for the reaction to proceed. w3’�3 in Figure 3). These two intermediates, species 4-w3�3 The next step, the water loss, pays a penalty in terms of and 4-w3’�3 are very similar in energy, just involving a solvent Gibbs energy barrier inside the cavity, although the effect is reorganization. The forming C C bond distance at this quite moderate. Its Gibbs energy barrier is 7.1 kcal/mol in transition state is 2.364 Å (Figure 5). The formation of the cyclic solution and 9.7 kcal/mol once encapsulated. Such a relative carbocationic intermediate, 5-w3�3, is highly exergonic with a increase in the barrier involves that this step also becomes rate Gibbs energy of 3.4 kcal/mol. Therefore, the reaction is limiting along with the cyclization. Finally, the Gibbs energy irreversible once the cyclized species is formed. barrier of the cyclization is lowered when occurring inside the To evaluate the magnitude of the stabilization of the cage by 1.2 kcal/mol. It suggests a relatively moderate effect of cyclization TS associated to the encapsulation, one needs to the supramolecular host on the rate acceleration regarding the compute and compare its barrier in solution and inside the cyclization step. metallocage. The relative Gibbs energy barrier associated to this Overall, despite preorganization of the substrate upon step in solution is 5.1 kcal/mol (from 4-w3 to TS4-w3), whereas encapsulation helps the process, the ability of the metallocage that for the cyclization step inside the metallocage is 3.9 kcal/ to stabilize the cationic species (protonated alcohol) is the mol (from 4-w3’�3 to TS4-w3’�3); the energy barrier decreases by crucial factor for the rate acceleration observed. The ensuing 1.2 kcal/mol. Thus, despite the metallocage favors the cycliza- two steps, water loss and cyclization, are both affected tion step, the effect on the overall process is relatively moderately by encapsulation. Understanding and evaluating moderate. these effects should help to the rational design of novel host- The increase of the Gibbs barrier for the water loss along guest catalyses. with the fact that the relative Gibbs energy barrier for the electrocyclization diminishes inside the metallocage, involves that the relative Gibbs energy value for the transition states of Computational Details water loss (TS2-w3�3) and cyclization (TS4-w3’�3) acquire both the All DFT calculations were performed using the Gaussian 09 same value of 21.8 kcal/mol. Therefore, both steps are involved software[49] with the B3LYP D3 functional.[50–52] The SDD pseudopo- in defining the reaction rate of the process, once again in very tential and related basis set, complemented with a set of d good agreement with experiment.[31] Moreover, the calculated polarization functions, were used for Ga,[53,54] while the 6-31G(d) Gibbs energy barrier of 21.8 kcal/mol for both rate determining basis set was used for all other atoms.[55] The structures of the reactants, intermediates, transition states and products were steps is also in excellent agreement with the estimated value of optimized in water solvent by using the SMD continuum model. [56] 21.0 kcal/mol (from the experimental rate constant of These calculations involve systems with more than 300 atoms. The 5.7 · 10 2 s 1 at 45 °C). quasi-rigid-rotor-harmonic-oscillator (quasi-RRHO) approach was used for thermal contributions to the Gibbs energies.[57] The standard state correction (1.9 kcal/mol to each of the compounds) Chem. Eur. J. 2022, e202201792 (7 of 9) © 2022 The Authors. Chemistry - A European Journal published by Wiley-VCH GmbH Research Article Chemistry—A European Journal doi.org/10.1002/chem.202201792 was also included. [58] Gibbs energies were calculated at full DFT [6] W. N. M. Van Leeuwen, M. Raynal, Supramolecular Catalysis; Wiley-VCH, level except the binding of the neutral and protonated substrate in Weinheim, 2008. the metallocage which were computed using the APR method (see [7] W. N. M. Van Leeuwen, M. Raynal, Supramolecular Catalysis: New below). Directions and Developments; Wiley-VCH, Weinheim, 2022. [8] M. Morimoto, S. M. Bierschenk, K. T. Xia, R. G. Bergman, K. N. Raymond, MD simulations were performed with the AMBER 16 package[59] F. D. Toste, Nat. Catal. 2020, 3, 969–984. using the CUDA version of the pmemd program. The MD simulation [9] D. M. Vriezema, M. Comellas Aragonès, J. A. A. W. Elemans, J. J. L. M. box, of size 46 × 49 × 47 Å and treated under periodic boundary Cornelissen, A. E. Rowan, R. J. M. Nolte, Chem. Rev. 2005, 105, 1445– 1490. conditions, contains the metallocage, the guest, 11 potassium [10] T. S. Koblenz, J. Wassenaar, J. N. H. Reek, Chem. Soc. Rev. 2008, 37, 247– counterions, and ~ 2500 water molecules (TIP3P). [60] The simu- 262. lations were performed at constant temperature (298.15 K, using a [11] M. Raynal, P. Ballester, A. Vidal-Ferran, P. W. N. M. van Leeuwen, Chem. Langevin thermostat) and pressure (1 bar, using a Monte Carlo Soc. Rev. 2014, 43, 1734–1787. barostat). Force field parametrization procedure was described in [12] S. H. A. M. Leenders, R. Gramage-Doria, B. de Bruin, J. N. H. Reek, Chem. our previous work (the parameters of the metallocage were directly Soc. Rev. 2015, 44, 433–448. used). [48] The absolute binding Gibbs energies (ΔGbind) were [13] H. Vardhan, F. Verpoort, Adv. Synth. Catal. 2015, 357, 1351–1368. computed using the attach–pull–release (APR) method.[61–63] The [14] D. L. Caulder, C. Brückner, R. E. Powers, S. König, T. N. Parac, J. A. Leary, simulation details were described in the previous work [48] and can K. N. Raymond, J. Am. Chem. Soc. 2001, 123, 8923–8938. [15] M. Yoshizawa, M. Tamura, M. Fujita, Science 2006, 312, 251–254. be summarized as: (i) derivation of non-standard parameters [16] J. L. Bolliger, A. M. Belenguer, J. R. Nitschke, Angew. Chem. Int. Ed. 2013, needed for the host and guest molecules, (ii) combining the 52, 7958–7962; Angew. Chem. 2013, 125, 8116–8120. derived parameters with the general AMBER force field, (iii) [17] H. Amouri, C. Desmarets, J. Moussa, Chem. Rev. 2012, 112, 2015–2041. calculating the binding Gibbs energy with the APR method. The [18] M. D. Pluth, R. G. Bergman, K. N. Raymond, Science 2007, 316, 85–88. adaptively biased molecular dynamics (ABMD) [64] simulations were [19] D. M. Kaphan, M. D. Levin, R. G. Bergman, K. N. Raymond, F. D. Toste, performed to analyze the conformational change of the carboca- Science 2015, 350, 1235–1238. tionic intermediate inside the metallocage (see Supporting Informa- [20] M. D. Levin, D. M. Kaphan, C. M. Hong, R. G. Bergman, K. N. Raymond, tion and Figure S5). F. D. Toste, J. Am. Chem. Soc. 2016, 138, 9682–9693. [21] G. Norjmaa, J D. Maréchal, G. Ujaque, J. Am. Chem. Soc. 2019, 141, 13114–13123. [22] G. Norjmaa, J D. Maréchal, G. Ujaque, Chem. Eur. J. 2020, 26, 6988– Acknowledgements 6992. [23] G. Norjmaa, J D. Maréchal, G. Ujaque, Chem. Eur. J. 2021, 27, 15973– 15980. We thank D. Tantillo and coworkers for generously sharing [24] Z. J. Wang, C. J. Brown, R. G. Bergman, K. N. Raymond, F. D. Toste, J. Am. results before publication (their work appeared in Ref. [45]). The Chem. Soc. 2011, 133, 7358–7360. authors acknowledge the financial support of the Spanish [25] D. H. Leung, R. G. Bergman, K. N. Raymond, J. Am. Chem. Soc. 2007, 129, 2746–2747. MINECO-FEDER (Grants PID2020-116861GB I00 and RED2018- [26] C. J. Brown, R. G. Bergman, K. N. Raymond, J. Am. Chem. Soc. 2009, 131, 102387-T). UAB is also acknowledged by a PIF grant to G.N., and 17530–17531. Generalitat de Catalunya for grant 2017SGR1323. P. Vidossich is [27] D. H. Leung, R. G. Bergman, K. N. Raymond, J. Am. Chem. Soc. 2006, 128, 9781–9797. acknowledged for his helpful comments. [28] C. M. Hong, R. G. Bergman, K. N. Raymond, F. D. Toste, Acc. Chem. Res. 2018, 51, 2447–2455. [29] C. J. Hastings, M. D. Pluth, R. G. Bergman, K. N. Raymond, J. Am. Chem. Conflict of Interest Soc. 2010, 132, 6938–6940. [30] A. Radzicka, R. Wolfenden, Science. 1995, 267, 90–93. [31] C. J. Hastings, R. G. Bergman, K. N. Raymond, Chem. Eur. J. 2014, 20, The authors declare no conflict of interest. 3966–3973. [32] C. Goehry, M. Besora, F. Maseras, ACS Catal. 2015, 5, 2445–2451. [33] Y. Ootani, Y. Akinaga, T. Nakajima, J. Comput. Chem. 2015, 36, 459–466. [34] “Catalysis by Metal–Organic Cages: A Computational Perspective”: G. Data Availability Statement Sciortino, G. Norjmaa, J D. Maréchal, G. Ujaque in Supramolecular Catalysis: New Directions and Developments, Wiley-VCH, Weinheim 2022; pp 271–285. The data that support the findings of this study are available in [35] C. Garciá-Simón, C. Colomban, Y. A. Çetin, A. Gimeno, M. Pujals, E. the supplementary material of this article. Ubasart, C. Fuertes-Espinosa, K. Asad, N. Chronakis, M. Costas, J. Jiménez-Barbero, F. Feixas, X. Ribas, J. Am. Chem. Soc. 2020, 142, 16051– 16063. Keywords: density functional theory · metallocage · molecular [36] A. Tarzia, K. E. Jelfs, Chem. Commun. 2022, 58, 3717–3730. dynamics · Nazarov cyclization · supramolecular catalysis [37] W L. Li, H. Hao, T. Head-Gordon, ACS Catal. 2022, 12, 3782–3788. [38] M. P. Frushicheva, S. Mukherjee, A. Warshel, J. Phys. Chem. B 2012, 116, 13353–13360. [39] V. Vaissier Welborn, T. Head-Gordon, J. Phys. Chem. Lett. 2018, 9, 3814– [1] A. J. McConnell, C. J. E. Haynes, C. Caltagirone, J. R. Hiscock, ChemPlu- 3818. sChem 2020, 85, 2544–2545. [40] F. Sebastiani, T. A. Bender, S. Pezzotti, W. L. Li, G. Schwaab, R. G. [2] W. Liu, J. F. Stoddart, Chem. 2021, 7, 919–947. Bergman, K. N. Raymond, F. D. Toste, T. Head-Gordon, M. Havenith, Proc. [3] J. L. Atwood, Comprehensive Supramolecular Chemistry II; Elsevier, Natl. Acad. Sci. USA 2020, 117, 32954–32961. Amsterdam, 2017. [41] T. A. Young, V. Martí-Centelles, J. Wang, P. J. Lusby, F. Duarte, J. Am. [4] Y. Voloshin, I. Belaya, R. Krämer, The Encapsulation Phenomenon: Chem. Soc. 2020, 142, 1300–1310. Synthesis, Reactivity and Applications of Caged Ions and Molecules; [42] H. Daver, J. N. Harvey, J. Rebek, F. Himo, J. Am. Chem. Soc. 2017, 139, Springer, Cham, 2016. 15494–15503. [5] C. García-Simón, M. Garcia-Borràs, L. Gómez, T. Parella, S. Osuna, J. [43] E. Pahima, Q. Zhang, K. Tiefenbacher, D. T. Major, J. Am. Chem. Soc. Juanhuix, I. Imaz, D. Maspoch, M. Costas, X. Ribas, Nat. Commun. 2014, 2019, 141, 6234–6246. 5, 5557. [44] D. Chakraborty, P. K. Chattaraj, J. Comput. Chem. 2018, 39, 151–160. Chem. Eur. J. 2022, e202201792 (8 of 9) © 2022 The Authors. Chemistry - A European Journal published by Wiley-VCH GmbH Research Article Chemistry—A European Journal doi.org/10.1002/chem.202201792 [45] Q. N. N. Nguyen, K. T. Xia, Y. Zhang, N. Chen, M. Morimoto, X. Pei, Y. Ha, [55] V. A. Rassolov, M. A. Ratner, J. A. Pople, P. C. Redfern, L. A. Curtiss, J. J. Guo, W. Yang, L P. Wang, R. G. Bergman, K. N. Raymond, F. D. Toste, Comput. Chem. 2001, 22, 976–984. D. J. Tantillo, J. Am. Chem. Soc. 2022, 144, 11413–11424. [56] A. V. Marenich, C. J. Cramer, D. G. Truhlar, J. Phys. Chem. B 2009, 113, [46] J. R. Pliego, J. M. Riveros, J. Phys. Chem. A. 2002, 106, 7434–7439. 6378–6396. [47] M. A. Ortuño, A. Lledós, J. Organomet. Chem. 2021, 949, 121957. [57] S. Grimme, Chem. Eur. J. 2012, 18, 9955–9964. [48] G. Norjmaa, P. Vidossich, J D. Maréchal, G. Ujaque, J. Chem. Inf. Model. [58] V. S. Bryantsev, M. S. Diallo, W. A. Goddard, J. Phys. Chem. B 2008, 112, 2021, 61, 4370–4381. 9709–9719. [49] Gaussian09, Revision D.01,M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. [59] D. A. Case, R. M. Betz, D. S. Cerutti, T. E. Cheatham, III, T. A. Darden, R. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Duke, T. J. Giese, H. Gohlke, A. W. Goetz, N. Homeyer, S. Izadi, P. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Janowski, J. Kaus, A. Kovalenko, T. S. Lee, S. LeGrand, P. Li, C. Lin, T. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, Luchko, R. Luo, B. Madej, D. Mermelstein, K. M. Merz, G. Monard, H. M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Nguyen, H. T. Nguyen, I. Omelyan, A. Onufriev, D. R. Roe, A. Roitberg, C. Honda, O. Kitao, H. Nakai, T. Vreven, J. A. Montgomery, Jr., J. E. Peralta, Sagui, C. L. Simmerling, W. M. Botello-Smith, J. Swails, R. C. Walker, F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. J.Wang, R. M. Wolf, X. Wu, L. Xiao, P. A. Kollman (2016), AMBER2016, Staroverov, T. Keith, R. Kobayashi, J. Normand, K. Raghavachari, A. University of California, San Francisco. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J. M. [60] P. Mark, L. Nilsson, J. Phys. Chem. A 2001, 105, 9954–9960. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, [61] J. Yin, N. M. Henriksen, D. R. Slochower, M. K. Gilson, J. Comput.-Aided R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Mol. Des. 2017, 31, 133–145. Pomelli, J. W. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, [62] J. Yin, N. M. Henriksen, D. R. Slochower, M. K. Gilson, AMBER Advanced G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels, O. Tutorial 29 Computing Binding Free Energy and Binding Enthalpy Using Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski, D. J. Fox, Gaussian, Inc., the Attach-Pull-Release (APR) Method Example. 2016, https://ambermd. WallingfordCT, 2013. org/tutorials/advanced/tutorial29/. [50] A. D. Becke, J. Chem. Phys. 1993, 98, 5648–5652. [63] G. Heinzelmann, N. M. Henriksen, M. K. Gilson, J. Chem. Theory Comput. [51] C. Lee, W. Yang, R. G. Parr, Phys. Rev. B: Condens. Matter Mater. Phys. 2017, 13, 3260–3275. 1988, 37, 785–789. [64] V. Babin, C. Roland, C. Sagui, J. Chem. Phys. 2008, 128, 134101. [52] S. Grimme, J. Antony, S. Ehrlich, H. Krieg, J. Chem. Phys. 2010, 132, 154104–154119. [53] T. H. Dunning, P. J. Hay, Modern Theoretical Chemistry; Schaefer, H. F., III, Eds.; Plenum: New York, 1977; Vol. 3, pp 1–28. [54] A. Höllwarth, M. Böhme, S. Dapprich, A. W. Ehlers, A. Gobbi, V. Jonas, Manuscript received: June 10, 2022 K. F. Köhler, R. Stegmann, A. Veldkamp, G. Frenking, Chem. Phys. Lett. Accepted manuscript online: July 20, 2022 1993, 208, 237–240. Version of record online: ■■■, ■■■■ Chem. Eur. J. 2022, e202201792 (9 of 9) © 2022 The Authors. Chemistry - A European Journal published by Wiley-VCH GmbH RESEARCH ARTICLE The Nazarov cyclization process Dr. G. Norjmaa, Prof. Dr. F. Himo, Dr. J.- performed inside a metallocage is D. Maréchal, Dr. G. Ujaque* much faster than in solution. This is 1 – 10 mainly due to the change of basicity of the substrate (an alcohol) once en- Catalysis by [Ga4L6]12 Metallocage capsulated. on the Nazarov Cyclization: The Basicity of Complexed Alcohol is Key