sign in sign up
<<
Home , Molecularity, Substitution reaction

FULL PAPER

A Self-Assembled Cage with Endohedral Acid Groups both Catalyzes Substitution Reactions and Controls their Paul M. Bogie, Lauren R. Holloway, Courtney Ngai, Tabitha F. Miller, Divine K. Grewal, and Richard J. Hooley[a]*

[16],[17] Abstract: A self-assembled Fe4L6 cage complex internally decorated many possibilities in controlled biomimetic , above with acid functions is capable of accelerating the thioetherification of and beyond simply increasing the effective concentration of activated , ethers and amines by up to 1000-fold. No product bound substrate. The incorporation of active functions in an inhibition is seen, and effective supramolecular catalysis can occur enclosed space enables reagent-controlled reactions to take with as little as 5 % cage. The substrates are bound in the host with place in enclosed cavities, as opposed to [18]-[20] or up to micromolar affinities, whereas the products show binding that is unimolecular rearrangements, [21],[22] which are still the most an order of magnitude weaker. Most importantly, the cage host alters common reactions studied in synthetic hosts. By internalizing the molecularity of the reaction: whereas the reaction catalyzed by reactive functional groups in a cage, the effect of substrate simple acids is a unimolecular, SN1-type substitution process, the rate binding on nucleophilic substitution reactions can be investigated. of the host-mediated process is dependent on the concentration of . The molecularity of the cage-catalyzed reaction is substrate-dependent, and can be up to bimolecular. In addition, the catalysis can be prevented by a large excess of nucleophile, where substrate inhibition dominates, and the use of tritylated anilines as substrates causes a negative feedback loop, whereby the liberated product destroys the catalyst and stops the reaction.

Introduction

Enzymes are commonly thought of as “perfect” catalysts, showing extremely high rate accelerations and high substrate selectivity compared to small molecule catalytic processes.[1] As well as providing a favorable environment for reaction, varying the molecularity of the rate determining step is possible. A common example is general acid-base catalysis,[2] whereby sidechains directly involve themselves in the .[3] Using synthetic host molecules to mimic a variety of types of enzymatic behavior has led to numerous successes in recent years,[4]-[7] including examples of rate accelerations[8]-[10] and binding affinities[11] that even exceed those of natural enzymes. However, altering the molecularity of a reaction with a synthetic host is far less common: cycloadditions and unimolecular rearrangements can be II accelerated by increased effective concentration upon binding, Figure 1. Enzymatic catalysis in a functionalized cage. a) Structure of Fe 4L6 acid cage 1 and a minimized structure of its S4 isomer (SPARTAN, semi- without the need for functional groups oriented towards an internal empirical calculations); b) control meso-helicate 2; c) summary of the acid cavity. catalyzed substitution processes tested. To achieve this type of reactivity, co-encapsulation of multiple substrates is required, in the presence of acidic and/or basic functional groups in a defined cavity. Self-assembled capsules Polar reactions can be challenging for host molecules to capable of co-encapsulation are often unfunctionalized, and do promote or catalyze, especially nucleophilic substitutions. Metal- not contain internal acidic or basic groups. [12] A solution lies in ligand cage hosts can be sensitive to strong , which endohedrally functionalized cage complexes, [13]-[15] which offer have a tendency to destroy the structural M-L contacts. However, there are some exquisite examples of host complexes directing

the outcome of SN2 processes in the literature: aromatic panels in Ga-catecholate tetrahedra invert the in [*] P. M. Bogie, L. R. Holloway, C. Ngai, T. F. Miller, D. K. Grewal, and [23] Prof. R.J. Hooley encapsulated substitutions, and Menshutkin reactions can be University of California - Riverside, Department of , accelerated in deep cavitands with internal acid groups. [24] Other Riverside, CA, 92521, U.S.A. E-mail: [email protected]. examples of polar reactions include eliminations, [8],[9] [25] [26] Supporting information for this article is given via a link at the end of Knoevenagel condensations, epoxide openings, and the document. additions to imines[27] or organic cations. [28],[29] Compared to the

FULL PAPER wealth of cycloadditions and rearrangements promoted or are well-known “SN1” substrates that can undergo various catalyzed by self-assembled hosts, though, polar reactions substitution reactions via their highly stabilized cationic remain rare. intermediates. [31] As cage 1 is sensitive to a variety of different nucleophiles, some as mild as chloride, [32] we focused on mild, neutral nucleophiles for the reactivity tests. The combination of tri- Results and Discussion or diphenylmethyl with thiols in highly acidic media is a well-precedented method of thioether synthesis, [33],[34] and We recently described the synthesis of endohedrally occurs via an acid-catalyzed dissociative substitution mechanism. functionalized acid cage 1, and investigated its ability to catalyze We initially used n-propanethiol (PrSH) as the nucleophile, paired the deprotection of acetals and effect tandem reactions. [30] The with different catalysts in CD3CN, and monitored the relative high rate accelerations observed, and the fact that the cage binds reaction rates by 1H NMR, as shown in Figure 2a. Significant rate 4 -1 benzaldehyde dimethylacetal with Ka = 1.3 x 10 M suggested accelerations were observed for the reaction of both 4a and 4b that cage 1 would be an effective supramolecular catalyst for with PrSH in the presence of 5 % cage 1 as catalyst. The reaction other acid-mediated reactions. Here we show that the cage can was complete after 8 h at 80 °C, and 100 % conversion was catalyze a , effecting rate acceleration and observed in both cases, with no evidence of product inhibition. variable molecularity on the process that is dependent on The conversion is clean, and only the cage, the reactants and substrate molecular recognition. propyl trityl sulfide product 5a are observed in the NMR spectra (see Supporting Information for full spectra). Most importantly, cage 1 remains intact throughout the process, and is completely tolerant to thiol nucleophiles, even at reflux. The characteristic peaks for the imine region of the C3 and S4 isomers of cage 1 at δ 8.9-9.1 ppm are shown in Figure 2b, [30] and no cage decomposition products are formed. To ensure that the cage was the active catalyst rather than small amounts of leached Fe2+ ions, the reaction was repeated with meso-helicate 2 as catalyst, as that assembly does not contain a defined cavity or acidic functions. In that case, no reaction was observed even after 48 h heating. The effectiveness of the acidic cage was then compared to an equivalent concentration of free acid groups by reacting PrSH with 4a and 4b in the presence of 30 % control acid 3. No conversion was observed after 10 h heating at 80 °C for either (Figure 2a). Even 24 h reflux only gave 1% conversion. The observed initial rates and relative rate accelerations of the thioetherification process are shown in Table 1. The self- assembled cage shows up to a 1023-fold acceleration in rate when compared to a “free” acid catalyst that contains the same functional groups (i.e. 3). The relative rates of substitution of the trityl electrophiles 4a and 4b were very similar, and showed similar (~1000-fold) accelerations. The difference in basicity between 4a and 4b (conjugated acid pKa of ~-3.5 vs -2) was not observed to be a determining factor, as the reactions rates are essentially identical. The far less basic trifluoroethyl ether 4c showed no reactivity, however, even after extended reaction times. Benzhydrol 4d was less reactive than 4a, and displayed Figure 2. Accelerated Substitution Catalyzed by Cage 1. a) Reaction only 58 % conversion after 72 h, but a rate acceleration of at least progress over time for the transformation of electrophiles 4a and 4b with either 1 5 % cage 1 or 30 % control acid 3 catalyst (CD3CN, 353 K). b) H NMR spectra 100-fold was observed with 5 % 1 as catalyst as compared to that of the reaction of 4a with PrSH catalyzed by 1 at various intervals (CD3CN, 400 with 30 % 3. MHz, 298 K). Blue = PrSH; Red = thioether product 5a; downfield inset shows The cage-catalyzed substitution reaction can also be the imine CH region of the C3/S4 isomers of 1, and that cage 1 remains intact throughout the reaction. performed with other mild nucleophiles. The more hindered cyclohexanethiol (CySH) showed a slightly slowed initial rate of reaction compared to PrSH, but both 4a and 4b were smoothly The reaction is shown in Figure 1: we chose a mild, acid- converted to product with 5 % 1. The reaction of p-tolylthiol catalyzed substitution reaction to prevent destruction of the cage (TolSH) was complicated by the formation of significant amounts complex. Four different activated electrophiles were tested that of oxidation byproduct p-tolyldisulfide. In the case of all the other thiols, no disulfide was observed at any point during the reaction, vary in reactivity; triphenylmethanol 4a, its ethyl (4b) and despite the fact that the reactions were performed in air. Only the trifluoroethyl (4c) ethers, and benzhydrol 4d. These electrophiles

FULL PAPER more easily oxidized TolSH was susceptible to disulfide formation, for a synthetic receptor. There are no large flat panels (like those but interestingly, minimal disulfide was observed with control acid in Raymond’s Ga-catecholate tetrahedron[35] or Fujita’s 3, indicating that the oxidation is catalyzed by the cage itself. tripyridyltriazine octahedron[36]) that limit the rate of guest egress. Despite this side reaction, the initial rate of thioetherification could In addition, the cage is not appreciably water-soluble, so no be established, and TolSH showed similar reactivity to CySH with hydrophobic driving forces can be exploited. Indeed, cage 1 does both 4a and 4b. Weaker nucleophiles such as EtOH were not form long-lived Michaelis complexes that can be observed on tolerated, but the was slower, and the bulky 1- the NMR timescale. All guests tested exhibit rapid in/out kinetics: adamantanethiol (AdSH) showed no reaction after 24 h (further small changes in chemical shift of the cage peaks could be heating was ineffective, and cage decomposition began to occur observed, but they were not substantial enough for quantitation, at long reaction times). While thiols and alcohols were good nor were any complexes sufficiently long-lived to confer nucleophiles for the process, carbon-based nucleophiles were appreciable NOEs. We therefore used UV-Vis titrations to unsuccessful. Dimethylmalonate and dimedone gave no determine the binding affinity of the various components of the conversion, and indole caused rapid decomposition of the cage. reaction. Binding affinities were calculated via linear regression This data shows that the cage is a highly effective acid catalyst, analysis (Nelder-Mead method) from the change in cage 1 exhibiting rate accelerations of over 1000-fold for certain absorbance at two points (300 nm/330 nm and 370 nm) upon substitutions. It also provides hints as to the host-mediated nature guest titration. [37],[38] For the large electrophiles 4a-4d and product of the catalysis. Substitution at trityl electrophiles occurs via an 5a, the data showed best fit to a 1:1 binding model (Figures S-42 [31] SN1 mechanism, whereby the nucleophile is not part of the rate to S-51). The smaller nucleophiles showed more variable fits, determining step (and thus the rate equation), only that of however. The two smallest thiols, PrSH and CySH, clearly fit best electrophile and acid. As such, the rate of reaction with different to 2:1 binding model, whereas TolSH only fit to the 1:1 model. nucleophiles should be identical. When catalyzed by cage 1, The only ambiguous guest was AdSH: the fitting was most however, the type of nucleophile changes the reaction rate. PrSH, accurate with a 1:1 binding, with only 8 % error. The error in fit in CySH, TolSH and EtOH all show large variations in reaction rate the 2:1 model was higher (27%), but not so high as to rule out 2:1 with 4a and 4b, suggesting that molecular recognition events are binding completely. The affinities (K11 (1) and K12 (1), where essential for catalysis. To determine this, we investigated the relevant) are shown in Table 2, and are unexpectedly high, up to binding affinity of the various reaction components in cage 1. 199,000 M-1 (AdSH). The thiol nucleophiles were the most strongly bound substrates, ranging from 80,000 M-1 to 199,000 M- 1 -1 Table 1. Supramolecular Substitution Catalysis. [a] . The larger electrophiles had weaker affinities (from 3200 M (trifluoroethylether 4c) to 20,100 M-1 (ethyl ether 4b)), but they are still strongly bound. Importantly, the affinity for the thioether product 5a (6500 M-1) is three times lower than that of the triphenylmethanol substrate and almost twenty times lower than that of PrSH nucleophile. The factors that control the affinity are not obvious: non-polar such as adamantane Initial Rate V(1), Initial Rate V(3), Substrate Nucleophile V(1)/V(3) showed no affinity. Presumably, a mix of polar interactions with x10-4 mM/min x10-4 mM/min the acid groups and CH-π or π-π interactions with the aromatic

4a PrSH 778 0.76 1023 cage walls are the key determinants. Shape-based cavity occupancy is not the predominant factor, as even the largest 4a CySH 541 1.4 386 guest 4b does not fill the cavity by itself.

4a TolSH 211 0.81 260 Table 2. Binding Affinities of Substrates and Products in Cage 1.[a] 4a AdSH n.r. n.r. n.d. 3 -1 3 -1 3 -1 Guest K11 (1), x 10 M Guest K11 (1), x 10 M K12(1) x 10 M 4a EtOH 77 n.r. n.d. 4a 15.8 ± 0.3 PrSH 114 ± 15 0.75 ± 0.008 4b PrSH 723 0.65 1112 4b 20.1 ± 1.2 TolSH 80.6 ± 9.7 N/A 4b TolSH 156 1.1 107 4c 3.2 ± 1.3 CySH 116 ± 10 4.0 ± 0.4 4b CySH 156 2.9 47 4d 6.9 ± 0.4 AdSH 199 ± 17 N/A 4c PrSH n.r. n.r. n.d. 5a 6.5 ± 1.3 4d PrSH 39 n.r. n.d. [a] in CH3CN, [1] = 3 μM, absorbance changes measured at 300/330 nm and 370 [a] [37],[38] 353 K, CD3CN, [4] = 15.8 mM, [Nu] = 19.8 mM, [1] = 0.8 mM; [3] = 4.74 mM, nm. concentrations confirmed using dioxane as standard (7.9 mM). n.r. = no reaction; n.d. = not determined. The strong affinities of the reactants for the cage confer a variety of novel outcomes for the catalyzed reaction. These are The large internal cavity and endohedrally oriented acid groups not simple SN1 processes, but show variable molecularity, suggest host-mediated catalysis, but cage 1 is relatively unusual

FULL PAPER depending on the nature of the individual components. Figures recognition in the catalytic cycle. If a large excess of PrSH is used 3a/b show the effect of varying nucleophile concentration on the (15 eq. with respect to 4a), substrate inhibition is observed, and rate of reaction of 4a or 4b with PrSH in the presence of 5 % cage significantly slowed reaction occurs. The higher affinity of PrSH 1. Simply varying the from OH to OEt changes the for 1 saturates the cage, preventing binding and activation of dependence on nucleophile concentration of the catalyzed either 4a or 4b. processes. Despite the fact that 4a and 4b react at very similar An important point is that turnover is very efficient here: only initial rates at identical [PrSH] (Figure 2a, Table 1), and that both 5 % cage 1 is used, and no product inhibition is seen. This is 4a and 4b have similar binding affinities for 1, the molecularity of unusual in bimolecular processes promoted by host molecules: the substitution is very different. The reaction of PrSH with usually, binding one large product guest is more favorable than triphenylmethanol 4a shows a strong dependence on nucleophile binding two smaller reactants due to entropy effects, so the concentration: increased [PrSH] increases the initial rate of the reactions tend to be stoichiometric rather than catalytic. [18],[39] reaction (Figure 3a). The measured order of reaction of Most examples of turnover in synthetic host-mediated bimolecular 4a with respect to nucleophile is ~1.2. reactions are only possible by changing the binding affinity of product some other way, most usually exploiting solubility effects in water. [22],[40] Here, product 5a is a significantly weaker guest for -1 - 1 (Ka = 6500 M ) than either reactant (Ka = 20,100 – 199,000 M 1). These results beg the question of why such variable rate profiles are observed for such similar reactants. Each of the components can bind in cage 1, and a number of pre-equilibria are possible in the reaction. In addition, the large cavity of 1 allows the formation of ternary complexes. Not only can the small nucleophiles form ternary complexes with 1, but there is sufficient space in the cavity to form hetero-ternary complexes. A minimized (SPARTAN, Hartree-Fock) structure of 1•4a•PrSH is shown in Figure 3d, and illustrates the ease in which both reactants can bind in the cavity. When PrSH is titrated into a preformed complex of cage 1 and excess 4a after the addition, further changes in cage absorbance occur, similar to those seen when the reactants are titrated into cage 1 together (see Figures S-60 and S-61). This illustrates that PrSH can still be bound in the system containing both 1 and 4a. Unfortunately, as the changes in cage absorbances are very similar for different guests (lowering of absorbance at 330 nm and increasing absorbance at 370 nm, with a slight red-shifting of absorbance frequency), differentiating between expulsion of 4a from the cage and formation of the heterocomplex is difficult. As such, the individual K11 Figure 3. Initial rates of thioetherification with varying [PrSH]. a) 4a with 1•4a•PrSH 5 % 1, 333 K; b) 4b with 5 % 1, 333 K; c) 4a with 5 % CF3CO2H, 273 K. [4] = and K12 heterocomplex formation constants cannot be 15.8 mM, [1], [TFA] = 0.8 mM in CD3CN; concentrations confirmed using unambiguously determined. However, these spectra, along with 1 dioxane as standard (7.9 mM), rates monitored by H NMR. d) Minimized the reaction rate data and modeling, are consistent with the structure of the S4 isomer of cage 1 co-encapsulating PrSH and possibility of a host:guest complex in solution. triphenylmethanol 4a (Hartree-Fock, SPARTAN). 1•4a•PrSH

In contrast, the reaction of ether 4b, which is more basic and larger than 4a, shows zero dependence on [PrSH] (Figure 3b), and displays initial rate characteristics that correspond to a unimolecular rate-determining step. Control reactions were performed to confirm that the acid-catalyzed mechanism in the absence of cage is truly SN1. As the reaction with control acid 3 was so slow, we analyzed the process catalyzed by 5 % CF3CO2H in CD3CN at room temperature over a period of hours. The reactions were run to < 20 % conversion to ensure that the initial rates could be determined. As can be seen in Figure 3c, the initial rate of the reaction had no dependence at all on nucleophile concentration: the nucleophile is not part of the rate equation, as Figure 4. Substrate-dependent thioetherification mechanisms in cage 1. would be expected for a simple SN1 process. Furthermore, cage 1 shows additional traits that indicate the importance of molecular

FULL PAPER

Figure 4 illustrates the possible equilibria present in the reaction cage 1, liberating diacid ligand 7 as well as forming the ML3 mechanism. The electrophile and nucleophile are competitive complex 8. The formation of product 5a is faster than the cage guests, with the affinity of 4a/4b for 1 being an order of magnitude destruction, and ~25 % conversion is observed before the 5 % lower than that of PrSH nucleophile. The cage is large enough to catalyst is destroyed. The reaction plot is bifurcated, though - conceivably encapsulate both 4a and PrSH, so four initial cage 1 catalyzes solvolysis rapidly until it is destroyed, and then equilibria are present: free 1 (expected to be minimal at these the reaction proceeds more slowly. The free diacid ligand concentrations), 1•(PrSH)2, 1•4a and 1•4a•PrSH. There are two catalyzes the reaction much more poorly than the cage itself, product-determining steps from these equilibrium states. The hence the slowing of the rate. The equivalent reaction with 30% electrophile can be directly activated upon binding, then released control acid 3 is shown in black in Figure 5a, and the post- for external reaction with PrSH, a classical SN1 process that destruction rate is similar to that (although the presence of a small shows a rate that is independent of nucleophile concentration. amount of weakly coordinated Fe2+ salts from decomposition of 8 Alternatively, this activation can occur in a 1•4a•PrSH ternary adds to the background rate somewhat). This process nicely complex, which would involve nucleophile in the rate equation. [41] illustrates the potency of cage 1 as a catalyst - the 12 internal acid Simply changing the electrophile from 4a to 4b switches between groups cooperatively accelerate the activation of the trityl these two mechanistic possibilities. The unimolecular mechanism substrate, and once liberated, the reaction is markedly slowed, can certainly occur with 4a and PrSH, as the cage-catalyzed even though the catalytic groups are mostly identical. By varying process is not strictly bimolecular, but the dependence of the rate the substrate, a “negative feedback loop” is possible, where the on the concentration of PrSH in this case suggests that some catalyst can be inhibited by a secondary coupled reaction that reaction is taking place in a 1•4a•PrSH complex. The differing effects its self-destruction. basicity of 4a and 4b cannot explain this variation in molecularity by itself, as the observed rate with identical [PrSH] is the same. Competitive binding of nucleophile before the rate determining step does not explain this, as the nucleophile is identical in each case, and the affinities of 4a and 4b are broadly similar (Table 2). The high affinity of PrSH does explain the substrate inhibition observed at high [PrSH] - the nucleophile saturates the cage, preventing electrophile binding and reaction. The simplest explanation for the variable molecularities is that the larger 4b limits the formation of a ternary complex with PrSH in the cavity of the cage and the active cationic intermediate must be expelled before reaction, whereas 4a is small enough to allow reaction with the ternary 1•4a•PrSH complex. The variable binding affinities of the different nucleophiles (PrSH, CySH, TolSH) bias the equilibrium populations of 1•4a, and in some cases 1•4a•RSH: these equilibrium populations control the observed rate. It is conceivable that the bimolecular reaction between PrSH and 4a can occur via a concerted process, as the structure shown in Figure 3d positions the reactive species in close proximity, but we have no evidence of that. The most likely mechanism involves loss of water in the ternary complex, followed by rapid collapse to product. The key to the bimolecularity of Figure 5. Self-destruct sequence. a,b) Detritylation of aniline 4e causes a reaction is the selective substrate molecular recognition inside the negative feedback loop, whereby; c) the product turns off the catalyst via active site of 1. These results, as well as the different reactivities transimination. [4e] = 15.8 mM, [PrSH] = 19.75 mM, [1] = 0.8 mM in CD3CN, shown by different nucleophiles, illustrate the importance of the 298 K; concentrations confirmed using dioxane as standard (7.9 mM), rates monitored by 1H NMR. cage in the . The cage catalyzed substitution process is not limited to alcoholic leaving groups: tritylanilines are more rapidly solvolyzed than the equivalent alcohols under acidic conditions, as they are Conclusions more basic and more rapidly protonated. However, amine leaving Here, we have shown that a self-assembled cationic M-L cage groups can obviously react with the cage via subcomponent complex is capable of accelerating, and controlling the exchange, [42] and this provides an opportunity to illustrate the molecularity of, the thioetherification of activated alcohols, ethers catalytic efficiency by “self-immolation”, i.e. destroying the and amines. Rate accelerations of up to 1000-fold are possible, catalyst as the reaction occurs. When N-trityl-4-bromo- relative to the rate catalyzed by “free”, small molecule acid phenylaniline 4e is reacted with PrSH and 5 % cage 1, the analogs. The catalysis is not troubled by product inhibition, unlike substitution reaction is extremely rapid, and occurs in minutes at many processes promoted by self-assembled cages; indeed the room temperature as shown in Figure 5. The product 4- reaction can be turned off in the presence of a large excess of bromoaniline 6 is capable of transimination of the iminopyridine substrate. Most importantly, the cage host alters the molecularity

FULL PAPER of the reaction. The TFA-catalyzed process is a normal, points (300nm/330nm and 370nm), the data was fit to either a 1:1 or 1:2 [37],[38] unimolecular SN1-type substitution process, whereas the rate of binding model and the variance used to determine best fit. the host-mediated reaction with electrophile 4a is dependent on the concentration of nucleophile. This variable molecularity is substrate dependent, and substrates as similar as Acknowledgements triphenylmethanol and O-tritylethyl ether show completely different reaction molecularity. The catalyst is also capable of a The authors would like to thank the National Science Foundation self-destruct sequence in the presence of tritylated amines, where (CHE-1708019 to R.J.H. and CHE-1626673 for the purchase of fast initial reaction is rapidly turned off by destruction of the Bruker NEO 600 and NEO 400 NMR spectrometers) for funding. catalyst by product, essentially a negative feedback loop. All of these properties require strong molecular recognition of multiple Supramolecular chemistry Self-assembly reactants in a catalytic pathway, indicating the potential of Keywords: · · functionalized cage hosts as enzyme-mimicking nanoreactors. Catalysis · Enzyme models · Molecular recognition

[1] A. J. Kirby, Angew. Chem. 1996, 108, 770-790; Angew. Chem., Int. Ed. Engl. 1996, 35, 707-724. Experimental Section [2] S. Nakano, D. M. Chadalavada, P. C. Bevilacqua, Science 2000,287, 1493-1497. [30] [43] General Information. Cage 1 and suberone meso-helicate 2 were [3] J. S. Weinger, K. M. Parnell, S. Dorner, R. Green, S. A. Strobel, Nat. synthesized according to literature procedures. See those publications for Struct. Mol. Biol. 2004, 11, 1101-1106. 1 13 full characterization. H, and C spectra were recorded on Bruker Avance [4] C. J. Brown, F. D. Toste, R. G. Bergman, K. N. Raymond, Chem. Rev. NEO 400 MHz or Bruker Avance 600 MHz NMR spectrometer. The 2015, 115, 3012-3035. spectrometers were automatically tuned and matched to the correct [5] M. D. Ward, P. R. Raithby, Chem. Soc. Rev. 2013, 42, 1619-1636. 1 13 operating frequencies. Proton ( H) and carbon ( C) chemical shifts are [6] T. R. Cook, P. J. Stang, Chem. Rev. 2015, 115, 7001-7045. reported in parts per million (δ) with respect to tetramethylsilane (TMS, [7] I. Sindha, P. S. Mukherjee, Inorg. Chem. 2018, 57, 4205-4221. δ=0), and referenced internally with respect to the protio impurity [8] W. Cullen, M. C. Misuraca, C. A. Hunter, N. H. Williams, M. D. Ward, Nat. 1 13 for CD3CN ( H: 1.94 ppm, C: 118.26 ppm). Deuterated NMR Chem. 2016, 8, 231-236. were obtained from Cambridge Isotope Laboratories, Inc., Andover, MA, [9] D. M. Kaphan, M. D. Levin, R. G. Bergman, K. N. Raymond, F. D. Toste, and used without further purification. Spectra were digitally processed Science 2015, 350, 1235-1238. (phase and baseline corrections, integration, peak analysis) using Bruker [10] W. Cullen, A. J. Metherell, A. B. Wragg, C. G. P. Taylor, N. H. Williams, Topspin 1.3 and MestreNova. All other materials were obtained from M. D. Ward, J. Am. Chem. Soc. 2018, 140, 2821-2828. Aldrich Chemical Company (St. Louis, MO), or Fisher Scientific (Fairlawn, [11] M. V. Rekharsky, T. Mori, C. Yang, Y. H. Ko, N. Selvapalam, H. Kim, D. NJ), and were used as received. Solvents were dried through a Sobransingh, A. E. Kaifer, S. Liu, L. Isaacs, W. Chen, S. Moghaddam, M. commercial solvent purification system (Pure Process Technologies, Inc.). K. Gilson, K. Kim, Y. Inoue, Proc. Natl. Acad. Sci. U.S.A. 2007, 104, UV/Vis spectroscopy was performed on a Cary 60 Photospectrometer 20737-20742. using the Varian Scans program to collect data. [12] R. Chakrabarty, P. S. Mukherjee, P. J. Stang, Chem. Rev. 2011, 11, 6810-6918. General procedure for substitution reactions. Electrophile 4 (6.3 μmol, [13] K. Suzuki, J. Iida, S. Sato, M. Kawano, M. Fujita, Angew. Chem. 2008, 1 mol.-eq.) was placed in an NMR tube followed by 5 mol % acid cage 1 120, 5864-5866; Angew. Chem. Int. Ed. 2008, 47, 5780-5782. (0.31 μmol, 2 mg) or 30 mol % control acid 3 (1.86 mmol, 0.9 mg). The [14] K. Suzuki, M. Kawano, S. Sato, M. Fujita, J. Am. Chem. Soc. 2007, 129, nucleophile (1.25 mol.-eq., 7.9 μmol, 3.9 μL of 2 M solution in CD3CN) was 10652-10653. then added followed by 1,4-dioxane as the internal standard (0.5 mol. -eq. [15] M. Krick, J. Holstein, C. Würtele, G.H. Clever, Chem. Commun. 2016, 52, 3.2 μmol, 1.6 μL of 2 M solution in CD3CN). 400 μL of CD3CN was added, 10411-10414. and the tube was capped and quickly shaken to dissolve all solids. Control [16] Q.-Q. Wang, S. Gonell, S. H. A. M. Leenders, M. Dürr, I. Ivanović- trials use 30 mol % of diacid to maintain the same concentration of COOH Burmazović, J. N. H. Reek Nat. Chem. 2016, 8, 225-230. 1 groups present in 5 mol % of the M4L6 cage 1. An initial H NMR spectrum [17] R. Gramage-Doria, J. Hessels, S. H. A. M. Leenders, O. Tröppner, M. of the reaction mixture was obtained to verify the of the Dürr, I. Ivanović-Burmazović, J. N. H. Reek, Angew. Chem. 2014, 126, sample. The sample was then heated at the specified temperature and the 13598-13602; Angew. Chem. Int. Ed. 2014, 53, 13380-13384. reaction progress monitored over time. Rate calculation trials were [18] Y. Nishioka, T. Yamaguchi, M. Yoshizawa, M. Fujita, J. Am. Chem. Soc. performed in triplicate. The percent conversion values were obtained via 2007, 129, 7000-7001. integration of the product and substrate peaks against the internal [19] M. Yoshizawa, M. Tamura, M. Fujita, Science 2006, 312, 251-254. standard and the calculated values of repeated trials were averaged. [20] V. Martí-Centelles, A. L. Lawrence, P. J. Lusby, J. Am. Chem. Soc. 2018, 140, 2862-2868. General procedure for binding affinity calculations. A 3 µM solution of [21] C. J. Hastings, D. Fiedler, R. G. Bergman, K. N. Raymond, J. Am. Chem. cage 1 was prepared in CH3CN (9 µmol in 3 mL) and an initial Soc. 2008, 130, 10977-1098. measurement was taken. To this solution was then added 1 µL aliquots [22] D. H. Leung, D. Fiedler, R. G. Bergman, K. N. Raymond, Angew. Chem. from a 9 mM solution of the corresponding guest molecule, equating to one 2004, 116, 981-981; Angew. Chem. Int. Ed. 2004, 43, 963-966. molar equivalent guest to cage. Following five additions of 1 µL aliquots, [23] C. Zhao, F. D. Toste, R. G. Bergman, K. N. Raymond, J. Am. Chem. Soc. the additions were increased to 5 µL at a time. These additions were 2014, 136, 14409-14412. continued until there was no observable change in the absorption [24] B. W. Purse, P. Ballester, J. Rebek, Jr. J. Am. Chem. Soc. 2003, 48, spectrum. Binding affinities were calculated via linear regression analysis 14682-14683. using the Nelder-Mead method from the change in absorbance at two [25] T. Murase, Y. Nishijima, M. Fujita, J. Am. Chem. Soc. 2012, 134, 162- 164.

FULL PAPER

[26] J. Jiao, C. Tan, Z. Li, Y. Liu, X. Han, Y. Cui, J. Am. Chem. Soc. 2018, [36] M. Fujita, K. Umemoto, M. Yoshizawa, N. Fujita, T. Kusukawa, K. 140, 2251-2259. Biradha, Chem. Commun. 2001, 6, 509-518. [27] J. Jiao, Z. Li, Z. Qiao, X. Li, Y. Liu, J. Dong, J. Jiang, Y. Cui, Nat. [37] Association constants calculated using BindFit software, found at Commun. 2018, 9, DOI: 10.1038/s41467-018-06872-0. http://supramolecular.org. [28] C. J. Hastings, M. D. Pluth, R. G. Bergman and K. N. Raymond, J. Am. [38] D. B. Hibbert, P. Thordarson, Chem. Commun. 2016, 52, 12792-12805. Chem. Soc. 2010, 20, 6938-6940. [39] J. M. Kang, J. Rebek, Jr. Nature 1996, 382, 239-241. [29] M. D. Pluth, R. G. Bergman, K. N. Raymond, Science 2007, 316, 85-88. [40] R. J. Hooley, S. M. Biros, J. Rebek, Jr., Angew. Chem. 2006, 118, 3597- [30] L. R. Holloway, P. M. Bogie, Y. Lyon, C. Ngai, T. F. Miller, R. R. Julian, 3599; Angew. Chem. Int. Ed. 2006, 45, 3517-3519. R. J. Hooley, J. Am. Chem. Soc. 2018, 140, 8078–8081. [41] E. V. Anslyn, D. Dougherty, in Modern Physical , [31] K. Manabe, S. Iimura, X.-M. Sun, S. Kobayashi, J. Am. Chem. Soc. 2002, University Science Books, 2004. 124, 11971-11978. [42] D. A. Roberts, B. S. Pilgrim, G. Sirvinskaite, T. K. Ronson, J. R. Nitschke, [32] L. R. Holloway, H. H. McGarraugh, M. C. Young, W. Sontising, G. J. O. J. Am. Chem. Soc. 2018, 137, 9616-9623. Beran, R. J. Hooley, Chem. Sci. 2016, 7, 4423-4427. [43] M. C. Young, A. M. Johnson, R. J. Hooley, Chem. Commun. 2014, 50, [33] R. K. Sharma, J. P. Tam, Org. Lett. 2011, 13, 5176-5179. 1378-1380. [34] T. M. Vishwanatha, E. Bergamaschi, A. Dömling, Org. Lett. 2017, 19, 3195-3198. [35] C. M. Hong, R. G. Bergman, K. N. Raymond, Acc. Chem. Res. 2018, 51, 2447-2455.

FULL PAPER

FULL PAPER

An acid-functionalized self-assembled Paul M. Bogie, Lauren R. Holloway, host catalyzes the thioetherification of Courtney Ngai, Tabitha F. Miller, Divine alcohols via high affinity molecular K. Grewal, and Richard J. Hooley* recognition and activation of the reactants. Page No. – Page No.

((Insert TOC Graphic here: max. A Self-Assembled Cage with Endohedral width: 5.5 cm; max. height: 5.0 cm)) Acid Groups both Catalyzes Substitution Reactions and Controls their Molecularity