Self-Assembled Tetrahedral Hosts as Supramolecular CatalystsPublished as part of the Accounts of Chemical Research special issue “Supramolecular Chemistry in ConfinedSpace and Organized Assemblies”.

Cynthia M. Hong, Robert G. Bergman,* Kenneth N. Raymond,* and F. Dean Toste*

Chemical Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, California 94720, United States

Department of Chemistry, University of California, Berkeley, California 94720, United States

CONSPECTUS: The field of supramolecular chemistry has itsfoundation in molecular recognition and selective binding of guestmolecules, often with remarkably strong binding affinities. Thefield evolved to leverage these favorable interactions between thehost and its guest to catalyze simple, often biomimetictransformations. Drawing inspiration from these early studies,self-assembled supramolecular hosts continue to capture asignificant amount of interest toward their development ascatalysts for increasingly complex transformations. Nature oftenrelies on microenvironments, derived from complex tertiarystructures and a well-defined active site, to promote reactions withremarkable rate acceleration, substrate specificity, and productselectivity. Similarly, supramolecular chemists have becomeincreasingly intrigued by the prospect that self-assembly ofmolecular components might generate defined and spatially segregated microenvironments that can catalyze complextransformations.Among the growing palette of supramolecular catalysts, an anionic, water-soluble, tetrahedral metal−ligand coordination hosthas found a range of applications in catalysis and beyond. Early work focused on characterizing and understanding this host andits various host−guest phenomena, which paved the path for exploiting these features to selectively promote desirablechemistries, including cyclizations, rearrangements, and bimolecular reactions. Although this early work matured intoachievements of catalysis with dramatic rate accelerations as well as enantioenrichment, the afforded products were typicallyidentical to those produced by background reactions that occurred outside of the host microenvironment.This Account describes our recent developments in the application of these anionic tetrahedral hosts as catalysts for organic andorganometallic transformation. Inspiration from natural systems and unmet synthetic challenges led to supramolecular catalysisdisplaying unique divergences in reactivity to give products that are inaccessible from bulk solution. Additionally, thesetetrahedral assemblies have been shown to catalyze a diverse range of transformations with notable rate acceleration over theuncatalyzed background reaction. The pursuit of complexity beyond supramolecular catalysis has since led to the integration ofthese tetrahedral catalysts in tandem with natural enzymes, as well as their application to dual catalysis to realize challengingsynthetic reactions.Variation in the structure, including size and charge, of these tetrahedral catalysts has enabled recent studies that provideinsights into connections between specific structural features of these hosts and their reactivities. These mechanistic studiesreveal that the solvent exclusion properties, hydrophobic effects, confinement effects and electrostatic effects play importantroles in the observed catalysis. Moreover, these features may be leveraged for the design of supramolecular catalysis beyondthose described in this Account. Finally, the supramolecular chemistry detailed in this Account has presented the opportunity toemulate some of the mechanisms nature engages to achieve catalysis; however, this relationship need not be entirelyunidirectional, as the examples describe herein can stand as simplified model systems for unravelling more complex biologicalprocesses.

■ INTRODUCTION

The canonical reaction within a flask limits synthetic chemists’control over reactivity to macroscopic parameters: temperature,solvent, reagent concentrations, and the exclusion of circum-stantially deleterious components (e.g., water, oxygen, andlight). This necessitates that all individual mechanistic steps andreaction intermediates must be amendable to the chosen

parameters and yield the intended outcome, which constrainsthe scope of transformations performed in synthetic chemistry.Supramolecular assemblies with well-defined cavities present

a rare opportunity for synthetic chemists to surpass these

Received: July 3, 2018Published: October 1, 2018

Article

pubs.acs.org/accountsCite This: Acc. Chem. Res. 2018, 51, 2447−2455

© 2018 American Chemical Society 2447 DOI: 10.1021/acs.accounts.8b00328Acc. Chem. Res. 2018, 51, 2447−2455

Dow

nloa

ded

via

UN

IV O

F C

AL

IFO

RN

IA B

ER

KE

LE

Y o

n Fe

brua

ry 6

, 201

9 at

01:

22:3

3 (U

TC

).

See

http

s://p

ubs.

acs.

org/

shar

ingg

uide

lines

for

opt

ions

on

how

to le

gitim

atel

y sh

are

publ

ishe

d ar

ticle

s.

limitations by generating internal microenvironments that aredistinct from bulk solution. Upon encapsulation, guests areinfluenced by the host at the molecular level via thesemicroenvironments, not parameters defined at the macroscopiclevel. Common manifestations of this control include (i)segregation of guests from solvent and nonencapsulatedconstituents, (ii) conformational control over the guest, (iii)shifts in equilibria to generate reactive species, and (iv) dramaticincreases in effective concentrations upon binding multipleguests. These phenomena, among others, control the relativeenergetics of the ground and transition states of substrates andultimately grant new opportunities to achieve a wealth ofhomogeneous applications including catalysis,1 reactive inter-mediate stabilization,2 chemical sensing,3 and the mimicry ofbiological processes such as signal transduction,4 multicatalyticcascades,5 configurationally adaptive binding,6 and informationtransfer.7

Supramolecular chemists draw inspiration from nature’sprecise chemistry.8 In biological systems, complex reactionsfeaturing high uncatalyzed energetic barriers proceed efficientlywith excellent selectivities, despite limited temperature controland an abundance of reactive species. Enzymes’ exquisitecommand over substrates is achieved by the isolation ofsubstrates into customized active sites. The parallels betweenenzymatic active sites and the microenvironments of supra-molecular hosts are self-evident, as both rely on molecularrecognition, substrate isolation, and conformational control.Thus, the development of supramolecular chemistry neces-sitates the interpretation and application of insights gleanedfrom enzymatic catalysis.The broad range of supramolecular architectures applied to

catalysis spans from early covalent structures to self-assembledsystems constructed via a number of assembly motifs.8−11 In thisAccount, we survey recent advancements within a fruitfulcollaboration between the Toste, Raymond, and Bergmangroups studying self-assembled metal−ligand tetrahedra withwell-defined, isolated cavities. In contrast to previous topicalreviews, we focus on the divergent reactivities achieved, as wellas the development of complex multicatalytic systems andmechanistic probes. To demonstrate the strategies employed inour pursuit of catalysis, we begin with a discussion of thestructural characteristics of the tetrahedron and revisit selecthost−guest studies to give context to crucial insights regardingthe effects of encapsulation. We then connect lessons learnedfrom those works to recent developments in utilizing the hostitself as a catalyst to access new products via divergentreactivities. Generally, our approaches aim to exploit intrinsichost properties established from earlier work, such as electro-static effects, confinement effects, and solvent exclusion. Finally,more complex examples that involve multicatalytic reactioncascades with biological and synthetic systems, as well asexamples where broader mechanistic insights can be obtainedfrom studying reactions with diversified catalysts, are discussed.

■ PROPERTIES OF THE HOSTThe featured supramolecular host is self-assembled from sixbisbidentate ligands 1 and four metal ions to form a cage withmetal−triscatecholate vertices. The tetrahedron has ligandsspanning the six edges, as evidenced by crystallographicallyobtained geometric parameters and its T symmetry (Figure1a).12 This architecture was predesigned by fixation of the C2-axis of symmetry of ligand 1 about the C3-axis of the metal−triscatecholates, as well as the geometry of the 1,5-naphthalene

substitution pattern in 1 (Figure 1b).13 While several III and IVoxidation state metal hosts have been prepared, Ga(III) is mostcommonly used to generate 2 [K12(Ga416)]. Because eachGa(III)−triscatecholate has a formal trianionic charge, 2 isoverall dodecaanionic. By virtue of this high charge, the host issoluble in water and polar organic solvents. In terms of size, thehost falls just within the nanoscale regime, measuring about 13 Åfrom vertex to vertex.For catalysis, the most pertinent feature of the host is the

microenvironment within. Like many enzymatic active sites, thismicroenvironment excludes the bulk solution by virtue of thetightly geared hydrocarbon walls. The 1,5-substitution pattern ofthe naphthalene in 1 sets the metal-binding catecholamide(CAM)moieties apart, enabling the arene walls of 2 to “breathe”via amide bond rotation about the C−N bond to accommodateguests of variable sizes and shapes, including spherical, prolate,and cylindrical, as evidenced by NMR spectroscopy, highresolution mass spectrometry, UV−vis spectroscopy, and single-crystal X-ray diffractometry.12 Below the size exclusion limit,cationic guests are often strongly bound (log Ka up to 4.61) dueto electrostatic enthalpic and solvation-related entropic drivingforces, while neutral guests are bound by the hydrophobiceffect.14,15 The exceptional flexibility of 2 bears consequencesfor the kinetics andmechanism of host−guest association. Guestencapsulation and self-exchange are known to proceed throughan aperture dilation mechanism, rather than host rupture bypartial ligand dissociation.16 This mechanism constitutes anadditional parallel with biological systems, wherein enzymesaccommodate their substrates by architectural deformation andconfigurationally adaptive binding.A useful metric for studying enzyme structures is the rates of

amide hydrogen−deuterium exchange. These exchange kineticsmeasurements provide a quantitative characterization of thestereoelectronic factors that influence dynamic protein−waterinteractions, as well as hydrophobic, noncovalent interactionswithin the active site. Although 2 is water-soluble, themicroenvironment of this host is highly hydrophobic due tothe naphthalene walls. To compare with proteins, we studiedamide hydrogen−deuterium exchange kinetics of 2 at variouspD values and with different guests (Figure 2).17 Theseexperiments confirmed that the hydrophobic character of theinterior mimics those of well-isolated active sites and revealedthat the internalized amide protons of 2 react with encapsulatedwater through acid-, base-, and water-mediated mechanisms.The acid-mediated pathway is significantly favored compared tonatural systems due to the high anionic charge of the host. Inmany biological systems, rates of proton transfer defineenzymatic activities and correlate with the degree of solventexclusion. This study not only divulged proton transfer reactionswithin 2 that mimic enzymes, but also revealed that thistechnique is useful for studying synthetic catalysts.

Figure 1. (a) Crystal structure of tetrahedral host 2. (b) Schematicrepresentation of host 2 [K12(Ga416)].

Accounts of Chemical Research Article

DOI: 10.1021/acs.accounts.8b00328Acc. Chem. Res. 2018, 51, 2447−2455

2448

■ REACTIVITY INSIDE THE HOSTAn understanding of the host structure and the driving forces forguest encapsulation provided a framework for the pursuit ofreaction promotion within the host. Specifically, the hydrogen−deuterium exchange studies revealed the potential to promotereactions of guests by acid−base chemistry, while the highaffinity for monocationic guests suggested that such inter-mediates may be stabilized. Indeed, effective pKa shifts of up to4.5 units were observed in encapsulated amine and phosphineBrønsted bases via significant degrees of protonation despite thebasic exterior bulk solution (Figure 3a).18 The indication that 2

drives the formation of monocationic species coupled with itssolvent exclusion led us to attempt shifts in other intrinsicequilibria by encapsulation. This outcome was achieved whenamines and aldehydes, which do not generate observablequantities of condensed iminium ions in aqueous conditions,were found to form quantitatively within 2 (Figure 3b).19

Similarly, phosphines were observed to react with ketones togenerate β-hydroxyphosphonium ions, which are undetected inbulk solution (Figure 3c).20

Next, the utilization of confinement effects within 2 wereprobed. The observation that guests bearing amide bonds havetheir barrier of C−N bond rotation lowered by up to 3.6 kcal/mol upon encapsulation in 2 was ascribed to stabilization of thedipole-minimized transition state by the hydrophobic interior ofhost 2 (Figure 4a).21 In contrast, C−C bond rotations inbenzylphosphines were found to have their barriers raised by 3−6 kcal/mol upon encapsulation within 2 (Figure 4b).22 Thisdifference in barrier correlated with guest size and shape and wasattributed to conformational “locking” effect on the guest uponencapsulation.

We next demonstrated that 2 could promote electro-cyclizations and sigmatropic rearrangements via molecularconfinement and constriction. These transition states aregenerally less polar than their precursors, and encapsulationconfines the substrate in a conformation resembling thetransition state relative to the linear conformation in bulksolution, destabilizing the ground state. We observed thatcationic allyl-enammonium substrates such as 3 are readilyencapsulated and can undergo aza-Cope rearrangements withrate accelerations of up to 850-fold.23 Furthermore, rapidhydrolysis of products such as 4 to corresponding aldehyde 5enabled turnover of 2, which was further developed into an earlyexample of true catalysis within the host (Figure 5).

Beyond rate accelerations, nature is extremely proficient atasymmetric transformations. While proteins owe their inherentchirality to the biased availability of L-amino acids, 2 iscomposed of achiral components. However, metal−triscatecho-lates occupy Δ or Λ configurations, and excellent mechanicalcoupling among all vertices leads to the formation of 2 asracemic mixtures ofΔΔΔΔ andΛΛΛΛ hosts.13 In early studies,we had observed that 2 is capable of enantiodiscrimination, aseach enantiomer of the host generates diastereomericallydistinct host−guest pairs with either R or S guest (Figure 6).To exploit this discrimination for asymmetric transformations, aprotocol was developed for the optical resolution of 2.24 Aza-Cope rearrangements such as those discussed above could beperformed by ΔΔΔΔ-2 with enantiomeric excesses as high as78%, demonstrating enantioinduction despite the lack of specificsubstrate−catalyst interactions.25A final example of stoichiometric reactivity in 2 highlights an

unprecedented photochemical reaction that is inaccessible inbulk solution. Upon encapsulation, cinnamylammonium sub-strates such as 6 rearrange from the lower energy linear isomer tothe higher energy branched isomer 7 upon photoexcitation of 2(Figure 7a).26 This unexpected reactivity arises from thecooperation of solvent exclusion and constrictive binding effectsthat facilitate and stabilize radical intermediates that form upon

Figure 2. Internal amide hydrogen bonds monitored for H−Dexchange.

Figure 3. (a) Effective pKa shifts of Brønsted bases. (b) Formation ofunfavored iminium ions. (c) Formation of unfavored phosphonium−ketone adducts.

Figure 4. (a) Acceleration of aryl amide bond rotations. (b) HinderedC−C bond rotations in benzyl phosphoniums.

Figure 5. Constrictive binding effects within 2 promote the aza-Coperearrangement of allyl enammonium ions.

Accounts of Chemical Research Article

DOI: 10.1021/acs.accounts.8b00328Acc. Chem. Res. 2018, 51, 2447−2455

2449

photoinduced electron transfer (PET) from host 2 to thesubstrate. The proposed mechanism, derived from spectro-scopic experiments (UV−vis absorption, fluorescence, andtransient absorption) and cyclic voltammetry, stipulates thathost 2 is initially excited by incoming light to generate theexcited charge transfer state 6⊂2*, which acts as the PET agentfor the 1,3-rearrangement. Donation of an electron from excited2* to acceptor 6 induces heterolytic C−N bond cleavage,generating free amine and radical ion pair 8. Although there is

some evidence that this amine can reversibly encapsulate, fastdecay of the radical ion pair by electron donation back to thehost generates carbenium ion 9, which is rapidly sequestered bythe amine to give product 7 (Figure 7b). Though this reaction iscurrently the only example of photosensitization via 2, itrepresents a new direction of research for our collaboration andjoins a growing field of synergy between photochemical andsupramolecular chemistry.27

■ DIVERGENT CATALYSIS WITHIN THE HOSTThe insights revealed by these early reactivity studies providedthe context and incentive for further advancement of 2 as acatalyst. As a result of continued work, 2 currently boasts aprolific portfolio as a catalyst in which reactions are promotedwith significant rate accelerations and excellent catalyticturnovers.1 Of these, perhaps the most intriguing cases arethose that feature divergent reactivities. Here, we discuss threesuch recent examples where newmodes of reactivity are enabledwithin the host that yield products that do not form in detectableamounts from analogous transformations in bulk solution.Though some advancements are owed to serendipity, wedescribe the rational approaches that led to explorations of thesereactivities. The factors that promote these divergent productsare generally biomimetic, drawing from solvent exclusion,confinement effects, and chiral discrimination propertiesrevealed by the early work discussed above.The first example is the 2-catalyzed Prins cyclization of

citronellal and its derivatives.28 Initial inspiration for this workwas the broad class of terpene synthases, which convert terpenesubstrates into numerous natural products ranging from simplemolecules to complex polycyclic structures. Terpene synthasesare distinguished by 1,5-diene cyclizations performed viacationic carbenium intermediates and remarkable degrees ofproduct selectivity achieved by substrate preorganization.Furthermore, these enzymes rely on a controlled terminationof such carbocation cascades via deprotonation or nucleophiliccapture by water, which is in turn governed by a controlleddegree of solvent exclusion.Given that substrate preorganization could facilitate the aza-

Cope rearrangement, we turned our attention to cyclizationreactions similar to those performed by terpene synthases. Thedemonstrated driving force for the generation of monocationiccharge within 2 was hypothesized to favor the formation ofcarbenium ions upon a nucleophilic olefin attack, which mightbe driven by amplified protonation of a heteroatom substrate. Byanalogy to dramatic pKa shifts in amines and phosphines, wepredicted that oxygen bases may undergo such shifts in pKa. To

Figure 6. Enantiodiscrimination of chiral guests by each enantiomer of2 leads to diastereomeric host−guest pairs.

Figure 7. Proposed mechanism for PET-induced rearrangements ofcinnamylammonium substrates within 2.

Figure 8. Catalysis of the Prins cyclization by 2 forms alkene products, in contrast to bulk solution acid catalysis.

Accounts of Chemical Research Article

DOI: 10.1021/acs.accounts.8b00328Acc. Chem. Res. 2018, 51, 2447−2455

2450

explore this possibility, terpene-based citronellal 10 was heatedwith catalytic amounts of 2 under basic conditions, and indeed,cyclized products were observed. Specifically, unsaturatedcyclization product 11 was catalytically formed via deprotona-tion as the carbenium ion terminating step. This generates a newalkene by means similar to terpene synthases, which reinforcesthe notion of excellent water exclusion within 2. Uncatalyzedcyclization was very minimal in the absence of 2 (kcat/kuncat of upto 1.9 × 105) without acidification of the bulk solution to pH <4.29 However, the uncatalyzed reactivation under acidicconditions favored capture of the carbenium ion by water togenerate the hydrated product 12 rather than 11 (Figure 8).Thus, 2 was shown to effectively divert the acid catalysis of aPrins reaction to give unsaturated products from basic solutiondespite the abundance of bulk water.Inspired by the success of the 2-catalyzed Prins cyclization, we

sought to expand this reaction manifold and discovered asurprising divergence in product selectivities. We hypothesizedthat akin to the pKa shifts of aldehydes in the Prins cyclization,the stabilization of transient iminium ions could be exploited toaffect a similar olefin capture event, followed by protonelimination. To explore this possibility, amine-tethered olefinsubstrates such as 12 were subjected to catalytic amounts of 2and excess formaldehyde.30 As expected, cyclized products weregenerated via nucleophilic olefin capture; however, the productswere unanticipated dealkylated piperidines such as 13 (Figure9a)! Further experiments supported a mechanistic rationalewhere iminium adduct 14 is captured by 2, followed by rapidnucleophilic alkene attack. The resulting carbenium ion 15,rather than being quenched by water or deprotonation,undergoes an unprecedented transannular 1,5-hydride shiftdue to confinement effects to generate iminium products, whichhydrolyze upon egress to furnish piperidine 13 (Figure 9b).Kinetic analysis of the reaction revealed that encapsulation wasthe rate-limiting step of the reaction, rather than the Michelis−Menten profile that most supramolecular catalysts take. Thisstudy provided the first observation of rate-limiting encapsula-

tion, which is attributed to the fast reactivity of the constrainediminium adduct. Similar to the 2-catalyzed Prins cyclization, thedriving force for monocationic charge, confinement effects, andexcellent solvent exclusion properties led to divergent catalysiswithin 2. However, the strong steric influence on internalizedconfigurations of 14 resulted in the serendipitous discovery ofthis unusual transannular 1,5-hydride shift, which has otherwisenever been observed in simple substrates.The final example highlights an unusual divergence in

stereochemical, rather than regiochemical outcomes. As one ofthe fundamental transformations in organic chemistry, nucleo-philic substitution at carbon sp3 centers is described in thecontext of two classic mechanisms, SN1 and SN2. For eithermechanism, there is an associated stereochemical outcome, withSN2 being stereospecific for inversion and SN1 mechanismsresulting in stereochemical erosion due to the intermediacy of anachiral carbenium ion. Initially, we were interested in utilizingthe stabilization of carbocations and the chirality of these hoststo explore asymmetric substitution reactions. In the course ofthese efforts, we discovered that our tetrahedral hosts catalyzedsolvolysis of trichloroacetimidates with retention of stereo-chemistry, a complete reversal from the inversion of solvolysis inbulk solution!31 While enzymes that excel at stereochemicalcontrol are commonplace, enzymes that can reverse the intrinsicstereochemical course of a reaction are less common, and this isapparently unprecedented in supramolecular catalysis.This work featured the modified tetrahedral host 16, which

substitutes CAM binding moieties for terephthalamides (TAM)functionalized with chiral directing groups (Figure 10). Thechirality of these distal amides directs the Δ or Λ gearing of themetal−triscatecholates such that enantiopure hosts can bedirectly prepared.32 Upon the discovery that solvolysis wasproceeding with stereochemical retention, we sought tounderstand the origin of this stereochemical reversal. Remark-ably, it was observed that either enantiomer of substrate 17 leadsto substantial retention of stereochemistry in both the ether andalcohol products, 18 and 19, and that absolute catalyst chirality

Figure 9. (a) The aza-Prins cyclization of amine-tethered olefins generates unprecedented piperidine products. (b) The proposedmechanism of the 2-catalyzed aza-Prins cyclizations features rate-limiting encapsulation.

Accounts of Chemical Research Article

DOI: 10.1021/acs.accounts.8b00328Acc. Chem. Res. 2018, 51, 2447−2455

2451

has no impact on the product stereochemistry (Figure 11a).This outcome contrasts with solvolysis catalyzed by an achiral

phosphoric acid, suggesting that the stereointegrity of thesubstrate is well-preserved throughout the course of thesubstitution reaction. We propose that there is a strongstabilizing effect at the backside of the developing carbocationby the electron rich naphthalene walls (Figure 11b). Thisstabilization leads to overall stereochemical retention caused byan effective double inversion: the first inversion by the host’snaphthalene wall, followed by a second inversion by theincoming solvent nucleophile. Cation−π interactions are oftenimplicated in many enzymes and supramolecular systems, butthis example constitutes one of the more unusual applications ofthis interaction.

■ THE HOST BEYOND SINGLE-SITE CATALYSISLooking beyond a single catalytic process, supramolecularchemists seek the development of complex, multicomponentsystems as well as deeper insights into the driving forces andmechanisms of microenvironment reactivity.33 Multicompo-nent systems offer the prospect of performing traditionallyincompatible synthetic steps in one pot, similar to biologicalsystems. Here, we present two examples of tandem catalysiswhere either biological or synthetic catalyst partners areintegrated.Our initial foray into tandem catalysis aimed to unify our

supramolecular approach with the reactivity of natural enzymes.We had previously shown that 2 can encapsulate, stabilize, andenhance the reactivity of Me3PAu

+ in the hydroalkoxylation ofalcohol-tethered allenes via encapsulation of the monocationicAu(I) species with solvent exclusion (Figure 12a).34 In this case,2 does not act as the catalyst but instead enhances reactivity andlongevity of Au(I) as the active catalyst complex Me3PAu

+⊂2.We envisioned that esterases and lipases could kinetically resolveand cleave amide- and ester-tethered allene substrates such as

20, which are otherwise unreactive to Me3PAu+⊂2. The

resolved alcohol products such as 21 would then undergoMe3PAu

+⊂2-catalyzed hydroalkoxylation to give enantioen-riched cyclized products 22 (Figure 12b). Indeed, we found thata range of natural enzymes can achieve this sequence withMe3PAu

+⊂2 where both catalysts are necessary for productformation.5 Enzyme and Me3PAu

+ in the absence of 2 led tocatalyst deactivation, indicating that the supramolecularsequestration of Me3PAu

+⊂2 inhibits adverse interactionsbetween free Me3PAu

+ and the enzyme.The successful integration of biological and synthetic

microenvironment catalysts under aqueous conditions encour-aged us to seek further tandem reactivity with transition metalsin 2. We were able to demonstrate such a union following theinitial discovery that 2 catalyzes reductive eliminations fromhigh-valent transition metal complexes. A common challenge incross-coupling chemistry is the judicious selection of a ligand setthat promotes all steps of the catalytic cycle. However, becauseconditions promoting one elementary step often perform at thedetriment of other steps, selective catalysis of one elementarystep is highly desirable.Transition metal complexes are encapsulated by 2 with

concomitant dissociation of an X-type ligand.35 Relatedly, it isknown that kinetically unfavorable reductive eliminations can beaccelerated by halide abstraction. The recognition of cationicintermediates within 2 manifests in the strong encapsulation ofappropriately sized transition metal complexes with concom-itant halide dissociation within 2. We discovered that complexessuch as Pt(IV) 23 and Au(III) 24 generate the cationicencapsulated complexes, leading to dramatic accelerations ofalkyl−alkyl reductive eliminations of up to 1.9 × 107-fold via aMichaelis−Menten-type mechanism (Figure 13a)!36 Notably,these accelerations may also be partially due to constrictivebinding effects, bringing the elimination partners closertogether. This example (at the time of this writing) is thelargest supramolecular rate acceleration to date, resulting fromthe unusual application of a biomimetic strategy to accelerate anentirely synthetic reaction.With a strategy in hand for enabling prohibitive reductive

eliminations, we sought to complete a challenging cross-coupling reaction with tandem dual catalysis.37 The catalyticsystem featuring Pt(IV) complex 23, which is prohibitively slowat reductive elimination of ethane, was successfully activated bysupramolecular tandem catalysis by 16 (Figure 13b). This

Figure 10. (a) Modified tetrahedral host 16, which features chiralterephthalmides that form enantiopure hosts. (b) The proposedreaction intermediate, which features a cation−π interaction between anaphthalene wall and the transient carbocation. Adapted withpermission from ref 31. Copyright 2014 American Chemical Society.

Figure 11. Solvolysis is catalyzed by either enantiomer of host 16,leading to retention of product stereochemistry rather than inversion.

Figure 12. (a) Me3PAu+⊂2 is an improved catalyst for hydro-

alkoxylation of alcohol-tethered allenes. (b) Kinetic resolution andcyclization of ester-tethered allenes via an integrated enzyme−supramolecular catalyst approach.

Accounts of Chemical Research Article

DOI: 10.1021/acs.accounts.8b00328Acc. Chem. Res. 2018, 51, 2447−2455

2452

unique approach to dual catalysis is distinct from enzymaticexamples, where two independent catalysts cooperate totransform one starting material. Instead, an arrested transitionmetal catalyst can be consumed as a substrate in a supra-molecular approach to catalysis of an elementary step, whichrestores overall catalytic activity in the transition metal catalyst.

■ THE HOST STRUCTURE AS A PROBE

A commonality between enzymes and supramolecular catalystsis the specificity of structure to reactivity. Thus, structure−activity relationship (SAR) studies of supramolecular catalystsoffer the opportunity to gain insights into the factors thatinfluence microenvironment catalysis and draw connectionsbetween specific structural features and reactivities. However,structural variation of self-assembled systems is significantlychallenging and there is a subsequent dearth of supramolecularSAR studies. We discuss here two examples that demonstratehow diversified tetrahedral hosts enable mechanistic studies toprovide fundamental insights into supramolecular reactivity.Studies such as these are rare opportunities to isolate anddeconvolute specific governing features of supramolecularreactivity, such as electrostatic effects, confinement effects, andhost substitution effects.Toward structural diversification of these hosts, we have

introduced a variation in chelators (CAM-2 to TAM-16) as wellas a variation in spacer size (naphthalene-2 to pyrene-24, andTAM-16 to TAM-pyrene-25) (Figure 14).29 In conjunctionwith structurally diversified substrates, this set of hosts was usedto study the terpene synthase-like Prins cyclization in furtherdepth. Although the chelator was not found to impact catalysisrate significantly, increases in spacer size led to improvedcatalytic efficiencies and selectivities for certain stereoisomers ofproduct. The trends that could be extracted by variation ofcatalysts as well as substrates enabled a better understanding ofsupramolecular microenvironment catalysis of the Prinscyclization in terms of chemo-, diastereo-, and enantioselectiv-ities.While studies regarding the effect of size, constrictive binding,

and binding moieties are important for understanding stericparameters of microenvironment catalysis, another unmetchallenge lies in characterizing the effect of electric fields and

Coulombic forces. The stabilization of monocationic charge hasbeen a staple in driving reactivity in 2, and this stabilization isascribed to the electrostatic effect resulting from its dodeca-anionic charge. Although electrostatic charge is implicated in awide range of enzymatic and supramolecular catalyst systems,the complete isolation and study of this effect is rare.38 Toexperimentally probe electrostatic charge and its implication inour catalysis, an isostructural octaanionic Si(IV)-based analog26 was developed and shown to be identical to 2 in the solidstate as well as the solution phase (Figure 15a).39 The similarityin structure between 26 and 2 was further supported by theidentical rate accelerationsmeasured in both 2- and 26-catalyzedaza-Cope rearrangements, which are net neutral and driven byconstrictive binding. In contrast, a Nazarov cyclization, which

Figure 13. (a) Challenging reductive eliminations are catalyzed from high-valent transition metal complexes. (b) A dual catalytic cross-couplingreaction enabled by a supramolecular approach to catalyzed reductive elimination.

Figure 14. Diversification of host 2 by modification of chelators andspacers.

Accounts of Chemical Research Article

DOI: 10.1021/acs.accounts.8b00328Acc. Chem. Res. 2018, 51, 2447−2455

2453

features protonation of substrate 27 and the generation of acarbocation, showed a 680-fold difference in rates of catalysis,standing as the first experimental probe for the dramatic effect ofanionic host charge in cation-mediated catalysis (Figure 15b).Moreover, the parallels between enzymatic actives sites andsupramolecular microenvironments generated an opportunityfor these studies to inform on the governing principles ofenzymatic catalysis by investigation of supramolecular catalysts.

■ CONCLUDING REMARKSThe recognition of how supramolecular microenvironments canisolate and influence molecular guests has been the key to thedevelopment of a wide range of synthetic catalysis, as well asmore complex systems such as tandem reaction cascades,integration with biological systems, and the development ofmechanistic tools to understand supramolecular reactivity. Inparticular, a thorough understanding of what drives molecularrecognition enabled us to leverage those stabilizing forces topromote and catalyze many reactions by ground statedestabilization effects, as well as transition state stabilization.These manifest as the recognition and electrostatic stabilizationof monocationic charge, excellent solvent exclusion, constrictivebinding effects, and compatibility with basic aqueous media. Wehave been fortunate to study a host that is amenable to theseapplications and features unique properties to dissect andexploit. As available supramolecular hosts grow in number, weanticipate that the field will progress with great strides withapplications surpassing those reported thus far.

■ AUTHOR INFORMATIONCorresponding Authors

*E-mail: rbergman@berkeley.edu.

*E-mail: raymond@socrates.berkeley.edu.*E-mail: fdtoste@berkeley.edu.

ORCID

Cynthia M. Hong: 0000-0002-2563-219XF. Dean Toste: 0000-0001-8018-2198Notes

The authors declare no competing financial interest.

Biographies

Cynthia M. Hong, Texan, earned her B.A./M.S. in chemistry fromNorthwestern University in 2013 before moving west to the Universityof California, Berkeley. There, she joined the collaborative supra-molecular project to obtain her Ph.D. in 2018 under the supervision ofProf. KennethN. Raymond and Prof. F. Dean Toste. She now continuesher career in chemistry at Merck & Co. in Rahway, NJ.

Robert G. Bergman received his B.A. at Carleton College and his Ph.D.with Jerome A. Berson at the University of Wisconsin. After apostdoctoral study with Ronald Breslow at Columbia, he began hisacademic career at the California Institute of Technology in 1967. Hemoved his research group to Berkeley in 1978, where he has held jointappointments at the University of California, Berkeley, and theLawrence Berkeley National Laboratory. His research interests lie atthe interface between organic and inorganic chemistry, with a primaryfocus on catalysis, supramolecular chemistry, and the study of reactionmechanisms.

Kenneth N. Raymond obtained a B.A. in chemistry at Reed College in1964 and Ph.D. from Northwestern University in 1968 and began hisfaculty appointment at the University of California at Berkeley in 1967.He served as Vice Chair of the Berkeley Chemistry Department (1982−1984 and 1999−2000) and Chair (1993−1996). He has received manyawards and honors, including election to the National Academy ofSciences and the American Academy of Arts and Sciences. He is acofounder (2001) of Lumiphore Inc. He is the author of 24 US Patents,11 International Patents, and 572 research publications.

F. Dean Toste received his B.Sc. and M.Sc. from the University ofToronto and completed his Ph.D. studies at Stanford University underthe guidance Professor Barry Trost. After a postdoctoral appointmentwith Professor Robert Grubbs at the California Institute of Technology,he took an Assistant Professorship at the University of California,Berkeley, in 2002. In 2006, he was promoted to Associate Professor andis currently Gerald E. K. Branch Distinguished Professor of Chemistry.

■ REFERENCES(1) Brown, C. J.; Toste, F. D.; Bergman, R. G.; Raymond, K. N.Supramolecular Catalysis in Metal−Ligand Cluster Hosts. Chem. Rev.2015, 115, 3012−3035.(2) Mal, P.; Breiner, B.; Rissanen, K.; Nitschke, J. R. WhitePhosphorus Is Air-Stable Within a Self-Assembled TetrahedralCapsule. Science 2009, 324, 1697−1699.(3) Anslyn, E. V. Supramolecular and Chemical Cascade Approachesto Molecular Sensing. J. Am. Chem. Soc. 2010, 132, 15833−15835.(4) Pilgrim, B. S.; Roberts, D. A.; Lohr, T. G.; Ronson, T. K.; Nitschke,J. R. Signal Transduction in a Covalent Post-Assembly ModificationCascade. Nat. Chem. 2017, 9, 1276−1281.(5) Wang, Z. J.; Clary, K. N.; Bergman, R. G.; Raymond, K. N.; Toste,F. D. A Supramolecular Approach to Combining Enzymatic andTransition Metal Catalysis. Nat. Chem. 2013, 5, 100−103.(6) Hong, C. M.; Kaphan, D. M.; Bergman, R. G.; Raymond, K. N.;Toste, F. D. Conformational Selection as the Mechanism of GuestBinding in a Flexible Supramolecular Host. J. Am. Chem. Soc. 2017, 139,8013−8021.

Figure 15. (a) Octaanionic Si(IV) catalyst 26 is an isostructural analogto dodecaanionic catalyst 2. (b) Differences in catalytic activities giveinsights on the role of anionic charge in catalysis.

Accounts of Chemical Research Article

DOI: 10.1021/acs.accounts.8b00328Acc. Chem. Res. 2018, 51, 2447−2455

2454

(7) Stross, A. E.; Iadevaia, G.; Nunez-Villanueva, D.; Hunter, C. A.Sequence-Selective Formation of Synthetic H-Bonded Duplexes. J. Am.Chem. Soc. 2017, 139, 12655−12663.(8) Raynal, M.; Ballester, P.; Vidal-Ferran, A.; van Leeuwen, P. W. N.M. Supramolecular Catalysis. Part 2: Artificial Enzyme Mimics. Chem.Soc. Rev. 2014, 43, 1734−1787.(9) Raynal, M.; Ballester, P.; Vidal-Ferran, A.; van Leeuwen, P. W. N.M. Supramolecular Catalysis. Part 1: Non-Covalent Interactions as aTool for Building and Modifying Homogeneous Catalysts. Chem. Soc.Rev. 2014, 43, 1660−1733.(10) Kang, J.; Rebek, J. Acceleration of a Diels−Alder Reaction by aSelf-Assembled Molecular Capsule. Nature 1997, 385, 50−52.(11) Kaanumalle, L. S.; Gibb, C. L. D.; Gibb, B. C.; Ramamurthy, V.Controlling Photochemistry with Distinct Hydrophobic Nanoenviron-ments. J. Am. Chem. Soc. 2004, 126, 14366−14367.(12) Pluth, M. D.; Johnson, D. W.; Szigethy, G.; Davis, A. V.; Teat, S.J.; Oliver, A. G.; Bergman, R. G.; Raymond, K. N. StructuralConsequences of Anionic Host−Cationic Guest Interactions in aSupramolecular Assembly. Inorg. Chem. 2009, 48, 111−120.(13) Caulder, D. L.; Powers, R. E.; Parac, T. N.; Raymond, K. N. TheSelf-Assembly of a Predesigned Tetrahedral M4L6 SupramolecularCluster. Angew. Chem., Int. Ed. 1998, 37, 1840−1843.(14) Sgarlata, C.; Mugridge, J. S.; Pluth, M. D.; Zito, V.; Arena, G.;Raymond, K. N. Different and Often Opposing Forces Drive theEncapsulation and Multiple Exterior Binding of Charged Guests to aM4L6 Supramolecular Vessel in Water. Chem. - Eur. J. 2017, 23,16813−16818.(15) Biros, S. M.; Bergman, R. G.; Raymond, K. N. The HydrophobicEffect Drives the Recognition of Hydrocarbons by an Anionic Metal−Ligand Cluster. J. Am. Chem. Soc. 2007, 129, 12094−12095.(16) Davis, A. V.; Raymond, K. N. The Big Squeeze: Guest Exchangein an M4L6 Supramolecular Host. J. Am. Chem. Soc. 2005, 127, 7912−7919.(17) Hart-Cooper, W. M.; Sgarlata, C.; Perrin, C. L.; Toste, F. D.;Bergman, R. G.; Raymond, K. N. Protein-like Proton Exchange in aSynthetic Host Cavity. Proc. Natl. Acad. Sci. U. S. A. 2015, 112, 15303−15307.(18) Pluth, M. D.; Bergman, R. G.; Raymond, K. N. Making AminesStrong Bases: Thermodynamic Stabilization of Protonated Guests in aHighly-Charged Supramolecular Host. J. Am. Chem. Soc. 2007, 129,11459−11467.(19) Dong, V. M.; Fiedler, D.; Carl, B.; Bergman, R. G.; Raymond, K.N.Molecular Recognition and Stabilization of Iminium Ions inWater. J.Am. Chem. Soc. 2006, 128, 14464−14465.(20) Brumaghim, J. L.; Michels, M.; Raymond, K. N. HydrophobicChemistry in Aqueous Solution: Stabilization and StereoselectiveEncapsulation of Phosphonium Guests in a Supramolecular Host. Eur.J. Org. Chem. 2004, 2004, 4552−4559.(21) Pluth, M. D.; Bergman, R. G.; Raymond, K. N. Acceleration ofAmide Bond Rotation by Encapsulation in the Hydrophobic Interior ofa Water-Soluble Supramolecular Assembly. J. Org. Chem. 2008, 73,7132−7136.(22) Mugridge, J. S.; Szigethy, G.; Bergman, R. G.; Raymond, K. N.Encapsulated Guest−Host Dynamics: Guest Rotational Barriers andTumbling as a Probe of Host Interior Cavity Space. J. Am. Chem. Soc.2010, 132, 16256−16264.(23) Fiedler, D.; Bergman, R. G.; Raymond, K. N. SupramolecularCatalysis of a Unimolecular Transformation: Aza-Cope Rearrangementwithin a Self-Assembled Host. Angew. Chem. 2004, 116, 6916−6919.(24) Davis, A. V.; Fiedler, D.; Ziegler, M.; Terpin, A.; Raymond, K. N.Resolution of Chiral, Tetrahedral M4L6 Metal−Ligand Hosts. J. Am.Chem. Soc. 2007, 129, 15354−15363.(25) Brown, C. J.; Bergman, R. G.; Raymond, K. N. EnantioselectiveCatalysis of the Aza-Cope Rearrangement by a Chiral SupramolecularAssembly. J. Am. Chem. Soc. 2009, 131, 17530−17531.(26) Dalton, D. M.; Ellis, S. R.; Nichols, E. M.; Mathies, R. A.; Toste,F. D.; Bergman, R. G.; Raymond, K. N. Supramolecular Ga4L612−Cage Photosensitizes 1,3-Rearrangement of Encapsulated Guest via

Photoinduced Electron Transfer. J. Am. Chem. Soc. 2015, 137, 10128−10131.(27) Ward, M. D. Photo-Induced Electron and Energy Transfer inNon-Covalently Bonded Supramolecular Assemblies. Chem. Soc. Rev.1997, 26, 365−375.(28) Hart-Cooper, W. M.; Clary, K. N.; Toste, F. D.; Bergman, R. G.;Raymond, K. N. Selective Monoterpene-like Cyclization ReactionsAchieved by Water Exclusion from Reactive Intermediates in aSupramolecular Catalyst. J. Am. Chem. Soc. 2012, 134, 17873−17876.(29) Hart-Cooper, W. M.; Zhao, C.; Triano, R. M.; Yaghoubi, P.;Ozores, H. L.; Burford, K. N.; Toste, F. D.; Bergman, R. G.; Raymond,K. N. The Effect of Host Structure on the Selectivity andMechanism ofSupramolecular Catalysis of Prins Cyclizations. Chem. Sci. 2015, 6 (2),1383−1393.(30) Kaphan, D. M.; Toste, F. D.; Bergman, R. G.; Raymond, K. N.Enabling New Modes of Reactivity via Constrictive Binding in aSupramolecular-Assembly-Catalyzed Aza-Prins Cyclization. J. Am.Chem. Soc. 2015, 137, 9202−9205.(31) Zhao, C.; Toste, F. D.; Raymond, K. N.; Bergman, R. G.Nucleophilic Substitution Catalyzed by a Supramolecular CavityProceeds with Retention of Absolute Stereochemistry. J. Am. Chem.Soc. 2014, 136, 14409−14412.(32) Zhao, C.; Sun, Q.-F.; Hart-Cooper, W. M.; DiPasquale, A. G.;Toste, F. D.; Bergman, R. G.; Raymond, K. N. Chiral Amide DirectedAssembly of a Diastereo- and Enantiopure Supramolecular Host and ItsApplication to Enantioselective Catalysis of Neutral Substrates. J. Am.Chem. Soc. 2013, 135, 18802−18805.(33) Ueda, Y.; Ito, H.; Fujita, D.; Fujita, M. Permeable Self-AssembledMolecular Containers for Catalysts Isolation Enabling Two-StepCascade Reactions. J. Am. Chem. Soc. 2017, 139, 6090−6093.(34) Wang, Z. J.; Brown, C. J.; Bergman, R. G.; Raymond, K. N.;Toste, F. D. Hydroalkoxylation Catalyzed by a Gold(I) ComplexEncapsulated in a Supramolecular Host. J. Am. Chem. Soc. 2011, 133,7358−7360.(35) Fiedler, D.; Leung, D. H.; Bergman, R. G.; Raymond, K. N.Enantioselective Guest Binding and Dynamic Resolution of CationicRuthenium Complexes by a Chiral Metal−Ligand Assembly. J. Am.Chem. Soc. 2004, 126, 3674−3675.(36) Levin, M. D.; Kaphan, D. M.; Hong, C. M.; Bergman, R. G.;Raymond, K. N.; Toste, F. D. Scope andMechanism of Cooperativity atthe Intersection of Organometallic and Supramolecular Catalysis. J. Am.Chem. Soc. 2016, 138, 9682−9693.(37) Kaphan, D. M.; Levin, M. D.; Bergman, R. G.; Raymond, K. N.;Toste, F. D. A Supramolecular Microenvironment Strategy forTransition Metal Catalysis. Science 2015, 350, 1235−1238.(38) Burschowsky, D.; van Eerde, A.; Okvist, M.; Kienhofer, A.; Kast,P.; Hilvert, D.; Krengel, U. Electrostatic Transition State StabilizationRather than Reactant Destabilization Provides the Chemical Basis forEfficient Chorismate Mutase Catalysis. Proc. Natl. Acad. Sci. U. S. A.2014, 111, 17516−17521.(39) Hong, C. M.; Morimoto, M.; Kapustin, E. A.; Alzakhem, N.;Bergman, R. G.; Raymond, K. N.; Toste, F. D. Deconvoluting the Roleof Charge in a Supramolecular Catalyst. J. Am. Chem. Soc. 2018, 140,6591−6595.

Accounts of Chemical Research Article

DOI: 10.1021/acs.accounts.8b00328Acc. Chem. Res. 2018, 51, 2447−2455

2455