Assembly and Encapsulation with Self-complementary Molecules Julius Rebek, Jr.+ Massachusetts Institute of Technology, Department of Chemistry, Cambridge, MA 02139, USA 1 introduction Molecular recognition is a branch of chemistry that is concerned with complementarity. By this term is meant complementarity of size, and shape and chemical surfaces: the 'goodness of fit' between two molecules like the fit between a foot and a hand-made shoe. Molecular recognition defines this goodness of fit, and explores the intermolecular forces, the weak attractions that act over short dis- tances between molecules. These forces -hydrogen bonds, aro- matic stacking, polar and van der Waal's interactions -are the ones that bring molecules together into complexes. Such complexes are temporarily and weakly held groups of two or more molecules. More complicated complexes -assemblies -are also possible.These structures may be made up of complementary molecules or multiple copies of a self-complementary molecule. They form and dissipate on a timescale that varies from microseconds to hours: time intervals long enough for many types of chemistry to occur between them. Assemblies have captured the attention of many research groups involved in molecular recognition because new phenomena fre- quently emerge whenever more than one copy of an entity is present. It is often as impossible to predict what emerges as it is to predict the hexagonal shape of a honeycomb from the study of a single bee. Biochemistry abounds with examples: allosteric enzymes, channel-forming peptides and viral coat proteins are all assemblies of multiple copies of a molecule that give rise to super- structures with functions that are unique to their assembled states: functions such as regulation, transport, replication and encapsula- tion.But simpler molecules, available through chemical synthesis, are also able to exhibit unique behaviour through assembly. This review is concerned with such molecules and is limited in its scope to systems that show sharply defined features of size and shape in solution. Specifically excluded are aggregates such as micelles, liquid crystals and the like as well as assemblies that emerge pri- marily in the solid state. A two-dimensional system of three molecules illustrates some of the features of self-complementarity involved in assembly.The structure designed and synthesized by Zimmerman (Fig. 1)' pre-sents a pattern of hydrogen bond donors and acceptors on one edge + Present address The Skaggs Institute for Chemical Biology, at the Scnpps Research Institute, 10550North Torrey Pines Road, La Jolla, California, 92037, USA Julius Rebek, Jr. received his undergraduate education at the University of Kansas 1966, and obtained the PhD degree fiom the Massachusetts Institute of Technology (1970) for studies in peptide chemistry with Professor D. S. Kemp. At the University of California at Los Angeles he used the 'three-phase test' to detect reactive intermediates. In 1976 he moved to the University of Pittsburgh where new systems for molecular recogni- tion were developed.In I989 he returned to M.I.T. and devised self-replicating and selj-assem- bling molecules. In June of 1996, he moved to La Jolla where he is the director of the new Skaggs Institute for Chemical Biology at Scripps. that is complementary to the pattern on the other functioning edge. Accordingly, assembly at one level can be predicted. The orienta- tion in space of the atoms capable of hydrogen bonding at the two edges of the molecule is fixed at almost exactly 120"Cby the rigid- ity of the aromatic centrepiece of the structure. The information, the code, for the assembly is written into the hydrogen bond pat- terns of the edges and their angular orientations with respect to each other.In solution a trimer is formed, and assembly takes place in the manner expected. But entropy also has its say: linear aggre- gates intrude on the well-ordered cyclic trimer. A cooperative, self-assembling trimer m R 3 H 'R R and R Figure 1 Self-complementary hydrogen bonding sites lead to trimers and oligomers Two-dimensional hydrogen bonded arrays are enormously popular in supramolecular chemistry? For the most part, the systems use rigid, flat, heterocyclic compounds. The resulting assemblies resemble practically infinite sheet^;^ crinkled tapes;4 twisted ribbon^;^ ingenious rosettes: to name just a few. The principles of molecular recognition can also be used to assemble three-dimensional structures, but additional structural requirements must be met.The most important is curvature. This is beautifully illustrated with the constructs of Ghadiri? These structures take the self-complementary hydrogen bonding pat- terns featured by peptide p sheets and wrap them into a macro- cyclic format (Fig. 2). The resulting assembly is a hollow cylinder, a peptide nanotube, and is an early example of an assem- bly that actually shows function. Transport of glucose through the nanotube and across a lipid bilayer membrane was demon- strated. The systems we have worked with at MIT are literally mini-malistic: they involve only two copies of the same molecule. But it is the single molecule that holds the key, loaded with the information and what it promises for assembly.Rene Wyler synthesized the first CHEMICAL SOCIETY REVIEWS, 1996 I I Complementary distances between acids and bases 4 )r 0‘ ‘0 \ Cyclizing a bsheet structure... gives a self-assembled peptide nanotube Figure 2 Curvature increases the complexity of self-complementary assemblies such molecule la from diphenylgl ycoluril and tetrabromodurene.8 The cis-fusion of the two five-membered rings of the glycoluril forces a fold at either end of the molecule. Curvature along the length of the skeleton is caused by seven-membered rings on either side of the central benzene unit. When all of the glycoluril substituents appear on the same face of the molecule, a low energy conformation that features self-complementarity can be obtained.Specifically, the 0-0 distance indicated is appropriate for two hydrogen bonds to form to the ends of another molecule. The stereochemical features impart the necessary curvature for the dimeric assembly, while the hydrogen bonds provide the cohesive force, just as the stitching along the seam of a baseball holds the two pieces together. Spectroscopic studies in CDCI, indicated that a single species was present in solution which showed extensive hydrogen bonding. We were fortunate that this solvent (the industrial standard for studies in molecular recognition) was a poor fit for the interior of the capsule, as most of the encapsulation phenomena described in this review owe their very existence to two characteristics of a solvent. Foremost, it must dissolve reasonable concentrations of the assembly’s components.Secondly, it must compete so poorly for the interior of the capsule that the encapsulation of solute guest molecules can be observed. In other words, the solvent’s superior concentration -relative to an intended guest’s -must be overcome by the goodness of fit. This observation was made earlier by Still and used with tremendous success by Cram in his incarceration of ‘convicts’ in covalently bound capsules .9 The IH NMR spectrum of this ‘baseball’ (la-la) in CDC1, saturated with methane is shown in Fig. 4to illustrate some of the advantages of NMR for the study of assembly and encapsulation. The spectrum displays a sharp singlet at the unusual chemical shift of 6 -0.9 representing encapsulated CH,,l0 in addition to the resonance for free methane at 6 0.2 (the ethane signal at 6 0.82 and the characteristic pattern of propane at 6 0.9 and 1.3 are a measure of the purity of natural gas ‘on line’ at MIT).The separa- tion of the free and bound signals for both host and guest places limits on the rates of exchange into and out of the capsule: the rates are slow on the NMR timescale but fast on the human time- scale. The crystal structure of dimeric lc was solved by Toledo,II and it is shown, shorn of its ethyl esters in Fig. 5. The crystal of 1.1 appears to contain a disordered guest species in its cavity, although it was not possible to determine unambiguously the identity of the 1-1 Figure 3 Two identical molecules give a closed-shell ‘baseball’ CH,CH, CH4 bound CbCH, bound CH4 1.2 0.8 0.4 -0.0 -0.4 -0.8 PPm Figure 4 Unusual chemical shifts are observed for encapsulated small molecules guest.It was possible, however, to locate and refine a carbon atom at two positions with assigned occupancy factors of 0.55 and 0.45. This leads us to believe that this guest species is a methanol mole- cule with a highly disordered oxygen atom. There are eight (car- bony1 oxygen) good hydrogen bond acceptors along the seam of the capsule, and it is possible that the hydroxy group of methanol is dis- ordered about eight positions. It was desirable to find more soluble versions of the glycoluril subunit and the family of esters available from dihydroxytartaric acid were settled on.Their enhanced solubility in organic media and the possibility of functional group manipulations (e.g.hydrazinolysis) to water-soluble glycolurils are promising. The p-dimethylamino- diphenylglycoluril derivative lb and the capsule derived from it offer control of the assembly process by events that take place on the periphery of the structure rather than on the inside. Encapsulation of xenon in dimethyl formamide (DMF) solution can be directly observed in the i29Xe NMR spectrum.’* Neil Branda showed that the presence of xenon actually causes nucleation, the formation of the capsule. The basic sites of the dimethylamino function are subject to ASSEMBLY AND ENCAPSULATION WITH SELF-COMPLEMENTARY MOLECULES -JULIUS REBEK, JR.Synthesis of glycolurils H"K~-H 0 1-1 Figure 5 Synthesis of soluble glycolurils and the methane-selective capsule protonation with strong acids, and at high acidities the guest is released. Neutralization with bases reverses the processes: the guest is again encapsulated (Fig. 6). The simplest interpretation (though still unproved) is that the multiple positive charges that build up on protonation of the periphery force the two halves of the capsule apart through coulombic repulsion. The influence of acidity on dimeriza- tion suggests that it may be possible to make multicomponent systems in which assembly is fine-tuned to subtle changes in pH.Xavier Garcias showed it was also possible to alter the environ- ment inside the cavity. The hydroquinone 2 and quinone 3 spacers present either electron rich or electron deficient surfaces to encap- sulated guests (Fig. 7). The affinity of some small molecules to the capsules derived from these, compared to the original capsule, are given in Table 1.13 The studies of the fluorinated methanes were inspired by recent calculations by KollmanI4 that predicted CF, to be an appropriate guest for these capsules, but experiment and computation for the affinity of CF, have yet to be reconciled. Qualitative evidence for the encapsulation of nitric oxides in 3-3 was also obtained. 2 Other Shapes and Sizes In molecules 1,2 and 3, the glycoluril functions at the end of the molecule and the connecting spacer elements, which determine the dimensions and the overall shape of the dimer, remain constant.It was not unexpected to find that heterodimers formed readily when 3:E = C02(~-CqHg) Figure 7 Electron rich and electron poor capsules show altered selectivities Table 1 Association constants KJmol -I I, (298 K), for encapsulation of guests in dimers lc-lc, 2-2, and 3-3 (K,,,, = [N-guest-Nj/[N-N] X [free guest]). Host lc -lc 2-2 3-3 Guest CH4 33 70 10 C,% 51 51 13 CH3F 1.0 17 <0.3 CF4 0.7 0.6 <0.2 the various components were present in the same solution. Can dimerization still take place if the spacers are varied? The monomers 4,5 and 6 contains ethylene, naphthalene and a bridged anthracene, respectively, as their spacer elements.If good hydro- gen bonds are to be formed in the dimers, the 'length' of the spacer should complement the 'width' of the glycoluril. The energy-min- imized structures of the corresponding dimers, as generated by an MM2 for~efield,'~ are provided in Fig. 8, and the calculated 0-0 distances of the monomers are shown.ll The dimerization leads to unusual pseudo-spherical structures with cavities smaller or larger than that of 1-1. Obviously, 4.4 is smaller than 1.1;calculations indicate that the cavity formed by 4.4 (41 A3) is approximately 18% smaller than that of 1.1 (50 A3). Carlos Valdks showed that dimer 4.4 does bind small molecules, and displays a remarkable selectivity: ethane, which binds to 1.1, was not measurably encapsulated by 4.4.However, a price is paid for the discrimination: the affinity of methane for 4.4 is approximately 70 times lower than for 1.1. What are the consequences for 'miscegenation'? Molecular models indicate that three heterodimeric combinations are espe- cially likely to form: hybrid structures 1-4, 1-6 and 5-6.16The energy-minimized heterodimers are shown in Fig. 9. In the experiment, Urs Spitz observed heterodimers when two different Figure 6 Control of encapsulation is possible with changes in acidity 258 4.09 A 0-Y Figure 8 Dimensions and shapes of new capsules dimers were present in the same solution. The formation of hybrid capsules showed that structurally related molecules that are self- complementary are often complementary to each other. The 'recombination' of the homodimers to form hybrids is an equilibrium process, and opens access to new capsule shapes.The hydrogen bonds in the homodimers are expected to be superior to those of the heterodimers; the disproportionation observed must, to some extent, be driven by the entropy of mixing. The recombina- tions (disproportionation equilibria) of the dimers depended strongly on the solvent.ll For example, in the case of 1 with 4, the CHEMICAL SOCIETY REVIEWS, 1996 6.14 A 0--0K 0 6 0 7.17 A 0' '0 disproportionation constant K,, is large in CDCl,, CDBr, and Cl,DCCDCl, but is an order of magnitude smaller in CD,Cl,.The former three solvents are too large to fit the cavity of any of the three species present in the equilibrium, but CD,CI, is of appropriate size for 1.1. The CD,CI, provides a motivation, the driving force, for the formation of 1.1. The goodness of fit again determines the favoured species. Of the various possible capsules, only 6.6 appears large enough to accommodate Cl,DCCDCl,. Accordingly, only traces of hetero- dimer 1.6 are present in solutions of this large solvent. The current ASSEMBLY AND ENCAPSULATION WITH SELF-COMPLEMENTARY MOLECULES-JULIUS REBEK, JR. Figure 9 Hybrid or heterodimeric capsules form when exposed to Complementary guest solvents state of predictions regarding the energetics and geometry of hydro- gen bonds is far from ideal, and to predict the magnitude and the sign of the thermodynamic parameters for hybrids 1-6 and 5.6 without knowledge of the number of species being encapsulated or released is tenuous.3 Larger Volumes Robert Meissner and Jongmin Kang, with the collaboration of Javier de Mendoza developed a new system shown as 7 in Fig. 10.’’ The overall architecture involves a tape-like structure of 13 fused and one bridged ring systems. The ring fusions provide the gentle curvature required for the pseudo-spherical assembly. When viewed on edge, as shown in the Fig. 10, the somewhat exaggerated curvature of the structure can be seen. Compared to the ‘baseball’ the new system is a larger ‘softball’. The compound did show the expected NMR spectroscopic earmarks of a dimer in some aromatic solvents.But in chloroform the molecule produces a gel-like phase. Apparently, a polymeric assembly takes place; the molecule expresses its self-complementarity in an unexpected arrangement. The large solvent [2H,,I-p-xylene permitted the encapsulation of guests of considerable size and shape in the dimeric form. Adamantane is a particularly good fit and even tetramethyl-adamantane can be accommodated. Again, the widely separated and relatively sharp signals for free and encapsulated guest in the NMR spectra indicate that exchange of guests in and out of the capsule is slow on the NMR timescale. The numerous polar and polarizable atoms that line the inner surface can also stabilize complementary functional groups on guests, such as those on adamantaneamine and adamantanedicarboxylic acid.These proved to be the most tightly bound guests. Because higher order aggregates were also observed with 7a, it was necessary to find modifications by which the assembly process 0 0 7a R = phenyl, X = H 7b R = COZ-i-pentyl, X = H 7c R = 4-n-heptylphenyl, X = OH Figure 10 Self-complementary ‘softballs’ feature up to sixteen hydrogen bonds could be better controlled. For example, in 7c additional hydrogen bonding sites are provided by the phenolic functions. As these sites are positioned along the seam and can provide up to eight additional hydrogen bonds to stabilize the capsule, dimerization is favoured over other processes.I The encapsulation behaviour of this system proved surprising. The data for adamantane and ferrocene with 7c-7~in [2H,,]-p-xylene are given in Table 2. Table 2 Thermodynamic parameters for guest encapsulation by the 7c-7~dimer in [2Hlo]-p-xylene. The temperature range was from 298 to 343 K (K = association constant as defined in ref. 18; 1 cal = 4.184 J). AGlkcal AHlkcal ASlcal Guest K,,, mol-' mol -I mol-I K-' Adarnantane 1.7 X lo3 2 90 -4.6 2 0.2 2.2 2 0.1 22.6 2 1.1 Ferrocene 3.3 X 10' 2 170 -5.0 -+ 0.3 2.3 2 0.1 24.4'2 1.2 The most significant observation is that guest encapsulation iricreases with temperature. Now, most processes involving host-guest association are entropically unfavourable but enthalpi- cally favourable, and much has been written about the compensat- ing effects of entropy and enthalpy in complex f0rmati0n.I~ For 7c.7~the inclusion of guests involves an enthalpic cost compen-sated by a larger entropic gairz.Such entropy-driven binding is rem- niscent of the classical hydrophobic effect, wherein release of bound water to the bulk solvent compensates for the association of solutes In organic media such behaviour is not frequently detected,I8 even though liberation of solvent must be a universal feature of molecular recognition phenomena. How can this anom- alous behaviour be interpreted? A single solvent molecule appears too small to fill the cavity of the capsule 7c -7c; more than one CDCl, or p-xylene is required to maximise the intermolecular forces -the van der Waal's interac- tions -between the convex surfaces of the guests and the concave surface of the host's interior.The answer came from a simple set of experiments. In either [2H,]benzene or fl~oro[~H,]benzene the NMR spectra of compound 7b are characteristic of a single dimeric species. When the spectra was recorded in a mixture of the two sol-vents it became clear that three species were present.20 The most economical interpretation is that the third species is the capsular form that contains one of each solvent molecule (Fig. 11). Figure 11 Three distinct softballs are observed in a mixture of two solvents It is reasonable that single molecular guests which fill the cavity and offer chemical and structural complementarity are preferred to multiple solvent molecules; one guest releases the two solvent mole- cules inside the capsule (as well as the retinue of solvents associated with the guest outside).Consequently, the encapsulation of adaman- tanes or ferrocenes by 7c.7~increases entropy, since more than one encapsulated solvent is released to the bulk solution (Fig. 12). What are the limits on the capacity, i.e. what fraction of the space in these capsules can be occupied? What is the volume? Using MACROMODELIS the capsule's volume is calculated to be about 400 A3,or 4 X dm3. Two benzenes or two toluenes can easily be accommodated in this space. This suggested that the cavity was CHEMICAL SOCIETY REVIEWS, 1996 Figure 12 Encapsulation is entropy driven since two solvent molecules are 1iberated roomy enough to accommodate the transition state geometry of a typical Diels-Alder reaction.When both p-quinone and cyclohexa- diene are present in solution with 7a, a single well-defined complex emerges in the NMR spectrum. Integration shows one molecule of quinone is present within this complex. but no signals unique to encapsulated cyclohexadiene can be assigned. Nonetheless it must also be present, since the encapsulated Diels-Alder adduct is observed by NMR within one day (Fig. 13).21 The rate of this reac- tion in bulk solution is so slow at these concentrations that reaction must be taking place within the capsule. The effective molarity of the reactants inside is ca. 1 mol 1-I.This is a promising figure for the future use of capsules as reaction chambers for bimolecular processes. Figure 13 The softball as a reaction vessel for a Diels-Alder cycloaddition 4 Flattened Spheres Other geometric changes that can be made on the original design involve higher symmetries and one such improvement was pro- vided by a triphenylene spacer. This dimer features D,, symmetry, but it resembles -in shape and size comparisons relative to the base- ball and softball -a jam ('jelly) doughnut. This structure 8 was syn- thesized by Robert Grotzfeld (Fig. 14).22 The ceiling and floor of the assembled capsule consist of aromatic r-surfaces; 12 strong hydrogen bonds hold the two halves together. Titration experiments with benzene in CDCI, reveal a direct competition between benzene and CDCl, for the cavity of 8-8.Titration in [2H,,]-p-xylene solu-tion (an uncomfortable guest for 8.8) with cyclohexane revealed a new upfield (6 -0.87) signal in the 1H NMR spectrum for the encapsulated aliphatic guest.ASSEMBLY AND ENCAPSULATION WITH SELF-COMPLEMENTARY MOLECULES -JULIUS REBEK, JR. 26 1 H H 11 x2 Figure 14 The ‘jelly doughnut’ 8.8 has an ideal cavity for cyclohexane encapsulation A depth-shaded view of the complex with cyclohexane reinforces the notional and functional aspects of the jam (jelly) doughnut description. Unlike the other capsules, which take up and release guests at a rate that is fast on the human timescale, 8.8 takes hours to equilibrate with cyclohexane.This guest probably requires a large fraction of the host’s hydrogen bonds to break before passage into and out of the cavity is permitted. 5 Assembly with a Macrocycle Ken Shimizu proposed and synthesized the self-complementary calixarene 9a. It forms a dimeric capsule of yet another shape. The overall architecture of the assembly 9a.9a is that of two hemi- spheres ‘zippered’ together along the equator by hydrogen-bonded ureas (Fig. 15)?3 This hydrogen-bonding pattern of ureas has been well established, particularly in the solid state, where X-ray crys- tallography has shown that the head-to-tail arrangement is a common geometry. Their inviting bowl-like shape and their syn- thetic availability have made calixarenes attractive scaffolds for applications in supramolecular chemistry.24 The calix[4]arene systems synthesized in the research groups of Reinh~udt?~ Ungaro26 and Shinkai27 to assemble by hydrogen bonding are par- ticularly relevant.For the dimerization at hand, the ureas can be hydrogen-bonded in this fashion with the carbonyl oxygens buried into the NH’s of the preceding urea. All eight ureas may be fixed in same direction forming up to 16 hydrogen bonds. The hydrogen-bonding slows rotation about the calixarene-urea bond resulting in an isomer of D s mmetry. 4dEvydence for the existence of calixarene as a hydrogen bonded dimer 9a.9a comes from the encapsulation of solvent molecules inside the assembled cavity. Inclusion was initially apparent in mixed solvent systems where two distinct calixarene assemblies were observed by lH NMR.Direct observation of the encapsulated guests by ‘H NMR is also possible, given some time.Z8 For example, when excess benzene is added to the solution of 9b-9bin [2H,,J-p-xylene, a new signal for encapsulated benzene appears at 6 4.02 and gradually grows in over the course of ca. 40 min. The interleaved geometry of the assembly prevents visiting guests from leaving or entering quickly. 9aR=H 9bR=F 9a-9a Ar = C6H5 9b-9b Ar = CeH4p-F Figure 15 The head-to-tail ureas drive the formation of new capsules 9.9 Other guests have also been directly observed, such as fluoro- benzene, p-difluorobenzene and pyrazine. Blake Hamann observed remarkably different chemical shifts for encapsulated fluorobenzene; the para and ortho protons are separated by > 2 ppm, suggesting that a particular orientation is favoured by this guest within the cavity.The orientation shown in 10 (Fig. 16), where the ortho and metu hydrogens are directed at the 7~ faces of the calixarene while the para hydrogen and the fluorine are directed at the belt of ureas, is consistent with the chemical shifts. Competition studies of guests with benzene were undertaken and the affinity of the calixarene dimer for these, relative to benzene, is given in Table 3. Addition of other ureas such as the phenyl-a-phenylethyl deriv- CHEMICAL SOCIETY REVIEWS, 1996 R R R RI I I /Figure 16 'Denaturation' of the capsule and liberation of p-fluorobenzene by competing urea functions ative to the solution results in the liberation of guests and a new calixarene species 11 appears.This is most likely a result of the 'denaturation' of the calixarene dimer by the urea through compet- itive hydrogen bonding (Fig. 16), a process analogous to the denat- uration of proteins with urea. Table 3 Relative affinities of guests in competition experiments with benzene Guest Affinity (25 "C) Guest Affinity (25 "C) C6H6 1.o C,H,OH 0.83 C,H,F 2.6 C,H5NH, 0.32 p-C,H,F, 5.8 Pyrazine 3.2 C,H,CI 0.30 Pyridine 1.2 C,H,CH3 <0.1 The high affinity observed for pyrazine is also suggestive of a conformation that directs the nitrogen lone pairs at the urea hydro- gens within the complex.A recent study by Sherman29 also finds pyrazine a favoured guest. The system involves a belt of strong hydrogen bonds (phenol with phenolate) and structures closely related to calixarenes (Fig. 17). k k k Figure 17 Hydrogen bonding holds hemispheres together for guest encapsulation The volume of the cavity was estimated to be ca. 210 A3, and most guests appear to fill ca. 50%or less of the available space in the cavity. By comparison, in a typical liquid 70% of space is filled. The shape of the cavity within 9.9 is difficult to visualize, but a geometric simplification is shown in Fig. 18. Each calix[4]arene is represented by a square pyramid with its tri- angular facets corresponding to the outside of the aromatic sur- faces.In the dimer, the pyramids are offset by 45", and the corners are cut off to represent the eight interleaved ureas which encircle the equator. 2 Square Pyramids ( 45"offset ) 7-8A 7-8& Figure 18 Schematic representations of the cavity formed by the dimeriza- tion of tetraureas 9,and vertical and horizontal cross sections Cross sections of the molecular model are in agreement with the gem-like geometric representation. A horizontal slice through the plane of the ureas yields an octagonal cavity, while a vertical slice gives a diamond-shaped cavity, having different angles and lengths for the top and bottom sections. The representations shown in Fig. 18 suggest that cubane is a ASSEMBLY AND ENCAPSULATION WITH SELF-COMPLEMENTARY MOLECULES- JULIUS REBEK, JR.reasonable three-dimensional complement for the cavity shape. The energy-minimized structures for the complexes with cubane are shown in Fig. 18, again with the cavity walls cut away for ease of visualization. In the experiment, cubane was added to a solution of 9b-9bin [*H,,]-p-xylene and the formation of a I: 1 complex was again observed by NMR. Figure 19 Cut-away views of CPK representations of the calixarene dimer with encapsulated cubane In conclusion, the behaviour and functions of molecular assem- blies can, to some extent, be controlled -either by solvation effects or by nucleation by guests. The energetics for such complexes invariably pit intermolecular forces against the decreased freedom of the included guest.These forces are the van der Waals' interac- tions and hydrogen bonds between the exterior surface of the guest and the interior surface of the host. In addition, special entropic effects can emerge when more than one solvent molecule is present in the capsules. Guests, or better, hostages, which more closely fit the host's cavity in size and shape and leave no empty space are favoured.30 The acceleration of a Diels-Alder reaction augurs well for the long-term goal of using these capsules as reaction chambers. In the meantime, we continue to explore the behaviour of 'mole- cules within molecules .'3I Acknowledgements I am indebted to the members of my research group for their inspired designs and experiments.Their names appear on the original publications. I am grateful to the National Institutes of Health for financial support. 6 References 1 S. C. Zimmerman and B. F. Duerr, J. Org. Chem., 1992,57,2215. 2 J.-M. Lehn,Angew. Chem., Int. Ed. Engl., 1990,29,1304;J. S. Lindsey, New J. Chem.,1991,15, 153; G. M. Whitesides, J. P. Mathias and C. T. Seto, Science, 1991,254, 13 12. 3 C. T. Set0 and G. M. Whitesides,J. Am. Chem. Soc.. 1990,112,6409:F. Garcia-Tollado, S. J. Geib, S. Goswami and A. D. Hamilton. J. Am. Chem. SOC., 1991,113,9265. 4 J. A. Zerkowski, C. T. Seto, D. A. Wierda and G. M. Whitesides, J. Am. Chem. SOC., 1990, 112, 9025; J. Zerkowski, J. P. Mathias and G. M. Whitesides, J. Am. Chem. SOC.,1994,116,4305. 5 M.Simard, D. Su and J. D. Wuest,J. Am. Chem. Soc., 1991,113,4696; J. Am. Chem. SOC., 1994, 116 1725; Y. Ducharme and J. D. Wuest, J. Org. Chem., 1988,53,5787;C. Fouque, J.-M. Lehn and A. M. Levelut, Adv. Mater., 1990,2,254; J. Yang, J.-L. Marendaz, S. J. Geib and A. D. Hamilton, Tetrahedron Lett., 1994,35,3665. 6 J. P. Mathias, C. T. Seto, E. E. Simanek and G. M. Whitesides, J. Am. Chem. Soc.. 1994, 116, 1725; J. P. Mathias, E. E. Simanek, J. A. Zerowski, C. T. Seto and G. M. Whitesides, J. Am. Chem. Soc., 1994, 116,4316. 7 M. R. Ghadiri, J. R. Granja, R. A. Milligan, D. E. McRee and N. Khazanovich, Nature, 1993,366,324; J. R. Granja and M. R. Ghadiri, J. Am. Chem. SOC., 1994,116, 10785; K. Kobayshi, J. R. Granja, R. K. Chadha and D. E. McRhee, Angew.Chem., Int. Ed. Engl., 1995.34, 93. 8 R. Wyler, J. de Mendoza and J. Rebek, Jr. Angew. Chem., Int. Ed. Engl., 1993,32,1699. 9 K. T. Chapman and W. C. Still, J. Am. Chem. Soc., 1989, 111,3075; M. L. C. Quan and D. J. Cram,./. Am. Chem. SOC.,113,2754. 10 N. Branda. R. Wyler and J. Rebek, Jr. Science, 1994, 263, 1267; for methane complexation inside a cryptophane: L. Garel, J.-P. Dutasta and A. Collet, Angew. Chem., Int. Ed. Engl., 1993,32, 1169. 11 C. ValdCs, U. P. Spitz, L. Toledo, S. Kubik and J. Rebek, Jr., J. Am. Chem. SOC., 1995,117,12733. 12 N. Branda, R. M. Grotzfeld, C. ValdCs and J. Rebek, Jr., J. Am. Chem. SOC.,1995,117,85. 13 X. Garcias and J. Rebek, Jr., Angew. Chem., Int. Ed. Engl., 1996, in press. 14 T. Fox, B.E.Thomas, M. McCarrick and P. A. Kollman,J. Phys. Chem., submitted for publication. We thank these authors for helpful corre-spondence and a preprint of their manuscript. 15 Molecular modelling was performed using MACROMODEL 3.5X (MM2" force field); F. Mohamadi, N. G. Richards, W. C. Guida, R. Liskamp, M. Lipton, C. Caufield, G. Chang, T. Hendrickson and W. C. Stil1,J. Comput. Chem., 1990,11,440. 16 C. Valdes, U. P. Spitz, S. Kubik and J. Rebek, Jr., Angew. Chem., Int. Ed. Engl., 1995.35.1885. 17 R. Meissner, J. de Mendoza and J. Rebek, Jr., Science, 1995,270.1485. 18 J. Kang and J. Rebek, Jr., Nature, in press. For other cases involving covalent J. Canceill, M. Cesario, A. Collet, J. Guilhem, L. Lacombe, B. Lozach and C. Pascard, Angew.Chem., Int. Ed. Engl.. 1989,28. 1246; D. J. Cram, H.-J. Choi. J. A. Bryant and C. B. Knobler, J. Am. Chem. SOC., 1992, 114. 7748; D. J. Cram, M. T. Blanda, K. Paek and C. B. Knobler, J. Am. Chem. Soc., 1992,114,7765. 19 J. D. Dunitz, Chem. Bio., 1995, 2,709; B. R. Peterson, P. Wallimann, D. R. Carcangue and F. Diederich, Tetrahedron, 1995,51,401; M. S. Searle, M. S. Westwell and D. H. Williams,J. Chem. Soc., Perkin Trans 2,1995,141. 20 R. Meissner, X. Garcias and J. Rebek, Jr., Angew. Chem.. submitted for publication. For other unique encapsulation phenomena, see: D. J. Cram and M. E. Tanner, Angew. Chem., Int. Ed. Engl., 1991, 30, 1024; P. Timmerman, W. Verboom, F. C. J. M. van Veggel, J. P. M. Duynhoven and D. N. Reinhoudt, Angew. Chem., lnt.Ed. Engl., 1994,34,2345; R. G. Chapman, N. Chopra,E. D. Cochien and J. C. Sherman,./. Am. Chem. Soc., 1994,116,369. 21 J. Kang and J. Rebek, Jr., unpublished. 22 R. Grotzfeld and N. Branda and J. Rebek, Jr., Science, 1996,271,487. 23 K. D. Shimizu and J. Rebek, Jr. Proc. Natl. Acad. Sci. USA, 1995,92, 12403. 24 For recent reviews see: V. Bohmer,Angew. Chem., Int. Ed. Engl.. 1995, 34, 713; P. Linnane and S. Shinkai. Chem. Ind., 1994, 811; C. D. Gutsche, Aldrichim. Acta, 1995, 28, 3; S. Shinkai, Tetrahedron. 1993, 49, 8933; P. Timmerman, W. Verboom and D. N. Reinhoudt. Tetrahedron, 1996,52,2663. CHEMICAL SOCIETY REVIEWS, 1996 25 J. D.van Loon, R. G. Janssen, W. Verboom and D. N. Reinhoudt, 30 For examples of covalently linked hemispheres see: T.A. Robbins, Tetrahedron Lett. 1992,33,5125;P.Timmerman,K.G.A.Nierop,E.A. C. B. Knobler, D. R. Bellow and D. J. Cram, J. Am. Chem. Soc., 1994, Brinks, W. Verboom, F. C. J. M. van Veggel, W. P. van Hoorn and D. N. 116,lll;A.IkedaandS.Shinkai,J.Chem.Soc.,PerkinTrans.I,1993, Reinhoudt, Chem. Eur. J., 1995,1, 132. 2671 ;L. Garel, J. P. Dutasta and A. Collet, Angew. Chem., fnt. Ed. Engl., 26 A. Arduini, F.Fabbi, M. Mantovani, L. Mirone, A. Pochini, A. Secchi 1993,32, 1169; J. Canceill, L.Lacombe and A. Collet, J.Am. Chem. and references therein. For a review of molecules and R. Ungaro, J.Org. Chem., 1995,60,1454. SOC.,1986,108,4230, 27 For two-component assemblies based on calixarenes see: K. Koh, K. with large interior cavities see: C. See1 and F. Vogtle,Angew. Chem., Int. Araki and S. Shinkai, Tetrahedron Lett., 1994,44,8255. Ed. Engl., 1992,31,528. 28 B. C. Hamman, K. D. Shimizu and J. Rebek, Jr., Angew. Chem. tnt. Ed. 31 D. J. Cram, Lecture at the C. David Gutsche Symposium, Washington Engl., 1996,35,1326. University, St. Louis, Missouri, May 5, 1990. 29 R. G.Chapman and J. C. Sherman, J.Am. Chem. Soc., 1995,117,9081.