Chapter 1: Applications of Synthetic Receptors for Biomolecules

This chapter describes the major classes of synthetic receptors for biomolecules and how they are employed for four different types of practically useful applications: separations, imaging and sensing, catalysis, and pharmaceutical activity. These applications play crucial roles within the frontier technology areas of health care, environmental remediation, nanotechnology, and advanced materials.¹  Each application requires a synthetic receptor that is endowed with a specific set of functional properties. In all cases, the functions are triggered by a molecular recognition event that involves association of the receptor (sometimes called a host) with the target biomolecule (sometimes called a guest, substrate, ligand, or analyte) (Scheme 1.1).²  The associated complex is held together by a collection of weak and reversible bonds, and typically the association alters the chemical and physical properties of the binding partners. It is this change in molecular properties that enables the subsequent functions to occur.

The association event can be characterized in terms of thermodynamics (the extent to which association occurs) and kinetics (the rate at which it occurs). The energy diagram in Scheme 1.1 defines the free energy of association (ΔG_a), which determines the association constant (K_a). Another important receptor binding property is thermodynamic selectivity, which is equal to the ratio of association constants for two separate guests. The energy diagram also defines the free energies of activation for complex association (ΔG_on^‡) and dissociation (ΔG_off^‡), which dictate the corresponding rate constants for these steps (k_on and k_off, respectively). Different applications have distinctive thermodynamic and kinetic requirements for optimal performance and the kinetic and thermodynamic properties of biological receptors are often informative benchmark values.

Scheme 1.2 is a simple workflow diagram that connects three important stages in supramolecular chemistry research. In the case of biomolecule recognition, the vast majority of research projects move through these stages from left to right. Researchers typically start out by conducting fundamental studies of model receptor/biomolecule association systems and work to develop systematic rules and generalizable concepts that rationalize experimental observations. Based on the new insight, a typical project progresses to the second stage by developing prototype examples of novel functional molecules. Only occasionally does a biomolecule recognition project progress all the way and generate robust new technology with useful applications.

Putting aside discussion concerning the mission of academic research, there are several scientific reasons for the modest number of current industrial and commercial processes that use synthetic receptors for biomolecules. The most obvious point is the relative newness of supramolecular chemistry as a sub-discipline within molecular science. There is strong consensus that the discovery of crown ethers in the late 1960s was the seminal event that coalesced supramolecular chemistry as a field of systematic study.³  Within few a years supramolecular chemists were pursuing synthetic receptors for small biomolecules. But molecular recognition in water is a complicated physical process, and it has taken some time for the field to elucidate the fundamental thermodynamic factors that produce selective non-covalent association. As discussed further in Chapter 2, relatively intuitive supramolecular concepts such as the hydrophobic effect, binding cooperativity, and enthalpy–entropy compensation are still under intellectual refinement.^4–6  Computational algorithms for rational design of effective receptors are in early-stage development, and so are alternative statistical discovery approaches that aim to uncover effective receptors by combinatorial synthesis and screening methods.

Nonetheless, examples of practically useful applications based on synthetic receptors are emerging at an accelerated pace. It is worth noting that many applications mimic biological processes that evolved by natural selection over billions of years. In comparison, the remarkable progress of supramolecular chemistry over the last 50 years is quite impressive. Continued expansion is expected for several reasons. One is the growing synergistic overlap of supramolecular chemistry with advances in nanoscience. The development of versatile nanoscale fabrication methods has led to new synthetic nanoparticle platforms for constructing multivalent biomolecule recognition systems with relatively large binding surface areas. In addition, major advances in biomolecular engineering now allow rapid construction of biopolymers with unnatural sequences, and new bioconjugation techniques enable ready modification of these biopolymers with synthetic components such as indicator groups, catalytic residues, and secondary binding sites, to produce natural/synthetic hybrid receptor molecules.

Indeed, the field of synthetic receptors for biomolecules is sufficiently advanced that researchers can realistically have the ambition to progress all the way through the workflow diagram in Scheme 1.2. It is quite feasible for a scientist or engineer to use the modern principles of receptor design and discovery to produce supramolecular systems that have practical applications. One of the first major decisions in this workflow process is whether the receptor system should be based on a small molecule, polymer, or nanoparticle platform and the next section in this chapter provides an overview of the attributes of each major choice.

The pool of protein-based receptors is a logical resource for researchers who wish to develop a biomolecule-binding receptor that will perform a specific function. Protein-based molecular recognition systems include antibodies, enzymes, lectins, signaling partners, transcription factors, and membrane transport proteins. Antibodies are the most widely utilized protein systems, with a wide range of binding affinities (K_a typically 10⁵–10¹² M⁻¹) and target selectivities. But for certain applications, antibodies and related binding proteins are not feasible. They may be too expensive to produce, or they cannot be stored for extended periods. It may be difficult or impossible to chemically convert them into reporting agents for certain types of imaging or sensing applications. A concern with pharmaceutical applications is the possibility of an undesirable immune response, or the challenge of optimizing the pharmacokinetic properties of a high molecular weight structure. In principle, all of these drawbacks can be circumvented by developing synthetic receptors.

This chapter classifies synthetic receptors into six categories. The majority of synthetic receptors have a binding pocket that is formed by a convergent arrangement of functional groups appended to either a macrocyclic or cleft-shaped scaffold (Scheme 1.3). The receptor structural factors that control guest affinity, binding selectivity, and binding kinetics are described in Chapter 2. As a rule, preorganized macrocyclic structures tend to bind their complementary guests with higher affinity and selectivity than flexible acyclic receptors. However, the higher affinity usually means that the rates of association and dissociation are slower, which can be an undesirable property for certain applications.⁴ 

Receptors can also be constructed by attaching multiple functional groups to the exterior of nanoscale scaffolds such as dendrimers or nanoparticles. This divergent arrangement of functional groups enables the receptors to form multiple interactions with the exposed surfaces of relatively large binding partners such as proteins or nucleic acids. But selective binding of biopolymer surfaces in aqueous solution is quite challenging. Exposed polar functional groups are heavily solvated, and thus association processes must overcome a high desolvation penalty. The presence of some nonpolar groups on the receptor surface is often needed to drive association, but too many hydrophobic groups can lead to undesired receptor self-aggregation.

The largest numerical group of synthetic receptors is comprised of organic molecules. There is a tendency to work with privileged organic molecular scaffolds that already have an inherent molecular recognition capability. Many of these scaffolds are macrocycles, such as crown ethers, calixarenes, cucurbiturils, cyclodextrins, or alternatively acyclic structures with cleft or tweezer shapes (Scheme 1.4, and a more detailed description is provided in Chapter 2). Some organic structures are relatively easy to covalently modify, such as crown ethers, while others are technically harder to transform and purify, such as cyclodextrins and cucurbiturils. The structural modifications include linkers to solubilize or immobilize the receptor, chemical functionality for enhanced binding affinity or recognition selectivity, catalytic residues, or reporter groups for imaging or sensing applications.

Pharmaceutical applications typically require a receptor structure that is biocompatible, and it is relatively easy to modify the structure of a small organic molecule and improve its pharmacokinetic or ADMET (absorption, distribution, metabolism, excretion, and toxicity) properties. Organic receptors are excellent scaffolds for binding ions or small biomolecule guests. But organic receptors have a small number of non-covalent interaction sites and low surface area for hydrophobic contact. Thus, it is difficult to produce organic receptors with high affinity for larger biomolecules such as proteins or cell membranes. One potential solution is to devise receptor self-aggregation methods that assemble a large number of small molecule receptors to produce a larger multivalent conglomerate with increased biomolecule affinity.⁷ 

While water is a strong hydrogen bonding molecule, it is a relatively weak Lewis base. Therefore, molecular recognition systems that utilize Lewis acid coordination in aqueous solvent are usually more effective than purely hydrogen bonding systems. Metal coordination complexes are often employed as synthetic receptors for electron-rich species such as inorganic anions or anionic biomolecules. The receptor typically comprises one or more transition metal cations that are chelated by an organic scaffold containing nitrogen donor atoms. At least one of the metal cation coordination sites is open and able to accept the electron-rich guest (Scheme 1.5a). Binding affinity can be enhanced by using multiple Lewis acidic centers to form a chelated binding site for the target biomolecule (Scheme 1.5b). The receptor scaffold can also be decorated with hydrogen bonding units that provide favorable secondary interactions (Scheme 1.5c). Metal coordination complexes often have prominent optical or redox properties, thus they are excellent fabrication platforms for indicators and reporters. As strong Lewis acids, they also can catalyze chemical reactions by polarizing ground states and stabilizing reactive intermediates. While some transition metal cations are potentially toxic, it is usually straightforward to alter the molecular structure of a coordination complex in order to improve solubility or ADMET properties.

In addition to molecular recognition strategies that employ metal centers with open coordination sites for direct contact with a Lewis basic guest, new tactics are emerging that use metal coordination as a structure organizing element. One approach employs metal cation coordination as a construction method to assemble multiple hydrogen bonding units (Scheme 1.5d) and create a convergent binding pocket. An alternative strategy is to combine metal centers with appropriately designed rigid organic building blocks and generate self-assembled soluble cages with large enough cavities to encapsulate molecular guests. This metal organic self-assembly approach has many attractive features, including rapid, often single-step, production of cage structures, which can be highly charged and water soluble. In many cases the cage interior is lined with aromatic surfaces and the hydrophobic effect can be used to drive guest association. The solid-state versions of these self-assembled cages are metal organic frameworks (MOFs). In many cases, MOFs are porous networks and they have typically been used for small molecule capture, as their large surface areas allow for an excellent storage capacity of gases and volatile solvents.⁸  Researchers are beginning to pursue additional applications by incorporating fluorescent reporters or catalytic sites within a MOF.⁹  It seems likely that porous MOFs will soon be used to selectively trap small biomolecules.

Proteins and nucleic acids have excellent molecular recognition abilities, and they can be synthesized using reliable and efficient chemical or biosynthetic procedures. Thus, it is logical to design or discover synthetic receptors that are based on biopolymers with unnatural sequences. The biopolymers can fold to produce structural features such as hairpins, bulges, loops, and binding pockets. Often there is concomitant coordination of a proximal metal cation that can enhance guest binding or reaction catalysis. Polypeptides have the benefit of a wider choice of naturally occurring building blocks (22 naturally occurring amino acids versus four naturally occurring ribonucleotides) to construct a wider and more diverse set of receptor candidates. But oligonucleotide systems have a significant advantage in that the binding properties of the receptor structures can be refined by in vitro selection methods using biotechnology.

Aptamers are single strands of oligonucleotides with 15–50 building blocks that can fold to form a tertiary structure with the capability to bind a guest molecule (Scheme 1.6).¹⁰  The sequences are usually ribonucleotides rather than deoxyribonucleotides, since the extra hydroxyl provides additional interactions. RNA aptamers are synthetic versions of riboswitches, which are segments within naturally occurring mRNA that can bind a small molecule (usually a metabolite) and change the production of proteins encoded by the mRNA. Aptamers can exhibit extremely high binding constants and high guest selectivities.¹¹  For example, an aptamer has been reported with 12 000 times higher affinity for l-arginine over its enantiomer, d-arginine.¹²  While aptamers can be generated for virtually any molecular target, there are some biomolecules with few functional groups, such as sugars, that cannot be recognized with high affinity. One way to circumvent this problem is to develop aptamers that form a high-affinity ternary complex with the biomolecule bound to a small organic receptor. An example of this strategy is an aptamer that can recognize glucose bound to a small boronic acid receptor.¹³ 

Most aptamers are produced using the SELEX (systematic evolution of ligands by exponential enrichment) method or a variant, which involves iterative cycles of oligonucleotide selection, amplification and mutation (see Chapter 2). The technology can be automated to produce an effective binding sequence within a few days. The fact that aptamers are well defined chemical entities with little or no immunogenicity makes them attractive for pharmaceutical development. But aptamers have limited stability in physiological solution due to nuclease degradation. In some cases the rapid degradation is an advantage, but often aptamer stability needs to be improved. Avoidance of the degradation enzymes can be achieved by either: (a) attaching long polyethylene glycol (PEG) chains to sterically protect the aptamer; (b) constructing a Spiegelmer, a sequence composed of non-natural l-ribonucleotides that are not recognized by enzymes; or (c) selective incorporation of a non-cleavable nucleoside, such as a 2′-fluoro-2′-deoxyribonucleoside, at specific sequence locations that are susceptible to enzyme attack.

Aptamers have been investigated extensively as pharmaceutical agents.^14,15  In many cases, the pharmaceutical action is due to selective recognition of a biological target, usually a small section of protein surface (epitope). At present, the only clinically approved aptamer-based therapeutic agent is pegaptanib (brand name MACUGEN^®), which is used for treating age-related macular degeneration. Clinical trials have investigated other aptamers for various molecular targets and they also show promise as delivery vehicles for therapeutic small interfering (si)RNA. A related pharmaceutical application is prodrug delivery. A recent study incorporated 30 copies of the cytotoxic nucleoside analogue 5-fluoro-2′-deoxyuridine (5-FUdr) into the sequence of an aptamer (comprised of 106 nucleotides) that binds strongly and specifically to a cell surface receptor.¹⁶  Receptor binding led to endocytosis and aptamer entry into the cell, where it was degraded by intracellular nucleases to release the 5-FUdr drug.

Aptamers are broadly useful in a range of other recognition-based technologies, such as separations, environmental purification, or diagnostics.^11,17,18  In most of these cases, the aptamers are immobilized by covalent attachment to a surface. Considerable effort has focused on developing aptamers as the recognition unit within chemical detection systems. Optical sensing with aptamers is used often for quantifying specific biomolecules in environmental samples, biomedical tissue, or cell culture. Molecular beacons are a broad family of fluorescent oligonucleotide sensors that exploit the conformational switching caused by guest binding to modulate the distance between a fluorophore and an energy accepting partner. Other aptamer design configurations use biomolecule binding to trigger association and enhanced emission of a reporter dye. Although highly effective and very popular, the sensitivity of molecular beacons is limited because of the 1 : 1 signaling stoichiometry. Approaches to produce amplified signaling have investigated ribozymes (or DNAzymes), which are aptamers having catalytic activity.¹⁹  Suitable combination of an aptamer with a ribozyme creates a sensing ensemble that can bind a guest molecule and turn on a catalytic process producing an amplified sensing signal.

Polypeptide versions of aptamers are known, but the combined process of synthesis and screening is much slower and there is no peptide equivalent of SELEX for in vitro selection and refinement of binding ability.²⁰  Peptide bonds are kinetically stable at physiological pH but they are susceptible to catalytic cleavage by protease enzymes. This has led researchers to develop alternative polyamide systems that do not have side chains connected to the α-carbon and thus resist protease action. These structures include peptoids with the side chain connected to the peptide backbone nitrogen,²¹  and β-peptides with the amino group bonded to the β-carbon.²² 

The general idea of molecular imprinted polymers (MIPs) is to generate cross-linked polymeric materials with template imprinted cavities that act as selective molecular recognition sites. The picture in Scheme 1.7 shows the method of non-covalent imprinting, where a polymerization reaction is conducted in the presence of a molecule template. The template is incorporated into the cross-linked polymer matrix by means of non-covalent interactions with the monomers and the growing polymer chains. An alternative approach is covalent imprinting, which uses a functional monomer that is comprised of the molecular template with polymerizable groups attached by cleavable covalent bonds. In both cases, the template is removed after the polymerization reaction is complete to reveal the MIP. Optimization of the template structure and the polymer cross-linking conditions produces MIPs with strong and selective affinity for many types of analytes, including various biomolecules. There are reports of MIPs for proteins and even complete micro-organisms. The most reproducible results are obtained when the biological template is a rigid structure like a virus. MIPs that use templates based on flexible protein chains or multiple membrane diffusible epitopes are often more problematic.

MIPs are attractive materials for a broad range of applications in separations, molecular sensors, and catalysis.²³  Although the basic concept of an MIP is easy to visualize, there are several technical challenges that have to be overcome for successful implementation. An obvious requirement is efficient removal of the template from the cross-linked polymer matrix. Sometimes complete removal of the template is not possible without damaging the imprinted cavity, leading to a lower binding activity. One way to overcome this limitation is to change from a bulk imprinting method to a surface imprinting approach.²⁴  A related problem is the inherent heterogeneity of the imprinting process, which produces a range of binding sites with different degrees of guest affinity and selectivity. This problem can be circumvented by using blocking methods that selectively eliminate the lower binding sites. Regardless of the fabrication method, the performances of MIPs are usually assessed by comparisons to control, non-imprinted polymers according to three quantitative binding factors: capacity, affinity, and selectivity.

Most early-generation MIPs were prepared in organic solvents, which were not compatible with most biological templates. But newer polymerization methods have been developed to make water-based structures such as sol–gel or hydrogel MIPs.²⁵  The latter are made of water compatible polymers like polyacrylate or polyacrylamide, and they can be prepared by copolymerizing mixtures of monomers containing different polar functional groups. In addition, methods to make imprinted inorganic surfaces, imprinted membranes, and imprinted nanoparticles are emerging.²⁶  The recent progress with imprinted polymer nanoparticles as “plastic antibodies” is quite exciting. For example, peptide imprinted nanoparticles have been prepared and shown to extract toxic peptides from the bloodstream of living animal models.²⁷ 

Dendrimers are branched polymers with symmetrical structures and monodisperse molecular weights. The tree-like architecture is usually visualized as a focal core with surrounding layers, or generations, of attached monomers. As the structures get larger (later generations) they begin to adopt a spherical shape with nanoscale dimensions, a sterically crowded surface, and internal pores. While dendrimers are aesthetically pleasing structures, the synthetic effort to make customized systems can be time-consuming and expensive. Therefore many researchers focus on commercially available dendrimers and dendron building blocks. Alternatively, there is increasing interest in hyperbranched polymers, which can be prepared on a large scale using one-pot synthetic methods. Although hyperbranched polymers are not monodisperse structures, they often exhibit many of the same supramolecular properties as dendrimers.²⁸ 

Dendrimers and hyperbranched polymers are investigated for multifarious purposes such as advanced materials, supports for inorganic catalysts, and drug delivery agents.^29,30  In terms of biomolecule recognition, dendrimers and hyperbranched polymers are often converted into multivalent targeting systems by covalent attachment of multiple affinity ligands to the exterior surface of the polymer. This type of coated nanoparticle will be discussed further in the following section. Here, the focus is on the ability of high-generation dendrimers and hyperbranched polymers to encapsulate guest biomolecules within their internal cavities (Scheme 1.8). They act like “unimolecular micelles” with potential applications in drug delivery and supramolecular catalysis.

Most designs of unimolecular micelles have an amphiphilic core–shell architecture, with either a hydrophobic core and hydrophilic corona for extraction of nonpolar molecules into water or vice versa.³¹  Early work with dendrimer systems demonstrated encapsulation of hydrophobic dyes or drugs, and in most cases the basis for the molecular recognition was a complementary match of guest size with the hydrophobic cavities inside the dendritic structure. Attempts to increase guest selectivity have investigated two related synthetic strategies to create a well-defined binding cavity. One approach is to append multiple dendritic branches to the periphery of a container molecule like a cyclophane or cyclodextrin.³²  The other approach is to use a templated imprinting process to produce a cross-linked dendrimer with a single binding site at the core.³³ 

More recent work has investigated new methods of triggering guest release from the dendrimer core. One approach uses “facially amphiphilic dendrimers”.³⁴  The dendritic structures incorporate biphenyl building blocks, each equipped with a polar arm and a nonpolar arm. The aryl groups respond to solvophobic forces by folding and minimizing interfaces with mismatched polarities. In water, the pores in the dendrimer core are lipophilic and can encapsulate hydrophobic guests, which can be subsequently released by different types of stimuli, including protein binding to the dendrimer surface. This causes a change in the dendrimer folding and concomitant polarity inversion within the pores.

The binding pockets within high generation dendrimers and hyperbranched polymers are reminiscent of enzyme active sites and researchers have worked to convert these synthetic receptors into enzyme mimetics (see below). The recent developments in organocatalysis have broadened the number of reactions that can be promoted by dendrimer catalysts.³⁵  In addition to bond forming and cleavage reactions, oxidations and reductions have been studied. The large size of high-generation dendrimer catalysts facilitates the process of catalyst separation and recycling. The expected continued growth in organocatalysis should lead to further opportunities to develop novel dendrimer catalysts with enzyme-like recognition properties.

Nanoparticles are increasingly valuable as spherical platforms that can be coated with synthetic receptors. Scheme 1.9 shows a nanoparticle with core–shell architecture.³⁶  The core of the nanoparticle can be filled with signaling materials for imaging (e.g. inorganic lattices such as gold atoms or transition metal salts, or organic polymers doped with dyes), magnetic materials for separations, or pharmaceutical agents for controlled release. The size and shape of the nanoparticles and the thickness of the surrounding shell are parameters that affect their physical, chemical, and biological performance.^37,38  An important point for molecular recognition is the loading density of affinity ligands on the particle surface and their surface mobilities.

Inorganic core–shell nanoparticles are usually prepared by stepwise processes that coat a preformed inorganic core with an organic shell. Organic particles can be prepared by microemulsion polymerization methods and surface functional groups are added either by copolymerization of monomer mixtures or covalent modification of the particle exterior. A wide variety of organic particles are commercially available with diameters ranging from microns to hundreds of nanometers. An alternative way to fabricate organic nanoparticles is through self-assembly of appropriately designed amphiphilic building blocks. Vesicles and micelles are the best known colloidal structures that are formed by amphiphile self-aggregation. Hollow vesicles composed of polar lipids are called liposomes and they are used widely as biocompatible delivery vehicles for drugs and imaging contrast agents. Liposome assembly methods are straightforward and it is easy to control particle size and loading capacity. In addition, modern conjugation methods allow the liposome surface to be functionalized with targeting ligands or reporter groups. The past decade has seen great expansion in the types of amphiphilic building blocks that can be assembled into well-defined nanoparticles. A wide range of amphiphilic block copolymers have been created with various chain architectures, including linear, graft, cyclic, and star polymers. This work has been greatly facilitated by the development of controlled polymerization techniques such as reversible addition–fragmentation chain transfer polymerization, atom transfer radical polymerization, and nitroxide-mediated polymerization.

While the technical simplicity of amphiphile self-assembly is very attractive, there are inherent quality control concerns that may limit translation into certain types of practical applications, especially in pharmaceutical settings. Most notably, batch-to-batch production and particle size control is often hard to reproduce. Multicomponent nanocomposites can be especially difficult to characterize at the atomic level. The dynamic nature of self-assembled systems makes it challenging to confirm size and structure throughout the entire time period of an application. If a self-assembled nanoparticle is not thermodynamically stable, it is important to evaluate its kinetic stability. For example, micelles typically dissociate into monomer components upon dilution below the critical micelle concentration, whereas vesicles are kinetically more stable and will remain intact for extended periods.

The following section describes in general terms the four major classes of applications (Scheme 1.10). Each application is illustrated with example receptor systems. The thermodynamic and kinetic properties that control receptor performance are also discussed.

Various biomolecule analysis and purification methods use a molecular recognition process to separate the target analyte from a mixture, often by translocating it into a separate phase of matter. Performance parameters include receptor/biomolecule affinity, biomolecule selectivity, rates of complex association/dissociation, separation efficiency, batch size, feasibility of continuous flow processes, and capacity for regeneration.

Affinity chromatography is a liquid chromatography process that employs an affinity agent as the stationary phase to selectively retain a target biomolecule.³⁹  In chromatography, a selectivity factor, α, is defined by eqn (1.1), where k′ represents the retention factors for two analytes X and Y.

A fundamental goal in affinity chromatography is to maximize the difference in binding affinity for X over Y, effectively increasing the value of α. Affinity columns are often used in combination with sensitive detection methods like mass spectrometry to create analytical detection assays. Shown in Scheme 1.11 is the common “step-elution” mode of operation which requires a selective recognition system with large α value. The separation mixture (usually a complex biological mixture) is passed through the column and the targeted biomolecule is selectively retained for subsequent elution under weakened binding conditions. A typical configuration employs a column filled with packing material that exhibits low non-specific binding such as agarose or cellulose. Newer methods incorporate silica particles or monolithic materials that are capable of withstanding high flow rates and pressures. The affinity agents range from biopolymers, such as proteins and nucleic acids, to small organic dyes and Lewis acids. Antibodies are often used to isolate and identify drug molecules or metabolites, and sugar binding lectins are effective for the separation of glycoconjugates.

The robustness of synthetic receptors makes them attractive candidates as stationary phases for chromatography, but only a few systems are presently utilized for biomolecule separations.³⁹  Boronate affinity chromatography takes advantage of the reversible covalent association of diol-containing carbohydrates with boronic acids (Scheme 1.12a). Boronate columns are used extensively for the clinical quantification of glycated hemoglobin as an assessment of long-term diabetes management. They also have been used to separate glycoproteins from complex mixtures. Immobilized metal ion affinity chromatography exploits the reversible coordination of metal cations like Ni²⁺, Zn²⁺, Cu²⁺, or Fe³⁺ by biomolecule targets such as amino acids, peptides, proteins, or nucleic acids.⁴⁰  The metal cations are immobilized on the stationary phase of the column by a covalently bonded chelating group (Scheme 1.12b), and the method is used extensively to purify proteins that contain multiple histidine residues (His-tagged proteins). The configuration of the stationary phase can be modified to produce new platforms for separations. A recent example employed magnetic beads coated with immobilized Zn²⁺ coordination complexes to remove bacterial cells from blood samples.⁴¹ 

The composition of the Lewis acid stationary phase can be varied. Metal-oxide affinity chromatography typically employs beads of titanium dioxide as the stationary phase to retain acidic biomolecules, especially phosphorylated peptides. Dye affinity chromatography was developed in the late 1960s after the serendipitous discovery that stationary phases coated with immobilized triazine dyes selectively retain kinase enzymes (Scheme 1.12c). Follow-up work screened a large number of organic dyes from the textile industry and found examples that retained blood proteins and various enzymes. In recent years molecularly imprinted polymers⁴²  and aptamers have been examined as stationary phases for affinity chromatography of biomolecules and it is likely that future studies will show other types of synthetic receptors to be effective.⁴³ 

There is a need in phosphoproteomics research for high-throughput methods that enrich digested samples containing phosphorylated proteins and peptides for subsequent analysis by mass spectrometry. In addition to the affinity chromatography methods described above, there are newer enrichment strategies using phosphoprotein-binding domains (PBDs). Commercial kits are available that use proteins with PBDs that specifically recognize the amino acid sequences surrounding phosphorylated Ser, Thr, or Tyr residues. The high cost and poor durability of these proteins has motivated efforts to develop synthetic receptors as PBDs. One example is the commercially available Phos-tag system (Scheme 1.12d), which has been used in a range of phosphopeptide separation and identification procedures, such as affinity chromatography, electroblotting, and gel electrophoresis.⁴⁰ 

Liquid/liquid extraction is a classic separation method that typically uses a lipophilic synthetic receptor to extract a water soluble analyte into a liquid organic phase or an aqueous phase that is rich in PEG.^44,45  The ability of a biomolecule to partition between two phases is determined by the partition coefficient, K_p, which is defined in eqn (1.2), where C_T and C_B are the concentration of the solute in the top and bottom phase, respectively.

For receptor-mediated extraction, the extraction constant K_ex is the product of the association constant, K_a, and the partition coefficient of the receptor/analyte complex, K_peqn (1.3).

Selectivity is the ratio of extraction constants, and receptor mediated extraction can be used to separate analytes with similar shapes and charges. For example, the cyclo[8]pyrrole receptor in Scheme 1.12e can selectively remove sulfate from an aqueous mixture containing other anions, such as nitrate, that are predicted by the Hofmeister series to be easier to extract (Chapter 4). Enantioselective extraction is another important goal and requires chiral receptors that can selectively bind to one enantiomer. The chiral receptor shown in Scheme 1.12f is able to enantioselectively extract aromatic amino acids using a three-point molecular recognition process.

Once extraction is achieved, a subsequent challenge is to reclaim the analyte and to do so in a continuous process. One solution is to convert the extraction process into a liquid membrane transport system. As shown in Scheme 1.13, the receptor resides in the organic membrane, which can be a bulk organic liquid or an organic liquid retained within a thin, porous polymer sheet. There are many examples of synthetic receptors that facilitate the transport of small hydrophilic biomolecules such as amino acids, nucleotides, sugars, and carboxylates, through liquid or supported liquid membranes.⁴⁶  With liquid membrane transport there is typically a bell shaped relationship between transport rate and affinity. High-affinity transporters can exhibit lower transport rates because release of the analyte from the membrane into the receiving phase becomes rate-limiting. Under these conditions the approximation that transport selectivity is proportional to extraction selectivity is no longer valid. Finally, co-transport and counter-transport schemes can be designed to couple the movement of a second analyte across the membrane and push transport of the original analyte uphill against a concentration gradient.

The related fields of biomolecule imaging, sensing, and diagnostics are growing rapidly and having a broad impact in many sectors of modern society. The remarkable growth is fueled by a synergistic cycle of innovation by academics developing new paradigms and commercial vendors producing turn-key instrumentation for easy operation. The ideal technology is label free and amenable to non-invasive evaluation of samples with no preliminary preparation. However, the opaqueness of many biomedical samples requires the development of exogenous contrast agents or indicators (collectively called probes) that can (a) be delivered to a desired internal site, and (b) interact with an external detection device to produce a signal that can be recorded and interpreted. The development of molecular or nanoparticle probes for imaging or sensing applications is a complex multi-parameter task that often involves close collaboration between different types of scientists and engineers. The various technologies have hugely different spatiotemporal requirements with length scales ranging from nanometer to centimeter and time scales ranging from a femtosecond to days.

The focus here is on probes with ability to recognize biomolecules and report a signal for imaging, sensing, and diagnostics. The most common technologies are microscopy and diagnostics methods and they predominantly use fluorescent reporter groups. The synthetic chemistry to make fluorescent probes is well established and they can be stored for long periods and utilized quite safely. Furthermore, the technology to produce and detect fluorescent signals is mature with many advantages, such as high sensitivity, wide spectral range of wavelengths, and miniaturized instrumentation. Fluorescence is an intellectually rich optical phenomenon, and it is possible to engineer a wide array of probe configurations that can report analyte quantities and also provide information about changes in analyte structure and dynamics. Each specific application will have its own performance requirements, such as signal detection sensitivity, image contrast and resolution, and perhaps a need for multiplex detection of multiple wavelengths. Furthermore, probe performance properties should be considered, such as synthetic accessibility, probe brightness, chemical stability, photochemical stability, photosensitization ability, and toxicity. The imaging performance requirements become more stringent as samples become larger and more complicated. Perhaps the most demanding situation is in the field of molecular imaging, where the sample is a living subject and the optical signal has to penetrate through skin and tissue. There are various signal interference problems including scattering of the light (decreases with longer wavelength), absorption of the light by endogenous biomolecules (also decreases with longer wavelength), and the subsequent autofluorescence. These problems are minimized when the optical imaging wavelength is 650–900 nm, the so-called near-infrared window. But even with an optimized near-infrared wavelength the useful signal penetration depths are a few millimeters and thus the imaging methods are only suitable for shallow sites. A particularly exciting emerging application is fluorescence-guided surgery, where fluorescent imaging probes are used to identify nerves and tumor margins and help guide surgeons during resection. But for deeper tissue penetration, other molecular imaging modalities are more appropriate, such as ultrasound, nuclear imaging, magnetic resonance imaging (MRI), or X-ray. Each modality has its strengths and weaknesses regarding sensitivity and resolution and there is a need for injectable contrast agents to improve image quality and provide molecular-level information. The contrast agents may be targeted or non-targeted and a current trend is to develop probes with multiple labels that enable imaging by two or more modalities (e.g., optical and MRI).

Shown in Scheme 1.14 are four of the most common ways that a synthetic receptor is used as a molecular or nanoparticle probe for imaging, sensing, or diagnostics. The first system is a targeted contrast agent (Scheme 1.14a). It is simply a synthetic receptor with an attached reporter group that continually emits a detectable signal. Targeted contrast agents are used increasingly for microscopic imaging of cells and tissues sections, and also for non-invasive molecular imaging of living subjects. Scheme 1.15a and b shows two probes that have been developed for the imaging of cancer and bacterial infection, respectively.^47,48  Targeted fluorescent probes are valuable in microarray diagnosis technology, where they are used to label and identify different species within a spatially arranged microarray. Even weakly selective receptors can produce unique thermodynamic or kinetic signal patterns that can be identified by appropriate pattern recognition algorithms.⁴⁹ 

The second molecular design (Scheme 1.14b) is a selective indicator (sometimes called a chemosensor). In this case the signal from an attached chromophore is modulated by a molecular recognition event. The example in Scheme 1.15c is a well-known fluorescent sensor for glucose.⁵⁰  Compared to a targeted contrast agent, a “switch-on” indicator exhibits improved sensitivity due to the lower background signal caused by the silent off-target probe. Technically it is better to produce indicator systems that change signal wavelength rather than intensity as this enables ratiometric detection which is less susceptible to systematic error. If the biomolecule association process is irreversible, then the sensing system acts as a dosimeter for batch-wise measurements. If the association is reversible, the system can be used as a longitudinal sensor for real-time monitoring.

With reversible chemosensors it is crucial that the receptor affinity for the analyte is commensurate with the analyte concentration. The rule of thumb is that the dissociation constant K_d (which equals 1/K_a) must be close to the expected concentration of analyte. A simple analogy is the choice of pH indicator to monitor acid/base titrations. The equivalence point for titrating the weak base NH₃ with the strong acid HCl is ∼pH 5.1 and can be visualized using methyl red which has a pK_a of 5.0. But methyl red is not a useful indicator for titrating acetic acid with the strong base NaOH, since it will not register a color change at the equivalence point of ∼8.8. Instead, an effective indicator would be thymol blue with a pK_a of 9.2. Now consider the design of an optical sensor for Ca²⁺ in biological samples.⁵¹  The concentration range of Ca²⁺ in blood plasma is 1.0–3.2 mM, which explains why the most common clinical fluorescent Ca²⁺ sensor for extracellular fluid has a K_d of 1.1 mM. But this fluorescent sensor does not have sufficient Ca²⁺ binding affinity for useful intracellular imaging where the concentration of Ca²⁺ is micromolar or less. Instead, a different Ca²⁺ sensor molecule is used and it has a K_d of ∼200 nM.

The third supramolecular imaging and sensing design (Scheme 1.14c) is an indicator displacement assay. In general terms, the biomolecule displaces an indicator from the receptor to produce an observable change in indicator signal. A major advantage is its technical simplicity, since the receptor does not have to be synthetically modified. The examples in Scheme 1.15d and e are ensembles based on a molecular receptor and functionalized nanoparticle, respectively.^49,52  Since the indicator and receptor are not covalently bonded, the ensemble is best suited for batch assays. Recent improvements have produced displacement systems that liberate a catalyst which enables signal amplification.^53,54 

The fourth biomolecule sensing system shown in Scheme 1.14d is a fluorescent indicator comprised of a receptor structure with several functional attributes. It is typically a flexible molecule with an appended fluorescent reporter group and complementary quencher group. Binding of a target biomolecule produces a change in receptor shape and modulation of the fluorescence signal. The best known examples of this design are molecular beacons, which are oligonucleotide receptors that target complementary sequences. The same concept has been used to create aptamers able to detect the presence of target biomolecules such as adenosine triphosphate (ATP) or pharmacological agents like cocaine (Scheme 1.15f).⁵⁵ 

Finally, it is worth noting that analytical chemosensing can be achieved using other forms of signal transduction, such as an electrochemical response or a mass response recorded by a quartz crystal microbalance.⁵⁶ 

The research topic of enzyme mimicry has a rich history within supramolecular chemistry, as researchers have spent several decades preparing and studying synthetic models of enzyme active sites.⁵⁷  In recent years, the academic topic of synthetic enzymes has begun to merge with transition-metal catalysis, and, more recently, with organocatalysis. The current state is a connected network of emerging technologies encompassing homo- and heterogeneous catalysis, asymmetric synthesis, and green chemistry.⁵⁸  Important catalyst performance parameters include the amount of reaction acceleration, catalyst turnover number, and degree of asymmetric induction.⁵⁹  Ideally, the enantioselectivity for an asymmetric reaction pathway should be >50, which is the ratio needed to obtain 99% ee.

The goal of “biomimetic chemistry” is to produce new chemistry using mechanisms that resemble processes exhibited by living systems. In the context of synthetic enzymes, the aim is to accelerate chemical reactions of selected substrates using concepts from enzymology, such as lock and key association to form a catalyst/substrate complex, transition state stabilization, covalent stabilization of intermediates, and general acid or base catalysis (Scheme 1.16).⁶⁰  The classic Michaelis–Menten model of enzyme catalysis includes two system-specific parameters, V_max, the maximum reaction velocity, and K_m, the Michaelis–Menten constant, which is equal to the substrate concentration when the rate is 1/2 V_max. For fast binding catalyst/substrate complexes, K_m is equal to 1/K_a. Thus, K_m for synthetic enzymes can be controlled by rational molecular design.

Classical efforts to produce synthetic enzymes de novo start with structural information about the crucial catalytic residues within enzyme active sites. The advent of advanced computer modeling methods has greatly enhanced the capability to rationally design an enzyme mimic, but it is worth emphasizing that it is much more challenging to model a catalyst/substrate transition-state than a ground-state receptor/guest complex. Catalyst fabrication has typically involved multistep synthetic procedures that append catalytic residues to organic container molecules such as cyclodextrins, crown ethers, calixarenes, or porous dendrimers. The illustrative example in Scheme 1.17 shows a high-generation dendrimer with a pyridoxamine cofactor at the core.⁶¹  The system is able to accelerate transamination reactions 1000 times faster than free pyridoxamine. In recent years the receptor scaffold options have expanded with the development of rapid self-assembly methods that use metal–ligand interactions to produce porous hollow cages that can encapsulate and constrain reactive substrates.

An alternative way to develop an effective catalyst based on a synthetic receptor is to use combinatorial synthesis and screening methods. In short, the goal is to create a large library of candidate structures and then screen them to find a member with appropriate catalytic performance. This approach works well for molecules that can be synthesized rapidly in combinatorial fashion, such as synthetic peptides⁶²  or oligonucleic acid aptamers.⁶³  An example of this approach to catalyst discovery is the work of Reymond and coworkers, who examined libraries of peptide dendrimers (Scheme 1.18a) with structures containing a judicious mixture of hydrophobic side chains to encourage guest binding and polar groups for reaction catalysis. Screening studies revealed library members with strong esterase activity, including enantioselective catalysts.⁶⁴ 

The most common biomolecule reactions that are promoted by synthetic catalysts are cleavage processes (Scheme 1.16). In contrast, examples of biomolecule ligation reactions are fairly rare, especially in water.⁶⁵  A common drawback with ligation catalysts is limited catalyst turnover due to product inhibition. This occurs often because the ligated product molecule has higher affinity for the catalyst than the two smaller reactant molecules. Product inhibition is typically not a problem in cleavage reactions and a lot of effort has worked towards artificial ribonucleases, esterases, and peptidases.⁶⁶  Many early artificial enzymes were based on macrocyclic scaffolds with attached nucleophilic groups that promoted bond cleavage events. A classic example is the azacrown receptor in Scheme 1.18b that catalyzes the hydrolysis of ATP in aqueous solution with impressive turnover.²  Another is a cyclodextrin with two appended imidazoles for general acid/base catalysis (Scheme 1.18c).⁶⁶  In general, however, synthetic organic receptors are not effective catalysts of biomolecule hydrolysis reaction in water. In comparison, synthetic receptors that incorporate one or more Lewis acidic metal centers are more successful, and most research has been focused on these metalloenzyme mimics.⁶⁷  Typically, the metal cation acts as a Lewis acid and activates hydrolysis reactions by polarizing bound substrates, stabilizing deprotonated nucleophiles, or stabilizing anionic intermediates and associated transition states. The more sophisticated catalyst designs have structures containing multiple metal centers or a combination of metal center and secondary hydrogen bonding sites that produce a cooperative catalysis effect (Scheme 1.18d).⁶⁸  Appropriately designed micellar assemblies⁶⁹  and coated nanoparticles (Scheme 1.18e)⁷⁰  can also effectively catalyze cleavage reactions in aqueous solution.

There are relatively few examples of supramolecular catalysts that promote functional group transformations, and the best known examples are oxidation/reduction processes.⁶⁶  Most of the catalysts are conjugates of container molecules with appended enzyme cofactors. The cofactor could be a redox active metal that promotes one electron transfer processes (an example is the cytochrome P450 mimic in Scheme 1.18f),⁶⁶  or alternatively it could be an organic structure like pyridoxamine, thiamine pyrophosphate, flavin, or nicotinamide adenine dinucleotide that promotes two electron processes.

Efforts to create MIPs with catalysis properties (plastic enzymes) have focused on templates that are transition-state analogues. That is, the shapes of the binding sites created by the template molecules have geometries and polarities that match reaction transition states.⁷¹  Most of the catalyzed reactions are cleavage processes that tend to avoid problems associated with product inhibition. Efforts to produce MIPs with increased catalytic power have incorporated transition metals as Lewis acid catalysts into the polymeric matrix.⁷¹  In addition, nanoscale particles composed of various organic and inorganic composites are emerging as promising catalysts for many types of reactions involving biomolecules (nanozymes).⁷² 

Future research into enzyme mimics is expected to include efforts to produce catalysts whose activity is regulated by effector molecules or external stimuli.⁷³  Another likely direction is next-generation designs of molecular self-replication due to autocatalysis; that is, a synthetic receptor that catalyzes a reaction that produces another copy of itself, leading to exponential growth. One of the motivations for this field of research is to find possible explanations for the emergence of life from the primordial soup of molecules that existed on early earth.

The majority of small-molecule pharmaceuticals (drugs) work by binding tightly and selectively to a crevice within a medically relevant protein. The protein crevice could be an enzyme active site, receptor binding site, allosteric site, or binding interface with a partner biopolymer. The drug discovery pathway to find the perfect compound that binds tightly and selectively to a protein crevice is long and arduous. The traditional medicinal chemistry approach employs an iterative cycle of synthesis and screening protocols to generate empirical trends that guide follow-up studies. The search can be augmented by a wide range of predictive tools that help narrow the choice of candidate molecules for synthesis and testing. Binding site docking algorithms utilize computational models to predict association strengths based on non-covalent interactions. There is no doubt that this part of the drug discovery process is enhanced by a deep understanding of the fundamental concepts that control supramolecular assembly.⁷⁴ 

Increasingly, the field of supramolecular chemistry is able to produce synthetic receptors that recognize biomolecules effectively under physiological conditions. Inspired by this success, researchers are beginning to devise novel supramolecular strategies to elicit pharmacological outcomes. Especially attractive are synthetic receptors with the capacity to modulate the four protein/biomolecule association systems illustrated in Scheme 1.19. In principle, this can be achieved by targeting the protein or the biomolecule partner. The affinity has to be strong enough to saturate the target biomolecule at dosing concentrations, which are typically micromolar or less. Selective binding of the target molecule is, of course, only one of the many necessary properties that have to be optimized in order to create a useful pharmaceutical agent. Eventually the pharmaceutical performance of a synthetic receptor has to be judged by the full array of classic ADMET parameters. In this regard it is inherently easier to optimize the pharmaceutical properties of a small molecule than a relative large nanoparticle, and the task becomes even more challenging if the nanoparticle is a non-covalent assembly of multiple components.

Examples of small-molecule pharmaceuticals that block protein/protein association are already known,⁷⁵  which supports the feasibility of future efforts to achieve this type of pharmacological effect in a rational way using synthetic receptors. Molecules that can cover a large area of exposed protein surface are expected to be most effective and the design criteria to achieve this are discussed further in Chapter 9. Prototype protein receptor systems include extended organic scaffolds such as porphyrins with appended arms, nanoparticle molecularly imprinted polymers,⁷⁶  and core–shell nanoparticles coated with a multicomponent monolayer. In terms of pharmaceutical performance, the most clinically advanced synthetic recognition systems for protein surfaces are aptamers. As stated above, pegaptanib was the first aptamer-based therapeutic agent approved for treating age-related macular degeneration. Pegaptanib works by binding tightly to vascular endothelial growth factor (VEGF) within the intravitreous region of the eye and inhibiting VEGF binding to cell surface receptors, thus preventing the formation of new blood vessels in the eye.⁷⁷ 

In certain cases, the pharmaceutical target is a cell surface that contains many copies of a protein and the pharmaceutical effect is enhanced if binding is increased. This situation invites a targeting strategy based on multivalency, which is expected to increase affinity and also improve cell selectivity. Most published examples of multivalent targeting use a nanoparticle that is coated with multiple copies of a targeting ligand such as a vitamin, peptide, or carbohydrate. The individual ligand may only have moderate affinity for the membrane-bound target, but the multivalent nanoparticle binds strongly to multiple copies of the target on the cell surface. Some of the structural factors that control affinity and selectivity of multivalent binding systems are discussed in Chapter 2. A key variable is receptor abundance on the cell surface. In some cases, binding alone produces a desired pharmaceutical outcome, like blocking viral infection using nanoparticles coated with carbohydrates. In other cases, the multivalent targeting triggers a cell entry process such as endocytosis and thus it can be used for selective drug delivery. A good example is the targeting of folate-coated liposomes to cancer cells that overexpress the folate receptor.⁷⁸ 

The same multivalency concepts are applicable if the targeting ligand on the nanoparticle is a synthetic receptor.⁷⁹  This strategy was employed in a recent study of micelle nanocarrier delivery of cis-platin drug to cancer.⁸⁰  The drug delivery was enhanced when the surface of the nanocarrier was decorated with multiple boronic acid groups (Scheme 1.20a) that targeted over-expressed sialic acid residues on the surface of cancer cells. The boronic acid coated nanocarrier was able to deliver comparatively higher doses of the anticancer drug to solid tumors and reduce the growth rate of melanoma in animal models.

Nucleic acids are attractive pharmaceutical targets for synthetic receptors since they are intimately involved in cell replication and transcription processes and ultimately cell proliferation. The surface of DNA and RNA presents a rich assortment of binding targets including the anionic phosphate backbone, the large hydrophobic surfaces of the nucleobases, and the hydrogen bonding motifs on the exposed edges of the major and minor grooves of the double-stranded helix. The many different types of synthetic receptors that achieve effective nucleic acid recognition are described in Chapter 7. Arguably the most impressive programmable system with significant biological activity is the nucleic acid binders produced by Dervan and coworkers. The group has developed a series of small molecule pyrrole-imidazole polyamides that bind to double-stranded nucleic acids with extremely high affinity and exquisite sequence selectivity. The recognition is based on hydrogen bonding patterns created by the alternating pyrrole and imidazole groups within the receptor structure (Scheme 1.20b). Sequence specificity is determined by pairing rules that maximize hydrogen bonding with the donor-acceptor sites that are exposed on the edges of the stacked nucleobases. For example, an imidazole-pyrrole pair distinguishes G·C from C·G and also the A·T/T·A base pairs, while a pyrrole–pyrrole binding pair is selective for A·T over G·C.⁸¹  These receptors have been used to block numerous RNA/DNA–protein interactions that control cell function, including RNA polymerase II,⁸²  nuclear factor-κB DNA binding,⁸³  and have been shown to control gene expression in a xenograft animal tumor model.⁸⁴  Studies of in vivo efficacy have encountered obstacles concerning delivery, toxicity, and slow clearance.⁸⁵  However, the modularity of the polyamide design allows manipulation of the ADMET properties, and thus this class of synthetic receptors has promise for continued pharmaceutical development.

Another way that synthetic receptors can produce pharmaceutical activity is to alter the cellular distribution of a specific biomolecule. In principle, it is an attractive way to treat cells that have become pathological due to an abnormal accumulation of lipophilic biomolecules. An example is 2-hydroxypropyl-β-cyclodextrin (2HPBCD) (Scheme 1.20c), a synthetic receptor for cholesterol with promising pharmaceutical activity. It was recently reported that 2HPBCD has efficacy against a mouse model of Niemann–Pick type C (NPC) disease, a rare but fatal neurodegenerative disorder caused by a defect in the NPC1 protein and characterized by widespread intracellular accumulation of cholesterol.⁸⁶  Currently, there is no effective treatment for NPC disease, which primarily affects children. Recent experiments using a mouse model of the disease found that systemic treatment with 2HPBCD significantly prolonged lifespan by allowing trapped cholesterol within the late endosome/lysosome to be released into the cell cytosol where it was normally metabolized.⁸⁷  Ongoing studies are attempting to determine exactly how 2HPBCD induces its therapeutic effect, but the reversal of abhorrent cholesterol trafficking is strongly implicated.

It is well known that healthy living cells maintain a tightly regulated transmembrane distribution of ion concentrations. This balance is maintained by the collective action of a series of endogenous active and passive transport systems, and disruption of this ion distribution leads to cell malfunction. There are logical reasons to develop pharmacological agents that can alter transmembrane ion gradients; they can either become a method to kill pathological cells, or a way to restore ion gradients in genetically abnormal cells. Ionophores are chemicals that associate with ions and many facilitate ion transport across cell membranes. Historically the term ionophore referred to cation binding molecules, but these days many synthetic anion binding ionophores are also known.⁸⁸  Quite a few natural products are known to kill cells by permitting unregulated ion flow across the plasma membrane.⁸⁹  These compounds are produced by microbes and include the ionophore antibiotics, a wide range of natural products that promote ion transport by mechanisms that resemble the classic channel and carrier processes shown in Scheme 1.21. In some cases, the ionophore antibiotics form well-defined 1 : 1 lipophilic complexes with metal cations (e.g., monensin A and valinomycin) and they act as membrane transport carriers. In other cases (e.g., gramicidin A) they exhibit classic channel behavior, whereas yet others self-assemble within membranes to create pores (e.g., gramicidin S and alamethicin). Sometimes the pore assembly process involves molecular components within the target membrane. An example is amphotericin B, an antifungal drug that binds with ergosterol, a component of fungal cell membranes, and forms a transmembrane channel that promotes ion leakage (Chapter 10).

Over the years, supramolecular chemists have prepared a large number of synthetic ionophores and determined if they transport ions across membranes by carrier, channel, or pore mechanisms (Chapter 3). Many of these compounds have been tested for cytotoxicity. In general, simple crown ethers are not highly toxic molecules. However, there are examples of channel-forming systems having significant toxicity.⁹⁰  For example, Gokel's hydrophiles (crown ether oligomers with channel transport behavior; Scheme 1.20d) are toxic to mammalian and bacterial cells.⁹¹  Amino-substituted cyclodextrins have also been shown to exhibit strong antimicrobial activity and an ability to promote K⁺ efflux from both Gram-positive and Gram-negative bacteria.⁹²  It appears that many synthetic cation ionophores are likely to be cytotoxic and the pharmaceutical challenge is to target them to sites of disease such as cancer or microbial infection.

Anion-binding ionophores have also been shown to exhibit a range of interesting pharmaceutical properties. For example, the naturally occurring prodigiosins exhibit antimicrobial, immunosuppressive, and anticancer activities.⁸⁹  These compounds are known to promote HCl co-transport, but how this relates to their pharmacological properties is not exactly clear. Nonetheless, synthetic receptors with HCl co-transport ability have been shown to exhibit interesting anticancer properties (Scheme 1.20e).⁸⁹ 

An alternative pharmaceutical goal using ionophores is not to kill cells, but rather to restore ion gradients in genetically abnormal cells. Certain genetic conditions, including Bartter syndrome, Best disease, and cystic fibrosis are caused by defects in endogenous cell membrane transport channels. It has been suggested that exogenous addition of synthetic chloride transporters may restore transmembrane chloride gradients to normal levels. A large number of synthetic anion transporters have been tested in model membranes and cell culture, and many have exhibited potent chloride transport ability (Scheme 1.20f).^93,94  However, there are formidable challenges to solve concerning tissue specific delivery and extended residence time in the appropriate apical membrane of the target cells.

It is not possible to predict with any certainty how the next generation of synthetic receptors for biomolecules will be employed. Instead, this concluding section briefly describes some of the emerging classes of functional molecules and supramolecular systems that will likely be incorporated within transformative new technologies.

The most complicated known machine in the universe is the human brain and there is an active community of scientists and engineers trying to create molecule-based computers.⁹⁵  The nanoscale dimensions of biomolecules make them attractive building blocks for bottom-up fabrication of high-density information devices for storage and calculation. There are already a number of biomolecule sensing systems that use logic gates to report analyte levels in biomedical samples.⁹⁶  The sensors only produce a signal when there is a precise number of multiple analytes, thus providing sophisticated information about the state of a living biological sample. There is also a growing number of synthetic receptor systems that can selectively modulate biomolecule association events.⁹⁷  Merging these topics will undoubtedly lead to biomolecular logic devices that mimic the complex signaling pathways of cells.

An extension of logic devices is the field of responsive materials for controlled release and drug delivery.⁹⁸  Current methods for drug release usually rely on passive diffusion of the drug molecules from an inert vehicle into the physiological environment. There are many prototype examples of release systems that can be triggered by external stimuli, such as light, heat, or change in pH, and increasingly there are new classes of advanced materials with properties that respond to the presence of biological molecules.⁹⁹  State of the art examples include glucose-responsive systems that trigger the release of pharmaceutically relevant proteins¹⁰⁰  or cells,¹⁰¹  and synthetically modified protein tubes that are activated by intracellular ATP to disassemble and release encapsulated guests.¹⁰² 

One of the unpleasant realities of drug development is the fact that many chemical entities that bind strongly to a validated pharmaceutical target in vitro do not become useful drugs because they cannot reach the target in vivo. A dogma within drug discovery is Lipinski's rule of five,¹⁰³  which is a set of simple structural criteria that help predict if a drug candidate is likely to be orally active. A corollary of these rules is that high molecular weight and/or highly charged molecules are unlikely to cross cell membranes, which means that they will be ineffective against intracellular targets. This includes biomolecular structures such as peptides, antibodies, enzymes, nucleotides, nucleic acids, aptamers, and many carbohydrates. Furthermore, many of these structures are susceptible to enzyme degradation in the bloodstream. Thus, there is a need for drug delivery vehicles that will protect these fragile agents and deliver them into the cytoplasm of target cells. To date, there has been considerable progress in the area of transfection and gene delivery in vitro using supramolecular assemblies such as cationic polymers and amphiphiles.⁵¹  The next major challenge is in vivo organ-specific delivery. A reason for great optimism is the rapid ongoing development of biocompatible, self-assembled nanocomposites with multiple attributes that allow cell targeting, in vivo imaging, and drug delivery.¹⁰⁴  Similarly, there are elegant new designs of delivery molecules, such as “molecular umbrellas”, which are facially amphiphilic receptors that can encapsulate biomolecules in aqueous solution and then invert polarity as they pass through the lipophilic core of a cell membrane.¹⁰⁵ 

The molecular machines that operate inside cells, like the ATP-powered motor proteins that drive the contraction of muscle fibers in animals or transport biological cargo along microfilaments, have captured the imaginations of many supramolecular chemists. The recent literature contains a diverse array of brilliantly designed molecules with machine-like motions such as rotors, switches, and shuttles.¹⁰⁶  But the functional performances of these primitive, early-generation molecules are very modest when compared to their much larger and more complicated biological counterparts. The challenge of fabricating biomimetic supramolecular systems that can convert biomolecule binding energy (input) into coherent mechanical movement (output) seems daunting.¹⁰⁷  At present the most advanced projects utilize base-pairing of nucleic acids as the binding event that triggers macroscale movement,^108,109  or they prepare nanomechanical devices using the rules of DNA origami.¹¹⁰ 

The long-term goal is to produce nanoscale robots for various applications in artificial life, synthetic biology, and nanomedicine.

We are grateful for funding support from the National Science Foundation (Arlington, VA, USA) (CHE1401783).