Introduction

What is now proved was once only imagined

William Blake, 1757–1827

The recognition of a target molecule by a synthetically prepared receptor has captured the imagination of supramolecular chemists. Since its inception, research in this area has been instrumental in elucidating the mode of action of a great many biological events concerning recognition and catalysis. The importance of this work was underlined with the award of the Nobel Prize in Chemistry to Cram, Lehn and Peterson in 1987 "for their development and use of molecules with structure-specific interactions of high selectivity." Since then the diversity of compounds studied under the umbrella of supramolecular chemistry has grown significantly. Of particular interest are chemical molecular sensors, single molecules with the ability to both recognise and signal analytes in real time.

The development of coherent strategies for the selective binding of target molecules, by rationally designed synthetic receptors, remains one of chemistry's most sought after goals. The research conducted to this end is driven by a fundamental inquisitiveness and need to monitor compounds of industrial, environmental and biological significance.

Within our research groups we have exploited the interaction between boronic acids and diols. The primary interaction of a boronic acid with a diol is covalent and involves the rapid and reversible formation of a cyclic boronate ester. The array of hydroxyl groups present on saccharides provides an ideal scaffold for these interactions and has led to the development of boronic acid-based sensors for saccharides (Scheme 1).

Many synthetic receptors developed for neutral guests have relied on non-covalent interactions, such as hydrogen bonding, for recognition. It is the case, however, that in aqueous systems neutral guests may become heavily solvated. While biological systems have the capacity to expel water from their binding pockets and sequester analytes wholly, using non-covalent interactions, synthetic monomeric receptors have not yet been designed where hydrogen bonding has been able to compete with bulk water for low concentrations of monosaccharides. However, it should be pointed out that progress is being made in this area and recently Davis reported a hydrogen-bonding receptor capable of binding D-glucose in water with a weak but significant stability constant.

The capacity of boronic acid receptors to function effectively in water is reflected by the number of published sensory systems designed around them. The most popular class of the fluorescent boronic acid-based sensors utilise an amine group proximal to boron. The Lewis acid-ewis base interaction between the boronic acid and the neighbouring tertiary amine has a dual role. First, it enables molecular recognition to occur at neutral pH. Second, it can be used to communicate binding by modulating the intensity of fluorescence emission through photoinduced electron transfer (PET), introducing an "off-on" optical response to the sensor.

The quality of the research in this area, particularly in the past few years, has led to a significant advance in the understanding of the basic science behind the generic mode of action of this class of receptor. This book will therefore bring together and critique the contemporary scientific understanding of the fundamental processes involved in the molecular recognition of saccharides by boronic acids. It should be noted that a comprehensive overview of this nature has not been previously reported in the scientific literature. A literature review will then investigate the application of these sensory systems. Although this review cannot be exhaustive, it is our intention to illustrate the current state of play in the field.

The Molecular Recognition of Saccharides

The expression "fits like a glove" is an odd one, because there are many different types of gloves and only a few of them are going to fit the situation you are in.

Lemony Snicket

A Series of Unfortunate Events: The Grim Grotto: Book the Eleventh

2.1 Molecular Recognition

Molecular recognition lies at the very heart of sensor chemistry. The process itself involves the interaction between two substances, often termed as a host and a guest, a lock and a key or a receptor and a substrate. Importantly, recognition is not just denned as a binding event but requires selectivity between the host and the guest.

Selectivity between host and guest is a premise of compatibility. It arises between compounds with carefully matched electronic, geometric and polar elements. For synthetic receptors, the potential, therefore, exists to engineer receptors for any chosen analyte through judicious structural design and functional group complementarity. The power of this concept is illustrated within Nature. Biological systems have evolved with exquisitely constructed active-binding sites, sequestering guest molecules with near perfect selectivity.

Nevertheless, for the recognition event at a receptor to be of practical use, a further element is required. A channel of communication must be established between the receptor and the outside world. This additional quality converts a receptor into a sensor.

For a sensor to function, it must, therefore, permit selective binding to occur between host and guest and also report these binding events by generating a tangible signal. By performing these two fundamental tasks, sensors have the potential to relay information on the presence and location of important species in a quantifiable manner, bridging the gap between events occurring at the molecular level and our own (see Scheme 2).

Chemical sensors can be broadly categorised as either biosensors, or synthetic sensors. Biosensors make use of existing biological elements for recognition. Many of the physiologically important analytes already have corresponding biological receptors with intrinsically high selectivity and if these receptors can be connected to a signal transducer, a biosensor can be developed.

Synthetic sensors incorporate a synthetically prepared element for recognition. While biomimetic receptors have been prepared, with synthetic receptors mimicking the active sites in naturally occurring biological molecules, synthetic receptors can, and often are, designed entirely from first principles.

2.2 The Importance of Saccharides

2.2.1 Saccharides and Carbohydrates

In keeping with convention, the term saccharide is used within this book to refer broadly to polyhydroxylated carbohydrates. The product of photosynthesis, carbohydrates single-handedly account for the most prolific class of organic compounds that can be found on the surface of the Earth. Within biology they are of fundamental significance. In their most ubiquitous roles, they endow Nature with structural rigidity, in the form of cellulose, and function as the energy store that sustains life, in the forms of starch and glycogen.

Not only are these compounds abundant they are also incredibly versatile. Oligosaccharides are involved in protein targeting and folding, as well as controlling the cell recognition events for infection, inflammation and immunity. From a medicinal perspective, the monitoring of D-glucose has proved of particular importance. D-glucose provides the metabolic energy for most cells of higher organisms. In humans, a breakdown in the transport pathways of Dglucose has been linked to conditions such as cancer, cystic fibrosis and renal glycosuria, but by far the most prevalent condition resulting from ineffective D-glucose transport is diabetes mellitus.

2.2.2 Diabetes Mellitus

Diabetes presents one of the largest health challenges to face us in the 21st century. Current reports indicate that diabetes affects 5% of the global population. In the UK, the increase in obesity, population age and a progressively more sedentary lifestyle has seen the prevalence of Type 1 diabetes double every 20 years since 1945. Diabetes is associated with chronic ill health, disability and premature mortality. From a physiological perspective, the debilitating long-term complications include heart disease, blindness, kidney failure, stroke and nerve damage leading to amputation.

At an economic level the repercussions are also serious. Within the UK, 5% of the National Health Service's budget is spent on treating diabetes and its complications. This equates to £3.5 billion per year or £9.6 million per day. Following extensive and widespread trials, unequivocal evidence exists that monitoring and adjusting diabetic blood sugar levels to maintain them within tight boundaries dramatically reduuces the health risks faced by diabetics.

2.2.3 Structure of Saccharides

The detection of saccharides presents a curious challenge. Bristling with linked arrays of hydroxyl groups, saccharides are structurally complex. The linear form of D-glucose, for example, contains four stereocentres. Considering just its immediate family, the aldohexoses, this presents us with 16 stereoisomers (see Figure 1).

In aqueous solution, this complexity is further compounded by mutorota-tion. Cleavage of the hemiacetal ring causes interconversion between the pyranose and furanose ring forms, via an acyclic intermediate, with inversion of configuration at the anomeric centre equilibrating the α- and β-anomers thus altering the optical rotation of the solution (see Figure 2). These properties make saccharides difficult to differentiate from each other. Furthermore, in water receptors for saccharides may be heavily solvated, differentiation between water and hydroxyl groups presenting a formidable challenge in its own right.

This point can be illustrated by considering the protein family, lectins. Aside from antibodies, lectins are largely responsible for carbohydrate recognition within biological systems. Lectin–oligosaccharide-binding constants are, by biological standards, unusually small, commonly in the range 103–104 M-1.

The crystal structure of the Escherichia coli galactose-binding protein with a bound molecule of (β-D-glucose was obtained. From the structure, it can be observed that recognition is achieved by sequestering the monosaccharide entirely beneath the protein surface and expelling bulk water from the active site. This not only fosters a favourable entropic stabilisation but also allows the surrounding amino acid residues unfettered access to the substrate. In the binding site of the molecule (β-D-glucose is sandwiched between two apolar phenylalanine and tryptophan residues axial to the ring. Equatorially, an array of 13 different complementary hydrogen-bonding interactions is provided by one advantageously sited water molecule and eight individual amino acid residues, all ideally located (see Figure 3).

2.2.4 Home Blood Glucose Monitoring

Commercially, the preferred tools for sensing complex molecular species have relied on the high specificity exhibited by antibodies and enzymes. Most clinical systems currently available for measuring blood glucose levels rely on the glucose oxidase enzyme (GOx, also commonly abbreviated to GOD or GO).

The majority of these home blood glucose monitoring tools rely on the invasive withdrawal of blood, typically from a pricked finger, followed by application of the sample to an amperometric enzymatic test strip allowing GOx to catalyse the oxidation of glucose to gluconic acid (see Scheme 3).

Early glucose monitors measured the production of hydrogen peroxide by oxidation at a single-working electrode, as in Equation (1). At a constant voltage, the current generated across the cell is proportional to the concentration of hydrogen peroxide, which is in turn proportional to the glucose concentration in the sample under investigation.

H2O2 [right arrow] O2 + 2H+ + 2e- (1)

As this method of monitoring glucose is heavily dependent on the oxygen concentration, a dimethylferrocene mediator was developed. At a set potential of +160 mV glucose is oxidised with concurrent reduction of the dime-thylferricinium cation; the applied potential serving to oxidise and thus recycle the resulting dimethylferrocene back to the dimethylferricinium cation (see Scheme 4).

There can be no question that the availability of affordable home blood glucose monitoring has revolutionised the quality of life experienced by diabetics. However, there are some inherent limitations with an enzymatic approach. The systems have to be stored appropriately, they are specific only for a few saccharides and in most cases they become unstable under harsh conditions and hence cannot be sterilised. For this reason, much work has been focused on the development of synthetic sensors with the capacity to monitor saccharides under a broad range of environmental conditions, and thus allow access to a wider spread of diagnostic applications.

2.3 Non-Boronic Acid Appended Synthetic Sensors for Saccharides

A great amount of attention is currently devoted to the development of synthetic molecular receptors with the ability to recognise saccharides. This book is primarily concerned with the role of boronic acids in that recognition process, although many systems have been developed which use non-covalent interactions for recognition. A few examples are given below to provide a brief illustration of research in this area. (For a more comprehensive insight, the reader is directed to two recently published reviews.)

The first synthetic saccharide receptor 9 was documented in 1988 by Aoyama et al. The calixarene framework is not large enough to actually permit the encapsulation of a saccharide within the central annulus of the bowl, but it does permit "face-to-face" recognition to occur between the calixarene's upper rim hydroxyls and those of the saccharide. However, in order to complex saccharides in water Aoyama needed to use very high concentrations of saccharides (100 mM-1 M).

Inspired by the structure of the saccharide-binding proteins such as the active site of the Escherichia coli galactose-binding protein, illustrated in Figure 3, Davis and Wareham developed the octaamide 10. The sophisticated architecture mimics Nature in that the binding pocket is designed so as to completely encapsulate a host saccharide. The planar hydrophobic surfaces of the dual diphenyl groups provide internal apolar contacts above and below the aliphatic ring. Located equatorially to the saccharide host an array of secondary amides induces favourable hydrogen-bonding interactions, anchoring the hydroxyls on the saccharide ring.

Even with such elegant design the reported stability constants (Ka) for sensor 10 in deuterated chloroform are reduced significantly on addition of competitive co-solvents. In the presence of octyl-β-D-glucoside 11 ~300-fold reduction was witnessed in the stability constant (Ka) of sensor 10 on addition of 8% deuterated methanol to the system. The stability constants (Ka) for sensor 10 with octyl-β-D-glucopyranoside 11 were 300,000 M-1 in deuterated chloroform and ~1000 M-1 in 92:8 deuterated chloroform/deuterated methanol. More recently, the dodecacarboxylate version of this receptor 12 has been prepared and this system does bind D-glucose in water with a stability constant (Ka) of 9.5M-1. This is much better than any of the previous systems but the binding is still to weak in water for these systems to be useful in detecting saccharides in biological systems.

As well as developing recognition sites through careful structural pre-organisation, non-covalent interactions of increased strength have also been studied. Manipulation of anionic centres has proved successful as functional groups such as phosphates, phosphonates and carboxylates can be powerful hydrogenbondacceptors. While the pre-organisation can be reduced, it has been found that these anionic centres provide a degree of structure in their own right. Flexible sensors with multiple anionic groups will conform to the tendency of the charged groups to repel each other so as to minimise electrostatic repulsions.

Das and Hamilton have employed phosphonate moieties in sensors such as the rigid chiral spirobifluorene 13. The racemate was used directly and although the exact nature of the hydrogen-bonding motif could not be determined excellent binding was observed with octylglucosides in organic solvents. The enantioselectivity was also determined and reported as ~ 5.1:1, the greatest value reported for saccharide enantioselective discrimination by non-covalent interactions. The binding constant (Ka) for sensor 13 with octyl-β-D-glucopy-ranoside 11 was 47,000 M-1 in deuterated acetonitrile.

Diederich and co-workers have made use of binaphthalene-derived macro-cycles appended with internal phosphate groups to provide a ring of hydrogen-bonding sites within a central recognition cavity. One of the first of these, sensors 14, was shown to have good selectivity for suitably sized saccharides such as the octyl-D-β-glucopyranoside 11, the calculated interphosphate distance of 7.2 Å being designed so as to accommodate monosaccharides exclusively. A stability constant (Ka) of 5200 M-1 was reported for the complexation of sensor 14 and octyl-D-β-glucopyranoside 11 in 98:2 deuterated acetonitrile/deuterated methanol.

The major obstacle faced by all of the synthetic receptors above, which are wholly reliant on non-covalent interactions, is that of solvent competition. In aqueous systems, neutral guests may become heavily solvated and are therefore unable to monitor glucose in media of specific interest to analysts such as blood, urine, tear fluid, beverages, foodstuffs and so forth. Because, monitoring saccharide interactions in water with these receptors is difficult they are typically evaluated in aprotic solvens such as chloroform. By using aprotic solvents such as chloroform simple saccharides can not be used because they are insoluble, when using these solvents it is necessary to use soluble O-alkylated saccharides. In overcoming this hurdle, the covalent interaction between boronic acids and saccharides has proved hugely advantageous.