Biology takes a back seat once the business of organic synthesis begins.

Biology is, of course, at the heart of medicine. But drugs, when all is said and done, are really just chemicals. Biology can give clues as to what compounds may or may not be bioactive, but hard core reaction chemistry is needed to synthesize proposed drug molecules in efficient and pure form. Pharmaceutical researchers must inevitably take their biological themes and combine them with old-fashioned chemical synthesis to create the multiplex of compounds needed for drug testing. Automated and streamlined, this process is called combinatorial chemistry. Although the process was born of peptide synthesis, it is now being adapted to the production of small bioactive molecules en masse. Questions concerning solution- versus solid-phase synthesis, type and size of libraries produced, and which of a multiplicity of possible scaffoldings and reactions can best be used are at the heart of this rapidly progressing field where test-tube organic synthesis transforms into drug discovery.

Solutions Aren’t the Solution?
Working in solution allows combinatorial researchers to “use the vast body of existing synthetic chemistry methodology, which provides this technology platform with the same synthetic tools as those available to a medicinal chemist” (1). Indeed, the list of reactions used in combinatorial synthesis would make an excellent table of contents for an elementary organic chemistry textbook, except for the fact that it would be far too extensive for any normal college student to absorb in a single semester. The chemistries used range widely, running an alphabetical and named-reaction gamut from acid hydrolysis to Diels–Alder to Wittig reactions.

But despite reactions in solution being the historical mainstay of organic chemistry, synthesis in solution makes up only some 30% of libraries produced, according to a recent review (2). There are many reasons for this seemingly contradictory state of affairs, but perhaps the most important one is the unique requirements of combinatorial chemistry; it is difficult to rapidly process compounds in solution from one set of reaction conditions to another. In larger libraries, there is the even greater problem of tracking and identifying individual compounds at the end of the process. Solid supports easily address these two difficulties.

Making It Go Faster

Because many of the researchers and companies most interested in using combinatorial libraries are often more biologically oriented than chemistry proficient, an entire industry has developed to do some of the complex synthesis for them. These contract research organizations will happily produce combinatorial libraries on demand, according to pharmaceutical specifications. For researchers who want to get their own hands dirty, metaphorically speaking, another host of companies have developed to provide an extensive set of tools from multiwell pipetters to automated workstations to help get the job done in-house. For an extensive list of companies involved in the production of combinatorial libraries and the automation of combinatorial chemistry, see www.5z.com/divinfo/company.html.

By far, the majority of modern libraries are created using solid-phase supports, generally a type of bead. These beads are most often made of polystyrene resin, although variations exist and more types are rapidly being developed, including some based on poly(ethylene glycol) derivatives. Beads perform several critical functions in combinatorial chemistry—from controlling the chemistry itself to purification and identification of the final products. Primarily because of the so-called “washout factor”, the use of beads permits a variety of beneficial chemical procedures, including, for example, the mass-action effect of using reagent excess to drive reactions favorably toward product. Beads also permit radical changes in the sequence of chemistries applied because the previous solutions can be completely removed with ease. The vast array of chemical possibilities and the often pronounced difficulties encountered in optimizing reactions, especially in multiple stages of solid-phase synthesis, have created one industry to do the work and another to automate it (see box, “Making It Go Faster”).

The typical bead uses a chemical linker to attach the compound to the bead. Properly designed, the linker serves as a cleavage point for releasing the created compound after the complete series of reactions is finished. Figure 1 (below, right) shows the multiplicity of identical compounds attached to a typical bead after synthesis is complete. Each bead is the geographical location of an individual compound (really a collection of copies of an individual compound) in a combinatorial library.

Going to the Libraries

Figure 1: Autoradiograph of a bead covered with peptide compounds.

There are three main combinatorial library formats. The choice of which format to use is dependent on several factors, from library size to the type of assays available to detect desired compounds.

Parallel synthesis is generally used for small libraries (from the hundreds to thousands of compounds produced), most often those evolving from preexisting template information in a rational drug design approach. In parallel synthesis, each reaction sequence is carried out separately and simultaneously with every other reaction sequence. The major benefit of parallel synthesis is that it is easy to determine the synthetic heritage of the individual compounds produced—all members in an identifiable well or tube are identical. Large libraries can be produced using massive parallel synthesis, but this generally requires a heavy investment in automation.

Split, or split-pool synthesis, is a mixing technique that can produce libraries in the thousands to tens-of-thousands range. Although the sequence can vary, typically, a variety of starting scaffolds are produced, generally on beads, reacted, pooled, and subjected to a subsequent round of chemistry. These products can be split and then reacted again, with or without repooling, in however many multiples are desired or required. The greater the number of steps, and/or the more frequent the pooling, the larger the potential number of library members produced. Tagging is frequently used to manage the size of the libraries.

Tagged or encoded methods of synthesis allow the largest possible libraries to be constructed—from thousands to hundreds of thousands and more. Tagging usually involves the use of either fluorescent compounds or radiolabels that are attached directly to beads or the starting framework of the molecules themselves.

The use of tagging techniques “allows free movement and transfer of in-process materials without loss of identity . . . [and] is highly compatible with solid-phase methods, but clumsy, at best, with solution methods” (1).

Solution-based, large-library parallel synthesis requires significant automation and is extremely expensive, especially when compared with solid-phase methods that can be “conducted in a standard chemistry laboratory with minimal automation” (1).

An alternative way to think of the different sizes of combinatorial libraries is to consider their function rather than the approach used. According to Dolle (2), libraries for lead discovery are typically large in size (>5000 members) and deal with those having no preconceived biological information. Then there are the so-called targeted libraries, “which contain a pharmacophore known to interact with a specific (or family of) molecular target.” Finally, there are optimization (or lead development) libraries “where a lead exists and an attempt is being made to improve its potency, selectivity, pharmaceutical profile, etc.” Both of these latter cases involve much smaller libraries.

Lipinski Rules!
In the short term, the most promising hope for new drug leads through combinatorial chemistry is to base the reactions on prior knowledge of bioreactive compounds. One obvious method is to attempt to emulate one particular known biologically active structure. Another approach, which is more general and exploratory, uses molecular structures in a less constrained fashion and simply follows a series of guesstimates to determine whether or not a potential compound produced or building block used can be considered “druglike”.

The druglike concept is important for deciding the molecular scaffoldings and potential building blocks that might prove useful to produce a combinatorial library with a reasonable chance of providing some useful leads. Selecting among druglike criteria is also critical to choosing whether to continue examining a potential test candidate that such a library might produce. Several criteria are currently in vogue for deciding whether a new and unknown compound is druglike. These include potential for adsorption and permeability into cells, as well as the presence of known bioactive functional groups.

One set of criteria is, for example, the Lipinski “rule of five” (named because of its emphasis on numbers that are fives or multiples of five), which predicts that poor adsorption and permeability of potential drug candidates will occur if there are more than 5 H-bond donors (expressed as the sum of -OHs and -NHs); the molecular weight is over 500; the logP is over 5; or if there are more than 10 H-bond acceptors (expressed as the sum of nitrogens and oxygens). For each of these criteria, 80–90% of the actual drugs examined fall below the cutoff range (3).

A variety of architectural criteria are also currently considered. Mark Murcko and colleagues, for example, examined some 5120 distinct drug molecules and found that 32 particular geometric frameworks (using atoms as vertices and bonds as edges) formed some 50% of these drugs (4). Such frameworks make an obvious start or a potential goal for combinatorial synthesis.

The Bonds That Tie
When synthesizing libraries using a so-called “rational” or knowledge-based approach, it is also obviously important to use building blocks on the basis of whether they might contribute to the formation of key frameworks, such as the formats detailed above, and whether they contain biologically active functional groups in appropriately accessible forms. Because “the vast majority of marketed pharmaceuticals are low-molecular-weight nonpeptide, nonpolymeric entities, . . . it is logical that small organic molecules that can display functional groups would surface as a scaffolding approach” (5).

A wide variety of specific small-molecule scaffolds have been designed for differential introduction of functional groups. For example, an all-cis-substituted cyclopentane library was developed such that “by clever use of a cyclic anhydride, a methyl ester, and a Boc-protected amine it is a relatively straightforward task to sequentially introduce four different functional groups as desired” in each of the scaffold’s four active sites (5).

Natural product templates such as scaffolding are common. Tri-substituted purine libraries, flavone-derivatives, benzofurans and benzopyrans, steroids, taxoids (based on the valued anticancer agent taxol), and a wide variety of natural alkaloid frameworks, to name a very few, have all been used as starting points (6). The key is to start with bioactive models and modify with known reactive groups to try and get something that will have a physiological effect better than or antagonistic to known drugs or hormones.

One Good Turn
The synthesis of peptide turn mimetics is a perfect example of the use of solid-phase combinatorial techniques that rely on preexisting biological patterns as a starting point. Beta-turn mimetics are nonpeptide small molecules that mimic the shape of the reverse turn in folded peptides. They have been an important focus of medical research for years, and mimetics have been found that act as integrin antagonists and inhibitors of the human neutrophile receptor, among other physiological effects (Figure 2). Recently, Ellman managed to produce a turn mimetic library using solid-phase synthesis to incorporate a wide variety of side-chain functionalities. Figure 3 shows the synthetic pathway used (2).

Mimetics have also been produced to the -pleated sheets of proteins. These sheet motifs are natural recognition elements for a variety of medically important enzymes such as HIV protease. Mimetics are also being examined for their potential as inhibitors of proteolytic activity.

From such examples, it is obvious that collaboration on combinatorial chemistry between biologists and synthetic organic chemists is likely to yield new drugs. Still, far too many biologists working in the field of drug discovery are likely to envision the typical 96- to 384-well plates used in combinatorial synthesis as myriad tiny black boxes in which alchemical magic occurs, transforming vague biological guesses into medical answers. To many, combinatorial chemistry can seem a magical place where carefully discovered metabolic clues are input to receive an oracular output of druglike compounds ready for testing in human cell culture or animal models.

As long as combinatorial chemistry remains shrouded in mystery, staff chemists—or chemists at the CRO where the work was done—will be the only ones who know what really happens between metabolic lead and animal trials. Which is probably best for the efficiency (and peace of mind) of all concerned.

References

1. Baldino, C., Ed. J. Comb. Chem. 2000, 2 (2), 89–103.
2. Dolle, R. E. J. Comb. Chem. 2000, 2 (5), 383–433.
3. Lipinski, C. A.; Lombardo, F.; Dominy, B. W.; Feene, P. J. Adv. Drug Deliv. Rev. 1996, 23, 3–25.
4. Bernis, G. W.; Murcko, M. A. J. Med. Chem. 1996, 39, 2887–2893.
5. Gordon, E. M., Kerwin, J. F., Jr., Eds. Combinatorial Chemistry and Molecular Diversity in Drug Discovery; Wiley & Sons: New York, 1999.
6. Hall, D. G.; Manku, S. J. Comb. Chem. 2001, 3 (2), 125–150.