Solution DEL and Bead DEL: Two Tools for Advancing Drug Discovery

In 1992, the late professors Brenner and Lerner were the first to publish the concept of encoding synthetic chemistry with nucleic acid. Their approach involved synthesis of “one-bead, one-compound” (OBOC) libraries with concomitant encoding. A decade later, a different approach to encoding was pioneered by David Liu, Pehr Harbury and the Praecis/GSK team (some of whom are now X-Chem team members), in which libraries were synthesized in solution-phase by templated, routed or recorded synthesis approaches, respectively. What set these techniques apart from earlier approaches to encoding was that the libraries were designed to be screened as mixtures. Influenced by the concepts of molecular evolution and in analogy with the selection methods of phage and mRNA display or aptamers/SELEX, these researchers confirmed that affinity-mediated selection of mixtures was a powerful and efficient approach to screening.

These early discoveries established the basic hallmarks of what is commonly referred to as DNA-encoded library (DEL) technology: solution-phase, split-and-pool synthesis with encoding, affinity-mediated selection of library mixtures in a single vessel and DNA sequencing to characterize the bound fraction. In the ensuing 15 years, this approach has become well established, practiced by numerous companies, described in many peer-reviewed articles and responsible for an increasing number of clinical candidates.

Despite its many successes, the fact that DEL screening is a binding experiment can decrease its applicability in certain situations. Many drug discovery projects target specific modes of action, such as activation. Others desire a particular phenotypic response, regardless of mechanism. In the former case, at least DEL could give hope of finding a binder that has activating properties; in the latter case, it is hard to imagine how a mixture library could be screened to deliver such an outcome. These limitations have given rise to what has been variously described as “functional DEL” or “bead DEL.” Pioneered by Brian Paegel and his team, these methods adhere more closely to Brenner and Lerner’s original concept. OBOC libraries are made by encoding, with the beads then spatially segregated for assay. A big advancement in this area has been the advent of microfluidics, which allow the beads to be captured in water droplets containing assay reagents and then sorted based on parameters such as fluorescence.

We can compare the characteristics of “bead DEL” and “solution DEL” and see that they are quite different. One relies on synthesis of OBOC libraries, spatial separation and activity assay, while the other is characterized by solution-phase synthesis of libraries, mixture-based screening and affinity-mediated selection.

Let’s compare them in more detail (Table 1). First, in terms of library synthesis, each has advantages and disadvantages. Bead DEL is limited by the number of synthesis beads, while DEL display isn’t bound by this constraint and, as a result, has the advantage of almost unlimited library size. The largest bead DEL libraries reported to date contain fewer than 100,000 members. On the other hand, since the beads can be suspended in non-aqueous solvents, the scope of chemistry is likely to be wider. DNA stability is still required, however, so bead DEL libraries still cannot tap into the full range of organic chemistry conditions.

It is in screening methods that differences between the two approaches really emerge. Solution DEL methods are biology-driven. That is, high concentrations of protein (or other biological targets) are presented to the library mixture in such excess that competition for binding sites is avoided. Thus stimulated by biology, the readout is chemical in nature, in the form of chemical structures selected by the target. In solution DEL screening, the main driver for false positives and artifacts is the quality of the biological reagents used in the screen.

On the other hand, like high-throughput screening (HTS) and other assay systems, bead DEL is chemistry-driven. High concentrations of (photocleaved) compound in the microdroplet stimulate the assay system, which can be biochemical or cellular. The readout is an assay signal of some sort, usually fluorometric. It is possible that — as observed with combinatorial chemistry screening in the 1990s, as well as in the early days of HTS with non-optimized screening decks — performing assays with high concentrations of unpurified and uncharacterized compounds will lead to a propensity for assay artifacts.

Since a solution DEL selection is a binding incubation, its experimental setup is quite simple and involves very low volumes (usually less than 100 µL). This allows solution DEL screens to be multiplexed, enabling assessment of selectivity, binding site and affinity all in a single experiment. In contrast, bead DEL screening is limited by the sorting rates of the specialized microfluidic and photochemical apparatus required. Multiplexing the screens in the same way would require running the experiment as a series, greatly increasing time and expense.

The two techniques are similar after the screen. Both rely on DNA sequencing, although solution DEL methods generate very large sequence sets that require extensive informatic analysis to deconvolute. Both require a follow-up synthesis to unambiguously establish the structures of the actives and confirm them in an orthogonal assay format.

The various attributes described above inform the optimal use case for each technology. Solution DEL is an ideal tool for novel hit identification. Its ever-growing set of libraries comprehensively survey the accessible chemical space, yielding novel solutions to difficult binding challenges. Furthermore, solution DEL libraries are screened at such a small scale that a single library synthesis provides enough material for thousands of screening experiments. Bead DEL, with its more focused libraries that are destroyed by the screening experiment, is better suited for surveying specific SAR around a known active pharmacophore.

In summary, two synthetic chemical techniques that utilize DNA encoding are currently being practiced. One is rooted in bead-based combinatorial chemistry and biochemical screening. The other is based in molecular evolution, compound display and affinity-mediated selection. Both are powerful approaches to molecular screening, but solution DEL has been practiced longer and has a longer track record of success. We look forward to seeing how both technologies impact the discovery of drugs in the future.

Back to Blog