Cellular Binder Trap Enrichment: Screening DNA-Encoded Libraries Inside a Living Cell

The problem with screening in a test tube

Conventional DEL screening requires a highly purified, active protein of interest (POI) diluted into a buffer of water and salts. Two things go wrong here. First, many pharmaceutically important target classes — proteins that fold poorly outside their native context, that need a membrane environment, or that depend on cofactors and partner proteins present only in the cell — simply cannot be produced as purified, active protein. Second, even when purification is possible, a dilute buffer solution is not a faithful stand-in for a cell: the cytosol is packed with 20–30% (by weight) of macromolecules, and screening in their absence is a nonphysiological condition that generates assay-specific artifacts and contributes to attrition later in a drug discovery program.

The logical fix is to screen the library where the target actually lives — inside a cell. The obstacle has always been delivery: getting a multimillion-member DNA-tagged library across a cell membrane without destroying its diversity. Electroporation, liposomes, chemical transfection, and cell-penetrating peptides are all options for moving DNA into cells, but all are inefficient (under 5% for cell-penetrating peptides), and DEL material is too precious to lose that way.

The oocyte solves the delivery problem

We used oocytes from the South African clawed frog, Xenopus laevis. A Xenopus oocyte is roughly 1 µL in volume and about 1 mm in diameter — more than 100,000 times the volume of a typical somatic cell. That size means a library can simply be injected with a microneedle, rather than coaxed across a membrane by chemical or electrical trickery. Oocytes are also easy to obtain, robust outside the animal, and have been a workhorse expression system in biology since the 1970s, reliably translating injected mRNA into functional protein.

How cellular binder trap enrichment (cBTE) works

cBTE builds on binder trap enrichment (BTE), a DEL screening technology we developed earlier for purified protein, and adapts it to run inside the oocyte. The method has five stages:

1. Express the target inside the cell. A DNA construct encoding the POI fused to a small “prey” protein is transcribed in vitro to mRNA, which is microinjected into the oocyte. Four days later, the oocyte is expressing POI-prey fusion protein.
2. Inject the library and a labeling “bait.” The DEL and a bait-DNA — a small molecule with high, specific affinity for the prey protein, tethered to a short duplex DNA — are co-injected into the same oocyte. Inside the cell, library members bind the POI (if they are true binders) while the bait binds the prey, generating a DNA-tagged complex specifically on the target of interest.
3. Lyse and dilute. After a short binding period, the oocyte membrane is broken by centrifugation and the supernatant is transferred to a fresh tube, highly diluted into a buffer containing T4 DNA ligase.
4. Compartmentalize in an emulsion. Oil and surfactant are added and the mixture is emulsified by shaking, generating far more water droplets than POI molecules. A DEL member genuinely bound to the POI stays co-located with it — and with its prey and bait-DNA — in the same droplet. A DEL member that is not bound only ends up sharing a droplet with the POI by chance.
5. Ligate, amplify, sequence, count. Adding Mg²⁺ activates the ligase, joining library DNA to bait DNA wherever both are present in the same droplet — which is overwhelmingly true for real binders and rare for everything else. The ligated DNA is PCR amplified (using a priming site that only ligated species carry, so background is not co-amplified), sequenced, decoded back into compounds in silico, and counted. True binders show up as high-frequency events against a background of low-frequency random noise.

What discriminates target binders from everything else

The oocyte cytosol is full of endogenous protein that library members could bind nonspecifically. The prey/bait pairing is what lets cBTE tell target-specific binding apart from that background: only complexes that include the POI (via its prey fusion) can pick up a bait-DNA tag, so only true target engagement gets carried through to the ligation and sequencing steps. We describe the choice of prey protein and bait molecule in the companion note on labeling-system design.

What we found

We ran nine parallel experiments — three protein targets (p38α, ACSS2, DOCK5), each screened at three different bait-DNA concentrations, each in triplicate — against Lib051, a 194-million-member DEL built on 1,102 building blocks across three diversity positions (average molecular weight 516 Da, average cLogP 0.9). Across all nine experiments we recovered 323 structurally unique hits above a mathematical significance threshold: 154 specific for p38α, 20 for ACSS2, 76 for DOCK5, and a further 70 promiscuous or prey-binding hits common across targets. For each target, several distinct chemical clusters emerged rather than a single family — a sign of real, diverse target engagement rather than one artifact.

For p38α specifically, 21 representative hits were pulled forward into an orthogonal biochemical assay: 15 of 21 showed IC₅₀ values below 10 µM, and 9 of 21 below 1 µM, with the best compound, VPC00628, inhibiting p38α with a cellular IC₅₀ of 7 nM. We cover the validation data and how cBTE’s results compare to classical, purified-protein BTE in a dedicated note.