Bait-DNA and Prey-Fusion Design for Target Labeling Inside a Living Cell
Screening a DNA-encoded library (DEL) inside an intact cell creates a labeling problem that does not exist in a test tube: the cytosol is full of endogenous protein that library members can bind nonspecifically, and unlike a purified-protein assay there is no easy way to pull down “the target” and nothing else. Cellular binder trap enrichment (cBTE) solves this with a prey/bait pairing that tags only the protein of interest (POI) with DNA, so only genuine target engagement survives into sequencing.
Why a fusion tag, not a direct label
Directly chemically labeling a POI with DNA inside a living oocyte is impractical — there is no way to introduce a purified, site-specifically labeled protein into the cell without reintroducing the delivery problem cBTE exists to avoid. Instead, we express the POI as a genetic fusion to a small “prey” protein that has a known, specific, high-affinity small-molecule ligand. That ligand — the “bait” — is what actually carries the DNA tag. Because the bait only binds the prey, and the prey only exists in the cell as part of the POI-prey fusion, a bait-DNA molecule only becomes associated with library members that are also bound to the POI.
Choosing the prey: four criteria
We chose the catalytic domain of human carbonic anhydrase IX (CA IX) as the prey protein. It satisfies four practical requirements simultaneously:
| Requirement | Why it matters | How CA IX catalytic domain meets it |
|---|---|---|
| Expressible in the oocyte | The fusion has to fold and function after mRNA injection and translation | Well established as recombinantly expressible in Xenopus oocytes |
| Not already present in the cell | Any endogenous copy would compete with the fusion for bait binding, diluting signal | Xenopus oocytes have undetectably low native carbonic anhydrase activity |
| Tolerates fusion at either end | The POI’s own N- or C-terminus may not be free for fusion, depending on the target | The CA IX catalytic domain accommodates both N- and C-terminal fusion partners |
| Has a known, specific, high-affinity small-molecule ligand | The ligand becomes the bait; weak or promiscuous binders would tag off-target complexes too | CA IX inhibitor chemistry is well characterized, with potent, selective binders already validated as functional in Xenopus oocytes |
For the bait itself, we used a modified version of a previously reported CA IX-binding small molecule already shown to be active inside Xenopus oocytes — an important detail, since a bait that works in a purified-protein assay does not automatically work in a living cell.
Building the bait-DNA and library DNA
The bait molecule is installed via a chemical linker onto one strand of a 58/60-mer duplex DNA. Two design features matter for function inside a cell rather than a test tube:
- Phosphorothioate protection. Both the bait-DNA and the library DNA are protected against nuclease degradation in the oocyte by phosphorothioate linkages — unprotected DNA would simply be degraded before the binding and ligation steps can occur.
- A directional cohesive end. Both DNAs carry a cohesive (sticky) end designed so that bait-DNA and library DNA can only ligate to each other in the correct orientation during the later T4 DNA ligase step, rather than to themselves or to each other’s blunt ends.
Any DEL, regardless of its original synthesis platform, can in principle be adapted for cBTE by adding these two features — phosphorothioate protection and a compatible overhang.
Target constructs: three targets, one prey
We fused three different pharmaceutically relevant targets to the same CA IX prey (28.1 kDa):
| Target | Construct | Fusion size | Biology |
|---|---|---|---|
| p38α | Full-length protein | 69.8 kDa | MAP kinase; cell signaling, inflammation, cancer |
| ACSS2 | Full-length protein | 106.9 kDa | Acetyl-CoA synthetase; implicated in cancer, obesity, fatty liver disease, alcohol use disorder |
| DOCK5 | Dock Homology Region 2 domain only | 79.5 kDa | Guanine nucleotide exchange factor for Rho GTPases; cancer and immune disease target |
Note that DOCK5 was expressed as an isolated functional domain (DHR2) rather than the full-length protein — a reminder that cBTE, like any fusion-based labeling strategy, may require some construct engineering when a target is very large or has domains that don’t tolerate fusion or oocyte expression.
Confirming the fusion actually expresses
After mRNA microinjection, oocytes were incubated for four days and expression was checked by a prey-specific ELISA run on lysates from oocytes injected in parallel with the screening samples — a necessary quality-control step, since screening an oocyte that failed to express the fusion would simply waste library material. Apparent expression levels were 1.8 ± 0.2 µM for p38α-prey, 1.5 ± 0.1 µM for ACSS2-prey, and 0.5 ± 0.1 µM for DOCK5-prey — roughly a fourfold spread across three unrelated targets fused to the identical prey tag, illustrating that expression level is target-dependent even when the labeling chemistry is held constant.
Dosing the oocyte
Each oocyte received 50 nL of library (1.2 × 10¹² molecules total, roughly 6,000 copies per individual library member on average) alongside bait-DNA dosed at 30, 100, or 250 fmol — equivalent to final concentrations of approximately 30, 100, and 250 nM, assuming distribution through the oocyte’s roughly 1 µL volume. Running three bait-DNA concentrations in parallel, each in triplicate, let us assess how label dose affects data quality; we cover that comparison, along with the rest of the enrichment and hit-calling workflow, in a companion note.
Source: Petersen, L. K.; Blakskjær, P.; Christensen, A. B.; Andersen, J.; Folkesson, C. G.; Kristensen, O.; Andersen, C.; Alzu, A.; Sløk, F. A.; Madsen, D.; Azevedo, C.; Micco, I.; Hansen, N. J. V. “Screening of DNA-Encoded Small Molecule Libraries inside a Living Cell.” J. Am. Chem. Soc. 2021, 143, 2751–2756. DOI: 10.1021/jacs.0c09213