Conventional pharmacology blocks a site. Induced proximity does something stranger: it creates an interface that did not exist, holding two proteins together long enough for one to act on the other. For peptide chemistry this is an unusually good fit, because the hard part of the problem is gripping a large flat protein surface — exactly what peptides do and small molecules usually cannot. This page covers how these molecules are built, the counterintuitive dose-response they produce, and what has to be designed rather than simply chosen.
An inhibitor has to stay bound to keep working, so its effect tracks how much of the target is occupied at any moment. A degrader does not. Once it has held the target against an E3 ubiquitin ligase for long enough for ubiquitin to be transferred, the target is marked for the proteasome and the degrader is free to go and do it again. The consequence is that degraders act catalytically and can work at substoichiometric concentrations relative to the target1. The pharmacology is driven by events rather than occupancy, which is why potency and target affinity decouple in ways that surprise people coming from inhibitor work.
The targets that matter most here are the ones conventional chemistry cannot reach: transcription factors, scaffolding proteins, nucleic-acid-binding proteins with no pocket to occupy. Peptide-based degraders are attractive precisely because they avoid the requirement for a deep binding site, engaging large interacting surfaces instead, and they bring high specificity2. They also extend the range of what can be targeted, including mutant variants and proteins that lack any potent small-molecule ligand, and — unlike a synthetic small molecule — they can be genetically encoded, which opens routes a chemist alone does not have.
The history is instructive. The original PROTAC, reported in 2001, was peptide-based: one domain was a phosphopeptide from IκBα (the ten-residue motif DRHDSGLDSM, which β-TRCP recognises once both serines are phosphorylated), the other was ovalicin, and the construct tethered methionine aminopeptidase-2 to the SCF complex so that it was ubiquitinated and degraded3. It worked, and the approach was then largely set aside for small molecules because of poor cell permeability and chemical stability2. Those two problems are still the central ones, and most subsequent progress on the peptide side has been about solving them rather than about finding new targets.
The two constraints are permeability and stability, and both are addressable by peptide chemistry rather than by biology. Conformational constraint — stapling or cyclisation — can improve protease resistance and cell entry, though neither is guaranteed for an individual peptide, and stapled peptide degraders have been reported against targets including β-catenin, estrogen receptor α, MDM2/MDMX and protein kinase A4. The trade-off is worth stating plainly: a stapled peptide degrader is a far larger molecule than a small-molecule PROTAC, and degraders of this class are commonly limited by poor cellular permeability arising from high molecular weight and large exposed polar surface area7. Appending a cationic cell-penetrating sequence is the other standard route2.
A three-body system does not behave like a two-body one. The active species is the ternary complex — target, degrader, ligase, all on one assembly — and that requires both ends of the degrader to be engaged simultaneously. At low concentration there is not enough degrader to bridge anything. At high concentration there is too much: every target molecule has its own degrader and every ligase molecule has its own, and the productive complex is competed away by the two unproductive binary ones. The result is a bell-shaped dose-response, known as the hook effect1. It follows from the equilibrium; whether it is actually seen in a cellular experiment depends on the system and on the concentration range tested.
This has three practical consequences that are easy to get wrong.
Two degraders can have identical binary affinities and perform completely differently, because the proteins themselves may like or dislike each other once brought together. That effect is cooperativity, written α, and it is defined as the ratio of the binary to the ternary binding affinity: the ternary complex affinity is the binary affinity divided by α. Positive cooperativity (α > 1) arises when favourable protein–protein contacts form at the new interface, making the ternary complex more stable than the binary binding alone would predict; negative cooperativity (α < 1) means the two proteins resist being put together5.
In the curves above, raising α from 1 to 20 does two things at once: it raises the peak amount of ternary complex about fivefold, and it widens the concentration window over which useful amounts exist — in this calculation the range giving at least half-maximal ternary complex broadens from roughly 26-fold to roughly 80-fold. That second effect is why a strongly cooperative degrader may show no visible hook effect at all within a practical concentration range — positive cooperativity suppresses the unproductive binary complexes across a wider window1. Work correlating these parameters with cellular outcome has found that ternary complex affinity and cooperativity track degradation potency and the initial rate of degradation5.
The same logic explains the relationship between bifunctional degraders and molecular glues. A glue is a small molecule that may bind one partner weakly or barely at all on its own, and works almost entirely by remodelling the interface; mathematically, it behaves like a bifunctional ligand with very high cooperativity1. They are two ends of one continuum rather than two mechanisms.
Three choices are made rather than inherited, and each is a real variable.
For the conjugation chemistry used to join the two ends, see crosslinker and linker selection; for the constraint strategies used to make peptide binders durable enough to be useful, see cyclic and constrained peptides.
Because the readout is loss of a protein, the ways to be fooled are specific. A proper control set includes an inactive-ligase variant of the degrader — identical except for a point change that abolishes E3 binding — so that degradation can be attributed to the intended pathway. A proteasome inhibitor should rescue the effect; failure to rescue is informative but not conclusive on its own, since inhibitor concentration, timing, target engagement and inhibitor toxicity all bear on the result. The unconjugated peptide binder on its own distinguishes degradation from plain inhibition. And a concentration series is not optional, for the reasons above. Our page on control peptides in experimental design covers the general principles.
| Active species | The ternary complex, not either binary complex |
| Dose-response | Bell-shaped; falls at high concentration (hook effect) |
| Optimum | Near the geometric mean of the two binary KD values |
| Cooperativity α | Binary affinity divided by ternary affinity; >1 is favourable |
| Peptide strengths | Large flat interfaces, no pocket required, genetically encodable |
| Peptide weaknesses | Cell entry and intracellular stability |
LifeTein synthesises the peptide components these constructs need: stapled and cyclised binders, sequences carrying a defined conjugation handle at a chosen position, cell-penetrating sequences, and the inactive-variant controls that make the result interpretable. If you are working out which attachment vector to try, we can make the series rather than one compound at a time.