Targeted protein degradation as a powerful research tool in basic biology and drug target discovery

Abstract

Controlled perturbation of protein activity is essential to study protein function in cells and living organisms. Small molecules that hijack the cellular protein ubiquitination machinery to selectively degrade proteins of interest, so-called degraders, have recently emerged as alternatives to selective chemical inhibitors both as therapeutic modalities and as powerful research tools. These systems offer unprecedented temporal and spatial control over protein function. Here we review recent developments in this field with a particular focus on the use of degraders as research tools to interrogate complex biological problems.

In living cells, proteins are constantly synthesized and degraded. Two pathways control the bulk of protein degradation in mammals: the ubiquitin-proteasome system (UPS) and the autophagy-lysosome pathway^1,2. The UPS is mainly responsible for degrading misfolded and damaged proteins, and, more importantly, short-lived regulatory proteins that control many critical cellular processes, including cell cycle progression, cell proliferation and differentiation, cell signaling and transcription³. The autophagy-lysosome pathway is primarily a stress response mechanism that degrades proteins, but also larger cellular structures, such as organelles, under conditions of starvation and other types of stress².

Since the discovery that specific protein features, termed degrons, which often bind ubiquitin ligases⁴ or the proteasome directly⁵, render a protein susceptible to UPS-mediated degradation, researchers have sought ways to use degrons for controlled destabilization of proteins of interest (POI)⁶. Compared with genetic techniques, such as CRISPR-Cas9 or RNA interference methods, targeted protein degradation (TPD) has several advantages. First, TPD commonly depletes the target proteins fast, in minutes to hours, rather than the days or weeks required when manipulating mRNAs or genomic DNA. This prevents molecular compensation and cellular adaptation, which otherwise often convolutes phenotypic readouts. Second, while genetic manipulation generally affects all protein copies regardless of post-transcriptional RNA processing or posttranslational protein modifications, methods targeting proteins directly allow certain protein variants to be selectively degraded. Third, protein targeting methods are rapidly reversible and offer more fine-grained control. Incorporation of additional functional groups in TPD approaches allows protein degradation to be controlled by chemical or physical stimulation^7,8. For example, photo-activable degradation tags enable control of the degradation of fusion proteins through irradiation⁹.

TPD approaches modify proteins with degrons or adapt existing scaffolding or enzymatic components of protein degradation machineries to redirect their activity^10–17. More recently, powerful small molecule degrader systems have been developed that allow exquisite temporal control of protein degradation¹⁸. Small molecule-induced protein degradation has gained widespread attention not only because of its ability to precisely modulate target protein levels, but also because of the potential to develop new therapeutics that differ from classical inhibitors^19–23. In this context, target proteins are often referred to as neo-substrates, because they are recruited to the protein degradation machinery only when the small molecule degrader is present. Induced degradation of therapeutic targets offers the advantage of activity at sub-stoichiometric doses²⁴, and opens up therapeutic opportunities for classically “undruggable” targets, such as transcription factors and scaffold proteins²⁵. Overall, three common TPD approaches have emerged: small molecule induced degradation, genetically encoded degradation tags and the use of protein conjugates, all of which will be discussed in detail below.

Over the last two decades, developments in the field of protein degradation have converged from two independent directions. Efforts in the early 2000’s established the concept of using two distinct binding moieties, one for the neo-substrate and one for the ubiquitin ligase, bridged by a linker (Fig. 1a). This approach is commonly referred to as PROteolysis TArgeting Chimeras (PROTACs) or bifunctional degraders^17,25,26. Parallel studies uncovered that thalidomide and its analogs lenalidomide and pomalidomide (all FDA approved drugs for the treatment of hematologic malignancies) exert their therapeutic activity by induced degradation of neo-substrates, such as the zinc-finger transcription factors Ikaros (IKZF1) and Aiolos (IKZF3), the kinase Casein Kinase 1 alpha (CSNK1A1), and many more^27–33. Strikingly, these molecules operate by a mechanism similar to the plant hormone auxin, which was coined a ‘molecular glue’ (Fig. 1b)³⁴. This term commonly refers to molecules that display ligand affinity for only one of the two binding partners, either the neo-substrate or, more often, the ubiquitin ligase, and that enable engagement of the neosubstrate with the ubiquitin ligase^35–40. These two lines of research converged and now thalidomide analogs are commonly used as ubiquitin ligase recruiters for bifunctional degraders^41,42. Moreover, structural work (Fig. 1) has revealed that both bifunctional and molecular glue degraders can use extensive protein-protein contacts between ubiquitin ligase and neo-substrate, and hence the differences between the two classes may be more subtle than initially thought^{30,31,35–37,40,43–45}. In this review, we therefore collectively refer to molecules that trigger protein degradation as ‘degraders’. Only if necessary do we further distinguish between bifunctional or molecular glue degraders. While the therapeutic potential of degraders is exciting and widely discussed^19–22, we will focus here on recent advances in using TDP approaches as research tools.

Fig. 1.

Schematic representations of key concepts in protein degradation. a, Bifunctional degrader (PROTAC). Model of CRL2^VHL E3 ligase in complex with MZ-1 (PROTAC) and BRD4_BD2 (as a model POI) PDB: 5T35 and 5N4W (chain A and R). b, ‘Molecular glue’ degrader. Model of CRL4^DCAF15 E3 ligase in complex with indisulam and RBM39 (POI). PDB: 6Q0R (chain B, C and D), PDB: 4A0C (chain C and D) and PDB: 4A0K (chain C). c, dTAG system as an example for a tagging strategy. Model of CRL4^CRBN E3 ligase in complex with dTAG-13 recruiting FKBP12^F36V-BRD4_BD1 (POI). PDB models as in B, FKBP12 PDB: 1FKJ. Structure visualization created with the Illustrate program¹⁰¹

Small molecule degraders

Control of ubiquitin ligase activity by a low-molecular-weight ligand was first discovered in plants in the late 1990s^18,34,46,47. The plant hormone auxin was found to exert its activity by recruiting auxin responsive transcriptional repressors, the AUX/IAA proteins, to the Cul1-Rbx1-Skp1-Tir1 (SCF^Tir1) ubiquitin ligase for degradation. The structural basis of auxin recognition by the SCF^Tir1 ligase revealed that auxin binds to the substrate receptor Tir1 and contributes a small additional hydrophobic patch between Tir1 and the AUX/IAA substrates that is sufficient for AUX/IAA recruitment to SCF^{Tir1 34}. As mentioned above, the molecular mechanism of auxin was described as a ‘molecular glue’, a term now widely used for small molecule degraders that primarily bind the ligase.

Immunomodulatory drugs (IMiDs).

Thalidomide and its close derivatives pomalidomide and lenalidomide, collectively known as immunomodulatory drugs, are clinically used in the treatment of multiple myeloma (MM), 5q-deletion associated myelodysplastic syndrome (del (5q)-MDS), and leprosy^48–50. Thalidomide is infamous for its teratogenic effects when given to pregnant women for morning sickness in the 1950s, but second generation analogs such as lenalidomide and pomalidomide are now effective cancer treatments⁵¹. Re-approval of these drugs as anti-cancer agents was granted despite a lack of understanding of their mechanism of action. It was only subsequently discovered that IMiDs work by hijacking the ubiquitin ligase CRL4^CRBN in a ‘molecular glue’ mechanism akin to auxin. IMiDs target the transcription factors IKZF1/3, ZFP91, ZNF692, SALL4, the kinase CSNK1A, and others, for degradation by CRL4^{CRBN 27,29,30,32,40,52,53} (Table 1). Importantly, this retrospective discovery of the mechanism of action provided clinical proof of concept for protein degradation therapeutics and has fostered widespread interest in this therapeutic modality²⁰.

Table 1.

Neo-substrates, E3 ligases and ligands for the current molecular glue systems.

Molecular glue	Known neo-substrates	Regulatory stage
IMiDs (CRL4CRBN)
Thalidomide32,40,52,102	ZNF692, SALL4, RNF166, ZFP91, IKZF1, IKZF3	FDA approved for leprosy, multiple myeloma
Lenalidomide27,28,32,40	CSNK1A1 (CK1α), ZNF692, SALL4, RNF166, IKZF1, IKZF3, ZNF827	FDA approved for multiple myeloma, del5q myelodysplastic syndrome
Pomalidomide32,40	ZNF692, SALL4, RNF166, IKZF1, IKZF3, ZNF827, ZFP91, ZBTB39, GZF1, ZNF98, ZNF276, FAM83F, WIZ, RAB28, DTWD1, ZNF692, ZNF653	FDA approved for multiple myeloma
Avadomide (CC-122)103	IKZF1, IKZF3	In clinical trials: Phase 1/2
Iberdomide (CC-220)32,104	IKZF1, IKZF3, ZFP91, RNF166, SALL4, RAB28, ZNF653, ZNF98	In clinical trials: Phase 1
CC-88531	GSPT1, GSPT2, IKZF1, IKZF2, (possibly others as translational effects convolute proteomics)	Preclinical
FPFT-2216105	CSNK1A1, IKZF1	Preclinical
Aryl-sulfonamides (CRL4DCAF15)
Indisulam(E7070)38,57–59	RBM39, RBM23, PRPF39	In clinical trials: Phase 2 – terminated
E782038,57	RBM39, RBM23	In clinical trials: Phase 2 - completed
Tasisulam38,58	RBM39	In clinical trials: Phase 1/2 - terminated
Chloroquinoxaline sulfonamide (CQS)38,58	RBM39	In clinical trials: Phase 2 - completed

While the therapeutic potential of CRL4^CRBN-binding degraders has been widely discussed^19,20, thalidomide analogs also serve as research tools. In contrast to most bifunctional degraders that are at the chemical porbe stage, the physicochemical properties and oral bioavailability of IMiDs make them great tools for in vivo studies^54,55. Initially, the use of IMiDs in rodents was hindered by polymorphisms in CRBN that prevented target degradation. Specifically, murine Crbn can bind IMiDs, but Ile391, a residue at the putative Crbn–neo-substrate interface, causes steric hindrance that prevents recruitment of the neo-substrates⁵⁶. However, a recently developed mouse model harboring a CRBN^I391V humanizing mutation has emerged as a useful tool to assess the effects of IMiD-induced protein degradation in genetically engineered or xenografted mouse tumor models⁵⁶.

Aryl-sulfonamides.

Reminiscent of IMiDs, aryl-sulfonamides, including indisulam, E7820, tasisulam and chloroquinoxaline sulfonamide (CQS), were shown to induce degradation of the splicing factor RBM39 via recruitment of the ubiquitin ligase CUL4-RBX1-DDB1-DCAF15 (CRL4^DCAF15) (Table 1)^38,57–59. We and others have recently established the molecular basis of aryl-sulfonamide activity^35–37 and demonstrated that E7820 and its analogs bind to a relatively shallow pocket on DCAF15, facilitating the recruitment of RBM39 and RBM23 through the second RNA recognition motif (RRM) present in RBM39 or RBM23. Aryl-sulfonamides act as interface stabilizers that preferentially bind to the complex of the ligase and neo-substrate, in contrast to IMiDs that tightly interact with CRBN even in the absence of neo-substrates⁶⁰. While the use of aryl-sulfonamides as effective research tools remains to be established, they expand our knowledge of how small molecules promote substrate recruitment and could encourage the development of novel molecular glue compounds as well as tagging strategies based on the RBM39 RRM domain.

Bifunctional degraders (PROTACs).

Fueled by discoveries of small-molecule ligands for ubiquitin E3 ligases, the field of bifunctional degraders (PROTACs)¹⁷ is rapidly evolving^24,26. Combining high-affinity ligands for ubiquitin ligases such as CRL4^CRBN, CUL2-RBX1-ElonginB/C-VHL (CRL2^VHL)⁶¹, IAPs⁶², and MDM2⁶³ or covalent binders of the ubiquitin ligases DCAF16⁶⁴, RNF144⁶⁵ and RNF4⁶⁶ with ligands targeting a variety of POI has yielded degraders with exquisite efficacy and selectivity. The ever expanding space of proteins that have been successfully targeted with bifunctional degraders includes kinases^67–70, nuclear receptors^24,71, and epigenetic regulators^72,73.

Many bifunctional degraders promote extensive protein-protein interactions between the neo-substrate and the ligase. These often provide an additional layer of substrate selectivity^41–44 beyond or distinct from the target binding ligand alone. For example, achieving CDK4/6 homolog selectivity with active site inhibitors has been challenging due to the highly conserved CDK active site. By adapting the CDK4/6 dual inhibitor palbociclib as a degrader, Brand et al. were able to achieve excellent CDK6 selectivity and subsequently used the degrader to parse the consequences of acute CDK6 depletion on downstream signaling networks and transcriptional programs in AML^74,75. Another application for bifunctional degraders is as probes for drug target identification (ID) and validation. For example, a bifunctional degrader was used to verify the transcription factor pirin as the target of a new chemical probe. The probe was discovered in an unbiased phenotypic screen designed to identify inhibitors of the HSF1 stress pathway⁷⁶. Attaching a thalidomide moiety to this compound generated a degrader that specifically depleted pirin in cells, thus providing evidence for intracellular engagement of pirin. While this approach adds to the chemical biology toolbox commonly used for drug target ID, it is important to remember that false positive and false negative results are common for any target-ID method, including degraders.

Adding another layer of spatiotemporal control, optical-controlled degraders were recently described. Two independent groups reported photo-cleavable caging groups that can mask the degrader compounds. Upon radiation, such cages break down and release the active degrader, leading to degradation of neo-substrates ^77–79. Photo-switchable bifunctional degraders were also designed and synthesized. Photo-PROTACs that contain an ortho-F4-azobenzene linker can switch between the cis- and trans-configurations upon visible light radiation, thus keeping the degrader in an active and inactive photostationary state, respectively. These probes allow very precise spatiotemporal control of protein degradation within the limits of intracellular diffusion of the activated degrader molecules^7,8,80.

The discovery of ligands that covalently bind E3 ligases, such as DCAF16, RNF114 and RNF4^64–66 offer exciting new opportunities to access new protein target space. Proof of concept has been established by synthesizing bifunctional molecules using covalent E3 ligands and established neo-substrate ligands to degrade targets such as BRD4 and FKBP^64,66. It remains to be shown, however, if covalent ligands will exhibit similar broad applicability as observed for CRBN and VHL ligands. Interestingly, covalent E3 binders can be discovered from fragment-based libraries^64,66, but can also be found in the vast realm of natural products. One such example is nimbolide, which recruits the E3 ligase RNF114⁶⁵.

One of the most exciting byproducts of the widespread therapeutic interest in degraders is the number of selective and well-characterized chemical probes that become available for biological research. While it is impressive how quickly the field has leveraged the technology, the development of bifunctional degraders will become increasingly difficult once the existing ligand space is exhausted and new ligands need to be discovered. In addition, while it is possible for some substrates to develop highly effective degraders in a few structure-activity relationship iterations, other potential substrates are refractory to degradation⁸¹. Substantial efforts are often required to obtain probes and ultimately therapeutics for some of the most promising therapeutic targets. Therefore, when available, target-specific degraders are ideal research tools, but one has to carefully consider whether development of a new degrader to answer a biological question is feasible or whether a degradation inducing fusion tag would provide a more adequate alternative.

Genetic fusion of degradation tags

While the degraders discussed above control endogenous protein levels, their development requires substantial synthetic efforts that are not guaranteed to succeed for every POI. Degradation tags offer an alternative and faster way to obtain precise control over protein abundance using established ligands (Fig. 1c). Here, we will cover a representative set of genetic tagging strategies as well as some general considerations, such as the size of the tag, selectivity and potential off-target effects. We will further discuss the feasibility of genome editing to achieve endogenous tagging and in vivo properties of ligands, which are relevant when choosing a particular approach for drug target validation and for studying biological pathways.

Auxin-inducible degron (AID).

As described earlier, the plant hormone auxin can bind to the ubiquitin ligase SCF^TIR1 and trigger the degradation of auxin responsive transcriptional repressors³⁴. The degron sequence of these repressors (Tag size: 7 kDa) can be fused to a POI, and in the presence of exogenous plant E3 ligase SCF^TIR1, the fused target protein is rapidly degraded upon auxin treatment¹³ (Fig. 2). An advantage of this system is that the plant hormone auxin has no known mammalian targets and is therefore assumed to have no off-target effects in mammalian cells. A major limitation of the system is the requirement for exogenous expression of the plant E3 ligase adapter TIR1, which complicates experimental design. In addition, TIR1 might have unknown activity in non-plant cells. Genetic knock-in of the AID sequence into a target locus to generate an endogenously tagged target protein has been reported, suggesting that this strategy is amenable to gene editing approaches⁸² (Table 2).

Fig. 2.

An overview of degradation tag fusion strategies for targeted protein degradation. Please refer to descriptions in the main text. POI based on BRD4 as an example, PDB: 6BOY; Auxin-inducible Degron based on TIR/IAA structure, PDB: 2P1Q (chain C), TIR structure, PDB: 2P1Q (chain B); Destabilzing Domain FKBP12 structure, PDB: 1FKJ; dTAG CRBN structure, PDB: 6BOY; IKZF3 tag based on structure of IKZF1 ZF2, PDB: 6H0F (chain C); HaloTag based apo HaloTag structure, PDB: 5UY1; VHL structure, PDB: 5T35 (chain D); NS3 protease structure in SMASh, PDB: 4WF8; NS4A degron based structure of NS4A peptide (shorter than the length of degron), PDB: 2OBO (chainB); Structure visualizations created with theIllustrate program¹⁰¹.

Table 2.

Genetic fusion strategies of degradation tags for targeted protein degradation.

System	Tag	Tag size (kDa)	Drug	Molecular machinery for degradation	Note	Applications
AID	Auxin-inducible degron	7	Indole-3-acetic acid (IAA)	CRL1TIR1	• CRISPR-Cas9 knock-In applicable • Exogenous expression of TIR1 is required	13,82,106–108
Destabilizing domain	FKBP12L106P mutant	12	Shld1	Unidentified quality control pathway	• Continuous treatment of Shld1 ligand is required to keep the POI from degradation	12,83–85
dTAG	FKBP12F36V	12	dTAG13	CRL4CRBN	• CRISPR-Cas9 knock-In applicable • Available in vivo	15,86,87,109,110
IKZF3-peptidic Degron	IKZF3 degron	3	Pomalidomide	CRL4CRBN	• CRISPR-Cas9 knock-In applicable • Available in vivo	14
HaloPROTACs	HaloTag	34	HaloPROTAC-3	CRL2VHL/cIAP	• CRISPR-Cas9 knock-In applicable • Available in vivo	16,91–93
Hydrophobic	HaloTag	34	HyT13	Unidentified quality control pathway	• Potential cellular perturbation from artificially unfolded protein	88,89
SMASh	NS3-NS4A	34	Asunaprevir	Unidentified quality control pathway	• No structural modification from tagging • Existing proteins cannot be degraded	11,111–114

The Destabilizing domain.

This approach uses an engineered FKBP12^L106P mutant protein as a fusion tag¹² (Tag size: 12 kDa). The tag induces rapid and constitutive degradation of the fusion protein when expressed in mammalian cells, unless the cells are treated with a synthetic ligand (Shld1) that binds to the destabilizing domain and shields the fusion protein from degradation (Fig. 2). While this approach allows the reversible and tunable control of target protein stability, it requires continuous treatment with the Shld1 ligand to maintain the desired POI level. Unlike other degradation tag strategies, this system has the unique feature that it stabilizes the POI and turns on its downstream signaling upon ligand addition. This method has been used in many studies, including the investigation of lipophosphoglycan implicated in parasite survival⁸³, P. falciparum protease falcipain-2⁸⁴, and PTEN⁸⁵ (Table 2).

dTAG tag.

This tag (Tag size: 12 kDa)¹⁵ combines fusion of the targeted POI to mutant FKBP12^F36V with the use of a degrader (dTAG13). The latter recruits the FKBP12^F36V-fusion protein to CRL4^CRBN and induces its degradation (Fig. 1c and Fig. 2). dTAG13 only binds the mutated FKBP12^F36V and not the wild-type protein, thus minimizing off-target effects. Coupled with CRISPR-mediated knock-in, this system has been applied in multiple studies to deconvolute biological functions of POIs, as well as in drug target validation^86,87. dTAG13 has been shown to have acceptable in vivo properties and in most cases triggers rapid protein degradation¹⁵. Empirically, dTAG appears to be a robust degradation tag, perhaps because the FKBP protein provides ubiquitination sites itself. Major limitations of this system are the rather large size for endogenous tagging (Table 2).

IKZF3-peptidic degron.

Another recent example of a degradation tag is the IKZF3-peptidic degron for IMiD-induced degradation¹⁴. Mapping of the boundaries in the zinc finger-2 domain of IKZF3 identified a 25 amino acid degron (Tag size: 3 kDa), which can be fused to a POI for IMiD-induced degradation (Fig. 2). Although it represents a single zinc-finger domain and not a linear peptide motif, this is still the smallest degron among all reported tags. The small size makes it less likely to interfere with the native function of target proteins and facilitates efficient knock-in. Another advantage is the good bioavailability of IMiDs, which cross the blood brain barrier and are active in the central nervous system (CNS), representing the only ligand-induced system suitable for modulation of proteins in the brain. An important consideration for this degron is the species-dependent sensitivity to IMiD-induced protein degradation discussed above. Many species, such as mice, harbor sequence variations in CRBN rendering them intrinsically resistant to IMiD-mediated recruitment of neo-substrates. The degron would thus be ineffective in wild-type animals, unless mouse CRBN is humanized to CRBN^I391V as in reported mouse models⁵⁶. To date, this degron has been reported to induce degradation for only ~ 50% of proteins that have been tagged, and further optimization may be required for it to be more widely applicable¹⁴ (Table 2). In addition, one has to take into account off-target effects: IMiDs induce the degradation of several transcription factors^27,28, potentially confounding the interpretation of experimental results.

Halo Tag.

The HaloTag (Tag size: 34 kDa) is a modified haloalkane dehalogenase that can be specifically and covalently labeled with chloroalkane ligands. A HaloTag fused to a POI and labeled with a covalent bifunctional chemical probe consisting of such a chloroalkane ligand and a hydrophobic adamantyl group creates a hydrophobic degron (Fig. 2). This degron targets the fusion protein for degradation through the unfolded protein response in cell lines, zebrafish and mice^88,89, but the precise mechanism of degradation triggered by the adamantyl group is not fully understood (Table 2). The HaloTag is widely used and the chemistry is thought to be bio-orthogonal (i.e. not exerting any off-target activity) in mammalian cells⁹⁰. HaloTag fusion proteins can also be degraded using HaloPROTACs, covalent bifunctional molecules that react with the HaloTag and recruit VHL¹⁶ or cIAP⁹¹ ligases (Fig. 2). Combined with CRISPR/Cas9 genome editing technology, this method has been used to degrade endogenous HaloTag fusion proteins⁹² and target proteins in vivo⁹³ (Table 2).

SMASh (Small Molecule Assisted Shutoff).

This system uses a destabilizing degron that is coupled to the NS3 protease from hepatitis C virus and a NS3 cleavage site. This cassette can be fused to a POI (Tag size: 34 kDa) (Fig. 2)¹¹. After translation, the NS3 protease cleaves the recognition site between the fused cassette and the POI, generating an intact, native target protein. Administering the selective NS3 protease inhibitor, asunaprevir, blocks the cleavage event and leads to retention of the degron on the fusion protein and ultimately degradation. A main advantage of this method is that it produces a native protein without structural modifications. However, degradation is limited to newly synthesized proteins after asunaprevir addition, therefore the kinetics of degradation follow the rate of protein synthesis and are often slow. As a consequence, this tag may not be suitable for studying a biological process with fast dynamics (Table 2).

In addition to tagging substrate proteins, modifying or tagging the E3 ligases with a specific protein-recognizing entity, such as nanobodies, creates a degradation system that targets an endogenous POI. Nanobodies recognizing GFP or other target proteins were successfully fused to an F-box motif creating a functional Cul1-Rbx1-Skp1-nanobody ligase to degrade GFP fusion proteins¹⁰. In another study, nanobodies were fused to SPOP, an adaptor protein of the CUL3 E3 ubiquitin ligase, to degrade specific nuclear proteins in mammalian cells and zebrafish embryos⁹⁴. Because of the general adaptability of nanobodies, this approach can potentially be used to target any endogenous protein, including specific post-translationally modified states or splice variants.

Macromolecular conjugates

Trim-away.

The Trim-away approach takes advantage of the fact that the ubiquitin ligase TRIM21 binds tightly to the Fc domain of antibodies and ubiquitinates itself, which results in degradation of the antibody-POI complex together with TRIM21⁹⁵. The Trim-away method has been shown to degrade endogenous proteins for which specific antibodies are available (Fig. 3a). A major limitation is that antibodies are not cell penetrant. This issue can be overcome through microinjection of individual cells or electroporation of large cell populations with the antibody and the TRIM21 protein, resulting in rapid target protein degradation with protein half lives in the range of minutes to hours. In cell lines with high levels of TRIM21, injection or electroporation of the antibody alone is sufficient for target protein degradation⁹⁵. An advantage of this method is its application in primary or non-dividing cells that are not amenable to genetic manipulation. Recently, this method was also used to degrade proteins in zebrafish embryos, acting faster than morpholinos⁹⁶. Limitations of this method include the potential degradation of binding partners of the POI, observed for certain targeted proteins⁹⁵. Further, as TRIM21 and antibodies are also degraded alongside the target protein, the method is prone to rapid recovery of POI levels⁹⁵.

Fig. 3.

Schematic representations of targeted protein degradation strategies mediated by macromolecular conjugates. a, Trim-away. b, Lysosome targeting chimera. c, Peptide-directed lysosomal degradation.

Lysosome targeting chimeras (LYTACs).

While most TPD methods utilize the UPS to degrade a POI, a recently developed method allows degradation of secreted and membrane proteins through the lysosome pathway (Fig. 3b). LYTACs are macromolecular conjugates that are able to simultaneously bind a cell surface lysosome targeting receptor and the extracellular domain of a target protein. They consist of an antibody fused to agonist glycopeptide ligands for the cation-independent mannose-6-phosphate receptor (CI-M6PR)⁹⁷. These LYTACs can induce degradation of proteins in a variety of cell lines. However, protein degradation using LYTACs generally requires days, compared to minutes or hours for methods directly engaging the UPS. Another potential limitation is the off-target degradation of interacting partners of the targeted membrane proteins.

Peptide-directed lysosomal degradation.

Another approach to degrade endogenous proteins uses a peptide fusion of a chaperone-mediated autophagy (CMA)-targeting motif (CTM) with a cell membrane-penetrating domain (CMPD) and a protein binding domain (PBD), to target a protein that is specifically recognized by the PBD⁹⁸ (Fig. 3c). By simply adding the peptide into the cell culture medium, this method was effective in targeting a number of proteins in established cell lines as well as in primary neuronal cultures. Inherent limitations of this method include the inability to target proteins involved in the CMA pathway and lysosomal integrity⁹⁸ and the limited stability of the peptide in cells and in animals. Recently, this method was used to target CDK5 to study the potential therapeutic effects of degrading CDK5 in prevention of strokes⁹⁹.

Validation methods

The cellular activity of degraders relies on multiple factors, such as cellular uptake, efficient recruitment of the E3 ligase, ubiquitination of the target protein, and finally recruitment to the proteasome for degradation. As with any complex process, validation of expected results at each step is paramount to success. While not exhaustive, the methods outlined in Table 3 and Fig. 4 together with the use of established UPS inhibitors represent some of the available techniques for quality control at the various stages of the protein degradation pathway and are broadly applicable to all protein degradation modalities. In our experience the following assays are critical to troubleshoot protein degradation systems: 1) testing ligase or target protein engagement in cells, which also provides a qualitative readout of cell permeability, 2) in vitro ubiquitination assays, 3) target protein degradation in cells using a quantitative assay, such as ratiometric GFP/RFP⁴⁰ or luminescence assays, and 4) assessment of the selectivity and efficacy for the targeted protein using mass spectrometry-based proteomics, which is a powerful profiling method for small molecule degraders.

Table 3.

Validation methods for targeted protein degradation.

Steps	Purpose	Example Assays
Dimerization – in vitro	Ternary complex formation in vitro	TR-FRET40,43, SPR/BLI115
Dimerization – cellular	Ternary complex formation in cells	NanoBit116 NanoBret116
Cellular engagement	Cellular permeability and cellular engagement of the E3 ligase	Degradation based engagement assay43 NanoBret based Fluoro-ligand displacement117
Ubiquitination – in vitro	Verify ubiquitination	In-vitro ubiquitination assay
Ubiquitination – cellular	Ubiquitination quantification and type	Western blot, NanoBret, TUBES118
	Identify ubiquitination sites	Proteomics
Cellular degradation – targeted approach	Verify degradation in cells	Western blot, GFP-fusion, mCherry reporter lines119 Endogenous CRISPR fusions (HiBit Tag116, Split GFP120)
Cellular degradation – selectivity profiling by proteomics	Verify degradation selectivity in cells	Proteomics approaches32,52, library based screens40
UPS inhibitors	Inhibitors for Cullin-RING family of ligases121,122	MLN4924 - a specific inhibitor of the NAE1/UBA3 Nedd8 activating enzyme CSN5i-3 – inhibitor of COP9 signalosome
	p97 inhibitor123	CB-5083
	Ubiquitin E1 (UBA1) inhibitor	MLN7243
	Proteasome inhibitors124	Bortezomib, Carfilzomib and MG132

Fig. 4.

Validation methods for targeted protein degradation at different stages of the ubiquitin-proteasome pathway. P97 PDB: 5FTJ; model of CRL4^CRBN E3 ligase apo structure and structure with BRD4 created with PDB: 4A0C (chain C and D), PDB: 4A0K (chain C) and 6BOY (chain B and C); Human 26S proteasome PDB: 5GJR;Ubiquitin PDB: 1UBQ;POI BRD4 PDB: 6BOY (chain C). Structure visualizations created with the Illustrate program¹⁰¹.

Concluding remarks

Protein homeostasis is an important field of study for understanding human physiology and disease, as evidenced by an increasing number of small molecule therapeutics targeting this process. Key contributions to our mechanistic understanding of protein modification and degradation make it possible to engineer these systems to control the turnover of a POI. Most current TPD methods use the UPS, but several approaches using large biomolecular conjugates target the autophagy/lysosome pathway. R ecently reported small molecule degraders based on the autophagy pathway (autophagy-targeting chimeras or AUTACs)¹⁰⁰ offer a glimpse into future possibilities. With this review focusing on the impact of TPD methods as research tools, we hope to encourage and inspire more efforts in creating new TPD methods and to provide readers with a resource to help navigate this expanding toolbox.