Assessing the Suitability of Deubiquitylases As Substrates For Targeted Protein Degradation

This study employs a chemical genetics system to demonstrate that while some deubiquitylases (DUBs) are susceptible to targeted degradation via PROTACs or molecular glues, others resist turnover either through auto-deubiquitylation or by being inherently poor proteasome substrates.

The Gist

The Big Problem: The "Look-Alike" Crowd

Imagine the human body is a massive city with about 100 different security guards called DUBs (Deubiquitylases). Their job is to act as "erasers." In our cells, there is a tagging system (Ubiquitin) that marks broken or unwanted proteins for the trash can (the proteasome). The DUBs are the guards who can wipe that tag off, saving the protein from being thrown away.

Sometimes, these guards go rogue and save the wrong proteins (like cancer-causing ones). We want to get rid of these rogue guards to treat diseases like cancer or Alzheimer's.

The Catch: All 100 of these guards look almost identical in their "hands" (their active sites). If you try to make a drug to stop one specific guard by grabbing their hand, you accidentally grab the hands of 99 other innocent guards. This is why making specific DUB drugs is so hard.

The New Idea: The "Molecular Glue" Trap

Instead of trying to stop the guard's hand, the scientists thought: What if we just grab the guard by their shoulder (a part of the body that looks different for everyone) and drag them straight to the trash can?

This is called Targeted Protein Degradation (using a tool called a PROTAC). You attach a "molecular glue" to the guard's shoulder. This glue has two ends: one grabs the guard, and the other grabs the trash can. The guard gets pulled into the trash and destroyed.

The Big Fear: There was a major worry. Since these guards are "erasers," what if they see the trash tag being put on them and just erase it themselves? If they can erase their own "trash tag," the trap won't work. The scientists wanted to know: Which guards are too smart to be caught this way, and which ones are easy targets?

The Experiment: The "Fake Guard" Test

Since they didn't have a perfect "shoulder-grabbing" drug yet, they built a test system. They took the guards and glued a special "handle" (called FKBP*) onto them. This handle is like a universal keyhole. They then used a special key (dTAG-13) that fits the handle and instantly drags the guard to the trash can.

They tested four different guards: USP11, USP4, USP15, and UCHL1.

Here is what they found, broken down into three groups:

1. The "Easy Targets" (USP11)

The Story: USP11 is like a guard who is very good at his job but doesn't notice when someone puts a "trash tag" on his back.
The Result: As soon as the scientists used the key to drag him, he vanished quickly. Even though he was still active (still erasing tags on other things), he couldn't stop his own destruction.
The Takeaway: USP11 is a great candidate for this new type of drug. We can likely build a "shoulder-grabber" that will successfully destroy it.

2. The "Self-Rescuers" (USP4 and USP15)

The Story: These two guards are very clever. As soon as the trash tag is put on them, they immediately use their "eraser" hands to wipe the tag off. They are essentially erasing their own execution order.
The Result: When the scientists tried to drag them to the trash, the guards kept fighting back and surviving.
The Twist: When the scientists made a version of these guards with "broken hands" (so they couldn't erase anything), the trash can dragged them away easily.
The Takeaway: These guards are hard to kill with this method unless we can find a way to stop their erasing ability first. However, since they are so similar to USP11, maybe we can find a drug that grabs all three, but because USP11 is "dumber" (can't erase itself), it gets destroyed while the others survive. This could be a way to get selective.

3. The "Stuck Box" (UCHL1)

The Story: This guard isn't smart enough to erase the tag, but he is also very tightly packed. Imagine a guard who is a solid, knotted ball of yarn with no loose ends. The trash can (proteasome) needs a loose string to grab onto to pull the guard inside.
The Result: The trash can tried to pull him, but couldn't get a grip. He sat there, stubbornly refusing to go.
The Fix: The scientists added a long, loose string (a disordered tail) to the back of this guard. Suddenly, the trash can could grab the string and pull him in easily.
The Takeaway: UCHL1 isn't fighting back; it's just physically too compact to be eaten by the trash can. To kill this one, we might need to attach a "handle" that gives the trash can something to grab.

Why Does This Matter?

This paper is like a safety checklist for drug developers.

Before spending millions of dollars trying to invent a drug to grab a specific guard's shoulder, you can run this test.

• If the guard is like USP11, go ahead! It's likely to work.
• If the guard is like USP4, be careful. It might erase the drug's signal.
• If the guard is like UCHL1, you might need to change your design to give the trash can a better grip.

In short: The scientists built a "simulation" to see which of these tricky enzymes can be successfully dragged to the trash can. They found that some are easy targets, some are too clever, and some are just too stiff. This helps scientists know which targets are worth pursuing for new, life-saving medicines.

Technical Summary

1. Problem Statement

Deubiquitylases (DUBs) are a family of ~100 enzymes in the human proteome that regulate protein stability by cleaving ubiquitin chains. Many DUBs are attractive drug targets for cancer and neurodegenerative diseases. However, developing selective small-molecule inhibitors is extremely difficult due to the high degree of structural homology in their catalytic active sites.

Targeted Protein Degradation (TPD) using PROTACs (Proteolysis-Targeting Chimeras) offers an alternative strategy by recruiting an E3 ligase to a less conserved region of the target protein, avoiding the active site. However, a critical "de-risking" question remains: Will DUBs be effective substrates for PROTACs?

• The Concern: Since DUBs remove ubiquitin chains, a PROTAC-induced ubiquitination event might be reversed by the DUB's own catalytic activity (auto-deubiquitylation), preventing proteasomal degradation.
• The Gap: There are virtually no drug-like ligands that bind DUBs outside their active sites, making it impossible to directly test this hypothesis with native proteins and specific PROTACs.

2. Methodology

To address this gap, the authors developed a chemical genetics workflow using the dTAG system to evaluate DUB degradability without inhibiting their catalytic activity.

• System Design: They fused a mutant FKBP tag (FKBP12F36V, referred to as FKBP*) to the C-terminus of specific DUBs (USP11, USP4, USP15, and UCHL1) and a control (HDAC1). These fusion proteins were stably expressed in Flp-In 293 cells.
• Degradation Mechanism: Treatment with dTAG-13 induces a ternary complex between FKBP*, the target DUB, and the E3 ligase Cereblon (CRBN). This recruits the ubiquitin-proteasome system to the DUB. Crucially, because dTAG-13 binds the FKBP* tag and not the DUB active site, the DUB remains catalytically active during the assay.
• Experimental Groups:
  1. Wild-type (WT) DUBs: To assess native degradability.
  2. Catalytically Dead Mutants: Cysteine-to-Serine/Alanine mutations (e.g., USP4-C311S) to test if auto-deubiquitylation is the barrier.
  3. Structural Modifications: Adding a disordered tail to UCHL1 to test proteasome engagement efficiency.
  4. Inhibitor Co-treatment: Using inhibitors for E1 (TAK-243), Proteasome (Bortezomib), and p97 (CB-5083) to confirm degradation mechanisms.
• Validation:
  • Functional Assay: Tested if FKBP*-USP11 could rescue DNA double-strand break (DSB) repair (measured by $\gamma$-H2AX foci) after endogenous USP11 knockdown.
  • Interactome Analysis: Mass spectrometry to verify FKBP*-USP11 retains native protein-protein interactions.

3. Key Results

The study categorized DUBs into three distinct groups based on their suitability for PROTAC-mediated degradation:

Category 1: Highly Degradable (e.g., USP11)

• Observation: FKBP*-USP11 was rapidly and efficiently degraded by dTAG-13 (reaching <10% of control levels within 4 hours).
• Validation: The fusion protein retained full catalytic function, rescuing DSB repair defects and maintaining native interaction partners (USP7, EIF4B, PRKDC).
• Conclusion: USP11 is a prime candidate for PROTAC development; its auto-deubiquitylation activity does not prevent degradation.

Category 2: Resistant via Auto-deubiquitylation (e.g., USP4, USP15)

• Observation: WT FKBP*-USP4 and FKBP*-USP15 showed only modest degradation (remaining ~60-65% of control levels).
• Mechanism: When the catalytic cysteine was mutated (inactivating the DUB), degradation efficiency increased dramatically (mutants dropped to <20% levels).
• Conclusion: These DUBs actively resist degradation by removing the ubiquitin chains added by the PROTAC/E3 complex. Their catalytic activity is the primary barrier.

Category 3: Poor Proteasome Substrates (e.g., UCHL1)

• Observation: WT and catalytically dead FKBP*-UCHL1 were both degraded poorly and slowly.
• Mechanism: UCHL1 is a tightly folded protein lacking an unstructured tail, which is required for the proteasome's AAA-ATPases to engage and unfold the substrate.
• Rescue: Adding a 35-amino acid disordered tail (from S. cerevisiae Cytochrome b2) to UCHL1 significantly improved degradation rates.
• Unexpected Finding: Both the native and tail-modified UCHL1 required the p97/VCP unfoldase for degradation, challenging the dogma that p97 is only essential for tightly folded proteins without disordered tails.

4. Key Contributions

1. De-risking Workflow: Established a generalizable chemical genetics platform to screen DUBs for PROTAC suitability before investing in difficult medicinal chemistry campaigns for non-active-site ligands.
2. Categorization of DUBs: Defined three mechanistic classes of DUBs regarding TPD:
   • Susceptible (e.g., USP11).
   • Resistant due to auto-deubiquitylation (e.g., USP4, USP15).
   • Resistant due to structural constraints (e.g., UCHL1).
3. Selectivity Mechanism: Proposed a novel mechanism for achieving selectivity among paralogs. Even if a PROTAC binds USP11, USP4, and USP15 equally, it might selectively degrade USP11 because USP4 and USP15 can "self-correct" via auto-deubiquitylation, while USP11 cannot.
4. Insight into p97: Provided evidence that p97 is essential for degrading even proteins with disordered tails if the core structure is knotted or difficult to unwind (UCHL1 has a 5-2 knot).

5. Significance

• Therapeutic Strategy: The study validates that TPD is a viable strategy for targeting DUBs like USP11, which are difficult to inhibit orthosterically.
• Strategic Guidance: It warns drug developers that for certain DUBs (like USP4/15), simply recruiting an E3 ligase may be insufficient; strategies to inhibit the DUB's catalytic activity simultaneously or engineer the PROTAC to overcome auto-deubiquitylation may be necessary.
• Fundamental Biology: The findings refine the understanding of how protein structure (knots, disordered tails) and enzymatic activity (auto-deubiquitylation) dictate proteasomal turnover rates.

Conclusion: The paper demonstrates that while some DUBs are excellent PROTAC targets, others are inherently resistant due to their enzymatic ability to reverse ubiquitination or their structural inability to be processed by the proteasome. This workflow allows researchers to prioritize targets (like USP11) and anticipate challenges for others before beginning ligand discovery.