FAF1 and FAF2 enhance human p97-UFD1-NPL4 complex unfoldase activity enabling rational design of p97 activators

This study identifies FAF1 and FAF2 as potent activators of the human p97-UFD1-NPL4 unfoldase complex, elucidating their structural mechanism for enhancing substrate unfolding and demonstrating how this knowledge enables the rational design of de novo proteins to boost p97 activity.

The Gist

Imagine your cell is a bustling, high-tech city. In this city, there is a massive, powerful garbage truck called p97. Its job is to grab broken, tangled, or unwanted proteins (the trash), pull them apart, and shove them into a shredder (the proteasome) to be recycled.

However, this garbage truck doesn't work alone. It needs a specific set of tools and a driver's assistant to know what to grab and how to pull it. In the scientific world, this assistant is a team called UFD1-NPL4.

The Problem: The Truck is Sluggish

Scientists discovered something strange: The human version of this garbage truck (p97) is much slower and less efficient than the yeast version. Even when the "trash" is clearly marked with a "Take Me Away" tag (a chain of ubiquitin molecules), the human truck often struggles to get the job done, especially if the tag is short or the trash is stuck tight. It's like having a powerful engine but a transmission that slips; the truck just can't get the traction it needs to pull the heavy load.

The Discovery: Finding the "Turbocharger"

The researchers asked: Is there something else in the human cell that helps this truck work better?

They screened a list of potential helpers and found two standout candidates: FAF1 and FAF2. Think of these as turbochargers or power boosters. When they attached FAF2 to the garbage truck and its assistant, the truck suddenly became incredibly fast and efficient. It could now pull apart even the toughest, most stubborn trash, even if the "Take Me Away" tag was relatively short.

How It Works: The "Velcro" Trick

How does FAF2 do this? The researchers used advanced imaging (like a super-high-resolution 3D camera) to see exactly what was happening.

1. The Bridge: FAF2 acts like a bridge. It grabs onto the truck's assistant (UFD1) and also grabs onto the "trash" (the ubiquitin chain).
2. The Stabilizer: Normally, the assistant struggles to hold onto short tags. FAF2 comes in and acts like a third hand or a clamp. It holds the tag steady, making it much easier for the truck to get a grip and start pulling.
3. The Shape-Shift: Interestingly, FAF2 is a bit of a shapeshifter. When it's floating around alone, it's floppy and unstructured (like a loose piece of string). But the moment it latches onto the truck and the trash, it snaps into a rigid, helical shape (like a stiff spring) that locks everything together perfectly.

The Big Breakthrough: Designing a Custom Tool

The most exciting part of the paper is what the scientists did next. They realized that FAF2 works because of a very specific, small section of its body (a tiny stretch of amino acids) that acts as the "clamping mechanism."

Instead of using the whole FAF2 protein, they asked: Can we build a brand-new, tiny tool from scratch that just has this clamping mechanism?

Using a powerful AI design tool (called RFdiffusion, which is like a 3D printer for proteins), they designed de novo proteins (completely new proteins that don't exist in nature). These new "mini-tools" were designed to look like the clamping part of FAF2.

The Result: These tiny, custom-designed tools worked just like the natural turbocharger! They successfully boosted the garbage truck's speed, proving that you don't need the whole complex protein to get the job done—you just need the right "key" to unlock the power.

Why This Matters

This discovery is a game-changer for medicine.

• The Disease Connection: When the garbage truck (p97) breaks down or slows down, trash piles up in the cell. This buildup is linked to terrible diseases like ALS (Lou Gehrig's disease) and other neurodegenerative disorders.
• The Solution: By understanding exactly how FAF2 boosts the truck, scientists can now design drugs or synthetic proteins that act as "turbochargers" for patients whose trucks are broken. Instead of just trying to fix the engine, we can give the truck a super-powerful assistant to help it clear the trash faster, potentially slowing down or stopping these diseases.

In short: The scientists found a natural "power-up" for the cell's cleanup crew, figured out exactly how it works, and then used AI to build a custom version of that power-up. It's like taking a slow, struggling car, finding a secret turbo kit, and then 3D-printing a perfect copy of that kit to help anyone with a broken engine.

Technical Summary

1. Problem Statement

• The Core Issue: VCP/p97 is an AAA+ ATPase essential for extracting ubiquitylated proteins from membranes and complexes for proteasomal degradation. While the yeast homolog (Cdc48) functions efficiently with its core cofactors Ufd1-Npl4 (UN), the human p97-UN complex exhibits significantly reduced unfoldase activity, particularly with shorter ubiquitin chains (4–10 ubiquitins), which are more physiologically relevant than the long chains often used in vitro.
• The Gap: Human p97 interacts with over 30 cofactors (compared to ~15 in yeast), suggesting a need for additional regulatory factors to fine-tune activity. However, the specific mechanisms by which these cofactors modulate human p97-UN activity, especially regarding substrate engagement and chain-length thresholds, were poorly understood.
• Therapeutic Context: Mutations in p97 cause neurodegenerative diseases (e.g., ALS, MSP). Enhancing p97 activity could help clear toxic protein aggregates, but current activators are often non-specific or allosteric. A rational, mechanism-based approach to activate p97 specifically in the context of its core machinery is needed.

2. Methodology

The authors employed a multidisciplinary approach combining biochemistry, structural biology, and computational protein design:

• In Vitro Reconstitution & Kinetic Assays:
  • Compared yeast Cdc48-UN and human p97-UN using a fluorescent reporter substrate (Ub-mEos3.2) with varying K48-linked ubiquitin chain lengths (Very Short, Short, Long).
  • Screened five UBX-domain containing cofactors (p47, UBXN7, FAF1, FAF2, UBXN1) for their ability to stimulate unfolding.
  • Measured ATPase activity (basal and substrate-stimulated) using phosphate release assays.
• Structural Biology & Biophysics:
  • AlphaFold3 (AF3): Used to model the p97-UN-FAF2 complex and the complex with polyubiquitin chains to predict interaction interfaces.
  • Cross-linking Mass Spectrometry (XL-MS): Validated AF3 predictions by mapping physical contacts between p97, UN, FAF2, and ubiquitin.
  • Hydrogen-Deuterium Exchange MS (HDX-MS): Analyzed conformational dynamics and solvent accessibility changes upon complex formation, specifically looking at the disorder-to-order transition of FAF2.
  • Mass Photometry: Confirmed the stoichiometry of the assembled complexes.
  • Isothermal Titration Calorimetry (ITC): Tested direct binding affinities between FAF2/UFD1 and ubiquitin.
• Computational Protein Design:
  • Used RFdiffusion and ProteinMPNN to design de novo mini-proteins.
  • Strategy: Scaffolding the identified "Activation Motif" (AM) of FAF2 (residues 286–312) onto novel protein backbones to mimic the interaction with UFD1 and ubiquitin, bypassing the need for the native FAF2 UBX domain.

3. Key Contributions & Results

A. Identification of FAF1 and FAF2 as Potent Activators

• Screening: Among tested cofactors, FAF2 (and to a lesser extent FAF1) was identified as the strongest activator of human p97-UN.
• Chain Length Specificity: While p47 and UBXN7 could stimulate unfolding of long ubiquitin chains, only FAF1 and FAF2 could efficiently drive the unfolding of substrates with short (Ub6) ubiquitin chains, effectively lowering the activation threshold to match yeast Cdc48 efficiency.
• Mechanism of Activation: FAF2 does not increase basal ATPase activity. Instead, it significantly boosts substrate-stimulated ATP hydrolysis, indicating it enhances the efficiency of substrate engagement and commitment to unfolding.

B. Structural Mechanism: The "Activation Motif" (AM)

• Complex Architecture: AF3 and XL-MS revealed that FAF2 binds the p97-UN complex in a UFD1-dependent manner. The FAF2 C-terminal UBX domain binds the p97 N-domain, while its central region (CR) bridges to the UT3 domain of UFD1.
• Ubiquitin Bridging: The FAF2 CR does not bind ubiquitin with high affinity on its own. Instead, it acts as a scaffold that stabilizes the interaction between the UT3 domain of UFD1 and the proximal ubiquitin (the one closest to the substrate).
• Disorder-to-Order Transition: HDX-MS showed that the FAF2 central region (residues ~275–348) is intrinsically disordered in the apo state but adopts a stable $\alpha$-helical structure upon binding to the p97-UN complex.
• Critical Residues: Mutagenesis identified a conserved Activation Motif (AM) within the FAF2 CR (residues 286–312). Key residues (e.g., D304, R290) mediate contacts with UFD1 and the ubiquitin linkage. Mutations here abolish activation.
• Cooperative Binding: The data supports a model where FAF2, UFD1, and ubiquitin form a multivalent, high-avidity complex. The AM stabilizes the "initiator" ubiquitin in the NPL4 groove, facilitating the rate-limiting step of substrate engagement.

C. Rational Design of De Novo p97 Activators

• Design Strategy: Using the FAF2 AM (residues 286–312) as a fixed "hotspot," the authors used RFdiffusion to generate novel protein scaffolds that mimic the AM's interaction with UFD1 and ubiquitin.
• Validation: Of 15 designed binders, 4 successfully activated p97-UN unfolding. Two binders (Binder 4 and Binder 6) showed potent activity comparable to FAF2 in maximal unfolding rates.
• Potency vs. Mechanism: While the de novo binders had higher EC50 values (lower affinity) than full-length FAF2 (likely due to the lack of the native UBX domain anchoring to p97), they proved that the AM alone is necessary and sufficient to confer activation. This demonstrates that the activation mechanism can be decoupled from the full-length protein architecture.

4. Significance

• Mechanistic Insight: The study resolves why human p97 is less efficient than yeast Cdc48 and identifies FAF1/FAF2 as critical adaptive cofactors that lower the ubiquitin chain length threshold for substrate processing.
• Novel Activation Paradigm: It reveals a mechanism where a cofactor acts not just as a recruiter, but as a structural stabilizer that bridges the core adaptor (UFD1) and the ubiquitin signal, overcoming a kinetic barrier in substrate engagement.
• Therapeutic Potential: The successful de novo design of p97 activators provides a proof-of-concept for rational drug design targeting protein homeostasis. These engineered proteins (or future small molecules mimicking the AM) could specifically boost p97 activity in disease contexts (e.g., ALS, tauopathies) without the off-target effects of non-specific allosteric activators.
• Methodological Advance: The work demonstrates the power of combining structural prediction (AF3), biophysics (HDX/XL-MS), and generative AI (RFdiffusion) to dissect complex molecular machines and engineer functional mimics.

In summary, this paper defines a specific molecular "switch" (the FAF2 Activation Motif) that fine-tunes human p97 activity and leverages this knowledge to engineer synthetic activators, offering a new avenue for treating proteinopathies.