Paradoxical non-catalytic kinase functions are driven by inhibitor-induced displacement of autoinhibitory domains

This study reveals that ATP-competitive kinase inhibitors frequently trigger paradoxical, non-catalytic effects by inducing structural rearrangements in autoinhibitory domains that alter protein-protein interactions and subcellular localization, a phenomenon identified through a multimodal proteomics approach.

The Gist

The Big Idea: The "Double Agent" Drug

Imagine you hire a security guard (a drug) to stop a specific worker (a kinase protein) from doing their job. You tell the guard, "Just make sure the worker stops typing on their keyboard."

In the world of cancer drugs, scientists have been doing exactly this for decades. They create drugs that block the "keyboard" (the catalytic activity) of proteins called kinases. These proteins are like the managers of a cell; they tell other parts of the cell what to do by sending chemical messages. When a kinase goes rogue (mutates), it causes cancer. So, the goal is to hit the "off" switch on the keyboard.

But here's the twist: This paper discovered that when you hit the "off" switch, you aren't just stopping the typing. You are also accidentally changing the worker's posture and clothing, which causes them to start hugging different people or running to different rooms in the office. These new behaviors can cause unexpected side effects, even though the worker is technically "stopped."

The Analogy: The Folded Origami Crane

To understand how this works, imagine a kinase protein is like a piece of paper folded into an origami crane.

1. The Autoinhibitory Domain (The Folded Wing): In its resting state, the protein is folded up tight. A specific part of the paper (the "Autoinhibitory Domain" or AID) is tucked under the body, acting like a safety lock. This keeps the protein inactive and prevents it from grabbing onto other things.
2. The Drug (The Hand that Unfolds): When the drug binds to the protein, it doesn't just break the keyboard. It acts like a hand that pulls the folded wing open.
3. The Paradox: The protein is now unfolded (open) and looks like it's ready to work, but the drug is still jamming the keyboard, so it can't actually send messages.
   • The Problem: Because the protein is now "open," it exposes new sticky surfaces that were previously hidden. It starts grabbing onto different coworkers or sticking to different walls in the cell.

The Three Case Studies

The researchers tested this idea on three different "workers" (kinases) to see what happened when the drugs opened them up.

1. CAMKK2: The Bodyguard Who Won't Let Go

• The Job: CAMKK2 usually helps activate a protein called AMPK (the cell's energy manager).
• The Drug: The drug SGC-CAMKK2-1 blocks the keyboard.
• The Surprise: Because the drug forced CAMKK2 to "open up," it grabbed onto AMPK and held it tight, like a bodyguard hugging a VIP.
• The Consequence: AMPK was trapped! It couldn't move around to get activated by other signals (like low energy). The drug didn't just stop CAMKK2; it hid AMPK from the rest of the cell. This is a "non-catalytic" effect—the drug changed the relationship between proteins, not just the activity.

2. CHEK1: The Detached Safety Pin

• The Job: CHEK1 is a DNA repair manager. It usually hangs out with a mitochondrial protein called CLPB (think of CLPB as a safety pin holding the cell's power plant together).
• The Drug: The drug rabusertib blocks the keyboard.
• The Surprise: The drug forced CHEK1 to open up, causing it to let go of the safety pin (CLPB).
• The Consequence: Without the safety pin, the cell's power plant (mitochondria) started breaking apart and fragmenting. This happened even though the drug successfully stopped the DNA repair work. The side effect came from the loss of a hug, not the loss of the job.

3. PRKCA: The Office Drifter

• The Job: PRKCA usually floats around the cell, only sticking to the walls when it gets a specific signal (calcium).
• The Drug: The drug Gö 6983 blocks the keyboard.
• The Surprise: The drug opened up a specific part of the protein (the C2 domain), which acts like a magnet. Suddenly, the protein didn't need a signal to stick to the wall. It rushed to the cell junctions (the glue holding cells together).
• The Consequence: The protein moved to the wrong place and started messing with the cell's ability to stick to its neighbors. Again, the drug didn't just stop the work; it changed the protein's address.

Why Does This Matter?

For years, scientists thought that if a drug stopped a protein from working, that was the only thing that mattered. This paper says: "No, that's just the tip of the iceberg."

• The Hidden Danger: Many drugs might be causing side effects not because they are too strong or weak, but because they are changing the shape of the protein, causing it to interact with the wrong things.
• The New Tool: The researchers used a high-tech "X-ray vision" method (combining mass spectrometry with protein digestion) to see these shape changes and new interactions. It's like using a drone to map out not just who is in the room, but how they are standing and who they are talking to.
• The Future: If drug developers test for these "shape-shifting" side effects early on, they can design better drugs that stop the bad work without accidentally changing the protein's posture or location.

The Takeaway

Drugs are like keys that fit into locks. We used to think the key just turned the lock to stop the machine. This paper shows that sometimes, the key also changes the shape of the doorframe, causing the door to swing open and hit a vase on the other side. To make safer medicines, we need to watch the whole room, not just the lock.

Technical Summary

1. Problem Statement

ATP-competitive kinase inhibitors (Type I inhibitors) are a major class of anti-cancer drugs designed to block the catalytic activity of kinases by binding to the ATP-binding site. However, clinical and cellular observations reveal "paradoxical" phenotypic effects that cannot be explained solely by the inhibition of substrate phosphorylation.

• The Gap: Current pharmacological understanding focuses on catalytic inhibition and target specificity. It largely overlooks how inhibitor binding alters kinase conformation and protein-protein interactions (PPIs), specifically regarding Autoinhibitory Domains (AIDs).
• The Challenge: AIDs are often intrinsically disordered or linked via disordered regions, making them difficult to characterize using classical structural methods (e.g., X-ray crystallography). Consequently, the extent to which inhibitors induce "open" (active-like) conformations that drive non-catalytic functions remains a "dark space" in kinome pharmacology.

2. Methodology

The authors developed a multimodal proteomics approach to systematically map inhibitor-induced conformational and interactome changes in a native cellular context.

• Target Kinases: Three disease-associated kinases with distinct AID mechanisms were selected:
  • CAMKK2: Linear motif AID; activated by Calmodulin.
  • CHEK1: Structured AID; activated by ATR phosphorylation.
  • PRKCA: Disordered pseudosubstrate AID; activated by Ca²⁺.
• Core Techniques:
  1. AP-LiP-MS (Affinity Purification - Limited Proteolysis Mass Spectrometry):
     • Target kinases were affinity-purified from human cells under native conditions.
     • Samples were treated with specific ATP-competitive inhibitors (SGC-CAMKK2-1, rabusertib, Gö 6983) or DMSO.
     • Proteins were treated with Proteinase K (PK), which cleaves accessible/flexible regions.
     • LC-MS analysis detected changes in peptide abundance, indicating structural rearrangements (increased accessibility = open conformation).
  2. AP-MS (Affinity Purification Mass Spectrometry): To identify changes in stable protein complexes.
  3. In Vivo Proximity Labeling (miniTurbo): To map dynamic changes in the local proteome and subcellular localization.
  4. Mutagenesis & Validation: Catalytically inactive (D312A, D130A, D463A) and constitutively active (T483D, L449R, R22A/A25E) mutants were used to decouple catalytic inhibition from conformational changes.
  5. Structural Modeling: AlphaFold3 was used to predict interaction interfaces (e.g., CAMKK2-PRKAA1).
  6. Phenotypic Assays: Live-cell imaging, super-resolution microscopy (mitochondrial morphology), and Western blotting (phosphorylation status).

3. Key Contributions

• Discovery of a General Mechanism: Demonstrated that ATP-competitive inhibitor binding universally displaces AIDs, driving kinases into an open, active-like conformation even though catalytic activity is blocked.
• Methodological Framework: Established a high-throughput, unbiased pipeline (AP-LiP-MS + Contextual Proteomics) to detect "off-mechanism" drug effects without prior structural knowledge.
• Neomorphic Functions: Identified specific, non-catalytic gain-of-function mechanisms for three distinct kinases that explain paradoxical cellular phenotypes.

4. Key Results

A. Structural Rearrangements (AP-LiP-MS)

• Inhibitor binding to CAMKK2, CHEK1, and PRKCA consistently caused increased PK accessibility at their respective AIDs.
• This indicates a transition from a closed (autoinhibited) state to an open conformation, mimicking physiological activation.
• For PRKCA, structural changes were also observed in the C2 domain (membrane-binding), suggesting allosteric disruption of C2-KD interactions.

B. Case Study 1: CAMKK2 & AMPK Sequestration

• Observation: Inhibition of CAMKK2 with SGC-CAMKK2-1 induced a strong, stable complex formation with the AMPK catalytic subunit (PRKAA1).
• Mechanism: The inhibitor locks CAMKK2 in an open conformation that binds PRKAA1.
• Consequence: This binding sequesters PRKAA1, physically blocking the T183 phosphorylation site (normally targeted by LKB1 during glucose starvation).
• Validation: The effect was independent of CAMKK2 catalytic activity (observed in D312A mutant) but abolished by the disease-associated R311C mutation, which disrupts the interface.
• Outcome: The drug acts as a dual inhibitor: blocking catalysis and preventing AMPK activation by other kinases via sequestration.

C. Case Study 2: CHEK1 & Mitochondrial Fragmentation

• Observation: Inhibition of CHEK1 with rabusertib caused the dissociation of the mitochondrial protein CLPB from CHEK1 complexes.
• Mechanism: The dissociation is driven by the conformational change (open state) induced by the inhibitor, not by catalytic inhibition (observed in D130A mutant).
• Phenotype: Loss of the CHEK1-CLPB interaction led to significant mitochondrial fragmentation.
• Validation: This fragmentation occurred even when the canonical CDK1-DRP1 axis was inhibited, ruling out the standard DNA damage response pathway. It suggests CHEK1 normally stabilizes CLPB to maintain mitochondrial integrity.

D. Case Study 3: PRKCA Relocalization

• Observation: Treatment with Gö 6983 caused a rapid (within 8 minutes) relocalization of PRKCA from the cytoplasm to cell junctions (tight junctions, adherens junctions, desmosomes).
• Mechanism: Inhibitor-induced structural changes in the C2 domain increased its affinity for membrane lipids (likely PIP2) or junctional proteins (e.g., ITGB1).
• Validation: The relocalization occurred in catalytically inactive mutants, confirming it is a conformational effect.
• Outcome: The drug traps inhibited PRKCA at cell junctions, potentially disrupting intercellular signaling and junctional architecture.

5. Significance

• Redefining Drug Action: The study challenges the view that kinase inhibitors act solely through catalytic blockade. It posits that the ATP-binding site is a central organizing hub where inhibitor binding dictates global conformation and interactome remodeling.
• Clinical Implications: These "on-target, off-mechanism" effects (neomorphic functions) likely contribute to drug side effects and resistance. For example, the sequestration of AMPK or disruption of mitochondrial homeostasis could explain toxicity profiles not predicted by catalytic inhibition models.
• Future Drug Development: The authors advocate for the systematic integration of structural and contextual proteomics (AP-LiP-MS) early in drug discovery. This approach can identify neomorphic liabilities before clinical trials, leading to the development of safer, more effective therapeutics by either avoiding these conformations or exploiting them for novel indications.
• Generalizability: Since most FDA-approved kinase inhibitors are Type I, this mechanism of AID displacement and conformational unlocking is likely a widespread phenomenon across the kinome.