In-cell cryo-electron tomography reveals differential effects of type I and type II kinase inhibitors on LRRK2 filament formation and microtubule association

This study utilizes in-cell cryo-electron tomography to demonstrate that type I kinase inhibitors (MLi-2) promote extensive microtubule-associated LRRK2 filament formation and bundling, enabling the construction of a full-length structural model, whereas type II inhibitors (GZD-824) prevent such assemblies.

The Gist

The Big Picture: A Broken Machine in the Cell

Imagine your body is a bustling city, and inside every cell, there are tiny highways called microtubules. These highways are used by delivery trucks (vesicles) to transport important cargo like food and waste.

Now, imagine a specific protein called LRRK2. In a healthy city, LRRK2 is like a maintenance worker who mostly hangs out in the office (the cytosol), checking his phone and waiting for instructions. However, in many people with Parkinson's disease, a mutation turns this worker into a hyperactive supervisor. This "hyperactive" LRRK2 gets too excited, starts shouting orders, and disrupts the delivery trucks, causing traffic jams that lead to cell death.

Scientists have been trying to build "brakes" (drugs) to calm this hyperactive supervisor down. But there's a catch: there are two different types of brakes, and they seem to work in very different ways. This paper investigates what happens when you use Type I brakes versus Type II brakes inside the actual cell.

---

The Two Types of "Brakes" (Inhibitors)

Think of the hyperactive LRRK2 protein as a shape-shifting robot. It has a "closed" mode (active/aggressive) and an "open" mode (inactive/calm).

1. Type I Inhibitors (The "Lock-Down" Brake):

   • Analogy: Imagine a security guard who grabs the robot and forces it into a closed, rigid box.
   • What happens: The robot is forced into a specific shape. Surprisingly, when the scientists used this drug (MLi-2), the hyperactive LRRK2 didn't just calm down; it went crazy and started clumping together. It wrapped itself tightly around the microtubule highways, forming massive, tangled bundles. It was like the maintenance workers decided to build a giant wall around the highway, blocking all traffic.

2. Type II Inhibitors (The "Open-Door" Brake):

   • Analogy: Imagine a different security guard who gently opens the robot's doors and tells it to relax and spread out.
   • What happens: When the scientists used this drug (GZD-824), the LRRK2 stayed scattered. It didn't clump together. It remained mostly in the "office," leaving the microtubule highways clear and free for the delivery trucks to pass.

The Discovery: Why the Difference Matters

The researchers used a high-tech camera (Cryo-Electron Tomography) to take 3D movies of these cells. Here is what they found:

• With Type I (The Clumping Drug): The LRRK2 formed a beautiful, highly organized lattice (like a honeycomb) around the microtubules. Because the drug forced the protein into a stable, rigid shape, the scientists could finally see the full blueprint of the protein.

  • The "Aha!" Moment: They discovered that in this "closed" state, the protein's "arms" (N-terminal repeats) were unhooked and floating freely. This explains how the protein becomes active: it unhooks its own safety locks to get to work. This is the first time scientists have seen this specific "unhooked" shape in a living cell.

• With Type II (The Scattering Drug): The LRRK2 was messy and disorganized. It barely touched the highways. Because it was so scattered and moving around too much, the scientists couldn't get a clear picture of its full shape.

The Takeaway: It's Not Just About Stopping the Engine

The most important lesson from this paper is that stopping the engine (kinase activity) isn't the only thing that matters.

• Type I drugs stop the engine but force the car to park in a giant, messy pile-up on the highway. This might actually make the traffic jam worse, even if the engine is off.
• Type II drugs stop the engine and let the car park neatly in the garage, leaving the highway clear.

Why This is a Big Deal

1. New Structural Map: By using the Type I drug to "freeze" the protein in a stable clump, the scientists built the first complete 3D model of the full-length LRRK2 protein in its active state. It's like finally getting the full instruction manual for a machine that was previously too wobbly to photograph.
2. Drug Safety: This study warns drug developers that just because a drug stops the protein from working, it doesn't mean it's safe. Some drugs might cause the protein to clump up in dangerous ways that block cell transport.
3. Future Treatments: Understanding exactly how these drugs change the shape of the protein helps scientists design better medicines that calm the hyperactive LRRK2 without causing it to clog up the cell's highways.

In short: The paper shows that different drugs change the shape of the Parkinson's-related protein in completely different ways. One drug forces it to build a wall around the cell's highways, while the other lets it stay scattered and harmless. This helps us understand the protein's structure better and choose safer drugs for the future.

Technical Summary

1. Problem Statement

Leucine-Rich Repeat Kinase 2 (LRRK2) is a major genetic risk factor for Parkinson's disease (PD). Pathogenic mutations (e.g., I2020T, G2019S) hyperactivate LRRK2 kinase activity, leading to the hyperphosphorylation of Rab GTPases and subsequent disruption of vesicular trafficking and cytoskeletal dynamics.

• The Gap: While LRRK2 is typically cytosolic, hyperactive mutants can form microtubule-associated filaments. However, the physiological relevance of this filamentation and the structural mechanisms by which different classes of kinase inhibitors (Type I vs. Type II) modulate LRRK2 architecture in situ remain unclear.
• The Challenge: Resolving the full-length structure of LRRK2 in its active state has been difficult due to the conformational heterogeneity of its flexible N-terminal repeats (LRR, ANK, ARM domains) and the dynamic nature of the protein in cellular environments.

2. Methodology

The study employed an integrated in-cell structural biology approach combining cryo-correlative light and electron microscopy (cryo-CLEM) with cryo-electron tomography (cryo-ET).

• Cell Model: 293T cells were transfected with a GFP-tagged, hyperactive LRRK2 mutant (LRRK2-I2020T).
• Pharmacological Treatment: Cells were treated with two distinct kinase inhibitors:
  • Type I Inhibitor (MLi-2): Stabilizes the kinase domain in a closed, active-like conformation.
  • Type II Inhibitor (GZD-824): Stabilizes the kinase domain in an open, inactive-like conformation.
• Sample Preparation: Cells were vitrified, and cryo-Focused Ion Beam (FIB) milling was used to create thin lamellae (150–200 nm) suitable for imaging.
• Imaging & Correlation:
  • Cryo-Fluorescence Microscopy (cryo-FM): Used to locate GFP-LRRK2IT signals within the vitrified cells.
  • Cryo-Electron Tomography (cryo-ET): Performed on a Titan Krios to acquire tilt-series of the lamellae, visualizing the ultrastructure at nanometer resolution.
• Data Analysis:
  • Subtomogram Averaging: Used to reconstruct 3D maps of LRRK2 filaments on microtubules.
  • Segmentation: Manual and automated tracing of microtubules, membranes, and LRRK2 lattices.
  • Model Building: Rigid body fitting of known LRRK2 domains (RCKW, LRR, ANK, ARM) into the in situ density maps to determine conformational states.

3. Key Results

A. Differential Effects on Localization and Filamentation

• Type I (MLi-2) Treatment: Induced extensive formation of LRRK2-I2020T decorated microtubule bundles.
  • LRRK2 formed highly ordered lattices coating microtubules.
  • Microtubules were bundled into groups of 15–30 microtubules.
  • The center-to-center distance between bundled microtubules increased to ~68.7 nm (compared to ~35 nm for bare bundles) due to the ~15 nm thick LRRK2 coat.
• Type II (GZD-824) Treatment: Resulted in minimal microtubule association.
  • LRRK2-I2020T remained largely cytosolic, appearing as discrete puncta.
  • Microtubule bundling was absent; only rare, isolated decorated microtubules were observed.
  • The few filaments present showed a disordered lattice with reduced nearest-neighbor precision compared to MLi-2 treated cells.

B. Structural Insights into Full-Length LRRK2

By stabilizing the filamentous state with MLi-2, the authors resolved the full-length closed-kinase conformation of LRRK2-I2020T in situ for the first time:

• Kinase Conformation: The catalytic RCKW domains adopted a closed, active conformation, confirmed by the absence of steric clashes in the COR-COR interface and successful fitting of active kinase models.
• N-Terminal Repeats (LRR-ANK-ARM):
  • Unlike previous in vitro intermediate states where N-terminal repeats were docked to the kinase, the in situ map revealed that the LRR and ANK domains were undocked from the catalytic core.
  • These domains shifted by ~19.8 Å away from the kinase, exposing the substrate-binding site.
  • The ARM domains were resolved extending outward, potentially mediating interactions between adjacent microtubules in the bundle.
• Microtubule Interaction: The ROC domain was observed interacting directly with tubulin, consistent with in vitro data but elusive in previous in situ studies.

C. Mechanism of Inhibitor Action

• MLi-2 (Type I): Stabilizes the active, closed kinase conformation, which promotes the undocking of N-terminal repeats and facilitates high-affinity binding to microtubules, leading to ordered filament assembly and bundling.
• GZD-824 (Type II): Stabilizes the inactive, open conformation. This prevents the specific conformational changes required for stable microtubule binding and filament formation. The rare filaments observed likely represent pre-existing assemblies that were not fully converted to the inactive state by the inhibitor.

4. Key Contributions

1. First In Situ Full-Length Structure: Provided the first structural model of full-length LRRK2 in a closed-kinase, active state within a cellular environment, resolving the previously disordered N-terminal repeats.
2. Conformational Coupling: Demonstrated that microtubule association is coupled to the kinase conformational state (closed/active) rather than just kinase activity levels. The undocking of N-terminal repeats is a prerequisite for filament assembly.
3. Inhibitor Specificity: Established a structural basis for the differential effects of Type I vs. Type II inhibitors, showing that Type I inhibitors promote a distinct, pathological filament architecture, while Type II inhibitors prevent it.
4. Methodological Advancement: Successfully applied cryo-CLEM and subtomogram averaging to resolve dynamic protein assemblies in vitrified cells, overcoming the heterogeneity challenges that have plagued previous structural studies.

5. Significance

• Therapeutic Implications: The findings suggest that while Type I inhibitors effectively block kinase activity, they may inadvertently promote the formation of LRRK2 filaments and microtubule bundling, which could disrupt intracellular transport. This highlights a potential "off-target" structural effect of Type I inhibitors that must be considered in drug development.
• Pathogenesis Insight: The study clarifies that PD-associated mutations (like I2020T) drive disease not just through kinase hyperactivity, but by altering the protein's structural landscape to favor microtubule binding and filamentation.
• Structural Biology: The work validates in-cell cryo-ET as a powerful tool for capturing transient, conformationally heterogeneous states of large multidomain proteins that are often lost in in vitro purification.

In summary, this paper provides a critical structural framework linking LRRK2 kinase conformation, inhibitor class, and cytoskeletal pathology, offering new insights for the design of safer and more effective Parkinson's disease therapeutics.