Branch-specific axon pruning induced by Dpr4/DIP-{Theta} transneuronal interactions

This study demonstrates that trans-neuronal interactions between Dpr4 in $\gamma$-Kenyon cells and DIP-$\theta$ in dopaminergic neurons specifically inhibit the pruning of vertical axonal branches in the *Drosophila* mushroom body via an N-Cadherin-dependent mechanism, revealing a novel branch-specific regulatory layer in neural circuit remodeling.

The Gist

The Big Picture: The Great Brain Renovation

Imagine your brain as a massive, chaotic construction site during a child's development. It builds way too many roads (neural connections) at first. As the child grows up, the brain needs to do some "renovation." It has to tear down the old, unused roads and build new, efficient highways for adulthood. This process is called neuronal remodeling.

In fruit flies (Drosophila), this happens during their "teenage years" (metamorphosis). A specific type of brain cell, called a γ-Kenyon cell, starts its life as a larva with two main branches (like a Y-shape):

1. A Vertical Branch.
2. A Medial Branch.

Normally, the cell knows exactly what to do: it completely demolishes BOTH the Vertical Branch and the Medial Branch. Once the site is cleared, it builds a brand new, single Adult-Specific Medial Branch. It's a precise, timed demolition of the entire larval structure followed by the construction of a new adult structure.

The Mystery: The "Glitch" in the System

The scientists in this paper were studying a group of proteins called Dpr and DIP. Think of these proteins as name tags or handshake signals that cells use to recognize each other. They are like the "Hello, I'm a neuron" badges that help cells find their correct partners.

The researchers decided to play a trick on the system. They forced the γ-Kenyon cells to wear too many of one specific name tag, called Dpr4.

What happened?
Instead of a total failure, something weird and specific happened:

• The Larval Vertical Branch refused to be demolished. It stayed stuck there.
• The Larval Medial Branch was demolished perfectly fine, exactly as it should be.
• The Adult-Specific Medial Branch was then rebuilt correctly on the cleared site.

This is like a construction crew that was supposed to knock down an old shed and an old house next door. They successfully demolished the house, but the shed refused to fall and stayed standing. Meanwhile, the crew successfully built a brand new skyscraper on the cleared lot. The cell ended up with a "tempo-chimera"—a mix of its baby self (the stuck shed) and its adult self (the new skyscraper).

Why is this a big deal?
Until now, every known mutation that messed up this pruning process caused both larval branches (vertical and medial) to fail and stay intact. This is the first time researchers have seen ONE specific branch stay intact while the other is demolished normally. This branch-specific inhibition is the major novelty of the study.

The Investigation: Who is the Culprit?

The scientists asked: Why did only the vertical branch get stuck?

Since the "Dpr4 name tag" was everywhere on the cell, the problem must be coming from the neighborhood. The vertical branch lives in a specific part of the brain where a different type of neuron (a Dopaminergic Neuron) hangs out.

The scientists discovered that these neighboring neurons have a specific "handshake partner" called DIP-θ.

• The Interaction: The overabundant Dpr4 on the γ-Kenyon cell grabbed onto the DIP-θ on the neighbor.
• The Result: This handshake sent a "STOP" signal to the demolition crew. Because the DIP-θ neighbors only hang out near the vertical branch, only that branch got the "Do Not Demolish" order. The larval medial branch, having no neighbors with DIP-θ nearby, was demolished normally, allowing the adult branch to be rebuilt.

Analogy: Imagine a construction crew (the pruning machinery) is about to tear down a shed and a house. Suddenly, a neighbor (DIP-θ) walks up and shakes hands with the foreman (Dpr4). The neighbor says, "Hey, don't touch this shed, it's my favorite!" The crew stops. But because the neighbor only stands near the shed and not the house next door, the house gets demolished anyway.

The Secret Mechanism: The "Ig2" Switch

The scientists then looked at how this handshake worked. They knew the main handshake happened at the first part of the protein (Ig1). But they found that the second part of the protein (Ig2) was also crucial.

They swapped the "Ig2" part of Dpr4 with a different protein's part.

• If they swapped it with a similar protein's part, the "Stop Demolition" signal still worked.
• If they swapped it with a totally different protein's part, the signal failed.

This suggests that Dpr4 needs a third partner (a co-receptor) to actually send the "Stop" signal inside the cell. The Ig2 part is the docking station for this third partner.

The Downstream Effect: The "N-Cadherin" Link

Since these proteins (Dpr/DIP) are stuck on the outside of the cell and can't talk to the inside directly, they need a messenger. The scientists suspected a protein called N-Cadherin was the messenger.

They tested this by breaking the N-Cadherin gene.

• Result: When N-Cadherin was broken, the "Stop Demolition" signal from Dpr4 disappeared. The vertical branch was demolished normally, even though Dpr4 was still overexpressed.

Analogy: Think of Dpr4 as a person shouting "STOP!" from outside a building. N-Cadherin is the intercom system inside the building. If you smash the intercom, the people inside (the demolition crew) never hear the shout, so they keep working.

Why Does This Matter?

1. Precision: It shows that the brain can control the demolition of one specific branch while leaving the other larval branch alone. It's not a "blunt instrument" approach; it's surgical.
2. Independence of Three Processes: This result proves that "tearing down" (pruning) and "building up" (regrowth) are not just two separate jobs, but three. You can block the pruning of one specific larval branch while both the pruning of the other larval branch AND the subsequent regrowth of the adult branch proceed normally.
3. Human Health: Since these proteins exist in humans too (and are linked to conditions like autism and schizophrenia), understanding how they control brain wiring could help us understand what goes wrong in neurodevelopmental disorders.

Summary in One Sentence

The researchers found that by overloading a brain cell with a specific "name tag" (Dpr4), they accidentally tricked a neighboring cell into sending a "Do Not Demolish" signal to just one specific larval branch (the vertical one) while the other larval branch (the medial one) was pruned normally and the adult branch was rebuilt, revealing a complex handshake involving a third messenger (N-Cadherin) that controls how our brains rewire themselves with unprecedented precision.

Technical Summary

1. Problem Statement

Neuronal remodeling is a conserved developmental mechanism essential for refining neural circuits, involving the precise elimination of exuberant connections followed by regrowth. While the spatiotemporal precision of this process is well-documented, the molecular mechanisms governing branch-specific pruning (why specific axonal branches are eliminated while others are spared) remain poorly understood.

• Context: In Drosophila melanogaster, $\gamma$-Kenyon cells ($\gamma$-KCs) in the mushroom body (MB) undergo stereotyped remodeling during metamorphosis, pruning both vertical and medial larval axonal branches before regrowing a single adult-specific medial branch.
• Gap: Previous studies identified dynamic expression of the Dpr (Defective proboscis extension response) and DIP (Dpr-interacting protein) families (Ig superfamily members) in $\gamma$-KCs. However, their functional mechanisms are unclear, particularly because they are GPI-anchored proteins lacking intracellular signaling domains. It is unknown how they achieve spatial specificity in pruning or what downstream effectors mediate their function.

2. Methodology

The study employed a combination of Drosophila genetics, protein engineering, structural biology, and high-resolution imaging:

• Genetic Manipulation:
  • Overexpression: Used the $\gamma$-KC-specific driver R71G10-Gal4 (and R71G10-QF2) to overexpress Dpr4 and various mutants/chimeras.
  • Knockdown/Suppression: Utilized RNAi to knock down DIP-θ in specific neuronal populations (PPL1 dopaminergic neurons) and used CRISPR/Cas9 to generate a null allele of N-Cadherin (cadNΔORF).
  • Mosaic Analysis: Employed MARCM (Mosaic Analysis with a Repressible Cell Marker) to generate single-cell clones, confirming cell-autonomous effects.
  • Sparse Labeling: Used "layered" SPARC (SPARC-L) to visualize single neurons and distinguish between larval and adult axonal branches.
• Protein Engineering & Structural Biology:
  • Designed point mutations (I87D, Y95D, Q130D) in the Ig1 domain of Dpr4 to disrupt binding interfaces based on crystal structures and AlphaFold3 models.
  • Created chimeric transgenes replacing the Ig2 domain of Dpr4 with that of Dpr12 or Fasciclin II (FasII).
  • Validated binding affinities using Surface Plasmon Resonance (SPR) and Free Energy Perturbation (FEP) simulations.
• Imaging & Quantification: Confocal microscopy was used to visualize axonal pruning at specific timepoints (e.g., 24h APF). Pruning severity was quantified via blinded ranking and intensity measurements of axonal lobes.

3. Key Contributions & Results

A. Discovery of Branch-Specific Pruning Inhibition

• Phenotype: Overexpression of Dpr4 in $\gamma$-KCs inhibits axon pruning. Crucially, this inhibition is branch-specific: the vertical larval branch fails to prune, while the medial branch prunes normally and regrows into the correct adult position.
• Significance: This is the first demonstration of branch-specific control in this system, proving that pruning and regrowth are independently regulated processes at the level of individual branches.

B. Mechanism of Action: Trans-neuronal Interaction

• Interaction Partner: The branch-specificity arises because the vertical lobe is innervated by PPL1 dopaminergic neurons (DANs), which endogenously express the binding partner DIP-θ. The medial lobe is innervated by PAM DANs, which do not express DIP-θ.
• Validation:
  • In vitro SPR confirmed that wild-type Dpr4 binds DIP-θ, while Ig1 domain mutants (I87D, Y95D) abolish this binding.
  • In vivo, overexpression of the binding-deficient mutants did not cause pruning defects.
  • Suppression: Knocking down DIP-θ specifically in PPL1-DANs suppressed the Dpr4-induced pruning defect, confirming that ectopic Dpr4 (in $\gamma$-KCs) interacting with DIP-θ (in PPL1-DANs) blocks pruning.

C. The Role of the Ig2 Domain and a Third Partner

• Hypothesis: Since Dpr/DIPs are GPI-anchored and lack intracellular domains, a third co-receptor must mediate downstream signaling.
• Ig2 Domain Importance: Replacing the Ig2 domain of Dpr4 with that of Dpr12 (a related Dpr) retained the strong pruning defect. However, replacing it with the Ig2 of Fasciclin II (an unrelated IgSF adhesion molecule) significantly weakened the phenotype.
• Implication: The Ig2 domain is critical for recruiting a conserved co-receptor, distinct from the Ig1-mediated Dpr-DIP binding.

D. Identification of N-Cadherin as a Downstream Mediator

• Genetic Link: Based on the hypothesis that a transmembrane adhesion molecule acts as the co-receptor, the authors tested N-Cadherin (CadN).
• Result: While cadN loss-of-function clones failed to regrow adult axons (consistent with previous work), loss of CadN within Dpr4-overexpressing clones completely suppressed the pruning defect. The vertical branches pruned normally despite Dpr4 overexpression.
• Conclusion: N-Cadherin is genetically required to mediate the Dpr4-induced inhibition of pruning, likely acting downstream of the Dpr4/DIP-θ interaction.

4. Significance

• Spatial Regulation of Remodeling: The study provides a mechanistic model for how neural circuits achieve spatial precision. It demonstrates that trans-neuronal interactions between specific ligand-receptor pairs (Dpr4/DIP-θ) can locally inhibit pruning in specific axonal compartments based on the innervation pattern of neighboring neurons.
• Decoupling Pruning and Regrowth: The "tempo-chimeric" neurons (retaining larval vertical branches while regrowing adult medial branches) prove that pruning and regrowth are separable events, offering a new tool to study the independent regulation of these processes.
• Signaling Pathway: The findings suggest a novel signaling cascade: Dpr4 (in $\gamma$-KC) + DIP-θ (in PPL1-DAN) $\rightarrow$ Recruitment of N-Cadherin $\rightarrow$ Cytoskeletal stabilization (inhibition of pruning).
• Broader Relevance: As Dpr/DIPs have vertebrate homologs (IgLONs) associated with neurodegenerative disorders, and N-Cadherin is conserved, these findings may illuminate mechanisms of circuit refinement and pathology in higher organisms.

Summary Model

1. Normal State: $\gamma$-KCs downregulate Dpr4 to allow pruning of both lobes.
2. Perturbed State (Dpr4 Overexpression): Dpr4 is present on $\gamma$-KC surfaces.
3. Vertical Lobe: Dpr4 binds to DIP-θ on PPL1-DANs. This interaction recruits N-Cadherin (via the Dpr4 Ig2 domain), stabilizing the cytoskeleton and blocking pruning.
4. Medial Lobe: PAM-DANs lack DIP-θ; thus, no ectopic interaction occurs, and pruning proceeds normally.
5. Regrowth: Once pruning is complete (or bypassed), the adult medial branch regrows independently, unaffected by the persistent vertical branch.