Conservation of extended sequence and structure in the branchpoint-to-3' splice site region upstream of neural microexons

This study reveals that neural microexons in both humans and chickens rely on conserved, accessible RNA secondary structures within the branchpoint-to-3' splice site region to facilitate spliceosome assembly and regulate stage-specific splicing during neurodevelopment.

The Gist

The Big Picture: The "Tiny Puzzle Piece" Problem

Imagine your DNA is a massive instruction manual for building a human. To build a specific protein (like a tool or a machine part), the cell reads this manual, but it has to cut out the "junk" pages (introns) and glue the "useful" pages (exons) together. This process is called splicing.

Usually, the "useful" pages are about the size of a standard paragraph (120–150 letters long). The cell's splicing machine (the spliceosome) is designed to grab the beginning and end of these paragraphs, hold them, and glue them together.

The Problem:
Some of these useful pages are microexons. They are tiny—sometimes only 3 to 27 letters long. Imagine trying to use a giant pair of industrial scissors to cut out a single grain of rice. It's too small for the machine to grab properly. If the machine can't grab it, it might skip the grain of rice entirely, leading to a broken protein. This is a huge problem for the brain, which relies heavily on these tiny pieces to build complex neural circuits.

The Discovery: How the Cell Solves the "Grain of Rice" Problem

This paper investigates how the brain manages to successfully glue these tiny microexons together, specifically looking at chick embryos (a great model for human development) and comparing them to humans.

The researchers found that the cell uses a clever "stretching" trick to make these tiny pieces workable. Here is how they do it:

1. The "Velcro" Stretch (The Polypyrimidine Tract)

Think of the splicing machine as a magnet that needs to stick to a specific spot on the DNA to start gluing. Usually, there is a strip of "Velcro" (a sequence of specific letters called a polypyrimidine tract) right next to the microexon.

• Normal Exons: The Velcro is right next to the piece.
• Microexons: The researchers found that for microexons, the cell stretches this Velcro strip farther away into the "junk" area (the intron).
• The Analogy: Imagine you are trying to pick up a tiny marble with a magnet. If the magnet is too close, it might miss. But if you attach a long, flexible string to the magnet and pull it back, you create a larger "catching zone." By stretching the Velcro strip, the cell gives the splicing machine more room to grab onto the area, even though the actual microexon is tiny.

2. The "Open Road" (RNA Structure)

DNA and RNA aren't just straight strings; they fold up like origami. Sometimes, they fold into tight knots that hide the instructions.

• The Finding: The researchers used chemical probes to see what the RNA looked like. They found that the area where the splicing machine needs to work (between the "Velcro" and the microexon) is unusually open and flat.
• The Analogy: Imagine a busy highway. If the road is blocked by construction or parked cars (knots in the RNA), the delivery trucks (splicing factors) can't get through. But for microexons, the cell keeps that specific stretch of road completely clear and empty. This "open road" allows the splicing machine to slide in easily and do its job without getting stuck.

3. The "Construction Site" Timing

The paper also looked at when this happens in a developing chick brain.

• The Analogy: Think of brain development as a construction site. In the early stages, the site is quiet. But as the brain starts building complex structures (around day 15 to day 27 in the chick), the "foreman" (a protein called SRRM4) shows up.
• The Result: When the foreman arrives, the construction crew starts gluing in these tiny microexons at a rapid pace. The study found that the more the foreman is present, the more these tiny pieces get included. This explains why the brain needs these specific proteins to develop correctly; without them, the "tiny puzzle pieces" get lost, leading to issues like autism or schizophrenia.

What They Didn't Find (The Twist)

The researchers expected that because these microexons are so important, the "folding patterns" (the 3D shape) of the RNA would be identical between humans and chickens, like a universal blueprint.

• The Reality: They found that the shapes were not perfectly identical. The cell doesn't rely on a specific, rigid 3D shape to make this work. Instead, it relies on the flexibility and the open space (the "open road" mentioned above). As long as the road is clear and the Velcro is stretched out, the exact shape of the paper doesn't matter as much.

Why This Matters

This study solves a mystery: How does the cell handle instructions that are too small for its machinery?

The answer is: It creates more space. By stretching out the binding sites and keeping the area clear of knots, the cell ensures that even the tiniest, most critical instructions for building a brain get read and used correctly.

In short: The brain is a complex machine that needs tiny, precise parts. This paper shows that the cell has a special "extension cord" and a "clearance crew" to make sure those tiny parts get installed correctly during development. If this system fails, the brain's wiring gets messed up.

Technical Summary

1. Problem Statement

Microexons are short exons (3–27 nucleotides) highly conserved in vertebrates and critical for neurodevelopment. Their small size presents a fundamental challenge to the canonical exon-definition model of splicing, which typically requires exons of 120–150 nucleotides to allow sufficient space for the assembly of the spliceosome and regulatory proteins (e.g., SRSF1) across the exon.

• The Gap: It is unclear how microexons overcome steric constraints to ensure efficient splicing.
• The Unknown: The specific sequence and RNA structural features that facilitate microexon recognition, particularly in the region between the branchpoint (BP) and the 3' splice site (3'ss), remain largely uncharacterized. Furthermore, it is unknown whether conserved RNA secondary structures exist between orthologous microexons in different species (e.g., humans and chickens).

2. Methodology

The authors employed a multi-faceted approach combining bioinformatics, developmental biology, and experimental RNA structure probing:

• Developmental Model System:
  • Used chick embryos (Gallus gallus) at five developmental stages (HH11, HH15, HH27, HH31, HH35).
  • Collected brain and heart tissues to compare neural vs. non-neural splicing patterns.
  • Performed RNA-seq to quantify splicing changes (PSI values) and expression of splicing factors (SRRM4, NOVA1, etc.).
• Bioinformatic Analysis:
  • Utilized VastDB to identify neural-specific microexons in humans, chickens, and mice.
  • Filtered for high-confidence alternative exons (2–27 nt) and analyzed conservation using phastCons scores.
  • Predicted branchpoint (BP) locations using SVM-BPfinder and analyzed polypyrimidine tract lengths and UGC motifs (SRRM4 binding sites).
  • Performed covariation analysis using multiple sequence alignments (100 vertebrates) to detect conserved RNA structures.
• Experimental RNA Structure Probing:
  • Synthesized in vitro transcripts for 13 selected human and chicken neural microexons (and midexons as controls).
  • Applied SHAPE-MaP (using 5NIA) and DMS-MaP chemical probing to map RNA flexibility and accessibility at single-nucleotide resolution.
  • Validated in vitro findings against in vivo cellular data from medulloblastoma cells.
  • Used RNAstructure and RNAalifold to model secondary structures and calculate base-pairing probabilities.

3. Key Contributions & Results

A. Dynamic Regulation in Chick Development

• Temporal Regulation: A subset of neural microexons (e.g., FRY, AGAP1, CPEB4) undergoes dramatic, stage-specific inclusion changes during chick brain development, specifically between HH15 and HH27.
• Correlation with Regulators: This inclusion spike correlates with the upregulation of known microexon regulators SRRM4 and NOVA1 in the brain (but not the heart).
• Motif Dependency: While many dynamic microexons contain UGC motifs (SRRM4 binding sites), a significant portion of regulated microexons lack these motifs, suggesting alternative or redundant mechanisms for regulation.

B. Sequence Features: Extended Branchpoint-to-Splice Site Distance

• Polypyrimidine Tract Shift: Neural microexons possess unusually extended polypyrimidine tracts located 15–30 nucleotides upstream of the 3' splice site, unlike midexons where these tracts are closer (5–15 nt).
• Increased BP-to-3'ss Distance: Consequently, the predicted distance between the branchpoint and the 3' splice site is significantly longer in neural microexons (median 45–47 nt) compared to midexons (34 nt) in both humans and chickens.
• Mechanism: This extension creates physical space, potentially alleviating steric hindrance for the binding of regulatory factors like SRRM4 and facilitating spliceosome assembly despite the short exon length.

C. Structural Features: Accessibility over Conserved Motifs

• Lack of Global Structural Conservation: Contrary to the hypothesis that RNA structure is strictly conserved via covariation, the study found little evidence of conserved secondary structures between human and chicken orthologs that are independent of sequence similarity. Covariation signals did not consistently match experimental reactivity data.
• Local Structural Similarity: Structural similarity between species is primarily driven by sequence similarity. Blocks of conserved structure exist but are localized and vary in position (upstream, downstream, or within the exon).
• High Accessibility: The most consistent structural feature across neural microexons is high structural accessibility (low base-pairing probability) specifically in the branchpoint-to-3' splice site region. This region is significantly more unpaired and accessible than in midexons.

4. Significance

• Redefining Exon Definition: The study proposes a revised mechanism for microexon splicing. Instead of relying solely on short-range exon definition, microexons utilize extended intronic regions (via shifted polypyrimidine tracts) and structural accessibility to create a "functional distance" that accommodates the spliceosome and regulatory proteins.
• Neurodevelopmental Insight: The findings link the specific splicing dynamics of microexons to critical developmental windows in the brain, mediated by SRRM4 and NOVA1. This provides a mechanistic basis for understanding how misregulation of these factors leads to neurological disorders like autism and schizophrenia.
• Methodological Advance: The work demonstrates that while global RNA structure conservation may be weak in microexons, local structural accessibility is a critical, conserved functional feature. It highlights the utility of the chick embryo model for studying dynamic splicing regulation that is difficult to capture in static human cell lines.

Conclusion

The paper concludes that neural microexons overcome their size constraints not through conserved global RNA secondary structures, but through sequence-driven extensions of the branchpoint-to-3' splice site region and the maintenance of structural accessibility in this critical zone. This architecture likely facilitates the recruitment of splicing factors (like SRRM4) and the assembly of the spliceosome, ensuring robust splicing fidelity during neurodevelopment.