Sequence and structure of protein binding sites in RNA impact biomolecular condensates

This study demonstrates that in *Ashbya gossypii*, the specific sequence and structural context of RNA-protein binding sites, rather than just their number (valence), critically determine the physical properties and biological function of biomolecular condensates.

The Gist

The Big Picture: The Cell's "Cloudy" Neighborhoods

Imagine a cell not as a solid factory, but as a busy city. Inside this city, there are no walls separating different departments. Instead, the cell creates "clouds" or "foggy neighborhoods" called biomolecular condensates. These are liquid-like droplets where specific molecules gather to get work done, like a coffee shop where only people with a specific job title are allowed to sit.

For a long time, scientists thought these clouds were formed mostly by proteins sticking to each other, like magnets. But this paper asks: What role does RNA (the cell's instruction manual) play in forming these clouds?

The Characters: The Protein and the RNA

• The Protein (Whi3): Think of Whi3 as a social butterfly or a club organizer. Its job is to gather things together to form a cloud.
• The RNA (CLN3): Think of CLN3 as a flyer or a membership card. It has specific words printed on it that the club organizer recognizes.
• The "Password" (UGCAU): The club organizer (Whi3) is looking for a specific 5-letter password: UGCAU. If the flyer has this password, the organizer grabs it and starts a party (a condensate).

The Experiment: Changing the Flyer

The researchers found that the CLN3 flyer has five of these passwords scattered across it. They wondered: Does it matter exactly where these passwords are, or is it just about having five of them?

To find out, they played "spot the difference." They took the CLN3 flyer and changed the letters of the passwords one by one, effectively erasing them. They created different versions of the flyer:

1. The Original: Has all 5 passwords.
2. The Mutants: Have 4, 3, or even 0 passwords.
3. The Twist: Some mutants had the same number of passwords as the original, but the passwords were in different spots or slightly different versions.

The Surprising Discoveries

If you think of these condensates like building a tower with LEGO bricks, you might expect that having 5 bricks (passwords) always makes a tower of the same size. This paper proves that is wrong.

Here is what they found:

1. Location, Location, Location
Even when two flyers had the exact same number of passwords (e.g., 4 passwords), they formed clouds that looked and acted completely differently.

• Analogy: Imagine two houses with the same number of front doors. In one house, the doors are all on the front porch, making it easy for guests to enter. In the other, the doors are hidden in the basement or blocked by furniture. Even though the count of doors is the same, the experience of entering the house is totally different.
• Result: Some mutants formed clouds easily (at low concentrations), while others struggled, even though they had the same number of binding sites.

2. The "Folding" Factor
RNA isn't just a straight string of letters; it folds up like a piece of origami. Sometimes, the paper folds in a way that hides the passwords.

• Analogy: Imagine a flyer where the "Password" is printed on the front, but the paper is folded so the password is stuck to the back. The club organizer can't see it!
• Result: The researchers melted the RNA (unfolding the origami) and found that the passwords became accessible again. This proved that the shape of the RNA matters just as much as the sequence of letters.

3. The "Goldilocks" Effect on Cell Cycles
In the real cell (the fungus Ashbya gossypii), these clouds control when the cell divides.

• When the cloud works perfectly, the cell nuclei divide at different times (asynchronously), which is healthy.
• When the researchers messed with the passwords, the clouds became "synchronous" (everyone dividing at once), which messed up the cell's schedule.
• Analogy: It's like a conductor leading an orchestra. If the sheet music (RNA) has the right notes in the right places, the musicians (nuclei) play in a beautiful, staggered rhythm. If you change the notes, everyone starts playing at the exact same time, creating a chaotic mess.

The Main Takeaway

The big lesson from this paper is that it's not just about how many connections you have, but where and how they are arranged.

• Old View: "If you have 5 binding sites, you get a big cloud."
• New View: "If you have 5 binding sites, but they are hidden by folding or placed in a bad spot, you might get a tiny cloud, or no cloud at all."

Why This Matters

This changes how we understand life. It tells us that tiny changes in our genetic code (like a single letter change in a word) can completely change how cells organize themselves, even if the "count" of important parts stays the same. It's a reminder that in biology, context is everything. Just like a word in a sentence can change meaning based on where it is placed, a binding site in RNA changes the cell's behavior based on its surroundings.

Technical Summary

1. Problem Statement

Biomolecular condensates, formed via liquid-liquid phase separation (LLPS), are critical for subcellular organization. While the role of protein multivalency (specifically intrinsically disordered regions) in driving phase separation is well-established, the specific contribution of RNA sequence and structure remains poorly understood.

• The Gap: It is unclear whether the mere number (valence) of RNA binding sites dictates condensate properties, or if the context (primary sequence flanking the motif, local secondary structure, and folding history) plays a decisive role.
• The Model System: The study focuses on the interaction between the RNA-binding protein Whi3 and the G1 cyclin mRNA CLN3 in the multinucleate fungus Ashbya gossypii. Whi3-CLN3 condensates regulate nuclear asynchrony during the cell cycle. Previous work suggested Whi3 binds a UGCAU motif, but the functional equivalence of multiple binding sites within the transcript was unknown.

2. Methodology

The authors employed a multi-scale approach combining in vitro biochemistry, structural biology, and in vivo cellular assays.

• RNA Bind-n-Seq (RBNS):
  • Used to determine the primary sequence preferences of Ashbya Whi3.
  • Tested both randomized 20-nt oligos and tiled 200-nt oligos across the CLN3 transcript.
  • Calculated enrichment scores (R values) to identify high-affinity motifs (UGCAU) and assess the influence of flanking nucleotides.
• Site-Directed Mutagenesis:
  • Generated six CLN3 variants: Wild Type (WT) and mutants where individual Whi3 binding sites (WBS1–5) were mutated (wbs1–5), plus a "no binding site" mutant (nwbs).
  • Mutations were designed to disrupt the UGCAU motif while preserving protein coding sequences where necessary.
• In Vitro Phase Separation Assays:
  • Reconstituted condensates using purified His-tagged Whi3 and in vitro transcribed CLN3 RNAs.
  • Measured Saturation Concentration ($C_{sat}$) to determine the threshold for phase separation.
  • Quantified dense phase composition (protein:RNA ratio) and dynamics (FRAP recovery times).
  • Structural Perturbation Experiments: Compared "unmelted" (co-transcriptionally folded), "melted" (heat-denatured), and "refolded" (slowly re-annealed) RNAs to decouple sequence effects from folding history.
• Structural Analysis:
  • DMS-MaP (Dimethyl Sulfate Mutational Profiling): Probed RNA secondary structure in solution to detect local changes in base-pairing probability (BPP) and global structural shifts.
  • Vienna RNA Predictions: Calculated Minimum Free Energy (MFE) and ensemble diversity.
  • Native Gel Electrophoresis: Assessed RNA dimerization propensity.
• In Vivo Cellular Assays:
  • Integrated mutant CLN3 constructs into A. gossypii strains expressing eGFP-Whi3.
  • Quantified Whi3 puncta formation (condensate number/intensity) via live-cell imaging.
  • Measured cell cycle synchrony by analyzing spindle pole body positioning in adjacent nuclei.
  • Performed RNA FISH to ensure observed phenotypes were not due to changes in transcript abundance.

3. Key Results

A. Sequence Context Dictates Binding Affinity

• RBNS confirmed Whi3 prefers the UGCAU motif. However, flanking nucleotides significantly modulate affinity (e.g., UUGCAU has a much higher R-value than GUGCAU).
• Scoring the CLN3 transcript revealed that not all UGCAU sites are functionally equivalent; their local sequence context creates a gradient of binding strengths.

B. Valence is Not the Sole Determinant of Phase Behavior

• Unexpected $C_{sat}$ Shifts: Contrary to the hypothesis that fewer binding sites would raise the $C_{sat}$ (requiring higher concentrations to condense), several single-site mutants (wbs1, wbs2, wbs4, wbs5) formed condensates at lower protein/RNA concentrations than WT.
• Dense Phase Composition: Mutants exhibited distinct protein:RNA ratios in the dense phase. For instance, wbs5 recruited significantly more RNA to the dense phase than WT, while others recruited less.
• Dynamics: Despite compositional differences, FRAP recovery times for Whi3 were largely unchanged, suggesting the material state (viscosity) remained liquid-like across mutants.

C. RNA Structure and Folding History are Critical

• Global Structure: DMS-MaP and MFE predictions showed no large-scale structural differences between WT and mutants (except nwbs).
• Local Accessibility: However, Base Pairing Probability (BPP) varied significantly between individual binding sites in the WT. Site 1 was more likely to be double-stranded than sites 2 and 4.
• Melting/Refolding Experiments:
  • Melting: Heating RNA to disrupt secondary structure reduced the differences between mutants and WT, suggesting that co-transcriptional folding occludes certain binding sites.
  • Refolding: Slowly refolded RNAs (at 25°C) showed different recruitment patterns compared to unmelted RNAs (folded at 37°C), indicating that folding pathways dictate which binding sites are accessible, thereby altering condensate properties.

D. In Vivo Functional Consequences

• Condensate Formation: Mutants wbs2 and nwbs showed significantly reduced Whi3 puncta formation in cells, correlating with cell cycle defects.
• Cell Cycle Synchrony: Mutants wbs1, wbs2, wbs5, and nwbs caused a shift from asynchronous to synchronous nuclear division, disrupting the normal cell cycle control.
• Decoupling Expression: RNA FISH confirmed that phenotypic changes were not caused by altered transcript levels (e.g., wbs2 had normal RNA levels but failed to form condensates).

4. Significance and Conclusions

• Beyond Simple Valence: The study challenges the prevailing view that condensate properties are determined solely by the number of binding sites (valence). Instead, it demonstrates that the sequence context and local structural accessibility of individual sites are critical determinants of phase behavior.
• Folding History Matters: The "folding history" of an RNA (co-transcriptional vs. post-transcriptional) dictates which binding sites are available for interaction, leading to distinct condensate compositions even with identical sequences.
• Biological Implication: Small, synonymous, or UTR mutations that alter the local sequence or structure of RNA binding sites can disrupt biomolecular condensates and cellular function (e.g., cell cycle asynchrony) without changing the overall number of binding sites.
• Evolutionary Insight: This mechanism suggests that evolution can tune condensate properties and cellular regulation through subtle sequence variations that alter RNA structure and binding site accessibility, rather than just gaining or losing entire motifs.

In summary, Cole et al. establish that the sequence and structural context of RNA binding sites act as a sophisticated regulatory layer in biomolecular condensation, where the specific arrangement of interactions, rather than just their quantity, defines the physical and functional properties of the condensate.