Combination of CG methylation and Chromomethylase 2-mediated RdDM-independent CHH methylation is required for chromosome-specific rRNA gene silencing

This study demonstrates that in *Arabidopsis thaliana*, chromosome-specific silencing of NOR2 rRNA genes relies on a synergistic mechanism combining MET1-mediated CG methylation and CMT2-mediated RdDM-independent CHH methylation, a requirement driven by the predominance of CHH sites in plant rRNA promoters that contrasts with the CG-dependent silencing observed in mammals.

The Gist

The Big Picture: The Cell's Library and the "Do Not Read" Signs

Imagine a cell as a bustling library. Inside this library, there are thousands of copies of the same very important book: the instruction manual for building ribosomes (the machines that make proteins). These books are called rRNA genes.

In the plant Arabidopsis (a small weed often used in science), these books are stored in two specific sections of the library, located on Chromosome 2 and Chromosome 4. Let's call them Section 2 and Section 4.

• Section 4 (NOR4): These books are open and active. The librarians are reading them constantly to build new machines.
• Section 2 (NOR2): These books are shut down and locked away. Even though they are identical to the books in Section 4, the cell decides, "We don't need these right now," and silences them.

The big question the scientists asked was: How does the cell know to lock Section 2 but leave Section 4 open?

The Lock and Key: DNA Methylation

The cell uses a chemical "lock" called DNA methylation to silence genes. Think of methylation as putting a heavy, sticky note on a book that says, "DO NOT READ."

There are three different colors of sticky notes (chemical contexts) the cell can use:

1. CG notes (The classic, well-known lock).
2. CHG notes (A secondary lock).
3. CHH notes (A third type of lock).

For a long time, scientists thought only the CG notes were important for locking these books. They thought the other two colors didn't matter much.

The Experiment: Breaking the Locks

To figure out which lock was the most important, the scientists created "broken" versions of the plant. They removed the workers responsible for putting down the sticky notes:

• One group couldn't make CG notes.
• One group couldn't make CHG notes.
• One group couldn't make CHH notes.
• One group couldn't make any notes at all.

Then, they checked the library to see if the "locked" books in Section 2 started getting read again (which would mean the silencing failed).

The Surprise Discovery:
They found that removing just the CG notes caused the books to open up. But, they also found that removing the CHH notes caused the books to open up just as much!

It turns out that both the CG notes and the CHH notes are essential. If you take away either one, the "Do Not Read" sign fails, and the cell starts making too many ribosomes from the wrong section.

The "Why": The Blueprint of the Locks

Why did the CHH notes matter so much? The scientists looked at the actual blueprint of the rRNA genes (the text of the books) to see where the sticky notes fit best.

They discovered a fascinating pattern:

• In the promoter (the "Start Reading" button at the beginning of the gene), the blueprint is almost entirely made of spots designed for CHH notes. There are very few spots for CG notes.
• In the gene body (the middle of the book), CHH spots are still the most common, followed by CG.

The Analogy:
Imagine the "Start Reading" button is a door with a very specific keyhole.

• In mammals (like humans), the door only has a CG keyhole. So, humans only need CG locks to keep the door shut.
• In plants, the door has mostly CHH keyholes and a few CG ones. If you only put a CG lock on a door that needs a CHH lock, the door won't stay shut!

This explains why the CHH lock is so critical in plants. The cell is trying to lock a door that is mostly designed for CHH keys.

The "CMT2" Worker

The paper also identified a specific worker named CMT2 (Chromomethylase 2).

• This worker is the specialist who puts down the CHH sticky notes.
• Unlike other workers who need a complex "radio signal" (called RdDM) to know where to work, CMT2 works independently.
• The study showed that if you fire CMT2, the CHH notes disappear, the locks break, and the silenced genes wake up.

The Human Comparison

The scientists also looked at human rRNA genes.

• Humans: The "Start Reading" button is full of CG spots. We lost the ability to make CHG and CHH locks millions of years ago. We rely 100% on CG locks.
• Plants: We kept the CHH locks because our "Start Reading" buttons are built differently.

The Takeaway

This paper changes how we understand how plants control their genes. It's not just about one type of lock (CG). It's a team effort.

To keep the "Do Not Read" sign on the rRNA genes in plants, you need two things working together:

1. The classic CG lock (managed by a worker named MET1).
2. The CHH lock (managed by a worker named CMT2).

If you lose either one, the silencing breaks, and the cell gets confused. It's like trying to secure a vault with two different types of deadbolts; if you remove just one, the vault is no longer secure.

In short: Plants use a unique, two-part locking system (CG + CHH) to silence specific genes, a strategy that is very different from how humans do it.

Technical Summary

1. Problem Statement

In Arabidopsis thaliana, ribosomal RNA (rRNA) genes are organized into two Nucleolus Organizer Regions (NORs): NOR2 on chromosome 2 and NOR4 on chromosome 4. While NOR4 genes are predominantly active, NOR2 genes are selectively silenced during development to regulate ribosome dosage.

• Known Mechanism: Cytosine methylation is known to silence rRNA genes, specifically in the CG context (mediated by MET1).
• Knowledge Gap: The relative contributions of the three cytosine methylation contexts (CG, CHG, and CHH) to this selective silencing were unclear. Specifically, it was unknown whether Chromomethylase 2 (CMT2)-mediated, RNA-directed DNA methylation (RdDM)-independent CHH methylation plays a critical role in silencing NOR2, distinct from the RdDM-dependent pathway (mediated by DRM2).
• Evolutionary Context: Mammalian rRNA silencing is driven almost exclusively by CG methylation. Since plants possess plant-specific CMT enzymes that methylate CHG and CHH, the study aimed to determine if these non-CG contexts are essential for plant-specific rRNA regulation.

2. Methodology

The researchers employed a multi-faceted approach combining genetic mutants, biochemical assays, and bioinformatic analysis:

• Genetic Models: They analyzed a panel of Arabidopsis DNA methylation-deficient mutants:
  • met1 (loss of CG methylation)
  • cmt3 (loss of CHG methylation)
  • cmt2 (loss of RdDM-independent CHH methylation)
  • drm2 (loss of RdDM-dependent CHH methylation)
  • ddm1 (loss of global methylation in all contexts due to chromatin remodeling defect)
  • Introgression Line: ColSf-NOR4, containing NOR2 from Col-0 (silenced) and NOR4 from the Sf-2 ecotype (active), to decouple sequence variation from chromosomal location effects.
• Expression Analysis: RT-PCR was performed on 4-week-old leaf tissues to quantify the expression of specific rRNA variants (VAR1, VAR2, VAR3, VAR4). VAR1 is primarily associated with the silenced NOR2.
• Methylation Status Assays:
  • HpaII Digestion: A methylation-sensitive enzyme (cuts unmethylated CCGG) used to distinguish hypomethylated (digested) vs. hypermethylated (undigested) DNA.
  • McrBC Digestion: A methylation-dependent enzyme (cuts methylated C) used to confirm hypermethylation.
  • HindIII CAPS Assay: Used to distinguish between Col-0 and Sf-2 rRNA variants based on restriction sites.
• Bioinformatic Analysis:
  • Heatmaps: Generated from published whole-genome bisulfite sequencing (WGBS) data (Stroud et al., 2013) to visualize methylation levels across the 45S rRNA gene body, promoter, and intergenic spacers (IGS) in various mutants.
  • Cytosine Distribution: Computational analysis of cytosine occurrence in CG, CHG, and CHH contexts within the 45S rRNA gene sequences of Arabidopsis, tomato (Solanum lycopersicum), and humans.

3. Key Results

A. Disruption of Silencing Correlates with CHH and CG Loss

• Expression Data: The cmt2, ddm1, and met1 mutants showed the highest levels of VAR1 (NOR2) reactivation. In contrast, cmt3 (CHG loss) and drm2 (RdDM-dependent CHH loss) mutants showed only modest reactivation.
• Conclusion: Loss of CG methylation (MET1) or RdDM-independent CHH methylation (CMT2) is sufficient to disrupt NOR2 silencing. Loss of CHG or RdDM-dependent CHH is insufficient.

B. Differential Methylation of NOR2 vs. NOR4

• Hypermethylation of NOR2: In wild-type (Col-0), NOR2 genes are hypermethylated in all contexts (CG, CHG, CHH) compared to the hypomethylated, active NOR4 genes.
• Location Independence: In the ColSf-NOR4 introgression line, NOR2 genes (even when carrying Sf-2 sequences) remained hypermethylated and silenced, while NOR4 genes (carrying Col-0 sequences) remained hypomethylated and active. This confirms silencing is determined by chromosomal location/chromatin state, not just DNA sequence.

C. CMT2 is the Primary Driver of rDNA CHH Methylation

• Heatmap Analysis: The cmt2 mutant displayed the most significant loss of CHH methylation across the entire rRNA gene body.
• Promoter vs. Gene Body: While drm2 mutants lost CHH methylation primarily at the promoter and upstream regions, cmt2 mutants lost CHH methylation throughout the gene body.
• Sufficiency of CHH Loss: Despite retaining near-wild-type levels of CG methylation, cmt2 mutants exhibited severe NOR2 derepression. This indicates that CHH methylation loss alone is sufficient to release silencing, and CG methylation alone is not sufficient to maintain it.

D. Cytosine Distribution Explains the Mechanism

• Plant rRNA Genes: Analysis of Arabidopsis and tomato 45S rRNA sequences revealed a striking predominance of CHH sites (approx. 45-50% of cytosines), followed by CG (35%), and the lowest in CHG (20%).
  • Promoters/Spacer Promoters: The gene promoter (GP) and spacer promoters (SP) are overwhelmingly rich in CHH sites (75-83% of cytosines in these regions are CHH).
  • IGS: The Intergenic Spacer (IGS) containing Sal repeats is CG-rich, explaining the importance of MET1.
• Human rRNA Genes: In contrast, human rDNA promoters and gene bodies are dominated by CG sites (~40%), with significantly fewer CHH sites. This correlates with the evolutionary loss of CMT enzymes in mammals.

4. Key Contributions

1. Identification of CMT2's Critical Role: The study establishes that CMT2-mediated, RdDM-independent CHH methylation is a primary mechanism for silencing plant rRNA genes, acting in concert with MET1-mediated CG methylation.
2. Gene Body Methylation Importance: It highlights that CHH methylation in the gene body (not just the promoter) is crucial for maintaining silencing, as evidenced by the severe phenotype of cmt2 (gene body loss) vs. drm2 (promoter loss).
3. Complementary vs. Synergistic Model: The authors propose that MET1 (CG) and CMT2 (CHH) act complementarily. The loss of either pathway is sufficient to break silencing, rather than requiring the loss of both.
4. Evolutionary Divergence: The paper provides a mechanistic explanation for the difference between plant and mammalian rRNA regulation: Plants rely on a combination of CG and CHH methylation due to the high density of CHH sites in their rRNA promoters and bodies, whereas mammals rely solely on CG methylation due to the predominance of CG sites and the absence of CMT enzymes.

5. Significance

• Mechanistic Insight: This work resolves the "missing link" in plant epigenetics regarding how non-CG methylation contributes to the silencing of highly repetitive, heterochromatic loci like NORs.
• Redefining RdDM: It clarifies that while RdDM (DRM2) targets promoters, the maintenance of silencing in heterochromatic rDNA relies heavily on the CMT2 pathway.
• Evolutionary Biology: It offers a compelling example of how the sequence composition of a gene family (high CHH content) drives the evolution of specific epigenetic machinery (CMTs) in plants, contrasting sharply with mammalian regulation.
• Future Directions: The findings suggest that gene body methylation is a critical, perhaps underappreciated, layer of transcriptional control in plants, with implications for understanding transposon silencing and genome stability.