A sequence-encoded promoter proximal super pause stabilizes an offline RNA polymerase II state

By developing the GATO-seq assay and utilizing cryo-EM, this study reveals that a specific nucleic acid sequence induces a unique, TFIIIS-resistant "super pause" in RNA polymerase II, which stabilizes a reversible, single-nucleotide backtracked "sidetracked" state through a threonine-lined pocket, thereby demonstrating how sequence directly encodes transcriptional pausing control.

The Gist

The Big Picture: The Great Gene Assembly Line

Imagine your cells are massive, bustling factories. Inside these factories, there is a critical assembly line that builds the instructions (proteins) needed to keep you alive. The machine doing the heavy lifting is called RNA Polymerase II (let's call it the "Builder Bot").

Usually, the Builder Bot moves smoothly down a DNA track, reading the instructions and building the product. However, sometimes it stops. It hits a "pause button." In the past, scientists thought these stops were mostly caused by traffic cops (proteins) telling the bot to wait. This paper reveals that sometimes, the road itself is the problem. Certain stretches of the DNA track are so bumpy or tricky that the Builder Bot gets stuck, and no amount of traffic control can easily get it moving again.

The New Tool: GATO-seq (The High-Speed Camera)

Before this study, watching the Builder Bot pause was like trying to film a race car from a blurry, slow-motion video taken from a moving train. You could see it was fast, but you couldn't see exactly where it stalled or why.

The researchers invented a new tool called GATO-seq.

• The Analogy: Imagine you have a library of 1,000 different DNA tracks. Instead of watching one car at a time, GATO-seq launches 1,000 Builder Bots onto these tracks simultaneously.
• The Magic: It uses a special camera (Direct RNA Sequencing) that takes a snapshot of exactly where every single bot is at specific moments in time. This lets the scientists see the "traffic jams" in high definition, down to the single letter of the DNA code.

The Discovery: The "Super Pause"

Using this high-speed camera, the team found something amazing. They discovered a specific DNA sequence that acts like a magnetic trap for the Builder Bot.

• The "Super Pause": When the Builder Bot hits this specific sequence, it doesn't just slow down; it gets stuck in a deep, sticky state.
• The Rescue Failure: Normally, if a bot gets stuck, a "rescue factor" called TFIIS (think of it as a tow truck) comes along, pushes the bot, and gets it moving again.
• The Twist: The "Super Pause" is so strong that the tow truck (TFIIS) cannot move the bot. It's like the bot has locked its wheels and the tow truck's hook won't even fit. The bot is stuck in an "offline" state.

The Secret Mechanism: The "Sidetracked" State

To understand why the bot was stuck, the scientists used a super-powerful microscope called Cryo-EM (which takes 3D pictures of molecules at near-atomic resolution).

They found the Builder Bot wasn't just "paused"; it had fallen into a weird, hidden position they call "Sidetracked."

• The Analogy: Imagine the Builder Bot is a train on a track. Usually, if it gets stuck, it slides backward a few inches (backtracking). But on this "Super Pause" track, the train slides backward just one single rail tie and then gets wedged into a tiny, custom-shaped pocket in the side of the tunnel.
• The Pocket: This pocket is lined with specific amino acids (like a velvet lining) that hold the bot perfectly still. Because the bot is wedged in this specific "sidetracked" angle, the tow truck (TFIIS) physically cannot reach the bot to pull it out. The bot is safe, but it's offline.

Why Does This Matter?

You might ask, "Why would a cell want a machine to get stuck in a way that can't be fixed?"

1. Quality Control: It acts as a strict checkpoint. If the DNA instructions are too messy or the conditions aren't right, the bot gets stuck in this "Super Pause." This prevents the factory from building bad products.
2. Decision Time: It gives the cell time to decide: "Do we want to finish this job?"
   • If the answer is YES, other powerful factors (like P-TEFb) can come in and force the bot out of the pocket to finish the job.
   • If the answer is NO, the stuck bot might be removed entirely, stopping the production of that specific protein.

The Takeaway

This paper changes how we think about gene regulation. We used to think genes were controlled mostly by external "traffic cops" (proteins). Now we know that the DNA code itself contains hidden "speed bumps" and "magnetic traps" that can freeze the machinery in a unique, unfixable state.

In short: The DNA track isn't just a passive road; it's an active participant that can grab the Builder Bot, lock it in a secret pocket, and force the cell to pause and think before proceeding.

Technical Summary

1. Problem Statement

Promoter-proximal pausing of RNA Polymerase II (Pol II) is a critical regulatory checkpoint in eukaryotic gene expression, controlled by protein factors (e.g., DSIF, NELF, P-TEFb) and DNA sequence. However, the specific contributions of nucleic acid sequence versus protein factors remain incompletely understood due to limitations in existing methodologies:

• Cell-based assays (NET-seq, PRO-seq): Operate under steady-state conditions with limited temporal resolution, making it difficult to measure pause duration, frequency, or the precise kinetic impact of individual factors.
• Single-molecule/biochemical assays: Often use non-physiological nucleotide concentrations and are restricted to a handful of model sequences, failing to capture the diversity of pausing behavior across the genome.
• Gap: There is a lack of high-throughput, reconstituted systems that can simultaneously analyze thousands of sequences under controlled, physiologically relevant conditions to define sequence-encoded pausing rules.

2. Methodology: GATO-seq

The authors developed GATO-seq (Gene-specific Analysis of Transcriptional Output), a high-throughput, reconstituted transcription assay combining direct RNA sequencing with massively parallel library synthesis.

• Library Design: A library of 1,000 human gene promoters (first 262 bp) was fused downstream of an Adenovirus Major Late (AdML) promoter and a G-less cassette. This standardized the initiation context while varying the elongation sequence.
• Reconstitution: Pre-initiation complexes (PICs) were assembled using highly purified porcine Pol II and human general transcription factors (TFIIA, TFIIB, TFIIE, TFIIF, TFIIH, TBP, CAK).
• Kinetic Control: Transcription was initiated, and elongation was allowed to proceed for specific time points (1 min, 5 min) before quenching. Reactions were performed with varying NTP concentrations (low vs. physiological 1 mM) and in the presence/absence of elongation factors (DSIF, NELF, TFIIS, P-TEFb).
• Sequencing: RNA products were polyadenylated and sequenced using Oxford Nanopore Direct RNA sequencing. This allowed for the mapping of the 3' ends of nascent transcripts at single-nucleotide resolution without PCR bias.
• Analysis: Signal intensity at each nucleotide position was normalized to calculate elongation rates, pause efficiency ($E_{max}$), and pause half-life ($t_{1/2}$).

3. Key Contributions

1. Development of GATO-seq: A scalable, reconstituted platform enabling the study of transcription kinetics across thousands of sequences with temporal resolution.
2. Discovery of the "Super Pause": Identification of a specific, conserved DNA consensus sequence that induces an exceptionally stable pause ("super pause") that is refractory to rescue by the elongation factor TFIIS.
3. Structural Insight ("Sidetracking"): Determination of the first cryo-EM structure of Pol II in a novel, reversible, single-nucleotide backtracked state termed "sidetracked," which explains the TFIIS insensitivity of the super pause.
4. Mechanistic Decoupling: Demonstration that DNA sequence alone can encode "offline" states that trap Pol II, independent of chromatin or complex cellular machinery.

4. Key Results

A. GATO-seq Characterizes Pausing Kinetics

• Velocity: Pol II elongation rates were measured at ~25.6 bp/min (low NTP) and ~101 bp/min (physiological NTP). Addition of DSIF-NELF slowed elongation by ~2-fold.
• Pause Dynamics: The assay identified thousands of pause sites. While DSIF-NELF increased the number of detectable pauses, it surprisingly decreased the median pause half-life ($t_{1/2}$), suggesting these factors facilitate rapid release from weak pauses rather than stabilizing long arrests.
• TFIIS Sensitivity: TFIIS rescued ~60% of pause sites (TFIIS-sensitive), shifting the 3' ends or reducing signal. However, a distinct subset of sites (approx. 8.9%) remained unchanged in position and signal intensity upon TFIIS addition (TFIIS-insensitive).

B. The "Super Pause" Consensus Sequence

• Sequence Motif: TFIIS-insensitive sites shared a conserved consensus: R-13...R-8G-7...M-10...C-2Y-1G+1C+2C+3. Key features include a purine-rich upstream region, a specific G at -8, a C at -2, and crucially, an incoming G (+1) followed by two Cytosines (+2, +3).
• Validation: Synthetic "super pause" sequences recapitulated the GATO-seq findings: they induced robust pausing lasting >10 minutes, were unaffected by high NTP concentrations, and were insensitive to TFIIS.
• Mutational Analysis: Substitution of the +1 G to U abolished the pause, while mutation of the -2 C to U rendered the pause TFIIS-sensitive, confirming the sequence specificity.

C. Structural Basis: The "Sidetracked" State

• Cryo-EM Structures: Structures of Pol II on the super pause sequence revealed three states: pre-translocated, post-translocated, and a novel sidetracked state.
• Sidetracked Mechanism:
  • The RNA 3' end is backtracked by exactly one nucleotide.
  • Unlike canonical backtracking (where RNA extrudes through the pore), the sidetracked RNA is bent ~50° and stabilized by a "threonine pocket" formed by RPB1 residues T850, T854, and trigger loop residues T1100, T1103.
  • This pocket specifically accommodates pyrimidines (like the C at -2) but sterically clashes with purines.
  • TFIIS Resistance: The position of the backtracked base in the sidetracked state physically clashes with TFIIS Domain III, preventing TFIIS from binding and stimulating RNA cleavage.
• Reversibility: Pyrophosphorolysis and activation by P-TEFb (forming the EC* complex) demonstrated that the super pause is a stable but reversible offline state, not a permanent arrest.

D. Genomic Relevance

• Scanning the human genome revealed ~4,455 hypothetical super pause sites in promoter-proximal regions of expressed genes.
• These sites are enriched in genes involved in development and stimulus response.
• ChIP-exo data showed focused Pol II accumulation at these sites, whereas PRO-seq (which requires nucleotide incorporation) failed to detect them, consistent with the "offline" nature of the sidetracked state.

5. Significance

• Sequence-Encoded Regulation: The study proves that specific DNA sequences can encode "offline" states that trap Pol II, acting as a regulatory layer independent of, or competitive with, protein factor-mediated pausing.
• New Structural State: The discovery of the "sidetracked" state expands the known conformational landscape of Pol II, revealing a mechanism for sequence-specific, TFIIS-resistant pausing.
• Regulatory Implications: The super pause may function as a quality control checkpoint or a timing mechanism. If DSIF/NELF fail to associate, Pol II may enter this sidetracked state, potentially leading to premature termination or attenuation, thereby preventing aberrant transcription of specific gene classes.
• Methodological Impact: GATO-seq provides a powerful, modular tool for dissecting transcription regulation, capable of testing the effects of any sequence, factor, or condition in a controlled, high-throughput manner.

In conclusion, Vazquez Nunez et al. have bridged the gap between sequence, structure, and function in transcription regulation, identifying a novel, sequence-encoded "super pause" that stabilizes a unique, TFIIS-resistant Pol II conformation.