CDK12 and CDK13 suppress distinct intronic polyadenylation sites

This study demonstrates that CDK12 and CDK13 play distinct roles in suppressing intronic polyadenylation to maintain transcript integrity, and reveals that the resulting prematurely terminated transcripts are translationally competent, generating novel intronic peptides that could serve as tumor-specific neoantigens in CDK12/13-deficient cancers.

The Gist

Imagine your body's cells are like massive, busy factories. Inside these factories, there is a master blueprint (DNA) that tells the workers how to build proteins, which are the machines and tools the cell needs to function. To read this blueprint, the factory uses a specialized scanner called RNA Polymerase II. This scanner moves along the blueprint, copying the instructions into a working draft.

However, this copying process is delicate. Sometimes, the scanner gets confused and stops too early, cutting the draft short before it's finished. If this happens inside a section of the blueprint that shouldn't be the end (an "intron"), it creates a broken, useless instruction manual. This mistake is called Intronic Polyadenylation (IPA).

The Guardians: CDK12 and CDK13

In a healthy factory, two very specific supervisors, named CDK12 and CDK13, keep the scanner moving smoothly. They act like "speed coaches" for the copying machine, ensuring it doesn't stop prematurely. They also make sure the draft is edited correctly as it's being written.

When these supervisors are missing or broken (a "loss-of-function" mutation), the scanner gets confused. It stops too early in the middle of the blueprint. This happens often in certain cancers, like ovarian cancer.

The Experiment: A "Switch" in the Factory

The researchers wanted to figure out exactly what CDK12 and CDK13 do differently. To do this, they used a special chemical "brake" called THZ531 that stops both supervisors.

But here's the clever part: they also created a special version of the factory (using CRISPR gene editing) where the CDK12 supervisor had a tiny, invisible "shield" (a specific mutation). This shield made CDK12 immune to the brake, while CDK13 was still stopped.

By using different amounts of the brake, they could see:

1. What happens when both supervisors are stopped.
2. What happens when only CDK13 is stopped (because CDK12 is shielded).

This allowed them to see that CDK12 and CDK13 are like two different types of safety nets; they catch different types of mistakes and stop the scanner from halting at different specific spots on the blueprint.

The Discovery: Broken Drafts Become New Tools

When the supervisors were stopped, the scanner produced thousands of these "broken drafts" (prematurely stopped transcripts). The researchers used a high-tech long-read camera (Oxford Nanopore) to take pictures of these drafts.

They found something surprising:

• Truncated Proteins: Some of these broken drafts were missing important parts, like a car missing its engine.
• New Peptides: But other broken drafts weren't just trash. Because they stopped in a weird spot, they created brand new, short protein sequences (peptides) that the factory had never made before. These were like "glitch art"—accidental new tools created from the mistake.

The Proof: The Factory Actually Uses the Glitches

To prove these new, glitchy sequences weren't just paper scraps, the researchers looked at the actual tools floating around in the factory (using mass spectrometry). They found physical evidence that the factory was building these new, intron-derived peptides.

The Big Picture

This study shows that:

1. CDK12 and CDK13 are distinct guardians that prevent the factory from stopping too early.
2. When they fail, the factory doesn't just stop; it starts making a flood of new, strange, short protein pieces from the middle of the blueprint.
3. These strange new pieces are real and can be found in the cell.

The paper suggests that because these new pieces are unique to the cancer cells (they don't exist in healthy cells), they could be used as a "flag" to help the body's immune system recognize and attack the cancer. The researchers call this a "therapeutic vulnerability," meaning the cancer's own mistake creates a weakness that could be exploited.

Technical Summary

Technical Summary

Problem Statement
Cyclin-dependent kinases CDK12 and CDK13 function as RNA polymerase II C-terminal domain Ser2 kinases, playing critical roles in transcriptional elongation and co-transcriptional RNA processing. Loss-of-function (LOF) mutations in these kinases are frequently observed in various cancers and are mechanistically linked to increased intronic polyadenylation (IPA). This phenomenon leads to prematurely terminated transcripts, which can alter coding potential and generate novel intronic peptides. However, the distinct contributions of CDK12 versus CDK13 in suppressing early transcriptional termination and the specific consequences of IPA on the proteome remain to be fully characterized.

Methodology
To dissect the distinct roles of CDK12 and CDK13, the study utilized OVCAR4 ovarian cancer cells and a dual CDK12/13 inhibitor, THZ531. A CRISPR-engineered CDK12 C1039S mutation was introduced to confer resistance to THZ531, thereby enabling the functional differentiation between CDK12 and CDK13 loss-of-function conditions. The researchers employed Poly(A) Click-seq (PACseq) across a six-point inhibitor dose curve to map thousands of IPA sites and classify them based on their dose-response behaviors into CDK12-dependent and CDK13-dependent subsets.

To validate these findings and assess transcript integrity, Oxford Nanopore long-read sequencing was utilized to confirm premature termination events and identify full-length isoform switches. Furthermore, the study integrated long-read transcriptome data with mass spectrometry to detect and confirm the translation of peptides derived from IPA isoforms.

Key Results
The study identified thousands of IPA sites exhibiting distinct dose-response profiles, allowing for the classification of specific subsets dependent on either CDK12 or CDK13. Long-read sequencing confirmed that the inhibition of these kinases results in premature transcriptional termination, leading to the truncation of protein domains and the production of novel, intronically encoded peptide sequences. Crucially, the integration of transcriptomic and proteomic data provided direct evidence that these intronic sequences are translationally competent, resulting in the production of peptides derived from IPA isoforms.

Significance and Claims
The paper establishes distinct roles for CDK12 and CDK13 in maintaining transcript integrity, demonstrating that their loss leads to specific patterns of early transcriptional termination. The findings confirm that IPA transcripts are not merely abundant but are also capable of being translated into novel peptide sequences. The authors posit a mechanistic link between transcriptional dysregulation in CDK12/13-deficient cancers and the generation of these novel peptide isoforms. Finally, the study highlights a potential therapeutic vulnerability, suggesting that these IPA-derived peptides could be exploited as tumor-specific neoantigens.