Removing GABPA from mouse stem cells within an hour showed that it recruits the Integrator complex to promoters, and that the machinery stalls before chromatin changes at all.

The textbook account of how a gene is switched on divides the job neatly. Transcription factors bind DNA near a gene and recruit the polymerase; the polymerase then travels along the gene making an RNA copy. The factor’s work is at the start, and what happens afterwards is somebody else’s department.

A study in Science Bulletin from a team led by Chunxia Zhang at the Institute of Genetics and Developmental Biology of the Chinese Academy of Sciences finds that one long-studied transcription factor does not stop at the start.

“We were surprised to find that GABPA does not stop working once transcription begins,” Zhang said. “Beyond recognizing DNA and initiating gene expression, it also recruits Integrator to help RNA Pol II move into productive elongation.”

The problem with slow experiments

The finding depended on how the protein was removed. GABPA has been studied for decades, and the reason this role was missed is largely methodological.

Conventional approaches — deleting a gene, or knocking down its RNA over days — are slow relative to the processes they disturb. By the time the protein is gone, the cell has reorganised around its absence: other genes have changed expression, chromatin has shifted, cells may be dying. What is then observed is the aftermath, and separating a direct effect from the cascade that followed becomes guesswork.

The team instead used an acute protein degradation system that removes GABPA within an hour, working in mouse embryonic stem cells. An hour is short enough to catch the cell before it has had time to compensate.

What the first hour showed

Within that first hour, transcription of ribonucleoprotein biogenesis genes dropped markedly. Integrator occupancy at their promoters decreased. Polymerase II progression into gene bodies was impaired, and phosphorylation of its Ser2 site was reduced.

That last detail is a specific readout rather than a general sign of trouble. The largest subunit of polymerase II carries a tail that gets phosphorylated at different positions as transcription proceeds, and Ser2 phosphorylation marks the switch into productive elongation. Reduced Ser2 means polymerases are at genes but not committing to the journey.

The control that makes the result interpretable is what did not change. Chromatin accessibility and active histone marks stayed the same over this early period. Since those are the features that would alter if the problem were a failure to open the gene for business, their stability points at elongation as the primary defect rather than a downstream consequence of a closed-up genome.

Prolonged depletion did eventually produce the broader damage: impaired chromatin accessibility and widespread transcriptional defects. The sequence is the argument. Elongation fails first, and the chromatin problems follow.

Working without its partner

GABPA does not normally act alone. It pairs with GABPB, and the functional unit is usually described as the GABP complex, with GABPA providing DNA binding and GABPB the activation function.

Here the partner turned out to be largely dispensable. Depleting GABPB1 and GABPB2 had only modest effects, in contrast to the rapid loss of cell viability caused by removing GABPA. If the new role required the partner, taking the partner away should have been nearly as damaging.

That asymmetry is the cleanest evidence that this is a distinct activity rather than the known one seen from a new angle. The protein appears to be doing two jobs, only one of which needs its usual collaborator.

What Integrator is

The complex being recruited is Integrator, and its reputation is the opposite of promotional. It carries an RNA endonuclease — an enzyme that cuts RNA — and is best known for terminating transcription early, cleaving nascent transcripts at paused polymerases and sending them for degradation. It has been characterised largely as a brake.

Biochemical, structural and functional analyses in this study show GABPA interacting directly with Integrator’s endonuclease module, built around the subunits INTS4, INTS9 and INTS11.

A factor recruiting a terminator in order to promote elongation sounds contradictory, and the resolution is presumably that Integrator does both depending on context — clearing stalled or unproductive polymerases so that functional ones can proceed. Evidence has been accumulating that the complex has positive roles alongside its destructive one, and a transcription factor that recruits it deliberately to promote elongation fits that revision. The mechanism by which recruitment produces elongation rather than termination is not something this release settles.

The genes that noticed first

The genes most immediately affected were those for ribonucleoprotein biogenesis — the assembly of complexes of RNA and protein, including the machinery that builds ribosomes.

That is a demanding category. Ribosome production consumes an enormous share of a dividing cell’s transcriptional output, and embryonic stem cells divide fast. A defect concentrated there explains why losing GABPA costs viability within hours while losing its partner does not, and it is a reminder that the genes that fail first in such experiments are often the ones with the least slack rather than the ones most specifically regulated.

The caveat is the usual one for stem cell work: mouse embryonic stem cells are a specific and unusual cell type, and a dependency that is acute there may be mild in a cell with slower demands. The release does not state the study’s funding sources.

---

Notes

1. “GABPA recruits Integrator to promote transcription elongation.” Science Bulletin, 2026. DOI: 10.1016/j.scib.2026.07.060.
2. Science China Press. “A new role for a classic transcription factor.” EurekAlert!, 3 August 2026. https://www.eurekalert.org/news-releases/1138643.