The U2AF1 Gene: Function, Mutations, and Disease

U2AF1 is a small but essential gene that encodes part of the molecular machinery cells use to process their genetic instructions, and when it picks up certain mutations, it becomes one of the earliest genetic drivers of blood cancers like myelodysplastic syndromes. The gene’s product helps the cell recognize where to cut and paste segments of newly copied RNA, a step so fundamental that even slight disruptions cascade through hundreds of downstream genes. Those disruptions show up not only in blood cancers but also in lung adenocarcinoma, and research over the past decade has revealed that U2AF1 mutations do considerably more than just scramble RNA splicing.

What U2AF1 Does in a Healthy Cell

Before a gene’s instructions can be used to build a protein, the cell has to edit the initial RNA copy. Segments that do not belong in the final message (introns) need to be snipped out, and the remaining coding pieces (exons) are stitched together. This editing process is called pre-mRNA splicing, and it is carried out by a large molecular machine called the spliceosome. U2AF1 plays a specific role early in that process: it helps the spliceosome find the correct cutting points at the downstream (3′) end of each intron. Working as part of a two-protein partnership, U2AF1 reads a short RNA sequence at the splice site and signals that the spliceosome should assemble there.

Knockout experiments in mice show just how critical this job is. When researchers deleted U2AF1 entirely from blood-forming cells, the animals developed catastrophic bone marrow failure. Stem and progenitor cells in the marrow were virtually wiped out, and the mice became severely pancytopenic, meaning all major blood cell types dropped dramatically. Colony-forming assays confirmed that U2AF1-deficient cells lost the ability to generate myeloid, erythroid, and megakaryocytic lineages.¹ In short, without U2AF1, blood production collapses.

The U2AF1–U2AF2 Partnership

U2AF1 does not work alone. It forms a tight complex with a larger partner protein, U2AF2 (also called U2AF65). Think of U2AF2 as the main RNA-binding engine and U2AF1 as a smaller subunit that fine-tunes how the engine grips the RNA. Single-molecule experiments have shown that when U2AF1 binds to U2AF2, it locks U2AF2 into a particular structural arrangement. Without U2AF1 present, U2AF2 flexes between open and closed shapes, but with U2AF1 attached, nearly all of the U2AF2 molecules settle into the closed conformation.² That structural switch matters because it is part of how the complex senses whether it is sitting on a genuine splice site or a decoy sequence.

When U2AF1 carries its most common cancer-associated mutation, this stabilizing effect is disrupted. Roughly half of the U2AF2 molecules fail to shift into the expected open conformation when RNA is present, suggesting the mutant complex misreads splice-site signals.² The downstream result is that the spliceosome starts choosing wrong splice sites across the genome.

Two Mutation Hotspots With Different Consequences

U2AF1 mutations in cancer are not scattered randomly across the gene. They cluster at two hotspots, both located in zinc-finger domains that contact RNA. The first and more common hotspot sits at amino acid position 34, where serine is most often replaced by phenylalanine (S34F). The second hotspot is at position 157, where glutamine changes to proline or arginine (Q157P or Q157R). In a large study of myeloid neoplasms, S34 variants accounted for about 60% of U2AF1 mutations and Q157 variants for roughly 35%, with the remainder scattered elsewhere.³

These two hotspots are not interchangeable. S34F mutations tended to co-occur with BCOR mutations and abnormal chromosome patterns, while Q157P mutations clustered with ASXL1 and TP53 mutations. Patients with Q157P were more likely to carry additional concurrent mutations overall.³ Mouse studies reinforce this distinction: animals engineered to express S34F showed more severe drops in blood and bone marrow cell counts and worse stem-cell repopulating ability than mice expressing Q157R.⁴ Researchers have argued that lumping the two mutations together in clinical trials may obscure important differences in how they drive disease.

How Mutations Rewire Splicing Decisions

Both hotspot mutations change which splice sites U2AF1 prefers, but they do so by reading RNA differently. The S34F mutation sits in the zinc finger that contacts the nucleotide at position −3 relative to the splice site (three bases upstream of where the exon begins). Normal U2AF1 strongly favors a thymine (T) at that position. The S34F mutant switches that preference toward cytosine (C) or adenine (A).⁵ Because thousands of exons have a T at that position under normal conditions, flipping the preference causes widespread reshuffling of which exons get included or skipped.

The Q157 mutation affects a different zinc finger and alters recognition at a different position within the splice-site sequence, which is why the two mutations produce overlapping but distinct sets of mis-spliced genes. A comparative study of MDS patients carrying mutations in the three most commonly mutated splicing factors (SF3B1, SRSF2, and U2AF1) found that each factor produced a largely unique set of aberrantly spliced genes, with 181 genes mis-spliced specifically in U2AF1-mutant cases.⁶ This gene-level specificity helps explain why different splicing-factor mutations lead to different clinical presentations even though they all disrupt the same general process.

Disrupted Blood Cell Production

The hallmark disease linked to U2AF1 mutations is myelodysplastic syndrome, a group of bone marrow disorders in which blood cells are produced abnormally and often in insufficient numbers. Experiments expressing the S34F mutation in human blood-forming progenitor cells showed that erythroid (red blood cell) differentiation was impaired: fewer red-cell colonies formed, the colonies were smaller, and hemoglobinization was defective. At the same time, the mutation skewed the production of white blood cells, pushing differentiation away from monocytes and toward granulocytes, particularly eosinophils.⁷

Mouse models tell a consistent story. Mice carrying a conditional knock-in of S34F developed persistent reductions in red blood cells, hemoglobin, and white blood cells. White blood cell counts dropped by nearly half compared to controls, with B cells hit hardest. Surprisingly, though, bone marrow cellularity and spleen size remained relatively normal, and the mice did not show the extramedullary hematopoiesis (blood production outside the marrow) that often accompanies severe marrow failure.⁸ Integration of mouse and human RNA-seq data showed that the splicing changes caused by the mutation were enriched in genes involved in RNA processing, ribosome function, and genes recurrently mutated in MDS and acute myeloid leukemia, pointing toward a self-reinforcing loop of dysfunction.⁹

U2AF1 Mutations in Lung Adenocarcinoma

Outside of blood cancers, the place where U2AF1 mutations appear most frequently is lung adenocarcinoma. In a dataset of over 2,100 lung adenocarcinomas, about 2.7% carried U2AF1 mutations, and the S34F hotspot dominated even more than it does in myeloid cancers, accounting for 81% of the lung mutations. This pattern contrasts with other tumor types, where mutations are more evenly split between the S34 and Q157 hotspots.¹⁰

Functionally, S34F in lung cells activates a gene expression program associated with the epithelial-to-mesenchymal transition, a shift that makes tumor cells more invasive. Cells expressing the mutation showed increased invasiveness in laboratory assays, and U2AF1 S34F was significantly co-associated with ROS1 gene fusions, suggesting the two alterations may cooperate in some lung cancers.¹⁰ However, S34F alone appears insufficient to produce tumors. In bronchial epithelial cell models, tumor formation required a cooperating oncogenic KRAS mutation. Cells carrying both U2AF1 S34F and KRAS G12V showed increased proliferation, anchorage-independent growth, and the ability to form tumors in mouse xenografts, while cells with S34F alone did not.¹¹

An intriguing finding from the same study is that cells carrying only S34F showed increased viability when exposed to cigarette smoke concentrate. The mutation appeared to upregulate splicing of stress-granule protein genes, potentially giving precancerous cells a survival advantage under environmental stress. The proposed model is that S34F may act as a permissive early event: it helps cells survive smoking-related damage long enough to pick up additional mutations like KRAS activation, which then synergizes with S34F to drive full transformation.¹¹

R-Loops, DNA Damage, and Genomic Instability

One of the more surprising consequences of U2AF1 mutations is that they destabilize the genome through a mechanism that has nothing to do with protein-coding errors. When splicing goes wrong, incompletely processed RNA can thread back onto the DNA template, forming structures called R-loops (three-stranded complexes of RNA, the template DNA strand, and the displaced non-template strand). Cells expressing U2AF1 S34F accumulate excess R-loops, which stall DNA replication forks and trigger the ATR stress-response pathway.¹²

Independent experiments confirmed R-loop induction genome-wide. Overexpression of normal (wild-type) splicing factors actually suppressed R-loops, but expression of mutant versions, including U2AF1 S34F, increased them. The R-loop buildup preceded the appearance of DNA damage markers, consistent with a model where the R-loops themselves trigger replication stress once cells enter S phase.¹³ This genomic instability may explain why U2AF1-mutant MDS sometimes progresses to acute leukemia: as R-loops cause more DNA breaks, the chance of acquiring additional cancer-driving mutations rises.

Beyond Splicing

U2AF1’s job description turns out to be broader than just splice-site recognition. Researchers discovered that U2AF1 also binds to fully processed, mature mRNA in the cytoplasm, where it acts as a negative regulator of translation. The S34F mutation changes which mRNAs U2AF1 binds, and polysome profiling showed that translation of hundreds of mRNAs was altered in mutant cells.¹⁴ This means that even genes whose splicing is unaffected by the mutation may still have their protein output altered because U2AF1 is mistranslating the instructions at the ribosome level. The finding complicates efforts to map all downstream effects of U2AF1 mutations, since looking only at splicing changes captures just part of the picture.

On the immune signaling front, U2AF1 mutations produce a particularly consequential mis-splicing event in the gene encoding IRAK4, a kinase that sits at the heart of innate immune signaling. Mutant U2AF1 promotes expression of a longer, active isoform of IRAK4 (called IRAK4-L), which drives chronic activation of inflammatory pathways including NF-κB. This creates a persistent inflammatory state in the bone marrow microenvironment.¹⁵ The IRAK4 connection matters for treatment, because IRAK4 inhibitors already exist and could offer a targeted way to dampen the inflammation without broadly suppressing the immune system.

Therapeutic Strategies Under Investigation

Because U2AF1 mutations make cells dependent on a defective spliceosome, one therapeutic approach is to stress the spliceosome further, pushing mutant cells past a functional tipping point while sparing normal cells. A small molecule called NSC 194308 does exactly this: it disrupts the interaction between U2AF2 and RNA, and cells carrying U2AF1 S34F are preferentially killed by the compound. In laboratory models, the selectivity was roughly two- to four-fold, meaning mutant cells died at drug concentrations that left wild-type cells relatively unharmed.¹⁶

The R-loop accumulation described earlier opens a second therapeutic angle. Because mutant cells already have elevated replication stress from R-loops, they become unusually dependent on the ATR kinase to hold their genome together. ATR inhibitors induced DNA damage and cell death specifically in U2AF1 S34F-expressing cells, and combining ATR inhibitors with splicing-modulating compounds enhanced the effect. Critically, overexpression of RNaseH1, an enzyme that degrades R-loops, rescued cells from ATR-inhibitor damage, confirming that the vulnerability stems from R-loop overload rather than some unrelated mechanism.¹²

The DNA damage repair pathway offers yet another vulnerability in lung cancer. U2AF1 S34F causes mis-splicing and downregulation of RAD51, a key protein in homologous recombination repair. Cells carrying the mutation were highly sensitive to ATR inhibitors, and adding even a low dose of a RAD51 inhibitor on top of ATR inhibition dramatically increased DNA damage and cell death.¹⁷ None of these approaches has reached late-stage clinical trials yet, but they illustrate a broader principle: the very splicing defects that help U2AF1-mutant cells become cancerous also create dependencies that can be exploited.

U2AF1 Mutations as a Diagnostic Signal

U2AF1 mutations are increasingly showing up in an unexpected clinical context: liquid biopsies performed on patients being treated for solid tumors. When circulating tumor DNA is sequenced from the blood of cancer patients, U2AF1 variants sometimes surface as an incidental finding, raising the possibility that the patient has an occult blood disorder on top of their solid tumor. Clinical protocols have begun flagging pathogenic U2AF1 variants detected at a variant allele frequency of 10% or higher as a trigger for hematology referral.¹⁸

In healthy individuals, low-level U2AF1 mutations can also appear as part of clonal hematopoiesis, a condition in which a mutant blood stem cell expands to contribute a detectable fraction of circulating blood cells without causing overt disease. Plasma cell-free DNA testing has detected pathogenic U2AF1 variants at high allele frequencies in individuals who turned out to have longstanding, previously unrecognized cytopenias.¹⁹ Because U2AF1 mutations are considered early driver events in myeloid cancer, their detection in otherwise healthy people can serve as an early warning of progression toward MDS.

Evolutionary Hints From a Distant Relative

How fundamental is U2AF1 to complex life? Genomic surveys across the animal kingdom show it is deeply conserved: from humans to fruit flies, the gene’s core structure and zinc-finger domains are recognizable. An unusual twist was found in the larvacean tunicate Fritillaria borealis, a tiny marine animal with one of the most compact genomes among multicellular organisms. This species has duplicated both U2AF1 and its partner U2AF2, producing α and β copies of each. The α copies closely resemble the single-copy genes found in other animals, while the β copies have diverged substantially outside the RNA-binding domains, hinting that they may have taken on specialized splicing roles.²⁰ This natural experiment underscores how tightly U2AF1’s function is constrained: even when the gene duplicates, the RNA-binding regions stay nearly unchanged, because altering them disrupts splice-site recognition in precisely the way that cancer-associated mutations do in humans.