One protein, two opposing fork outcomes
The paper's central claim is that PrimPol does not have a single job under oncogenic KRAS. It executes two distinct replication stress tolerance modes, and which mode is active depends on how long the stress has lasted. During acute KRAS-induced heterochromatin-associated replication stress (K-HaRS), PrimPol cooperates with SMARCAL1 to slow fork progression. During adapted or chronic K-HaRS, an ATR-dependent switch makes phosphorylated PrimPol promote repriming instead, and fork slowing is attenuated.
These two modes pull in opposite directions on the same measurement. Fork slowing means the replication fork moves more slowly; repriming means the fork restarts DNA synthesis downstream of a blocking lesion, leaving single-stranded DNA behind. The paper argues that cells need the first mode to stabilize challenged forks early and the second to keep replication moving across heterochromatin barriers later.
The switch is not explained by changing protein abundance. Total PrimPol and SMARCAL1 levels were broadly comparable across the cell models representing acute and adapted/chronic K-HaRS states, so the authors attribute the functional transition to regulation of PrimPol rather than to more or less PrimPol protein.
KRAS builds a heterochromatin barrier to forks
The study uses epithelial cell systems in which KRAS G12V expression can be switched on with 4-hydroxytamoxifen. In immortalized human small airway epithelial cells (SAEC), acute KRAS G12V expression reduced fork speed by approximately 30% in a DNA fiber assay, which tracks labeled replication tracts to measure how fast forks advance.
Three mechanistically different treatments restored fork speed and point to the source of the stress. Inhibiting RNA polymerase II elongation with DRB, inhibiting EZH2 with GSK126 to suppress PRC2-mediated maintenance of H3K27me3, and depleting HP1α each rescued fork progression. The authors read this as evidence that oncogenic KRAS generates replication stress through transcription-dependent facultative heterochromatin, the K-HaRS state.
The same pattern appeared when KRAS G12V was induced in KRAS-wild-type lung adenocarcinoma lines H1819 and H322, and in RPE cells. That matters for scope: active fork slowing in response to this heterochromatin-associated stress is not confined to one cell type, at least among the immortalized epithelial models tested.
PrimPol and SMARCAL1 slow forks together acutely
During acute K-HaRS, depleting PrimPol or SMARCAL1 reversed the KRAS-induced fork slowing, and the two proteins could be co-immunoprecipitated in SAEC undergoing acute K-HaRS. The authors interpret this as a PrimPol-SMARCAL1 fork-slowing configuration that is available in normal epithelial cells under acute stress.
The functional consequence is a limit on error-prone synthesis. When this fork-slowing mode is compromised during acute K-HaRS, the paper reports that Rev1-Polζ-dependent synthesis maintains fork speed. In other words, the slow-fork mode is not passive damage; it is a strategy that keeps the fork stable while avoiding a switch to a more mutagenic bypass pathway.
This is the paper's most counterintuitive point. PrimPol is usually discussed as a repriming enzyme that helps forks restart, yet here its unphosphorylated form is required for active slowing. The authors frame this as a pathway-choice problem: cells choose among fork slowing, repriming, and translesion synthesis rather than relying on one tolerance mechanism.
ATR phosphorylation flips PrimPol toward repriming
As K-HaRS persists, the balance shifts. The paper links the repriming mode to phosphorylation of PrimPol at Ser255. Because endogenous Ser255 phosphorylation was hard to detect by conventional immunoblotting, likely reflecting low stoichiometry or transient signal, the authors generated a phospho-specific antibody and used a proximity ligation assay (PLA) to detect the modification in cells.
PrimPol Ser255 phosphorylation-associated PLA signals were readily detected in the KRAS-driven lung adenocarcinoma lines H358 and H2009, and these signals fell after ATR inhibition or PrimPol depletion. Signals were limited in SAEC and in non-KRAS-driven H1819 and H322 cells. The authors conclude that this phosphorylation mark is enriched in KRAS-driven, repriming-active cells and depends on ATR pathway activity.
Consistent with a switch, ATR inhibition restored detectable PrimPol-SMARCAL1 association in H358 cells. Expressing the non-phosphorylatable PrimPol S255A mutant suppressed the elevated fork speed seen when PrimPol depletion was combined with ATR inhibition, and depleting SMARCAL1 alleviated that slowing. The paper reads this as ATR inhibition re-engaging a PrimPol-SMARCAL1-dependent fork-slowing mode in KRAS-driven cancer cells.
Global ATR activity does not define the mode
A reader could easily conflate general replication stress response activation with the specific PrimPol repriming state. The paper separates them. Rad17 Ser645 and Chk1 Ser296/Ser345 phosphorylation were more prominent in established lung cancer lines than in SAEC regardless of KRAS mutation status, consistent with cancer cells experiencing elevated endogenous replication stress and broadly engaging ATR-Chk1 signaling.
ATR Thr1989 phosphorylation, however, was more prominent in the adapted RSTC model and in the KRAS-driven H358 and H2009 cells that show PrimPol-dependent repriming. The authors therefore state that global ATR/RSR activation is associated with, but does not fully define, the PrimPol-dependent repriming state.
This distinction is the practical core of the paper. A marker of general checkpoint activation is not the same measurement as the PrimPol Ser255 phosphorylation-associated signal, and the two should not be used interchangeably when judging whether a cell is in the repriming mode.
Losing PrimPol pushes cells onto Polζ
When ATR-dependent repriming and the PrimPol-SMARCAL1 fork-slowing axis are both compromised in KRAS-driven cancer cells, the paper reports that fork progression shifts to Rev1-Polζ-dependent synthesis. In H358 cells treated with ATR inhibitor and depleted of SMARCAL1, restored fork progression was restrained by Rev1 inhibition or depletion and by Rev3 depletion.
Loss of PrimPol itself enforces chronic reliance on this error-prone Rev1-Polζ pathway and compromises cellular fitness under KRAS-driven stress. The authors connect this to their earlier observation that loss of PrimPol compromises oncogenic KRAS-driven proliferation, proposing that chronic Polζ dependence would otherwise excessively fuel mutagenesis and impair fitness.
The paper also examines human tumor mutation data. In a TCGA-LUAD cohort, it compares mutation numbers and single-base substitution spectra between groups classified as repriming-prone or Polζ-prone, and extracts mutational signatures with cosine similarity thresholds. These analyses are correlative groupings of tumor data, not a direct demonstration that the switch causes the observed mutation patterns in patients.
Why the timing of the switch matters
The paper's distinctive claim is that the same protein must do opposite things at different stages. Acute K-HaRS needs PrimPol-SMARCAL1-dependent fork slowing to stabilize challenged forks and limit error-prone synthesis. Adapted or chronic K-HaRS needs ATR-dependent PrimPol repriming to keep replication moving across heterochromatin barriers.
The authors describe the outcome as a compromise rather than pure protection. ATR and PrimPol fine-tune replication stress responses to create a state of genome instability that is tolerable enough for precancerous cells to survive oncogenic stress, diversify, and evolve. That framing places the value of the switch in the balance it strikes, not in maximizing fork speed or minimizing mutations.
The scope conditions are explicit. The work is built on epithelial cell models, including SAEC, RPE, and lung adenocarcinoma lines, with KRAS G12V induction and K-HaRS driven by H3K27me3-enriched heterochromatin. The authors note that oncogenic RAS in non-epithelial cells typically produces R-loop-associated transcription-replication conflicts handled by topoisomerases, so the ATR-PrimPol pairing described here is tied to this epithelial, heterochromatin-associated stress context.
The paper also leaves mechanism open. ATR-mediated phosphorylation of SMARCAL1 at Ser652 has been reported to negatively regulate its activity, so ATR may attenuate fork slowing through SMARCAL1 as well, but the precise contribution of that phosphorylation to the fork-slowing versus repriming balance during K-HaRS tolerance remains to be determined. The most useful takeaway is therefore conditional: when ATR-dependent repriming dominates, blocking ATR re-engages PrimPol-SMARCAL1 fork slowing, and if that slowing is also lost, replication falls back on Rev1-Polζ. That ordering, rather than any single marker, is what the paper's evidence supports for judging which tolerance mode a KRAS-driven cell is using.
Sources used in this answer
PrimPol toggled by ATR modulates fork slowing and repriming under oncogenic KRAS-induced replication stress
PrimPol executes two distinct replication stress tolerance modes under oncogenic KRAS: unphosphorylated PrimPol cooperates with SMARCAL1 to slow forks during acute K-HaRS, while ATR-dependent Ser255 phosphorylation shifts it to repriming during chronic K-HaRS, and loss of PrimPol enforces error-prone Rev1-Polζ dependence.
