PSF3 recruits DNA polymerase theta (Polθ) to stalled replication forks to promote DNA repair and survival of myeloid malignant cells. Free

DNA replication is constantly challenged by DNA damage and other intrinsic and extrinsic stresses collectively termed replication stress. When replication forks encounter DNA damage, they can stall or even collapse to generate DNA double-strand breaks (DSBs), triggering a DNA damage response, and potentially leading to cell death if not adequately repaired.

FLT3(ITD)-positive AML cells and JAK2(V617F)-positive MPN cells, in comparison to normal cells, accumulate high numbers of stalled/collapsed replication forks due to elevated replication stress caused by metabolic alterations and induced by the treatment. We reported that DNA-polymerase theta (Polθ)-dependent microhomology-mediated end-joining (TMEJ) might play a pivotal role in the repair of DSBs arising from stalled replication forks, but the mechanistic aspect of this process is not known.

The CMG (Cdc45-MCM-GINS) complex is the eukaryotic replicative helicase, the enzyme that unwinds double-stranded DNA ahead of the moving replication fork. Eukaryotic GINS complex (contains the four subunits, PSF1, 2, 3, and 4) also links with other key proteins at the fork to maintain an active replisome progression complex. Here, we report that PSF3 interacts directly with Polθ and recruits it to stalled replication forks to maintain genomic integrity by promoting a fork restart by TMEJ-mediated repair of replication fork-associated DSBs.

PSF3 was detected in Polθ pulldown followed by LC-MS/MS. Co-immunoprecipitations form cell lysates, of the recombinant purified proteins, and immunofluorescent co-localization assays demonstrated that Polθ specifically interacts directly with PSF3. In silico analysis followed by site-directed mutagenesis revealed that Polθ forms a “claw-like structure” to engage the C-terminal portion of PSF3.

Polθ, which is recruited to the stalled replication forks by PSF3, exerts TMEJ to repair the breaks resulting from collapsed forks to restart the forks. Based on AlphaFold2 prediction we generated PSF3 mutant (PSF3mut) which lost its capability to interact with Polθ but retained the interaction with other GINS proteins (PSF1, PSF2 and PSF4). Expression of the PSF3mut mutant induced DSBs and abrogated survival of FLT3(ITD) and JAK2(V617F)-positive AML and MPN cells, respectively, but not the parental counterparts.

In conclusion, PSF3 – Polθ interaction promotes the proliferation of AML and MPN cells. Overall, this study highlights an important mechanism of PSF3-mediated recruitment of Polθ to the stalled replication forks to maintain Polθ-dependent fork protection in leukemia cells. In addition, our studies indicate that Polθ inhibitors may selectively target AML and MPN cells displaying enhanced replication stress.