ABSTRACT

Pif-1 proteins are 5′→3′ superfamily 1 (SF1) helicases that in yeast have roles in the maintenance of mitochondrial and nuclear genome stability. The functions and activities of the human enzyme (hPif1) are unclear, but here we describe its DNA binding and DNA remodeling activities. We demonstrate that hPif1 specifically recognizes and unwinds DNA structures resembling putative stalled replication forks. Notably, the enzyme requires both arms of the replication fork-like structure to initiate efficient unwinding of the putative leading replication strand of such substrates. This DNA structure-specific mode of initiation of unwinding is intrinsic to the conserved core helicase domain (hPifHD) that also possesses a strand annealing activity as has been demonstrated for the RecQ family of helicases. The result of hPif1 helicase action at stalled DNA replication forks would generate free 3′ ends and ssDNA that could potentially be used to assist replication restart in conjunction with its strand annealing activity.

A double-strand DNA break was introduced at a specific site within the lacZ gene of plasmid pUC18 using one of several restriction enzymes, and the plasmid exposed to nuclear extracts from human cell lines. Physical rejoining of DNA was monitored by Southern analysis after gel separation, and the fidelity of rejoining by expression of the lacZ gene after bacterial transformation with the treated plasmid. Breaks at the SalI and EcoRI sites were rejoined by extracts to form circular monomers, but the efficiency of rejoining was much higher at the SalI site. Measurement of rejoining at several adjacent sites having different types of termini, consistently showed a range of efficiencies with 5' 4-base greater than 3' 4-base overhangs and 4-base greater than 2-base greater than no overhang. Similar efficiencies were found for nuclear extracts from transformed cell lines, both from a 'normal' individual and an ataxia-telangiectasia (A-T) patient, and from a non-transformed normal cell culture. In contrast at some sites, especially those with a low rejoin efficiency, the fidelity of rejoining was very much lower for the A-T extracts than for normal cell extracts. Mis-rejoining was, however, unrelated to rejoin efficiency at other sites, suggesting that factors such as the exact sequence at the break site on the molecule may also influence the fidelity of rejoining.

DNA molecules carrying a site-specific double-strand break were exposed to nuclear extracts from human cell lines. It was shown previously that breaks could be rejoined correctly by human extracts, but that a proportion of the rejoined molecules had suffered deletions and insertions. The 'mis-rejoined' proportion was higher with cell extracts from an individual with the disorder ataxia-telangiectasia than with normal cell extracts. We now show by sequence analysis that deletions in extract-treated molecules occur exclusively between short direct repeats (2-6 base pairs). A mis-rejoined molecule containing an insertion of 300 bp also had a repeat-based deletion at the same site. A number of different direct repeats are involved; however, some clustering of these occurs especially on the upstream side of the initial breakpoint. These data are most simply interpreted in terms of a model of deletion formation involving single-strand exposure and repair, perhaps with the action of other DNA-metabolising enzymes influencing the frequency with which some repeats are involved.

The nature of mutations occurring in two colorectal carcinoma cell lines deficient in mismatch repair and displaying mutator phenotypes was determined. One of the lines (HCT116) exhibited a higher level of microsatellite instability than the second (DLD-1), although the rate of mutation at the selectable locus encoding the purine salvage enzyme hypoxanthine guanine phosphoribosyl transferase (HPRT) was equally elevated (about 350-450-fold relative to mismatch repair proficient cell lines). Transitions were the major class of mutations in the two mutator lines. In DLD-1 these mutations recurred at several sites that appeared to be hotspots. Frameshifts at a run of six guanine residues in the coding sequence for HPRT constituted 35% of mutations in HCT116. These frameshifts were highly unstable and reverted to wild type at high frequency. Larger deletions were also detected in HCT116. Although these deletions constituted a small proportion of mutations compared with the other types, our data suggest that the rate of deletion is elevated relative to mismatch repair proficient (hMLH1+) cell lines. These observations suggest that the gene(s) altered in DLD-1 may preferentially affect the repair of base mismatches while the alteration(s) in HCT116 may affect the repair of both mismatches and frameshifts.

Cyclin-dependent kinases (CDKs) coordinate cell cycle checkpoints with DNA repair mechanisms that together maintain genome stability. However, the myriad mechanisms that can give rise to genome instability are still to be fully elucidated. Here, we identify CDK18 (PCTAIRE 3) as a novel regulator of genome stability, and show that depletion of CDK18 causes an increase in endogenous DNA damage and chromosomal abnormalities. CDK18-depleted cells accumulate in early S-phase, exhibiting retarded replication fork kinetics and reduced ATR kinase signaling in response to replication stress. Mechanistically, CDK18 interacts with RAD9, RAD17 and TOPBP1, and CDK18-deficiency results in a decrease in both RAD17 and RAD9 chromatin retention in response to replication stress. Importantly, we demonstrate that these phenotypes are rescued by exogenous CDK18 in a kinase-dependent manner. Collectively, these data reveal a rate-limiting role for CDK18 in replication stress signalling and establish it as a novel regulator of genome integrity.

Defining the processes that sustain telomere maintenance is critical to our understanding of cancer and longevity. PIF1 is a nonprocessive 5 0 ->3 0 human DNA helicase that exhibits broad substrate specificity. In vitro studies have implicated PIF1 in maintaining telomeres and processing stalled DNA replication forks, but disruption of the murine Pif1 gene did not yield any apparent phenotype. In this study, we evaluated the function of the PIF1 gene in human cells by using siRNA knockdown strategies to gauge its role in the response to DNA replication stress. We found that PIF1 depletion reduced the survival of both p53-deficient and p53-proficient human tumor cells by triggering apoptosis. In contrast, nonmalignant cells were unaffected by PIF1 depletion. Apoptosis induction in tumor cells was augmented by cotreatment with replication inhibitors (thymidine, hydroxyurea, or gemcitabine). When sensitive PIF1-depleted cells were released from a thymidine-induced S-phase arrest, there remained a subpopulation of cells that failed to enter S-phase. This cell subpopulation displayed an increase in levels of cyclin E and p21, as well as a deficiency in S-phase checkpoint markers that were induced with thymidine in PIF1 expressing cells. Specifically, CHK1 activation was suppressed and we detected no consistent changes in ATM S1981 autophosphorylation, gH2AX induction, or RPA hyperphosphorylation. Death in PIF1-depleted cells was detected in late G 1 /early S-phase and was dependent on caspase-3 activity. Taken together, our findings suggest roles for PIF1 in S-phase entry and progression that are essential to protect human tumor cells from apoptosis. Cancer Res; 71(14); 4998-5008. '2011 AACR.

Unwinding duplex DNA is a critical processing step during replication, repair and transcription. Pif1 are highly conserved non-processive 5′->3′ DNA helicases with well-established roles in maintenance of yeast genome stability. However, the function of the sole member of Pif1 family in humans remains unclear. Human PIF1 is essential for tumour cell viability, particularly during replication stress, but is dispensable in non-cancerous cells and Pif1 deficient mice. Here we report that suppression of PIF1 function slows replication fork rates and increases arrested forks during normal cycling conditions. Importantly, PIF1-dependent replication impediments impair S-phase progression and reduce proliferation rates of RAS oncogene-transformed fibroblasts, where replication fork slowing is exacerbated, but not parental, non-cancerous cells. Disrupted fork movement upon PIF1-depletion does not enhance double-stranded break formation or DNA damage responses but affects resumption of DNA synthesis after prolonged replication inhibitor exposure, accompanied by diminished new origin firing and mainly S-phase entry. Taken together, we characterised a functional role for human PIF1 in DNA replication that becomes important for cell growth under oncogenic stress. Given that oncogenes induce high levels of replication stress during the early stages of tumorigenesis, this function of PIF1 could become critical during cancer development.