Okazaki Fragment Processing-independent Role for Human Dna2 Enzyme during DNA Replication*

Background: Dna2 and FEN1 putatively cooperate in Okazaki fragment (OF) processing in yeast. Results: FEN1 depletion leads to OF maturation defects that do not impact replication fork kinetics, whereas hDna2 depletion leads to DNA damage independent of OF maturation. Conclusion: hDna2 participates in DNA replication in a manner distinct from FEN1. Significance: hDna2 participates in replication in a FEN1, OF maturation-independent manner. Dna2 is an essential helicase/nuclease that is postulated to cleave long DNA flaps that escape FEN1 activity during Okazaki fragment (OF) maturation in yeast. We previously demonstrated that the human Dna2 orthologue (hDna2) localizes to the nucleus and contributes to genomic stability. Here we investigated the role hDna2 plays in DNA replication. We show that Dna2 associates with the replisome protein And-1 in a cell cycle-dependent manner. Depletion of hDna2 resulted in S/G2 phase-specific DNA damage as evidenced by increased γ-H2AX, replication protein A foci, and Chk1 kinase phosphorylation, a readout for activation of the ATR-mediated S phase checkpoint. In addition, we observed reduced origin firing in hDna2-depleted cells consistent with Chk1 activation. We next examined the impact of hDna2 on OF maturation and replication fork progression in human cells. As expected, FEN1 depletion led to a significant reduction in OF maturation. Strikingly, the reduction in OF maturation had no impact on replication fork progression, indicating that fork movement is not tightly coupled to lagging strand maturation. Analysis of hDna2-depleted cells failed to reveal a defect in OF maturation or replication fork progression. Prior work in yeast demonstrated that ectopic expression of FEN1 rescues Dna2 defects. In contrast, we found that FEN1 expression in hDna2-depleted cells failed to rescue genomic instability. These findings suggest that the genomic instability observed in hDna2-depleted cells does not arise from defective OF maturation and that hDna2 plays a role in DNA replication that is distinct from FEN1 and OF maturation.

Dna2 is an essential helicase/nuclease that is postulated to cleave long DNA flaps that escape FEN1 activity during Okazaki fragment (OF) maturation in yeast. We previously demonstrated that the human Dna2 orthologue (hDna2) localizes to the nucleus and contributes to genomic stability. Here we investigated the role hDna2 plays in DNA replication. We show that Dna2 associates with the replisome protein And-1 in a cell cycledependent manner. Depletion of hDna2 resulted in S/G 2 phasespecific DNA damage as evidenced by increased ␥-H2AX, replication protein A foci, and Chk1 kinase phosphorylation, a readout for activation of the ATR-mediated S phase checkpoint. In addition, we observed reduced origin firing in hDna2-depleted cells consistent with Chk1 activation. We next examined the impact of hDna2 on OF maturation and replication fork progression in human cells. As expected, FEN1 depletion led to a significant reduction in OF maturation. Strikingly, the reduction in OF maturation had no impact on replication fork progression, indicating that fork movement is not tightly coupled to lagging strand maturation. Analysis of hDna2-depleted cells failed to reveal a defect in OF maturation or replication fork progression. Prior work in yeast demonstrated that ectopic expression of FEN1 rescues Dna2 defects. In contrast, we found that FEN1 expression in hDna2-depleted cells failed to rescue genomic instability. These findings suggest that the genomic instability observed in hDna2-depleted cells does not arise from defective OF maturation and that hDna2 plays a role in DNA replication that is distinct from FEN1 and OF maturation.
Genomic maintenance requires coordination of several processes including DNA replication, DNA repair, transcription, and cell cycle progression. Given the importance of genomic stability for ongoing cell proliferation, it is not surprising that DNA replication and repair proteins have evolved multiple functions and participate in several replication and repair pathways. Previously, we demonstrated that human Dna2 (hDna2) 2 localizes to the nucleus and mitochondria and participates in DNA maintenance in both compartments (1). Dna2 is a highly conserved helicase/nuclease, originally discovered in Saccharomyces cerevisiae and found in all organisms from yeast to humans (2)(3)(4)(5)(6). Dna2 possesses a 5Ј to 3Ј ATP-dependent helicase activity and a flap endonuclease activity (7). Genetic and biochemical experiments conducted in yeast support a model in which Dna2 contributes to Okazaki fragment (OF) maturation in lagging strand DNA replication (8 -10), although in vivo evidence of Dna2 participating in this process is lacking.
During OF processing, flap endonuclease 1 (FEN1) is the major endonuclease that acts in a coordinated manner with DNA polymerase ␦ strand displacement activity to remove short RNA/DNA flaps formed on the previous OF (short flap pathway OF processing) (9 -12). Flaps that escape FEN1 cleavage and are longer than 27 nucleotides are subsequently coated by replication protein A (RPA), which inhibits FEN1 nuclease activity (long flap pathway OF processing) (9,12,13). RPAbound OF flaps recruit Dna2, which cleaves the RPA-coated DNA and displaces RPA, leaving a short 5-6-nucleotide RNAfree DNA flap that is further processed by FEN1 to produce a ligatable nick (9, 10, 14 -18). In support of this long flap model, both PIF1 helicase and polymerase ␦ processivity subunit (Pol 32) promote long flap formation in vitro, and their deletion rescues the lethality associated with yeast Dna2 loss, presumably because long flaps no longer form (12, 19 -22). Conversely, mutations that increase polymerase ␦ processivity are synthetically lethal with dna2 mutations (19).
In addition to its role in OF processing, Dna2 functions in DNA repair and telomere maintenance through its DNA endresection activity. In yeast, Dna2 undergoes a dynamic localization where it is present at telomeres in G 1 , relocalizes throughout the genome in S phase, and moves back to the telomeres in late S/G 2 phase (23). In addition, upon bleomycin treatment, yeast Dna2 leaves the telomere and localizes to sites of double strand breaks (DSBs) (23). Recent studies have demonstrated that yeast Dna2 nuclease activity participates in the formation of a 3Ј single strand DNA overhang essential to initiate the homologous recombination process or to maintain telomeric stability (24 -30). Furthermore, Nimonkar et al. (31) elegantly reconstituted DNA end resection in vitro using purified human proteins and demonstrated that hDna2 physically interacts with Bloom to resect 5Ј DNA ends in a process that depends on hDna2 nuclease but not hDna2 helicase activity. Nonetheless, in vivo and in vitro studies in yeast and humans indicate that Exo1 can compensate for Dna2 nuclease activity in this process (24,29,31). This suggests that the essential function of Dna2 is not its resection activity during DSB repair but rather its function in removing long flaps during DNA replication.
Our previous work revealed that hDna2 contributes to genomic stability (1). Here, we provide evidence that hDna2 ensures genomic stability by virtue of a critical role in DNA replication that is independent of FEN1 and OF processing.
For FEN1 ectopic expression experiments, viral production and infections were carried out as described previously (32). U-2-OS cells were infected with pLKO.1 shSCR or pLKO.1 shDna2 lentiviruses for 5 h. Cells were then counted and reseeded in the presence of media containing Adeno-FEN1 or Adeno-GFP (33). 48 h post-infection cells were re-plated at 0.8 ϫ 10 6 cells/10-cm plate and then analyzed 24 to 48 h later by hypotonic propidium iodide staining and flow cytometry as previously described (1). FEN1 ectopic expression was confirmed by Western blot and immunofluorescence (data not shown).
Immunoprecipitation and Western Blot Analysis-For co-immunoprecipitation studies, cells were washed in PBS and lysed in TBS-Triton X-100 buffer containing 50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, and protease and phosphatase inhibitors. 2 mg of protein extracts were immunoprecipitated with 2.4 g of anti-hDna2 or IgG control antibodies using protein A beads overnight at 4°C. The following morning beads were washed 3 times in 1 ml of TBS-Triton X-100 buffer before eluting bound proteins in 2ϫ Laemmli buffer by boiling for 5 min at 95°C. Anti-hDna2 antibody was produced as previously described (5), and anti-And-1 was kindly provided by Dr. Anindya Dutta (34).
For the cell cycle experiments, cells were enriched in G 0 /G 1 and G 1 /S by serum starvation and double thymidine block, respectively, as previously described (35). Briefly, for G 0 /G 1 enrichment, exponentially growing cells were cultured in DMEM medium containing 0.6% FBS for 38 h, then harvested, analyzed by FACS to ensure arrest, and used for immunoprecipitation of hDna2 followed by Western blotting for And-1. For G 1 /S enrichment, exponentially growing cells were cultured in medium supplemented with 3 mM thymidine for 16 h and washed twice in pre-warmed PBS, and fresh medium without thymidine was added for 12 h. Then, fresh medium with 3 mM thymidine was added, and cells were grown for another 16 h. Arrest was monitored by flow cytometry analysis, and cells were harvested and used for co-immunoprecipitation studies as described above.
For RPA immunofluorescence, cells were pre-extracted before fixing as previously described (37). Briefly, U-2-OS cells were grown for 1-2 days on coverslips. Cells were washed in ice-cold cytoskeleton (CSK) buffer (10 mM HEPES/KOH, pH 7.4, 300 mM sucrose, 100 mM NaCl, 3 mM MgCl 2 ) and then extracted for 6 min on ice with 0.5% Triton X-100 in CSK buffer supplemented with protease and phosphatase inhibitors. After extraction, cells were fixed in 3.7% formaldehyde at room temperature for 25 min followed by immunofluorescence staining using anti-replication protein A antibody (RPA-70-9, #NA13, Calbiochem).
Internuclei Chromatin Bridges and Micronuclei Counts-hDna2-depleted (shDna2 and shDna2Ј) and control (shSCR) U-2-OS cells were seeded in a 12-well plate at 5 ϫ 10 4 cells per well 4 days post-infection. After 36 -48 h, cells were fixed and stained with DAPI, and the number of internuclei bridges (ICBs) and micronuclei were counted. At least 1000 nuclei were counted per well, and 6 wells were quantified per experiment. Two independent experiments were quantified for micronuclei counts in Fig. 1C and ICB counts in Fig. 5D. To inhibit Chk1, cells were treated with either 300 nM Gö 6976 (Calbiochem) (38) or 100 nM AZD7762 (Axon Medchem BV, The Netherlands) (39) for 8 h before fixing and quantification of ICBs. Inhibition of Chk1 was confirmed by flow cytometry and Western blot ( Fig. 5C and data not shown). Images were processed using Photoshop 7.0 gray scale and invert function (Adobe, San Jose, CA).
S Phase Progression Assay-U-2-OS cells were cultured for 30 min in the presence of 20 M bromo-2-deoxyuridine (BrdU) in the dark as previously described (33). Briefly, cells were washed, cultured in fresh medium, and harvested by trypsinization at the indicated times. Cells were washed in PBS and fixed in 4% paraformaldehyde and 0.1% Triton X-100 in PBS for 20 min at room temperature before DNase I treatment (30 g DNase I (Sigma) at 37°C for 1 h). BrdU was detected with an Alexa Fluor 488-conjugated anti-BrdU antibody (A21303, Invitrogen), and cellular DNA content was determined by propidium iodide staining followed by flow cytometric analysis.
Metaphase Preparation and Chromosome FISH-Subconfluent U-2-OS cells were incubated for 3.5-4 h with 0.1 g/ml colcemid to isolate mitotic cells by mitotic shake-off. After hypotonic swelling in 75 mM KCl for 10 min at 37°C, cells were fixed in methanol/acetic acid (3:1), dropped onto glass slides, and aged at room temperature for 3 days. FISH was performed as previously described (32). Briefly, slides were rehydrated for 10 min in PBS, fixed with 4% paraformaldehyde in PBS for 2 min, then hybridized with 0.3 g/ml telomeric PNA probe (Cy3-(CCCTAA) 3 ) and a centromeric probe (Flu-OO-CTT-CGTTGGAAACGGGA) in 70% formamide, 10 mM Tris HCl, pH 7.2, plus blocking reagent (Roche Applied Science). DNA was denatured for 3 min at 80°C, and hybridizations were carried out at 37°C for 4 h in a moist chamber. Slides were subsequently washed, dehydrated, and mounted using VectaShield (Vector Labs, Burlingame, CA) containing DAPI. Images were taken using a Nikon 90i microscope and analyzed using the ISIS FISH imaging software (Metasystems, Altlussheim, Germany).
BrdU-comet Assay-The BrdU-comet assay was performed as previously described (41). Briefly, U-2-OS cells were plated at 5 ϫ 10 5 cells/6-cm plate and grown for 36 h at 37°C. Cells were then pulsed for 30 min with 100 M BrdU (Sigma) and chased for 1-8 h in growth medium lacking BrdU. Cells were then trypsinized and embedded in low melting point agarose at 37°C before spreading onto comet slides (Trevigen, Gaithersburg, MD). Cells on slides were then lysed, denatured, and run through electrophoresis under denaturing conditions (200 mM NaOH, 1 mM EDTA) before immunostaining with anti-BrdU antibody (#555627, BD Biosciences) for 1 h in the dark. The primary antibody was detected using anti-mouse IgG-Alexa-Fluor 488 (Invitrogen), and cells were counterstained with DAPI. At least 40 comet tails were quantified per sample time point using CometScore TM (TriTek). A total of three independent experiments were conducted.
Microfluidic-assisted Replication Track Analysis (maRTA) Assay-Replication fork progression rates were determined using maRTA (42). Briefly, hDna2-depleted (shDna2) or control (shSCR) U-2-OS cells were labeled for 30 min each with 50 M IdU followed by 50 M CldU (Sigma). For FEN1-depleted cells, U-2 OS-hTert cells were used to avoid potential telomeric defects induced by FEN1 depletion (43). Cells were then collected by trypsinization and used to prepare agarose plugs as previously described (42). High molecular weight DNA was isolated from cells embedded in agarose by brief heating to 75°C to melt the agarose followed by agarose digestion. The resulting high molecular weight DNA was then loaded by capillary tension into microchannels to uniformly stretch and capture long, high molecular weight DNA molecules on glass coverslips for immunostaining and fluorescence microscopy. Origin firing efficiency was determined by counting the fraction of origin firing events among all active replication events. Replication elongation efficiency was determined by measuring the mean length of first-label replication tracks in double-labeled tracks to analyze active/ongoing fork rates. Track lengths were measured in digital images of track using the AxioVision software package (Carl Zeiss). Three replicate samples of hDna2-depleted U-2-OS cells or mock-depleted U-2-OS cells (hDna2 experiment) or FEN1 depleted U-2 OS-hTert versus mock-depleted U-2 OS-hTert cells (FEN1 experiment) were analyzed for each determination. A total of 250 -450 replication tracks were measured in each sample.

RESULTS
hDna2 Contributes to Genomic Stability-We previously reported that hDna2 depletion in U-2-OS cells leads to genomic instability characterized by the appearance of aneuploid cells, ICBs, and an accumulation of cells in the late S/G 2 phase of the cell cycle (1). Here we further report that hDna2 depletion results in an increase in ␥-H2AX, a well characterized marker of DNA damage including DSBs, and the appearance of micronuclei indicative of aberrant mitosis. U-2-OS cells were transduced with one of two shRNA hairpins that led to a greater than 70% reduction in hDna2 mRNA and protein levels (Fig. 1, A and  B). Analysis of hDna2-depleted cells revealed a 2-fold increase in micronuclei compared with cells expressing a control shRNA (Fig. 1C). Furthermore, hDna2-depleted cells displayed an increase in ␥-H2AX foci that was confirmed by Western blot analysis (Fig. 1, B and D). In addition to the appearance of ␥-H2AX foci, we also observed phosphorylated-ATM foci in hDna2-depleted cells (Fig. 1D), confirming that hDna2 depletion results in the generation of DNA damage with activation of the DNA damage checkpoint.
ICBs arise from unresolved replication intermediates, defective mitosis, and/or telomere fusions that form upon loss of telomeric integrity (44). Because Dna2 plays an important role in telomere stability in yeast, we investigated whether the ICBs observed upon hDna2 depletion were the result of telomeric fusions. To address this possibility, we analyzed metaphases from control or hDna2-depleted cells. No increase in chromosomal end-to-end fusions was observed in cells expressing shRNAs targeting hDna2, suggesting that telomere dysfunction is not responsible for the formation of ICBs in these cells (supplemental Fig. 1). Interestingly, analysis of metaphases isolated from hDna2-depleted cells revealed the appearance of abnormal chromosomes (Fig. 1E). These metaphases are reminiscent of chromosomes isolated from cells entering mitosis with incompletely replicated DNA (45) that could lead to the generation of ICBs, DSBs, and micronuclei similar to what is observed in hDna2-depleted cells. Together these experiments suggest that the genomic instability observed in hDna2-depleted cells arises from incomplete DNA replication.
hDna2 Nuclease and Helicase Activities Are Essential-Dna2 is a highly conserved enzyme that possesses nuclease and helicase/ATPase activities that are postulated to contribute to its function in vivo. Although both activities are essential for viability in S. cerevisiae (46), recent in vitro biochemical studies have called into question the significance of the helicase activity in the human protein (5, 6). Therefore, we addressed whether hDna2 nuclease and/or helicase activities contribute to genomic stability in mammalian cells. To assess the roles of these activities, we carried out a series of genetic knockdownrescue experiments utilizing an shRNA targeting the 3Ј-untranslated region (UTR) of endogenous hDna2 while expressing a FLAG-tagged shRNA-resistant hDna2 cDNA. We expressed either a control vector (ctrl) or wild-type (wt) hDna2, nuclease-deficient D294A (nuc) hDna2, or helicase/ATPasedeficient K671E (hel) hDna2 cDNA (5). To distinguish between the endogenous and exogenous hDna2 mRNAs, we designed specific PCR primers that amplify a region encompassing the last exon (exon 21) and the 3Ј-UTR of the endogenous hDna2 gene that is absent in the cDNA constructs. This allowed us to FIGURE 1. hDna2 contributes to genomic maintenance. A, knockdown of hDna2 in U-2-OS cells was determined by quantitative RT-PCR. Results for two independent hairpins targeting hDna2 (shDna2 and hDna2Ј) are normalized to the shSCR control cell line expression levels. Note: shDna2 induces a better knockdown of hDna2 mRNA than shDna2Ј. B, hDna2 knockdown was determined by Western blot analysis. Lysates were also probed for ␥-H2AX to assess DNA damage and DSBs. C, shown are relative micronuclei counts and representative images depicting micronuclei (red arrows) after hDna2 depletion in cells. Results were normalized to shSCR control cells. D, shown is immunofluorescence staining of ␥-H2AX (red) and phosphorylated-ATM (P-ATM) (green) in U-2-OS cells depleted of hDna2 (shDna2, bottom panels) or a control hairpin (shSCR, top panels). Nuclei were stained with DAPI (blue). E, left, shown is metaphase spread of control cells and hDna2-depleted cells with abnormal chromosomes. Right, quantification of 200 metaphases from 2 independent experiments is shown. All error bars represent S.E., and * denotes p Ͻ 0.01 compared with shSCR.
confirm the knockdown of endogenous hDna2 while verifying ectopic expression of the cDNAs by Western blot using a FLAG antibody (Fig. 2, A and B).
We previously demonstrated that depletion of endogenous hDna2 results in a reduced G 1 population, a late S/G 2 cell cycle arrest, and the appearance of aneuploid cells (1). To assess the role of the helicase and nuclease activity of hDna2, we utilized flow cytometry to measure DNA content and determine the cell cycle profile of cells expressing wild-type or mutant hDna2. Expression of the wild-type allele significantly rescued the cell cycle defects observed upon depletion of hDna2 (Fig. 2C). In contrast, expression of the nuclease-deficient or helicase-deficient alleles did not rescue the cell cycle defects, indicating that both activities are essential to maintain genomic stability (Fig.  2C). Furthermore, cells expressing the nuclease-deficient hDna2 protein at even lower levels than the wild-type or helicase-deficient mutants displayed cell cycle defects that were more severe than those observed in cells depleted of endogenous hDna2 alone (Fig. 2, B and C, compare G 1 /(S ϩ G 2 ) ratio, and cells with Ͼ4N DNA content). Interestingly, cells express-ing high levels of the nuclease deficient allele were selected against over the course of the experiment as evidenced by a significant reduction in protein level, suggesting that the nuclease-deficient mutant is either toxic or functions as a dominant negative (Fig. 2, A-C).
To address whether hDna2 helicase activity contributed to the deleterious properties of the nuclease-deficient Dna2 allele, we expressed a double mutant (dm) allele lacking both the hDna2 nuclease and helicase activities. As seen in supplemental Fig. 2B, expression of the double mutant was maintained at levels similar to the helicase-deficient protein in contrast to the nuclease-deficient mutant whose expression was rapidly lost (supplemental Fig. 2, A and B). These results suggest that the toxic effect of nuclease-deficient hDna2 depends on its helicase activity and that both activities are coupled in vivo. Together, these results demonstrate that the nuclease and helicase activities of hDna2 are essential to maintain genomic stability.
hDna2 Interacts with Replisome Protein And-1 in a Replication-dependent Manner-Above, we demonstrate that hDna2 depletion leads to DNA damage and the appearance of meta- phases reminiscent of cells entering mitosis with incomplete DNA replication. In Xenopus laevis extract, Dna2 is recruited to DNA shortly after replication licensing where it interacts with the replisome protein And-1 (47). To determine whether Dna2 associates with the replisome in human cells, we investigated interactions between hDna2 and And-1. We found that endogenous And-1 co-immunoprecipitated with hDna2 from asynchronous cells, suggesting that hDna2 associates with the replisome ( Fig. 3A and data not shown). Furthermore, comparisons of the And-1/hDna2 interaction in cells arrested in G 0 /G 1 by serum starvation and cells blocked at the G 1 /S border by a double thymidine treatment revealed a significant increase in And-1/hDna2 interaction in cells arrested at the G 1 /S transition (Fig.  3B). These observations demonstrate that hDna2 specifically interacts with And-1 in replicating cells and are similar to observations made in Xenopus extract (47). Together they suggest that hDna2 interacts with And-1 shortly after licensing of the pre-replication complex as part of the replisome.
hDna2 Depletion Leads to Replication Checkpoint Activation-Dna2 plays an essential role in DNA replication in yeast. Our results indicate a corresponding, important role in human DNA replication. Therefore, we next tested whether hDna2 depletion altered S phase progression. Control or hDna2-depleted U-2-OS cells were pulsed with BrdU for 30 min and chased for 12 h. Although hDna2-depleted cells incorporated BrdU at the same rate as control cells, they displayed a marked delay in completing the S/G 2 phase and consequently took longer to appear in the next G 1 phase after mitosis (Fig. 4, A and B) (compare 8-, 10-, and 12-h time points). These observations support the original cell cycle profiling data and suggest that hDna2 depletion alters late S/G 2 replication rather than bulk DNA replication rates during S phase. Furthermore, Western blot analysis revealed a reduction in the phosphorylation of the mitotic marker histone H3 at serine 10 in depleted versus control cells, confirming that the block occurs before mitosis (Fig. 5A, bottom panel).
Complete replication of the genome relies on efficient replication initiation, progression, and maturation. When replication is inhibited at any of these steps, cells respond by activating the S-phase checkpoint. Therefore, given the characteristic damage and cell cycle arrest observed after hDna2 depletion, we next investigated the activation status of Chk1. Western blot analysis of hDna2-depleted cells revealed a significant increase in phosphorylation of Chk1 at serine 317 compared with control cells, indicative of DNA replication stress (Fig. 5A, top  panel). In agreement with this finding, we observed an accumulation of single-stranded DNA, a trigger of the S-phase checkpoint, as evidenced by a significant increase in the number of RPA foci per cell in hDna2-depleted cells compared with control cells (Fig. 5B).
Chk1 activation prevents cells that have not completed DNA replication from moving into mitosis (48). When Chk1 is inhibited in cells undergoing replication stress, they can move into mitosis with incompletely replicated DNA that if unrepaired will lead to unresolved replication intermediates. To address whether checkpoint activation blocked hDna2-depleted cells from entering mitosis with under-replicated chromosomes and/or unresolved damage, we treated cells with one of two Chk1 inhibitors, Gö 6976 or AZD7762 (38,39). Cells depleted of hDna2 were grown for 8 h in the presence of the Chk1 inhibitor. Flow cytometric analysis revealed a drastic alteration in the cell cycle profile of hDna2-depleted cells treated with a Chk1 inhibitor compared with untreated cells (Fig. 5C). hDna2-depleted cells treated with the Chk1 inhibitor and released from the G 2 arrest entered mitosis but displayed an increase in ICBs compared with untreated hDna2-depleted cells, indicating that cells released from the G 2 block underwent aberrant mitosis (Fig. 5D). Together these experiments establish that cells depleted of hDna2 display replication defects and arrest in late S/G 2 due to checkpoint activation. Upon checkpoint override, cells resume cell cycle progression despite incomplete replication and/or unresolved damage and undergo aberrant mitoses.
Above we demonstrate that hDna2 depletion leads to Chk1 activation and a late S/G 2 cell cycle arrest. Chk1 activation after replicative stress occurs to stabilize ongoing replication forks FIGURE 3. hDna2 interacts with And-1 in a replication-dependent manner. A, shown is the interaction between hDna2 and And-1 in asynchronous cells. hDna2 was immunoprecipitated from asynchronous HeLa cells followed by Western blot analysis with antibodies against And-1 or hDna2 as indicated. B, interaction between hDna2 and And-1 occurs during G 1 /S transition. HeLa cells were arrested in G 0 /G 1 by serum starvation or during G 1 /S by a double thymidine block. hDna2 was immunoprecipitated from the arrested cells, and Western blot analysis was performed with antibodies against And-1 or hDna2 as indicated. and to inhibit firing of additional forks (48). Therefore, to examine the impact of Chk1 activation on origin firing we used maRTA, wherein DNA is sequentially labeled in vivo with short pulses of the base analogues IdU and CldU (42). After labeling, DNA is isolated and stretched on coverslips for detection by IdU and CIdU immunostaining (Fig. 6A). Using maRTA, we quantified the fraction of DNA origin firing events present among all tracks labeled with CldU and IdU. We found that hDna2-depleted cells displayed a 20 -25% reduction in origin firing events compared with control cells (Fig. 6B). Together, these experiments establish that hDna2 depletion activates the replication checkpoint, which in turn inhibits the firing of replication forks.
Defects in Okazaki Fragment Processing Do Not Alter Replication Fork Progression-Given the type of DNA damage we observed after hDna2 depletion and the model established in yeast suggesting Dna2 participates in OF maturation (9, 10), we next examined the role of hDna2 in DNA replication and fork progression using maRTA. By measuring the lengths of first label segments within replication tracks containing both labels, we determined the rate of DNA replication elongation. Using this method, we found that hDna2 depletion did not alter replication fork rate. Indeed, we found that DNA track lengths were similar in control and hDna2-depleted cells (Fig. 7A). This finding suggests that the damage that arises from unprocessed DNA fragments accumulating behind the replication fork does not impede continued replication fork progression and is consistent with our result showing no defects in S phase progression in hDna2-depleted cells (Fig. 4, A and B).
To further address the possibility that unprocessed OFs do not impact replication kinetics, we also depleted cells of FEN1 (supplemental Fig. 3) (10). Because FEN1 is the major OF processing nuclease in the cell, we expected to find that FEN1 depletion would alter replication fork progression. Strikingly, we found that FEN1 depletion also failed to slow replication fork progression. In fact, we found that DNA track lengths were slightly longer in FEN1-depleted cells compared with control cells (Fig. 7B and supplemental Fig. 3). Together these observations strongly suggest that the presence of unprocessed OFs and the associated damage they generate behind the replication fork is not sufficient to slow replication fork progression in vivo. FIGURE 5. hDna2 depletion activates the replication checkpoint. A, Western blot analysis of control (shSCR) and hDna2-depleted (shDna2) U-2-OS cells is shown. Whole cell lysates were probed for total versus phosphorylated-Chk1 (P-Chk1 S317) (top panels). Samples were also probed for phosphorylated-histone H3 (P-H3 S10). B, immunofluorescence staining of RPA-70 in shSCR and shDna2 U-2-OS cells is shown. Cells were pre-extracted before fixing and staining to eliminate cytoplasmic RPA-70 (37). Quantification of RPA foci: average number (left) and total number (right) of foci per cell. * denotes p Ͻ 0.01. C, cell cycle distribution of shSCR and shDna2 ϩ/Ϫ treatment with Chk1 inhibitor (Gö 6976) for 8 h. Similar results were obtained when AZD7762 was used to inhibit Chk1 (data not shown). D, ICBs (1) were quantified in control (shSCR) and hDna2-depleted (shDna2) U-2-OS cells Ϯ Chk1 inhibitor. A representative image is shown with an ICB indicated by an arrow. * denotes p Ͻ 0.01. FIGURE 6. hDna2 depletion reduces origin firing. A, shSCR or shDna2 U-2-OS cells were labeled consecutively with IdU (green) and CldU (red) for 30 min each before isolating and stretching DNA for immunostaining. B, origin firing events among all tracks labeled were identified as CldU-only (red only) or CldU-IdU-CldU (red-green-red) triple-segment tracks. The mean percentages of new origin firing events defined by these two track types among all labeled tracks are shown for three independent experiments in which 200 -450 tracks/experiment were analyzed for control shSCR or shDna2. Error bars show S.E. between three independent experiments. * denotes p Ͻ 0.05.
hDna2 Depletion Does Not Lead to Detectable Defects in Maturation of Newly Synthesized DNA-Long flaps requiring Dna2 for processing are predicted to arise in only a small percentage of OFs, and in specific regions of the genome (4,12,20). Therefore, the need for Dna2 should be dispensable for processing the vast majority of OFs, in contrast to the short flaps that are cleaved by FEN1. To determine whether hDna2 participates in OF processing, we measured the maturation kinetics of newly replicated DNA using a BrdU-comet assay. We reasoned that if hDna2 were necessary to process only a minority of the flaps, it would be difficult to observe maturation differences in cells depleted of hDna2 alone. Therefore, we also depleted FEN1 in addition to hDna2 in an effort to increase the accumulation of long flaps that would require Dna2 activity. In parallel, we also depleted cells of DNA ligase I (LigI) as a positive control; LigI is required to ligate OF, which is required to complete OF proc-essing (41). Western blot analysis demonstrated that we successfully depleted FEN1, LigI, or hDna2 alone and co-depleted FEN1 and hDna2 (Fig. 8A).
To assess the maturation of newly replicated DNA, control cells and cells depleted of hDna2, FEN1, or LigI alone or co-depleted of both hDna2 and FEN1 were pulsed with BrdU and analyzed by an alkaline comet assay. Immunofluorescence against BrdU was used to assess the integrity of newly replicated DNA by measuring its migration from the tail (unligated DNA fragments, i.e. unprocessed OFs) to the head (ligated DNA, i.e. processed OFs) of the comet (Fig. 8B compare 1-h to 8-h images). As expected, LigI depletion slowed the maturation of newly replicated DNA from comet tails to heads (Fig. 8C) (41). In contrast, we failed to observe a significant difference in OF maturation in hDna2-depleted compared with control cells (Fig. 8C). Furthermore, FEN1-depleted cells displayed a reduction similar to that of LigI-depleted cells, indicating that FEN1 is the major flap endonuclease responsible for processing OFs during lagging strand replication. Finally, OF maturation did not differ in FEN1-hDna2 co-depleted cells compared with cells depleted of FEN1 alone (Fig. 8C), suggesting that hDna2 does not compensate for FEN1 loss in OF processing. These results indicate that the DNA damage that arises upon hDna2 depletion (Fig. 9C, compare shDna2 to shFEN1 ␥-H2AX levels) is unlikely to arise from a global OF maturation defect. However, it does not rule out a role for hDna2 in the maturation of a small fraction of OFs.
To further assess whether hDna2 primary function in DNA replication involves OF processing, we ectopically expressed FEN1 in hDna2-depleted cells (Ref. 33 and data not shown). In yeast, FEN1 overexpression compensates for phenotypes caused by hypomorphic mutations of Dna2 (8). However, ectopic expression of FEN1 in human cells did not improve the cell cycle defects observed in hDna2-depleted cells (Fig. 9, A  and B). These results further reinforce the contention that the key role of hDna2 in DNA replication is independent of FEN1 or OF processing.

DISCUSSION
In this study we provide evidence that hDna2 is essential to ensure faithful replication of the genome like its yeast homologue and that unlike in yeast this essential role is independent of a requirement for hDna2 in global OF processing. We further demonstrate that both the nuclease and helicase activities of hDna2 contribute to genomic stability in human cells as they do in yeast. Finally, we report the unexpected finding that defective OF processing has no detectable impact on replication fork progression. This surprising result indicates that fork movement is decoupled from damage that arises behind the replication fork.
In S. cerevisiae, temperature-sensitive dna2 mutant alleles arrest cells in G 2 with a 2C DNA content when shifted to the restrictive temperature (49). In Schizosaccharomyces pombe, temperature-sensitive mutants also arrest in late S phase but display no defects in bulk DNA synthesis (4). When combined with a checkpoint inhibitor, these cells bypass the arrest and undergo aberrant mitosis. Similarly, here we demonstrate that hDna2-depleted cells harbor DNA damage and arrest in late FIGURE 7. Replication fork progression is unperturbed in hDna2-or FEN1depleted cells. A, quantification of three independent maRTA experiments is shown on the right. The bar graph summarizes mean lengths of first-label segments labeled for 30 min in two-segment tracks (i.e. tracks labeled consecutively with IdU and CIdU) to ensure that fork rate measurements were made from active replication forks. Error bars show 95% confidence intervals for sample means. No statistical difference in mean track lengths was observed between shSCR and shDna2 cells using a two-sample Kolmogorov-Smirnov test. B, shown are quantification of three independent experiments summarizing mean track length of the first label in cells depleted of FEN1 (shFEN1, shFEN1Ј) or control cells (shSCR).
S/G 2 phase. Upon Chk1 inhibition, cells bypass the replication checkpoint and progress through mitosis, displaying aberrant mitotic structures (increase in ICBs; Fig. 5D). These pheno-types recapitulate observations in yeast and strongly suggest a conserved role for Dna2 in DNA replication in humans. Supporting our conclusions, ectopic expression of hDna2 suppresses the growth defects of the replication mutant dna2-1 in S. cerevisiae (50).
The requirement for Dna2 helicase activity for yeast viability in S. cerevisiae indicates that this function is required in vivo (46). Indeed, different studies have demonstrated that Dna2 helicase activity is essential for efficient cleavage of long DNA flaps that may form secondary structures (7,51). However, several studies demonstrated that the helicase activity of hDna2 is weak or undetectable and might not serve a corresponding role in human cells (5,6). Our observation that neither nuclease nor helicase-deficient hDna2 mutants rescue hDna2 depletion demonstrates that both activities are essential to ensure the integrity of the genome. Furthermore, we provide evidence that as in yeast, hDna2 nuclease-deficient expression is toxic to the cell (46,51). This toxicity depends on its helicase activity, suggesting that both activities need to be coupled within a single Dna2 molecule to act on its substrates in vivo. This is in contrast to Dna2 function in 5Ј end resection during DSB repair where only the nuclease activity of Dna2 participates in this process (24,28,29,53).
OF processing is an essential process that ensures a continuous lagging DNA strand and avoids single and double strand break formation. Several prior studies have suggested that defects in OF processing do not impact the rate of DNA syn-  thesis. Indeed, we previously demonstrated that FEN1 depletion did not slow S-phase progression or SV40 LargeT antigendependent DNA replication in vitro (33). Similarly, LigI mutant cells (46BR.1G1) that exhibit low LigI activity do not activate the intra-S phase checkpoint (41). Although protein extracts from these cells are deficient in OF ligation during SV40 DNA replication in vitro, they support incorporation of [␣-32 P] dATP into plasmid DNA with similar kinetics as extracts from control fibroblasts (56). In addition, yeast deficient in LigI completely replicate their genomes despite being unable to ligate OF (54,55). Together these studies suggest that replication fork progression is unperturbed in the absence of LigI. Here, using the sensitive maRTA technique, we demonstrate that depletion of an essential OF maturation protein, FEN1, does not alter replication fork progression in vivo. These observations strongly suggest that flaps that persist behind the replication fork do not affect the synthesis rate of DNA during replication.
FEN1 is postulated to be the primary endonuclease that processes flaps produced during lagging strand synthesis. However, both genetic and biochemical data indicate the presence of an additional long-flap processing pathway. Long DNA flap formation is promoted by PIF1 in vitro and requires Dna2 activity to create shorter flaps that are cleaved by FEN1 to generate a ligatable nick (9,10,14). To address the role of hDna2 in longflap processing in human cells, we depleted FEN1 and hDna2 independently or together and then measured the maturation of nascent DNA. We found that maturation occurs with slower kinetics in FEN1-depleted cells, thus confirming FEN1 as the primary endonuclease responsible for processing OFs in human cells. In contrast, we did not observe a defect in maturation in hDna2-depleted cells. This indicates that if hDna2 processes long flaps, it does not represent a significant fraction of OFs in accordance with the model. However, our finding that FEN1-hDna2 co-depletion did not result in a significant difference in OF maturation rate compared with cells depleted of FEN1 alone was surprising, as in yeast dna2 and rad27, mutations are synergistic (8). These results suggest either long flaps do not form readily in the absence of FEN1 in human cells or that another nuclease can process flaps to render them ligatable. In yeast, other nucleases such as RNaseH2 or Exo1 can compensate for FEN1 loss by processing short flaps (20,(57)(58)(59). It remains to be seen whether their functions are conserved in humans. Despite the apparent functional dominance of FEN1, we were surprised to find that hDna2-depleted cells display more DNA damage than FEN1-depleted cells as assessed by ␥-H2AX levels and abnormal cell cycle profile (Fig. 9B and data not shown). One explanation for this is that even infrequent unprocessed long flaps are a source of DNA damage and genetic instability in the absence of hDna2. However, we find this to be unlikely because ectopic FEN1 expression had no impact on the defects observed in hDna2-depleted cells. Thus a more likely and interesting possibility is that hDna2 has an important, as yet to be elucidated, role in replication or DNA repair beyond its postulated role in OF processing. This possibility is strengthened by our observation that hDna2 interacts with And-1 specifically in replicating but not resting human cells. This interaction is in accordance with yeast and Xenopus observations (47, 52, 60, 61) and suggests that hDna2 is a component of the replication fork. The late S/G 2 phase arrest and appearance of aberrant post-mitotic structures could suggest a role for hDna2 in completing DNA replication by aiding in the resolution of replication intermediates. Interestingly, the Bloom RecQ helicase can resolve late replication intermediates, and its ectopic expression rescues Dna2 mutants in yeast (50). Alternatively, Xenopus Dna2 has been proposed to play a role in the early steps of DNA replication (3,47). It remains to be addressed whether hDna2 actively participates in any of these processes.