Error-prone, stress-induced 3’ flap-based Okazaki fragment maturation supports cell survival

Abstract

How cells with DNA replication defects acquire mutations that allow them to escape apoptosis under environmental stress is a long-standing question. Here, we report an error-prone Okazaki fragment maturation (OFM) pathway that is activated at restrictive temperatures in rad27Δ yeast cells. Restrictive temperature stress activates Dun1, facilitating transformation of un-processed 5’ flaps into 3’ flaps, which are removed by 3’ nucleases including Polδ. However, at certain regions, 3’ flaps form secondary structures that facilitate 3’ end extension rather than degradation, producing alternative duplications with short spacer sequences. Once such mutations occur at POL3, it fails to displace 5’flaps, thus rescues rad27Δ cells. Our study defines a stress-induced, error-prone OFM pathway that generates mutations that counteract replication defects and drive cellular evolution and survival.

Understanding the mutagenesis mechanisms that are active in cells under stress conditions is crucial for developing strategies to intervene in microbial pathogenesis, tumorigenesis, and drug resistance (1, 2). Lagging-strand DNA synthesis is particularly vulnerable to stress and environmental factors. During replication, lagging-strand DNA is synthesized as discrete Okazaki fragments (3), which contain short primase/Polα-synthesized RNA-DNA primers at their 5’ ends (4-6). During Okazaki fragment maturation (OFM), the RNA portion and any Polα-synthesized DNA with high incorporation errors are removed, via Polδ-mediated displacement DNA synthesis, which produces a 5’ RNA-DNA flap (4-6). The 5’ flap structure is removed by flap endonuclease 1 (FEN1) or through the sequential actions of DNA2 nuclease/helicase and FEN1 (7-9). FEN1 deficiency leads to accumulation of unprocessed 5’ flap structures, which may prevent ligation of Okazaki fragments, leaving DNA nicks or gaps that lead to collapse of replication forks and DNA double-strand breaks. In yeast, deletion of the FEN1 homolog RAD27 (rad27)Δ results in slow growth at permissive growth temperatures (30°C) and death at restrictive growth temperatures (37°C) (10).

Nevertheless, we discovered that a small population of rad27Δ yeast cells, which we called revertants, could grow at a similar rate as wild-type (WT) cells at 37°C (Fig. 1A). To determine if the revertants acquired somatic mutation(s) that permitted growth and to identify any such mutation(s), we conducted whole-genome sequencing (WGS) of WT, parental rad27Δ, and a revertant strain of yeast cells. We identified 21 somatic DNA mutations specific to one revertant colony (Table S1). A mutation in POL3, the DNA polymerase delta (Polδ) catalytic subunit (11), was the only nonsynonymous mutation that had 100% allele frequency in the revertant. Subsequent DNA sequencing analysis of the POL3 gene in independent rad27Δ revertant colonies (n = 31) revealed that each colony harbored a pol3 mutation (Fig. 1B). This suggests that these pol3 mutations, which map onto POL3 functional motifs (Fig. 1B, Supplementary text S1) and possibly affect its biochemical activities, might provide a survival advantage for rad27Δ cells grown under restrictive temperature stress. Furthermore, knock-in of the 458–477 internal tandem duplication (ITD) mutation, which occurred in 19 of the 31 independent colonies, or any of the four representative point mutations (R470G, R475I, A484V, and S847Y) successfully reversed the restrictive temperature-induced lethality phenotype of rad27Δ cells (Fig. 1C and fig. S1). rad27Δ cells are sensitive to methyl methanesulfonate (MMS) (10). Although rad27Δ revertant cells and rad27Δ pol3 ITD knock-in mutant cells were resistant to a low concentration (0.005%) of MMS, they were sensitive to higher concentrations (≥0.01%) of MMS (fig. S2). We observed that pol3 ITD cells in a WT RAD27 background were also sensitive to high concentrations of MMS (fig. S2). This at least partially explains why the pol3 ITD could not suppress MMS-induced lethality of rad27Δ cells at high MMS concentrations. In addition, pol3 ITD did not rescue the synthetic lethality that occurs in the context of rad27Δ coupled with deficiency of the 5’ nucleases EXO1 or DNA2 nuclease/helicase (Tables S2, S3, Supplementary text S2).

Fig. 1. Polδ internal tandem duplication (ITD) and missense mutations drive resistance to rad27Δ-induced conditional lethality.

(A) Spot assays of WT, rad27Δ, or rad27Δ revertant (Rev) yeast cells grown at 30°C (optimal temperature), 25°C (sub-optimal temperature), or 37°C (restrictive temperature) for 48 h. rad27Δ::URA3 and rad27Δ::LEU2 represent the rad27Δ allele with a linked URA3 or LEU2 selection marker gene, respectively. (B) pol3 mutations detected in independent revertant strains (n=31). Circles and diamonds represent base substitution and ITD mutations, respectively. The domain structures were defined as previously described (23). (C) Spot assays of WT, rad27Δ, or rad27Δ yeast cells with indicated pol3 knock-in mutations grown at 30°C, 25°C, or 37°C for 48 h. POL3::HIS3 represents the POL3 (WT or mutant) alleles with a linked HIS3 selection marker gene.

Two types of duplications were present in the revertants: pol3 591–598 ITD, a previously reported classic duplication resulting from re-alignment and ligation of unprocessed 5’ flaps (12), and pol3 458–477 ITD, which contained a 55 bp duplication with a 5 bp spacer between the duplicated units (fig. S3). We named the duplication with an intervening spacer an “alternative duplication.” Both pol3 591–598 ITD and pol3 458–477 ITD resembled ITDs detected in human cancer (13-15). To determine how the alternative duplication pol3 458–477 ITD originated, we conducted WGS of WT and rad27Δ cells grown at 37°C or 30°C for 4 h. The mutation frequency of WT cells was the same at both temperatures (Fig. 2A). In contrast, restrictive temperature stress increased the mutation frequency of rad27Δ cells by 2-fold; in particular, the frequency of duplications and base substitutions was increased (Fig. 2A). In addition, duplication insertions in rad27Δ cells grown at 37°C were considerably longer than those in rad27Δ cells grown at 30°C (Fig. 2B). The duplications revealed that rad27Δ cells grown at 37°C exhibited alternative duplications that were similar to the pol3 458–477 ITD. The alternative duplications were not detected in WT cells (30°C or 37°C) or in rad27Δ cells grown at 30°C (Fig. 2C), suggesting that alternative duplications were induced by restrictive temperature stress.

Fig. 2. Restrictive temperature stress induces 3’ flap-based OFM and results in alternative duplications.

(A) Somatic mutation frequencies and types as determined by WGS, in WT and rad27Δ cells grown at 30°C or 37°C for 4 h. (B) Lengths of inserted DNA sequences in duplications in rad27Δ cells grown at 30°C or 37°C for 4 h. (C) Top, diagram of classic and alternative duplications. Bottom, frequencies of classic and alternative duplications. (D) Predicted structures leading to three types of alternative duplications. Red and green lines: DNA sequences in red and green in fig. S4A; Orange lines: yellow highlighting in fig. S4A. (E) Schematic for specific labeling of 3’ flaps in genomic DNA. Green dots: dideoxyribonucleotide; red star: ³²P-deoxyribonucleotide. (F) Levels of 3’ flaps in genomic DNA from WT, rad27Δ, or rad27Δ pol3 ITD (rad27Δ-ITD) cells grown at 30°C or 37°C for 4 h. (G) Levels of 3’ flaps in genomic DNA from rad27Δ cells grown at 37°C with or without pre-treatment with Polδ. (H and I) Reconstitution assays using ³²P-labeled 3’ flap substrate S3 (H) or ³²P-labeled secondary structure-forming 3’ flap substrate S4 or S5 (I). Substrate structures are shown above a representative image of 8% denaturing PAGE. DNA substrates (S3, S4, S5), cleavage products (Cleaved S3, S4), unligated extended 3’ flap intermediates (Extended S4, S5), and ligated extended products (Ligated P3, P4, P5) are indicated. dNTP: deoxyribonucleotides.

We further noted that the sequences of these alternative duplications suggested formation of three different types of hairpin structures (Fig. 2D, fig. S4A-4D, Supplementary text S3). This supports a model of sequential actions, including conversion of a 5’ Okazaki fragment flap to a 3’ flap, annealing of the flap to a complementary sequence, extension of the 3’ flap, realignment, and ligation of the extended 3’ flap to produce an alternative duplication, including pol3 458–477 ITD. Consistent with this model, our WGS data indicated that 40% of the alternative duplications also carried base substitutions at the duplication unit. These substitutions most likely resulted from failure to remove Polα-generated errors on the 5’ flap. To determine if the restrictive temperature induced 3’ flap formation in rad27Δ cells, we developed an approach to specifically label the OH group on the 3’ flap on genomic DNA, in which 3’ OH at the nick or at the DNA end was pre-blocked with dideoxyribonucleotides (Fig. 2E). We detected a considerable number of 3’ flaps in rad27Δ cells grown at 37°C; in contrast, we detected few flaps in rad27Δ cells grown at 30°C, in WT cells grown at either temperature, or in rad27Δ cells carrying pol3 458–477 ITD grown at either temperature (Fig. 2F). Furthermore, pre-incubation of Polδ with genomic DNA from rad27Δ cells grown at 37°C could effectively remove the 3’ flaps (Fig. 2G), suggesting that Polδ may process 3’ flaps during OFM.

To define the proposed 3’ flap-based OFM mechanism, we reconstituted the sequential reactions of 3’ flap cleavage, DNA synthesis, and ligation of oligo-based DNA substrates (S) with a simple 3’ flap (S2 or S3; Fig. 2H and fig. S5B) or a secondary structure-forming 3’ flap (S4 or S5; Fig. 2I) for formation of type I or type II alternative duplications. In the presence of deoxyribonucleotide, Polδ could effectively cleave 3’ flap substrates S2 and S3 and stop at the junction of the 3’ flap and DNA duplex, generating ligatable DNA nicks for DNA Lig I (Fig. 2H and fig. S5, Supplementary text S4). However, deoxyribonucleotides inhibited cleavage of hairpin-forming 3’ flaps, and promoted extension of the annealed 3’ flap, producing ligated extended products (Fig. 2I, Supplementary text S4); this process resembled formation of alternative duplications. When extension of the annealed 3’ flap could not generate ligatable nicks, only unligatable extended products were produced (fig. S6A-6D), leading to failure of 3’ flap-based OFM. The single-stranded DNA (ssDNA) binding protein RPA had little effect on Polδ-mediated 3’ flap cleavage or subsequent nick ligation (Fig. 2H), and it slightly enhanced formation of the ligated extended products (Fig. 2I).

Using reconstitution assays, we showed that the 3’ nuclease activities of Polδ and Lig I were sufficient to complete 3’ flap processing for OFM. Other nucleases in the nuclear extract (NE) might also be important in processing 3’ flaps, especially the hairpin-forming 3’ flap (fig. S7A, S7B, Supplementary text S5). However, NE from rad27AΔ cells, particularly those grown at 37°C, had reduced 3’ flap processing activity (fig. S7A, S7B, Supplementary text S5). Because we observed no significant changes in expression of major 3’ nucleases in yeast (fig. S8), we postulated that restrictive temperature stress could also induce molecular changes to inhibit 3’ flap processing, allowing 3’ flaps to invade into nearby homologous sequences, leading to alternative duplications.

We next determined how the pol3 458–477 ITD enabled rad27Δ cells to overcome lethal stress. Because the pol3 458–477 ITD did not change Polδ protein levels in rad27Δ cells (fig. S9), we tested if it affected biochemical properties of Polδ. We assayed the DNA polymerase and 3’ nuclease activities of a purified recombinant protein Polδ complex containing either a WT Pol3 subunit (WT Polδ) or a 458–477 ITD Pol3 subunit (hereafter called Polδ-ITD). Polδ-ITD could catalyze DNA synthesis but was less processive than WT Polδ during primer extension (Fig. 3A). Similarly, Polδ-ITD could effectively fill the gap, but it was less active than WT Polδ in displacing the downstream DNA oligo (Fig. 3B). In addition, Polδ-ITD had relatively weak 3’ exonuclease activity on DNA duplexes, compared to WT Polδ (fig. S10). However, Polδ-ITD had similar activity to WT Polδ in cleaving the 3’ flap and generating a ligatable nick (fig. S11). This activity likely allows cells carrying the pol3 458–477 ITD to have a similar capacity as WT cells for catalyzing 3’ flap processing for OFM. In contrast, a 3’ exonuclease-dead mutant, Polδ D520E, did not cleave the 3’ flap (fig. S11), which may explain why the Polδ D520E mutation is lethal at restrictive temperature and synthetically lethal with rad27Δ (16).

Fig. 3. Polδ-ITD suppresses 5’ flap formation.

(A and B) In vitro assays of primer extension (A) and displacement DNA synthesis (B) by WT Polδ or Polδ-ITD. DNA substrates and primer extension products in panel A, and DNA substrates, gap filling products, or displacement DNA synthesis products in panel B are indicated. (C) Mean Can^r mutation rates of WT (n=5), rad27Δ (n=5), or rad27Δ yeast cells with knock-in of pol3 458-477 ITD (n=3), pol3 R470G (n=2), pol3 R475I (n=3), pol3 A484V (n=2), and pol3 S847Y (n=2). Error bars indicate s.d. p values were calculated using student’s t test. (D) Can^r mutation spectra of the indicated yeast strains. Values shown are percentages of the specific type of Can^r mutation in WT (n=22), rad27Δ (n=20), rad27Δ with knock-in of pol3 458-477 ITD (n=21), pol3 R470G (n=10), pol3 R475I (n=21), pol3 A484V (n=19), or pol3 S847Y (n=21). (E) Mutation frequencies and types present across the genome, as determined by WGS, in WT, rad27Δ, or rad27Δ cells with indicated pol3 knock-in mutations grown at 30°C (n=1).

We further revealed that knock-in ofpol3 mutations significantly reduced the mutation rate of rad27Δ cells, as measured by Canavanine resistance (Can^r) (Fig. 3C) but did not affect the mutation rate of yeast cells with WT Rad27 (fig. S12). These pol3 mutations nearly completely suppressed the occurrence of duplications (Fig. 3D). Consistent with the Can^r assay results, our WGS data confirmed that pol3 mutations reduced the frequency of duplications and the overall mutation frequency (Fig. 3E). Duplication mutation rate correlates with the level of 5’ flap formation (12). Thus, our biochemical and genetic results demonstrate that pol3 ITD and other point mutations can reverse the conditional lethality phenotype by limiting 5’ flap formation in rad27Δ cells.

To identify the signaling pathways that induced 3’ flap- mediated OFM and led to generation of pol3 ITD, we compared the transcriptomes of WT and rad27Δ cells grown at 37°C or 30°C. We observed that genes regulated by the checkpoint kinases Mec1, Rad53, and Dun1 were significantly up-regulated in rad27Δ cells, especially those grown at 37°C (Fig. 4A); consistent with this, western blot analysis confirmed that chromatin-associated Dun1 protein was increased in rad27Δ cells grown at 37°C (Fig. 4B). These results suggest activation of the Mec1-Rad53-Dun1 axis, the major signaling pathway that is activated to counteract genotoxic stress (17, 18). We further showed that downstream targets of the upregulated genes, including the stress response genes HUG1, RNR2, RNR3, and RNR4, and the DNA repair gene RAD51, were synergistically induced by rad27Δ and restrictive temperature stress (fig. S13). RAD51 is associated with inhibition of 3’ ssDNA degradation, which at least partially explains why degradation of 3’ flaps induced by NE from rad27Δ cells grown at 37°C was markedly less than degradation induced by WT NE (fig. S7A, 7B).

Fig. 4. Restrictive temperature stress activates signaling that facilitates error-prone OFM and generation of rad27Δ revertants.

(A) Up-regulated or down-regulated gene ratios in Dun1-, Rad53-, Mec1-, or Tell-controlled pathways in WT or rad27Δ yeast cells exposed to 30°C (4 h) or 37°C (4 h). p values were calculated using the hypergeometric test; n.s., not significant. (B) Top, western blot of chromatin-associated Dun1 protein in WT or rad27Δ cells exposed to 30°C (4 h) or 37°C (4 h). Histone H2B was used as a loading control for the chromatin fraction in each sample. dun1Δ is a negative control. Bottom, quantification of chromatin-associated Dun1 relative to the loading control. The Dun1 level in rad27Δ cells grown at 30°C was arbitrarily set as 1, and the relative Dun1 levels in other samples were calculated by dividing their Dun1 levels by that in rad27Δ cells grown at 30°C. Error bars indicate s.e.m (n=4 biological replicates). (C) Levels of 3’ flaps in genomic DNA from WT, dun1Δ, rad27Δ, or rad27Δ dun1Δ double-mutant cells grown at 30°C or 37°C for 4 h. (D) Mean Can^r mutation rates of WT, rad27Δ, dun1Δ, and rad27Δ dun1Δ cells exposed to 30°C (4 h) or 37°C (4 h). Error bars indicate s.d. (n=3 independent assays). (E) Median revertant frequencies of rad27Δ or rad27Δ dun1Δ cells (n=3 independent assays). (F) Percentage of rad27Δ or rad27Δ dun1Δ revertants that carry a pol3 mutation (n=19 for each strain). (G) Schematic illustrating error-free 5’ flap-mediated OFM in WT cells, error-prone 3’ flap-mediated OFM and the corresponding consequences in rad27Δ cells under restrictive temperature stress (37°C), and the impact of Polδ-ITD on OFM in the revertant. Blue and pink segments, primase/Polα-synthesized RNA primer and DNA. Red segment, Polδ-synthesized DNA, replacing Polα-synthesized DNA. Green segment, Polδ-synthesized spacer DNA as part of alternative duplication. Black dots, Polα incorporation errors.

To define the role of Dun1 signaling in stress-induced mutation and generation of revertants, we deleted the DUN1 gene in WT and rad27Δ cells. We observed that knockout of DUN1 (dun1)Δ in WT or rad27Δ cells had little effect on their survival (fig. S14), 3’ flap formation (Fig. 4C), or mutation rate at 30°C (Fig. 4D). However, DUN1 deletion markedly reduced restrictive temperature stress-induced 3’ flap formation (Fig. 4C) and abolished restrictive temperature stress-induced mutations in rad27Δ cells (Fig. 4D). Consistent with this, DUN1 deletion inhibited generation of rad27Δ revertants (Fig. 4E, Supplementary text S6). Furthermore, all rad27Δ revertants in this experiment had pol3 mutations, predominantly the pol3 458–477 ITD, but none of the rad27Δ dun1Δ revertants had pol3 mutations (Fig. 4F). These findings suggest that Dun1 activation plays an important role in the development of restrictive temperature stress-induced mutations that can reverse the lethal phenotype of rad27Δ cells. Consistent with this finding, blocking activation of Chk1, a Dun1 functional analogue, significantly inhibited spontaneous lung cancer development in FEN1 mutant mice but not in WT mice (fig. S15, Supplementary text S7). An important function of Dun1 activation is to induce overexpression of HUG1, RNR2, RNR3, and RNR4 for deoxyribonucleotide production. Increased deoxyribonucleotide concentrations changed the mode of action of Polδ and promoted generation of ligated extended products in vitro (fig. S5, fig. S16, S17, Supplementary text S8). However, when we deleted SML1, the protein inhibitor of ribonucleotide reductase (19), to increase deoxyribonucleotide production, we did not observe increased mutation rates in rad27Δ cells (fig. S18), suggesting that an up-regulation of deoxyribonucleotide alone is not sufficient to promote alternative duplications.

To demonstrate the relevance of stress-induced 3’ flap-based OFM and alternative duplications in rad27Δ cells to human cancers, we used whole-exome sequencing (WES) to analyze alternative duplications in human tumors and mutant mice modeling human FEN1 mutations. Alternative duplications, similar to those in rad27Δ cells grown at restrictive temperature (i.e., 3’ flap OFM-related alternative duplications), were frequent in human B cell acute lymphoblastic leukemia (fig. S19A-19C, Supplementary text S9). In addition, FEN1 A159V mutation, which occurs in human lung cancers (20), promoted 3’ flap OFM-related alternative duplications in mice (fig. S19D, Supplementary text S9). Therefore, mutations in FEN1 or other OFM genes may lead to 3’ flap-based OFM, and play crucial roles for cancer cell evolution, tumor growth, and resistance.

Our current study defines error-prone processing of RNA-DNA primers during OFM (Fig. 4G). Induction of this mechanism generates alternative duplications and base substitutions. In WT cells, the displaced 5’ RNA-DNA flap is effectively cleaved by either Rad27 alone or by Dna2, which first cleaves the 5’ RNA-DNA flap in the middle, leaving a shorter 5’ DNA flap for Rad27 to subsequently cleave. When Rad27 is not available, other 5’ nucleases such as Dna2 alone or Exo1 are involved in inefficient 5’ flap processing (21, 22). Resolution of 5’ flaps also requires an alternative pathway that is mediated by the 3’ exonuclease activities of Polδ, which removes nucleotides from the 3’ end of an upstream Okazaki fragment, generating a gap for the unprocessed 5’ flap to re-anneal for ligation (16, 23). Restrictive temperature stress activates Dun1 signaling and stimulates de novo production of deoxyribonucleotides, which in turn inhibits the 3’ exonuclease activity, but not the flap nuclease activity of Polδ, and induces other DNA damage responses. These molecular changes push OFM toward transformation of an unprocessed 5’ flap into a 3’ flap, either through flap equilibration (24) or the actions of helicases such as Sgs1 or Pif1, leading to a secondary structure that may result in alternative duplications, including Polδ-ITD, in revertant strains. In the revertants, Polδ mutations limit DNA displacement, thus suppressing 5’ flap formation or allowing more time for Dna2 or Exo1 to act on the 5’ flap and bypass the requirement for Rad27 (Fig. 4G).

Data and materials availability:

All data is available in the manuscript or the supplementary materials. Accession numbers for mouse and yeast genomics datasets are GSE181154 and GSE178876, respectively.