Structure-specific nucleases: role in Okazaki fragment maturation

Abstract

Proper function of structure-specific nucleases is key for faithful Okazaki fragment maturation (OFM) process completion. Deregulation of such nucleases leads to aberrant OFM and causes a spectrum of mutations, some of which may confer survival outcomes under specific stresses and serve as attractive targets for therapeutic intervention in human cancers.

Introduction

Despite its incredible importance, the OFM (see Glossary) process is presently an understudied area in modern biology and medicine, as are the consequences of perturbation of the process and the potential to identify new targets for cancer therapeutics. During the genome duplication process, while the leading strand is replicated continuously, the lagging strand is synthesized discontinuously in forms of DNA fragments called Okazaki fragments (OFs). Each OF contains an RNA-DNA primer that is synthesized by the primase/DNA polymerase alpha (Pol α) complex and a DNA fragment that is synthesized by DNA polymerase delta (Pol δ). The RNA primer must be removed to join OFs into intact lagging-strand DNA. Meanwhile, Pol α has no proofreading function, leaving a high incidence of errors, which must be removed to avoid DNA mutations. RNA primer removal and Pol α error editing primarily depend on the 5′ flap endonuclease family, typified by flap endonuclease 1 (FEN1). During OFM, Pol δ-driven DNA synthesis of the upstream OF displaces the RNA-DNA primer into a 5′ flap (Figure 1), which is effectively cleaved by FEN1 to create a ligatable DNA nick for DNA ligase I (LIG1) to join the OFs. In all of these reactions, proliferating cell nuclear antigen (PCNA) is important in coordinating the sequential enzymatic actions. In some cases, a long 5′ flap forms due to PIF1 helicase activity. The long 5′ flap is bound by the single-strand DNA (ssDNA) binding protein RPA (replication protein A), which inhibits FEN1 activity but recruits and stimulates another structure-specific nuclease, DNA2. Cleavage of the long 5′ flap by DNA2 produces a short flap substrate for FEN1. In the absence of FEN1, the 5′ exonuclease 1 (EXO1) may substitute FEN1 for OFM (Figure 1) [1]. Defective OFM leads to a genome-wide high incidence of mutations. Deletion of the Saccharomyces cerevisiae FEN1 homolog RAD27 gene results in a strong mutator phenotype at the permissive growth temperature 30°C but lethality at the restrictive temperature of 37°C, suggesting that the stress condition may have an influence on OFM [2]. Intriguingly, rad27Δ is synthetically lethal with genes encoding 3′ nucleases including Pol3, Mus81, Rad1, and Mre11. It suggests an alternative OFM pathway mediated by these 3′ nucleases. The mutant is also synthetically lethal with double-strand-break (DSB) repair genes such as RAD51 and RAD52. It was previously proposed that the homology-directed repair (HDR) pathway may play an important role in repairing the DSBs resulting from unligated 5′ flaps in FEN1-deficient cells.

Figure 1. Structure-specific nuclease activity during Okazaki fragment maturation (OFM) processes in human cells.

OFM (light-blue background): lagging strand is synthesized discontinuously in the form of Okazaki fragments (OFs), comprising an RNA-DNA primer and a DNA polymerase δ (Pol δ)-synthesized DNA segment. During unperturbed OFM, Pol δ-driven DNA synthesis of the upstream OF displaces the RNA-DNA primer into a 5′ flap. These 5′ flaps are cleaved by flap endonuclease 1 (FEN1) to create ligatable DNA nicks, which are then joined by DNA ligase I (LIG1) to yield a continuous DNA strand. Long 5′ flaps are trimmed by DNA2 for subsequent processing by FEN1. Exonuclease 1 (EXO1) has both 5′-to-3′ exonuclease activity and flap endonuclease activity and may back up FEN1 activity. FEN1-defective stress-induced error-prone OFM (gray background): in the absence of FEN1, the 5′ flap bearing DNA polymerase α (Pol α)-incorporated errors can generate duplication mutations potentially through microhomology-mediated annealing [2]. Additionally, Pol δ with internal tandem duplications (Pol δ•) fails to generate a 5′ flap. RNase H2 can remove the RNA segment of the RNA-DNA primer leaving behind Pol α-incorporated errors. The resulting gap is then filled by Pol δ. Under restrictive temperature stress, activation of Dun1 facilitates the transformation of these unprocessed 5′ flaps into 3′ flaps potentially by the action of 5′-to-3′ helicases and translocases. 3′ Flaps can either be removed by 3′ flap nucleases to generate ligatable nicks for ligation or form secondary structures that give rise to alternative duplication mutations. If the 3′ flap is coated with RAD51 and protected from the nuclease attack, it may invade the sister chromatin DNA in the fork via the template switching (TS) pathway. 5′-to-3′ flap transformation fails to remove Pol α-incorporated errors and generates point mutations independent of the mechanism of 3′ flap resolution. DNA gap repair via TS (light-green background): G-quadruplexes (G4s) obstruct lagging-strand synthesis, causing gaps in the nascent strand. 5′ and 3′ nucleases expand the gaps, allowing helicases and translocases to generate 3′ flaps. These flaps may then invade into the sister chromatid via homology search and complete DNA synthesis using the undamaged sister chromatid as the template (TS). The resulting Holliday junctions (HJs) can be resolved by HJ resolvases. G4s may eventually resolve when the extended daughter strand pairs. Additionally, G4s can be resolved by G4 resolvases [17]. Poly (ADP-ribose) polymerase (PARP) trapping by PARP inhibitors can also induce DNA lesions that may be resolved by TS. Figure created with BioRender.com. Abbreviation: PCNA, proliferating cell nuclear antigen.

3′ Flap-based OFM mediated by 3′ flap nuclease is an error-prone alternative pathway for cell survival under DNA replication stress

Shen group has recently reported a 3′ flap-based OFM pathway [3]. It was observed that a restrictive temperature activates the cell cycle checkpoint protein Dun1, facilitating the transformation of unprocessed 5′ flaps into 3′ flaps, which are further removed by 3′ flap nucleases to generate ligatable nicks for ligation (Figure 1). Pol δ 3′ nuclease activity may play the primary role in degrading 3′ flaps for OFM. It is speculated that 3′ nucleases including Mre11, the Rad1/Rad10 complex, Mus81, and their human counterparts may participate in the removal of the 3′ flap with or without secondary structures such as hairpins. This is consistent with previous findings indicating synthetic lethality or sickness between Rad27 deletion and deficiency in either of these 3′ nucleases in yeast. However, stress conditions also induce overexpression of Rad51, which has been shown to protect 3′ ssDNA from degradation. If the 3′ flap is coated with Rad51 and protected from the nuclease attack, it may invade the sister chromatin DNA in the fork and initiate OFM via the template-switching (TS) pathway, which requires homolog recombination proteins. However, at certain regions, 3′ flaps may form secondary structures before Rad51 coating or undergo microhomology-mediated annealing. Extension of such intermediate DNA structures and subsequent DNA ligation produce duplications with short spacer sequences, such as an internal tandem duplication (ITD) in Pol3 (human Pol δ). These Pol3 mutations hinder DNA displacement and suppress 5′ flap formation, which ultimately rescues rad27Δ cells. In either case, the 3′-flap-mediated OFM failed to remove the Pol α errors, generating point mutations in the genome (Figure 1). Characterization of the induced di-directional OFM pathway represents our expanded knowledge on the cellular strategy for survivorship under stringent DNA replication stresses.

Nuclease and helicase complexes process unligated OFs at G-quadruplex (G4) sites via the TS pathway

G-rich regions in the genome, such as telomeres and centromeres, can form secondary structures such as G4s when the DNA duplex opens, posing an additional challenge for lagging-strand DNA synthesis and OFM [4]. Daughter strands (lagging strands) are now synthesized in the manner of being trans-lesion, missing genetic information contained in the unresolved secondary structures. Use of the undamaged sister chromatid as a template (TS) for lesion bypass is an effective strategy to complete DNA replication without loss of genetic information in the daughter strand [5]. Various structure-specific nucleases play essential roles in the proper completion of TS DNA synthesis. The TS process in the DNA replication fork involves multiple nucleases in four steps. The first step is the nuclease-driven widening of the gap. The Sae2–MRX (human CtIP–MRN) endo- and exonuclease complex initiates resection at short range, while Exo1 (human EXO1), Sgs1-Dna2 (human BLM-DNA2) and possibly EXO5 redundantly generate ssDNA of up to several tens of kilobases [6,7]. The second step is the helicase-catalyzed 3′ ssDNA flap generation. The 5′–3′ helicase PIF1 may bind to the template strand of the ssDNA gap and migrate in the 5′-to-3′ direction to generate a 3′ flap in the complementary strand. Next, the 3′ flap may invade into the sister chromatin to initiate HDR (Figure 1). Ubiquitinated PCNA and Rad51 protect the single-strand 3′ flap from nuclease cleavage to facilitate this process [8]. Recently, it was found that RNA coated with RAD51AP1 opens the sister chromatin in the fork forming a bubble structure called the R-loop, which is important for TS [9]. TS and DNA synthesis result in four-way DNA intermediates called Holliday junctions (HJs). A group of ubiquitous and highly specialized structure-selective endonucleases catalyze the cleavage of HJs into two disconnected DNA duplexes in a reaction called HJ resolution. These enzymes, called HJ resolvases, have been identified in bacteria and their bacteriophages, archaea, and eukaryotes. In humans, the HJ resolvases include GEN1, SLX1-SLX4, and MUS81-EME1 [10]. Meanwhile, the G4 may be resolved as the extended daughter strand comes back to pair with it (Figure 1).

Coordination and regulation of OFM dynamics

Efficient, accurate processing of up to 50 million OFs per cell cycle is primarily regulated via post-translational modifications (PTMs) and protein–protein interactions. It was elucidated that sequential PTMs program the actions and degradation of FEN1 in OFM during cell-cycle progression [11]. In addition, PCNA mediates OFM through tight coordination of the activities of Pol δ, FEN1, and Lig1 by sequential binding [12]. The generation of numerous OFs during lagging-strand DNA synthesis also results in genome-wide nick sites that serve as the binding sites for poly (ADP-ribose) polymerases (PARPs), which have recently emerged as important in OFM and gap repair in S phase [13]. Chemical inhibition of these enzymes in the clinic selectively kills HDR-defective cancer cells (e.g., those harboring mutations in BRCA1 or BRCA2). In recent years, an increased number of the genes involved in the HDR pathway have shown similar cellular responses to PARP inhibitors. This phenomenon is termed BRCAness. In cells with BRCAness, HDR is inefficient. The DNA single-strand gaps, DSBs, and aberrant DNA replication fork structures that can arise from unligated OFs are potential sources of chromosome degradation, rearrangement, and mitotic catastrophe. This concept also raises possible new avenues for anticancer therapy. For example, perhaps the use of PARP inhibitors can be extended beyond HDR-defective tumors to include cancers in which OFM is perturbed [14]. The observations that FEN1-deficient cells are hypersensitive to PARP inhibitors and that PARP-deficient cells are hypersensitive to FEN1 or DNA2 inhibitors are consistent with this idea [15]. Therefore, even the judicious use of combinations of inhibitors that target both PARP enzymes and FEN1/DNA2 may have therapeutic value, especially for HDR-defective cancers.

Concluding remarks

The recent elucidation of an error-prone, stress-induced OFM pathway is analogous to the mechanism in cancer cell evolution and drug resistance. The tumor mutation burden (TMB), caused due to the hijacking of DNA replication machinery components by drug treatment [16], is always associated with drug resistance in the clinic. Therefore, the information obtained from the yeast model system may give us hints to develop combination cancer therapeutic regimens to prevent the activation of mutagenesis pathways and drug resistance.

Glossary

BRCAness

defects in the HDR pathway that mimic the phenotypic traits of the loss of BRCA1 and BRCA2.

Holliday junction (HJ)

a four-way intermediate generated during homologous recombination. Resolution of Holliday junctions can lead to genetic crossover.

Homology-directed repair (HDR)

repair of DNA DSBs using templates with sufficient sequence identity to the damaged sequence.

Internal tandem duplication (ITD)

in-frame insertions of duplicated sequences in coding regions of genes.

Okazaki fragment maturation (OFM)

a multistep process by which RNA primers are removed and DNA Pol α-incorporated errors are edited out to yield an intact lagging-strand DNA.

Poly (ADP-ribose) polymerases (PARPs)

a family of enzymes that catalyze the transfer of ADP-ribose to target proteins, which has important roles in cellular processes such as DNA replication, repair, recombination, and transcription.

Structure-specific nucleases

nucleases that recognize specific DNA structures rather than specific DNA sequences or modifications as substrates.

Template switching (TS)

method of lesion bypass in which the DNA strand with the lesion invades into the sister chromatid allowing the extension of the stalled nascent strand using the newly synthesized sister chromatid as the template.

Tumor mutation burden (TMB)

frequency of sporadic somatic mutations in cancer tissues.