Tracking of progressing human DNA polymerase δ holoenzymes reveals distributions of DNA lesion bypass activities

Cryo-EM structure of apo-form human DNA polymerase δ elucidates its minimal DNA synthesis activity without PCNA

DNA polymerase δ (Pol δ) is a key enzyme in eukaryotic DNA replication and genome maintenance, essential for lagging strand synthesis, leading strand initiation, and DNA repair. While human Pol δ exhibits high activity and processivity in its holoenzyme form complexed with proliferating cell nuclear antigen (PCNA), it shows minimal DNA synthesis activity without PCNA, the molecular basis of which remains unclear. Here, we present the cryo-EM structure of the apo-form human Pol δ, comprising the catalytic subunit p125 and regulatory subunits p66, p50, and p12, at an overall resolution of 3.65 Å. We identified an acidic α-helix at the N terminus of p125, which occupies the single-stranded DNA-binding cavity within the polymerase domain in the apo-form Pol δ. This interaction likely inhibits DNA binding in the absence of PCNA, explaining the low activity of apo-form Pol δ. The acidic α-helix is absent in yeast Pol δ, providing a molecular explanation for species-specific differences in PCNA-independent Pol δ activity. These findings provide critical insights into the regulatory mechanisms of Pol δ and its reliance on PCNA for efficient DNA synthesis.

A human high-fidelity DNA polymerase holoenzyme has a wide range of lesion bypass activities

During replication, lagging strand lesions are initially encountered by high-fidelity DNA polymerase (pol) holoenzymes comprised of pol δ and the PCNA sliding clamp. To proceed unhindered, pol δ holoenzymes must bypass lesions without stalling. This entails dNMP incorporation opposite the lesion (insertion) and the 5' template nucleotide (extension). Historically, it was viewed that high-fidelity pol holoenzymes stall upon encountering lesions, activating DNA damage tolerance pathways that are ultimately responsible for lesion bypass. Our recent study of 4 prominent lesions revealed that human pol δ holoenzymes support insertion and/or bypass for multiple lesions and the extents of these activities depends on the lesion and pol δ proofreading. In the present study, we expand these analyses to other prominent lesions. Collectively, analyses of 10 lesions from both studies reveal that the insertion and bypass efficiencies of pol δ holoenzymes each span a complete range (0 - 100%). Consequently, the fates of pol δ holoenzymes upon encountering lesions are quite diverse. Furthermore, pol δ proofreading promoted holoenzyme progression at 7 of the 10 lesions and did not deter progression at any. Altogether, the results significantly alter our understanding of the replicative capacity of high-fidelity pol holoenzymes and their functional role(s) in lesion bypass.

Replication of [AT/TA]25 Microsatellite Sequences by Human DNA Polymerase δ Holoenzymes Is Dependent on dNTP and RPA Levels

Fragile sites are unstable genomic regions that are prone to breakage during stressed DNA replication. Several common fragile sites (CFS) contain A+T-rich regions including perfect [AT/TA] microsatellite repeats that may collapse into hairpins when in single-stranded DNA (ssDNA) form and coincide with chromosomal hotspots for breakage and rearrangements. While many factors contribute to CFS instability, evidence exists for replication stalling within [AT/TA] microsatellite repeats. Currently, it is unknown how stress causes replication stalling within [AT/TA] microsatellite repeats. To investigate this, we utilized FRET to characterize the structures of [AT/TA]25 sequences and also reconstituted lagging strand replication to characterize the progression of pol δ holoenzymes through A+T-rich sequences. The results indicate that [AT/TA]25 sequences adopt hairpins that are unwound by the major ssDNA-binding complex, RPA, and the progression of pol δ holoenzymes through A+T-rich sequences saturated with RPA is dependent on the template sequence and dNTP concentration. Importantly, the effects of RPA on the replication of [AT/TA]25 sequences are dependent on dNTP concentration, whereas the effects of RPA on the replication of A+T-rich, nonstructure-forming sequences are independent of dNTP concentration. Collectively, these results reveal complexities in lagging strand replication and provide novel insights into how [AT/TA] microsatellite repeats contribute to genome instability.

PRIMPOL ensures robust handoff between on-the-fly and post-replicative DNA lesion bypass

The primase/polymerase PRIMPOL restarts DNA synthesis when replication is arrested by template impediments. However, we do not have a comprehensive view of how PRIMPOL-dependent repriming integrates with the main pathways of damage tolerance, REV1-dependent 'on-the-fly' lesion bypass at the fork and PCNA ubiquitination-dependent post-replicative gap filling. Guided by genome-wide CRISPR/Cas9 screens to survey the genetic interactions of PRIMPOL in a non-transformed and p53-proficient human cell line, we find that PRIMPOL is needed for cell survival following loss of the Y-family polymerases REV1 and POLη in a lesion-dependent manner, while it plays a broader role in promoting survival of cells lacking PCNA K164-dependent post-replicative gap filling. Thus, while REV1- and PCNA K164R-bypass provide two layers of protection to ensure effective damage tolerance, PRIMPOL is required to maximise the effectiveness of the interaction between them. We propose this is through the restriction of post-replicative gap length provided by PRIMPOL-dependent repriming.

Replication of [AT/TA]25 microsatellite sequences by human DNA polymerase δ holoenzymes is dependent on dNTP and RPA levels

Difficult-to-Replicate Sequences (DiToRS) are natural impediments in the human genome that inhibit DNA replication under endogenous replication. Some of the most widely-studied DiToRS are A+T-rich, high "flexibility regions," including long stretches of perfect [AT/TA] microsatellite repeats that have the potential to collapse into hairpin structures when in single-stranded DNA (ssDNA) form and are sites of recurrent structural variation and double-stranded DNA (dsDNA) breaks. Currently, it is unclear how these flexibility regions impact DNA replication, greatly limiting our fundamental understanding of human genome stability. To investigate replication through flexibility regions, we utilized FRET to characterize the effects of the major ssDNA-binding complex, RPA, on the structure of perfect [AT/TA]25 microsatellite repeats and also re-constituted human lagging strand replication to quantitatively characterize initial encounters of pol δ holoenzymes with A+T-rich DNA template sequences. The results indicate that [AT/TA]25 sequences adopt hairpin structures that are unwound by RPA and pol δ holoenzymes support dNTP incorporation through the [AT/TA]25 sequences as well as an A+T-rich, non-structure forming sequence. Furthermore, the extent of dNTP incorporation is dependent on the sequence of the DNA template and the concentration of dNTPs. Importantly, the effects of RPA on the replication of [AT/TA]25 sequences are dependent on the concentration of dNTPs, whereas the effects of RPA on the replication of an A+T-rich, non-structure forming sequence are independent of dNTP concentration. Collectively, these results reveal complexities in lagging strand replication and provide novel insights into how flexibility regions contribute to genome instability.

Expression Profile, Molecular Association, and Clinical Significance of POLD4 in Glioblastoma

The POLD4 gene encodes a subunit (δ4) of DNA polymerase delta, which is a key enzyme involved in DNA replication and repair. Recent studies have suggested that POLD4 plays a crucial role in developing certain cancers. However, there is a lack of knowledge regarding the role of POLD4 in the context of glioblastoma (GBM). Therefore, in this study we have used various cancer bioinformatics tools to explore the role of POLD4 in glioblastoma. Data from various sources were accessed to analyze POLD4 gene expression and estimate tumor-infiltrating immune cells in glioblastoma. Methylation data were retrieved using the MEXPRESS web browser and analyzed. UALCAN webserver was used to analyze the protein expression of POLD4. Gene correlation and pathway enrichment analysis were performed using cBioPortal and GSEA software, respectively. Afterward, survival analysis was performed. POLD4 was significantly upregulated in glioblastoma at both gene and protein levels in GBM, and ROC curve analysis revealed it as a potential biomarker in glioblastoma. GSEA analysis of TCGA-GBM pan-cancer study exhibited that POLD4 expression was associated with critical pathways, such as interferon-gamma response, G2M checkpoint, inflammatory response, E2F targets, EMT transition, and KRAS signaling pathways. Furthermore, POLD4 expression was positively correlated with DNA methylation at 3 CpG sites, including Cg16509978, with a Pearson correlation coefficient value of 0.398 (p-value ≤ 0.01), while the promoter region had a positive correlation but was not significant. In addition, POLD4 is significantly linked with poor OS, PFS, and DFS. We also found association of POLD4 expression with altered immune cell infiltration. In conclusion, POLD4 is significantly upregulated in glioblastoma and may be used as a potential diagnostic or prognostic biomarker for GBM patients. However, to establish the same a large cohort study is needed. Using TCGA data and various cancer bioinformatics tools mentioned above we observed very high level of gene and protein expression of POLD4 in glioblastoma patients. The expression of POLD4 was significantly correlated with inflammatory and oncogenic pathways and it also has a significant correlation with adverse outcome in patients with glioblastoma.

OGG1 and MUTYH repair activities promote telomeric 8-oxoguanine induced cellular senescence

Telomeres are prone to formation of the common oxidative lesion 8-oxoguanine (8oxoG), and the acute production of 8oxoG damage at telomeres is sufficient to drive rapid cellular senescence. OGG1 and MUTYH glycosylases initiate base excision repair (BER) at 8oxoG sites to remove the lesion or prevent mutation. Here, we show OGG1 loss or inhibition, or MUTYH loss, partially rescues telomeric 8oxoG-induced senescence, and loss of both glycosylases results in a near complete rescue. Loss of these glycosylases also suppresses 8oxoG-induced telomere fragility and dysfunction, indicating that single-stranded break (SSB) intermediates arising downstream of glycosylase activity impair telomere replication. The failure to initiate BER in glycosylase-deficient cells suppresses PARylation at SSB intermediates and confers resistance to the synergistic effects of PARP inhibitors on damage-induced senescence. Our studies reveal that inefficient completion of 8oxoG BER at telomeres triggers cellular senescence via SSB intermediates which impair telomere replication and stability.