Evidence that DNA polymerase δ contributes to initiating leading strand DNA replication in Saccharomyces cerevisiae

Abstract

To investigate nuclear DNA replication enzymology in vivo, we have studied Saccharomyces cerevisiae strains containing a pol2-16 mutation that inactivates the catalytic activities of DNA polymerase ε (Pol ε). Although pol2-16 mutants survive, they present very tiny spore colonies, increased doubling time, larger than normal cells, aberrant nuclei, and rapid acquisition of suppressor mutations. These phenotypes reveal a severe growth defect that is distinct from that of strains that lack only Pol ε proofreading (pol2-4), consistent with the idea that Pol ε is the major leading-strand polymerase used for unstressed DNA replication. Ribonucleotides are incorporated into the pol2-16 genome in patterns consistent with leading-strand replication by Pol δ when Pol ε is absent. More importantly, ribonucleotide distributions at replication origins suggest that in strains encoding all three replicases, Pol δ contributes to initiation of leading-strand replication. We describe two possible models.

Fig. 1

Pol ε catalytic domains are critical for yeast growth. a Schematic representation of DNA Polymerase ε (Pol ε). The S. cerevisiae holoenzyme consists of the catalytic subunit (Pol2p) and three auxiliary subunits: Dpb2p, Dpb3p, and Dpb4p^–. Cryo-electron microscopy has shown that Pol2p has two lobes tethered by a flexible linker. Active polymerase and exonuclease domains are in the N-terminal lobe. The pol2-16 mutant has an in-frame deletion of the fragment of catalytically active lobe (amino acids 176–1134). b, c Tetrad analysis of pol2-16/POL2 and pol2-4/POL2 heterozygous diploids in two yeast backgrounds, ∆7, and W303, at 23 °C b and 30 °C c. 1–12 are dissected tetrads, A–D, and a–d are haploid spore colonies. Images were taken after 3 and 12 days. Genotypes were confirmed via PCR (pol2-16, red circles) or sequencing (pol2-4, blue). Wild-type colonies are circled in green. The lack of both Pol ε catalytic domains (pol2-16) causes severe growth defects. Exonuclease inactivation alone (pol2-4) does not. d Microscopic images of pol2-16 colonies taken 3 days after tetrad dissections. e Doubling times of pol2-16 and pol2-4 mutants compared to wild-type yeast. Doubling times were estimated from optical density at 600 nm of cultures grown at 23 °C. Error bars represent standard deviations (n = 4–6 yeast cultures, two or three from two independent isolates). Unpaired two-tailed t tests with Welch’s correction yielded p values (P). The doubling time of the pol2-16 mutant is about threefold longer than of the wild-type and pol2-4 yeast in both ∆7 and W303 backgrounds. The difference in the doubling times between the wild-type ∆7 and W303 backgrounds may be due to one or more of over 10,000 SNPs detected by the whole-genome sequencing. f Western blot detection of Pol2p level in whole-cell extracts. Presented are bands for three independent isolates of strains bearing POL2 or pol2-16 in fusion with TAP-tag. Immunoblotting was performed using an antibody to TAP-tag or PSTAIR (loading control). g Relative band intensity. Error bars represent standard deviations (n = 6–7 independent yeast isolates). Unpaired two-tailed t tests with Welch’s correction yielded p values (P)

Fig. 2

Phenotypes of wild-type, pol2-4, and pol2-16 mutants. a Confocal microscope images of DAPI stained, exponentially growing yeast in the synthetic complete media, fixed with 70% ethanol. The pol2-16 cells are much bigger than wild-type and pol2-4 cells and have aberrantly distributed DNA. b Flowcytometric analysis of cell cycle progression. Yeast cells were α-factor arrested in G1 phase, then released into complete media, sampled every 20 min, fixed with 70% ethanol, stained with propidium iodide, and then analyzed with flow cytometry (Methods). 1C and 2C indicate the DNA contents. c Colony size heterogeneity in the pol2-16 outgrowths. Resuspended pol2-16 and POL2 spore colonies, were plated on solid complete media and incubated at 23 °C for 6 and 4 days, respectively

Fig. 3

Ribonucleotides as biomarkers of replicase activity. Meta-analysis of ribonucleotides abundance in the vicinity of 214 replication origins analyzed in previous study, in bins of 5 bp, with 10-bin moving average trend lines. Read counts were normalized as described in Supplementary Methods. Shown are ratios of RER-deficient (RER⁻) and RER-proficient (RER⁺) strains bearing a pol1L868M, b pol2M644G, c pol3L612M, e pol3LM pol2-4, and f pol3LM pol2-16 mutations. Values on the right above each chart are average lagging-over-leading strand biases (r_b) (see calculations in Supplementary Methods). For each panel, a representative of two independent measurements is presented. d Schematic representation of DNA replication in the vicinity of replication origin, depicting the leading and lagging strands orientations

Fig. 4

HydEn-seq-derived evidence for DNA polymerase δ participation in leading-strand synthesis at yeast origins. Red, green, and blue denote polymerases α, δ, and ε, respectively, or DNA tracts synthesized by same. Canonical Okazaki fragments (yellow), synthesized by Pols α (dark), and δ (light) are approximately positioned exemplars. S. cerevisiae origins (identified by Smith et al.) are oriented such that ARS consensus sequences (ACS) are 5′–3′ beginning at position 0. a Diamonds represent the fraction of DNA strand synthesized by DNA Pols α, δ, and ε (5 bp bins; calculated from rescaled and background subtracted HydEn-seq end densities in pol1-L868M rnh201Δ, pol3-L612M rnh201Δ, and pol2-M644G rnh201Δ strains; see Supplementary Methods). Data for both strands were averaged (opposite strand reflected around +45 bp, the axis of strand symmetry; gray dashed line). Solid curves are regression models. b The fraction of inter-polymerase transfer events outside of canonical Okazaki fragment synthesis (extracted from regression models). The mode of each curve (vertical black line) suggests the most frequent synthesis tract (colored bars above). c–i Schematics of two non-exclusive models of polymerase action at yeast replication origins. DNA strands (colored bars) have the same horizontal scale as in a and b; polymerases (ellipses) and CMG helicases (gray polygons) are exaggerated; other components are omitted. εN and εC indicate N-terminal catalytic and C-terminal CMG-binding Pol ε domains. c Head-to-head dsDNA-binding CMG helicases. d Helicases transition to ssDNA-binding and translocate past one another, N-termini facing directions of travel (gray arrows). e, f Model 1. e Pol α associated with each replisome primes the leading strand that will be synthesized by the other (0.1% probability per bp translocated, from regression). Pol δ extends each nascent-leading strand. f Pols δ collide with respective replisomes, releasing 3′-termini to Pols ε. They assuming synthesis conformation to extend the leading strand. g, h Model 2. g Unidentified Pols α prime leading-strand replication (*; extended by Pols ε). The Pol α associated with each replisome primes the first Okazaki fragment (extended by Pol δ; destined to be longer than average). h Pol δ displaces the 5′ primer terminus of the nascent-leading strand, allowing nick translation or flap excision. i Synthesis patterns from both pathways indicate apparent Pol δ synthesis of both nascent strands at the origin