The role of RNase H2 in processing ribonucleotides incorporated during DNA replication

Abstract

Saccharomyces cerevisiae RNase H2 resolves RNA-DNA hybrids formed during transcription and it incises DNA at single ribonucleotides incorporated during nuclear DNA replication. To distinguish between the roles of these two activities in maintenance of genome stability, here we investigate the phenotypes of a mutant of yeast RNase H2 (rnh201-RED; ribonucleotide excision defective) that retains activity on RNA-DNA hybrids but is unable to cleave single ribonucleotides that are stably incorporated into the genome. The rnh201-RED mutant was expressed in wild type yeast or in a strain that also encodes a mutant allele of DNA polymerase ε (pol2-M644G) that enhances ribonucleotide incorporation during DNA replication. Similar to a strain that completely lacks RNase H2 (rnh201Δ), the pol2-M644G rnh201-RED strain exhibits replication stress and checkpoint activation. Moreover, like its null mutant counterpart, the double mutant pol2-M644G rnh201-RED strain and the single mutant rnh201-RED strain delete 2–5 base pairs in repetitive sequences at a high rate that is topoisomerase 1-dependent. The results highlight an important role for RNase H2 in maintaining the genome integrity by removing ribonucleotides incorporated during DNA replication.

1. Introduction

Genome stability is ensured by the chemical structure of DNA that includes the absence of a 2′ hydroxyl group on the ribose sugar that renders DNA approximately 100,000 times less susceptible than RNA to spontaneous hydrolysis under physiological conditions [1]. Nonetheless, ribonucleoside monophosphates (rNMPs) are occasionally incorporated by DNA polymerases [2], including during DNA replication [3, 4]. These newly incorporated ribonucleotides are ultimately removed by cleavage on the 5′ side of a ribonucleotide in DNA by RNase H2 [3], to initiate a pathway known as ribonucleotide excision repair (RER) [5]. In addition, RNase H2 also resolves RNA-DNA hybrids (R-loops), which can form during transcription when the nascent mRNA hybridizes to the complementary DNA strand and creates a displaced non-transcribed strand (reviewed in [6–8]).

Ribonucleotides incorporated into DNA during replication and RNA-DNA hybrids formed during transcription can impede an advancing DNA or RNA polymerase [9–13]. They can also be incised by topoisomerase 1 (Top1), an enzyme critical for resolving DNA supercoils that accumulate during both replication and transcription [14] Top1-incision at ribonucleotides generates single-strand DNA breaks harboring DNA ends containing a 5′-OH and a 2′-3′-cyclic phosphate. This nick cannot be sealed by ligases, but it can be reversed by Top1 [15, 16]. However, when nicking occurs in a short tandemly repeated sequence, a second Top1-dependent cleavage can occur and result in loss of one of the repeated sequences [17]. Thus, deletion of the gene encoding the catalytic subunit of yeast RNase H2 (RNH201) causes a high rate of 2–5 base pair deletions in short tandem-repeat sequences [18, 19]. The mutation rate for such events is enhanced in a strain with a mutation (pol2-M644G) in DNA polymerase ε (Pol ε) that catalyzes an increased level of ribonucleotide incorporation into the nascent leading strand DNA [3]. In addition to short deletion mutagenesis, complete loss of RNase H2 activity in yeast also causes other forms of genome instability that include altered distribution of cells in the cell cycle, sensitivity to replication stress, spontaneous checkpoint activation, elevated recombination and elevated rates of copy number variation and gross chromosomal rearrangements. These phenotypes depend on Top1 [18, 20–22]. Because complete loss of RNase H2 activity elicits all these phenotypes and abolishes its activity on both single ribonucleotides in DNA and on R-loops, it has been difficult to assign the genome instability phenotypes to a specific ribonucleotide-containing DNA structure.

Here we examine the phenotypes and the point mutation rates and specificity in strains of Saccharomyces cerevisiae using an RNase H2 mutant lacking the ability to remove single ribonucleotides embedded in DNA (rnh201-RED) but retaining the ability to process RNA-DNA hybrids [23]. The objective is to further define the biological consequences of failure of RNase H2-dependent removal of single ribonucleotides incorporated into genomic DNA. Specifically, we present data demonstrating that several phenotypes observed in a yeast strain containing a high density of unrepaired single ribonucleotides in the nascent leading strand are the consequence of Top1 processing of single unrepaired genomic ribonucleotides incorporated into DNA by Pol ε.

2. Materials and methods

2.1. Yeast strains

Saccharomyces cerevisiae strains are isogenic derivatives of strain Δ|(-2)|-7B-YUNI300 (MATa CAN1 his7-2 leu2Δ::kanMX ura3Δ trp1-289 ade2-1 lys2Δ GG2899-2900) [24]. Strain genotypes are listed in Supplemental Table S1. The URA3 reporter gene was introduced in orientation 2 (OR2) at position AGP1 [25]. The pol2-M644G mutator allele has been described previously [3, 26]. top1Δ strains were constructed by deletion-replacement of the endogenous TOP1 gene via transformation with a product containing the nourseothricin-resistance cassette (natMX4) amplified from pAG25 and flanked by 60 nucleotides of sequence homologous to the intergenic regions upstream and downstream of the TOP1 open reading frame. A plasmid containing Rnh201 C-terminally tagged with 5x-FLAG and flanked by 600 bp of upstream and downstream sequences was mutagenized using the Quikchange II site-directed mutagenesis kit (Agilent Technologies). Following PCR amplification, this construct was transformed into yeast to replace RNH201. Gene replacement was confirmed by marker selection and PCR analysis. The presence of the P45D and Y219A mutations was confirmed by sequencing and protein expression was confirmed by immunoblotting.

2.2. Phenotypic characterization

Strains were grown in rich medium (YPDA: 1 % yeast extract, 2 % bacto-peptone, 250 mg/L adenine, 2 % dextrose, 2 % agar for plates). Doubling time (D_t) values were calculated from cultures in the logarithmic phase of growth in rich medium at 30 °C. The experiment was performed in triplicate, with data displayed as the mean D_t ± standard deviation. Spot assays were performed by plating 10-fold serial dilutions of exponentially growing cells onto YPDA plates in the absence or presence of 150 mM hydroxyurea (HU; Sigma H8627). Plates were photographed after 3 days incubation at 30 °C. Flow cytometry to determine DNA content was performed as previously described [20]. Histograms and plots are representative of two independent experiments.

2.3. Immunoblotting

Whole-cell extracts were prepared from exponentially growing cells as described [20]. Proteins were resolved on a 10 % Bis-Tris gel (Life Technologies) and western blotting was performed using Anti-FLAG M2 (Sigma F1804), anti-histone H3 (Abcam: ab1791) or anti-Rnr3 (Agrisera AS09 574) antibodies at a dilution of 1:1000. The anti-PSTAIR antibody (Sigma P7962) was used at 1:5000.

2.4. Spontaneous mutation rates and sequence analysis

Strains containing the URA3 reporter placed in orientation 2 (URA3-OR2) adjacent to ARS306 on chromosome III were used in all mutation rate measurements and sequencing analysis. Spontaneous mutation rates were measured by fluctuation analysis as described [25]. PCR amplification and sequencing of ura3 was performed using DNA isolated from single, independent 5-FOA-resistant colonies. Rates of individual mutation classes were calculated by multiplying the fraction of that mutation type by the total mutation rate for each strain.

2.5. Statistical analysis

All statistical analysis was performed using GraphPad Prism. Statistical analysis to determine P-values for doubling time measurements was performed using the unpaired Students t-test. Statistical analysis of comparisons between overall mutation rates was performed using a one-sided nonparametric Mann Whitney test.

3. Results

3.1. The ribonucleotide excision-defective (RED) mutant of RNase H2 is expressed equivalently regardless of Pol epsilon or Top1 status

We replaced the genomic copy of the gene encoding the catalytic subunit of yeast RNase H2, RNH201 with a C-terminal 5xFLAG epitope tag in either a wild-type (wt) or RER-defective version (P45D Y219; RED) of the protein. This allowed for detection of either version of Rnh201 in the strain backgrounds used in this study. As shown in Figure 1, Rnh201 was expressed equivalently in strains expressing wild-type Pol ε (POL2) and in strains expressing a mutator variant of Pol ε, pol2-M644G, that incorporates a high density of ribonucleotides into nascent leading strand DNA [20, 27, 28]. Because Top1 mutagenically processes ribonucleotides not properly removed from nascent leading strand DNA by RNase H2 [18, 20, 29], we also tested whether Top1 status affects Rnh201-wt or Rnh201-RED protein levels. It does not (Fig. 1). These data suggest that the ribonucleotide-dependent phenotypes seen in subsequent experiments (see below) are due to the inability of Rnh201 to cleave at single ribonucleotides rather than a consequence of reduced protein expression or stability.

Fig. 1.

Monitoring protein expression of a variant of RNase H2 that is unable to initiate cleavage at single genomic ribonucleotides. The Rnh201-wt and Rnh201-RED proteins are expressed at equivalent levels in the POL2 versus pol2-M644G strains +/− TOP1. Western blot detection of Rnh201-FLAG was performed on whole cell extracts prepared from the indicated strains using an antibody to the FLAG epitope or histone H3 (loading control).

3.2. Topoisomerase 1-dependent genome instability phenotypes are initiated by cleavage at unrepaired single genomic ribonucleotides

We previously demonstrated that a pol2-M644G rnh201Δ strain grows slowly and has several phenotypes associated with the accumulation of RNA-DNA damage, many of which are Top1-dependent. To test the contribution of unrepaired single ribonucleotides versus persistent RNA-DNA hybrids to slow growth and sensitivity to replication stress, we first measured doubling time in the rnh201-RED strain in a POL2 or pol2-M644G mutator allele background. In parallel, we also tested growth of the rnh201Δ mutant. In the presence of wild type Pol ε, growth was not affected in either RNase H2 mutant strain (Fig. 2A), but the doubling time was increased in the pol2-M644G rnh201-RED strain, similar to that of the pol2-M644G rnh201Δ mutant (Fig. 2A and Supplemental Fig. S1A). This suggests that in an RNase H2-deficient strain with a high level of ribonucleotides being incorporated by Pol ε during leading strand synthesis, slow growth is caused by the processing of unpaired single ribonucleotides rather than the accumulation of RNA-DNA hybrids.

Fig. 2.

Top1 cleavage at unrepaired single ribonucleotides affects multiple cellular phenotypes in a strain containing a high density of ribonucleotides in the nascent leading strand. (A) Expression of the RER-defective RNase H2 (Rnh201-RED) causes an increase in doubling time in the pol2-M644G leading strand mutator variant similar to that observed in an rnh201Δ strain. *p = 0.046; **p = 0.024 (two-tailed Student’s t test). (B) Histograms displaying cell cycle distribution in the indicated RNase H2-deficient strains. The horizontal axis corresponds to the fluorescence parameter and the vertical axis displays the number of cells. Plotted black lines represent the raw data and the smoothed data generated using ModFit software. The gray shaded areas represent cells in G₁ or G₂/M phases and the striped area represents cells in S phase. (C) Ten-fold serial dilutions of exponentially growing cells were spotted onto YPDA agar plates (untreated) or YPDA agar plates containing 150 mM HU. Two independent isolates of pol2-MG rnh201-RED and pol2-MG rnh201-RED top1Δ are displayed. pol2-MG; pol2-M644G. (D) Genome integrity checkpoint activation occurs in a pol2-M644G rnh201-RED strain in a TOP1-dependent manner. Western blot detection of Rnr3 was performed on whole cell extracts prepared from the indicated strains (all harboring the pol2-M644G allele) using an antibody to Rnr3 or PSTAIR (loading control). Increased Rnr3 protein level is a sensitive indicator of genome integrity checkpoint activation [42]. The positive control is an extract prepared from wild-type cells treated with 200 mM HU for 3h.

Cell cycle distribution was then measured in these strains by flow cytometry. The pol2-M644G and pol2-M644G rnh201Δ cells accumulate in S phase and G₂ [3, 20], and this is also observed for the pol2-M644G rnh201-RED mutant (Fig. 2B). Deleting TOP1 in a pol2-M644G rnh201-RED mutant results in a more pronounced G₁ peak and reduces the percentage of cells in G₂, similar to what is observed for a pol2-M644G rnh201Δ strain (Fig. 2B; [20]). This Top1-dependent effect on cell cycle distribution in the rnh201-RED mutant is consistent with the idea that Top1-processing of single unrepaired ribonucleotides causes replicative stress.

We then monitored sensitivity of the RNase H2-deficient strains to the replication inhibitor hydroxyurea (HU), which causes depletion of dNTP pools and replication fork stalling [30, 31]. Although all strains expressing wild-type Pol ε (POL2) grew normally in the presence of HU (Supplemental Fig. S1B), growth of the pol2-M644G rnh201-RED mutant was impaired on the plate containing 150 mM HU (Fig. 2C). Also, as previously demonstrated for a pol2-M644G rnh201Δ mutant [20], deletion of TOP1 alleviates HU-sensitivity in the pol2-M644G rnh201-RED strain (Fig. 2C).

Top1 activity in a pol2-M644G rnh201Δ strain is associated with spontaneous checkpoint activation [20, 32], presumably due to the release of toxic DNA repair intermediates following Top1 cleavage at unrepaired ribonucleotides. To test whether this genome integrity checkpoint activation is caused by unrepaired single ribonucleotides or unresolved RNA-DNA hybrids, the protein level of Rnr3, a subunit of ribonucleotide reductase, was measured by immunoblotting of whole cell extracts prepared from the pol2-M644G rnh201-RED strain. Similar to the pol2-M644G rnh201Δ strain, Rnr3 is modestly elevated in the rnh201-RED mutant, and this increase is alleviated by deletion of TOP1 (Fig. 2D). The protein level of Hug1, which also increases during S-phase checkpoint activation [33] was also monitored and found to be modestly increased in the pol2-M644G strain background in the rnh201Δ and rnh201-RED mutants in a Top1-dependent manner (Supplementary Fig. S1C,D). Taken together, these immunoblotting results are again consistent with the hypothesis that Top1 cleavage at single unrepaired ribonucleotides causes lesions that activate the checkpoint to halt cell cycle progression and promote DNA repair.

3.3. Short deletion mutagenesis in the Rnh201-RED mutant

Finally, we measured rates of spontaneous mutation to 5-FOA-resistance caused by changes in the URA3 reporter gene. As demonstrated previously [3], the mutation rate of the pol2-M644G rnh201Δ strain is elevated compared to that of the pol2-M644G single mutant (Table 1 and Supplemental Table S2). This increased rate is largely accounted for by a large increase in the rate of 2–5 base deletions in repetitive sequences in URA3. A similar increase in 2–5 base pair deletion mutagenesis was previously reported in the pol2-M644G rnh201-RED strain [23] using the CAN1 reporter gene. That result is reproduced here using the URA3 reporter gene (Table 1 and Fig. 3A). Importantly, the rate of these 2–5 base pair deletions in repetitive sequences was greatly reduced in the TOP1 deletion background. Equally importantly, a similar reduction in the rate of 2–5 base pair deletions was observed in the rnh201-RED top1Δ mutant strain containing the wild type POL2 gene (Table 1 and Fig. 3B). These data indicate that Top1 cleavage at single ribonucleotides incorporated during replication initiates a 2–5 base pair deletion mutagenesis in an RNase H2-defective background.

Table 1.

A comparison of overall and specific mutation rates for individual mutation classes between the pol2-M644G rnh201Δ and pol2-M644G rnh201-RED strains ± TOP1.

A comparison of overall and specific mutation rates for individual mutation classes between the pol2-M644G, the rnh201-RED, the pol2-M644G rnh201Δ and the pol2-M644G rnh201-RED strains ± TOP1. For overall mutation rates, the pol2-M644G strain is statistically significantly different than the pol2-M644G rnh201Δ or pol2-M644G rnh201-red mutant (p < 0.0001). This is also true for the comparisons between TOP1 and top1Δ strains that are pol2-M644G rnh201Δ or pol2-M644G rnh201-RED (p < 0.0001). Specific mutation rates were calculated as the proportion of each type of event among the total mutants sequenced, multiplied by the overall mutation rate for each strain (using the data in Figure 3 and Tables S2 and S3). The sequencing data for the pol2-M644G strain is from [27]. The sequencing data for the pol2-M644G rnh201Δ strain is from [3] The sequencing data for the pol2-M644G rnh201Δ top1Δ strain is from [20]. BPS; base pair substitutions, 1 bp Indel; ±1 base insertions/deletions, Δ2–5bp; deletions of between 2 and 5 bases in size in perfect or imperfect repeats.

Genotype	Mutation rate (× 10−8)
	BPS	1 bp Indel	Δ2–5 bp	Overall
rnh201-RED	1.1	0.20	3.3	4.9
rnh201-RED top1Δ	2.6	0.55	0.036	3.5
pol2-M644G	5.2	0.74	0.16	13
pol2-M644G rnh201Δ	3.5	3.0	53	60
pol2-M644G rnh201Δ top1Δ	2.1	1.0	0.09	5.4
pol2-M644G rnh201-RED	4.2	3.7	46	53
pol2-M644G rnh201-RED top1Δ	4.2	2.1	0.09	8.7

Fig. 3.

Short deletion mutagenesis in a strain expressing the Rnh201-RED mutant is initiated by Top1 cleavage at unrepaired single genomic ribonucleotides. (A) The URA3-OR2 mutation spectra of the pol2-M644G rnh201-RED mutant +/− TOP1. The coding strand of the 804 base pair URA3 open reading frame (ORF) is shown with every tenth base indicated by a circle below the DNA sequence. Letters indicate single base pair substitutions, open or closed triangles indicate single base deletions or single base additions, respectively. The spectrum for the pol2-M644G rnh201-RED TOP1 strain is depicted above the URA3 sequence and the spectrum for the pol2-M644G rnh201-RED top1Δ strain is depicted below. (B) The URA3-OR2 mutation spectra of the rnh201-RED mutant ± TOP1. The spectrum for the rnh201-RED TOP1 strain is depicted above the URA3 ORF and the spectrum for the rnh201-RED top1Δ strain is depicted below.

4. Discussion

This study was motivated by an interest in the biological consequences of unrepaired ribonucleotides in DNA, and by the fact that the loss of RNase H2 activity causes the accumulation of both single unrepaired ribonucleotides and RNA-DNA hybrids in a cell. Using the RNase H2-RED mutant that specifically abolishes its ability to cleave at single ribonucleotides in DNA, the data support the hypothesis that replication stress and genome instability are initiated by Top1 cleavage at unrepaired single ribonucleotides that have been incorporated during leading strand synthesis by Pol ε. It is important to note that although the excision activity of the Rnh201-RED mutant on a single ribonucleotide in a duplex DNA substrate is completely abolished, this mutant also has reduced cleavage activity on RNA-DNA hybrid substrates, with in vitro measurements ranging from 3.1 % activity versus the wild type enzyme on a stretch of 6 ribonucleotides to up to 40 % activity on a poly rA-dT substrate [23]. This raises the possibility that some of the phenotypes observed in the rnh201-RED mutant may be related to less efficient processing of R-loops by RNase H2, although we have attempted to minimize this possibility by using strains that are proficient in RNase H1-dependent processing of RNA-DNA hybrids. Nonetheless, clearly both ribonucleotide-processing activities of RNase H2 are critical in a cell.

The fact that the Rnh201 protein was equivalently expressed in each of the strains used in these experiments (Fig. 1) indicates that the phenotypic differences in these strains are due to disruption of RNase H2 activity rather than altered expression of its catalytic subunit. The pol2-M644G rnh201-RED mutant cells accumulate in G₂/M phase, are sensitive to replication stress and have an activated genome integrity checkpoint, all of which are alleviated by deletion of TOP1 (Fig. 2B–D). In a pol2-M644G strain containing a high level of genomic ribonucleotides, the rate of 2–5 bp deletions in the rnh201-RED mutant is elevated to a level comparable to an rnh201Δ strain (Table 1 and Fig. 3), and these deletions are no longer seen in a strain that is defective in Top1.

In addition to short deletion mutations, larger forms of genome instability are associated with RNA-containing structures in the genome. These include DNA double-strand breaks (DSBs), gross chromosomal rearrangements (GCRs), gene conversion, gene duplication, chromosome rearrangement or loss, loss-of-heterozygosity (LOH) and copy number variation [4, 21, 22, 34–39]. The extent to which these events are due to failure to remove single ribonucleotides or RNA-DNA hybrids has been investigated in recent studies using independent approaches, and the results have shown that RNase H2-dependent resolution of both types of substrates is critical for preventing DNA breaks and recombination and for maintenance of genome stability. As is demonstrated in the accompanying manuscript from Cornelio et al. [40], LOH is elevated in a pol2-M644G rnh201-RED strain in a Top1-dependent manner, indicating that in the cellular context of a high density of genomic ribonucleotides, Top1-cleavage at unrepaired single ribonucleotides is an important contributor to chromosome instability.

In the future, it will be informative to determine whether the hallmarks of DNA damage in RNase H2-deficient mouse embryonic stem cells, including γH2AX foci and micronuclei formation [4], are also observed in cells expressing the analogous RER-defective allele of RNase H2. As both of the amino acids mutated in the RED variant of the catalytic subunit of RNase H2 are conserved in mice, the mouse mutant should be useful for studying the consequences of unrepaired genomic ribonucleotides in mammals. Can the p53-dependent embryonic lethality of an RNase H2 knockout mouse [4, 41] be attributed to failure of RNase H2-initiated removal of single ribonucleotides from DNA? Use of the RNase H2-RED mutant may provide an answer, and has the potential to provide mechanistic information regarding RNA-DNA damage phenotypes that arise due to failure of repair of a common spontaneous lesion in DNA.

Highlights.

• RNase H2 removal of single ribonucleotides is critical for genome maintenance.

• Biological consequences result from failure to remove single ribonucleotides.

• Top1 incision at single ribonucleotides causes mutagenesis and genome instability.