Removal of misincorporated ribonucleotides from prokaryotic genomes: an unexpected role for nucleotide excision repair

Abstract

Stringent steric exclusion mechanisms limit the misincorporation of ribonucleotides by high-fidelity DNA polymerases into genomic DNA. In contrast, low-fidelity Escherichia coli DNA polymerase V (pol V) has relatively poor sugar discrimination and frequently misincorporates ribonucleotides. Substitution of a steric gate tyrosine residue with alanine (umuC_Y11A) reduces sugar selectivity further and allows pol V to readily misincorporate ribonucleotides as easily as deoxynucleotides, whilst leaving its poor base-substitution fidelity essentially unchanged. However, the mutability of cells expressing the steric gate pol V mutant is very low due to efficient repair mechanisms that are triggered by the misincorporated rNMPs. Comparison of the mutation frequency between strains expressing wild-type and mutant pol V therefore allows us to identify pathways specifically directed at ribonucleotide excision repair (RER). We previously demonstrated that rNMPs incorporated by umuC_Y11A are efficiently removed from DNA in a repair pathway initiated by RNase HII. Using the same approach, we show here that mismatch repair and base excision repair play minimal back-up roles in RER in vivo. In contrast, in the absence of functional RNase HII, umuC_Y11A-dependent mutagenesis increases significantly in ΔuvrA, uvrB5 and ΔuvrC strains, suggesting that rNMPs misincorporated into DNA are actively repaired by nucleotide excision repair (NER) in vivo. Participation of NER in RER was confirmed by reconstituting ribonucleotide-dependent NER in vitro. We show that UvrABC nuclease-catalyzed incisions are readily made on DNA templates containing one, two, or five rNMPs and that the reactions are stimulated by the presence of mispaired bases. Similar to NER of DNA lesions, excision of rNMPs proceeds through dual incisions made at the 8(th) phosphodiester bond 5' and 4(th)-5(th) phosphodiester bonds 3' of the ribonucleotide. Ribonucleotides misinserted into DNA can therefore be added to the broad list of helix-distorting modifications that are substrates for NER.

Figure 1. Effect of inactivating RER, MMR, BER and NER on spontaneous mutagenesis in recA730 lexA(Def) ΔdinB strains ΔumuDC.

We constructed a series of isogenic recA730 lexA(Def) ΔdinB ΔumuDC rnhB ⁺/ΔrnhB strains in which ribonucleotide excision repair (RER; ΔrnhA), mismatch repair (MMR; ΔmutL, ΔmutH, ΔmutS, ΔuvrD), base excision repair (BER; Δung, Δxth, Δnfo), or nucleotide excision repair (NER; ΔuvrA, uvrB5, ΔuvrC, Δcho, ΔuvrD) were inactivated. The extent of spontaneous mutagenesis in strains harboring the control vector, pGB2, or plasmids expressing wild-type pol V, or umuC_Y11A was determined by assaying reversion of the hisG4 (ochre) allele (leading to histidine prototrophy). The average number of spontaneously arising His⁺ mutants per plate ± standard error of the mean (a) and the relative spontaneous mutagenesis in cells containing pGB2 vector or expressing the umuC_Y11A variant (b) are shown in the table below the graph. The data plotted on the graph represent the relative extent of spontaneous mutagenesis in cells expressing umuC_Y11A. The level of mutagenesis was calculated by subtracting the number of pre-existing mutants that grew on plates lacking histidine from the number of His⁺ mutants spontaneously arising on plates containing 1 µg/ml histidine. The relative amount of pol V-independent or umuC_Y11A-dependent mutagenesis was calculated as a percentage of the mutagenesis in cells expressing wild-type pol V. All experiments were performed in triplicate. Standard errors were calculated taking into account variability in spontaneous mutagenesis in each strain. The horizontal blue and red strips on the graph provide a reference baseline corresponding to the level of umuC_Y11A -dependent mutagenesis in repair-proficient rnhB ⁺ and ΔrnhB cells, respectively. Data for wild-type and ΔrnhA strains was taken from and is shown for comparison. As clearly observed, inactivation or rnhA and NER (ΔuvrA, uvrB5, ΔuvrC) in a ΔrnhB background leads to a dramatic increase in umuC_Y11A-dependent mutagenesis, indicating potential roles in RER.

Figure 2. Spectra of spontaneously arising rpoB mutations in recA730 lexA(Def) ΔumuDC ΔdinB strains expressing umuC_Y11A and proficient- or deficient- in MMR and RNase HII- mediated RER.

(A), Types of base-pair substitutions generated in the rpoB gene of mismatch repair defective recA730 lexA(Def) ΔdinB ΔumuDC mutL rnhB+/− strains. The arrows indicate mutagenic hot spots within rpoB. The spectra for the rnhB ⁺ strain (RW710) are taken from and are shown for direct comparison with the spectra for the isogenic ΔrnhB strain (RW942). The forward mutation rates in the MMR⁻ strains was assayed by measuring resistance to rifampicin and were calculated as 1.22±0.2×10⁻⁶ for RW710 and 1.82±0.2×10⁻⁶ for RW942. As expected, because of defects in MMR, the spectra are dominated by transition mutations. However, we have previously reported that low-fidelity pol V makes a significant number of transversion mutations compared to other E.coli DNA polymerases and we note a 4-fold increase in the number of transversion mutations in the ΔrnhB mutL strain expressing umuC_Y11A compared to the mutL rnhB ⁺ strain (see Table S1) consistent with pol V-dependent errors. The rnhB ⁺ mutL strain lacks these transversion mutations, as it undergoes active pol I-dependent RER and as a consequence, the spectrum generated in this strain actually reflects uncorrected pol I-dependent errors, rather than pol V-dependent errors. (B), Types of base-pair substitutions generated in the rpoB gene of mismatch repair proficient recA730 lexA(Def) ΔdinB ΔumuDC rnhB+/− strains. The forward mutation rates in the MMR⁺ strains assayed by measuring resistance to rifampicin and were calculated as 4.8±0.8×10⁻⁸ for RW698 and 2.5±0.4×10⁻⁷ for RW838. As expected, because the strains are proficient in MMR and transitions are repaired efficiently, the spectra are dominated by the poorly repaired transversion mutations. Approximately 300 Rif^R mutants were analyzed for each strain (Table S1). However, the spectrum of the rnhB ⁺ strain has been normalized to reflect the overall ∼6-fold lower mutation rate compared to the ΔrnhB strain. One can clearly observe that the spectra are very different in the two strains, with changes in both the types of mutations and locations of mutagenic hot-spots. Again, we believe the data support the model that in the MMR⁺ strains, pol I is responsible for low-levels of mutagenic events generated during RNase HII-pol I dependent RER, whilst in the absence of RER, umuC_Y11A mutations characterized by frequent transversion events persist and lead to a 6-fold increase in mutation rate.

Figure 3. Effect of deleting rnhA, rnhB and/or uvrA alone, or in various combinations, on the extent of umuC_Y11A-dependent spontaneous mutagenesis in recA730 lexA(Def) ΔumuDC ΔdinB strains.

We constructed a series of isogenic recA730 lexA(Def) ΔdinB ΔumuDC strains with ΔrnhA, ΔrnhB or ΔuvrA alleles alone, or in various combinations and assayed pol V-dependent spontaneous mutagenesis in strains harboring plasmids expressing wild-type pol V or umuC_Y11A by assaying reversion of the hisG4 (ochre) allele. The data plotted on the graph represent the relative extent of spontaneous mutagenesis in cells expressing umuC_Y11A calculated as a percentage of the spontaneous mutagenesis in cells expressing wild-type pol V. All experiments were performed in triplicate. Standard errors were calculated taking into account variability in spontaneous mutagenesis in each strain. Data for the wild-type, ΔrnhA, ΔrnhB and the ΔrnhA ΔrnhB strains were taken from and are shown for comparison. As clearly observed, in contrast to the ΔrnhA or ΔuvrA strains, which exhibit roughly the same extent of umuC_Y11A-dependent mutagenesis as the wild-type strain, the ΔrnhB strain exhibited significantly higher levels of spontaneous mutagenesis, suggesting that the rnhB-encoded RNase HII repair pathway is the primary defense against errant ribonucleotide incorporation in E.coli. However, in contrast to the rnhB ⁺ strains, deletion of either rnhA or uvrA in the ΔrnhB strain background leads to a further increase in umuC_Y11A-dependent spontaneous mutagenesis, suggesting that both enzymes participate in back-up pathways of ribonucleotide repair. Furthermore, umuC_Y11A-dependent spontaneous mutagenesis in the ΔrnhA ΔrnhB ΔuvrA triple mutant strain is actually higher than in the isogenic strain expressing wild-type pol V. These data imply that all major pathways specific for ribonucleotide repair are blocked in this strain background.

Figure 4. NER cleavage reaction products generated using various DNA-RNA-DNA hybrid substrates.

I; Cartoon of the synthetic substrates used in the in vitro assays, with the sites of incision and expected product size indicated, along with the DNA sequence containing rNMP(s) and mismatched nucleotides. II; The 50-mer duplexes (10 nM) in which the modified strand was 5′ end-labeled (indicated by *), were incubated with the NER proteins at concentrations of 40 nM (UvrA), 200 nM (UvrB), and 100 nM (UvrC) for 60 min at 55°C in the presence of 1 mM ATP. UvrABC-dependent incision on the fluorescein adducted 50-bp duplex (fT) was used as a control NER activity of the purified proteins (panel A). The DNA duplexes (panel B) as well as DNA-RNA-DNA hybrids either containing a single rAMP (panels C), two consecutive rAMPs (panel D), or five rNMPs (panels E) were assayed. The reaction products were separated under denaturing conditions by 15% polyacrylamide gel electrophoresis (PAGE). The efficiency of UvrABC-dependent incision was determined as a percentage of the radioactivity in the incised products relative to the total signal of the substrate (% inc.). The data below the gels are mean values calculated from at least two independent experiments. The DNA sequence containing rNMP(s) is shown alongside the gels where DNA and RNA are represented by uppercase and lowercase letters, respectively. Orange arrows indicate the cleavage sites. The red bracket indicates spontaneous ribonucleotide cleavage. UDS, refers to the undamaged strand, and DS, is damage/ribonucleotide-containing strand. The in vitro assays reveal that the NER proteins incise the DNA backbone 8 base pairs 5′ of ribonucleotides and that the reactions are stimulated by base mispairs.

Figure 5. RNase HII cleavage reaction products generated using various DNA-RNA-DNA hybrid substrates.

I; Cartoon of the synthetic substrates used in the in vitro assays, with the sites of incision and expected product size indicated, along with the DNA sequence containing rNMP(s) and mismatched nucleotides. II; The 50-mer duplexes (10 nM) in which the modified strand was 5′ end-labeled (indicated by *), were incubated with Rnase HII for 60 min at 37°C. The DNA duplexes (panel A) as well as DNA-RNA-DNA hybrids containing either single rAMP (panel B), two consecutive rAMPs (panel C), or five rNMPs (panel D) were assayed. The reaction products were separated under denaturing conditions by 15% polyacrylamide gel electrophoresis (PAGE). The efficiency of RNase HII-dependent incision was determined as a percentage of the radioactivity in the incised products relative to the total signal of the substrate (% inc.). The data below the gels are mean values calculated from at least two independent experiments. The DNA sequence containing rNMP(s) is shown alongside the gels where DNA and RNA are represented by uppercase and lowercase letters, respectively. Orange arrows indicate the cleavage sites. The in vitro assays confirm that E.coli RNase HII nicks the DNA backbone 5′ of ribonucleotides embedded in DNA and shows that the efficiency of the reaction is largely unaffected by base-mispairs.

Figure 6. NER cleavage reaction products generated using DNA templates with 3′ end-labeled rNMP-containing strands.

I; Cartoon of the synthetic substrates used in the in vitro assays, with the sites of incision and expected product size indicated, along with the DNA sequence containing rNMP(s) and mismatched nucleotides. II; The 50-mer duplexes (10 nM) in which the modified strand was 3′ end-labeled (indicated by *), were incubated with the NER proteins at concentrations of 40 nM (UvrA), 200 nM (UvrB), and 100 nM (UvrC) for 60 min at 55°C in the presence of 1 mM ATP. The DNA duplexes DNA-RNA-DNA hybrids either containing a single rAMP (panel A), two consecutive rAMPs (panel B) were assayed. The reaction products were separated under denaturing conditions by 15% polyacrylamide gel electrophoresis (PAGE). The efficiency of UvrABC-dependent incision was determined as a percentage of the radioactivity in the incised (% inc.) products relative to the total signal of the substrate. The data below the gels are mean values calculated from at least two independent experiments. The DNA sequence containing rNMP(s) is shown alongside the gels where DNA and RNA are represented by uppercase and lowercase letters, respectively. The full-sized DNA template (50-mer) and incision products of 19–20 bp are indicated. Orange arrows indicate the cleavage sites. The in vitro assays reveal that the NER proteins incise the DNA backbone 4–5 base pairs 3′ of ribonucleotides and that the efficiency of the reaction is largely unaffected by base-mispairs.

Figure 7. Various DNA repair pathways compete, cooperate, or substitute for each other in order to sanitize the E.coli chromosome from mispairs, uracils and incorporated rNMPs.

The cartoon helps to explain why spontaneous mutagenesis induced by wild-type pol V (A) differs from Y11A_UmuC-dependent mutagenesis (B) and illustrates the respective roles of MMR (red box), RER^A (stands for RNase HI-initiated ribonucleotide excision repair and is indicated in yellow), RER^B (RNase HII-initiated ribonucleotide excision repair, indicated in blue), NER (green), and BER (grey) (The competing pathways are indicated by boxes with gradient colors). Misincorporated nucleotides are shown in red (where ts are transitions, tv are transversions), correctly paired ribonucleotides are indicated in blue. For simplicity, all the transversions are shown refractory to MMR and NER while in reality they could be repaired by both pathways although less efficiently than transitions. Both wild-type and mutant pol V make frequent base-substitution errors. The transition mutations are rapidly repaired by MMR, while any errant ribonucleotides (correctly-paired or mispaired) are also efficiently removed by RER^B. Ung-dependent BER only operates on dU, not rU, incorporated into the DNA and therefore has no role in RER. In contrast, NER is able to remove rNMPs misincorporated by either wild-type or mutant pol V. Since umuC_Y11A is able to incorporate multiple consecutive rNMPs into DNA, RER involving RNase HI is limited to strains expressing the pol V variant. RER^B is normally required for highly efficient removal of errant ribonucleotides, however, in its absence the role of NER and RNase HI becomes apparent. The fact that the level of spontaneous mutagenesis in strains expressing umuC_Y11A with an “unlocked” sugar steric gate is 90% lower than mutagenesis in strains harboring wild-type pol V, implies that numerous errant ribonucleotides are very efficiently excised by the collaborative actions of rNMP-specific repair pathways which concomitantly remove mispaired dNMPs positioned within the repair patch (such as for example two D^tvs in the panel B). In the absence of RNase HI, RNase HII and NER proteins, the majority of the misincorporated rNMPs remains embedded in the chromosomal DNA. As a result, spontaneous mutagenesis in the ΔrnhA ΔrnhB ΔuvrA strain expressing umuC_Y11A is higher than in the isogenic strain expressing wild-type pol V.