Role of RNase H1 in DNA repair: removal of single ribonucleotide misincorporated into DNA in collaboration with RNase H2

Several RNases H1 cleave the RNA-DNA junction of Okazaki fragment-like RNA-DNA/DNA substrate. This activity, termed 3’-junction ribonuclease (3’-JRNase) activity, is different from the 5’-JRNase activity of RNase H2 that cleaves the 5’-side of the ribonucleotide of the RNA-DNA junction and is required to initiate the ribonucleotide excision repair pathway. To examine whether RNase H1 exhibits 3’-JRNase activity for dsDNA containing a single ribonucleotide and can remove this ribonucleotide in collaboration with RNase H2, cleavage of a DNA8-RNA1-DNA9/DNA18 substrate with E. coli RNase H1 and H2 was analyzed. This substrate was cleaved by E. coli RNase H1 at the (5’)RNA-DNA(3’) junction, regardless of whether it was cleaved by E. coli RNase H2 at the (5’)DNA-RNA(3’) junction in advance or not. Likewise, this substrate was cleaved by E. coli RNase H2 at the (5’)DNA-RNA(3’) junction, regardless of whether it was cleaved by E. coli RNase H1 at the (5’)RNA-DNA(3’) junction in advance or not. When this substrate was cleaved by a mixture of E. coli RNases H1 and H2, the ribonucleotide was removed from the substrate. We propose that RNase H1 is involved in the excision of single ribonucleotides misincorporated into DNA in collaboration with RNase H2.

RNase H2 is required to initiate the ribonucleotide excision repair (RER) pathway both in eukaryotes 11,12,18,19 and prokaryotes 20 because of its junction ribonucelase (JRNase) activity. This activity cleaves dsDNA containing a single rNMP (dsDNA R1 ) at the 5'-side of the ribonucleotide [21][22][23][24] . Subsequent removal of the ribonucleotide is catalyzed by several enzymes that all together form the RER pathway 11,18,20 . RNase H1 does not exhibit JRNase activity 21,[25][26][27][28] . However, RNases H1 from Halobacterium sp. NRC-1 (Halo-RNase H1) 29 , Sulfolobus tokodaii (Sto-RNase H1) 30 , Thermotoga maritima 26 , and E. coli 26,31 exhibit a weak activity that catalyzes the cleavage of an Okazaki fragment-like substrate at the RNA-DNA junction. To distinguish this activity from JRNase activity of RNase H2, that catalyzes the cleavage of an RNA-DNA/DNA substrate at the 5'-side of the ribonucleotide of the RNA-DNA junction 31 , this activity and JRNase activity of RNase H2 will be designated as 3'-and 5'-JRNase activities respectively hereafter ( Fig. 1a,b). The RNA-DNA junction of an Okazaki fragment-like substrate containing a single ribonucleotide is not cleaved by Halo-RNase H1 29 and Sto-RNase H1 30 , suggesting that an upstream duplex structure is necessary for recognition of this junction by RNase H1. However, it remains to be determined whether the RNA-DNA junction of this substrate is cleaved by RNase H1 by supplying a short RNA fragment that facilitates the formation of an upstream duplex structure. It also remains to be determined whether RNases H1 cleave dsDNA R1 at the (5')RNA-DNA(3') junction. We used E. coli RNase H1 as a representative member of type 1 RNases H in this study to analyze the 3'-JRNase activity of type 1 RNase H, because the structure and function of E. coli RNase H1 have been extensively studied 32 .
In this report, we showed that the RNA-DNA junction of the RNA1-DNA9/DNA18 substrate is not cleaved by E. coli RNase H1 but is cleaved by the enzyme when a short RNA fragment is supplied to facilitate the formation of an upstream duplex structure. We also showed that E. coli RNase H1 exhibits 3'-JRNase activity for dsDNA R1 in the presence of manganese ions much more effectively than in the presence of magnesium ions, regardless of whether this substrate is cleaved by 5'-JRNase activity of E. coli RNase H2 in advance or not. From these results, we propose that single rNMPs misincorporated into the genomes can be excised by a cooperative work of RNases H1 and H2 in prokaryotic cells.
Cleavage of R9-D9/D18 substrate by E. coli RNase H1. It has been reported that E. coli RNase H1 cleaves an Okazaki fragment-like substrate most effectively at R(−2)-R(−1) and less effectively at the  , indicating that E. coli RNase H1 exhibits a weak 3'-JRNase activity for this substrate in the presence of manganese ions. However, two conflicting results that E. coli RNase H1 can 26,31 and cannot 29,30 cleave this substrate at the RNA-DNA junction in the presence of magnesium ions have been reported. Therefore, the R9-D9*/D18 substrate was first cleaved by E. coli RNase H1 either in the presence of 0.1 mM MnCl 2 or 10 mM MgCl 2 , to examine whether E. coli RNase H1 exhibits 3'-JRNase activity for this substrate in the presence of magnesium ions. The results are shown in Figs. 3a, b. The D9* fragment is detected as one of the major products when the substrate is extensively cleaved by the enzyme. This band is produced more effectively in the presence of manganese ions than in the presence of magnesium ions. These results indicate that E. coli RNase H1 cleaves the R9-D9*/D18 substrate at the RNA-DNA junction either in the presence of manganese or magnesium ions, but more effectively in the presence of manganese ions than in the presence of magnesium ions. This result is consistent with that previously reported 31 . Thus, E. coli RNase H1 exhibits 3'-JRNase activity for an Okazaki fragment-like substrate either in the presence of manganese or magnesium ions, but more strongly in the presence of manganese ions. This metal ion preference is opposite to that for the RNase H activity of this enzyme determined using an RNA/DNA substrate 33 .
E. coli RNase H1 cleaves the R9-D9*/D18 substrate almost exclusively at R(−2)-R(−1) and RNA-DNA junction in the presence of 0.1 mM MnCl 2 , whereas it cleaves this substrate preferentially at R(−3)-R(−2) and R(−2)-R(−1) in the presence of 10 mM MgCl 2 (Fig. 3a,b). It has been reported that hydrolysis of this substrate by E. coli . These results suggest that the R9-D9*/D18 substrate is cleaved by E. coli RNase H1 preferentially at the upstream region of the RNA-DNA junction in the presence of magnesium ions due to its RNase H activity. This substrate is not cleaved by E. coli RNase H1 at the upstream region of the RNA-DNA junction in the presence of 0.1 mM MnCl 2 , except for R(−2)-R(−1), probably due to a very weak RNase H activity. It has been reported that the RNase H activity of E. coli RNase H1 determined using an RNA/DNA substrate in the presence of the optimum concentration of manganese ions (2-4 μM) is lower than that determined in the presence of the optimum concentration of magnesium ions (5-10 mM) by only 5 fold 33 , whereas the RNase H activity of E. coli RNase H1 determined using a 12 base pair RNA/DNA substrate in the presence of 5 mM MnCl 2 is lower than that determined in the presence of 5 mM MgCl 2 by 1000 fold 31 . As a result, E. coli RNase H1 cleaves the R9-D9*/D18 substrate more effectively in the presence of 10 mM MgCl 2 than in the presence of 0.1 mM MnCl 2 (Fig. 4a).
Cleavage of R1-D9/D18 and R8:R1-D9/D18 substrates by E. coli RNase H1. The R1-D9* fragment is detected as one of the major products when the R9-D9*/D18 substrate is extensively cleaved by E. coli RNase H1 (Fig. 4a), suggesting that the R1-D9*/D18 substrate is not cleaved by E. coli RNase H1. To examine whether E. coli RNase H1 does not cleave this substrate, but cleaves it at the RNA-DNA junction when 8 b RNA (R8) complementary to the single stranded region of the R1-D9*/D18 substrate is supplied, the R1-D9*/D18 and R8:R1-D9*/D18 substrates were cleaved by E. coli RNase H1 either in the presence of 0.1 mM MnCl 2 or 10 mM MgCl 2 . The results are shown in Figs. 3c, d. E. coli RNase H1 does not cleave the R1-D9*/D18 substrate regardless of the metal cofactors. In contrast, it cleaves the R8:R1-D9*/D18 substrate at the RNA-DNA junction, but only in the presence of manganese ions. This result suggests that E. coli RNase H1 cleaves an Okazaki fragment-like RNA-DNA/DNA substrate at the RNA-DNA junction regardless of whether this substrate contains a nick at R(−2)-R(−1). The RNA-DNA junction of the R8:R1-D9*/D18 substrate is not fully cleaved by E. coli RNase H1, probably because the upstream region of the RNA-DNA junction is cleaved before the RNA-DNA junction is completely cleaved. The RNA-DNA junction of the R9-D9*/D18 substrate is cleaved by the 3'-JRNase activity of E. coli RNase H1 less effectively in the presence of magnesium ions than in the presence of manganese ions, probably because the upstream region of the RNA-DNA junction is cleaved by the RNase H activity of E. coli RNase H1 more effectively in the presence of magnesium ions than in the presence of manganese ions. The RNA-DNA junction of the R8:R1-D9*/D18 substrate is not cleaved by E. coli RNase H1 in the presence of magnesium ions, probably because the presence of a nick at R(−2)-R(−1) alters the interaction between the substrate and metal ion, in such a way that the scissile phosphate group of the substrate and magnesium ions are not arranged ideally.
coli RNase H1 to cleave the R1-D9*/D18 substrate at the RNA-DNA junction indicates that multiple upstream ribonucleotides are necessary for the cleavage of an Okazaki fragment-like substrate by the enzyme at the RNA-DNA junction. To examine whether two upstream ribonucleotides are sufficient for this cleavage, the R2-D9*/D18 substrate was cleaved by E. coli RNase H1 either in the presence of 0.1 mM MnCl 2 or 10 mM MgCl 2 . The results are shown in Figs. 3e, f. E. coli RNase H1 cleaves the R2-D9*/D18 substrate at the RNA-DNA junction either in the presence of manganese or magnesium ions, but more effectively in the presence of manganese ions. This result indicates that the presence of two upstream ribonucleotides is sufficient for the cleavage of an Okazaki fragment-like substrate by E. coli RNase H1 at the RNA-DNA junction. To examine whether this cleavage site is shifted by supplying an RNA strand that facilitates the formation of an upstream duplex structure, the R7:R2-D9*/D18 substrate was also cleaved by E. coli RNase H1. The results are shown in Fig. 3e, f. E. coli RNase H1 most effectively cleaves this substrate at R(−2)-R(−1) in the presence of manganese ions. This site is also cleaved in the presence Scientific RepoRts | 5:09969 | DOi: 10.1038/srep09969 of magnesium ions, but much less effectively, probably because the RNA/DNA region of the R7:R2-D9*/ D18 substrate is effectively cleaved by the RNase H activity of the enzyme in the presence of magnesium ions. Thus, the primary products of the R7:R2-D9*/D18 substrate upon cleavage with E. coli RNase H1 in the presence of manganese and magnesium ions are probably the R7:R1-D9*/D18 and R2-D9*/D18 duplexes respectively. The RNA-DNA junctions of these primary products are cleaved only when the concentration of the enzyme is greatly elevated.
Cleavage of D8-R1-D9/D18 substrate by E. coli RNase H1. To examine whether E. coli RNase H1 exhibits 3'-JRNase activity for dsDNA R1 , the D8-R1-D9*/D18 and *D8-R1-D9/D18 substrates were cleaved by E. coli RNase H1 either in the presence of 0.1 mM MnCl 2 or 10 mM MgCl 2 . The results are shown in Fig. 5. The D9* and *D8-R1 fragments are detected as the major products more clearly in the presence of manganese ions, indicating that E. coli RNase H1 cleaves these substrates mainly at the (5')RNA-DNA(3') junction and much more effectively in the presence of manganese ions than in the presence of magnesium ions to produce the D8-R1-D9/D18 duplex containing a nick at the (5') RNA-DNA(3') junction. The D9* fragment produced from D8-R1-D9* upon cleavage with E. coli RNase H1 did not migrate in the urea gel equally with the synthetic D9* fragment having the 5'-OH terminus but migrated equally with the synthetic D9* fragment phosphorylated at the 5'-terminus. The synthetic D9* fragment phosphorylated at the 5'-terminus migrated in the urea gel faster and slower than the synthetic D9* and D8* fragments with the 5'-OH terminus respectively (data not shown). This indicates that the 3'-JRNase activity of E. coli RNase H1 hydrolyzes the phosphodiester bond (PO-3') of the (5'  has been shown to exhibit 5'-JRNase activity for an Okazaki fragment-like substrate 31 . A crude extract from E. coli cells exhibits 5'-JRNase activity for dsDNA R1 , whereas that from RNase H2-deficient E. coli cells does not exhibit this activity 22 , suggesting that E. coli RNase H2 also exhibits 5'-JRNase activity for dsDNA R1 . However, it remains to be determined whether the purified protein of E. coli RNase H2 exhibits this activity for dsDNA R1 . Therefore, the *D8-R1-D9/D18 and D8-R1-D9*/D18 substrates were cleaved by E. coli RNase H2 either in the presence of 10 mM MnCl 2 or 10 mM MgCl 2 , to examine whether this enzyme exhibits 5'-JRNase activity for dsDNA R1 . The R9-D9*/D18 substrate was also cleaved by this enzyme for comparative purpose either in the presence of 10 mM MnCl 2 or 10 mM MgCl 2 . As shown in Fig. 3a, b, E. coli RNase H2 cleaved the R9-D9*/D18 substrate almost exclusively at R(−2)-R(−1) either in the presence of manganese or magnesium ions, but more effectively in the presence of magnesium ions. Likewise, as shown in Fig. 5, E. coli RNase H2 cleaved the *D8-R1-D9/D18 and D8-R1-D9*/D18 substrates almost exclusively at the (5')DNA-RNA(3') junction to produce the D8-R1-D9/D18 duplex containing a nick at the (5')DNA-RNA(3') junction either in the presence of manganese or magnesium ions, but more effectively in the presence of magnesium ions. Thus, E. coli RNase H2 exhibits 5'-JRNase activity for dsDNA R1 either in the presence of manganese or magnesium ions, but more strongly in the presence of magnesium ions. This metal ion preference is opposite to that for the RNase H activity of this enzyme determined using an RNA/DNA substrate 31,35 . The RNase H activity of E. coli RNase H2 determined in the presence of 5 mM MgCl 2 is lower than that determined in the presence of 5 mM MnCl 2 by 10 fold. The RNA/DNA hybrid region of the R9-D9*/D18 substrate is not cleaved by E. coli RNase H2, probably because the RNase H activity of this enzyme is very low. It has been reported that the RNase H activity of E. coli RNase H2 determined in the presence of 5 mM MnCl 2 using an oligomeric RNA/DNA substrate is lower than that of E. coli RNase H1 determined in the presence of 5 mM MgCl 2 and 5 mM MnCl 2 by 2 × 10 4 and 20 fold respectively 31  Cleavage of D8-R1-D9/D18 substrate by mixture of E. coli RNases H1 and H2. To examine whether single ribonucleotides can be removed from dsDNA R1 by a cooperative work of E. coli RNases H1 and H2, the D8-R1-D9/D18 substrate was cleaved by a mixture of these enzymes. The results are shown in Fig. 7. When the *D8-R1-D9/D18 substrate was cleaved by the mixture of E. coli RNase H1 (20 ng μL −1 ) and E. coli RNase H2 (10 ng μL −1 ) in the presence of 1 mM MgCl 2 and 0.1 mM MnCl 2 , the *D8 fragment was detected as the major product. Likewise, when the D8-R1-D9*/D18 substrate was cleaved by the mixture of these enzymes in the same condition, the D9* fragment was detected as the major product. These results suggest that single ribonucleotides embedded in dsDNA can be removed by a cooperative work of E. coli RNases H1 and H2.
Possible involvement of E. coli RNase H1 in DNA repair. DNA (deoxyribonucleic acid) is a universal hereditary material that encodes genetic information essential for the existence of all cellular life and some viruses. It is characterized by its ability to store and transfer information and to self-replicate, which is catalyzed by enzymatic machinery. To keep the integrity of the transferred information, DNA should be kept unmodified. However, DNA is frequently subject to various modifications that can render it unstable if left unrepaired. Of these modifications, the presence of single ribonucleotide monophosphates (rNMPs) misincorporated into the DNA backbone shows both negative and positive consequences on the genome 36 . While the positive roles shown up to now are limited to nascent strand discrimination in mismatch repair [37][38][39] and the use of ribonucleotides by polμ in non-homologous end-joining (NHEJ) pathways [40][41][42] , more studies emphasized the negative roles such as the increase in mutation rate [43][44][45] , chromosomal abnormality 11,12 , mammalian embryonic lethality 11,12 , replication fork barrier 18,43,46 , and autoimmune diseases 13,47 . Recent studies indicate that RNase H2 saves the genome by initiating the pathway which removes those intruders and restores the DNA back to its original form with the assistance of several other enzymes 11,12,[18][19][20]38,39 .
In this study, we showed for the first time that E. coli RNase H1 exhibits 3'-JRNase activity for dsD-NA R1 much more effectively in the presence of manganese ions than in the presence of magnesium ions, regardless of whether this substrate is cleaved by 5'-JRNase activity of E. coli RNase H2 in advance or not, and can excise the single ribonucleotide in collaboration with E. coli RNase H2. These results suggest that not only RNase H2 but also RNase H1 is involved in the RER pathway. Two possible RER pathways, in which both enzymes are included, are schematically shown in Fig. 8. According to these pathways, removal of the single ribonucleotides misincorporated into DNA is initiated by the 5'-JRNase activity of RNase H2 and completed by the 3'-JRNase activity of RNase H1, or vice versa. However, human RNase H1 did not exhibit 3'-JRNase activity for dsDNA R1 either in the presence of magnesium or manganese ions (E. Tannous, unpublished result), suggesting that RNase H1 is involved in the RER pathway only in the prokaryotic cells. This result and the result that Halo-RNase H1 exhibits 3'-JRNase activity for dsDNA R1 in the presence of Mn 2+ ions (E. Tannous, unpublished result) may exclude the possibility that the 3'-JRNase activity of E. coli RNase H1 is derived from another E. coli enzyme contaminated, which is not detected as a band on SDS-PAGE (Fig. 4), because human RNase H1 is a basic protein similar to E. coli RNase H1 and is purified using cation-exchange column chromatography (E. Tannous, unpublished result), whereas Halo-RNase H1 is an acidic protein and is purified by anion-exchange column chromatography 48 . Because E. coli RNase H1 is purified using cation-exchange column chromatography 49 , it is unlikely that another E. coli enzyme with 3'-JRNase activity is co-purified with E. coli RNase H1 and Halo-RNase H1 but not co-purified with human RNase H1. Disruption of the rnhA gene has been reported to increase a basal level of SOS expression in E. coli, probably due to persistence of R-loops on the chromosome 50 . However, the 3'-JRNase activity of E. coli RNase H1 may not be involved in SOS response, because this activity may not be required for R-loop resolution.
The concentration of magnesium ions (10 mM) optimum for in vitro 5'-JRNase activity of E. coli RNase H2 is comparable to that in E. coli cells, whereas the concentration of manganese ions (0.1 mM) optimum for in vitro 3'-JRNase activity of E. coli RNase H1 is higher than that in E. coli cells. However, the fact that the presence of magnesium ions is not inhibitory for the 3'-JRNase activity of E. coli RNase H1 may suggest that E. coli RNase H1 exhibits this activity inside the cells. The in vivo studies showing that E. coli RNase H1 helps in sanitizing the genome from errant rNMPs misincorporated in dsDNA when the rnhB gene encoding E. coli RNase H2 is inactivated and thus limiting the consequences of the excessive accumulation of ribonucleotides in the E. coli genome 20 supports this hypothesis. Thus, the significance of this work is that it highlights a new possible role for bacterial type 1 RNases H in DNA repair, which should be further investigated in vivo, as well as for RNases H from different organisms.
Enzymatic activity. The R9-D9*/D18, R2-D9*/D18, R1-D9*/D18, *D8-R1-D9/D18, D8-R1-D9*/ D18, and *D18/D18 duplexes were prepared by hybridizing R9-D9*, R2-D9*, R1-D9*, *D8-R1-D9, D8-R1-D9*, and *D18 with a 2 molar equivalent of D18 respectively. The R7:R2-D9*/D18 and R8:R1-D9*/ D18 duplexes were prepared by hybridizing R2-D9* and R1-D9* with a 2 molar equivalent of D18 in the presence of a 2 molar equivalent of R7 and R8 respectively. These duplexes were used as substrates. Hydrolysis of the substrate at 30 °C for 15 min and separation of the products on a 20% polyacrylamide gel containing 7 M urea were carried out as described previously 54 . The reaction buffer contained 10 mM Tris-HCl (pH 8.5), 1 mM 2-mercaptoethanol, 0.01% BSA, 10 mM NaCl, and MnCl 2 or MgCl 2 at the Figure 8. A new RER pathway directed by RNases H1 and H2. dsDNA containing a single ribonucleotide (dsDNA R1 ) is schematically shown. DNA strand and a single ribonucleotide are shown by blue and red boxes respectively. Single ribonucleotide and flanking deoxyribonucleotides are labeled R and D respectively. Two possible pathways, in which a single ribonucleotide embedded in dsDNA is removed by RNases H1 and H2 in a stepwise manner, are shown. In pathway A, dsDNA R1 is first cleaved by RNase H2 at the 5' side of the ribonucleotide to produce dsDNA R1 with a nick at the 5' side of the ribonucleotide. Then dsDNA R1 with this nick is subsequently cleaved by RNase H1 at the 3' side of ribonucleotide of the (5')RNA-DNA(3') junction to release a single ribonucleotide. In pathway B, dsDNA R1 is first cleaved by RNase H1 at the 3' side of the ribonucleotide to produce dsDNA R1 with a nick at the 3' side of the ribonucleotide. Then dsDNA R1 with this nick is subsequently cleaved by RNase H2 at the 5' side of the ribonucleotide to release a single ribonucleotide. The gaps in dsDNAs produced by removal of a single ribonucleotide from dsDNA R1 in pathway A or B are probably filled by DNA polymerase and ligase.