Processing of matched and mismatched rNMPs in DNA by archaeal ribonucleotide excision repair

Abstract

Ribonucleoside monophosphates (rNMPs) are the main non-canonical nucleotides in genomic DNA, and their incorporation can occur as mismatches or matches in vivo. To counteract the mutagenic potential of rNMPs in DNA, all organisms evolved ribonucleotide excision repair (RER), a mechanism initiated by type 2 RNase H. Here, we describe the in vitro reconstitution of matched and mismatched rNMP repair using archaeal RER enzymes. Our data suggest two types of RER pathways, including the classical flap RER and a backup RER with the order of reactions changed for Fen1 and Pols. The genomic rNMP level in RER-deficient or PolB-deficient archaeal cells along with in vitro reconstitution of RER suggests an in vivo role of PolD in RER. Our results provide insights into how matched and mismatched rNMPs may be processed by RER.

Keywords: Bacteriology; Biochemistry; Molecular biology.

Graphical abstract

Figure 1

Enzymatic characterization of RER enzymes on matched or mismatched substrates Lanes L contain oligonucleotide ladders. The structure of matched or mismatched dsDNA substrates are shown at the top of each gel, and oligonucleotide sequences are in Table S3. Bar graphs are the means ± standard deviation of cleavage products of at least three independent experiments. p values were determined by unpaired t test and are reported: ∗∗∗p value <0.0001, ∗p value <0.05 and non-significant (ns). The p values represented above the bars when PabPCNA or PabRNaseHII are added to the reactions correspond to the difference observed with the condition of the same enzyme concentration without PabPCNA or PabRNase HII. (A) Ribonuclease activity of PabRNase HII. The indicated amounts of PabRNase HII were incubated with matched L34_rA/L34_dT (lanes 1–9), mismatched L34_rA/L34_dC (lanes 10–18), or L34_rG/L34_dA_reverse (lanes 19–27). (B) Strand-displacement activity of PabPolB and PabPolD. The indicated amounts of PabPolB or PabPolD were incubated with matched L30_rA:L57RC/L87 (lanes 29–34 and 45–50, respectively), mismatched L30_rA:L57RC/L87_dC (lanes 37–42, 53–58, respectively) or with PabPCNA (lanes 35, 51, 43, 59). (C) 5′-3′ exonuclease and 5′-flap endonuclease activities of PabFen1. The indicated amounts of PabFen1 were incubated with matched L30_rA:L57RC/L87 (lanes 60–67) or mismatched L30_rA:L57RC/L87_dC (lanes 68–75) without PabPCNA or with 0.3 μM PabPCNA (lanes 66, 74) or 0.5 μM PabRNase HII (lanes 67, 75). (D) Ligation activity of PabLig1. The indicated amounts of PabLig1 were incubated with matched L30_rA:L57RC/L87 in the absence or presence of 0.3 μM PabPCNA (lanes 76–77, respectively) or with mismatched rA/dC L30_rA:L57RC/L87_dC in the absence or presence of 0.3 μM PabPCNA (lanes 80–81, respectively) and 1 mM ATP.

Figure 2

PabPolB and PabPolD can act in RNase HII-initiated RER pathway (A–F) (A) Reactions were performed in presence of a dual-labeled matched L87_rA/L87_dT (Table S3) (A–C) or mismatched L87_rA/L87_dC (Table S3) (D–F) rNMP-containing dsDNA template. Reactions were performed in the presence of various combinations of PabPolB, PabPolD, PabFen1, PabPCNA, PabRNase HII and with (B and E) or without (C and F) PabLig1 in buffer containing 1 mM ATP and physiological level of dNTPs as described in STAR methods. Lanes L contain 5′-FAM-labeled oligonucleotide ladders (57- and 87-nt) in FAM panel and 3′-Cy5-labeled oligonucleotide ladders (30- and 87-nt) in Cy5 panel. (G) Bar graphs representing the means ± standard deviation of full-length ligation products (87-nt) obtained after RER reconstitution with PabPolB (lanes 6, 24) or PabPolD (lanes 7, 25) or both Pols (lanes 8, 26) of at least three independent experiments. p values were determined by unpaired t test and are reported: ∗p value <0.05 and non-significant (ns). The p values represented above the bars for rA/dC conditions correspond to the difference observed with the same enzyme on the rA/dT substrate.

Figure 3

Functional conservation of RER enzymes in Archaea (A) Relative quantification of polB and rnhB mRNA expression by RT-qPCR in RER mutants (Table S1) compared to TbaWT in at least three independent experiments. (B) Growth curves of RER mutants in TRM medium. The growth curves are the means ± standard deviation of three independent experiments. (C) Alkali sensitivity of genomic DNA from RER mutants. Separation by agarose gel of genomic DNA after alkaline (NaOH +) or control treatment with NaCl (NaOH -) under neutral (Formamide -) or denaturing (Formamide +) conditions. (D) Densitometry traces of NaOH conditions from (C) are shown, and sizes of the DNA markers are indicated on the left.

Figure 4

Proposed RER model for matched or mismatched rNMP removal in Archaea The archaeal RER pathway is efficient to repair matched (rA/dT) or mismatched (rA/dC) rNMP-containing dsDNA. STEP 1 involves the incision at the 5′-side of the matched or mismatched embedded rNMP by RNase HII. In the canonical RER pathway, STEP 2 implies strand-displacement synthesis mainly by PolB creating the 5′-terminal rNMP-containing flap subsequently cut by Fen1 endonuclease activity in STEP 3. PolD might replace PolB in STEP 2 for strand-displacement synthesis of mismatched 5′-rNMP-containing flap. In case of defective strand-displacement activity by Pol, Fen1 5′-3′ exonuclease or 5′-flap endonuclease activity could release the 5′-terminal rNMP after RNase HII incision (STEP 2B) but before PolB or PolD extension in STEP 3B. STEP 4 involves the sealing by Lig1 of DNA ends. Schematic illustrations are created with Biorender (https://biorender.com).