A novel endonuclease that may be responsible for damaged DNA base repair in Pyrococcus furiosus

Abstract

DNA is constantly damaged by endogenous and environmental influences. Deaminated adenine (hypoxanthine) tends to pair with cytosine and leads to the A:T→G:C transition mutation during DNA replication. Endonuclease V (EndoV) hydrolyzes the second phosphodiester bond 3' from deoxyinosine in the DNA strand, and was considered to be responsible for hypoxanthine excision repair. However, the downstream pathway after EndoV cleavage remained unclear. The activity to cleave the phosphodiester bond 5' from deoxyinosine was detected in a Pyrococcus furiosus cell extract. The protein encoded by PF1551, obtained from the mass spectrometry analysis of the purified fraction, exhibited the corresponding cleavage activity. A putative homolog from Thermococcus kodakarensis (TK0887) showed the same activity. Further biochemical analyses revealed that the purified PF1551 and TK0887 proteins recognize uracil, xanthine and the AP site, in addition to hypoxanthine. We named this endonuclease Endonuclease Q (EndoQ), as it may be involved in damaged base repair in the Thermococcals of Archaea.

Figure 1.

Cleavage activity of P. furiosus cell extracts toward dI-containing DNA. 5′-³²P-labeled 49-I25 dsDNA (5 nM) was incubated at 60°C for 1 h with no enzyme (lane 2), 10-nM PfuEndoV (lane 3) and aliquots of SP column fractions from P. furiosus cell extracts (lanes 4–7), in the reaction solution described in the Materials and Methods section. The product markers (mixture of 24, 25 and 26 nt) are shown in lanes 1 and 8.

Figure 2.

Purification scheme for the enzyme exhibiting cleavage activity on dI-containing DNA. Pyrococcus furiosus cell extracts (from 4 l of culture) were fractionated on five kinds of columns: P11, SP, Heparin, Hydroxyapatite and MonoS. An aliquot of each fraction was assayed for its activity on 5′-Cy5-labeled 49-I25 dsDNA. The fractions indicated by arrows showed the cleavage activity, and were subjected to downstream purification. The cleavage products were separated by 8-M urea-12% PAGE.

Figure 3.

SDS-10%PAGE analysis of fraction 48 from the MonoS-column chromatography, followed by silver staining. The sizes of markers (lane 1) are shown on the left side of the panel. The proteins in fraction 48 (lane 2) were cut into eight slices, as shown by the dots on the gel, and were subjected to the MS analysis.

Figure 4.

Cleavage properties of PfuEndoQ. (A) The nucleotide sequence of the substrate containing dI. The PfuEndoQ cleavage site is indicated by an arrow. (B) 5′-Cy5-labeled substrates (10 nM) were incubated at 75°C for 15 min, as described in the Materials and Methods section. *I, 5′-Cy5-labeled 45-I25 ssDNA; *I/T 5′-Cy5-labeled 45-I25 annealed with 45R; I/*T, 5′-Cy5-labeled 45R annealed with 45-I25. M, 5′-Cy5-labeled ssDNA (23, 24, 25 and 26 nt) (lane 1); −, no enzyme (lanes 2, 4 and 6); +, 20-nM PfuEndoQ (lanes 3, 5 and 7). (C) 3′-FITC-labeled substrates (10 nM) were incubated. I*, 3′-FITC-labeled 45-I25 ssDNA; I*/T 3′-FITC-labeled 45-I25 annealed with 45R. M, 3′-FITC-labeled and 5′-phosphorylated 21-nt-ssDNA as a marker (lane 1); −, no enzyme (lanes 2 and 4); +, 20-nM PfuEndoQ (lanes 3 and 5). Cleavage products were separated by 8-M urea-12% PAGE.

Figure 5.

Electrophoresis mobility shift assay of PfuEndoQ. The dsDNA (5 nM) was incubated with various concentrations of PfuEndoQ (0, 2.5, 5, 12.5, 25 and 50 nM) in the presence (+) (lanes 1–12) or absence (−) (lanes 13–18) of 1-mM MgCl₂. *A/T, 5′-Cy5-labeled dsDNA without dI (lanes 1–6); *I/T, 5′-Cy5-labeled dsDNA containing dI (lanes 7–18). The free DNA and the protein–DNA complex were separated by 4% native-PAGE.

Figure 6.

Substrate specificity of PfuEndoQ. (A) 5′-Cy5-labeled ssDNA (ss) or dsDNA (ds) substrates containing dI (lanes 2–5), dU (lanes 7–10), AP (lanes 12–15) and a mismatch (lanes 17 and 18) were subjected to reactions with PfuEndoQ. The normal DNAs (lanes 20–23) were used as the controls. The mixture of 5′-Cy5-labeled ssDNA (23, 24, 25 and 26 nt) was loaded on lanes 1, 6, 11, 16 and 19. The marker sizes are shown on the left of the panel. (B) 5′-³²P-labeled ssDNA (ss) or dsDNA (ds) substrates containing dX (lanes 2–5) were subjected to the reactions with PfuEndoQ. The mixture of 5′-³²P-labeled ssDNA (24 and 25 nt) was loaded on lane 1. The marker sizes are shown on the left of the panel.

Figure 7.

Re-ligation of the cleavage products from TkoEndoQ. *U/G, 5′-Cy5-labeled 45-U24 dsDNA (lanes 2–5). *I/T, 5′-Cy5-labeled 45-I25 dsDNA (lanes 7–10). Lanes 1 and 6, size marker (23, 24, 25 and 26 nt); lanes 2 and 7, cleaved products by TkoEndoQ (negative controls); lanes 3 and 8, annealed DNA without ligation; lanes 4 and 9, ligated products; lanes 5 and 10, intact substrates (positive controls).

Figure 8.

Estimation of the amount of TkoEndoQ in T. kodakarensis cells. The T. kodakarensis cells at exponential phases were harvested, and the whole cell extracts from the indicated numbers of the cells were subjected to western blot analysis. To create a standard curve, the serially diluted recombinant TkoEndoQ protein was subjected to the western blot analysis in parallel, and the band intensities were quantified by an image analyzer.

Figure 9.

Schematic drawing of the domain structure, and an unrooted phylogenetic tree of the EndoQ proteins. (A) The domain structures of the EndoQ proteins from Archaea and Bacteria are shown, with their total amino acid numbers. The conserved PHP domains (shaded in gray) and zinc finger motifs (black bar) are indicated in each protein. The highly conserved motifs I–IV (black bars), Glu and His (white lines) are indicated in the PHP domains of each protein. (B) An unrooted phylogenetic tree of PHP domains showing five clusters. Twenty one sequences were retrieved from the protein database at NCBI. The sequences were aligned by MAFFT (http://mafft.cbrc.jp/alignment/software/) and were further adjusted by hand. The alignment was used in generating a phylogenetic tree based on the neighbor-joining method. For a detailed description and the full names of organisms and accession numbers of proteins, see Supplementary Figure S3.

Figure 10.

Models for the EndoQ-mediated dI repair pathway. Left pathway: EndoQ recognizes dI in DNA and incises on its 5′-side. A DNA polymerase is recruited to the nicked site and synthesizes a new DNA strand coupled with the 5′–3′ strand displacement of the forward strand, and then Fen1 cuts off the resulting flap before DNA ligase seals the nick. Right pathway: EndoQ and EndoV cooperate in excising the damaged DNA portion, by cleaving on its 5′- and 3′-sides, respectively, followed by removal by helicase activity. Subsequently, a DNA polymerase fills the gap and DNA ligase seals the nick.