Pa-AGOG, the founding member of a new family of archaeal 8-oxoguanine DNA-glycosylases

Abstract

Oxidative damage represents a major threat to genomic stability, as the major product of DNA oxidation, 8-oxoguanine (GO), frequently mispairs with adenine during replication. In order to prevent these mutagenic events, organisms have evolved GO-DNA glycosylases that remove this oxidized base from DNA. We were interested to find out how GO is processed in the hyperthermophilic archaeon Pyrobaculum aerophilum, which lives at temperatures around 100 degrees C. To this end, we searched its genome for open reading frames (ORFs) bearing the principal hallmark of GO-DNA glycosylases: a helix-hairpin-helix motif and a glycine/proline-rich sequence followed by an absolutely conserved aspartate (HhH-GPD motif). Interestingly, although the P.aerophilum genome encodes three such ORFs, none of these encodes the potent GO-processing activity detected in P.aerophilum extracts. Fractionation of the extracts, followed by analysis of the active fractions by denaturing polyacrylamide gel electrophoresis, showed that the GO-processing enzyme has a molecular size of approximately 30 kDa. Mass spectrometric analysis of proteins in this size range identified several peptides originating from P.aerophilum ORF PAE2237. We now show that PAE2237 encodes AGOG (Archaeal GO-Glycosylase), the founding member of a new family of DNA glycosylases, which can remove GO from single- and double-stranded substrates with great efficiency.

Figure 1

Substrate specificity of GO-processing activity in crude extracts of P.aerophilum: 1 pmol of the 60mer substrates was incubated with 0, 2 or 10 μg of WCE as described in Materials and Methods. The positions of the full-length 60mer substrate and of the 23mer product are indicated; F indicates the fluorescein-labeled strand; ss, single-stranded GO 60mer.

Figure 2

Fractionation of P.aerophilum extracts and estimation of the molecular weight of the GO-DNA glycosylase. (A) Profile of Fraction V eluted from a HiTrap SP column with a linear gradient from 0.04 to 0.6 M NaCl. Inset shows the GO/G activity in 8 μl of the load (L), flow-through (F), wash (W) and fractions 6–12. Incubations were carried out for 30 min at 60°C, using 1 pmol of the GO/G substrate. The positions of the full-length 60mer substrate and of the 23mer product are indicated. (B) Recovery and renaturation of the protein eluted from 11 slices of the 15% SDS–PAGE gel (left panel; M, molecular weight standards; VI, 25 μl of fraction VI). The GO/G-processing activity was found to reside predominantly in slice 9 (right panel).

Figure 3

Identification of the P.aerophilum GO-glycosylase (PAE2237) by mass spectrometry. (A) Left panel: TCA precipitate from 2 ml of Fraction VI (Materials and Methods) was resolved on a 12.5% SDS–PAGE gel and stained with Coomassie blue. M, Bio-Rad SDS–PAGE molecular weight standards. The two bands (1 and 2) detected at ∼30 kDa were cut out and subjected to a tryptic digest (Materials and Methods) followed by MALDI-TOF Mass Spectrometry (MS) analysis of the peptide fragments. Right panel: MALDI-TOF MS analysis of band 1. Measured masses and amino acid sequences of the tryptic peptides are listed in the table. ‘a’ indicates measured minus calculated mass. (B) The peptide sequences identified by MS (shown in boldface) could be assigned to ORF PAE2237 in the P.aerophilum genome that encodes a hypothetical protein of 256 amino acids (29.5 kDa).

Figure 4

Pa-AGOG is a member of a new family of DNA glycosylases. (A) Partial sequence alignment of the putative ‘extended’ helix–hairpin–helix motif of AGOG homologues. PAE2237, P.aerophilum (TrEMBL:Q8ZVK6); APE0710, Aeropyrum pernix (TrEMBL: Q9YE60); PF0904, Pyrococcus furiosus (TrEMBL: Q8U2D, for simplicity, the sequence of only one of the three Pyrococcus species is shown); MK0541, Methanopyrus kandleri (TrEMBL: Q8TXW8); MMP0304, Methanococcus maripaludis (Trnew: CAF29860); NEQ515, Nanoarchaeum equitans (Trnew: AAR39356). Identical residues are shaded and the putative active site residues (K140Q, K147Q and D172N) are indicated by arrowheads. The sequence alignment was generated using the MultAlin software (23), available at www.toulouse.infra.fr. (B) Substrate specificity comparison of Pa-AGOG, human OGG1 and E.coli MutM. The GO/C, GO/G and GO/A substrates (1 pmol) were incubated for 15 min at 60°C in the absence of enzyme (lanes 1, 5, 10 and 14) or with 1 pmol of Pa-AGOG (Pae, lanes 2, 6, 8, 11 and 15) or at 37°C with 10 pmol of hOGG1 (Hs, lanes 3, 7, 9, 12 and 16) or 1 pmol of MutM (Ec, lanes 4, 13 and 17). Samples were then either directly precipitated (lanes 1–4), treated with 100 mM NaOH (OH) for 10 min at 90°C (lanes 5–7) or incubated with 0.5 pmol of Pa-EndoIV (nfo) for 10 min at 60°C (lanes 8 and 9). The 60mer substrate and the positions of the three possible 23mer products are indicated on the left. NB: Some GO is lost spontaneously at 60°C and the resulting AP-sites (∼10%) are cleaved by the hot alkali treatment (lane 5). (C) The highly efficient DNA glycosylase activity of Pa-AGOG is modulated by the base opposite GO; 10 pmol of the labeled GO/G and GO/C substrates were incubated for 15 min at 60°C with 1 pmol of Pa-AGOG in the presence or absence of 100 pmol of unlabeled 60mer competitor duplexes GO/C (lanes 3 and 7) or GO/G (lanes 4 and 6). The reactions were immediately quenched by the addition of 100 mM NaOH (10 min at 90°C) to measure DNA glycosylase activity. Lane 1, no enzyme; lanes 1, 2 and 5, no competitor DNA added. (D) Comparison of DNA-glycosylase (dashed lines) and DNA-glycosylase/lyase (solid lines) activities of Pa-AGOG on GO/G (squares) and GO/C (triangles) substrates. Pa–AGOG (1 pmol) was incubated with 10 pmol of the labeled substrates. After 5, 15, 30 and 60 min, aliquots of the reaction mixtures were removed and either immediately treated with hot alkali to measure glycosylase activity or directly ethanol-precipitated to measure combined glycosylase/AP-lyase activity.

Figure 5

Effect of mutations of the conserved lysines 140 and 147 and aspartate 172 on the glycosylase/lyase activity of Pa-AGOG. The labeled GO/G substrate (1 pmol) was incubated for 15 min at 60°C alone (lane 1), or with 1 pmol of either wild-type (wt) Pa-AGOG (lanes 2 and 3), Pa-AGOG K140Q (lanes 4 and 5), Pa-AGOG K147Q (lanes 6 and 7) or Pa-AGOG D172N (lanes 8 and 9). HT indicates that the proteins were heat-treated for 15 min at 80°C prior to addition of the substrate (lanes 3, 5, 7 and 9). Heat treatment did not affect the glycosylase/lyase activity of the wild-type enzyme (lane 3). The K140Q (lanes 4 and 5) and D172N (lanes 8 and 9) mutations severely attenuated the enzymatic activity of Pa-AGOG on the GO/G substrate, while substitution of K147 with glutamine resulted in a substantial reduction of the catalytic activity of the protein, while making the polypeptide also thermolabile (lanes 6 and 7).

Figure 6

Unrooted Phylogenetic Tree of the HhH-Superfamily of DNA glycosylases. All 35 protein sequences were aligned using the ClustalW program (http://www.ebi.ac.uk/clustalw). Distance analysis of the phylogenetic tree was performed using the program TreeView 1.6.6 (31). The abbreviations used in the figure are as follows: Bacteria: bh, Bacillus halodurans; bs, Bacillus subtilis; ec, E.coli; hi, Haemophilus influenzae; hp, Helicobacter pylori; tm, Thermotoga maritima; tt, Thermus thermophilus; Archaea: af, Archaeoglobus fulgidus; mth, Methanobacterium thermoformicicum; mj, M.jannaschii; mk, Methanopyrus kandleri; Eukaryotes: at, Arabidopsis thaliana; hs, Homo sapiens; sc, S.cerevisiae; sp, S.pombe. Members of the AGOG family are indicated in Figure 4A. Swiss-Prot/TrEMBL accession numbers are also indicated.