Repair of hydantoin lesions and their amine adducts in DNA by base and nucleotide excision repair

Abstract

An important feature of the common DNA oxidation product 8-oxo-7,8-dihydroguanine (OG) is its susceptibility to further oxidation that produces guanidinohydantoin (Gh) and spiroiminodihydantoin (Sp) lesions. In the presence of amines, G or OG oxidation produces hydantoin amine adducts. Such adducts may form in cells via interception of oxidized intermediates by protein-derived nucleophiles or naturally occurring amines that are tightly associated with DNA. Gh and Sp are known to be substrates for base excision repair (BER) glycosylases; however, large Sp-amine adducts would be expected to be more readily repaired by nucleotide excision repair (NER). A series of Sp adducts differing in the size of the attached amine were synthesized to evaluate the relative processing by NER and BER. The UvrABC complex excised Gh, Sp, and the Sp-amine adducts from duplex DNA, with the greatest efficiency for the largest Sp-amine adducts. The affinity of UvrA for all of the lesion duplexes was found to be similar, whereas the efficiency of UvrB loading tracked with the efficiency of UvrABC excision. In contrast, the human BER glycosylase NEIL1 exhibited robust activity for all Sp-amine adducts irrespective of size. These studies suggest that both NER and BER pathways mediate repair of a diverse set of hydantoin lesions in cells.

Figure 1. Structures of lesions used as substrates for NER and BER in this study

Figure 2. Representative reaction profiles of UvrABC with hydantoin lesion-containing DNA

Reactions conditions consisted of 20 nM UvrA, 100 nM UvrB, and 50 nM UvrC, 2 nM DNA duplex in assay buffer (50 mM Tris;HCl, pH 7.5, 50 mM KCl, 10 mM MgCl₂, 5 mM DTT, and 1 mM ATP) at 55 °C. Reactions and plots with UvrABC and DNA containing Sp;GPRP:A and Sp;GlcN:A were essentially identical to the reaction with Sp;GPRPGP:A.

Figure 3. Location of UvrABC incision sites near the lesion site

(A) Storage phosphor autoradiogram of UvrABC 5′;side incision site on 5′;³²P;F, OG, Gh, Sp;containing strand of 50 bp duplex substrate. The Maxam;Gilbert G + A sequencing reaction (lane G + A) was used to determine the location of the lesion site (X26) and the nucleotide cleaved by UvrABC (C19). Reactions containing 2 nM DNA duplex with 20 nM UvrA, 100 nM UvrB, and 50 nM UvrC were incubated for four hours at 55°C. (B) Storage phosphor autoradiograph of the same experiment using 3′;end;labling to visualize 3′;cleavage site. Highlighted nucleotides are the lesion site (X 26) and the UvrABC 3′incision site (C30). (C) The sequence of the 50 bp duplex containing F:A with nucleotides that are hydrolyzed by UvrABC indicated with arrows and scissors. Sites of UvrABC incision were the same for all lesion;containing duplexes.

Figure 4. Representative EMSA storage phosphor autoradiogram illustrating the formation of the UvrB*DNA pre-incision complex with F, OG, Gh, Sp, and Sp-amine adducts containing DNA

Reaction mixtures contained 200 pM DNA duplex, 100 nM UvrB, 0.5 nM UvrA, 50 mM Tris;HCl, pH 7.5, 50 mM KCl, 10 mM MgCl₂, 5 mM DTT, and 1 mM ATP, incubated at 55 °C for 20 min.

Figure 5. Sp-amine adduct removal by human and bacterial BER glycosylases

Qualitative glycosylase assays were performed with Fpg, E3Q Fpg, Nei, NEIL1, E3Q NEIL1 and hOGG1 using the 30 base pair duplex containing Gh, Sp1, Sp2, Sp;Lys, Sp;GlcN, and Sp;GPRPGP lesions base paired to C. ^,,, The asterisk indicates the lesion;containing strand containing the 5′;³²Phosphate. The enzyme reactions were performed at 37 °C and the reactions were quenched with 0.5 M NaOH after 20 minutes. The Sp;GPRPGP oligonucleotide degraded slightly to approximately 3% Sp, as evident in the slightly faster mobility in the substrate band (Lanes 36;42).

Figure 6. Structural rationale for potent activity of NEIL1 on Sp-amine adducts

(A) Structure of MvNei1E3Q bound to Tg;containing DNA. The enzyme;DNA structure surface is mapped according to hydrophobicity (ranging from yellow (high) to blue (Low)) with Tg structure shown as ball and stick model. The enzyme has very few amino acid interactions with the Tg lesion, which appears to be solvent exposed. (B) Modeled S;Sp;Lys in active site: A ring of Sp;Lys in similar orientation as Tg, while the B ring and Lys adduct are accommodated in open space at back of binding pocket. (C) Structures of Tg (red) overlaid on structure of Sp;Lys adduct. Figures in (A) and (B) were generated using the MvNei;Tg structure in the protein database (PDB 3VK8) reported by Doublie and co;workers.