Genomic DNA in all organisms is constantly threatened by different kinds of endogenous and exogenous agents, yielding a variety of DNA lesions 1, 2. Among them, the apurinic/apyrimidinic (abasic, AP, Fig. 1A) site produced from the hydrolysis of a nucleotide’s glycosidic bond is one of the most frequent spontaneous DNA lesions, and induced by DNA alkylating and oxidative agents, ionizing radiation, and UV light 3, 4, 5, 6. AP sites are cytotoxic because they stall DNA replication and transcription 7, 8. The AP site can undergo β-elimination spontaneously or catalyzed by enzymes or polyamine molecules to yield a 3′-phospho-α,β-unsaturated aldehyde (3′-PUA, Fig. 1B) within a single-strand break 9, 10, 11, 12. The AP site is mainly repaired by base excision repair (BER) initiated by AP endonucleases or lyases [13]. AP endonucleases are subdivided into two different families, i.e., exonuclease III (Exo III) and endonuclease IV (Endo IV), based on structural folding and amino acid sequence similarity [14]. Escherichia coli (E. coli) and Saccharomyces cerevisiae (S. cerevisiae) have both types while humans only have the former one. S. cerevisiae AP endonuclease 1 (yApn1) belongs to the Endo IV family that accounts for greater than 97% of the total AP endonuclease activities [15]. This enzyme is also a 3′ to 5′ exonuclease, 3′-phosphodiesterase, and has nucleotide incision activity that incises oxidized DNA lesions, such as 5,6-dihydro-2′-deoxyuridine 16, 17, 18. To our knowledge, the known substrates (e.g., AP site and 3′-PUA) of yApn1 are small DNA lesions. In this study, we reported that yApn1 repairs two types of bulky DNA-peptide cross-links (DpCs) and DNA-protein cross-links (DPCs) arising from AP sites and 3′-PUA.

AP sites, if unrepaired, will react with nucleobase amines and nucleophiles (e.g., Lys and Cys residues) within proteins to form DNA-DNA and DNA-protein cross-links, respectively 19, 20, 21, 22. DPCs derived from AP sites can be classified into two types based on the flanking DNA structures. Type 1 DPCs (AP-DPCs) involve conjugating proteins to AP sites within uncleaved DNA through Schiff base (Fig. 1C), thiazolidine, or S-glycosidic linkages [22]. In type 2 DPCs, the DNA strands within Schiff base AP-DPCs are incised via β-elimination that yields 3′-PUA-DPCs (Fig. 1D), which are also produced from direct conjugation of 3′-PUA to several DNA-binding proteins 22, 23, 24, 25, 26. Schiff base DPCs are chemically reversible but long-lived with a half-life of several hours under physiological pH and temperature 27, 28. AP-DPCs need to be repaired because they physically block DNA replication [29]. 3′-PUA-DPCs need to be eradicated as 3′-hydroxyl groups are required for gap-filling DNA synthesis and strand ligation [30]. How these DPCs are repaired is not well understood. An emerging model suggests that DPCs are proteolyzed by the 26 S proteasome and/or proteases, such as SPRTN in humans and Wss1 in S. cerevisiae, followed by removal of the remaining DpCs by canonical DNA repair pathways 22, 31, 32, 33, 34.

The reversibility of Schiff base DPCs has impeded the study of their repair mechanisms. Recently, we chemically synthesized stable and site-specific AP-DpCs (Fig. 1E) and AP-DPCs (Fig. 1F) and found that E. coli Endo IV, but not Exo III, incises the adducts followed by DNA polymerase I-mediated strand-displacement DNA synthesis and flap removal to eradicate the cross-links [35]. In another study, we synthesized Schiff base 3′-PUA-DPC analogs (Fig. 1G-H) and demonstrated that they were removed by E. coli Exo III and its human homolog, AP endonuclease 1 (hAPE1) [28]. In this study, we report the repair of these two types of DNA-peptide/protein cross-links by yApn1.