Back-Up Base Excision DNA Repair in Human Cells Deficient in the Major AP Endonuclease, APE1

Abstract

Apurinic/apyrimidinic (AP) sites are abundant DNA lesions generated both by spontaneous base loss and as intermediates of base excision DNA repair. In human cells, they are normally repaired by an essential AP endonuclease, APE1, encoded by the APEX1 gene. Other enzymes can cleave AP sites by either hydrolysis or β-elimination in vitro, but it is not clear whether they provide the second line of defense in living cells. Here, we studied AP site repairs in APEX1 knockout derivatives of HEK293FT cells using a reporter system based on transcriptional mutagenesis in the enhanced green fluorescent protein gene. Despite an apparent lack of AP site-processing activity in vitro, the cells efficiently repaired the tetrahydrofuran AP site analog resistant to β-elimination. This ability persisted even when the second AP endonuclease homolog, APE2, was also knocked out. Moreover, APEX1 null cells were able to repair uracil, a DNA lesion that is removed via the formation of an AP site. If AP site hydrolysis was chemically blocked, the uracil repair required the presence of NTHL1, an enzyme that catalyzes β-elimination. Our results suggest that human cells possess at least two back-up AP site repair pathways, one of which is NTHL1-dependent.

Repair and DNA Polymerase Bypass of Clickable Pyrimidine Nucleotides

“…When both phosphodiester bonds around dU were replaced with phosphorothioate linkages (sUs modification), which inhibit strand cleavage by AP endonucleases and AP lyases [ 62 , 63 , 64 ], the fluorescence signal was elevated significantly, although the magnitude of this increase was only about 4.2-fold. These results are consistent with our earlier observations in HEK293 cells [ 34 ] and suggest the existence of a still-to-be-identified backup repair pathway for genomic uracil. However, the outcome for EtU and C8-AlkU was strikingly different.…”

“…Plasmids pZAJ_Q205* carrying modified or unmodified nucleotides in the eGFP gene were generated as previously described [ 33 , 34 ]. HeLa cells were seeded in 12-well plates at 10 5 /well density, grown for 24 h and transfected with 100 ng of the construct and 300 ng of pDsRed-Monomer-N1 (Takara Bio, Kusatsu, Japan) using the Effectene reagent (Qiagen, Venlo, The Netherlands).…”

Probing the Conformational Restraints of DNA Damage Recognition with β-L-Nucleotides

“…If βLdA was placed at this position instead, we observed a severalfold increase in fluorescence (to ~7%), indicative of the error-prone bypass ( Figure 6 b). This level was slightly below the fluorescence of F modified with a 5′-phosphotioate linkage (sF), a model abasic lesion resistant to APE1 but subject to less efficient removal by some back-up pathways, possibly nucleotide excision repair (NER) [ 35 , 38 ]. Even larger fluorescent cell populations (~20% of the G-construct) were produced by βLdC and βLdT, suggesting either slower repair of these nucleotides or their higher miscoding potential ( Figure 6 b).…”

“…From the behavior of DNA polymerases studied here, we do not expect significant preference for non-instructive UMP incorporation opposite βLdNs by RNA polymerases, and repair remains the most likely reason for low fluorescence. The fluorescence of βLdG-transfected cells was significantly lower than that of the cells transfected by sF, a known modification resistant to BER but slowly repaired through an unidentified pathway, possibly NER [ 35 , 38 ]. We suggest that at least βLdG is subject to repair in human cells, although NER is unlikely to play a role here since the human NER system does not recognize single βLdG residues [ 76 ].…”

“…Plasmids pZAJ_Q205* carrying βLdN modifications or canonical nucleotides in the eGFP reporter gene between two Nb.Bpu10I nickase recognition sites were generated as previously described [ 38 , 92 ]. The pZAJ_5c plasmid with the original EGFP sequence [ 37 ] was used for fluorescence data normalization.…”