The excision of 3' penultimate errors by DNA polymerase I and its role in endonuclease V-mediated DNA repair

Base preference for inosine 3'-riboendonuclease activity of human endonuclease V: implications for cleavage of poly-A tails containing inosine

Deamination of bases is a form of DNA damage that occurs spontaneously via the hydrolysis and nitrosation of living cells, generating hypoxanthine from adenine. E. coli endonuclease V (eEndoV) cleaves hypoxanthine-containing double-stranded DNA, whereas human endonuclease V (hEndoV) cleaves hypoxanthine-containing RNA; however, hEndoV in vivo function remains unclear. To date, hEndoV has only been examined using hypoxanthine, because it binds closely to the base located at the cleavage site. Here, we examined whether hEndoV cleaves other lesions (e.g., AP site, 6-methyladenine, xanthine) to reveal its function and whether 2'-nucleoside modification affects its cleavage activity. We observed that hEndoV is hypoxanthine-specific; its activity was the highest with 2'-OH modification in ribose. The cleavage activity of hEndoV was compared based on its base sequence. We observed that it has specificity for adenine located on the 3'-end of hypoxanthine at the cleavage site, both before and after cleavage. These data suggest that hEndoV recognizes and cleaves the inosine generated on the poly A tail to maintain RNA quality. Our results provide mechanistic insight into the role of hEndoV in vivo.

Human Mitochondrial Protein HSPD1 Binds to and Regulates the Repair of Deoxyinosine in DNA

The generation of deoxyinosine (dI) in DNA is one of the most important sources of genetic mutations, which may lead to cancer and other human diseases. A further understanding of the biological consequences of dI necessitates the identification and functional characterizations of dI-binding proteins. Herein, we employed a mass spectrometry-based proteomics approach to detect the cellular proteins that may sense the presence of dI in DNA. Our results demonstrated that human mitochondrial heat shock protein 60 (HSPD1) can interact with dI-bearing DNA. We further demonstrated the involvement of HSPD1 in the sodium nitrite-induced DNA damage response and in the modulation of dI levels in vitro and in human cells. Together, these findings revealed HSPD1 as a novel dI-binding protein that may play an important role in the mitochondrial DNA damage control in human cells.

Development of an Endonuclease V-Assisted Analytical Method for Sequencing Analysis of Deoxyinosine in DNA

Deoxyinosine (dI) is a highly mutagenic lesion that preferentially pairs with deoxycytidine during replication, which may induce A to G transition and ultimately contribute to carcinogenesis. Therefore, finding the site of dI modification in DNA is of great value for both basic research and clinical applications. Herein, we developed a novel method to sequence the dI modification site in DNA, which utilizes endonuclease V (EndoV)-dependent deamination repair to specifically label the modification site with biotin-14-dATP that allows the affinity enrichment of dI-bearing DNA for sequencing. We have achieved efficient determination of the location of the modified nucleotide in dI-bearing plasmid DNA with the assistance of EndoV-dependent deamination repair. We have also successfully applied this approach to locate the dI modification sites in the mitochondrial DNA of human cells. Our method should be generally applicable for genome-wide sequencing analysis of dI modifications in living organisms.

Arabidopsis thaliana endonuclease V is a ribonuclease specific for inosine-containing single-stranded RNA

Endonuclease V is highly conserved, both structurally and functionally, from bacteria to humans, and it cleaves the deoxyinosine-containing double-stranded DNA in Escherichia coli, whereas in Homo sapiens it catalyses the inosine-containing single-stranded RNA. Thus, deoxyinosine and inosine are unexpectedly produced by the deamination reactions of adenine in DNA and RNA, respectively. Moreover, adenosine-to-inosine (A-to-I) RNA editing is carried out by adenosine deaminase acting on dsRNA (ADARs). We focused on Arabidopsis thaliana endonuclease V (AtEndoV) activity exhibiting variations in DNA or RNA substrate specificities. Since no ADAR was observed for A-to-I editing in A. thaliana, the possibility of inosine generation by A-to-I editing can be ruled out. Purified AtEndoV protein cleaved the second and third phosphodiester bonds, 3' to inosine in single-strand RNA, at a low reaction temperature of 20-25°C, whereas the AtEndoV (Y100A) protein bearing a mutation in substrate recognition sites did not cleave these bonds. Furthermore, AtEndoV, similar to human EndoV, prefers RNA substrates over DNA substrates, and it could not cleave the inosine-containing double-stranded RNA. Thus, we propose the possibility that AtEndoV functions as an RNA substrate containing inosine induced by RNA damage, and not by A-to-I RNA editing in vivo.

Proofreading and DNA Repair Assay Using Single Nucleotide Extension and MALDI-TOF Mass Spectrometry Analysis

The maintenance of the genome and its faithful replication is paramount for conserving genetic information. To assess high fidelity replication, we have developed a simple non-labeled and non-radio-isotopic method using a matrix-assisted laser desorption ionization with time-of-flight (MALDI-TOF) mass spectrometry (MS) analysis for a proofreading study. Here, a DNA polymerase [e.g., the Klenow fragment (KF) of Escherichia coli DNA polymerase I (pol I) in this study] in the presence of all four dideoxyribonucleotide triphosphates is used to process a mismatched primer-template duplex. The mismatched primer is then proofread/extended and subjected to MALDI-TOF MS. The products are distinguished by the mass change of the primer down to single nucleotide variations. Importantly, a proofreading can also be determined for internal single mismatches, albeit at different efficiencies. Mismatches located at 2-4-nucleotides (nt) from the 3' end were efficiently proofread by pol I, and a mismatch at 5 nt from the primer terminus showed only a partial correction. No proofreading occurred for internal mismatches located at 6 - 9 nt from the primer 3' end. This method can also be applied to DNA repair assays (e.g., assessing a base-lesion repair of substrates for the endo V repair pathway). Primers containing 3' penultimate deoxyinosine (dI) lesions could be corrected by pol I. Indeed, penultimate T-I, G-I, and A-I substrates had their last 2 dI-containing nucleotides excised by pol I before adding a correct ddN 5'-monophosphate (ddNMP) while penultimate C-I mismatches were tolerated by pol I, allowing the primer to be extended without repair, demonstrating the sensitivity and resolution of the MS assay to measure DNA repair.

Application of single nucleotide extension and MALDI-TOF mass spectrometry in proofreading and DNA repair assay

Proofreading and DNA repair are important factors in maintaining the high fidelity of genetic information during DNA replication. Herein, we designed a non-labeled and non-radio-isotopic simple method to measure proofreading. An oligonucleotide primer is annealed to a template DNA forming a mismatched site and is proofread by Klenow fragment of Escherichia coli DNA polymerase I (pol I) in the presence of all four dideoxyribonucleotide triphosphates. The proofreading excision products and re-synthesis products of single nucleotide extension are subjected to MALDI-TOF mass spectrometry (MS). The proofreading at the mismatched site is identified by the mass change of the primer. We examined proofreading of Klenow fragment with DNAs containing various base mismatches. Single mismatches at the primer terminus can be proofread efficiently. Internal single mismatches can also be proofread at different efficiencies, with the best correction for mismatches located 2-4-nucleotides from the primer terminus. For mismatches located 5-nucleotides from the primer terminus there was partial correction and extension. No significant proofreading was observed for mismatches located 6-9-nucleotides from the primer terminus. We also subjected primers containing 3' penultimate deoxyinosine (dI) lesions, which mimic endonuclease V nicked repair intermediates, to pol I repair assay. The results showed that T-I was a better substrate than G-I and A-I, however C-I was refractory to repair. The high resolution of MS results clearly demonstrated that all the penultimate T-I, G-I and A-I substrates had been excised last 2 dI-containing nucleotides by pol I before adding a correct ddNMP, however, pol I proofreading exonuclease tolerated the penultimate C-I mismatch allowing the primer to be extended by polymerase activity.

Insights into the role of endonuclease V in RNA metabolism in Trypanosoma brucei

Inosine may arise in DNA as a result of oxidative deamination of adenine or misincorporation of deoxyinosine triphosphate during replication. On the other hand, the occurrence of inosine in RNA is considered a normal and essential modification induced by specific adenosine deaminases acting on mRNA and tRNA. In prokaryotes, endonuclease V (EndoV) can recognize and cleave inosine-containing DNA. In contrast, mammalian EndoVs preferentially cleave inosine-containing RNA, suggesting a role in RNA metabolism for the eukaryotic members of this protein family. We have performed a biochemical characterization of EndoV from the protozoan parasite Trypanosoma brucei. In vitro, TbEndoV efficiently processes single-stranded RNA oligonucleotides with inosine, including A to I-edited tRNA-like substrates but exhibits weak activity over DNA, except when a ribonucleotide is placed 3' to the inosine. Immunolocalization studies performed in procyclic forms indicate that TbEndoV is mainly cytosolic yet upon nutritional stress it redistributes and accumulates in stress granules colocalizing with the DEAD-box helicase TbDhh1. RNAi-mediated depletion of TbEndoV results in moderate growth defects in procyclic cells while the two EndoV alleles could be readily knocked out in bloodstream forms. Taken together, these observations suggest an important role of TbEndoV in RNA metabolism in procyclic forms of the parasite.

Deoxyinosine repair in nuclear extracts of human cells

BACKGROUND

Deamination of adenine can occur spontaneously under physiological conditions generating the highly mutagenic lesion, hypoxanthine. This process is enhanced by ROS from exposure of DNA to ionizing radiation, UV light, nitrous acid, or heat. Hypoxanthine in DNA can pair with cytosine which results in A:T to G:C transition mutations after DNA replication. In Escherichia coli, deoxyinosine (hypoxanthine deoxyribonucleotide, dI) is removed through an alternative excision repair pathway initiated by endonuclease V. However, the correction of dI in mammalian cells appears more complex and was not fully understood.

RESULTS

All four possible dI-containing heteroduplex DNAs, including A-I, C-I, G-I, and T-I were introduced to repair reactions containing extracts from human cells. The repair reaction requires magnesium, dNTPs, and ATP as cofactors. We found G-I was the best substrate followed by T-I, A-I and C-I, respectively. Moreover, judging from the repair requirements and sensitivity to specific polymerase inhibitors, there were overlapping repair activities in processing of dI in DNA. Indeed, a hereditable non-polyposis colorectal cancer cell line (HCT116) demonstrated lower dI repair activity that was partially attributed to lack of mismatch repair.

CONCLUSIONS

A plasmid-based convenient and non-radioisotopic method was created to study dI repair in human cells. Mutagenic dI lesions processed in vitro can be scored by restriction enzyme cleavage to evaluate the repair. The repair assay described in this study provides a good platform for further investigation of human repair pathways involved in dI processing and their biological significance in mutation prevention.