DNA Double-Strand Break And Its Repair: A Potential Molecular Mechanism For HBV Integration

Objective The process of HBV DNA replication does not require viral DNA integration into host chromosomes, which, however, can be detectable in majority (up to 80%) of HBV-associated HCC. Studies have shown that integrated HBV DNA can result in, or be accompanied by, insertional mutagenesis, chromosomal deletions, secondary rearrangement and lead to genomic instability and finally to the development of HCC. It is also generally accepted that the HBV in integrated form has, in itself, some oncogenic activities. Most research in the past mainly focused on the exploration of highly preferred integration sites and those target cellular genes such as oncogenes, tumor suppressor genes, signal transducers and cell cycle regulators by a large scaled analysis in HBV-induced hepatocellular carcinomas. So far, more than sixty genes targeted by viral integration have been identified, involving retinoic acid receptor, cyclin A, human telomerase reverse transcriptase (hTERT) etc, most of which may play an important role in hepatocarcinogenesis. The viral insertion site was distributed over all chromosomes except 13, X, and Y. But some studies have shown that hTERT is a recurrent site of viral integration. Nevertheless, our understanding of molecular mechanism of integration is incomplete, largely due to its complexity and multiplicity displayed by the results of those research, for integration can take place at multiple sites on various chromosomes and a rather fixed integrated form has not yet been detected. It is commonly held that the integration appears to occur at random sites within host genome. How does HBV integrate? What has happened to host cell before integration? Can the integration be controlled? To clarify those questions is essential for the development of effective interventions of HBV DNA integration and carcinogenesis progress. Over the past few years, there have been some clues for understanding the process of hepatitis B virus DNA integration into host cell genome by DNA double-strand breaks (DSB) and its repair. Some have hypothesized that genomic DNA double-strand breaks are targets for hepadnaviral DNA integration and suggested that the integration at sites of chromosomal DSB is mediated by DNA repair pathway of NHEJ. In mammalian cells, DSB can be repaired either by homologous recombination (HR) or non-homologous end joining (NHEJ). The basic mechanism and factor requirements of the two pathways are different. HR is the error-free repair pathway following DNA replication, using an undamaged sister chromatid as a template to accurately repair the damage. In contrast, NHEJ repairs DSB by joining two nonhomologous DNA segments together. Since there are potential risks of gene deletion, insertion, indirect or direct repeats during the process of NHEJ, it is known as an error-prone repair pathway. The major factor of HR is the Rad52 protein, which can protect the DNA from exonuclease activity and activate end-to-end interactions when it binds to the site of DSB. And the central factor of NHEJ in organisms from yeast to man is the Ku protein, a heterodimer of two sequence-related subunits (Ku70 and Ku80). Ku and Rad52 compete for available DNA ends generated at break sites and trigger different repair pathway. Therefore, Rad52 protein is regarded as the―gatekeeper‖for HR, while Ku protein is the―gatekeeper‖for NHEJ. The two kinds of gatekeeper genes have shown the potential to regulate the ratio of HR to NHEJ in many research projects. Since molecular evidence clearly implicated DNA double-strand breaks as potential targets for hepadnavirus DNA integrations, our aim was to investigate the natural history of integrations and the effect of gatekeepers (Rad52, Ku70 and Ku80) on the frequency of viral integration.Methods To induce specific DNA breaks, I-SceⅠsystem of Saccharomyces cerevisiae that consists of a target plasmid and an expression vector, pEGFP2 and pCMV(3×NLS)I-SceⅠ, was used. The starting plasmid for pEGFP2 construction was pEGFP-C1. There are two mutated GFP genes in the plasmid pEGFP2, Sce-GFP and iGFP. Sce-GFP is mutated through the insertion of the I-SceⅠsite at a BcgⅠrestriction site by substituting 11 bp of wild-type GFP sequence. And the downstream of SceGFP in the same orientation is a 0.4 kb truncated GFP gene (iGFP) inserted between the site of EcoRⅠand BamHⅠ. The plasmid pEGFP2 was firstly transfected into Human fetal livel cell line L-02 and Human hepatoma cell line HepG2. The positive neomycin-resistant transfected cell clones were generated by G418 selection. Then the positive cells containing an 18-bp I-SceⅠendonuclease site were transfected with pCMV(3×NLS)I-SceⅠ, an I-SceⅠexpression plasmid. At 24 h post-transfection with pCMV(3×NLS)I-SceⅠ,γ-H2AX, a phosphorylated histone at its C terminus on serine 139, as an early cellular marker of DSB, was detected using immunocytochemistry analysis and Western blot. To further confirm that DSB occurs at exact site of I-SceⅠ, a nested PCR was designed to determinate it. Serum samples from 45 patients with chronic hepatitis B virus infection were collected. All of the samples contained high copies of HBV-DNA. Twenty-four hours after transient transfection with pCMV(3×NLS)I-SceⅠ, Cells were cultured with HBV serum. In order to release low-density lipoprotein receptor (LDL-R) to incept more HBV particles, the cell-bound lipoproteins in surface of L-02 and HepG2 should be removed prior to the addition of the viral inoculum. The cells were treated on ice with high molecular weight dextran sulphate plus EGTA in ice-cold PBS for 10 min. After incubation with HBV DNA on ice for 1h, the cells were washed extensively with ice-cold PBS and cultured in fresh medium. 48 h later, the infected cells were grown in the presence of additional human insulin and dexamethasone, both of which accelerated HBV integration. HBsAg and HBeAg in the culture supernatant were measured using ELISA (Enzyme Linked Immunosorbent Assay) kits. The results were illustrated with P/N value (P/N=sample A/negative control A; A stands for the amount of light absorbent). Genomic DNA was extracted from the virus-infected cells at indicated time. PCR amplification and DNA sequencing were used to identify the sequences of site-specific integration. By performing BLAST search, nucleotide sequence of insertion was compared with sequence databases of HBV DNA. According to the general guidelines for siRNA target selection, two target sequences to every gene were identified with the help of siRNA Wizard (www.sirnawizard.com). The seven pairs of oligonucleotides were synthesized and inserted into the plasmid psiRNA-hH1neo. After sequence was identified, six vector-derived siRNAs (denoted psiRNA1-6) and one mocking siRNA, psiRNA7 were constructed. Among them, psiRNA1 and psiRNA2 were targeted to Rad52, psiRNA3 and psiRNA4 to Ku70, psiRNA5 and psiRNA6 to Ku80, respectively. The mocking psiRNA7 was used as control. The psiRNAs were delivered into cells using Lipofectamine? 2000 reagent. The cells were incubated at 37°C in a humidified CO2 incubator for 48 hours until genes have been silenced. siRNA-induced silencing of gatekeeper genes was determined by using RT-PCR at RNA level and Western Blot at protein level. The more efficient siRNAs targeting to every gatekeeper gene were selected for the further study. To explore the effect of psiRNAs on the ratio of HR to NHEJ, the more effective psiRNAs were used to down-regulate the expression of gatekeeper genes prior to the addition of HBV serum. At 24 h after interference with psiRNAs, the repair of DSB by successful homologous recombination was assessed by the fluorescence intensity using fluorescence microscope and FCM assay. Moreover, HBV DNA integrated into site-specific DSB was quantified by a TaqMan-based real-time PCR assay at the 8th day post-interference with siRNAs.Results After three weeks of G418 selection, a cleavage site for the rare-cutting endonuclease I-SceⅠwas introduced into the genome of L-02 and HepG2 cells. When these cells were exposed to the plasmid pCMV(3×NLS)I-SceⅠ, a DSB would be induced at the site of I-SceⅠrecognition site. To investigate DSB in cells transfected with the two plasmids successfully, monoclonal anti-phospho-H2AX antibody against a C-terminal peptide containing phosphorylated Ser-139 was used. The results of immunofluorescence studies and Western Blot showed significant phosphorylation of histone H2AX in cells 24 h after transfection with pCMV(3×NLS)I-SceⅠ. And the result of nested PCR exhibited a visible band of approximately 600 bp that could not be digested by I-SceⅠ, which indicated that the EGFP sequence had been successfully introduced into the genome of transfected cells and a potential DSB at the I-SceⅠrecognition sequence had been undergone repair by HR. When insulin and dexamethasone were added to the basal cell culture medium, HBsAg and HBeAg can be detected in culture supernatant of HepG2 and L-02. During the first three days, the expression of HBsAg and HBeAg decreased significantly in both cell lines, especially in L-02. On the fourth day post-incubation under hormones condition, expression of HBsAg was almost not detectable in L-02 but still positive in HepG2 cell line. On the fifth day, HBeAg was negative in both cells. Genomic DNA was extracted from cell pellets on the 8th day post-infection with HBV inoculum and digested with I-SceⅠto enrich DNA lacking I-SceⅠrecognition site. Using the primers spanning the both flanks of DSB, multiple fragments with different sizes were generated by PCR amplification. It was suggested that the inserted sequence was not unique. Here, the densest bands extracted and purified from 1% agarose gel were subjected to direct sequencing. By performing BLAST search, a tool for nucleotide sequences analysis, inserted nucleotide sequence was compared with sequence databases of HBV DNA. It is exciting to see that the nucleotide contains part of x region of Hepatitis B virus DNA and the percent identity calculated between HBx and DNA sequence within the similar HBx domain is 65 %. The result demonstrated directly that DNA double-strand break is a potential and preferred target site for HBV DNA integration. The inhibitory effects of psiRNAs at mRNA Level were determined by semi-quantity RT-PCR. The transcript quantities of targeted genes were normalized by the GAPDH mRNA. psiRNA1 and psiRNA2 inhibited the Rad52 mRNA by 83.75% and 56.50% , psiRNA3 and psiRNA4 inhibited the Ku70 mRNA by 62.45% and 71.92% , psiRNA5 and psiRNA6 inhibited the Ku80 mRNA by 77.59% and 60.41% , respectively. psiRNA7 could not affect the three genes expression at mRNA level. Inhibitory Effects of psiRNAs at Protein Level were detected by Western Blot: psiRNA1 and psiRNA2 inhibited the Rad52 protein by 70.92% and 51.65%, psiRNA3 and psiRNA4 inhibited the Ku70 protein by 54.02% and 65.24%, psiRNA5 and psiRNA6 inhibited the Ku80 protein by 67.14% and 66.83%, respectively. To sum up, the results of RT-PCR and Western Blot were fully compatible to show that psiRNA1, psiRNA4 and psiRNA5 were more effective than their counterparts. The three plasmids were selected to investigate their effect on the two pathways of DSB repair and site-specific integration of HBV DNA. The result of fluorescence microscopy showed that less EGFP was expressed in cells treated with Rad52 siRNA ( psiRNA1) but the expression of EGFP increased significantly in cells treated with Ku70 and Ku80 siRNA (psiRNAkus, a tandem Ku-shRNA-encoding plasmid). The result of flow cytometry further verified the above findings, one major peak of fluorescence can be observed in HepG2 cell line treated with psiRNAkus. Results of the real-time PCR data are represented as Ct (threshold cycle) values and the Value of Ct is inverse ratio to the level of template. It is apparent that the Ct value 0.65±0.09 of the sample derived from psiRNA1 is much lower than that 2.71±0.33 of sample from psiRNAkus, indicating that site-specific integration of HBx DNA is enhanced by siRNA targeting against Rad52 and down-regulated by siRNA targeting Kus gene.Conclusions These findings presented direct evidence that DNA double-strand breaks are potential and preferred targets for HBV integration. The study has also shown that shRNAs targeted against gatekeeper genes for HR or NHEJ can regulated the ratio of the two DSB repair pathways and even the frequency of HBV integration, which proved the competition between HR and NHEJ and the gatekeeper status of Rad52 and Ku protein. Furthermore, our research has provided a novel strategy for interfering HBV DNA integration or even hepatocarcinogenesis.