Branched oligonucleotides induce in vivo gene conversion of a mutated EGFP reporter

The origin of unwanted editing byproducts in gene editing

The rapid development of CRISPR-Cas genome editing tools has greatly changed the way to conduct research and holds tremendous promise for clinical applications. During genome editing, CRISPR-Cas enzymes induce DNA breaks at the target sites and subsequently the DNA repair pathways are recruited to generate diverse editing outcomes. Besides off-target cleavage, unwanted editing outcomes including chromosomal structural variations and exogenous DNA integrations have recently raised concerns for clinical safety. To eliminate these unwanted editing byproducts, we need to explore the underlying mechanisms for the formation of diverse editing outcomes from the perspective of DNA repair. Here, we describe the involved DNA repair pathways in sealing Cas enzyme-induced DNA double-stranded breaks and discuss the origins and effects of unwanted editing byproducts on genome stability. Furthermore, we propose the potential risk of inhibiting DNA repair pathways to enhance gene editing. The recent combined studies of DNA repair and CRISPR-Cas editing provide a framework for further optimizing genome editing to enhance editing safety.

Strand bias influences the mechanism of gene editing directed by single-stranded DNA oligonucleotides

Gene editing directed by modified single-stranded DNA oligonucleotides has been used to alter a single base pair in a variety of biological systems. It is likely that gene editing is facilitated by the direct incorporation of the oligonucleotides via replication and/or by direct conversion, most likely through the DNA mismatch repair pathway. The phenomenon of strand bias, however, as well as its importance to the gene editing reaction itself, has yet to be elucidated in terms of mechanism. We have taken a reductionist approach by using a genetic readout in Eschericha coli and a plasmid-based selectable system to evaluate the influence of strand bias on the mechanism of gene editing. We show that oligonucleotides (ODNs) designed to anneal to the lagging strand generate 100-fold greater 'editing' efficiency than 'those that anneal to' the leading strand. The majority of editing events (∼70%) occur by the incorporation of the ODN during replication within the lagging strand. Conversely, ODNs that anneal to the leading strand generate fewer editing events although this event may follow either the incorporation or direct conversion pathway. In general, the influence of DNA replication is independent of which ODN is used suggesting that the importance of strand bias is a reflection of the underlying mechanism used to carry out gene editing.

Enhanced gene repair mediated by methyl-CpG-modified single-stranded oligonucleotides

Gene editing mediated by oligonucleotides has been shown to induce stable single base alterations in genomic DNA in both prokaryotic and eukaryotic organisms. However, the low frequencies of gene repair have limited its applicability for both basic manipulation of genomic sequences and for the development of therapeutic approaches for genetic disorders. Here, we show that single-stranded oligodeoxynucleotides (ssODNs) containing a methyl-CpG modification and capable of binding to the methyl-CpG binding domain protein 4 (MBD4) are able to induce >10-fold higher levels of gene correction than ssODNs lacking the specific modification. Correction was stably inherited through cell division and was confirmed at the protein, transcript and genomic levels. Downregulation of MBD4 expression using RNAi prevented the enhancement of gene correction efficacy obtained using the methyl-CpG-modified ssODN, demonstrating the specificity of the repair mechanism being recruited. Our data demonstrate that efficient manipulation of genomic targets can be achieved and controlled by the type of ssODN used and by modulation of the repair mechanism involved in the correction process. This new generation of ssODNs represents an important technological advance that is likely to have an impact on multiple applications, especially for gene therapy where permanent correction of the genetic defect has clear advantages over viral and other nonviral approaches currently being tested.

Cell death caused by single-stranded oligodeoxynucleotide-mediated targeted genomic sequence modification

Targeted gene repair directed by single-stranded oligodeoxynucleotides (ssODNs) offers a promising tool for biotechnology and gene therapy. However, the methodology is currently limited by its low frequency of repair events, variability, and low viability of "corrected" cells. In this study, we showed that during ssODN-mediated gene repair reaction, a significant population of corrected cells failed to divide, and were much more prone to undergo apoptosis, as marked by processing of caspases and PARP-1. In addition, we found that apoptotic cell death triggered by ssODN-mediated gene repair was largely independent of the ATM/ATR kinase. Furthermore, we examined the potential involvement of the mismatch repair (MMR) proteins in this "correction reaction-induced" cell death. Result showed that while defective MMR greatly enhanced the efficiency of gene correction, compromising the MMR system did not yield any viable corrected clone, indicating that the MMR machinery, although plays a critical role in determining ssODN-directed repair, was not involved in the observed cellular genotoxic responses.

Specific targeted gene repair using single-stranded DNA oligonucleotides at an endogenous locus in mammalian cells uses homologous recombination

The feasibility of introducing point mutations in vivo using single-stranded DNA oligonucleotides (ssON) has been demonstrated but the efficiency and mechanism remain elusive and potential side effects have not been fully evaluated. Understanding the mechanism behind this potential therapy may help its development. Here, we demonstrate the specific repair of an endogenous non-functional hprt gene by a ssON in mammalian cells, and show that the frequency of such an event is enhanced when cells are in S-phase of the cell cycle. A potential barrier in using ssONs as gene therapy could be non-targeted mutations or gene rearrangements triggered by the ssON. Both the non-specific mutation frequencies and the frequency of gene rearrangements were largely unaffected by ssONs. Furthermore, we find that the introduction of a mutation causing the loss of a functional endogenous hprt gene by a ssON occurred at a similarly low but statistically significant frequency in wild type cells and in cells deficient in single strand break repair, nucleotide excision repair and mismatch repair. However, this mutation was not induced in XRCC3 mutant cells deficient in homologous recombination. Thus, our data suggest ssON-mediated targeted gene repair is more efficient in S-phase and involves homologous recombination.

DNA breakage associated with targeted gene alteration directed by DNA oligonucleotides

Understanding the mechanism by which single-stranded oligonucleotides (ODNs) elicit targeted nucleotide exchange (TNE) is imperative to achieving optimal correction efficiencies and medical applicability. It has been previously shown that introduction of an ODN into cells results in the activation of DNA damage response pathways, but there has been no evaluation of the damage created at the level of the DNA. The activation of H2AX, a hallmark protein of DNA breakage, suggests that a double-strand break (DSB) could be occurring during the targeted gene alteration (TGA) reaction. Using the human HCT116 cell line with a single integrated mutant eGFP gene as our model system, we demonstrate that the DNA strand breakage occurs when a specific ODN, designed to direct TGA, is transfected into the cells. Both single- and double-stranded DNA cleavage is observed dependent on the level of ODN added to the reaction. Possible mechanisms of ODN-dependent DSB formation, as a function of TGA, are discussed herein.

Use of internally nuclease-protected single-strand DNA oligonucleotides and silencing of the mismatch repair protein, MSH2, enhances the replication of corrected cells following gene editing

BACKGROUND

Gene editing is potentially a powerful technology for introducing genetic changes by using short single-stranded DNA oligonucleotides (ssODNs). However, their efficiency is reduced by the mismatch repair system, especially MSH2, which may suppress gene editing, although findings vary depending on readout and type of oligonucleotide used. Additionally, successfully edited cells are reported to arrest at the S- or G2-phase. In the present study, we evaluate whether a novel ssODN design and down-regulation of MSH2 expression allows the isolation of replicating gene-edited cells.

METHODS

Cultured Chinese hamster ovary cells expressing mutated enhanced green fluorescent protein were targeted with ssODNs of varying design, all capable of restoring fluorescence, which allows the monitoring of correction events by flow cytometry. Converted cells were isolated by cell sorting and grown to determine colony formation efficiencies. MSH2 expression was suppressed with small interfering RNA and the cell cycle distribution of cells transfected with ssODN was quantified by flow cytometry, following propidium iodide or DRAQ5 staining.

RESULTS

Although efficiency was higher using ssODN end-protected with phosphorothioate, the potential of edited cells to form colonies was lower than those targeted with unmodified ssODN. We established that ssODN transfection itself perturbs the cell cycle and that MSH2 gene silencing increases correction efficiency. In both cases, however, the effect was dependent on the positioning of the protected nucleotides. Importantly, when internally protected ssODN was used in combination with MSH2 suppression, a higher proportion of G1-phase corrected cells was observed 48-64 h after transfection.

CONCLUSIONS

Use of internally protected ssODN and downregulating cellular MSH2 activity may facilitate isolation of viable, actively replicating gene-edited cells.

Targeted gene knock in and sequence modulation mediated by a psoralen-linked triplex-forming oligonucleotide

Information from exogenous donor DNA can be introduced into the genome via homology-directed repair (HDR) pathways. These pathways are stimulated by double strand breaks and by DNA damage such as interstrand cross-links. We have employed triple helix-forming oligonucleotides linked to psoralen (pso-TFO) to introduce a DNA interstrand cross-link at a specific site in the genome of living mammalian cells. Co-introduction of duplex DNA with target region homology resulted in precise knock in of the donor at frequencies 2-3 orders of magnitude greater than with donor alone. Knock-in was eliminated in cells deficient in ERCC1-XPF, which is involved in recombinational pathways as well as cross-link repair. Separately, single strand oligonucleotide donors (SSO) were co-introduced with the pso-TFO. These were 10-fold more active than the duplex knock-in donor. SSO efficacy was further elevated in cells deficient in ERCC1-XPF, in contrast to the duplex donor. Resected single strand ends have been implicated as critical intermediates in sequence modulation by SSO, as well as duplex donor knock in. We asked whether there would be a competition between the donor species for these ends if both were present with the pso-TFO. The frequency of duplex donor knock in was unaffected by a 100-fold molar excess of the SSO. The same result was obtained when the homing endonuclease I-SceI was used to initiate HDR at the target site. We conclude that the entry of double strand breaks into distinct HDR pathways is controlled by factors other than the nucleic acid partners in those pathways.

Simultaneous targeted exchange of two nucleotides by single-stranded oligonucleotides clusters within a region of about fourteen nucleotides

Background

Transfection of cells with gene-specific, single-stranded oligonucleotides can induce the targeted exchange of one or two nucleotides in the targeted gene. To characterize the features of the DNA-repair mechanisms involved, we examined the maximal distance for the simultaneous exchange of two nucleotides by a single-stranded oligonucleotide. The chosen experimental system was the correction of a hprt-point mutation in a hamster cell line, the generation of an additional nucleotide exchange at a variable distance from the first exchange position and the investigation of the rate of simultaneous nucleotide exchanges.

Results

The smaller the distance between the two exchange positions, the higher was the probability of a simultaneous exchange. The detected simultaneous nucleotide exchanges were found to cluster in a region of about fourteen nucleotides upstream and downstream from the first exchange position.

Conclusion

We suggest that the mechanism involved in the repair of the targeted DNA strand utilizes only a short sequence of the single-stranded oligonucleotide, which may be physically incorporated into the DNA or be used as a matrix for a repair process.

Recovery of cell cycle delay following targeted gene repair by oligonucleotides

We have previously shown that activation of the homologous recombinational repair pathway leads to a block of cell division in corrected cells, possibly through the activity of checkpoint proteins Chk1 and Chk2. In this study, we examine the long-term impact of this stalling on the growth of cells that have enabled gene repair events. Using a mutated eGFP gene as an episomal reporter, we show that corrected (eGFP-positive) cells contain only a few active replication templates 2 weeks after electroporation, yet do not display an apoptotic or senescent phenotype. By 6 weeks after electroporation, cells resume active replication with a cell cycle profile that is comparable to that of the non-corrected (eGFP-negative) population. These results indicate that the initial stalling is transient and eGFP-positive cells eventually resume a normal phenotypic growth pattern, allowing for passaging and expansion in vitro.

Manipulation of cell cycle progression can counteract the apparent loss of correction frequency following oligonucleotide-directed gene repair

Background

Single-stranded oligonucleotides (ssODN) are used routinely to direct specific base alterations within mammalian genomes that result in the restoration of a functional gene. Despite success with the technique, recent studies have revealed that following repair events, correction frequencies decrease as a function of time, possibly due to a sustained activation of damage response signals in corrected cells that lead to a selective stalling. In this study, we use thymidine to slow down the replication rate to enhance repair frequency and to maintain substantial levels of correction over time.

Results

First, we utilized thymidine to arrest cells in G1 and released the cells into S phase, at which point specific ssODNs direct the highest level of correction. Next, we devised a protocol in which cells are maintained in thymidine following the repair reaction, in which the replication is slowed in both corrected and non-corrected cells and the initial correction frequency is retained. We also present evidence that cells enter a senescence state upon prolonged treatment with thymidine but this passage can be avoided by removing thymidine at 48 hours.

Conclusion

Taken together, we believe that thymidine may be used in a therapeutic fashion to enable the maintenance of high levels of treated cells bearing repaired genes.

Reduction of gene repair by selenomethionine with the use of single-stranded oligonucleotides

Background

The repair of single base mutations in mammalian genes can be directed by single-stranded oligonucleotides in a process known as targeted gene repair. The mechanism of this reaction is currently being elucidated but likely involves a pairing step in which the oligonucleotide align in homologous register with its target sequence and a correction step in which the mutant base is replaced by endogenous repair pathways. This process is regulated by the activity of various factors and proteins that either elevate or depress the frequency at which gene repair takes place.

Results

In this report, we find that addition of selenomethionine reduces gene repair frequency in a dose-dependent fashion. A correlation between gene repair and altered cell cycle progression is observed. We also find that selenium induces expression of Ref-1 which, in turn, modifies the activity of p53 during the cell cycle.

Conclusion

We can conclude from the results that the suppression of gene repair by introduction of selenomethionine occurs through a p53-associated pathway. This result indicates that the successful application of gene repair for treatment of inherited disorders may be hampered by indirect activation of endogenous suppressor functions.

Genomic sequence correction by single-stranded DNA oligonucleotides: role of DNA synthesis and chemical modifications of the oligonucleotide ends

BACKGROUND

Single-stranded oligonucleotides (ssODN) can induce site-specific genetic alterations in selected mammalian cells, but the involved mechanisms are not known.

METHODS

We corroborate the potential of genomic sequence correction by ssODN using chromosomally integrated mutated enhanced green fluorescent protein (mEGFP) reporter genes in CHO cell lines. The role of integration site was studied in a panel of cell clones with randomly integrated reporters and in cell lines with site-specific single copy integration of the mEGFP reporter in opposite orientations. Involvement of end modification was examined on ssODN with unprotected or phosphorothioate (PS) protected ends. Also ssODN containing octyl or hexaethylene glycol (HEG) end blocking groups were tested. The significance of DNA synthesis was investigated by cell cycle analysis and by the DNA polymerases alpha, delta and epsilon inhibitor aphidicolin.

RESULTS

Correction rates of up to 5% were observed upon a single transfection of ssODN. Independent of the mEGFP chromosomal integration site and of its orientation towards the replication fork, antisense ssODN were more effective than sense ssODN. When ssODN ends were blocked by either octyl or HEG groups, correction rates were reduced. Finally, we demonstrate a dependence of the process on DNA synthesis.

CONCLUSIONS

We show that, on a chromosomal level, the orientation of the replication fork towards the targeted locus is not central in the strand bias of ssODN-based targeted sequence correction. We demonstrate the importance of accessible ssODN ends for sequence alteration. Finally, we provide evidence for the involvement of DNA synthesis in the process.

Targeted gene repair activates Chk1 and Chk2 and stalls replication in corrected cells

Oligonucleotides (ODNs) can direct the exchange of single nucleotides at specific sites in the mammalian genome. It is generally believed that the ODN aligns in homologous register with its complementary site in the target gene and provides a template for the endogenous repair machinery to alter the sequence of the gene. We have been studying the initial phase of the reaction with particular emphasis on the cellular events that occur when the oligonucleotide enters the cell. Our results show that, following introduction of the oligonucleotide, the DNA-damage response pathway is activated, evidenced by the presence of phosphorylated p53, Chk1 and Chk2, respectively. As a result, progression of some of these cells through the cell cycle is slowed and those bearing corrected genes all contain phosphorylated Chk1 and Chk2. In contrast, uncorrected cells contain much lower levels of these proteins in the activated state and pass through the cell cycle in a normal fashion. We suggest that gene repair directed by oligonucleotides activates a pathway that prevents corrected cells from proliferating in cell culture through the activation of Chk1 and Chk2. Our results impact the future use of gene repair for ex vivo gene therapy applications.

Factors affecting SFHR gene correction efficiency with single-stranded DNA fragment

A 606-nt single-stranded (ss) DNA fragment, prepared by restriction enzyme digestion of ss phagemid DNA, improves the gene correction efficiency by 12-fold as compared with a PCR fragment, which is the conventional type of fragment used in the small fragment homologous replacement method [H. Tsuchiya, H. Harashima, H. Kamiya, Increased SFHR gene correction efficiency with sense single-stranded DNA, J. Gene Med. 7 (2005) 486-493]. To reveal the characteristic features of this gene correction with the ss DNA fragment, the effects on the gene correction in CHO-K1 cells of the chain length, 5'-phosphate, adenine methylation, and transcription were studied. Moreover, the possibility that the ss DNA fragment is integrated into the target DNA was examined with a radioactively labeled ss DNA fragment. The presence of methylated adenine, but not the 5'-phosphate, enhanced the gene correction efficiency, and the optimal length of the ss DNA fragment (approximately 600 nt) was determined. Transcription of the target gene did not affect the gene correction efficiency. In addition, the target DNA recovered from the transfected CHO-K1 cells was radioactive. The results obtained in this study indicate that length and adenine methylation were important factors affecting the gene correction efficiency, and that the ss DNA fragment was integrated into the double-stranded target DNA.

Correction of the neuropathogenic human apolipoprotein E4 (APOE4) gene to APOE3 in vitro using synthetic RNA/DNA oligonucleotides (chimeraplasts)

Apolipoprotein E (apoE) is a multifunctional circulating 34-kDa protein, whose gene encodes single-nucleotide polymorphisms linked to several neurodegenerative diseases. Here, we evaluate whether synthetic RNA/DNA oligonucleotides (chimeraplasts) can convert a dysfunctional gene, APOE4 (C, A and E, T, Cys112Arg), a risk factor for Alzheimer's disease and other neurological disorders, into wild-type APOE3. In preliminary experiments, we treated recombinant Chinese hamster ovary (CHO) cells stably secreting apoE4 and lymphocytes from a patient homozygous for the epsilon 4 allele with a 68-mer apoE4-to-apoE3 chimeraplast, complexed to the cationic delivery reagent, polyethyleneimine. Genotypes were analyzed after 48 h by routine polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) and by genomic sequencing. Clear conversions of APOE4 to APOE3 were detected using either technique, although high concentrations of chimeraplast were needed (> or =800 nM). Spiking experiments of PCR reactions or CHO-K1 cells with the chimeraplast confirmed that the repair was not artifactual. However, when treated recombinant CHO cells were passaged for 10 d and then subcloned, no conversion could be detected when >90 clones were analyzed by locus-specific PCR-RFLP. We conclude that the apparent efficient repair of the APOE4 gene in CHO cells or lymphocytes 48 h post-treatment is unstable, possibly because the high levels of chimeraplast and polyethyleneimine that were needed to induce nucleotide substitution are cytotoxic.

Stable transmission of targeted gene modification using single-stranded oligonucleotides with flanking LNAs

Targeted mutagenesis directed by oligonucleotides (ONs) is a promising method for manipulating the genome in higher eukaryotes. In this study, we have compared gene editing by different ONs on two new target sequences, the eBFP and the rd1 mutant photoreceptor βPDE cDNAs, which were integrated as single copy transgenes at the same genomic site in 293T cells. Interestingly, antisense ONs were superior to sense ONs for one target only, showing that target sequence can by itself impart strand-bias in gene editing. The most efficient ONs were short 25 nt ONs with flanking locked nucleic acids (LNAs), a chemistry that had only been tested for targeted nucleotide mutagenesis in yeast, and 25 nt ONs with phosphorothioate linkages. We showed that LNA-modified ONs mediate dose-dependent target modification and analyzed the importance of LNA position and content. Importantly, when using ONs with flanking LNAs, targeted gene modification was stably transmitted during cell division, which allowed reliable cloning of modified cells, a feature essential for further applications in functional genomics and gene therapy. Finally, we showed that ONs with flanking LNAs aimed at correcting the rd1 stop mutation could promote survival of photoreceptors in retinas of rd1 mutant mice, suggesting that they are also active in vivo.

Gene therapy progress and prospects: targeted gene repair

The capacity to correct a mutant gene within the context of the chromosome holds great promise as a therapy for inherited disorders but fulfilling this promise has proven to be challenging. However, steady progress is being made and the development of gene repair as a viable and robust approach is underway. Here, we present some of the recent advances that are helping to shape our thinking about the feasibility and the limitations of this technique. For the most part, these advances center on understanding the regulation of the reaction and validating its application in animal models.

Implications of cell cycle progression on functional sequence correction by short single-stranded DNA oligonucleotides

Oligonucleotide-based sequence alteration in living cells is a substantial methodological challenge in gene therapy. Here, we demonstrate that using corrective single-stranded oligonucleotides (ssODN), high and reproducible sequence correction rates can be obtained. CHO cell lines with chromosomally integrated multiple copy EGFP reporter genes routinely show rates of 4.5% targeted sequence correction after transfection with ssODN. We demonstrate that the cell cycle influences the rates of targeted sequence correction in vivo, with a peak in the early S phase during ssODN exposure. After cell division, the altered genomic sequence is predominantly passed to one daughter cell, indicating that targeted sequence alteration occurs after the replication fork has passed over the targeted site. Although high initial correction rates can be obtained by this method, we show that a majority of the corrected cells arrest in the G2/M cell cycle phase, although 1-2% of the corrected cells form viable colonies. The G2/M arrest observed after targeted sequence correction can be partially released by caffeine, pentoxifylline or Go6976 exposure. Despite substantial remaining challenges, targeted sequence alteration based on ssODN increasingly promises to become a powerful tool for functional gene alterations.

Camptothecin enhances the frequency of oligonucleotide-directed gene repair in mammalian cells by inducing DNA damage and activating homologous recombination

Camptothecin (CPT) is an anticancer drug that promotes DNA breakage at replication forks and the formation of lesions that activate the processes of homologous recombination (HR) and nonhomologous end joining. We have taken advantage of the CPT-induced damage response by coupling it to gene repair directed by synthetic oligonucleotides, a process in which a mutant base pair is converted into a wild-type one. Here, we show that pretreating DLD-1 cells with CPT leads to a significant stimulation in the frequency of correction of an integrated mutant enhanced green fluorescent protein gene. The stimulation is dose-dependent and coincident with the formation of double-strand DNA breaks. Caffeine, but not vanillin, blocks the enhancement of gene repair suggesting that, in this system, HR is the pathway most responsible for elevating the frequency of correction. The involvement of HR is further proven by studies in which wortmannin was seen to inhibit gene repair at high concentrations but not at lower levels that are known to inhibit DNA-PK activity. Taken together, our results suggest that DNA damage induced by CPT activates a cellular response that stimulates gene repair in mammalian cells.