Improving Accurate Genome Editing Using a Hybrid of Donor DNA and Guide RNA

Escherichia coli strain DH5α (F⁻, Φ80dlacZΔM15, Δ(lacZYA-argF)U169, deoR, recA1, endA1, hsdR17 (r_k⁻, m_k⁺), phoA, supE44, λ⁻, thi-1, gyrA96, relA1) (Toyobo, Osaka, Japan) was used as the host for recombinant DNA manipulation. Transformed DH5α cells were grown in LB + A medium (1% (w/v) tryptone, 0.5% (w/v) yeast extract, and 1% (w/v) sodium chloride) containing 100 μg/mL ampicillin. S. cerevisiae BY4741 (MATa, his3Δ1, leu2Δ0, met15Δ0, ura3Δ0) was used as the host strain for genome editing. Yeast cells were cultured in the following media: YPD medium (1% (w/v) yeast extract, 2% (w/v) peptone, and 2% (w/v) glucose); SDC medium (0.67% (w/v) yeast nitrogen base without amino acids, 2% (w/v) glucose, and 0.5% (w/v) casamino acids); SD + HLM medium (0.67% (w/v) yeast nitrogen base without amino acids, 2% (w/v) glucose, 0.002% (w/v) histidine, 0.003% (w/v) leucine, and 0.003% (w/v) methionine); SD + HLM with canavanine medium (0.67% (w/v) yeast nitrogen base without amino acids, 2% (w/v) glucose, 0.002% (w/v) histidine, 0.003% (w/v) leucine, 0.003% (w/v) methionine, and 0.001% (w/v) l-canavanine sulfate (Nacalai Tesque, Kyoto, Japan)).

All primers used for plasmid construction are listed in Table S1. All DNA fragments were amplified by polymerase chain reaction (PCR) using KOD-One DNA polymerase (Toyobo). Fusion of the Cas9 gene from Streptococcus pyogenes and the nuclear localization signal of the SV40 large T antigen was amplified from p414-TEF1p-Cas9-CYC1t [10] using primers (Cas9-F and Cas9-R). Amplified Cas9_NLS was inserted into p416 GPD [23], which was linearized by amplification using primers (p416GPD-F and p416GPD-R). Insertion was performed using an In-Fusion HD Cloning kit (Takara Bio, Otsu, Japan). The resulting plasmid was named p416GPD_Cas9_NLS. The plasmid sequences were confirmed using Sanger sequencing.

The plasmid p416GPD_Cas9_NLS was introduced into yeast using the Frozen-EZ Transformation Kit II (Zymo Research, CA, USA). The transformants were selected on SDC solid medium.

gRNA and other extended gRNAs, which are necessary for DGybrid preparation, were synthesized using the CUGA^®7 gRNA Synthesis Kit (NIPPON GENE, Tokyo, Japan). DNA templates with T7 promoter and gRNA sequences were prepared from three oligonucleotide primers by overlap extension PCR using KOD-One DNA polymerase and then purified using the QIAEX Gel Extraction Kit (Qiagen, Hilden, Germany). All primers used for preparing gRNA template are listed in Table S1. Furthermore, gRNA synthesis was performed according to the manufacturer’s protocol. The synthesized gRNAs were purified using a spin column provided with the CUGA^®7 gRNA Synthesis Kit.

Extended gRNA was synthesized as described previously. The gRNA and ssODN sequences are listed in. Equimolar amounts of each complementary extended gRNA and ssODNs were mixed in TE buffer with salt (10 mM Tris-HCl (pH 8.0), 1 mM EDTA, and 10 mM NaCl). DGybrid was formed by incubating the mixture at 95 °C for 5 min and then cooling slowly to 25 °C. Table S2

DGybrid formation was confirmed by size fractionation using 10% native-PAGE. For evaluation of formation, DGybrid was formed from 150 pmol each of extended gRNA and ssODNs by the method described above. The samples were mixed with equal volumes of native loading buffer (30% (v/v) glycerol, 80 mM HEPES-KOH (pH 7.9), 100 mM KCl, and 2 mM magnesium acetate) and electrophoresed in 1× TBE buffer at 100 V and 25 °C for 80 min in a vertical gel tank. Dilutions of each gRNA and ssODN in RNase-free water were also performed to calculate the efficiency of hybrid formation. Moreover, 1 kb Plus DNA Ladder (New England BioLabs, Ipswich, MA, USA) was loaded as a reference for nucleic acid size. The DGybrid and its counterparts were stained with ethidium bromide and visualized using a UV transilluminator. Band intensities were quantified with ImageJ software (ver.1.53e; National Institute of Health, MS, USA).

For genome editing, DGybrid was transformed into BY4741 cells harboring the p416GPD_Cas9_NLS plasmid by electroporation. The DGybrid used for genome editing was formed from 200 pmol of each of the extended gRNA and ssODNs by the method described above. First, BY4741 cells harboring the p416GPD_Cas9_NLS plasmid were grown overnight in 10 mL SDC liquid medium. Cells that reached the stationary phase were inoculated into 100 mL of YPD liquid medium to obtain a starting OD₆₀₀ of 0.2 and grown to an OD₆₀₀ of 1.4~1.7. The cells were collected by centrifugation at 900× g for 5 min. The collected cells were washed twice with sterilized water and once with electroporation buffer (1 M sorbitol and 1 mM CaCl₂). The cells were then resuspended in 100 mM LiAc and 10 mM DTT solution and incubated at 30 °C for 30 min with shaking at 230 rpm. After washing again with electroporation buffer, the cells were resuspended in electroporation buffer to a final volume of 1.0 mL. Cell suspensions (300 µL) were mixed with 200 pmol/30 µL of DGybrid or other nucleic acid solutions and electroporated once at 2500 V in a 0.2 cm cuvette. The electroporated cells were quickly transferred to 4 mL of a 1:1 mixture of 1 M sorbitol and YPD medium. The cells were incubated at 30 °C for 2 h with shaking at 230 rpm.

After incubation of the electroporated cells for 2 h, the cells were inoculated into 10 mL of SDC liquid medium to obtain an OD₆₀₀ of 0.1 and grown for 24 h. Then, the cells were collected and diluted to an OD₆₀₀ of 0.001. Diluted cultures (100 µL) were spread onto SD + HLM solid medium or SD + HLM with canavanine solid medium. After incubation at 30 °C for 72 h, images of each solid medium were captured, and the number of colonies was counted using ImageJ software (ver.1.53e). Genome-editing efficiency was evaluated by the rate of appearance of canavanine-resistant colonies relative to the total number of cells identified in the canavanine-free medium.

To evaluate the efficiency of electroporation-based introduction of nucleic acids into yeast cells, fluorescently labeled nucleic acids () were used. Fluorescently labeled DNA/RNA hybrids were formed from 60 pmol each of RNA_60 and ssODN_60 using the method described above, whose 3′ end was labeled with Alexa Fluor 647 (AF647) and fluorescein isothiocyanate (FITC), respectively. ssODN_20, which was also labeled with FITC at the 3′ end, lacks the sequence for annealing to RNA_60 and was used as a control. AF647-labeled RNA was purchased from Ajinomoto Bio-Pharma Services (Tokyo, Japan). FITC-labeled DNA was purchased from Eurofins Genomics (Tokyo, Japan). Table S2

A fluorescently labeled DNA/RNA hybrid was introduced into BY4741 cells harboring the p416GPD_Cas9_NLS plasmid by electroporation. For electroporation, nucleic acids (20 pmol/3 µL) were introduced into 30 µL of the cell suspension. Electroporated cells were incubated in a 1:1 mixture of 1 M sorbitol and YPD medium at 30 °C for 10 min with shaking at 230 rpm and then collected at 3500× g for 2 min. The cells were washed twice with PBS (pH 7.4) and resuspended in PBS. The fluorescence intensity of the cells was measured using flow cytometry (JSAN; Bay Bioscience, Kobe, Japan), and the introduction efficiency was evaluated based on the percentage of fluorescent cells. The fluorescence of AF647 was detected with excitation at 640 nm and emission at 661 ± 10 nm (FL5), and FITC was detected with excitation at 488 nm and emission at 535 ± 23 nm (FL1). The fluorescence intensities of 20,000 yeast cells were plotted as a density plot. Data were analyzed using Kaluza software (ver. 2.1; Beckman Coulter, Brea, CA, USA). In the density plot, the ratios of yeast cells that richly took up the nucleic acids were quantified.

To achieve higher HDR efficiency in CRISPR-based genome editing, we attempted to develop a novel tool that combines donor ssODNs and gRNA through base pairing between complementary DNA and RNA sequences. For this purpose, we designed a unique gRNA (5′-40b-gRNA), in which the 5′ end of a conventional gRNA is extended by 40 bases complementary to the 5′ end of the donor ssODN (Figure 1). Compared to the 3′ end, the 5′ end of gRNA is known to be more tolerant to modification [20,24]. Furthermore, ssODNs (121 bases) were used as donor DNA. When 5′-40b-gRNA and ssODNs are mixed in equal molar amounts, their complementary 40 nucleotide sequences form base pairs, producing a donor DNA/gRNA hybrid duplex (DGybrid) that has both gRNA and donor DNA functions.

We performed native-PAGE analysis to examine DGybrid formation. The DGybrid samples were prepared in four different conditions: (1) TE buffer, (2) TE buffer + 10 mM NaCl, (3) TE buffer + 100 mM NaCl, and (4) ethanol precipitation of the sample prepared in condition (3). In the four lanes to which the formed DGybrids were applied, the bands corresponding to 5′-40b-gRNA and ssODNs almost completely disappeared. Instead, a new intense band was observed on the high-molecular-weight side at the top of the gel (Figure 2), indicating the formation of DGybrid. In addition, based on the band intensities of the diluted solutions of 5′-40b-gRNA and ssODNs, the amount of both counterparts remaining in the DGybrid sample prepared under condition (2) is less than 20%. This indicates that more than 80% of the 5′-40b-gRNA and ssODN contributed to the hybrid formation in the DGybrid sample prepared under condition (2).

We evaluated whether DGybrid can achieve HDR-mediated editing with higher efficiency than conventional methods. DGybrid was introduced into S. cerevisiae constitutively expressing Cas9 from the p416GPD_Cas9_NLS plasmid, and the editing efficiency was evaluated based on the phenotype (canavanine resistance). The DGybrid was designed to target the CAN1 gene as a function of gRNA and to cause nonsense mutations in the PAM sequence as a function of donor DNA. When a stop codon is placed in the CAN1 gene by HDR-mediated editing, the function of Can1 (arginine transporter) is abolished, and yeast cells acquire improved canavanine resistance [25,26].

Cells introduced with DGybrid were inoculated on an agar medium containing canavanine, and resistant colonies were obtained (Figure 3). The results show that the genome-editing efficiency of the sample using DGybrid (62%) was significantly higher than that of the conventional method (35%), in which gRNA and donor ssODNs worked separately (Figure 4). Sanger sequencing of the CAN1 gene from the obtained canavanine-resistant colonies showed that the stop codon was inserted precisely at the target site (Figure S1). These results indicate that the developed DGybrid can achieve precise HDR-mediated editing more efficiently than conventional methods.

The CRISPR/Cas9 system, which induces DSBs, is potentially cytotoxic. To investigate the effect of DGybrid on cytotoxicity, we monitored the growth of yeast cells introduced with DGybrid by measuring OD₆₀₀. The growth of DGybrid-introduced cells was not significantly different from that of cells in which donor ssODN and gRNA were introduced separately (Figure S2). Therefore, it is suggested that the DGybrid would not affect CRISPR toxicity.

Potentially, the length of the annealing sequence between gRNA and donor ssODNs may affect the stability of the DGybrid and HDR-mediated editing efficiency in yeast cells. Thus, the effect of DGybrid design on HDR-mediated editing efficiency was examined by extending the length of the annealing sequence to 70 bases (). In the first case (DGybrid 70-A), the annealing sequence was extended to 70 bases, but the homologous sequence was correspondingly reduced to 51 bases. In the second case (DGybrid 70-B), the annealing sequence was similarly extended to 70 bases, and 81 bases of the homologous sequence were used, as in the original design. These modified DGybrid showed editing efficiencies comparable to those of the initial DGybrid (). These results suggest that, regardless of the length of the annealing sequence, the combination of donor ssODNs and gRNA into a single unit is critical for improving editing efficiency. Figure S3A Figure S3B

Our DGybrid successfully improved the HDR-mediated editing efficiency in yeast cells. We assumed that the improved HDR-mediated editing efficiency could be due to the following two advantageous properties of the DGybrid: First, the donor DNA can be localized near the cleavage site using DGybrid. The donor ssODN can be recruited to the cleavage site by combining it with the gRNA that recognizes the target sequence as reported in some previous studies [14,15,16,17,18,19,20]. Thus, ssODNs are quickly available before the cleavage site is subjected to repair mechanisms such as NHEJ or degradation other than HDR, leading to a positive effect on editing efficiency. Second, DGybrid has the advantage of introduction when incorporating gRNA and donor DNA into a cell. When two oligonucleotides are introduced into a cell separately using the conventional method, the efficiency of the cells in taking up both oligonucleotides is low. Using DGybrid, both gRNA and donor ssODNs can be delivered together into the cell. Thus, the cell obtains both donor ssODNs and gRNA with only one molecule uptake and more HDR-mediated editing opportunities. This can also lead to improvements in editing efficiency.

To assess the contribution of the second advantage, we analyzed the introduction efficiency of the DNA/RNA hybrid into yeast cells. To detect nucleic acids introduced into cells, the 3′ ends of RNA_60 (60 bases) and ssODN_60 (60 bases) were labeled with Alexa Fluor 647 and FITC, respectively (Figure 5A). They were annealed via the same 40-base complementary sequence as DGybrid described in Figure 1, and their successful formation was confirmed by native-PAGE analysis (Figure S4). Flow cytometry analysis of the cells introduced with nucleic acids showed that cells to which the formed DNA/RNA hybrids were introduced were distributed in an approximate linear pattern on the density plot, indicating that the cells took up RNA and ssODN together (Figure 5B). The cell population localized in the upper right of the plot, which took up very large amounts of both RNA and ssODN, was assumed to have more HDR-mediated editing opportunities, and the percentages were compared. In samples to which the DNA/RNA hybrid formed from RNA_60 and ssODN_60 were introduced, 14.4% of the cell populations were rich in both RNA and ssODN. Such a cell population was virtually absent (0.2%) when RNA_60 and ssODN_20 (20 bases), which do not form DNA/RNA hybrids, were introduced together. These results suggest that use of DGybrid also increases the number of cells rich in both gRNA and donor ssODNs, which would contribute to the increased editing efficiency via HDR.