Efficiently Editing the Vaccinia Virus Genome by Using the CRISPR-Cas9 System

Abstract

Vaccinia virus (VACV) continues to be used in immunotherapy for the prevention of infectious diseases and treatment of cancer since its use for the eradication of smallpox. However, the current method of editing the VACV genome is not efficient. Here, we demonstrate that the CRISPR-Cas9 system can be used to edit the VACV genome rapidly and efficiently. Additionally, a set of 8,964 computationally designed unique guide RNAs (gRNAs) targeting all VACV genes will be valuable for the study of VACV gene functions.

TEXT

Since the eradication of smallpox, vaccinia virus (VACV) has been developed as a vector for vaccines against infectious diseases and immunotherapies for cancer (1–4), including oncolytic virotherapies (5–8). The renewed interest in VACV has driven a number of vaccine and therapeutic candidates to clinical trials, showing especially encouraging results for cancer treatment (5, 7). To improve VACV as a vector for vaccine or cancer therapy, a flexible system is required to delete viral genes or arm the VACV with therapeutic genes. Such a system would also expedite discoveries in cell biology, such as dissection of the signaling pathways used by VACV for its actin-based motility (9). Several strategies have been developed to construct VACV vectors (10–12); the current method for modification of VACV is based on homologous recombination in mammalian cells, but only 1% to 5% of the recombinant plaques contain the inserted DNA (13).

The clustered regularly interspaced short palindromic repeat (CRISPR)-Cas9 system is a natural microbial immune mechanism against invading viruses and other genetic elements (14–16). The CRISPR-Cas9 system, consisting of the RNA-guided Cas9 endonuclease (from Streptococcus pyogenes), a single guide RNA (sgRNA), and the trans-activating CRISPR RNA (tracrRNA), has been adapted for genome editing in eukaryotic cells (17, 18). The system has been used successfully for efficient generation of genetically modified cells and animal models (17, 19–21). Recently, the genomes of adenovirus and type I herpes simplex virus were edited using the gRNA-guided Cas9 system (22). We hypothesized that the CRISPR-Cas9 system could specifically generate double-strand breaks (DSBs) in the target DNA sites of VACV, increasing the efficiency of editing VACV genomes and constructing new VACV vectors expressing therapeutic genes.

Given that VACV replicates in the cytoplasm of infected cells, the Cas9 gene without the nuclear localization signal (NLS) was cloned into the pST1374 vector (Fig. 1A). The expression of Cas9 was confirmed (Fig. 1A). Given the important role of the N1L gene in virulence and in regulating the host immune response to VACV (23–25), we chose this as an example to validate the application of CRISPR-Cas9 for editing VACV. Two individual sgRNAs targeting the VACV N1L (LO24) gene were cloned into this guide RNA (gRNA) cloning vector (Fig. 1B) and designated gRNA N1 and N2. To prove that the CRISPR-Cas9 system could improve the efficiency of constructing a VACV vector expressing a therapeutic gene, a shuttle vector (pSh-N1LRFP/TRP2) was constructed to insert red fluorescent protein (RFP) and the tumor-associated antigen TRP2 driven by H5 promoters flanked by homologous sequences targeting both sides of the N1L region (Fig. 1C). The two gRNA plasmids targeting the N1L gene were used separately with Cas9 to generate mutant VACV in the presence of pSh-N1LRFP/TRP2 (donor vector). The success rates of N1L gene deletion in purified RFP-positive plaques (Fig. 1D) were 14.3% (1/7) and 62.5% (5/8) for gRNA N1 and gRNA N2, respectively (Fig. 1E, Table 1), whereas no deletion was detected in the 16 control plaques without prior exposure to CRISPR-Cas9 and the guide RNA (Fig. 1E). Thus, gRNA N2-guided Cas9 significantly increased (about 12 to 60 times) the homologous recombination efficiency in the N1L region. Furthermore, the transgene TRP2 was efficiently expressed in the virus-infected cells (Fig. 1F). For the detailed methods, see the Supplemental Methods.

FIG 1.

gRNA-guided Cas9 induces homologous recombination in the N1L gene. (A) The expression cassette for Cas9 was cloned into the pST1374 vector (pCas9 construct). Cas9 expression was confirmed by reverse transcriptase PCR. (B) Sequences of N1L gRNAs N1 and N2, as well as their alignment on the N1L gene. The PAM is underlined. (C) Schematic of the homologous recombination cassette of the shuttle vector (repair donor vector) and repaired target region in the vaccinia virus genome. X indicates homologous recombination. (D) Image of pure plaques of mutant VACV expressing RFP (magnification, ×400). (E) N1L gene deletion was verified by PCR in pure plaques of mutant VACV obtained from N1L gRNA N2-guided, Cas9-induced homologous recombination. Amplification of the N1L and L026 genes (Target) was carried out by PCR. A52R gene amplification was used as a DNA control. (F) Transgene Trp2 expression in CV1 cells after infection with N1L-deleted VACV was detected by Western blot assay. C, control virus-infected cells; M, mutant virus-infected cells. Actin was used as a loading control.

TABLE 1.

Efficiency of gRNA-guided Cas9-induced homologous recombination

Target site	Sequence (5′–3′)a	% efficiency (no. of plaques with deletion/total no. of plaques)
N1L gRNA N1	CGTCTTCTATATTATTAACTGG	14.3 (1/7)
N1L gRNA N2	GATTCAATCTATCTAGCAATGG	62.5 (5/8)
A46R gRNA A1	TTCCAGAGGAAACTAATATTGG	80 (12/15)
A46R gRNA A2	ACAAGTAAGTCATACTAACCGG	80 (12/15)
A46R gRNA A3	ATGACACCATAGAGATAAGAGG	85 (6/7)

^aUnderlining shows the protospacer adjacent motif (PAM) sequence located on the immediate 3′ end of the gRNA recognition sequence.

To test the versatility of gRNA-guided Cas9 in modifying VACV genomes, three gRNAs targeting the VACV A46R gene (Fig. 2A) were cloned into the gRNA cloning vector, as A46R is another important gene that regulates the host immune response and, therefore, the deletion of A46R may further improve the efficacy of VACV therapeutics (26, 27) (see the Supplemental Methods). The three gRNA plasmids were transfected separately with Cas9 to generate mutant VACV in the presence of the A46R homologous recombination shuttle vector (donor vector), using RFP as the plaque purification marker (Fig. 2B). The success rates of A46R gene deletion in RFP-positive plaques were 80% (12/15), 80% (12/15), and 85% (6/7) for gRNA A1, gRNA A2, and gRNA A3, respectively (Fig. 2C; Table 1), whereas there was no deletion detected in the 28 control plaques without prior exposure to CRISPR-Cas9 and the guide RNA (data not shown).

FIG 2.

gRNA-guided, Cas9-induced homologous recombination in the A46R gene. (A) Cloning cassette for gRNA, sequences of A46R gRNAs A1 to A3, and their alignments on the A46R gene. PAM is underlined. (B) Schematics of homologous recombination cassette of shuttle vector (repair donor vector) and repaired target region on VACV genome. X indicates homologous recombination. (C) A46R gene deletion in pure plaques of mutant VACV obtained by using A46R gRNA was verified by PCR. A1-, A2- and A3-guided Cas9, respectively, induced homologous recombination. Amplification of the A46R and A47L genes (Target) was carried out by PCR. A52R gene amplification was used as a DNA control.

In certain cases, there is a need to modify two or more genes in VACV to generate a mutant virus. Using the current method, more than 4 weeks are required to achieve modification of two genes in the virus. To test whether multiple genes of VACV can be edited simultaneously by gRNA-guided Cas9, the N1L region and the A46R region were targeted. N1L gRNA N2 and A46R gRNA A2 were used to guide Cas9 editing of the VACV genomes (Table 1). The shuttle vector carrying green fluorescent protein (GFP) for the N1L region (Fig. 3A) and donor vector carrying RFP for the A46R region were used for homologous recombination (Fig. 3A). Plaques that were double positive for both GFP and RFP were purified over 3 to 5 rounds (Fig. 3B). N1L and A46R dual deletions were found in 9 of 15 double-positive plaques (60%) (Fig. 3C). This result demonstrates that gRNA-guided Cas9 can simultaneously and rapidly induce homologous recombination across multiple target sites on the VACV genome. For the detailed methods, see the Supplemental Methods.

FIG 3.

Modification of two VACV genes simultaneously using gRNA-guided Cas9. (A) N1L gRNA N2 and A46R gRNA A2 were cotransfected with pCas9 into CV1 cells. A shuttle vector (repair vector) carrying a GFP selection marker was used for N1L region homologous recombination (HR), and an RFP selection marker was used for A46R region homologous recombination. Target regions and repaired regions in the targeted VACV genome region are shown. X indicates homologous recombination. (B) Top, images of cells after N1L gRNA N2 and A46R gRNA A2 were cotransfected with pCas9, followed by VACV infection and transfection of repair vectors with GFP and RFP selection markers. Bottom, images of pure plaques expressing both GFP and RFP markers (magnification, ×200). (C) Deletion of the N1L gene and the A46R gene were verified by PCR in pure plaques of mutant VACV expressing both GFP and RFP. Amplification of the targeted N1L or A46R region was carried out by PCR. A52R gene amplification was used as a DNA control. *, results show double deletion of the N1L gene and the A46R gene in pure GFP-and-RFP dual-positive plaques.

Finally, using a bioinformatics approach for computing all VACV genome targets, we generated a set of 8,964 gene sequences so that gRNA can maximally target specific genes in VACV but minimally target other locations in the genome (see Tables S1 and S2 in the supplemental material). Maximally efficient targeting of Lister clone VACV107 by a gRNA was achieved by identifying 19-nucleotide sequences followed by the nucleotides NGG (protospacer adjacent motif [PAM]), where N can be A, G, C, or T. Identical motif sequences present in multiple genes/locations (usually at either end of the genome) were further excluded, and thus, only unique motif sequences were retained in our final data set (see Table S1 in the supplemental material). The same search and exclusion criteria were also applied to find viral gene sequences for designing gRNA in the reverse complementary strand of coding regions of VACV107. Unique motif sequences are shown in our final data set (see Table S2). Of note, the two N1L gRNAs and all three A46R gRNAs that were functionally validated in this study are covered by the data set (see Tables S1 and S2).

To generate a new mutant of VACV, the current approach can take up to 10 rounds of plaque purification (lasting 4 to 6 weeks) and often results in an unsuccessful outcome in obtaining the desired virus with a purification marker in the target region. In this study, we demonstrate that the gRNA-guided Cas9 system greatly increases the efficiency of generating mutant VACVs without evident off-target effects on the VACV genome. This system can also efficiently engineer two viral genes simultaneously. This approach is likely to have a significant impact by expanding the application of VACV in biomedical research and, potentially, in clinical medicine.