DNA replicons for plant genome engineering

Abstract

Sequence-specific nucleases enable facile editing of higher eukaryotic genomic DNA; however, targeted modification of plant genomes remains challenging due to ineffective methods for delivering reagents for genome engineering to plant cells. Here, we use geminivirus-based replicons for transient expression of sequence-specific nucleases (zinc-finger nucleases, transcription activator-like effector nucleases, and the clustered, regularly interspaced, short palindromic repeat/Cas system) and delivery of DNA repair templates. In tobacco (Nicotiana tabacum), replicons based on the bean yellow dwarf virus enhanced gene targeting frequencies one to two orders of magnitude over conventional Agrobacterium tumefaciens T-DNA. In addition to the nuclease-mediated DNA double-strand breaks, gene targeting was promoted by replication of the repair template and pleiotropic activity of the geminivirus replication initiator proteins. We demonstrate the feasibility of using geminivirus replicons to generate plants with a desired DNA sequence modification. By adopting a general plant transformation method, plantlets with a desired DNA change were regenerated in <6 weeks. These results, in addition to the large host range of geminiviruses, advocate the use of replicons for plant genome engineering.

Figure 1.

Development of GVRs for Protein Expression in Tobacco. (A) Approach to deliver GVRs to plant cells. Rep is delivered using a separate T-DNA expression plasmid (pREP; data not shown). LB, left T-DNA border; RB, right T-DNA border. The red rectangle indicates the region where heterologous sequence is cloned. P35S, 2x35S promoter from the cauliflower mosaic virus. Splice donor and acceptor sequences flank the LIR within circularized replicons (data not shown). Blue and red arrows indicate primer binding sites for PCR to detect circularized GVRs. (B) Images of leaf tissue expressing GFP and GUS. As a negative control for GVR-mediated protein expression, leaf tissue was infiltrated with a single strain of Agrobacterium containing the LSL T-DNA plasmid. Tissue delivered pLSLGFP was imaged 3 d after infiltration. Tissue-delivered pLSLGUS was stained in a solution with X-Gluc 7 d after infiltration. (C) PCR-based detection of circularized GVRs within plant cells. The PCR control (LSL T-DNA) used primers designed to amplify sequence within linear LSL T-DNA.

Figure 2.

GVR-Mediated Expression of ZFNs for Targeted Mutagenesis. (A) LSL T-DNA encoding Zif268:FokI (top). PCR-based detection of circularized GVRs within plant cells (bottom). (B) PCR-based detection of ZFN-induced mutations at the Zif268 target sequence. Numbers beneath the gel image indicate the percentage of cleavage-resistant amplicons. NTC, nontransformed control. (C) Sequences of individual amplicons containing NHEJ-induced mutations. Letters in front of gray background indicate the Zif268 binding sequence. (D) Quantification of NHEJ-induced mutations. Error bars represent se of three experiments. *P < 0.05.

Figure 3.

GVRs Promote High-Frequency Gene Targeting. (A) Approach to repair a nonfunctional gus:nptII gene through homologous recombination. GT, gene targeting; TS, Zif268:FokI target site. (B) Images of leaf tissue stained in a solution with X-Gluc. To better visualize stained cells, chlorophyll was removed from leaf tissue, and the green and blue channels were removed from the image. The Neg. Ctrl. image is of nontransgenic leaf tissue transformed with two strains of Agrobacterium containing pLSLZ.D and pREP. (C) Relative frequencies of gene targeting using GVRs (pLSLZ.D + pREP; blue bars) and conventional T-DNA (p35SZ.D; red bars). The x axis represents four different plant lines with the gus:nptII gene integrated at different chromosomal positions. Error bars represent se of at least three biological replicates. *P < 0.05 and **P < 0.005.

Figure 4.

Synergism between Double-Strand Breaks, Replication of Repair Templates, and Pleiotropic Activity of Rep and RepA on Gene Targeting. (A) Illustration of the approach to directly compare gene targeting frequencies between two Agrobacterium samples. (B) Effect of targeted double-strand breaks on gene targeting. As indicated, red arrows represent coding sequence for Zif268:FokI; black boxes depict the us:NPTII repair templates (RT); small black boxes represent the left and right T-DNA borders. Error bars represent se of six experiments. ***P < 0.001. Data were normalized to sample 1. (C) Effect of replicating ZFN and repair template sequence on gene targeting (left graph). Error bars represent se of four experiments. ***P < 0.001. Effect of replicating repair template sequence on gene targeting (right graph). Error bars represent se of three experiments. **P < 0.005. Data were normalized to sample 1 in both graphs. (D) Effect of Rep and RepA expression on gene targeting. Measure of center is mean. Error bars represent se of four experiments. *P < 0.05. Data were normalized to sample 1.

Figure 5.

Regeneration of Cells with the Repaired GUS:NPTII Transgene. (A) Image of a leaf disc stained in X-Gluc 0 d after plating. (B) Image of a leaf disc stained in X-Gluc 7 d after plating. (C) Image of leaf discs stained in X-Gluc 14 d after plating. (D) Image of a leaf disc stained in X-Gluc 21 d after plating. (E) Shoots with GUS activity 42 d after plating. (F) Shoots with GUS activity 49 d after plating. (G) PCR-based detection of the repaired GUS:NPTII transgene. Total DNA was extracted from four calli that stained blue. As a negative control, genomic DNA was extracted from callus that did not stain blue (Callus #5). Primers were designed to be homologous to sequence within the 600-bp modification sequence and ∼1 kb downstream of the homologous DNA region carried by the repair template. The genomic DNA control used primers designed to amplify sequence within the endogenous F-box gene. (H) DNA sequences of amplicons generated from calli #1 to #4.

Figure 6.

Single-Component Geminivirus Vectors Enable Efficient Genome Editing. (A) Illustration of single-component GVR vectors. The blue box represents SIR sequence. (B) Selected images of leaves transformed with pLSLZ.D and pREP, pLSLZ.D.R, pZLSLD.R, or p35SZ.D. Green and blue channels were removed from the images to better visualize GUS staining. (C) Quantification of GUS activity in leaf tissue delivered GVR constructs. The left graph represents data from three leaves transformed with pLSLZ.D and pREP on one side and pLSLZ.D.R on the other side. The right graph represents data from two leaves transformed with pLSLZ.D and pREP on one side and pZLSLD.R on the other. Error bars represent se. P values for both graphs were >0.05.