New Tools for Engineering the Arabidopsis Plastid Genome^,

Abstract

New transformation-competent Arabidopsis lines, with new plastid transformation vectors and a protocol for measuring plastid transformation efficiency, will advance the engineering of the plastid genome in Arabidopsis.

Dear Editor,

Arabidopsis (Arabidopsis thaliana) is the best-characterized model plant and is used to study all aspects of basic science. A notable exception is that studies involving plastid genome engineering are carried out in tobacco (Nicotiana tabacum), the only vascular plant species in which plastome engineering is routine (Bock, 2015). Plastid transformation in Arabidopsis was reported in 1998, but only one transplastomic event was obtained per 100 bombarded samples, an efficiency 100 times lower than that in tobacco (Sikdar et al., 1998). A clue to why Arabidopsis plastid transformation was inefficient came years later from a study on nuclear genes essential for survival in the absence of chloroplast translation. Parker et al. (2014) determined that most Arabidopsis ecotypes are tolerant to spectinomycin, the selective agent used in plastid transformation, due to a duplication of the ACCase enzyme. Spectinomycin is an inhibitor of plastid translation that curtails callus proliferation and greening in all species for which it has been used to successfully recover transplastomic events (Maliga, 2012). In most Arabidopsis ecotypes, spectinomycin does not impair callus proliferation unless the ecotype carries a mutation in the nuclear ACC2 gene. When the nuclear ACC2 gene is not functional, the cells depend on the heteromeric ACCase enzyme for fatty acid biosynthesis, one subunit of which is encoded on the plastid genome (Parker et al., 2014).

We hypothesized that Arabidopsis plastid transformation inefficiency was due to tolerance of the RLD ecotype to spectinomycin. The finding that plastid transformation in the spectinomycin-hypersensitive Col-0 acc2-1 and Sav-0 backgrounds proved 100-fold more efficient (Yu et al., 2017) corroborated this notion. However, the transplastomic plants regenerated from leaves failed to produce viable seed. Ruf et al. (2019) confirmed that it is essential to carry out plastid transformation in an ACC2 knockout background in the C24 ecotype and showed that switching to root explants instead of leaves as the bombardment material in that ecotype yields fertile plants. We report here the creation of spectinomycin-hypersensitive lines in the Arabidopsis ecotypes RLD and Wassilewskija (Ws). We also extend the previous observations by showing that it is essential to use the ACC2 knockout background to transform the RLD and Ws ecotypes, indicating that the need to use a knockout recipient is probably true of all Arabidopsis ecotypes. We also report new Arabidopsis plastid transformation vectors that are suitable for the insertion of transgenes. In addition, we document significant differences in plastid transformation rates among different Arabidopsis ecotypes using a leaf system that enables rapid scoring of transplastomic callus events.

To enable plastid transformation in additional Arabidopsis ecotypes, we chose RLD and Ws because efficient plant regeneration protocols are available for these ecotypes (Márton and Browse, 1991; Chupeau et al., 2013; Zhao et al., 2014; Lutz et al., 2015). We designed a 20-bp single guide RNA (sgRNA) to target to the ACC2 N-terminal chloroplast transit peptide, thereby avoiding mutagenesis of the cytoplasmic ACC1 coding region (Fig. 1A; Supplemental Fig. S1). The same sgRNA was used to edit ACC2 in the conserved region of RLD and Ws ecotypes (Fig. 1B). Putative mutants were identified by the T7 Endonuclease I assay that detects imperfectly matched DNA, and mutations were verified by direct sequencing of PCR products (Fig. 1C). The most readily identifiable mutations in ACC2 are the 17- and 10-nucleotide deletions in the RLD acc2-3 and Ws acc2-3 lines, respectively. The nucleotide insertions and deletions created a stop codon close to the target site, which resulted in early translation termination in the ACC2 reading frame (Supplemental Fig. S1). Spectinomycin hypersensitivity of homozygous knockout lines was confirmed by germinating the seedlings on a selective spectinomycin medium, where wild-type seedlings develop leaves and the knockout lines fail to develop any structure at the shoot apex (Fig. 1D; Parker et al., 2014). Under standard growth conditions, the knockout plants are fully fertile and produce seed.

Figure 1.

Generation of ACC2 knockout lines using the CRISPR/Cas9 system. A, Map of the Arabidopsis ACC2 gene. The region targeted by sgRNA is boxed. A single-nucleotide polymorphism in Col-0 is highlighted in green. The protospacer-adjacent motif is underlined. B, Schematic diagram of the sgRNA and Cas9 genes (Mao et al., 2013). The first nucleotide of the sgRNA was changed from C to G (in red) to facilitate transcription from the U6 promoter by RNA polymerase III. C, Targeted mutations induced by CRISPR/Cas9 in the ACC2 gene. We show sequences in the targeted region aligned by the MultAlin software (Corpet, 1988). D, ACC2 mutations make RLD and Ws hypersensitive to spectinomycin (100 mg L⁻¹). Note the arrested growth at the cotyledon stage and lack of shoot apex in the acc2 lines and the outgrowth of true leaves and development of shoot in wild-type (WT) plants.

Testing of plastid transformation efficiency was carried out with the pATV1 vector used in our earlier study (Yu et al., 2017) and the newly constructed pMEK14 vector (Fig. 2B). Vector pMEK14 is based on the pMEK18/pMEK19 backbone that carries a multiple cloning site to facilitate the insertion of passenger genes (Fig. 2A). The left targeting region of the pMEK vectors is truncated at the HindIII site in the plastid rrn16 gene to enable utilization of the EcoRI-HindIII multiple cloning site. The multiple cloning site is adopted from the pUC18/pUC19 cloning vectors, which are routinely used for gene assembly in chloroplast engineering laboratories (Maliga, 2002). The pATV1 and pMEK14 vectors carry the same dicistronic aadA-gfp operon and target insertions between the plastid trnV gene and the 3′-rps12/rps7 operon in the inverted repeat region of the plastid genome (Fig. 2A). The first open reading frame of the operon encodes aminoglycoside-3′′-adenylyltransferase, which confers spectinomycin resistance to transformed cells. The second open reading frame encodes the GFP used for visual identification of transplastomic clones under UV light and confocal microscopy (Yu et al., 2017). Truncation of the left targeting region in the pMEK vectors does not appear to have an impact on plastid transformation efficiency (Table 1).

Figure 2.

Plastid transformation in Arabidopsis. A, The plastid-targeting region of the pMEK18 (GenBank accession no. MN326017) and pMEK19 (MN326018) Arabidopsis plastid transformation vectors. The map positions of plastid rrn16 and trnV genes and the promoter region of the 3′-rps12/7 operon are shown along with the relevant restriction enzyme sites. B, Arabidopsis plastid genome transformed with the pATV1 (MF461355) and pMEK14 vectors. The vectors carry an identical dicistronic operon with aadA and gfp genes. The positions of the PrrnLatpB promoter (P), TpsbA terminator (T), and ribosome-binding site (semiovals) are indicated. The black box at the aadA N terminus marks the atpB downstream box sequence (Kuroda and Maliga, 2001). Vector pMEK14 has a shorter left flanking region relative to pATV1 (1,055 versus 1,665 bp; targeting regions are shown as dashed lines). The thick black line indicates the probe used for the DNA gel-blot analyses. C, Col-0 acc2 transplastomic event (white arrows; fluoresces under UV light) and spontaneous spectinomycin-resistant mutant or aadA nuclear insertion line (black arrows; green in normal light) 1 month after leaf bombardment. D, DNA gel blot using the rrn16 probe (see B) confirms plastid transformation in GFP-expressing Col-0 acc2 clones. Plastid DNA transformed with vector pATV1 and pMEK14 yields 4.7- and 1.8-kb EcoRI fragments, respectively. The 2.7-kb fragment is present in the wild-type (WT) and heteroplastomic samples. E, Confocal microscopy to detect GFP accumulation in Col-0 acc2 chloroplasts. Arrows point to a stromule connecting two chloroplasts. Merged 1 refers to the overlap of GFP and chlorophyll channels; Merged 2 refers to the overlap of GFP, chlorophyll, and bright fields. Bars = 10 μm.

Table 1. The frequency of transplastomic events (TP) in Arabidopsis leaf culture.

Accession	Vector	No. of Plates	No. of TP	No. of TP per Plate
Col-0	pATV1	6	0	0.0
	pMEK14	2	0	0.0
RLD	pATV1	6	0	0.0
Ws	pATV1	7	0	0.0
	pMEK14	2	1	0.5
Col-0 acc2	pATV1	9	76	8.4
	pMEK14	9	69	7.7
RLD acc2	pATV1	4	9	2.2
	pMEK14	2	6	3.0
Ws acc2	pATV1	4	11	2.8
	pMEK14	2	1	0.5
Sav-0	pATV1	10	4	0.4
	pMEK14	12	2	0.2

Plastid transformation efficiency was evaluated by bombarding leaves and culturing them on the one-step selective SPEED medium containing 100 mg L⁻¹ spectinomycin for evaluation of transformation efficiency. Spectinomycin at this concentration severely inhibits callus proliferation in the ACC2-knockout leaves (Supplemental Fig. S2). Spectinomycin-resistant cells divide and form pale-green and dark-green clusters, but no shoots, on the SPEED callus medium (Supplemental Methods S1). Transplastomic cells are readily identified by GFP fluorescence 5 to 6 weeks after bombardment (Fig. 2C). Plastid transformation in these clones was confirmed by DNA gel-blot analyses (Fig. 2D) and confocal microscopy (Fig. 2E). We note that calli numbers 25, 27, 33, and 40 obtained with vector pATV1 and number 86 obtained with vector pMEK14 were homoplastomic after one round of selection. This contrasts with the situation in tobacco, where multiple rounds of selection are typically required to obtain homoplasmy. Spontaneous mutations in the plastid 16S rRNA genes or nuclear insertions of aadA may also confer resistance to spectinomycin (Svab and Maliga, 1993). These cell clusters appear green due to chlorophyll accumulation but do not fluoresce under UV light (Fig. 2C).

A comparison of plastid transformation efficiency in the wild-type and knockout backgrounds for Col-0, RLD, and Ws confirmed the necessity for an ACC2-defective background for the recovery of transplastomic events (Table 1). Only one transplastomic event was recovered in 23 samples with wild-type backgrounds, but many were obtained for all knockout backgrounds. Herein, transformation efficiencies varied significantly. The Col-0 acc2 line yielded an average of eight transplastomic events per bombarded sample. On the other extreme, Sav-0 yielded one event for every two to four bombarded samples. The poor transformation efficiency observed in Sav-0 may be due to the presence of a full-length ACC2 open reading frame with 15 missense mutations that may have maintained some residual enzymatic activity (Parker et al., 2016). The reasons for higher plastid transformation efficiency in the Col-0 background are not understood and remain to be explored.

In conclusion, we have advanced efforts to develop plastid transformation in Arabidopsis by developing a system for rapid evaluation of plastid transformation efficacy, complete with new transformation vectors, culture media, and new RLD and Ws ACC2-knockout lines that expand the range of accessions available for experimentation. While the root system has proven its value by yielding fertile plants, leaf explants offer the advantage of recovering transplastomic events in 35 to 42 d (5–6 weeks), instead of the 120 to 240 d required in the root system (Ruf et al., 2019). Here, we exploited the system to evaluate the impact of truncating the left targeting region of the pMEK vectors. This rapid scoring system could also be used to evaluate new selective marker genes. However, endoreduplication in leaf tissues likely fostered the sterility that we encountered in plants regenerated from leaf tissue (Galbraith et al., 1991; Melaragno et al., 1993; Yu et al., 2017). An alternative approach to overcome this issue may be to work with very young leaves or to explore the use of the BABY BOOM gene, which has the potential to preserve the diploid state in leaves (Lutz et al., 2015). Systematic research in the two systems should lead to reproducible protocols for plastid transformation in Arabidopsis, allowing the production of fertile transplastomic plants in an optimized time frame. This will enable full exploitation of Arabidopsis as a model system.

Accession Numbers

The nucleotide sequences of Arabidopsis plastid transformation vectors are available under GenBank accessions MF461355 (pATV1), MN326017 (pMEK18), and MN326018 (pMEK19).

Supplemental Data

The following supplemental materials are available.

• Supplemental Figure S1. DNA sequence and translation of the region targeted by the sgRNA in exon 1 of the ACC2 gene.

• Supplemental Figure S2. Abundant callus forms on the wild-type leaf explants while spectinomycin (100 mg L⁻¹) effectively suppresses callus formation of acc2 knockout leaf explants on the selective SPEED medium.

• Supplemental Methods S1. Detailed information on experimental procedures.