Highly efficient and genotype-independent barley gene editing based on anther culture

Abstract

Recalcitrance to tissue culture and genetic transformation is the major bottleneck for gene manipulation in crops. In barley, immature embryos of Golden Promise have typically been used as explants for transformation. However, the genotype dependence of this approach limits the genetic modification of commercial varieties. Here, we developed an anther culture-based system that permits the effective creation of transgenic and gene-edited plants from commercial barley varieties. The protocol was tested in Golden Promise and four Australian varieties, which differed in phenology, callus induction, and green plant regeneration responses. Agrobacterium-mediated transformation was performed on microspore-derived callus to target the HvPDS gene, and T0 albinos with targeted mutations were successfully obtained from commercial varieties. Further editing of three targets was achieved with an average mutation rate of 53% in the five varieties. In 51 analyzed T0 individuals, Cas9 induced a large proportion (69%) of single-base indels and two-base deletions in the target sites, with variable mutation rates among targets and varieties. Both on-target and off-target activities were detected in T1 progenies. Compared with immature embryo protocols, this genotype-independent platform can deliver a high editing efficiency and more regenerant plants within a similar time frame. It shows promise for functional genomics and the application of CRISPR technologies for the precise improvement of commercial varieties.

Keywords: Agrobacterium; CRISPR; Hordeum vulgare; genetic transformation; off-target; targeted mutation.

Figure 1

Flowchart of anther culture-based gene editing in barley. The procedures linked by dashed arrows show the typical process of barley anther culture for doubled haploid (DH) production, which takes approximately 35 weeks from the sowing of F1 seeds. The platform is modified for barley genetic transformation and gene editing in commercial varieties, with three steps (framed in boxes) skipped but Agrobacterium infection added. After infection, the plantlets regenerate in 6–8 weeks and are suitable for mutant identification. For details, see Methods.

Figure 2

Divergent phenology responses for anther culture in barley. (A) Days from seeding to optimum anther dissection time in five commercial varieties. The box plot represents data from five sowing dates in 2017 and 2018. Donor plants were grown in a controlled environment. Different letters indicate significant differences at p < 0.05. (B) Inter-ligule intervals between the flag leaf and the top-second leaf of tested varieties when microspores reach the optimum late-uninuclear stage. Scale bars, 20 μm.

Figure 3

Induction response in five barley varieties. Each Petri dish contained anthers from one spike, and the response was visually scored after 6 weeks. The bar chart represents the average proportions of dishes in each category and comprises all dishes/spikes from two sowing dates in 2018. The number of dissected spikes for each variety is indicated at the bottom of each column. Letters indicate significant differences in the proportion of spikes rated category 2 to 4 in each variety (p < 0.10).

Figure 4

Regenerated green and albino plants from anther culture and chromosome doubling in five barley varieties. (A and B) All induced embryos and calli were transferred for regeneration after 6 weeks’ induction, and the numbers of green (A) and albino (B) plants were counted before transplanting. The number of dissected spikes for variety screening is indicated at the bottom of each column. (C) Deaths (%) and chromosome doubling (% DH) in the transplanted plants of five barley varieties. The number of plants transplanted for each variety is indicated at the bottom of each column. Data are presented as means ± SE from three sowing dates in 2017. Different letters indicate significant differences at p < 0.05.

Figure 5

Genetic and phenotypic characterization of the barley hvpds mutants. (A) Schematic of the HvPDS gene in Golden Promise (used as wild type) and the target sequence for gene editing. Boxes and lines represent exons and introns, respectively. The predicted domain of phytoene desaturase is shaded in gray. Scale bar, 500 bp. (B) Nested PCR amplification for the gene target and flanking sequence from barley T0 materials, and the restriction digestion of PCR products. (C and D) Representatives of regenerated albino mutants (C) and gene mutations determined by Sanger sequencing (D). Photos were taken 3 months after anther dissection. (E) Editing rates in the T0 events.

Figure 6

Characterization of targeted gene editing in multiple sites in different barley varieties. (A) Schematic of the HORVU3Hr1G090980 gene and the target sequence for gene editing. Boxes and lines represent exons and introns, respectively. The predicted domain of oxoglutarate/iron-dependent dioxygenase is shaded in gray. Scale bar, 250 bp. (B) Editing efficiencies of different targeting sites in barley T0 lines. (C) Overall editing efficiencies of HvPDS and HORVU3Hr1G090980 in different barley varieties. (D) Mutation sites and types in HvPDS and HORVU3Hr1G090980 editing events. n, number of T0 individuals.

Figure 7

Off-target analysis in barley T1 progenies. (A) Alignment of Target 2 sequence with a highly similar hit in the barley Morex genome. The PAM sequence is highlighted in red, and the BspHI recognition site is in italics. (B) PCR–RE assay of eight T1 representatives with off-target mutations. WT, Golden Promise. (C) On-target and off-target mutation frequencies in four independent T1 populations.