CRISPR-Cas9-mediated mutagenesis of kiwifruit BFT genes results in an evergrowing but not early flowering phenotype

Abstract

Phosphatidylethanolamine-binding protein (PEBP) genes regulate flowering and architecture in many plant species. Here, we study kiwifruit (Actinidia chinensis, Ac) PEBP genes with homology to BROTHER OF FT AND TFL1 (BFT). CRISPR-Cas9 was used to target AcBFT genes in wild-type and fast-flowering kiwifruit backgrounds. The editing construct was designed to preferentially target AcBFT2, whose expression is elevated in dormant buds. Acbft lines displayed an evergrowing phenotype and increased branching, while control plants established winter dormancy. The evergrowing phenotype, encompassing delayed budset and advanced budbreak after defoliation, was identified in multiple independent lines with edits in both alleles of AcBFT2. RNA-seq analyses conducted using buds from gene-edited and control lines indicated that Acbft evergrowing plants had a transcriptome similar to that of actively growing wild-type plants, rather than dormant controls. Mutations in both alleles of AcBFT2 did not promote flowering in wild-type or affect flowering time, morphology and fertility in fast-flowering transgenic kiwifruit. In summary, editing of AcBFT2 has the potential to reduce plant dormancy with no adverse effect on flowering, giving rise to cultivars better suited for a changing climate.

Keywords: Actinidia chinensis; BFT; CRISPR-Cas9; dormancy; flowering; kiwifruit.

Figure 1

The kiwifruit PEBP gene family and gene expression. (a) A phylogenetic tree with the kiwifruit PEBP proteins. (b) RNA‐seq expression analysis of the three kiwifruit BFT genes in different kiwifruit organs, axillary buds collected from field‐grown plants at monthly intervals and axillary buds collected from excised canes exposed to cold over 4 weeks. The bars and circles represent mean FPKM ± SE of three biological replicates. The period of no visible growth during winter is shaded grey. Fruit T1 and T2 represent fruit at 20 and 40 days after anthesis respectively. (c) Localization of AcBFT2 expression. A 2.5‐kb fragment upstream from the AcBFT2 translational start site (AcBFT2p) was fused to uidA (GUS) and introduced into Arabidopsis. Histochemical localization of GUS activity in transgenic Arabidopsis identified promoter activity in the vascular tissue of cotyledons and rosette leaves (arrows, left) and axillary buds (arrowheads, right).

Figure 2

Design of the construct for CRISPR‐Cas9‐mediated mutagenesis of kiwifruit AcBFT2. (a) AcBFT2 (GenBank accession number KX11601, gene model Acc28037) was used to design primers and single guide RNA (sgRNA) sequences. Exons are presented in grey, sgRNA target regions are indicated in red, and primer positions are indicated by blue arrowheads. (b) The sequences targeted by the gene editing construct. The target sequences used for sgRNAs are underlined, PAM sites are highlighted in blue, and mismatches between AcBFT1, AcBFT2 and AcBFT3 alleles (I and II) and the sgRNAs are indicated in red.

Figure 3

Delayed dormancy and early budbreak in Acbft kiwifruit lines. (a) A dormant bud in a control line. (b–d) Examples of delayed leaf senescence (b), delayed budset (c), and emergence of new shoots (d) in Acbft lines. Photographs (a–d) taken on 11 July 2021. (e, f) Early budbreak in Acbft lines. Frequency of dormant buds (D), budbreak (BB) and shoot outgrowth (SO) in Acbft and control plants. The plants were defoliated and data were recorded after 4 weeks using 20 distal buds from 18 lines with bi‐allelic edits in AcBFT2 (Acbft) and 20 control and non‐edited lines (Control). The white dot is the median, the black bar is the interquartile range, the whiskers represent the minimum and maximum, and the violin represents the distribution of the data. (g) Earlier budbreak of a representative Acbft edited line (right) compared with the control (left) in year 3. Photograph taken on 17 August 2021. Scale bars represent 5 cm.

Figure 4

AcBFT regulates shoot architecture in kiwifruit. (a, b) Vigorous main shoot in a representative control and multiple shoots in a representative Acbft kiwifruit mutant line (bft#9) after establishment in the glasshouse. Shoots are indicated with black arrows. (c–f) Architecture of kiwifruit lines after first year's growth. The plants have been defoliated. Branches are indicated with white arrows. (g) Branching on representative fast‐flowering U6‐CEN4#7 (cen4) and double Acbft cen4 mutant lines (cen4‐bft). Shoots are indicated with black arrows. (h) The number of axillary shoots on cen4 and double mutant lines. (i, j) The numbers of leaves and flowers on the main (longest) shoot of cen4 and double mutant plants. The bars represent a mean ± SE of 6 cen4‐bft and 13 cen4 lines. **Indicates statistical significance (P < 0.05). (k, l) Appearance of flowers. Flowers on the main (longest) shoot are indicated with black arrowheads. Orange arrowheads mark flowers on the side shoot. Scale bars represent 5 cm.

Figure 5

Transcriptome analysis. (a) Representative kiwifruit bud samples. (b) Venn diagrams of differentially expressed genes (DEGs) in Acbft mutants relative to controls in May and July. (c, d) Hierarchical clustering (One minus Pearson's correlation, average linkage, using mean FPKMs of three biological replicates) of gene expression of DEGs in May (c) and July (d) with gene expression in wild‐type buds collected in different months in the field. The phenological stages of wild‐type buds are indicated below. (e, f) Examples of genes identified as co‐expressed with AcBFT2 (e) or with expression opposite to that of AcBFT2 (f). The graphs represent expression (mean FPKM ± SE of three biological replicates) of different classes of candidate genes in axillary buds collected from field‐grown plants at monthly intervals. The kiwifruit gene identification and closest Arabidopsis homologue and its function are indicated.