Regulation of wound ethylene biosynthesis by NAC transcription factors in kiwifruit

Abstract

Background: The phytohormone ethylene controls many processes in plant development and acts as a key signaling molecule in response to biotic and abiotic stresses: it is rapidly induced by flooding, wounding, drought, and pathogen attack as well as during abscission and fruit ripening. In kiwifruit (Actinidia spp.), fruit ripening is characterized by two distinct phases: an early phase of system-1 ethylene biosynthesis characterized by absence of autocatalytic ethylene, followed by a late burst of autocatalytic (system-2) ethylene accompanied by aroma production and further ripening. Progress has been made in understanding the transcriptional regulation of kiwifruit fruit ripening but the regulation of system-1 ethylene biosynthesis remains largely unknown. The aim of this work is to better understand the transcriptional regulation of both systems of ethylene biosynthesis in contrasting kiwifruit organs: fruit and leaves.

Results: A detailed molecular study in kiwifruit (A. chinensis) revealed that ethylene biosynthesis was regulated differently between leaf and fruit after mechanical wounding. In fruit, wound ethylene biosynthesis was accompanied by transcriptional increases in 1-aminocyclopropane-1-carboxylic acid (ACC) synthase (ACS), ACC oxidase (ACO) and members of the NAC class of transcription factors (TFs). However, in kiwifruit leaves, wound-specific transcriptional increases were largely absent, despite a more rapid induction of ethylene production compared to fruit, suggesting that post-transcriptional control mechanisms in kiwifruit leaves are more important. One ACS member, AcACS1, appears to fulfil a dominant double role; controlling both fruit wound (system-1) and autocatalytic ripening (system-2) ethylene biosynthesis. In kiwifruit, transcriptional regulation of both system-1 and -2 ethylene in fruit appears to be controlled by temporal up-regulation of four NAC (NAM, ATAF1/2, CUC2) TFs (AcNAC1-4) that induce AcACS1 expression by directly binding to the AcACS1 promoter as shown using gel-shift (EMSA) and by activation of the AcACS1 promoter in planta as shown by gene activation assays combined with promoter deletion analysis.

Conclusions: Our results indicate that in kiwifruit the NAC TFs AcNAC2-4 regulate both system-1 and -2 ethylene biosynthesis in fruit during wounding and ripening through control of AcACS1 expression levels but not in leaves where post-transcriptional/translational regulatory mechanisms may prevail.

Keywords: Biosynthesis; Ethylene; Kiwifruit; NAC; Regulation; Transcription factors; Wounding.

Fig. 1

Different models of ripening behavior in kiwifruit and tomato. In kiwifruit, the competence to ripen occurs well before ripening initiation. Ripening initiation and the initial softening period (ripening phase 1) are accompanied by non-autocatalytic ethylene production (system-1) and are separated from the late ripening period (ripening phase 2) that is accompanied by autocatalytic ethylene production (system-2). Kiwifruit ripening stages are defined in Richardson et al. (2011) [27] and tomato ripening stages in Feller et al. (1995) [28]. The BBCH plant development scale is described in Hess et al. (1997) [29]. In tomato, the competence to ripen (responsiveness to exogenous ethylene) coincides with the mature green (MG) stage and is closely followed by ripening initiation and autocatalytic ethylene production and softening. MG: mature green, B: breaker, O: orange, R: red

Fig. 2

Ethylene production in wounded immature A. chinensis ‘Hort16A’ kiwifruit. (A) Fruit were wounded with two incisions (open symbols) or four incisions (closed symbols) and transferred to a 1.5 L sealed jar at 2 L h^− 1 air flow. Ethylene production was measured continuously over the time course shown. Control unwounded fruit (open diamond) produced no ethylene over the course of the experiment. Three biological repeats (consisting of three fruit each) were harvested per time point from fruit with two incisions. (B) Soluble solids concentration (SSC, % Brix) and firmness (kgF) changes in immature fruit after wounding. Control: unwounded fruit; Wounded: fruit cut with 2 incisions were measured at 120 h after wounding; Ripe: eating ripe fruit. Data are the mean ± SE, n = nine biological replicates. **/++: statistically different compared to control in two/four cut respectively (p < 0.01, ANOVA (A) /Student’s t-test (B)

Fig. 3

Consensus phylogram of aminocyclopropane-1-carboxylic acid synthase (ACS) proteins from Arabidopsis, tomato and kiwifruit. Predicted ACS proteins from Arabidopsis (At), tomato (Sl), and kiwifruit (Ac/blue dot) were aligned using the Geneious MUSCLE alignment tool and a consensus UPGMA bootstrap phylogram was generated using 1000 replicates (Jukes-Cantor distance matrix). Only branches with over 50% support threshold are displayed. AT = aminotransferase. Type I, II and III ACS proteins were assigned based on presence of conserved C-terminal sequences (see Supplemental Fig. S1 for alignment). Type I (red) = RLSF/SLSF only; type II (green) = WVF, RLSF and RDE rich domains (TOE/target of ETO1 domain); type III (blue) = absence of type I/II domains (based on Yoshida et al. 2006) [57]. (III) = Both SlACS4 and 13 cluster with type I, but show absence of typical type I residues in the C-terminus (see Supplemental Fig. S1 for alignment). S1, S2 = involved in system-1, − 2 ethylene production. T = involved in transition between system-1 and -2

Fig. 4

(A) Ethylene production after mechanical wounding of immature A. chinensis ‘Hort16A’ fruit. Time course as per Fig. 2A with ripe fruit for comparison. (B) qRT-PCR expression analysis of ethylene biosynthetic (ACS, ACO) genes and NAC TFs in wounded fruit. Data are mean ratio calibrated ± SD, n = three biological replicates and expressed as a ratio compared to the PP2A reference gene. **: p < 0.01 significantly different compared to unwounded fruit, based on mixed model statistics (R packages nlme and emmeans, as described in Materials and Methods

Fig. 5

Phylogenetic analysis of putative NAC transcription factors from kiwifruit. (A) The 147 putative NAC TFs identified in the A. chinensis ‘Red5’ genome (see Supplemental Table S2) were initially aligned using Geneious MUSCLE alignment tool, then manually curated. The DNA binding site was extracted and realigned using Muscle. Phyml [59] was used to construct the tree shown using the JTT (Jones, Taylor & Thornton) substitution method with default calculation parameters and was rooted with Acc22424.1 and Acc22425.1 as outgroup. (B) A UPGMA consensus tree of kiwifruit (Ac) and tomato (Solyc) NAC TFs (complete ORFs). The tree was generated using 1000 replicates (Jukes-Cantor distance matrix). Only branches with over 50% support threshold are displayed. Solyc10g55760 is used as outgroup

Fig. 6

(A) Ethylene production in wounded kiwifruit leaves. Expanding A. chinensis ‘Hort16A’ leaves were wounded with a 96-well pin blotter and ethylene production was measured continuously over the time course shown (0, 1.5, 3, 6 h after wounding). Data are mean ± SD, n = 3 biological replicates. (B) qRT-PCR of ethylene biosynthetic (ACO and ACS) genes and NAC TFs in wounded leaves and control leaves (0, 3, 6 h after wounding). Data are mean ratio calibrated ± SD, n = three biological replicates and expressed as a ratio compared to the PP2A reference gene. **: p < 0.01 significantly different compared to unwounded leaves, based on Student’s t-test (A) or mixed model statistics (B) (R packages nlme and emmeans, as described in the Materials and Methods)

Fig. 7

Promoter activation by NAC (NAM, ATAF1/2, CUC2) and EIL (Ethylene-insensitive3-like) TFs using deletions of the A. chinensis ‘Hort16A’AcACS1 promoter. Different sized AcACS1 regulatory regions (< 1000 bp) upstream of the ATG (AcACS1pro + length in bp) were cloned upstream of the LUC reporter gene of pGreenII-0800LUC in frame with the start ATG and tested for transient activation in N. benthamiana. LUC/REN luminescence ratio values of transcription factors AcNAC1-AcNAC4 and a pool of AcEIL1–4 (equal mixture of AcEIL1–4) were compared to a GUS control construct which was set to 1. Statistical differences were determined by Tukey’s honest significant difference test (HSD) after analysis of variance (ANOVA) analysis compared to GUS. Data are mean ± SE, n = three biological replicates (plants), * different at p < 0.05

Fig. 8

Electrophoretic mobility shift assays (EMSA) of AcACS1 promoter fragments (27 bp) with recombinant NAC1–4 proteins. The DNA binding domains of NAC1–4 (described in Nieuwenhuizen et al. 2015 [52]) were over-expressed in Escherichia coli as Maltose Binding Protein (MBP)-tagged fusion proteins and purified by amylose resin affinity purification and EMSA was run according to Nieuwenhuizen et al. (2015) [52]. Wt = wildtype double-stranded DNA probe with putative NAC palindromic binding site (underlined): CATTATACGTATAGTCAACCACATAAC. Mut = mutated double-stranded DNA probe with randomly mutated NAC binding site (italic/underlined): CATCGATCCATCTGTCAACCACATAAC. NAC = MBP-NAC1, −NAC2, −NAC3 or -NAC4