The AP2/ERF transcription factor SlERF52 functions in flower pedicel abscission in tomato.

In plants, abscission removes senescent, injured, infected, or dispensable organs. Induced by auxin depletion and an ethylene burst, abscission requires pronounced changes in gene expression, including genes for cell separation enzymes and regulators of signal transduction and transcription. However, the understanding of the molecular basis of this regulation remains incomplete. To examine gene regulation in abscission, this study examined an ERF family transcription factor, tomato (Solanum lycopersicum) ETHYLENE-RESPONSIVE FACTOR 52 (SlERF52). SlERF52 is specifically expressed in pedicel abscission zones (AZs) and SlERF52 expression is suppressed in plants with impaired function of MACROCALYX and JOINTLESS, which regulate pedicel AZ development. RNA interference was used to knock down SlERF52 expression to show that SlERF52 functions in flower pedicel abscission. When treated with an abscission-inducing stimulus, the SlERF52-suppressed plants showed a significant delay in flower abscission compared with wild type. They also showed reduced upregulation of the genes for the abscission-associated enzymes cellulase and polygalacturonase. SlERF52 suppression also affected gene expression before the abscission stimulus, inhibiting the expression of pedicel AZ-specific transcription factor genes, such as the tomato WUSCHEL homologue, GOBLET, and Lateral suppressor, which may regulate meristematic activities in pedicel AZs. These results suggest that SlERF52 plays a pivotal role in transcriptional regulation in pedicel AZs at both pre-abscission and abscission stages.

Introduction

In plants, organ abscission specifically detaches senescent, injured, infected, or dispensable leaves or flower organs to maintain the healthy growth of the main body. Abscission also detaches mature seeds or fruits to disperse the plant’s progeny. To abscise an organ, plants generally develop a specialized tissue, the abscission zone (AZ), at a predetermined site on the organ to be abscised. Under normal conditions, the AZ firmly attaches the organ to the plant body; after initiation of abscission, the AZ tissues weaken, allowing the organ to detach. Plant hormones act in opposition to regulate organ separation: ethylene promotes abscission and auxin inhibits abscission, in an ethylene-antagonistic manner (Taylor and Whitelaw, 2001; Meir ). Abscission involves the activation of cell-wall-degradation machinery in the AZ, including cell-wall hydrolytic enzymes such as endo-β-1,4-glucanase (also referred as cellulase (Cel)), polygalacturonase (PG), expansin, and xyloglucan endotransglucosylase/hydrolase (Roberts ; Nakano and Ito, 2013). These enzymes degrade the primary cell wall or middle lamella pectin of AZ tissues so that abscising organs detach easily from the parent plant. Marked changes in transcription activate cell-wall degradation and other abscission processes (Meir ; Wang ); therefore, unveiling the mechanisms of transcriptional regulation will enable a more clear understanding of the onset of abscission. In Arabidopsis thaliana, various transcription factors (TFs) positively or negatively regulate abscission of floral organs, including stamens, petals, and sepals. These TFs include members of the KNOTTED-LIKE HOMEOBOX (KNOX) family, the DNA BINDING WITH ONE FINGER (DOF) family, the MADS-box family, the ETHYLENE-RESPONSIVE FACTOR (ERF) family, the AUXIN RESPONSE FACTOR (ARF) family, and the ZINC FINGER family (Fernandez ; Ellis ; Cai and Lashbrook, 2008; Wei ; Chen ; Shi ). However, the relationships among these TFs and the resulting transcriptional cascades remain incompletely understood.

Tomato (Solanum lycopersicum) plants develop AZs at the midpoint of the flower pedicels. The AZs have a knuckle-like structure with a groove on the surface. If pollination fails, the flower will senesce and eventually abscise from the plant at the AZ. During flower pedicel abscission, expression of PG and Cel greatly increases (Meir ; Nakano ; Wang ). Programmed cell death also occurs during flower pedicel abscission (Bar-Dror ). In tomato, several mutations can inhibit development of pedicel AZs, causing a ‘jointless’ phenotype. For example, jointless (j) is a mutation of a MADS-box TF gene and lateral suppressor (ls) is a mutation of a GRAS family TF gene (Schumacher ; Mao ). The locus for another ‘jointless’ mutation, j-2, has not yet been identified, but a sequencing analysis has identified a candidate gene encoding C-terminal domain (CTD) phosphatase-like 1 (ToCPL1) (Yang ). In addition, the current study group has showed that the MADS-box TF MACROCALYX (MC) regulates pedicel AZ development and that a heterodimer of MC and J functions as a unit for this regulation (Nakano ). Recent work identified another tomato MADS-box TF gene, SlMBP21, as a regulator of pedicel AZ development and showed that the encoded protein also interacts with MC and J (Liu ). To identify more genes involved in pedicel abscission, Nakano , 2013) identified genes that are regulated by both MC and J and are expressed specifically in pedicel AZs. Interestingly, the results of this screen suggested that the tomato WUSCHEL homologue (LeWUS), GOBLET (GOB), Ls, and BLIND (Bl), which regulate meristem activity, also regulate pedicel AZ activity. However, their detailed roles in AZs remain unknown. The screen also identified several other TF genes: OVATE, SlERF52, and a zinc finger-homeodomain (ZF-HD) family protein.

Based on the previous study, the current work focused on an ERF family TF gene, SlERF52. The ERF family TFs constitute one of the largest TF families in the plant kingdom (Riechmann ); for example, the tomato genome includes at least 85 genes for ERF family proteins, most of which remain uncharacterized (Sharma ). The ERF family members contain a single DNA-binding domain, the APETALA2 (AP2)/ERF domain (Ohme-Takagi and Shinshi, 1995), and, as monomers, recognize the GCC-box or CRT/DRE (for C-repeat/dehydration responsive element) cis-acting DNA elements (Allen ; Hao ; Yang ). The AP2/ERF domain was identified in proteins binding to ethylene-responsive gene promoters (Ohme-Takagi and Shinshi, 1995), but subsequent studies revealed that the ERF family TFs function in diverse aspects of plant growth, development, and physiology, such as meristem activity, floral organ abscission, lipid metabolism, alkaloid biosynthesis, and responses to environmental stress (extreme temperature, water deficit, salinity, low oxygen, and pathogen infection) (Stockinger ; Liu ; Solano ; van der Fits and Memelink, 2000; Banno ; Berrocal-Lobo ; Gu ; Kirch ; Komatsu ; Broun ; Xu ; Shoji ; Iwase ). The current study used gene suppression to investigate the function of SlERF52. The results demonstrate that SlERF52 is required for activation of cell-wall-degrading enzymes during abscission as well as pedicel-specific gene expression at the pre-abscission stage.

Materials and methods

Plant materials

The tomato cultivar Ailsa Craig was used to make transgenic plants. The jointless mutant (TK3043) and the MC-suppressed transgenic plants were described previously (Nakano ). Plants were grown in a controlled growth room under a 16/8 light/dark cycle at 25 °C.

Plasmid construction

Oligonucleotide primers used for gene amplification are listed in Supplementary Table S1 (available at JXB online). To obtain the SlERF52 gene fragments, cDNAs were synthesized from flower pedicel total RNA and used as templates for PCR amplification. A plasmid for RNA interference (RNAi) targeting SlERF52 was constructed as follows. A 315-bp fragment of SlERF52 was amplified with a pair of gene-specific primers, AK327476-F2 and AK327476-R2, and then cloned into the pENTR/D-TOPO Gateway entry vector (Invitrogen). The cloned fragment was transferred into a binary vector for RNAi, pBI-sense, anti sense-GW (Inplanta Innovations, Japan) using Gateway LR Clonase Enzyme Mix (Invitrogen). The resultant plasmid was designated pBI-GW-SlERF52-RNAi.

Plasmids for the transactivation assay were constructed as follows. The full-length open reading frame of SlERF52 was amplified with the primer pair NcoI-SlERF52-F1 and BamHI-SlERF52-R1 and inserted into the NcoI and BamHI sites of pGBKT7 (Clontech), which carries an auxotrophic marker gene (TRP1). The resulting plasmid was designated pGBK-SlERF52. Sequencing analysis revealed that SlERF52 from Ailsa Craig possesses five single-nucleotide polymorphisms in comparison with the genome sequence of the cultivar Heinz 1706 (accession no. AB889741). A series of partial SIERF52 fragments were amplified using NcoI-SlERF52-F1 and BamHI-SlERF52-R2 for amino acids 1–74, NcoI-SlERF52-F1 and BamHI-SlERF52-R3 for amino acids 1–98, NcoI-SlERF52-F1 and BamHI-SlERF52-R4 for amino acids 1–133, and NdeI-SlERF52-C3 and BamHI-SlERF52-R1 for amino acids 133–162. Each amplified DNA fragment was inserted into pGBKT7, resulting in pGBKT7-SlERF521–74, pGBKT7-SlERF521–98, pGBKT7-SlERF521–133, and pGBKT7-SlERF52133–162, respectively.

Plant transformation

The plant transformation vector pBI-GW-SlERF52-RNAi was introduced into Agrobacterium tumefaciens EHA105 by the freeze–thaw method (Cindy and Jeff, 1994). Cotyledons of tomato seedlings were used for transformation by Agrobacterium infection according to the previously described method (Sun ).

Transactivation assay

Transactivation assays in yeast cells were conducted according to the previously described method (Cho ). The yeast strain AH109 (Clontech), which carries two auxotrophic marker genes (ADE2 for adenine biosynthesis and HIS3 for histidine biosynthesis) under the GAL4 cis-regulatory element, was used for the experiment. Yeast transformation was performed using the Frozen EZ Yeast Transformation II kit (Zymo Research, Irvine, CA, USA), and transformants were selected on SD media lacking tryptophan (SD/–Trp, Clontech). Assays for transactivation activity were performed on SD media lacking tryptophan, adenine, and histidine (SD/–Trp/–Ade/–His). In the experiment, a target protein was expressed as a fusion protein with the GAL4 DNA-binding domain (GAL4DBD), and if the protein had the potential to activate transcription, the auxotrophic marker genes (ADE2 and HIS3) were expressed and the yeast cell was able to grow on the adenine- and histidine-deficient selection medium.

Sequence analysis

Multiple sequence alignment was performed with ClustalW version 1.83 and the phylogenetic tree was constructed by the neighbour-joining method. GENETYX version 10 (GENETYX, Japan) was used for the analysis. Supplementary Table S2 shows the accession numbers for the sequences used in the analysis.

Reverse-transcription PCR and quantitative reverse-transcription PCR

Total RNAs were extracted using the RNeasy plus mini kit (Qiagen) in combination with the QIA shredder spin column (Qiagen). First-strand cDNA was synthesized using PrimeScript II 1st strand cDNA Synthesis Kit (Takara Bio, Japan). PCR amplifications were performed using the ExTaq polymerase (Takara Bio). qRT-PCR was carried out with a 7300 Real-Time PCR System (Applied Biosystems) using THUNDERBIRD SYBR qPCR MIX (Toyobo, Japan). Data were normalized to the expression of the SAND gene (SGN-U316474) as an internal control (Exposito-Rodriguez ). Relative quantification of expression of each gene was performed using the 2–ΔΔCT method (Livak and Schmittgen, 2001).

Flower pedicel abscission assay

Flower pedicels were harvested at anthesis. The flower was removed from the pedicel using a sharp blade, the pedicel end was inserted into a 1.0% agar plate, and the plate was placed in a glass chamber to maintain high humidity. An abscission event was defined by pedicel detachment that occurred naturally or in a response to vibration applied to the distal portion of the explant.

Results

SlERF52 is a member of the ERF transcription factor family

As described previously, SlERF52 expression is strictly limited to the AZ region in the pedicel and SlERF52 expression is suppressed in plants that lack an AZ, namely MC-knockdown plants and j mutants (Nakano , 2013; Fig. 1A and B). No or very low expression of SlERF52 was detected in other organs, including roots, leaves, stems, flowers, sepals, and fruits (Fig. 1C). These results suggest that SlERF52 plays a specific role in pedicel abscission.

Fig. 1.

Expression specificity of SlERF52. (A) Expression analysis of SlERF52 in a jointless mutant, a MC-suppressed transgenic plant (AS-MC), and the wild type (WT). (B) Expression specificity of SlERF52 within flower pedicel parts, the distal (Dis), proximal (Prox), and abscission zone (AZ) regions in WT anthesis flowers. (C) Expression analysis of SlERF52 among various organs. Expression analysis was performed by reverse-transcription PCR using SAND (A and B) and SlActin-51 (C) as the internal control.

Phylogenetic analysis of the AP2/ERF domain revealed that SlERF52 belongs to group Va of ERFs (Fig. 2A). This group includes: Arabidopsis WAX INDUCER 1 (WIN1)/SHINE1 (SHN1), SHN2, and SHN3, which regulate cutin biosynthesis and abscission of floral organs (Aharoni ; Broun ; Shi ); the tomato homologue of SHN3 (SlSHN3) (Shi ); barley (Hordeum vulgare) NUDUM (NUD), which regulates lipid biosynthesis for hull-caryopsis adhesion of grain (Taketa ); tomato LeERF1, which regulates ethylene signalling (Li ); Medicago truncatula ERF REQUIRED FOR NODULE DIFFERENTIATION (EFD) (Vernie ); and poplar (Populus tremula × P. alba) PtaERF003, which is involved in adventitious and lateral root formation (Trupiano ). Group Va ERFs have three conserved domains: the AP2/ERF domain, conserved motif V (CMV)-1, and CMV-2 at the C-terminus (Fig. 2B). Group Va includes two subgroups, a subgroup with the normal CMV-1 domain (including WIN1/SHN1 and its orthologues), and another subgroup with an incomplete CMV-1 domain (including LeERF1, the AT5G15190-encoding protein, EFD, PtaERF003, and SlERF52; Fig. 2B).

Fig. 2.

Phylogenetic analysis and sequence alignment of SlERF52. (A) Phylogenetic relationships of SlERF52 with ERF proteins; the phylogenetic tree was constructed using amino acid sequences of the AP2/ERF domains in each ERF protein (Supplementary Table S2). (B) Multiple sequence alignment of group Va ERF proteins. Following the highly conserved AP2/ERF domain, there are two conserved motifs, CMV-1 and CMV-2 (Nakano ). Based on the CMV-1 structure, the group Va ERFs are further classified into two subgroups: the normal CMV-1 subgroup including Arabidopsis SHNs, SlSHN3, and NUD and the incomplete CMV-1 subgroup including SlERF52, LeERF1, PtaERF003, AT5G25190, and EFD. The CMV-2 motif is conserved in both subgroups. Arrowhead indicates the amino acid substituted in the nud mutant (nud1.b; Taketa ).

SlERF52 acts as a positive regulator of flower pedicel abscission

To analyse the biological role of SlERF52, RNAi was used to knock down SlERF52 expression. To that end, transgenic plants with an RNAi vector targeting SlERF52 were generated, 15 independent transgenic plants were obtained, and the three plants with the lowest expression levels of SlERF52 (plants 7, 18, and 20) were selected for further analysis (Fig. 3A and Supplementary Fig. S1). The three SlERF52-suppressed plants appeared similar to wild-type plants and developed pedicel AZs normally (Fig. 3B), indicating that SlERF52 does not regulate differentiation of pedicel AZs. To examine the pedicel abscission behaviour of the transgenic plants, flower pedicel abscission was induced by removing the flower from the pedicel, which stimulates ethylene production and restricts auxin supply from the flower (Meir ) and observing the frequency of abscission in the flower-removed pedicels for 3 d (Fig. 3C). The abscission frequency of pedicels from plants 7 and 20 at 3 d after flower removal was significantly lower than that of wild type, indicating that the pedicels of the two suppression lines showed decreased abscission potential compared to the wild type (Fig. 3D). The pedicels from plant 18 exhibited significant reduction of abscission frequency at 1 d after flower removal, although the abscission eventually occurred at the same level as the wild type at 3 d after flower removal (Fig. 3D). These observations indicate that the suppression of SlERF52 impaired activation of pedicel abscission.

Fig. 3.

Suppression of SlERF52 partially inhibited flower pedicel abscission. (A) Transcript levels of SlERF52 in three SlERF52-suppressed transgenic plants (plants 7, 18, and 20). Transcript level was examined in anthesis flower pedicels by quantitative reverse-transcription PCR. The levels are shown as fold-change values relative to that of WT; error bars indicate standard deviation of biological triplicates. (B) Suppression of SlERF52 did not affect pedicel abscission zone development. (C) Pedicel abscission was inhibited in the SlERF52-suppressed transgenic plants; arrows indicate the abscission zone. (D) Rate of flower pedicel abscission in SlERF52-suppressed transgenic plants. Pedicel abscission was induced by removing anthesis flowers. Asterisks indicate significant differences by chi-square test (*P<0.05, **P<0.01, respectively; n=206 for WT, n=112 for plant 7, n=119 for plant 18, and n=110 for plant 20).

Suppression of SlERF52 inhibits induction of genes for cell-wall hydrolytic enzymes

Expression of genes encoding cell-wall hydrolytic enzymes, including PG and Cel, is induced in response to the abscission stimulus (Roberts ). Because suppression of SlERF52 decreased the rate of pedicel abscission, the current work investigated whether it also affected the transcript levels of genes encoding PG (TAPG1, TAPG2, and TAPG4) and Cel (Cel1 and Cel5) during flower pedicel abscission. In accord with previous reports (Meir ; Nakano ; Wang ), in wild-type plants, removal of the flower induced the expression of TAPG1, TAPG2, TAPG4, Cel1, and Cel5 in AZs, but SlERF52 was expressed at constant levels before and after the onset of abscission (Fig. 4). In SlERF52-suppressed plants 7 and 20, TAPG1, TAPG2, TAPG4, and Cel5 were induced to significantly lower levels than in the wild type (Fig. 4), and the levels of these four genes corresponded to the abscission rates in the suppressed transformants (Fig. 3). The suppression was more severe for PG genes than for Cel5. Meanwhile, the levels of Cel1 expression did not correspond to the abscission rate.

Fig. 4.

Expression analysis in SlERF52-RNAi plants during abscission. Pedicel abscission was induced by anthesis flower removal and gene expression was investigated for 2 d by quantitative reverse-transcription PCR. For single RNA sample preparation, 3–24 pedicel abscission zones, which include both attached and abscised pedicels, were harvested in bulk and used for the analysis. Levels of transcripts are shown as fold-change values relative to the 0 d sample of WT (for Cel1, Cel5, TAPG4, SlERF52, Bl, GOB, LeWUS, and Ls). Because TAPG1 and TAPG2 transcript levels for the 0 d sample of WT were below detection limit (shown as ND), the level of the two genes are shown relative to the sample of SlERF52-suppressed plant 20 at 1 d. Data are mean±SD of biological triplicates.

Suppression of SlERF52 reduces expression of transcription factor genes LeWUS, GOB, and Ls in pedicel AZs

Previously, this study group reported that LeWUS, GOB, Ls, and Bl, four TF genes associated with shoot apical meristem or axillary meristem function, might also be involved in the regulation of pedicel AZ activity (Nakano , 2013). To investigate whether SlERF52 affects the expression of these four TF genes, their transcript levels in the SlERF52-suppressed plants were analysed. As observed previously, in wild-type plants, the expression of LeWUS, GOB, and Ls decreased markedly in response to flower removal, an abscission stimulus. In the SlERF52-suppressed plants, however, the transcript levels of these three genes were much lower than the wild type before flower removal (0 d) and their levels remained low after flower removal (1 d and 2 d) (Fig. 4). By contrast, the expression of Bl increased during abscission similarly in the SlERF52-suppressed plants and the wild type (Fig. 4). The transcript level of Bl in the suppressed plants was slightly lower than that in wild type throughout the examined period but the difference was not significant, except in the d-1 samples. The expression pattern of these four TF genes was not correlated with the expression of SlERF52 in shoot apices or leaf axillae of wild type plants and also was not affected by suppression of SlERF52 (Supplementary Fig. S2), which is consistent with the normal vegetative growth of the suppressed plants. The results suggest that the SlERF52-mediated regulation of LeWUS, GOB, and Ls is specific to pedicel AZs.

SlERF52 functions as a transcriptional activator

ERF proteins can activate or repress transcription of target genes (Fujimoto ; Ohta ). This study investigated the transcriptional activation potential of SlERF52 using a yeast system, with the GAL4 DNA-binding domain (DBD) fused to SlERF52 and marker genes expressed under the control of the GAL4 target-binding site. The results showed that the construct with the full-length SlERF52 coding region (GAL4DBD-SlERF521–162) induced expression of the marker genes (Fig. 5), indicating that SlERF52 can activate transcription. To identify which region of SlERF52 is necessary for the activity, three truncated SlERF52 proteins (SlERF521–74, SlERF521–98, and SlERF521–133) were assayed, but no activity was detected in any of the C-terminal truncated proteins (Fig. 5). By contrast, this work did detect activity in a construct with the C-terminal 30 amino acids (GAL4DBD-SlERF52133–162) (Fig. 5). These results indicated that the transcriptional activation activity of SlERF52 requires the C-terminal 30-amino-acid region that contains the CMV-2 motif.

Fig. 5.

SlERF52 functions as a transcriptional activator. (A) Schematic of full-length and truncated SlERF52 proteins examined in this assay; fused products of these proteins with GAL4DBD were expressed in yeast cells. (B) Results of transactivation assays; yeast cells expressing the fusion proteins were inoculated on SD/–Trp (control medium) and SD/–Trp–His–Ade (selection medium). Yeast cells expressing GAL4DBD were used as a negative control.

Discussion

SlERF52 functions as a positive regulator of flower pedicel abscission

These data showed that suppression of SlERF52 reduced the rate of pedicel abscission and repressed induction of the genes for cell-wall hydrolytic enzymes PG and Cel (Cel5, TAPG1, TAPG2, and TAPG4). Abscission of flower pedicels and leaf petioles in tomato requires the activity of these enzymes (Lashbrook ; Jiang ). Therefore, these results suggest that SlERF52 induces pedicel abscission through upregulation of these enzyme genes. In contrast to the low induction of Cel5, TAPG1, TAPG2, and TAGP4 in the suppressed plants, the expression of Cel1 was not correlated with suppression of SlERF52 or the abscission rate. These results also indicate that the transcript level of Cel1 in the wild type peaked at 1 d after flower removal and then declined, but the transcript levels of Cel5, TAPG1, TAPG2, and TAPG4 continuously increased (Fig. 4). In addition, Cel1 is expressed in a pedicel region distinct from the region where TAPG1 and TAPG4 are expressed (Bar-Dror ). These results imply that the transcriptional regulation of Cel1 is independent of the regulation mediated by SlERF52. Therefore, these results indicate that SlERF52 acts as a key positive regulator of flower pedicel abscission, but abscission also involves a SlERF52-independent pathway.

Interestingly, SlERF52 is necessary, but not sufficient, for the upregulation of PG and Cel genes; before the onset of abscission, SlERF52 is also expressed at a similar level to that observed after flower removal, but this expression does not induce PG and Cel gene expression (Fig. 4). Post-transcriptional regulation may explain the transcription-independent activity of SlERF52 (as will be discussed).

SlERF52, a positive regulator of abscission, has an opposite role to that of the Arabidopsis homologues, SHNs, which act as negative regulators of abscission of floral organs such as sepals, stamens, and petals (Shi ). Simultaneous suppression of all SHN genes induces earlier abscission of floral organs, possibly due to decreased cutin deposition and altered cell-wall composition of structural proteins and pectin (Shi ). SlERF52 and SHNs belong to the group Va ERF family, but belong to different subgroups based on their CMV-1 motif structures: SlERF52 belongs to the subgroup with an incomplete CMV-1, and the SHNs belong to the subgroup with normal CMV-1 structure (Fig. 2B). The biological function of CMV-1 has not been identified, but the structural difference in CMV-1 between SlERF52 and SHNs may be a possible cause of their functional diversity. Also, the latter half of the CMV-1 motif, which is lost in the incomplete-type Va ERFs, contains an important active site, as demonstrated in a study of mutants of NUD, a barley orthologue of WIN1/SHN1 (Fig. 2B; Taketa ). Elucidation of the function of the CMV-1 motif will provide insights into the functional diversity between the subgroups within the Va ERFs, including SlERF52 and SHNs.

Several group Va ERF proteins, including WIN1/SHN1, SHN2, SHN3, and EFD, act as transcriptional activators (Vernie ; Shi ). However, the domain for transcriptional activation was not identified. The current study demonstrated that the C-terminus of SlERF52, which contains the CMV-2 motif, acts as an activation domain. As shown in Fig. 2B, the CMV-2 motif is highly conserved in WIN1/SHN1, SHN2, SHN3, and EFD, suggesting that the conserved motif functions as a transcription activation domain in these proteins.

SlERF52 is involved in the expression of TF genes for shoot apical meristem and axillary meristem function in flower pedicel AZs

LeWUS, GOB, Ls, and Bl, key TF genes for meristem-associated functions, are expressed specifically in flower pedicel AZs, suggesting that these four TFs may have an additional function in control of organ abscission through regulation of meristem-like activity in the cells within the AZ (Nakano , 2013). The current study found that LeWUS, GOB, and Ls were expressed at significantly lower levels in the SlERF52-suppressed plants, implying that SlERF52 may be involved in the regulation of these TF genes. Expression of SlERF52, LeWUS, GOB, and Ls is reduced in pedicels of MC-suppressed plants, SlMBP21-suppressed plants, and j mutants (Nakano ; Liu ; Fig. 1A), indicating that SlERF52 may mediate the effect of MC, J, and SlMBP21 on these meristem-associated regulators. Two SlERF52 homologues that belong to the incomplete CMV-1 type subgroup regulate plant development through modulation of meristem activity: medicago EFD controls formation of root nodule meristems (Vernie ) and poplar PtaERF003 controls formation and growth of adventitious and lateral root meristems (Trupiano ). Therefore, the control of meristem-associated regulation may be a conserved biological function for the group Va ERFs with incomplete CMV-1 motifs. PtaERF003 functions in an auxin-regulated pathway that regulates root meristems (Trupiano ). Similar to root meristem regulation, expression of the shoot meristem-associated TF genes in the AZs may be regulated by a signalling pathway that requires auxin supplied from the flower before the onset of abscission, and SlERF52 may function in the auxin signalling pathway in the AZs.

The expression analyses revealed that SlERF52 activates the expression of LeWUS, GOB, and Ls in the AZ cells, but the expression of these three TF genes was suppressed after stimulation of abscission, even though SlERF52 expression remained constant (Fig. 4). By contrast, the cell-wall hydrolytic enzyme genes were suppressed before the stimulation of abscission, even though SlERF52 expression remained constant, a reverse pattern to that of the three TF genes. This partial dependence on SlERF52 is discussed in the next section.

Of the four TF genes for meristem-associated functions, Bl exhibits significant upregulation after flower removal, an expression pattern distinct from LeWUS, GOB, and Ls (Fig. 4). Thus, Nakano hypothesized that an independent pathway controls Bl expression, although MC and J are involved in the expression of all four TF genes. In the current study, the suppression of SlERF52 did not significantly affect Bl expression, indicating that a SlERF52-independent pathway regulates Bl. Also, the intense induction of Bl after flower removal suggests that Bl may be involved in pedicel abscission (Nakano ). The induction of Bl in the SlERF52-suppressed plants may help explain the partial progression of abscission in the suppressed lines.

Functional switching of SlERF52 before and after the onset of abscission

SlERF52 functions in the regulation of pedicel abscission and regulates transcription of distinct sets of genes before and after the onset of abscission. In the pre-abscission stage, the expression of LeWUS, GOB, and Ls requires SlERF52, either directly or indirectly. In response to an abscission-inducing stimulus, the expression of Cel5, TAPG1, TAPG2, and TAPG4 was also regulated by SlERF52, directly or indirectly. However, after the onset of abscission, the induction of LeWUS, GOB, and Ls ceases. To explain how SlERF52 is involved in the regulation of distinct sets of genes before and after the onset of abscission, it is postulates that coregulators specify the function of SlERF52 in the different states. In this hypothesis, SlERF52 recruits state-specific TFs and each state-specific TF complex activates expression of a distinct set of target genes. Several ERFs are predicted to require cofactors to bind target genes (Chakravarthy ; Kannangara ; Cheng ). As another possibility, repressor proteins or chromatin remodelling at SlERF52-binding sites may restrict the transactivation activity of SlERF52 in a stage-specific manner.

This work used knockdown experiments to examine SlERF52 function. A recent study using overexpression of SlMBP21 provided substantial insights on SlMBP21 gene function, adding to the results of the knockdown assay (Liu ). However, unlike the study of SlMBP21, overexpression of SlERF52 may not be effective to clarify SlERF52 function because the activity of SlERF52 in AZs is likely determined by other factors associated with SlERF52, not by the expression level of SlERF52.

In conclusion, the results of this study demonstrated that SlERF52 regulates pedicel AZ-specific transcription at both pre-abscission and abscission stages and that the regulation during the latter stage includes some of the genes required for abscission. The functional switching between before and after the onset of abscission, by a still-unknown mechanism, raises the possibility that SlERF52 serves as a hub TF that regulates the phase transition between the two stages. The identification of the switching mechanism will further improve the understanding of abscission.

Supplementary material

Supplementary data are available at JXB online.

Supplementary Table S1. Sequences of the oligonucleotide primers used in this study.

Supplementary Table S2. Accession numbers of ERFs used for construction of the phylogenetic tree.

Supplementary Fig. S1. Expression analysis of SlERF52-RNAi transgenic plants.

Supplementary Fig. S2. Expression of SlERF52 and meristem-associated TF genes in shoot apex and leaf axilla.