Ethylene-induced differential gene expression during abscission of citrus leaves.

The main objective of this work was to identify and classify genes involved in the process of leaf abscission in Clementina de Nules (Citrus clementina Hort. Ex Tan.). A 7 K unigene citrus cDNA microarray containing 12 K spots was used to characterize the transcriptome of the ethylene-induced abscission process in laminar abscission zone-enriched tissues and the petiole of debladed leaf explants. In these conditions, ethylene induced 100% leaf explant abscission in 72 h while, in air-treated samples, the abscission period started later and took 240 h. Gene expression monitored during the first 36 h of ethylene treatment showed that out of the 12 672 cDNA microarray probes, ethylene differentially induced 725 probes distributed as follows: 216 (29.8%) probes in the laminar abscission zone and 509 (70.2%) in the petiole. Functional MIPS classification and manual annotation of differentially expressed genes highlighted key processes regulating the activation and progress of the cell separation that brings about abscission. These included cell-wall modification, lipid transport, protein biosynthesis and degradation, and differential activation of signal transduction and transcription control pathways. Expression data associated with the petiole indicated the occurrence of a double defensive strategy mediated by the activation of a biochemical programme including scavenging ROS, defence and PR genes, and a physical response mostly based on lignin biosynthesis and deposition. This work identifies new genes probably involved in the onset and development of the leaf abscission process and suggests a different but co-ordinated and complementary role for the laminar abscission zone and the petiole during the process of abscission.

Introduction

Abscission is a cell separation process by which plants can shed some of their organs. It is the consequence of a breakdown in adhesion between a group of specialized cells differentiated at the site of organ shedding, named the abscission zone (AZ; Roberts ). From an evolutionary point of view, abscission is a highly favourable process that has several advantages such as seed dispersal as well as the shedding of unwanted, infected or damaged organs. In an agricultural context, however, abscission may become a major limiting factor of yield.

For convenience, abscission can be divided into four major steps (Patterson, 2001). Steps one and two involve the ontogeny of the abscission zone and the acquisition of competences to respond to abscission signals, respectively, while the third phase triggers the onset of the cell separation that leads to final organ shedding. The fourth step provides the differentiation of a protective layer that preserves the remaining tissue in the main body of the plant. During the third and fourth stages of the process, cells within the AZ may undergo elongation, although it remains unclear whether this is an essential component of the pathway or a consequence of it (Patterson, 2001). It is well established that plant growth hormones are deeply involved in abscission and that among them ethylene is thought to be its natural regulator (Jackson and Osborne, 1970). Although there is as yet no clear evidence for a direct link between the ethylene perception and the onset of abscission (Addicott, 1982; Abeles ; Patterson and Bleecker, 2004), it is well documented that the progress of the process is concomitant with an increase in the production of the hormone. It has also been shown that ethylene induces the synthesis and secretion of several cell wall and middle lamella hydrolytic enzymes involved in the separation of the abscinding organs (Tucker ; Kalaitzis ; del Campillo and Bennet, 1996; Burns ; Brummell ) and in some instances their own gene expression. The availability of ethylene-insensitive mutants has provided an excellent material with which to dissect the role of the hormone in the regulation of abscission (Roberts ). The delay in abscission displayed by ethylene mutants affected at the level of perception (etr1, ers2) or sensitivity (ein2 and ein3) has helped in the assessment of a role for the hormone in accelerating and synchronizing rather than activating the process (Bleecker and Patterson, 1997; Chao ). In addition, the occurrence of ethylene-dependent and -independent pathways in abscission has been suggested by the Arabidopsis dab (delayed floral organ abscission) mutant that exhibits a more significant delay in abscission than the ethylene mutants, although no other ethylene response pathway is affected (Patterson and Bleecker, 2004).

Most of the current molecular knowledge on the abscission process has been obtained from the model plant system Arabidopsis thaliana. However, there is an increasing economic interest in developing molecular approaches focused on the abscission of commercial crops. In addition, it has been noted that the reduced number of cell layers that constitute the AZ (Roberts ) has been a general impediment to discover the molecular mechanisms that bring about abscission. This restriction has led researchers to use, apart from Arabidopsis (Roberts ), other anatomically more convenient organisms such as Sambucus nigra, which possesses 30 layers of cells in its leaflet AZ, Phaseolus vulgaris, Solanum lycopersicum, and Citrus spp. In citrus, abscission has traditionally received much attention since fruit setting and growth and hence final yield are rather dependent upon many environmental and endogenous factors triggering abscission (reviewed in Iglesias ). The laminar abscission zone (LAZ) of this species for instance that comprises 15–20 cell layers located at the blade and petiole junction of the leaf has been extensively used to study cell separation (Ratner ; Iwahori and van Stevenink, 1976; Addicott, 1982; Burns ). In vitro ethylene treatments on leaf explants from these different species has allowed the study of the molecular and biochemical features associated with abscission (Sexton and Roberts, 1982). In citrus, this strategy has facilitated, for example, the cloning of pivotal genes encoding cell wall-degrading enzymes, such as cellulases (Burns ) and has rendered many other key biochemical results (Goren, 1993). In addition, recent advances in citrus research including the rapid development of functional genomics and molecular biology resources (reviewed in Tadeo ; Talon and Gmitter, 2008) has provided an innovative set of valuable tools to address new challenges. Thus, critical functional and global expression studies through microarrays with several platforms have recently been published (Forment ; Shimada ; Cercós ; Fujii ) and analyses of large EST collections in public databases have been initiated (Terol ). In this work, advantage has been taken of these new resources to investigate the ethylene-induced transcriptome of the laminar abscission zone of citrus leaves. Attention has also been focused on the petiole, the tissue that remains attached to the plant body since research in abscission has mostly been centred on the events triggered within the abscission zone. Ethylene has been proved to have an unequivocal promotive effect on citrus abscission (Goren, 1993) and increases in its endogenous level are associated with enhanced abscission of both reproductive (Gomez-Cadenas et al., 2000; Iglesias ) and vegetative organs under natural and stress conditions (Tudela and Primo-Millo, 1992; Gómez-Cadenas ). Current thinking in this area postulates that other hormones such as ABA (Zacarias ; Agustí et al., 2007), auxins and gibberellins (Talon ) may act as intermediates in the citrus abscission process while ethylene operates instead as the final hormonal activator of abscission (Tudela and Primo-Millo, 1992; Iglesias ).

The goal of this work was to identify and classify genes involved in the process of citrus leaf abscission. For this purpose, time-course microarray comparisons of global expression profiles between the laminar abscission zone and the petiole were performed. The selected time points to be analysed were based on the leaf abscission kinetics displayed under controlled in vitro ethylene treatments.

Materials and methods

Plant material and treatments

One-year-old mature hardened uniform leaves of ‘Clementina de Nules’ (Citrus clementina Hort. Ex Tan.) were harvested from adult trees grown in a homogeneous experimental orchard under normal cultural practices at the Instituto Valenciano de Investigaciones Agrarias. In order to enhance responsiveness to ethylene and prevent abscission-protective effects of the leaf blade (Jackson ), leaf explants consisting of a petiole carrying the laminar abscission zone (LAZ) were generated, cutting out a high portion of the leaf blade (90%) with a razor blade. Leaf explants were inserted by their petiole end in 1% (w/v) agar (see Supplementary Fig. S1 at JXB online) in 9 cm Petri dishes (Sterilin) and incubated in sealed 10 l containers at 22 °C with a 16 h light period under fluorescent lighting. For each time point, three independent pools of 100 explants were distributed in two Petri dishes. Laminar abscission zones and contiguous petiolar tissues were manually dissected with a razor blade and separated in approximately 2–3 mm2 section samples. Thus, these were composed of tissue either strongly enriched in LAZs or exclusively petiolar (Pet). Leaf blade and petiolar wings containing blade-like tissue characteristic of the Clementine petiole were also completely removed and discarded.

For abscission kinetic studies, leaf explants were incubated for up to 240 h in the presence or absence of ethylene (10 μl l−1). For gene expression analyses, laminar abscission zones and contiguous petiolar tissues from leaf explants harvested at 0, 6, 12, 24, and 36 h after the onset of the ethylene treatment, were manually dissected with a razor blade and separated in approximately 2–3 mm2 section samples. Thus, these were composed of tissue either strongly enriched in LAZs or exclusively petiolar (Pet). After harvest, all samples were immediately immersed in liquid nitrogen and stored at –80 °C for RNA isolations. Abscission was expressed as the percentage of leaf explants that shed with a gentle touch.

RNA isolation, sample labelling, and microarray hybridization

Total RNA was isolated from frozen tissue using the RNeasy Plant Mini Kit (Qiagen, Hilden, Germany). RNA samples were treated with RNase free DNase (Qiagen) through column purification following the manufacturer's instructions. RNA quality was tested by OD260/OD280 ratio and gel electrophoresis. The same total RNA samples were used for RNA amplification, for subsequent microarray hybridization, and real-time RT-PCR analyses. The first and second cDNA strands were generated from 5 μg of total RNA. The double-stranded cDNA, in vitro transcription, RNA labelling, microarray hybridization, and slide washes were performed as previously described in Cercós . For each sample, RNA was extracted from three biological replicates and independently processed, labelled, and hybridized to different microarrays. Total RNA from each sample was extracted and used to synthesize Cy5-labelled antisense cRNA. This was co-hybridized with Cy3-labelled antisense cRNA synthesized from a reference sample containing a mixture of equal amounts of RNA from all experimental samples. This experimental design is useful for reducing the number of hybridizations necessary to make all the possible pairwise comparisons between samples. It has clear advantages over other experimental designs, specially when the number of samples is relatively high (Churchill, 2002; Novoradovskaya ), since differential expression in a sample pair can be calculated directly from the ratios of their hybridizations with the reference, avoiding the multi-step calculations that are necessary when other designs, such as the loop design, are used (König ). All individual samples were labelled with Cy5 and the reference with Cy3 to avoid dye artefacts. Three independent biological replicates for both LAZ and Pet were used for each time point (0, 6, 12, 24, and 36 h after ethylene treatment) and a cDNA citrus microarray containing 12 672 probes obtained from several citrus species and hybrids and corresponding to 6875 putative unigenes representing about a quarter of the citrus transcriptome was utilized (Forment ).

Data acquisition and analysis

Hybridized arrays were scanned with a Scanarray Gx scanner (PerkinElmer) equipped with the Scanarray Express software, following the manufacturer's instructions in order to obtain an appropriate photomultiplier gain ratio for the two channels and a percentage of 1% of saturated spots. The software used to transform the intensity into numeric data was the GenePix 4.1 (Axon Instruments), compatible with the Scanarray software used for the data acquisition. Data from spots flagged as ‘not found’ or ‘bad’ during the scanning, as well as those with a signal-to-background ratio lower than 2 were discarded. Data were normalized to compensate for differences in sample labelling and other non-biological sources of variability. In order to obtain a robust normalization and, since the comparisons performed revealed a low percentage of probes showing significant differences in expression, normalization was carried out using the Lowess method. Probes showing significant differential gene expression between samples were identified using the Linear Models in Microarrays (LIMMA) library (Smith, 2004) of the Bioconductor software package (Gentleman ). Gene expression differences were considered to be significant when they were associated with a P-value lower than 0.05 and an M contrast cut-off value of ±1 being M=log2 [LAZ/Pet]. Probes with positive values represented genes preferentially expressed in the LAZ and those associated with negative values were preferentially expressed in the Pet. The raw microarray data of 30 hybridizations as well as the protocols used to produce the data and the normalized data were deposited in the ArrayExpress database under the accession number E-MEXP-1428. Functional classification of the selected genes was carried out through MIPS (Munich Information Center for Protein Sequences, http://www.mips.gsf.de) categorization. Gene expression data obtained from microarray hybridization were confirmed through real-time RT-PCR analysis.

Real-time RT-PCR

Total RNA from the same samples as described above was used for real-time RT-PCR. Accurate RNA concentration values were determined by fluorometric assays with the RiboGreen dye (Molecular Probes) according to the manufacturer's instructions. Three fluorometric assays per RNA sample were performed. Quantitative one-step real-time RT-PCR was performed with a LightCycler 2.0 Instrument (Roche) equipped with Light-Cycler Software version 4.0 as previously described (Agustí et al., 2007). Specificity of the amplification reactions was assessed by post-amplification dissociation curves and by sequencing the reaction products. Transformation of fluorescence intensity data into relative mRNA levels was carried out using a standard curve constructed with a 10-fold dilution series of a single RNA sample. Relative mRNA levels were normalized to total RNA amounts as previously described (Bustin, 2002; Hashimoto ). Sequences of forward and reverse primers and the sizes of the resulting fragments are listed in Supplementary Table S1 at JXB online.

Phloroglucinol lignin staining

Phloroglucinol staining of lignin in leaf explants was carried out according to Tadeo and Primo-Millo (1990). The leaf blade was removed and the petiole containing the LAZ cut longitudinally in order to facilitate the staining procedure and further image acquisition. A saturated solution of phloroglucinol (Sigma-Aldrich) was prepared in 20% HCl and applied directly to fresh cut tissue protected from light. Observation was performed with a high performance zoom system microscope (Leica Microsystems).

Ruthenium red pectin staining

Ruthenium-red staining of pectins in leaf explants was performed according to Vercher . The solution was applied directly on the fresh tissue and observation was performed as above.

Results

Abscission kinetics of citrus leaf explants

In order to elucidate the kinetics of ethylene-induced leaf abscission in citrus, comparative assays between ethylene-treated and air-treated explants were performed. Overall, in vitro ethylene-treated explants showed an acceleration of the abscission process that was completed in no more than 72 h after the beginning of the treatment, while in control samples 100% abscission took a period of 240 h (Fig. 1). The onset of abscission was also dissimilar. Ethylene-induced abscission started as soon as 24 h after the beginning of the treatment, at least 2 d before air-treated samples did. Based on these data, a time-course experiment covering the first 36 h period of the treatment was designed for subsequent gene expression analyses of ethylene-induced abscission. At this moment, the percentage of abscised explants was 21% (Fig. 1). Samples consisting of either laminar abscission zone-enriched tissue (LAZ) or petioles (Pet) were collected at 0, 6, 12, 24, and 36 h and used to compare the gene expression profile associated with each one of the two tissues. In this 36 h period, two phases were distinguished: the first one prior to the onset of explant abscission (phase I) including samples at 6 h and 12 h, and the second one, when organ shedding had started that included samplings at 24 h and 36 h (phase II).

Fig. 1.

Abscission kinetics of Citrus clementina leaves under ethylene or air treatments. Data are the average of three independent experiments and error bars show SE.

Gene expression between LAZ and Pet and functional classification of differentially expressed genes

Differential gene expression profiling between LAZ and Pet after 0, 6, 12, 24, and 36 h of treatment with ethylene was assessed through microarray data comparisons of each tissue versus a common reference sample. Results showed that ethylene regulated 725 (6%) of the 12 672 cDNA probes printed in the microarray. Out of these 725 probes, 216 (29.8%) were preferentially expressed in the LAZ, and 509 (70.2%) in the Pet (Fig. 2). A low number of these ESTs (35) were differentially expressed during the first phase of ethylene treatment (6–12 h) although differential expression strongly increased (693 ESTs) during the second one (24–36 h). In the Pet, only 10 ESTs were significantly expressed during the first phase while the bulk of them (499) were regulated thereafter.

Fig. 2.

Venn diagrams including the number of ethylene regulated ESTs in both the laminar abscission zone (LAZ) and the petiole of Citrus clementina leaves after 6, 12, 24, and 36 h of ethylene treatment. The total EST number is also shown in boxes. Data are based on microarray analyses.

Clustering of ESTs into MIPS functional categories indicated that transcriptional activation that takes place in both tissues during ethylene-induced abscission mostly involved ‘defence’, ‘protein biosynthesis’, and ‘metabolism’ categories. Interestingly, other categories involving non-functionally classified proteins (‘non-classified proteins’, ‘No MIPS classification for the Arabidopsis orthologue’ and ‘No orthologue in Arabidopsis thaliana’) were also highly represented (Fig. 3). Almost all MIPS functional categories contained more probes preferentially expressed in the Pet than in the LAZ.

Fig. 3.

Ratio and number of ethylene regulated ESTs in laminar abscisión zone (LAZ) (open box) or petiole (Pet) (filled box) of Citrus clementina leaves assigned to MIPS (Munich Information Center for Protein Sequences, http://www.mips.gsf.de) categories. Positive and negative values indicate the ESTs fraction preferentially expressed in LAZ and Pet, respectively. The total number of ESTs included in each one of the MIPS category is shown in the vertical axis. Data are based on microarray analyses.

Identification and differential expression of genes involved in ethylene-induced leaf abscission

With regard to the abscission process, differential expression of selected genes and gene families of particular biological interest is illustrated in Figs 4–9 (see Supplementary Tables S4–S7 at JXB online) and further discussed in detail. The selected gene families included those related to cell wall modification, lignin formation and deposition, control of transcription, signal transduction, transport, and defence.

In addition, a list of selected candidate genes associated with other important processes, such as protein metabolism, hormone biosynthesis and signal transduction, transport, and regulation of transcription preferentially expressed in the LAZ or in the Pet during abscission is also shown in Table 1.

Table 1.

Selected candidate genes associated to protein metabolism, hormone biosynthesis, and signal transduction, transport, regulation of transcription and protein phosphorylation preferentially expressed in laminar abscission zone (LAZ, positive values of gene expression ratio) and petiole (Pet, negative values of gene expression ratio) after 0, 6, 12, 24, and 36 h of treatment of citrus leaf explants with ethylene

ID	Putative gene identification [Species] (Accession No.)	E-value	Putative Ath orthologue	Expression ratio [log2 (LAZ/Pet)]
				Time after C2H4 treatment
				0 h	6 h	12 h	24 h	36 h
Protein metabolism
C18006E05	Ribosomal protein L4/L1e [Medicago truncatula] (ABP03135)	1e-53	AT3G09630	0.33±0.22	0.49±0.22	0.55±0.22	1.29±0.22	0.37±0.22
C07010D01	40s Ribosomal protein S23 [Oryza sativa (japonica cultivar-group)] (BAB92932)	3e-69	AT5G02960	−0.24±0.17	−0.43±0.17	−0.09±0.17	0.58±0.17	1.09±0.17
C05002E08	Chloroplast ribosomal L1-like protein [Arabidopsis thaliana] (CAB87802)	5e-113	AT3G63490	−0.35±0.19	−0.17±0.19	−0.17±0.19	0.12±0.19	−1.02±0.19
C20001D08	Ribosomal protein S19 [Solanum tuberosum] (1909359A)	3e-58	AT3G04920	−0.17±0.16	−0.41±0.16	−0.42±0.16	−0.82±0.16	−1.00±0.16
C01003D06	Ubiquitin extension protein [Capsicum annuum] (ABK42077)	6e-70	AT2G47110	0.12±0.17	0.19±0.17	0.09±0.17	1.01±0.17	0.06±0.17
C18016E03	RING-finger protein [Capsicum annuum] (AAX20031)	3e-130	AT5G14420	0.03±0.22	–0.17±0.24	1.08±0.22	–0.31±0.22	1.63±0.33
C07007F08	F-box/LRR-repeat protein 15 [Arabidopsis thaliana] (Q9SMY8)	2e-36	AT4G33210	−0.08±0.25	0.07±0.25	−0.55±0.25	−1.25±0.25	−0.67±0.25
C07010F08	Hypothetical protein (small ubiquitin-like modifier) [Vitis vinifera] (CAN79327)	1e-42	AT4G26840	−0.12±0.17	−0.13±0.17	−0.16±0.17	−0.32±0.17	−1.12±0.17
Hormone biosynthesis and signal transduction
C03005C12	Allene oxide synthase [Citrus sinensis] (AAO72741)	9e-65	AT5G42650	0.12±0.18	0.86±0.18	1.15±0.18	0.54±0.18	−0.25±0.18
C18010E03	SAM dependent carboxyl methyltransferase	3e-138	AT1G19640	0.10±0.26	0.81±0.26	1.59±0.26	0.47±0.26	−0.09±0.26
C20003C11	[Medicago truncatula] (ABO83264)			0.18±0.20	0.83±0.20	1.44±0.20	0.56±0.20	−0.02±0.20
C07002D05	TSB1 (tryptophan synthase β-subunit) [Arabidopsis thaliana] (NP_200292)	0.0	AT5G54810	−0.01±0.22	0.51±0.22	0.23±0.22	0.04±0.22	−1.07±0.22
C02012B01	Hypothetical protein (auxin-independent growth promoter) [Vitis vinifera] (CAN73548)	3e-149	AT5G15740	0.22±0.20	0.35±0.20	0.23±0.20	1.34±0.20	1.96±0.20
C02013E07	Putative auxin-repressed/dormancy-associated protein [Citrus cv. Shiranuhi] (ABL67651)	6e-51	AT1G28330	0.06±0.33	0.45±0.33	−0.69±0.33	−1.38±0.33	−2.77±0.33
C03003A11	Auxin response factor-like protein [Mangifera indica] (AAP06759)	0.0	AT5G62000	−0.21±0.19	0.03±0.19	0.31±0.19	−1.03±0.19	−0.40±0.19
C04015B03				−0.28±0.17	0.18±0.17	−0.03±0.17	0.04±0.17	−1.03±0.17
C08003D01	Gibberellin 2-oxidase 1 [Lactuca sativa] (BAB12442)	1e-47	AT1G30040	−0.01±0.16	0.28±0.16	0.22±0.16	−0.05±0.16	−1.21±0.16
C08005F06	Gibberellin regulated protein [Medicago truncatula] (ABN08074)	2e-24	AT3G02885	−0.07±0.23	0.60±0.23	0.46±0.23	2.83±0.23	3.62±0.23
C07001A05	S-adenosyl-L-methionine synthetase [Petunia×hybrida] (AF170798)	0.0	AT1G02500	−0.15±0.17	0.11±0.19	0.08±0.17	−0.09±0.17	−1.11±0.17
C08014A07	ACC synthase [Citrus sinensis] (CAB60831)	0.0	AT3G61510	0.17±0.15	−0.08±0.15	0.28±0.15	−1.06±0.15	0.19±0.15
C04002H01	Ethylene response factor [Manihot esculenta] (AAX84670)	4e-122	AT3G14230	0.22±0.16	0.72±0.16	0.50±0.16	0.49±0.16	1.10±0.16
Transport
C01005E02	TGF-beta receptor, type I/II extracellular region [Medicago truncatula] (ABD32293)	2e-79	AT1G52190	−0.20±0.18	−0.06±0.16	0.10±0.16	1.09±0.16	0.47±0.16
C04016H02	Stigma/style ABC transporter [Nicotiana tabacum] (AAR06252)	0.0	AT5G13580	0.10±0.18	0.51±0.18	0.54±0.18	2.32±0.18	2.85±0.18
C04002E06	Hypothetical protein (ABC transporter) [Vitis vinifera] (CAN77838)	8e-105	AT1G66950	0.04±0.16	1.05±0.16	0.34±0.16	0.17±0.16	−0.31±0.16
C08011G12	Hexose transporter [Vitis vinifera] (AAT09979)	0.0	AT5G26340	−0.39±0.19	0.02±0.19	−0.37±0.19	−1.13±0.19	−1.54±0.19
C07009F09	Hypothetical protein (triose phosphate/phosphate translocator) [Vitis vinifera] (CAN65381)	0.0	AT5G46110	−0.12±0.25	0.17±0.25	−0.18±0.25	−1.22±0.25	−0.84±0.25
C04019E08	ATCAX5 (calcium exchanger 5); [Arabidopsis thaliana] (NP_175969)	2e-77	AT1G55730	0.01±0.23	0.43±0.23	0.09±0.23	−0.78±0.23	−1.74±0.23
C02001G05	Plasma intrinsic protein 2,2 [Juglans regia] (AAO39008)	5e-142	AT2G37170	0.08±0.17	0.44±0.17	0.08±0.17	−0.76±0.17	−1.23±0.17
C02017H12	Unnamed protein (lipid transfer related domain) [Vitis vinifera] (CAO62888)	1e-158	AT1G64720	0.12±0.15	1.02±0.15	1.06±0.15	0.26±0.15	0.36±0.15
Transcription factors
C03001D06	Hypothetical protein (NAC domain protein) [Vitis vinifera] (CAN62145)	4e-128	AT4G27410	−0.51±0.15	−0.10±0.15	−0.17±0.15	−1.83±0.15	−1.30±0.15
C18015B04				0.22±0.23	0.36±0.23	0.02±0.23	−1.19±0.23	−0.60±0.23
C05001E08	Histone-fold/TFIID-TAF/NF-Y [Medicago truncatula] (ABD32395)	2e-80	AT1G08970	−0.03±0.16	0.05±0.16	−0.58±0.16	−0.62±0.16	−1.28±0.16
C20003B09	MYB transcription factor MYB93 [Glycine max] (ABH02845)	6e-143	AT5G47390	−0.25±0.16	1.09±0.16	0.05±0.16	−0.05±0.16	−0.46±0.16
Protein phosphorylation
C07005E02	ATMKK4 (MITOGEN-ACTIVATED PROTEIN KINASE KINASE 4) [Arabidopsis thaliana] (NP_175577)	3e-23	AT1G51660	−0.10±0.15	0.11±0.15	1.18±0.15	−0.01±0.15	0.15±0.15
C04015A08	Unnamed protein product (protein kinase) [Vitis vinifera] (CAO48860)	8e-10	AT5G01850	0.00±0.20	1.49±0.20	1.32±0.20	0.39±0.20	−0.31±0.20

Significant gene expression ratio (M contrast cut-off value of ±1; P <0.05) is shown in bold. Data are expression ratio values ±SE.

It is also worth noting that, among the 35 ESTs that were significantly regulated by ethylene and preferentially expressed in the LAZ during the first phase of abscission, there were ESTs representing genes involved in the jasmonic acid biosynthesis (allene oxide synthase and SAM-dependent methyltransferase). Other interesting genes preferentially expressed in the LAZ or in the Pet were those potentially involved in phospholipid-mediated signal transduction and transport, such as a protein containing a lipid transfer related domain and related kinases, as well as a MYB transcription factor. Data for these genes are contained in Table 1.

Cell wall metabolism and lignin biosynthesis and deposition

A number of ESTs corresponding to genes encoding cell wall modification enzymes were found to be over-represented in LAZ-enriched tissue during the second phase of the process (Figs 4, 5; see Supplementary Tables S2 and S3 at JXB online). These included, for instance, all ESTs representing cell wall hydrolytic enzymes such as two different polygalacturonases, two different xyloglucan-endotransglycosylases, a pectin-methylesterase, a pectate lyase, and a β-galactosidase, in addition to an acidic cellulase that was previously associated with abscission in citrus (Burns ). Similarly, all significant expressed genes encoding cell wall biosynthesis proteins, namely, a UDP-glucose:protein transglucosylase, involved in cellulose biosynthesis, and two genes involved in nucleotide-sugar metabolism, (a UDP-glucose dehydrogenase and a UDP-glucuronic acid decarboxylase), except a cellulose synthase, were also preferentially expressed in LAZ during the second phase of ethylene-mediated abscission (Fig. 5).

Fig. 4.

Expression ratio between laminar abscission zone and petiole (log2 LAZ/Pet) of ESTs representing cell-wall modifying enzymes with significant changes during the ethylene treatment of Citrus clementina leaves based on microarray analyses. Positive values, preferentially expressed in LAZ. Negative values, preferentially expressed in Pet. Data are the average of three independent comparisons and error bars show SE. Coloured lines represent the expression profiles of different ESTs printed in the cDNA microarray and assembled in the same contig.

Fig. 5.

Expression ratio between laminar abscission zone and petiole (log2 LAZ/Pet) of ESTs representing cell-wall biosynthesis proteins with significant changes during the ethylene treatment of Citrus clementina leaves based on microarray analyses. Positive values, preferentially expressed in LAZ. Negative values, preferentially expressed in Pet. Data are the average of three independent comparisons and error bars show SE.

Furthermore, laccase, the enzyme catalysing the last step of the lignin biosynthesis, was represented by four different ESTs that were clearly over-represented in Pet after 24 h (Fig. 6; see Supplementary Table S4 at JXB online). During the first phase, statistically significant changes were only found for one out of the four ESTs, although the pattern of gene expression change of all four ESTs strongly suggested over-representation of laccase transcripts at 12 h.

Fig. 6.

Expression ratio between laminar abscission zone and petiole (log2 LAZ/Pet) of ESTs representing laccases with significant changes during the ethylene treatment of Citrus clementina leaves based on microarray analyses. Positive values, preferentially expressed in LAZ. Negative values, preferentially expressed in Pet. Data are the average of three independent comparisons and error bars show SE. Coloured lines represent the expression profiles of different ESTs printed in the cDNA microarray and assembled in the same contig.

Lipid-transfer proteins

A set of 25 ESTs corresponding to three different lipid transfer proteins (LTPs) genes (namely, LTP1, LTP2, and LTP3) were preferentially expressed in the LAZ during the second phase of ethylene-mediated abscission (Fig. 7; see Supplementary Table S5 at JXB online).

Fig. 7.

Expression ratio between laminar abscission zone and petiole (log2 LAZ/Pet) of ESTs representing lipid transfer proteins with significant changes during the ethylene treatment of Citrus clementina leaves based on microarray analyses. Positive values, preferentially expressed in LAZ. Negative values, preferentially expressed in Pet. Data are the average of three independent comparisons and error bars show SE. Coloured lines represent the expression profiles of different ESTs printed in the cDNA microarray and assembled in the same contig.

Pathogen-related proteins

A set of pathogen-related protein genes including a stress-related protein, four different chitinases (10 ESTs), two different glutathione-S-transferases (GSTs) and a putative β-1,3-glucanase (Fig. 8; see Supplementary Table S6 at JXB online), was found. All of these genes displayed preferential expression in the Pet during the second phase of abscission (Fig. 8).

Fig. 8.

Expression ratio between laminar abscission zone and petiole (log2 LAZ/Pet) of ESTs representing pathogenesis-related proteins with significant changes during the ethylene treatment of Citrus clementina leaves based on microarray analyses. Positive values, preferentially expressed in LAZ. Negative values, preferentially expressed in Pet. Data are the average of three independent comparisons and error bars show SE. Coloured lines represent the expression profiles of different ESTs printed in the cDNA microarray and assembled in the same contig.

Oxidative stress

ESTs corresponding to five different genes encoding oxidative stress-related enzymes were also preferentially expressed in the Pet after 12–24 h of ethylene treatment (Fig. 9; see Supplementary Table S7 at JXB online). These were a catalase, a glutathione dehydrogenase, an ascorbate peroxidase, and two independent peroxidases. Another peroxidase showed, in contrast, a significant transient over-representation in the LAZ.

Fig. 9.

Expression ratio between laminar abscission zone and petiole (log2 LAZ/Pet) of ESTs representing oxidative stress-related proteins with significant changes during the ethylene treatment of Citrus clementina leaves based on microarray analyses. Positive values, preferentially expressed in LAZ. Negative values, preferentially expressed in Pet. Data are the average of three independent comparisons and error bars show SE. Coloured lines represent the expression profiles of different ESTs printed in the cDNA microarray and assembled in the same contig.

Protein metabolism

Differential expression of probes corresponding to genes encoding structural ribosome proteins was found in both tissues, mostly during the second phase of ethylene-mediated abscission (Table 1). In LAZ, a 40S and a L4/L1 ribosomal protein and a ubiquitin extension protein gene were over-represented, while a chloroplast ribosomal L1 and a ribosomal S19 protein and an F-box/LRR repeat protein were preferentially expressed in the Pet. Interestingly, different genes related to selective ubiquitin-mediated protein degradation (i.e. RING-finger protein and a F-box/LRR repeat protein) were also differentially expressed in the LAZ and in the Pet, respectively (Table 1).

Hormone biosynthesis and response

A number of representative cDNA probes corresponding to genes related to hormone biosynthesis and response were detected in both tissues (Table 1). Thus, a tryptophan synthase, an auxin biosynthesis related gene, and two different probes encoding the same auxin response factor were preferentially expressed in the Pet after 24–36 h of ethylene treatment. During this period, an auxin-repressed protein showed preferential expression in the Pet, while an auxin-independent growth factor was over-represented in LAZ. Ethylene biosynthesis, represented by S-adenosyl-L-methionine synthase and 1-aminocyclopropane-1-carboxylate synthase, was apparently over-represented in the Pet, in contrast to an ethylene response factor preferentially expressed in LAZ. During the entire second phase of the treatment, a gibberelin-regulated gene was preferentially expressed in the LAZ and at the end of the period, gibberellin 2-oxidase, a gene involved in gibberellin catabolism, was over-represented in Pet. During the first phase, two genes encoding jasmonic acid biosynthetic enzymes (allene oxide synthase and SAM dependent carboxyl methyltransferase) were preferentially expressed in the LAZ (Table 1).

Transport

Expression of genes encoding proteins with a predicted transport activity was also found in both tissues, especially during the second phase of ethylene-mediated abscission (Table 1). Two lipid ABC transporters were preferentially expressed in the LAZ, and displayed a complementary profile since one of them was over-represented during the first phase of the treatment (6 h), and the other during the second one. In the Pet, a hexose transporter and a triose phosphate/phosphate translocator were over-represented together with a calcium exchanger and a plasma membrane intrinsic protein (Table 1).

Regulation of transcription

Four cDNA probes encoding three transcription factors showed an interesting pattern of expression change after ethylene treatment (Table 1). Those preferentially expressed in the Pet (NAC transcription factor and histone-fold/TFIID protein) were detected during the last phase of the abscission process whereas the only transcription factor preferentially expressed in LAZ during the first phase was a MYB transcription factor.

Lignin and pectin staining

Presence of lignin and pectin at the abscission zone after ethylene treatment was assessed by direct observation after phloroglucinol and ruthenium red staining, respectively (Fig. 10). Lignin deposition was clearly distinguished 36 h after ethylene treatment in LAZ and Pet (Fig. 10A), an observation that confirms laccase induction in both tissues (Fig. 6; see Supplementary Table S4 at JXB online). Moreover, Fig. 10B shows ruthenium red staining for pectins in LAZ 36 h after the treatment, indicating activation of cell wall metabolism and the occurrence of degradation in this tissue.

Fig. 10.

Phloroglucinol staining for lignin (A) and ruthenium-red staining for pectins (B) in laminar abscission zone (LAZ) and petiole (Pet) before and after ethylene treatment of Citrus clementina leaves. Note that leaf blade was completely removed and petiole was cut longitudinally.

Discussion

It is well established that loss of cell adhesion and breakdown of primary cell walls as a result of the action of cell wall hydrolases are prominent metabolic and physiological processes taking place in AZs during abscission (Roberts ). Furthermore, transcripts related to cell wall hydrolases have been reported to be higher in the tissues supporting cell wall degradation such as the AZs than in the adjacent tissues where those processes are not induced. In a recent report, Cai and Lashbrook (2008) also highlighted new physiological processes and extracellular regulators and transcription factors probably involved in developmentally-regulated activation of the stamen AZ in Arabidopsis. However, information other than that in this field is very scarce. For instance, the differential metabolic and physiological processes activated by ethylene in AZs and in the tissues remaining on the plant after abscission are mostly unknown. In this regard, the genes identified in this differential gene expression survey define pivotal processes associated with ethylene-induced leaf abscission including cell wall hydrolysis and modification, lignin biosynthesis and deposition, lipid transference, hormone regulation, transport, transcriptional control, protein metabolism, defence, and oxidative stress scavenging.

Previous work on citrus abscission has mostly focused on the molecular characterization of individual genes. For instance, Burns cloned an acidic and a basic cellulase (CEL-a1 and CEL-b1) and studied cellulase and polygalacturonase activity in the fruit abscission zone of Valencia sweet orange (Citrus sinensis). This work revealed that these enzyme activities were complementary and essential for the process of abscission to occur. Other studies in citrus have reported isolation and characterization of genes encoding pectinmethylesterases (CsPME1, CsPME2, and CsPME3; Nairn ) and a lipid transfer protein (Wu and Burns, 2003). Moreover, using a suppression subtractive hybridization approach combined with differential mRNA display, Burns et al. (2002) depicted expression changes in a number of genes in laminar, fruit, and flower abscission zones under treatments with abscission agents such as ethephon and CMNP (5-chloro-3-methyl-4-nitro-1H-pyrazole). The results mostly yielded a number of candidate genes potentially involved in cell wall degradation.

The results presented in this manuscript showed that ethylene regulated 725 of the 12 672 cDNA probes printed in the microarray, 509 were preferentially expressed in the Pet and 216 in the LAZ-enriched tissues (Fig. 2). The LAZ to Pet gene expression contrast produced no significant differential expression in non-ethylene treated samples (0 h) while during the first phase of the ethylene treatment (6–12 h) only a few over-represented genes in both LAZ and Pet were found. The number of differentially over-represented genes in both tissues considerably increased as the ethylene treatment progressed during the second phase (24–36 h), when the onset of cell separation was observed (Fig. 1). The potential involvement of a number of selected genes on the abscission process is discussed below.

Cell wall metabolism

The results suggested that ethylene strongly modifies cell wall metabolism mostly in the LAZ through both degradation and new biosynthesis. Co-activation of these two mechanisms that has been widely documented in growth and developmental processes is also apparently necessary during the cell elongation process that takes place through abscission (Addicott, 1982). Indeed, a link between these processes and the generation of a protective layer has previously been proposed (Patterson, 2001).

Outstanding genes related to cell wall metabolism showing preferential expression in the LAZ were an acidic cellulose, two polygalacturonases (PGs), a pectin methylesterase (PMEs), two different xyloglucan-endotransglycosylases, and a pectate lyase among others (Fig. 4).

Cellulases and PGs are probably the most active enzymes in the cell separation process associated with abscission (Gonzalez-Carranza et al., 2002). In citrus, cellulase and PG activities in fruit abscission zone have been previously demonstrated (Burns ) and, in addition to the known acidic cellulase, in this work, evidence is provided of significant preferential expression in LAZ of two PGs that may be potentially involved in citrus leaf abscission. Induction of different cellulases and PGs in a number of species such as peach (Bonghi ) or oilseed rape (Gonzalez-Carranza et al., 2002) and Arabidopsis (Gonzalez-Carranza et al., 2007a) have been reported to occur during the processes of abscission and dehiscence. Moreover, four out of the seven members of the PG family in tomato were reported to be involved in abscission (Kalaitzis ; Hong and Tucker, 2000).

Pectin methylesterases (PMEs) catalyse the de-estherification of pectin into pectate and methanol and participate in the degradation of the middle lamella (Addicott, 1982). Although abscission has generally been related to PME enzyme activity decreases (Addicott, 1982), in citrus there is evidence that PME gene induction is associated with fruit abscission (Nairn ). In the present study, a PME gene different from the previously one reported by Nairn was preferentially expressed in LAZ (Fig. 4). Over-representation of a β-galactosidase transcript (β-gal) that hydrolyses the non-reduced terminal residues of β-galactosyl of β-D-galactoside (Smith ) was also found in LAZ. An additional member of this family has previously been reported to be induced during citrus fruit abscission (Wu and Burns, 2004).

Xyloglucan-endotransglycosylases are enzymes with important roles in the cell wall mechanical alteration and reorganization, cutting and fusing intermicrofibrilar chains (Fry ). The over-representation of these genes in LAZ is an interesting observation that can be linked to the induction of expansins, another family involved in the mechanical alteration of the cell wall, during the onset of leaflet abscission in Sambucus nigra (Belfield ). Another over-represented gene of cell wall modification was a pectate lyase, an endo-acting depolymerizing enzyme that cleaves α-1,4-glycosidic linkages in homogalacturonan of pectate (Solbak ).

Pivotal genes involved in cell wall polymer biosynthesis were a Type IIIa cellulose synthase gene belonging to the RGP1 (reversible glycosylation polypeptide) family, implicated in the biosynthesis of hemicelluloses (Dhugga ) and two different genes of the UDP-glucose dehydrogenase family, involved in the biosynthesis of cell wall polysaccharides (Kärkönen ; Fig. 5).

In addition, evidence of ethylene-induced cell wall degradation was provided by direct observation of pectin deposition on LAZ after 36 h of treatment (Fig. 10B). Taken together, these results suggest that cell wall polymer biosynthesis in citrus LAZ cells is coupled with cell wall degradation and that this reorganization may be related to the generation of protective layers in addition to the cell separation inherent to abscission.

Lignin biosynthesis

Although the role of lignin and lignified tissues in abscission has not yet been clarified, it is well known that during leaf abscission of woody species, ‘ligno-suberization’ of the protective layers is very common (Addicott, 1982). On the other hand, lignification in mechanical cell wall breakage has also been pointed out to be important in abscission (Sexton, 1979). The pattern of change of preferential expression of several ESTs representing laccase showed a sigmoid model with an initial trend to over-representation in LAZ while, during the last steps of abscission, expression was clearly higher in Pet than in LAZ (Fig. 6). This observation is compatible with the picture presented in Fig. 10A that shows lignin deposition in both the Pet and the LAZ. Overall, the data suggest that deposited lignin may act as a physical barrier of the protective layer developed on the part of the organ that remains attached to the plant and also as a component of a fracture line favouring shedding.

Lipid transfer proteins (LTPs)

Three different LTP genes were preferentially expressed in the LAZ at the end of the studied period (Fig. 7), a moment characterized by a high rate of secretion of primary cell wall- and middle lamella-degrading enzymes and the biosynthesis of protective sealing and reinforcement layers. Although it is known that the main function of LTPs is to transfer phospholipids through the membrane (Bourgis and Kader, 1997), the function of these proteins in plants is still a matter of controversy. All LTPs characterized in plants are directed to the secretory pathway (Horvath ), but different patterns of expression from several members of the LTP family in Arabidopsis and tomato suggest a role in a wide range of processes (Trevino and O'Connell, 1998; Clark and Bohnert, 1999). It has been suggested that LTPs could be involved in tissue enforcement, fatty acid interactions, and the provision of suberin and cutin monomers for further polymer biosynthesis, and hence in wax production (Sterk ). Moreover, the isolation and characterization of a tomato non-specific lipid transfer protein has recently revealed a role in polygalacturonase-mediated pectin degradation (Tomassen ). In citrus, it has previously been suggested that a member of the LTP family expressed in the fruit abscission zone may be involved in the transport of cutin monomers to the fracture plane of the abscission zone (Wu and Burns, 2003). Since in woody plants, suberin is a common constituent of protective layers (Addicott, 1982), an alternative explanation is proposed for these observations, suggesting that LTPs over-represented in LAZ may also be involved in the secretion of suberin to protect the part that remains in the main body of the plant after an organ has abscised.

Pathogenesis-related proteins

The shedding of plant organs provides an ideal target for pathogen invasion and cell separation has also been associated with the accumulation of pathogenesis-related (PR) proteins (Roberts ). This accumulation is apparently related to a programme of defence organized to protect the future wound. The results presented in Fig. 8 show over-representation in the Pet of several pathogenesis-related protein transcripts during the last phase of the ethylene treatment. Among them, the preferential expression in the Pet of four chitinases (Fig. 8), a type of gene involved in defensive mechanisms, is especially remarkable. Furthermore, previous reports have already associated the expression of chitinases in the abscission zone with the activation of a preventive defence programme against potential pathogen infections (Lim ; Bleecker and Patterson, 1997; Roberts ).

Several reactive oxygen species (ROS) scavenging genes such as peroxidases, catalases and glutation dehydrogenases were preferentially expressed by ethylene in the Pet (Fig. 9). ROS are versatile molecules related to a wide range of cellular processes, including programmed cell death, development, tropisms, and hormonal signalling (Kwak ) and previous reports have also linked ethylene with ROS signalling events (D'Haeze et al., 2003; Desikan ). Peroxidases in particular are very active in ROS detoxification and can play an important role in the inactivation of indole-acetic acid. Increases in peroxidase activity in the abscission zone during the onset of abscission are well documented (Hinman and Lang, 1965; Gahagan ; Henry, 1975) although their roles in the process of abscission are still unclear. Marynick (1977) concluded that oxidative respiration is essential to trigger abscission, while Michaeli demonstrated that whole plant antioxidant treatments significantly reduced low temperature-induced leaf abscission in Ixora coccinea. On the other hand, Sexton and Roberts (1982), proposed that peroxidases may be involved in the co-ordination of gene expression in response to pathogens.

The preferential expression detected in the Pet during the latter phases of the abscission suggest that ROS scavenging is part of a more general defensive programme including defence and PR genes and the generation of physical barriers.

Particular sets of genes related to protein synthesis and degradation were found to be preferentially expressed in both the LAZ and the Pet (Table 1). Protein population change was represented for instance by ribosomal structural proteins, and ubiquitin related proteins such as F-box proteins. In recent work, activity of the F-box protein HAWAIIAN SKIRT has been linked to the control of petal abscission in Arabidopsis thaliana (Gonzalez-Carranza et al., 2007b). Therefore, the data suggest that ethylene may elucidate major protein metabolism changes through the activation of specific gene expression in LAZ and Pet cells.

Hormones

Among the transcripts showing preferential expression in both the LAZ and the Pet there were various genes encoding hormone related proteins (Table 1). Although the role of hormones on abscission remains to be clearly defined and large-scale analyses of transcription profiles have demonstrated a huge network of hormone interactions in signalling events (Nemhauser ), it is generally accepted that ethylene operates as an activator while, instead, auxins act as retardants (Roberts ). In agreement with this supposition, a gene encoding an ethylene response factor was preferentially expressed in the LAZ, while several genes involved in auxin biosynthesis and auxin response were over-represented in the Pet.

Gene expression associated with ethylene treatment during the first phase of the abscission process preferentially found in the LAZ was mostly related to the biosynthesis of the hormone jasmonic acid and also to other non-hormone pathways such as protein kinase-mediated signal transduction (Table 1). On the other hand, jasmonic acid signal transduction has generally been associated with pathogen response regulation (Nandi ) and with other physiological processes such as flowering and senescence. It is also well documented that there is a cross-talk between ethylene and jasmonic acid signalling pathways and a positive feedback of ethylene on the jasmonic acid biosynthesis has been suggested (Sasaki ). Since methyl jasmonate (MeJa) has been reported to promote abscission in citrus (Hartmond ), these results suggest that ethylene may mediate the involvement of jasmonic acid biosynthesis during the first stages of citrus leaf abscission.

During the second phase of the process, over-representation of transcripts coding for S-adenosyl-L-methionine synthase and ACC synthase genes were found in the Pet, an observation probably related to a decrease in ethylene production concomitant with preferential expression of genes related to ethylene transduction in the LAZ. In this period, a gibberellin-regulated protein and an auxin-independent growth factor were also preferentially expressed in the LAZ while gibberellin-2-oxidase, a gene of gibberellin catabolism, and a putative auxin repressed protein were over-represented in the Pet. The initial events of ethylene-induced abscission in the LAZ of citrus leaf explants are apparently characterized by an active jasmonic acid biosynthesis while, during the second part of the process, ethylene, auxin, and gibberellin action are predominant in the Pet. Collectively, these results are compatible with the idea that ethylene, auxins, and gibberellins may function as defence mediators in the Pet during cell separation.

In addition to lipid transfer proteins (LTPs) the results revealed over-representation of a number of genes with potential transport activity in both LAZ and Pet (Table 1). Ion and sugar transport appeared to be preferentially expressed in the Pet whereas lipid transport in LAZ and genes belonging to protein or cell transport categories were activated in both tissues.

Ethylene exposure of citrus leaf explants induced preferential expression of a hexose transporter and a triose phosphate/phosphate translocator in the Pet consistently with previous observations that proposed an activation of sugar transport stimulated by mechanical wounding (Wilson and Lucas, 1988) and ethylene treatments. Recent reports have demonstrated that pathogen infections also are capable of activating the sugar transport (Azevedo ).

Lipid transport mediated by other proteins, namely ABC transporters, as well as LTPs appears to be very active also in LAZ (Table 1). Although ABC transporters have been related to a wide range of biological processes, little is known about their functions and roles (Campbell ). It is well documented, for instance, that they can translocate a large variety of substrates across extra- and intracellular membranes, including metabolic products, lipids, sterols, and drugs. Experiments carried out with ethylene Arabidopsis mutants for reception or signalling have suggested the implication of the hormone in the stimulation of certain ABC transporters (Campbell ). The results showed that two distinct ABC lipid transporter putative genes were preferentially expressed in the LAZ, one after 6 h of ethylene treatment and the other one during the second phase. Early gene expression of lipid transporters may be associated with phospholipid signal transduction events and/or with the characteristic increase of membrane systems that occurs in cells undergoing abscission (Addicott, 1982).

Transcription factors

Three transcription factors showed an interesting pattern of preferential expression during ethylene-induced abscission (Table 1). One of these ESTs representing a MYB transcription factor was preferentially expressed in the LAZ. MYB factors play a central role in the control of gene transcription involved in many processes including cell-separation. For instance, the involvement of AtMYB26 in the regulation of the swelling and lignification of the endothecium cell layer in the anther, which is essential to force a proper opening of the stomium and pollen release (Steiner-Lange ) has been demonstrated. Other transcripts preferentially expressed in the Pet encode a putative protein containing a histone-fold domain that may suggest histone modification and, therefore, a potential epigenetic regulation of the process and a NAC/jasmonic acid-related transcription factor. The putative Arabidopsis orthologue of this NAC gene, RD26 (AT4g27410), is known to be inducible by ABA and a number of stresses (Fujita ).

Conclusion

In this work, a large-scale transcriptional profile study on the ethylene-induced abscission of citrus leaf explants is presented. The data obtained provide a large catalogue of genes potentially involved in the regulation of the abscission process. The results also highlighted over-representation of specific transcripts in the abscission zone and the petiole of the leaf, enabling a first insight on the global modifications induced by ethylene during the activation and progression of abscission. These changes suggested the occurrence of major adjustments in cell wall and protein metabolism, hormone signal transduction, control of transcription, and lipid and sugar transport. In addition, the petiole probably undergoes a specialized defensive programme based on the generation of physical barriers and proper responses against ROS and pathogen attack.

Supplementary data

Supplementary data are available at JXB online.

shows morphological and anatomical features of the citrus leaf material used as source of RNAs for microarray studies of ethylene-induced abscission.

shows quantitative RT-PCR validation of microarrys.

contains sequences of primers used for the quantitative RT-PCR validation.

show expression ratio values for the genes presented in Figs 4–9.