Nội dung

Pirrello et al. BMC Plant Biology 2014, 14:341 http://www.biomedcentral.com/1471-2229/14/341 RESEARCH ARTICLE Open Access Transcriptional and post-transcriptional regulation of the jasmonate signalling pathway in response to abiotic and harvesting stress in Hevea brasiliensis Julien Pirrello1, Julie Leclercq1, Florence Dessailly1, Maryannick Rio1, Piyanuch Piyatrakul1,2, Kuswanhadi Kuswanhadi3, Chaorong Tang4 and Pascal Montoro1* Abstract Background: Latex harvesting in Hevea brasiliensis amounts to strong abiotic stress that can cause a halt in production in the most susceptible clones. Although the role of jasmonic acid has been suggested in laticifer differentiation, its role in latex production and in the response to harvesting stress has received very little attention. Only a few key genes acting in the COI-JAZ-MYC module have been isolated and studied at transcriptional level. Results: Use of a reference transcriptome obtained on rubber clone PB 260 covering a large number of tissues under different environmental conditions enabled us to identify 24 contigs implicated in the jasmonate signalling pathway in the rubber tree. An analysis of their expression profile by qPCR, combined with hierarchical clustering, suggested that the jasmonate signalling pathway is highly activated in laticifer cells and, more particularly, in the response to harvesting stress. By comparison with their genomic sequences, the existence of regulation by alternative splicing was discovered for JAZ transcripts in response to harvesting stress. Lastly, positive transcriptional regulation of the HbJAZ_1405 gene by MYC was demonstrated. Conclusion: This study led to the identification of all actors of jasmonate signalling pathway and revealed a specific gene expression pattern in latex cells. In-depth analysis of this regulation showed alternative splicing that has been previously shown in Arabidopsis. Interestingly, genotypic variation was observed in Hevea clones with contrasting latex metabolism. This result suggests an involvement of jasmonate signalling pathway in latex production. The data suggest that specific variability of the JA pathway may have some major consequences for resistance to stress. The data support the hypothesis that a better understanding of transcriptional regulations of jasmonate pathway during harvesting stress, along with the use of genotypic diversity in response to such stress, can be used to improve resistance to stress and rubber production in Hevea. Keywords: Latex, Tapping panel dryness, Jasmonic acid, Alternative splicing, Rubber, Transcriptional regulation Background Jasmonic acid (JA) is a key hormone in the development of plant responses to abiotic stress, such as wounding [1]. This plant hormone also plays a key role in the biosynthesis of secondary metabolites [2]. It is particularly involved in the response to oxidative stress since it activates the ascorbate-glutathion cycle for the reduction of these major antioxidants [3,4]. * Correspondence: pascal.montoro@cirad.fr 1 CIRAD, UMR AGAP, F-34398 Montpellier, France Full list of author information is available at the end of the article The jasmonate biosynthesis and signalling pathways have been very widely studied and described in Arabidopsis using the analysis of mutants [5]. The jar1 mutant is deficient in certain responses to JA, including ascorbate production to prevent oxidative stress [6]. The JAR1 enzyme catalyses the conjugation of jasmonate to isoleucine (JA-Ile) [7]. JA-Ile is the bioactive form of JA [8]. It was demonstrated recently that COI1 is the JA receptor and can interact directly with JA-Ile [9]. For its part, the coi1 mutant also shows insensitivity to JA, thereby causing male sterility, resistance to the inhibition of root development by JA, and a defect in the expression of genes regulated by JA ? 2014 Pirrello et al.; licensee BioMed Central. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Pirrello et al. BMC Plant Biology 2014, 14:341 http://www.biomedcentral.com/1471-2229/14/341 [10]. COI1 is an F-box protein [11] forming, with some other partners, a complex of the E3 ubiquitin ligase type, SCFCOI1 [12,13]. This type of complex is involved in the ubiquitination of target proteins, leading to their subsequent degradation by the 26S proteasome [14]. JASMONATE ZIM DOMAIN (JAZ) proteins are recognized by the SCFCOI1 complex. JAZ degradation enables the release of transcription factors such as MYC2,3 and 4 [15]. The latter bind to cis-acting elements of the G-box type [16] present in the promoters of JA response genes and they initiate their transcription [17-19]. In addition to the MYC transcription factors, JAZ are able to interact with other transcription factors of the MYB, EIN3, EIL, ERF, GAI, RGA and RGL1 types, suggesting that they play a role in the interaction of the JA signalling pathway with those of other hormones [20,21]. In Arabidopsis, JAZ proteins comprise 12 members which are characterized by the presence in the C-terminal position of the highly conserved domains, Jas and ZIM (TIFY). The Jas domain enables interaction with COI1 and with transcription factors, while the TIFY domains involved in dimerization and in the interaction with NINJA [21-23]. At the moment, the most likely model concerning the repression of genes induced by JA is an interaction of JAZ with TOPLESS (TPL). This interaction may necessitate the presence of the NINJA protein as an adapter, unless the JAZ protein possesses an EAR domain (ERF-associated amphiphilic repressor) to which TPL binds [1]. The transcriptional regulation of JAZ can occur via MYC2 [16], but other components might be involved. Indeed, in the myc2 mutant, most JAZ were expressed following infection with Pseudomonas syringae [24]. The sub-unit of the mediator complex, MEDIATOR25 (MED25), was recently identified as an integrative node of JA-dependent gene expression [25]. At post-transcriptional level, most JAZ genes can be regulated by alternative splicing. In Arabidopsis, it has been found that intron/exon organization in the region of the Jas domain is conserved in the majority of JAZ genes [26]. In Arabidopsis, the spliced form of JAZ10 has a Jas domain that is partially (JAZ10.3) or totally truncated (JAZ10.4) [23]. Ectopic expression of JAZ10.3 and JAZ10.4 affords dominant insensitivity to JA as a consequence of the stabilization of the JAZ protein [19,23]. The physiological role of isoforms with a truncated Jas domain in attenuating hormonal response has been suggested by various studies [19,27]. Many studies have shown the importance of alternative splicing in plant development, but also in the response to stress (for review [28]), though that link has never been demonstrated in the case of post-transcriptional regulation of JAZ genes. Hevea brasiliensis is the only commercial source of natural rubber (NR). NR is synthesized in the rubber particles of latex cells. Those cells differentiate from the cambium then anastomose to form an articulated network: laticifers. Page 2 of 17 Latex is harvested by tapping the tender bark (phloem tissues). After latex flow, rubber particles agglomerate and clog the tapping zone. Latex is regenerated within a few days. In order to stimulate latex production 2.5% ethephon (an ethylene releaser) is applied to the tapping panel. Tapping and stimulation frequency is usually adapted to the metabolic activity of Hevea clones. In extreme cases of environmental or man-made stress due to rubber tree tapping, an oxidative burst occurs in the laticifers. That oxidative stress leads to peroxidation of lipids in the membrane of lutoids (poly dispersed lysosomal vacuome), which contain agglutinins such as heveins. The release of these factors leads to the in situ coagulation of rubber particles: this is the physiological syndrome known as Tapping Panel Dryness (TPD) [29]. TPD causes substantial NR production losses (10-40%) and economic models predict that NR production will not be sufficient by 2020. The genetic variability existing within the different cultivated rubber clones shows that the intrinsically most productive and fast growing clones are also the most susceptible to abiotic stress and TPD [30]. Some rubber clones with a slow laticifer metabolism are more tolerant of TPD. To date, few selected clones have the latter characteristics [31]. For instance, clone PB260 is a clone with an active metabolism that is highly susceptible to TPD, while INC53 and RRIM600 are clones with a slower metabolism and are more resistant to TPD. The physical wounding caused by tapping, the osmotic shock within the laticifers due to latex flow, and metabolic activation linked to latex regeneration between two tappings amount to harvesting stress which causes diverse responses, including the production of hormones such as JA. In fact, wounded tissues produce systemin, which induces JA production [32]. Interestingly, jasmonate and wounding are also involved in laticifer differentiation [33]. It was recently shown that jasmonate acts as a signal molecule in rubber biosynthesis [34]. Although the jasmonate signalling pathway has been studied in the rubber tree in connection with rubber biosynthesis and laticifer differentiation, little is known about its role in the response to harvesting stress. To date, only the ethylene biosynthesis and signalling pathways have been largely studied in this connection, and notably the transcriptional regulation of ERFs (Ethylene Response Factor) [35-39]. In Hevea, one or two members of the multigene families encoding COI [40], JAZ [41] and MYC [42] have been identified. Studies on the expression profile of those genes suggest the importance of JA in latex production. Indeed, HbCOI1 is strongly expressed in laticifers [40], the transcripts of HbJAZ1 accumulate in response to tapping and wounding [41], and HbMYC1 and HbMYC2 are more abundant in latex. HbMYC1 is induced by regular tapping and wounding, while the expression of HbMYC2 is not altered by those stimuli [42]. Pirrello et al. BMC Plant Biology 2014, 14:341 http://www.biomedcentral.com/1471-2229/14/341 In this study, the availability of Hevea transcriptomic and genomic resources made it possible to identify all the different genes acting in the jasmonate signalling pathway and characterize their implication during development, and in response to abiotic stress. Among the transcriptomes sequenced on different tissues (latex, bark, leaves and stem tips) [43-48], a reference transcriptome covering a large number of tissues and environmental conditions has been published for rubber clone PB 260 [35]. An analysis of the structure of JAZs genes based on the genomic sequences of rubber clone CATAS 7-33-97 (BIG-CATAS Project, unpublished data), and an analysis of the corresponding transcripts revealed the existence of a mechanism of alternative splicing of JAZ gene transcripts. Thus, this study suggested that the stress caused transcriptionally and post-transcriptionally by latex harvesting regulates the jasmonate signalling pathway in Hevea. Methods Plant material For the transcript analyses, plant material of clone PB 260 grew in accordance with the conditions described in Duan and coll. [35]. Samples corresponding to reproductive tissues (immature and mature, male and female flowers, zygotic embryos) were added to this study. Flowers were collected from 15-year-old trees. Zygotic embryos were collected from 5-year-old trees. In vitro plantlets of clone PB 260 were obtained by somatic embryogenesis with line CI07060. The plantlets were acclimatized and grown for 3 months in a greenhouse at a temperature of 27?C with 45% humidity. Several treatments mimicking abiotic stress were carried out, such as wounding, methyl jasmonate (MeJA), ethylene (ET) and dehydration at 1, 4, 8 and 24 hours. Drought stress response was controlled by following HbERFIVa3 (orthologue to DREB2A) transcript accumulation during this stress [49], whereas efficiency of wounding, ethylene and MeJA treatment were controlled using HbERF-IXc4 and HbERF-IXc5 (orthologue to ERF1) [37,50]. The bark was wounded every 0.5 cm by scarification with a razor blade. The ET and MeJA treatments were carried out by placing the plants in a 300-l open-door Plexiglas box overnight before treatment. Five parts per million of pure ET gas (1.5 ml/300 l) was injected into the sealed air-tight box. A concentration of 100 μl of liquid ≥95% MeJA solution (Sigma, St Louis, MO) was diluted in 500 μl of absolute ethanol and then placed on Whatmann paper inside the box for gas release. Control plants used for the ET and MeJA treatments were placed in the box and exposed to air only. The dehydration treatment was carried out by taking the plants out of their pots and leaving them to dry under laminar air flow. Each sample was collected 1 h, 4 h, 8 h and 24 h after treatment. RNA samples were collected and prepared at the CRRC, RRIT, Sanam Chaikhet District, Chachoengsao 24160, Page 3 of 17 Thailand (13.39?N latitude and 101.26?E longitude). These locations and our activities did not require any specific permission. The field studies did not involve endangered or protected species. Total RNA extraction Total RNAs were extracted from one gram of fresh matter using the caesium chloride (CsCl) cushion method adapted from Sambrook and coll. (Sambrook et al. [51]) by Duan and coll. (Duan et al. [37]). DNA contamination was checked by PCR amplification using primers of the actin gene. Primer design and transcript abundance analysis by qPCR Total RNA integrity was checked by electrophoresis. For each candidate gene a blast was performed against the transcriptome of PB 260, to define highly conserved regions. The qPCR primers were designed outside these regions using the Primer 3 module of Geneious Pro software version 5.3.6 (Biomatters Ltd., New Zealand). Each primer pair was blasted against the transcriptome library. qPCR and the fusion curve of the PCR amplicon were done using a mix of cDNAs. In addition, the specificity of each primer pair was checked by sequencing the PCR amplicon. The sequences of the primers used are listed in Additional file 1. cDNAs were synthesized from 2 μg of total RNA to a final reaction volume of 20 μl using a Revert AidTM MMuLV Reverse Transcriptase (RT) kit according to the manufacturer? s instructions (MBI, Fermentas, Canada). Full-length cDNA synthesis was checked for each sample by PCR amplification of the actin. Quantitative gene expression analysis was carried out by qPCR using a Light Cycler 480 (Roche, Switzerland) as described by Putranto and coll. [52]. Real-time PCR was carried out for eleven housekeeping genes in order to select the most stable gene as the internal control for all the compared tissues (HbelF1Aa, HbUBC4, HbUBC2b, HbYLS8, HbRH2b, HbRH8, HbUBC2a, HbalphaTub, Hb40S, HbUbi, HbActin). HbRH2b was selected as the best reference gene due to its stability in the different tested tissues [39] (Additional files 2 and 3). The transcript abundance of each gene was relatively quantified by normalization relative to the transcript abundance of the HbRH2b reference gene. All the normalized ratios corresponding to transcript accumulation were calculated automatically by Light Cycler software version 1.5.0 provided by the manufacturer. Heatmap representation was carried out on normalized and centered ΔCt values, using the heatmap2 function of the R software gplots package [53,54]. Statistical analyses qPCR reactions were carried out with 3 biological replicates. The statistical analyses were ANOVAs carried out Pirrello et al. BMC Plant Biology 2014, 14:341 http://www.biomedcentral.com/1471-2229/14/341 on raw data after logarithmic transformation. An ad-hoc Tukey HSD (Honestly Significantly Different) test was carried out for the analysis of transcripts in the different organs and different clones. Values with the same letter were not significantly different. For the analysis of transcripts in response to abiotic stress and to hormone treatments, a comparison of means test (Student or Wilcoxon test, depending on the data) was carried out between the control and the treatment at each point of the kinetics. If, at one point, at least one significant difference was found an ? s? was indicated, otherwise ? ns? was indicated? . Phylogenetic analysis of JAZ genes Multiple alignment was carried out on the total protein sequences of the JAZs of Arabidopsis and Hevea. Alignment with Gblocks software [55] led to the isolation of conserved blocks at least 10 amino acids long. The blocks were then used to construct the phylogenetic tree using PhyML software [56] which implements a maximum likelihood tree reconstruction method using the LG + gamma model, starting from a BioNJ tree [57]. A RAP-Green analysis was carried out using the original tree from PhyML to predict duplications [58]. The final tree was visualized with the Archaeoptheryx program [59]. Test of transcriptional activity by transient expression in BY-2 cells of tobacco Transactivation experiments were carried out following the procedure described by Chaabouni and coll. [60]. The pJAZ_1405(−267) and pJAZ_1405(−955) promoters were cloned to the pMDC107 vector [61], thereby controlling the GFP reporter gene. The MYC_771 and MYC _94937 genes were cloned under control of the 35S promoter to the pMDC32 vector. Transactivation assay was carried with several controls. First, a window excluding debris and define auto-fluorescence was defined from protoplast solution without transformation. Second, a negative control was obtained by protoplast transformation with the empty reporter vector and effector vector. Third, protoplast transformation with two vectors pJAZ and pHbMYC was carried out. The negative control is used as reference for the calculation of transactivation activity as follows: (pJAZ_1405(? 955) + pHbMYC)/(pJAZ_1405 (? 267) + pHbMYC). This method avoids any risk of disturbance due to the gateway cassette present in pMDC32 vector. The primer sequences used for gateway cloning are listed in Additional file 4. Transformation was replicated 6 times independently. After 16 h, GFP expression was quantified by flow cytometry (FACS Calibur II instrument, BD Biosciences) on the Montpellier Rio Imaging (MRI) platform. Data were analysed using Flowing Software version 2.5.0. Page 4 of 17 Accession numbers Reads from this library are those published by Duan and coll. (Accession: PRJNA235297 ID: 235297) available on NCBI database. Results Identification of the different genes acting in the jasmonate signalling pathway The different genes acting in the jasmonate signalling pathway were identified from the ? comprehensive transcriptome? published by Duan and coll. [35] using TBLASTN with the corresponding Arabidopsis sequences as the query. For each gene, we selected the one that came out with the best score. For each gene identified in that way, we checked for the conservation of the domains characteristic of each family using INTERPROSCAN. We thus identified 6 contigs corresponding to JAR, 2 for COI, 10 for JAZ and 3 for MYC. The JARs belonged to the GH3 multigene family [62]. GH3s are generally characterized by the presence of 3 small conserved motifs [63]. Motif III was found to be highly conserved in the 6 Hevea sequences. Motifs I and II were found in JAR_5108, JAR_14894 and JAR_59958, while JAR_20347 and JAR_20244 did not display motif I. The absence of motif I in the N terminal part of the proteins deduced from the JAR_20347, JAR_21367 and JAR_20244 contigs was linked to the fact that the contig sequences were incomplete at 5′ (Additional file 5). An analysis of the protein sequences revealed that the two HbCOI, HbCOI_2304 and HbCOI_3058, were characterized by the presence of an F-box domain, along with 3 leucine-rich repeat (LRR) domains (Additional file 6). All of the JAZ proteins identified had the 2 characteristic domains, TIFY and Jas (Additional file 7) [64]. HbMYC_424, HbMYC_771 and HbMYC_94937 all had a bHLH DNA binding domain, along with the 4 conserved regions (I to IV) (Additional file 8) [65]. An analysis of the transcriptome of rubber clone PB260 also enabled us to identify the partners HbNINJA_6328 (Additional file 9) and HbTPL_7591 (Additional file 10). HbNINJA displayed 46% of identical residues with At4g28910, while HbTPL displayed very strong homology with At1g15750 as 85% of the amino acids were identical. We were also able to identify an orthologue of AtMED25, HbMED25_16787, which displayed 50% identity (Additional file 11). Expression profile of the different genes acting in different developing tissues In order to characterize the JA signalling pathway in different Hevea tissues, the transcript abundance of all the genes identified was analysed by qPCR in zygotic embryos (embryo body and cotyledon), roots (TR: taproot and SLR:secondary lateral root), bark, leaves, latex, and male and female flowers at the immature and mature stages (Figure 1). Heatmap representation combined with Pirrello et al. BMC Plant Biology 2014, 14:341 http://www.biomedcentral.com/1471-2229/14/341 Page 5 of 17 8 4 0 Count Color Key and Histogram -2 -1 0 1 2 Row Z-Score Male mat Fem mat Fem imm Male imm Latex Leaf SLR Bark TR MFE. body MF. cotyl JAZ_14313 MED25_16787 JAZ_19967 JAZ_26925 JAZ_2001 COI_3058 JAZ_863 JAR_5108 JAR_14894 JAZ_1229 COI_2304 JAZ_1660 JAZ_1405 JAZ_17062 MYC_771 MYC_424 NINJA_6328 TPL_7591 JAR_21367 JAZ_29511 JAR_20244 JAR_59958 MYC_94937 JAR_20347 Figure 1 Heatmap representation of the expression of genes acting in the jasmonate pathway in different rubber tree tissues. The data obtained by quantitative RT-PCR corresponded to the levels of the actor transcripts in total RNA samples extracted from mature fruit cotyledon (MF. cotyl), mature fruit embryo bodies (MFE.body), taproots (TR), bark, secondary lateral roots (SLR), leaves, latex, immature male flowers (Male imm), immature female flowers (Fem imm), mature female flowers (Fem mat), and mature male flowers (Male mat). The data presented correspond to averages of 3 independent biological replicates. The red and green colours correspond to low and high gene expression, respectively. The heat map was generated using R software. an analysis of variance (ANOVA) was used to class the genes according to their level of expression in each tissue (Additional file 12). In addition, hierarchical clustering was used to identify tissues that displayed the same transcriptional signature. The different parts of the embryo displayed the same transcriptional landscape. Overall the jasmonate-related genes implicated in flowers were those that were not expressed in the embryo. Male and female flowers, whether mature or immature, were grouped together, suggesting that the same jasmonate-related genes were involved, whatever the flower gender and development stage. The transcriptional signature in the taproot (TR) was grouped with the embryo, while most of the genes expressed in the secondary lateral roots (SLR) were also expressed in bark. Surprisingly, latex displayed an original profile, with a large number of over-expressed signalling genes (JAZ_14313, COI_3058, COI_2304, JAZ_1660, JAZ_1405, JAZ_17062, MYC_771, MYC_424, MYC_85795, MYC_94937). These results confirmed the importance of the jasmonate pathway in latex. In addition, it was interesting to see that TPL_7591 and NINJA_6328, which, a priori function together, were in the same cluster. MED25_16787 had an expression profile similar to that of JAZ_14313. Regulationby abiotic stress of genes acting in the jasmonate pathway The relative transcript abundance was monitored over time for the 24 contigs corresponding to the genes of the jasmonate signalling pathway in response to wounding, dehydration, ethylene and methyl jasmonate (MeJA) (Figure 2). In the aim to validate efficiency of treatments, Hevea orthologous of Arabidopsis genes, known to be regulated by wounding, dehydration, ethylene and MeJA were used as positive control. Orthologous to DREB2 and ERF1 Pirrello et al. BMC Plant Biology 2014, 14:341 http://www.biomedcentral.com/1471-2229/14/341 Figure 2 (See legend on next page.) Page 6 of 17 Pirrello et al. BMC Plant Biology 2014, 14:341 http://www.biomedcentral.com/1471-2229/14/341 Page 7 of 17 (See figure on previous page.) Figure 2 Expression pattern of the jasmonate transduction pathway members in Heveaplantlets in response to abiotic and hormonal stress. The data obtained by quantitative RT-PCR corresponded to the transcript accumulation of genes acting in the JA signalling pathway in response to wounding, dehydration, ethylene and methyl jasmonate (MeJA) in stems of 3-month-old plantlets. RNAs were collected after 1 h, 4 h, 8 h and 24 h of treatment.,Gene expression values were normalized using Rh2B as internal control, and then, ratio between treated and untreated plant have been calculatedfor each time point. A comparison of means test was carried out to compare the transcript abundance in stressed plants and control plants. Significant changes were indicated by ? s? above error bars. (A) Expression pattern of jasmonate pathway genes in control plant. (B) Expression pattern of jasmonate pathway genes in response to wounding. Expression pattern of HbERFIXc4and HbERFIXc5 have been tested to validate efficiency of wounding treatment. (C) Expression pattern of jasmonate pathway genes in response to dehydration. Expression pattern of HbERFIV-a3 has been tested to validate efficiency dehydration treatment. (D) Expression pattern of jasmonate pathway genes in response to ethylene. Expression pattern of HbERFIXc4and HbERFIXc5 have been tested to validate efficiency ethylene treatment. (E) Expression pattern of jasmonate pathway genes in response to jasmonate. Expression pattern of HbERFIXc4and HbERFIXc5 have been followed to validate efficiency jasmonate treatment. The data presented correspond to the meanof 3 independent biological replicates. were identified in Hevea by Piyatrakul and coll. [37]. Expression of the DREB2 genes is induced by drought stress and is involved in drought-responsive gene expression [49]. HbERF-IVa3 is an orthologous of DREB2A and DREB2B [37]. ERF1 is induced by various stresses including wounding and hormonal treatments, acts at the crosstalk of jasmonate and ethylene signalling pathways. Its two putative orthologous genes in Hevea, HbERF-IXc4 and HbERF-IXc5 [37], were used to validate wounding, ethylene and jasmonate treatments in this study. The transcript abundance of all the genes of the jasmonate signalling pathway was modified between 1 h and 24 h, whatever the treatment. The transcripts of JAR_14894, JAR_20347, JAR_5108 accumulated significantly in response to wounding, while the transcript abundance of JAR_20244 was significantly lower than in the control (Figure 2A,B). Only the transcript of JAR_5108 accumulated significantly in response to dehydration. For each type of stress, at least one transcript of each partner of the COI-JAZ-MYC module was highly activated (Figure 2C). Interestingly, several genes acting in the signalling pathway were also regulated by ethylene (Figure 2D). Statistical analyses showed that JAR_5108, JAZ_863, JAZ_1405, JAZ_2001, JAZ_14313, JAZ_26925 and MYC_94937 were regulated by ethylene. It should be noted that, among the JAR and COI, only the transcripts of JAR_5108 and COI_3058 were significantly regulated by JA (Figure 2E). Of the 10 JAZ genes studied, the transcripts of eight of them were significantly accumulated in response to JA (JAZ_863, JAZ_1229, JAZ_1405, JAZ_1660, JAZ_14313, JAZ_19967, JAZ_26925, JAZ_29511). The transcripts of MYC_771 and MYC_94937 were significantly accumulated in response to JA. TPL_7591, NINJA 6328 and MED25_16787 were regulated overall in the same manner. In response to wounding, their transcript abundance increased rapidly in 1 h, while in response to ET and MeJA, they only accumulated after 24 h. Taken together, these results suggest that the genes involved in JA signalling are expressed in response to abiotic stress, but it was not always the same JARs, COIs, JAZs and MYCs that were involved depending on the stress in that signalling. Genotypic regulation of genes involved in the JA pathway in response to harvesting stress in mature trees The transcript abundance of genes belonging to the COI-JAZ-MYC module were studied in the latex and bark of tapped rubber trees (Figure 3, Additional file 13). During harvesting, the transcripts of the COI-JAZ-MYC module were regulated in both latex and bark. Our results showed that the expression level for COI_2304 was significantly higher in latex than in bark (p < 0.01). In latex, the transcripts of COI_2304 were accumulated in response to tapping, while in bark the transcript abundance decreased in response to that stress (p < 0.01). Treatment with ethephon tended to reduce the transcript abundance of COI_3058 (p < 0.01) in latex and in bark. The results suggested that the transcript regulation of JAZ genes was specific to each gene and that no general tendency could be brought out for the family as a whole. For example, JAZ_1229, JAZ_2001 and JAZ_19967 were significantly more expressed in bark than in latex, while the reverse was true for JAZ_1405, JAZ_1660, JAZ_17062 and JAZ_29511. In latex, JAZ_1405 was the only JAZ to be significantly repressed by tapping (p < 0.01) (Additional file 13), while JAZ_863 (p < 0.01), JAZ_1660 (p < 0.01), JAZ_17062 (p < 0.01), JAZ_29511 (p < 0.01) and JAZ_26925 (p < 0.01) were induced by tapping when there was no ethephon treatment. Interestingly, the transcripts of JAZ_17062 and JAZ_26925 were accumulated in an opposite manner in latex and bark in response to tapping. In bark, JAZ_863 (p < 0.01), JAZ_1229 (p < 0.01), JAZ_19967 (p < 0.01), JAZ_1660 (p < 0.01), and JAZ_29511 (p < 0.01) were induced by tapping, while the expression of JAZ_2001 was repressed by tapping (p < 0.01). In the majority of cases, ethephon reduced JAZ transcript abundance. However, in latex, the transcripts of JAZ_29511 and JAZ_863 were accumulated in the presence of ethephon (p < 0.01 and p < 0.01, respectively). In bark, when the tree was Pirrello et al. BMC Plant Biology 2014, 14:341 http://www.biomedcentral.com/1471-2229/14/341 Page 8 of 17 Figure 3 Regulation of COI-JAZ-MYC module members in Hevea in response to harvesting stress in latex and in bark. The accumulation of each COI-JAZ-MYC member was studied in the latex (dark grey bar) and bark (light grey bar) of adult trees, during latex production. During latex production the trees were tapped (3,4) every 2 days, or not (1,2), and stimulated by 2.5% ethephon, once per month (2,4), or not (1,3). The data presented correspond to averages of 3 independent biological replicates. tapped, the transcripts of JAZ_26925 were accumulated in the presence of ethephon (p < 0.01). The 3 MYC factors tested displayed the same expression profile. In fact, they were expressed more in latex than in bark and were strongly induced by tapping. Together, these results suggested that JA played a role during latex harvesting. Clonal regulation of genes acting in the jasmonate signalling pathway In order to find out whether the jasmonate signalling pathway might be involved in the differences in laticifer metabolism observed at clonal level, we studied the transcript abundance of each of the genes acting in the pathway in clones PB260, INC53 and RRIM600. Of the 24 genes studied, only 8 displayed a transcript abundance that was significantly different between clones (Figure 4, Additional file 14). JAR_59958 and MYC_771, positive regulators of the JA signalling pathway, were highly expressed in INC53 and RRIM600. On the other hand, the transcripts of the JAZ negative regulators were more expressed in clone PB260. These results showed that the positive regulators of the JA pathway are preferentially accumulated in Pirrello et al. BMC Plant Biology 2014, 14:341 http://www.biomedcentral.com/1471-2229/14/341 Page 9 of 17 Figure 4 Expression of genes acting in the jasmonate pathway in 3 Hevea clones. The transcript abundance of the 24 identified genes was studied by qPCR on latex from PB 260, INC 53 and RRIM 600. According to an analysis of variance 8 transcripts were significantly over-represented in one or two clones (Additional file 14). The data presented correspond to averages of 3 independent biological replicates. Different letters correspond to significantly different values after theTukey HSD test (p < 0.05). clones with a slow laticifer metabolism, while the negative regulators are accumulated more in clones with a rapid laticifer metabolism. Conservation of the structure HbJAZ genes and that of their Arabidopsis orthologue A study of the phylogeny and structure of the JAZs of Hevea was carried out on 7 JAZs (JAZ_1660, JAZ_19967, JAZ_1405, JAZ_29511, JAZ_26925, JAZ_863, JAZ_1229) which were of interest because they are regulated both by JA and by tapping. A phylogenetic analysis based on the protein sequences of the HbJAZ sand AtJAZs revealed an organization primarily in three sub-classes B, C and D previously defined by Chung [26], with sub-class A not being represented by any JAZ member of Hevea (Figure 5A). AtJAZ10, previously assigned to sub-class C, revealed stronger homology with HbJAZ_29511 to form a new sub-class that we called sub-class E. HbJAZ_26925 was not grouped in the previously described sub-classes and formed a new sub-class. The gene structure of the 7 JAZs of interest was analysed using the scaffold sequences corresponding to the genome of rubber clone CATAS 7-33-97 (Figure 5B). Most rubber tree JAZs possess an intron in the Jas domain, which is similar to what has been described in Arabidopsis. The sequence analysis predicted that retention of the 5′ region of the Jas intron splicing site during pre-mRNA processing of JAZ_1660, JAZ_29511 and JAZ_26925 generated a premature termination codon immediately or very rapidly (Figure 5C). By using both primers overlapping the splicing zone and primers inside the intron, it was possible to confirm alternative splicing by sequencing of the qPCR amplicons for JAZ_1229, JAZ_863, JAZ_1660, JAZ_29511 and JAZ_26925. Alternative splicing of HbJAZ is regulated by tapping stress In order to gain a clearer understanding of JAZ posttranscriptional regulation in response to tapping stress in rubber trees, we used qPCR to measure the spliced Pirrello et al. BMC Plant Biology 2014, 14:341 http://www.biomedcentral.com/1471-2229/14/341 Page 10 of 17 Figure 5 HbJAZ gene structure is conserved with the Arabidopsis orthologue. (A) Phylogenetic tree constructed from the amino acid sequence of 12 full-length AtJAZs and 7 full-length HbJAZ, showing five (A? E) subclades of proteins. (B) Intron/exon organization of the corresponding genes. Thick green and blue bars indicate coding regions and non-coding untranslated regions in exons, respectively. The Jas motif coding region is depicted in red. Introns are depicted by a thin black horizontal line or, in the case of the Jas intron, a red line. (C) Sequence of the Jas exon/intron junction in five HeveaJAZ genes containing the intron. Exon sequences are highlighted in black with the predicted amino acid sequence. Illustration adapted from [26]. form/non-spliced form ratio of transcripts in the latex and bark of tapped or untapped rubber trees, treated or not with ethephon (Figure 6, Additional file 15). Despite the existence of non-spliced forms for the transcripts of JAZ_29511, JAZ_863 and JAZ_26925, this was not detected by qPCR in the samples tested. The alternative splicing of JAZ_1229 and JAZ_1660 was significantly regulated by tapping and by ethephon treatment. Our results suggested opposite regulation between the splicing of JAZ_1229 and JAZ_1660. In fact, tapping induced the non-spliced form of JAZ_1229, suggesting an increase in the repressive form of that JAZ, while the same stress induced the spliced form of JAZ_1660. MYC_771 and MYC_94937 regulate the transcription of JAZ_1405 The transcripts of certain JAZs were not subjected to alternative splicing like JAZ_1405. Hierarchical clustering