Natural rubber accounts for 42.3% of the 23.9 million tons of rubber consumed worldwide . Hevea brasiliensis is the sole commercial source of natural rubber. The increasing demand for natural rubber calls for improved productivity in rubber plantations. Cis-1,4-polyisoprene chains are synthesized in the rubber particles of latex cells. Rubber particles account for up to 90% of the dry matter in latex cytoplasm, which is harvested by tapping the soft bark of rubber trees . Latex production depends on genetic, environmental and harvesting components. Harvesting systems use ethephon, an ethylene (ET) releaser applied to the tapping panel, to stimulate latex production by improving the flow and regeneration of latex. Tapping and ethephon stimulation frequencies are adjusted to Hevea clones according to their metabolism . Given the high pressure in the phloem tissue, latex is expelled after tapping. Tapping and ethephon are likely to be sources of stress conducive to the production of secondary metabolites and consequent rubber, but over a certain stress limit they lead to tapping panel dryness (TPD) . Mechanical wounding and osmotic stresses related to tapping trigger the production of endogenous ethylene and oxylipins such as jasmonic acid (JA) [5, 6]. Both mechanical wounding and methyl-jasmonate treatments induce the differentiation of secondary latex cells in extended young stems [7–9]. In trees, secondary latex cells are differentiated from cambium and then anastomosed to create laticifer vessels . Ethephon application also induces several biochemical processes in laticifers, such as sucrose loading, water uptake, nitrogen assimilation or synthesis of defence proteins [11, 12], involving a large number of ethylene-response genes [13–18], whereas its direct role in rubber biosynthesis is controversial .
Given the major role of ethylene and jasmonic acid in regulating latex cells, Ethylene-Response Factors are greatly expected to be involved in latex cell functioning. Indeed, ET and JA signalling pathways involve transcription factors such as Ethylene-Responsive Element Binding Proteins (EREBP), also called the Ethylene-Response Factors (ERF) family . ERFs have been shown to act as activators or repressors of additional downstream ethylene responsive genes. ERFs are a control point for crosstalk with other signals and they function as an integrator of the ethylene and jasmonic acid pathway . Multiple signalling pathways converge on ERFs by transcriptional and post-transcriptional regulation . Ethylene and jasmonate pathways converge in the transcriptional activation of ETHYLENE RESPONSE FACTOR1 (ERF1), which regulates in vivo the expression of a large number of genes responsive to both ethylene and jasmonate. ERF1 acts downstream of the intersection between the ethylene and jasmonate pathways suggesting that this transcription factor is a key element in the integration of both signals for the regulation of defence response genes [23, 24]. In biotic stress, AP2/ERF transcription factor ORA59 acts as the integrator of the JA and ET signalling pathways and is the key regulator of JA- and ET-responsive PDF1.2 expression [21, 25].
The ERF family was first discovered in Nicotiana tabaccum by Ohme-Takagi and Shinshi . The ERF family is one of the largest families of transcription factors with 122 genes out of the 2016 predicted transcription factors from 58 families in Arabidopsis[26, 27]. The ERF family belongs to the AP2/ERF superfamily. This superfamily encodes transcriptional regulators that serve a variety of functions in plant development and responses to biotic and abiotic stimuli [28–30]. Members of the AP2/ERF superfamily contain at least one AP2 domain, which consists of about 60 amino acids. This domain is involved in DNA binding to a conserved AGCCGCC sequence called the GCC-box [20, 31] or to a dehydration response element (DRE: TACCGACAT) containing the C-repeat [32, 33]. The structure of the AP2 domain was first reviewed by Riechmann and coll. . Initially, the APETALA2 (AP2) gene was isolated by T-DNA insertional mutagenesis in Arabidopsis. This gene encodes a 432-amino acid protein with two copies of a 68-amino acid direct repeat called the AP2 domain. The AP2 domain consists of three anti-parallel β-sheets and one α-helix. Two conserved elements, YRG and RAYD, have been identified. The latter is an 18-amino acid core region that is predicted to form an amphipathic α-helix . In some AP2 genes, the AP2 domain contains a 37-amino acid serine-rich acidic domain putatively functioning as an activation domain, and a 10-amino acid domain including a putative nuclear localization sequence KKSR . While previously thought to be plant-specific transcription factors, AP2 domain-containing genes were recently found in bacteria and viruses, which are predicted to be HNH endonucleases [37, 38].
Several ways of classifying the AP2/ERF superfamily have been proposed in plants. Although all of them are based on the number of AP2 domains, some differences exist. Firstly, Sakuma et al. described five subfamilies including AP2, RAV, Dehydration Responsive Element Binding Proteins (DREB), Ethylene-Responsive Element Binding Proteins (EREBP), also called the Ethylene-Response Factors (ERF) family , and others based on a homology of the DNA binding domain, and the DNA sequences that bind it, namely the DRE element or GCC-box separately . The AP2, ERF/DREB and RAV subfamilies have two AP2 domains, one AP2 domain, or one AP2 and one B3 domain, respectively. Groups A1 to A6 and B1 to B6 have been assigned to the DREB and ERF families . Secondly, Nakano et al. classified these proteins in only three major families: AP2, ERF and RAV . The ERF family was then divided into ten groups according to the structure of the AP2 domain, with groups I to IV corresponding to the DREB subfamily in Sakuma’s classification. To date, Nakano’s classification method has remained a reference for organizing the AP2/ERF superfamily in three families (AP2, ERF, RAV) and the ERF family in ten groups. In the construction of phylogenetic trees, methods for multiple sequence alignment and tree reconstruction have to be considered with caution. In the analyses by Sakuma and Nakano, ClustalW followed by a neighbor-joining method was chosen. Currently, although computationally intensive, the multiple sequence alignment software MUSCLE followed by a maximum likelihood method (PhyML) is more relevant [40–44].
The availability of the whole genome sequence of several plant species has made it possible to confirm a relatively well-conserved organization of the AP2/ERF superfamily with 147, 149, 202, 180 and 146 genes in Arabidopsis thaliana, Vitis vinifera, Populus trichocarpa Oryza sativa, and Solanum lycopersicon, mostly represented by the ERF family [27, 28, 45–47]. Transcript sequencing is also an alternative for identifying such gene families. For instance, 156 AP2/ERF genes consisting of 148 ERFs, 4 AP2s and 4 RAVs were identified in Gossypium hirsutum from EST databases . In Hevea brasiliensis, transcriptome sequencing has been carried out on latex, bark, leaf and shoot apex tissues using various methods [49–54]. A few number of ERF sequences have been released in the Genbank (HbEREBP1: HQ171930.1; EREBP2: HQ171931.1; DREB1p: HQ902146.1; CBF1: AY960212.1) . As preliminary experiment, we identified AP2/ERF unigenes from latex and leaf tissues of the Hevea clone Reyan 7-33-97 members . This analysis revealed 45 AP2/ERF with partial AP2 domain that did not allow gene classification (Additional file 1). Given the involvement of wounding, jasmonate and ethylene in natural rubber production, we developed a reference transcriptome that covers a large number of tissues and environmental conditions to have a fully comprehensive Hevea transcriptome and we examined the organization of the AP2/ERF superfamily in Hevea. Firstly, RNAs were isolated from different tissues of plants at several stages of development growing under various conditions, and transcripts were sequenced using GS-FLX next-generation sequencing (NGS) technologies. Secondly, contigs harbouring at least one AP2 domain were identified in tissue-type libraries for leaves, bark, latex, roots and embryogenic tissues and from a global library which pooled reads from all tissue-type libraries. AP2 domain-containing genes were aligned with the Arabidopsis AP2/ERF sequences and classified according to Nakano’s method based on a phylogenetic analysis of the conserved AP2 domain, which was optimized using a maximum likelihood method (PhyML) [40–42]. Post-transcriptional regulation was checked by predicting microRNA-targeted AP2/ERF genes. This study suggested that some HbAP2/ERF genes expressed in latex cells could be involved in specific biological processes.