1. Introduction
The European sweet chestnut (
Castanea sativa) is a profitable and cost-effective woody species in the Mediterranean basin [
1,
2], highly relevant for biomass, timber and fruit production [
3,
4]. Chestnut trees have a great ecological value, as they have been related to an increase in plant species richness and they are important carbon sinks aiding to mitigate climate change [
5,
6]. Besides, chestnut trees are involved in the maintenance of traditional landscapes, significantly contributing to the environmental and cultural heritage [
7,
8]. Nonetheless, potential applications are hindered by its recalcitrant behaviour, thus driving to the use of biotechnological approaches to overcome this limitation. Selected chestnut genotypes have been successfully micropropagated by using both juvenile and mature material, although acceptable rooting rates have only been attained in juvenile-like material (reviewed in [
9]). However, rooting ability is also greatly influenced by the genotype, thus limiting the pool of available material for the vegetative propagation of this species.
Adventitious rooting (AR) is a complex post-embryogenic process through which roots are formed in tissues such as stems, leaves and hypocotyls, from cells not previously determined to form roots. AR is modulated by several internal factors such as mother plant status, the chronological and ontogenetic age, the genetic makeup, the hormonal balance, as well as external conditions like temperature, light or mineral nutrition [
10,
11,
12]. This process is a common occurrence in plant development and can be triggered as a response to different stressors such as flooding, nutrient deficiency, wounding or oxidative stress [
13,
14]. The formation of adventitious roots (ARs) is a key step in the vegetative propagation of plants. Many species, including several trees, exhibit a recalcitrant behaviour and their ability to form roots decreases dramatically with age as they go through the phase change from the juvenile to the mature stage, with low responses to rooting stimuli [
11,
15,
16]. Reversion of mature-related traits, particularly the improvement of AR, represents a challenge for the propagation of species whose desirable traits are only seen in the mature stage. The lack of rooting or the deficiencies in root architecture reduce the survival rate of plants and hamper the mass propagation of selected genotypes, causing heavy economic losses for the related industries.
AR is divided into three sequential stages, named induction, initiation and expression (outgrowth). In chestnut, the most limiting stage is the induction phase, in which specific cells respond to the stimuli and initiate a root developmental program in root-prone tissues/genotypes, takes place in the first 24-48 hours after the beginning of the induction treatment [
9,
17]. Every step in the AR process is subject to a dynamic and specific hormone modulation, although auxin is the key hormone in AR [
18]. The decline of rooting ability after the transition from the juvenile to the mature stage is linked to changes in auxin homeostasis [
11,
16], modifying the expression of auxin-responsive genes [
19,
20]. However, other plant hormones are involved in the regulation of AR. Plant growth regulators such as cytokinins [
21] or gibberellins [
22] seem to inhibit AR, whereas jasmonates [
23,
24] or abscisic acid [
25] are ARs stimulators in some species. Nonetheless, the role played by different hormones seems to have a species-specific component, while at the same time their function might be dependent on the particular stage of the process.
Ethylene (ET) is a gaseous plant growth regulator involved in different plant developmental processes, including close links to plant aging and phase change [
26,
27,
28]. Moreover, ET has been shown to interact with auxin in primary root development and other rooting processes [
29,
30]. In chestnut, a transcriptomic analysis revealed that ET signalling is upregulated in mature tissues in response to auxin and wounding when compared to juvenile tissues. This finding suggests that a high concentration of ET in mature tissues could be one of the causes underlying its recalcitrant behaviour [
20]. ET biosynthesis and signalling pathways have been properly characterized in plants, leading to the development of specific strategies that allow for their modulation. For instance, ET biosynthesis might be enhanced or reduced by the application of its precursor 1-aminocyclopropane-1-carboxylate (ACC) or aminoethoxyvinylglycine (AVG), respectively. On the other hand, the use of silver ions allows for the inhibition of ET perception (reviewed in [
31]). To test the effects of ET on the root system development of chestnut, ACC, silver nitrate (AgNO
3) and AVG were applied to the root induction media of juvenile and mature chestnut microshoots. Results obtained prompted us to further characterize the responses in mature chestnut tissues, where specific gene expression related to ET synthesis and signalling, auxin responses and epigenetic processes was evaluated by quantitative PCR. The results suggest a negative effect of ET on the induction of ARs in mature tissues and a link to recalcitrance, which might at least partially exert its action through the interaction with auxin transport and specific epigenetic processes.
3. Discussion
The successful development of ARs requires a change in the fate of specific cells not previously determined to form a meristem that will eventually drive the generation of the new organ [
32]. In the case of rooting-competent chestnut microshoots, cells close to the vascular bundles in the stem are able to respond to the rooting stimuli, auxin and wounding, then reprogramming their ongoing genetic pattern and entering a root developmental pathway [
9]. Therefore, the plasticity of those rooting-competent cells allows them to modify their gene expression patterns in response to wound stress and external auxin supply and switch their fate to AR founder cells. However, this ability is drastically reduced during maturation, with adult microshoots showing a recalcitrant behaviour that severely hampers their ability to form roots. Despite recent advances, the molecular basis of this recalcitrant behaviour is still poorly understood. Several players are believed to take part in this connection between AR and recalcitrance, particularly hormones, epigenetic mechanisms and their crosstalk, which may underlie developmental plasticity in plants [
33]. A recent analysis showed that transcriptomic responses to auxin and wounding vary greatly between the two types of microshoots used in the present study [
20]. Among the differences found, mature shoots exhibited an increased response related to ET, including biosynthesis-related genes and ET-responsive transcription factors. To gain deeper insight into the putative role of ET in the modulation of AR in chestnut and its relation to recalcitrance, treatments were designed where the ET content was either increased or decreased, or its perception was blocked. In juvenile shoots, no significant phenotypical changes were detected, suggesting that ET does not influence AR in these shoots. However, wounding is a necessary step for AR induction in cuttings and microshoots [
9,
10], and injury-related stress induces a temporary increase in ET and jasmonic acid (JA). These two hormones develop an antagonistic relation in which JA stimulates the expression of wound-responsive genes, while ET blocks that expression in order to locally and temporarily restrict the repairing response [
34]. However, the lack of ET effect in P2BS shoots suggests that other mechanisms might be active in these tissues to drive tissue repair. On the other hand, ET modulation severely impacted the rooting response of mature shoots, including rooting rates, root number and root length. Therefore, our results suggest that repairing mechanisms and root induction processes might be different according to the ontogenetic state of the tissues.
ET has been shown to present contrasting effects in the formation of ARs in different species. It was described as an AR inhibitor in peach [
35], Eucalyptus [
36] and Malus x domestica [
37]. On the other hand, there are species such as petunia [
38], cucumber [
39], marigold [
40] or woody plants such as Pinus thunbergii [
41] or Citrus sinensis [
42] in which ET stimulates ARs formation. However, the effect of this hormone in AR has not been analyzed regarding to the ontogenetic state of the tissues in woody species. In Arabidopsis, ET-related signalling was shown to be more active in old leaves, which was linked to a low ability for de-novo root regeneration [
43]. Therefore, ET influence on regeneration processes and particularly on AR might be directly connected to the age and development of the plants.
A remarkable effect of ET inhibition in P2BS and P2CR shoots was the ability to induce lateral roots (
Figure 1), a trait that improves root system performance. The formation of these lateral roots took place long after the AgNO
3 and AVG treatments ended, thus suggesting a lasting effect on microshoots performance. The inhibition of lateral root formation by ET was also described in Arabidopsis [
44,
45] and tomato [
46]. The possible mechanism of this effect in our system is unknown, although a role for auxin transport modulation might be a plausible explanation (see below).
According to the results of the present study, ET signalling is involved in the induction of AR, at least in mature shoots. ET content and signalling modulation were shown to directly influence genes involved in ET synthesis. Particularly, ACC-treated P2CR shoots showed a significant increase in the expression of CsACS1-like and CsACO1-like, and this increase was related to a lower rooting response. Surprisingly, ACC triggered a positive feedback loop in the expression of CsACS1-like, that lasted up to five days after the beginning of the treatment. This gene codes for an enzyme that catalyzes the conversion of S-adenosyl-l-methionine (SAM) into ACC, which is later transformed in ET by the activity of ACO enzymes. Thus, in mature shoots, ACC induced an increase in its own synthesis, whose presence seems to prevent specific developmental processes, at least in this system. In recent years, several reports have suggested that ACC might exert signalling effects on its own, not only due to its role as an ET precursor. Those activities have been related to developmental processes and stress responses, and putative transporters for ACC have been identified [
47,
48]. However, the ACC effect on the expression of ET-related genes suggest that this might be its main route of action in the system used here.
On the other hand, both IBA, IBA+AVG and IBA+AgNO
3 treatments reduced the expression of ET synthesis genes, particularly at 24h, thus indicating that lowering ET synthesis is necessary to induce rooting responses in mature shoots. Expression of CsEIN2, a core component of ET signalling, was modified according to treatments, with greater levels of expression in samples treated with the precursor of ethylene biosynthesis, while lower levels were detected when ET perception was inhibited. EIN2 acts downstream of ET receptors and modulates the activity of ET-Responsive transcription Factors (ERFs), also integrating inputs from other pathways in its expression. Here, CsEIN2 mRNA levels are clearly related to the rooting behaviour of mature shoots, with improved rooting responses by blocking ET signalling and thus decreased CsEIN2 expression. In Arabidopsis, an ACC+IBA treatment reduced the number of ARs, with the antagonistic relation between AtEIN2 and the JA signalling gene AtCOI1 showing to be particularly relevant in response to the IBA induction [
49]. Moreover, EIN2 also seems to control ET-related gene expression by inducing histone acetylation, in what is suggested to be a rapid transcriptional regulation process [
50]. Our data support this idea of a fast and dynamic control because CsERF3 showed a parallel expression pattern when compared to CsEIN2. However, previous reports suggested a positive role of ERF3 in the formation of ARs in Populus under normal and low-phosphorus conditions, which seems not to match results in our system, and a link to auxin signalling [
51,
52]. Probably, ET-related variations in the control of gene expression between both species underlie the differences found.
On the other hand, expression of the ET-responsive CsRAP2.12 transcription factor also slightly resembled CsEIN2 expression pattern at 24h. Previously, Valladares et al. [
53] analyzed the expression of this gene in chestnut and oak tissues during AR induction. Authors suggested that it might be implicated in the establishment of new developmental programs in an ontogenetic-related mode. This gene belongs to group VII of ERFs, whose activity has been linked to specific responses like low-oxygen conditions or oxidative stress [
54]. Here, levels of expression were lower than in T0 samples except for ACC+IBA treatment at 24 h, therefore its expression does not seem to relate to improved rooting responses.
ET and the key AR inducer, auxin, are known to interact in many processes and at different levels. Particularly, ET has been suggested to influence auxin movement through the modulation of the expression of auxin transporters [
55]. The activity of these transporters is essential for the generation of hormone gradients in the tissues, which eventually trigger the process of AR [
56]. ET-perception and synthesis inhibition dramatically induced the expression of CsPIN1 at 24h, while its induction was much more modest in IBA-treated shoots. Thus, ET seems to block auxin transport in mature chestnut tissues, preventing the establishment of hormone gradients needed for the induction of specific developmental responses. The activity of PIN transporters seems crucial in the early steps of AR for the successful outcome of the process, as seen for example in apple and tea nodal cuttings [
57,
58]. In other experimental systems ET also seems to influence regeneration processes by influencing auxin distribution, as seen in de novo shoot organogenesis in tamarillo [
59].
In response to the establishment of auxin gradients, specific related gene expression is triggered. CsIAA29 resembled the expression pattern of CsPIN1 at 24h, when AR induction is taking place. Aux/IAA proteins are auxin-responsive and work as repressors of Auxin Responsive Factors, modulating their activity [
60]. Therefore, CsIAA29 might be involved in AR induction playing a role in the fine-tuning of the auxin signalling events, at least in chestnut. Arabidopsis AtIAA29 has been suggested to be involved in the modulation of root system development [
61], and ET was shown to target this gene through the EIN3 proteins, which work downstream of EIN2 [
62]. Therefore, ET negatively influences AR induction in mature chestnut shoots by putatively preventing the establishment of auxin gradients and impeding further related signalling.
As previously mentioned, EIN2 influences histone acetylation. P2CR rooting might also be blocked through epigenetic mechanisms governed by ET, and the present results support this hypothesis. However, CsHDA14, a histone deacetylase that reduces the accessibility of the transcriptional machinery to DNA, showed no significant differences among treatments despite a higher level of expression in IBA+AgNO
3 treated samples. On the other hand, the expression of CsJMJ30 clearly resembled the ET signalling status of the tissues, with expression being up-regulated in ACC-treated shoots and down-regulated by the inhibition of ET perception or synthesis. JMJ30 is a histone demethylase that generally acts in conjunction with its paralog JMJ32. In Arabidopsis, AtJMJ30 has been shown to lead to callus formation by inducing the genes AtLBD16 and AtLBD29 [
63]. Moreover, it seems to control root elongation in response to abscisic acid [
64]. In a previous report, abscisic acid-related gene expression was found to be more active in mature than in juvenile shoots of chestnut [
20], thus suggesting that CsJMJ30 might be integrating different cues into its expression that eventually relate to the recalcitrant behaviour of P2CR shoots, as low levels of activity of this demethylase might be increasing the accessibility of the transcriptional machinery to genes whose expression is needed for the development of ARs.