Control of adventitious root formation: insights into synergistic and antagonistic hormonal interactions.

Plants have evolved sophisticated root systems that help them to cope with harsh environmental conditions. They are typically composed of a primary root and lateral roots (LRs), but may also include adventitious roots (ARs). Unlike LRs, ARs may be initiated not only from pericycle cells, but from various cell types and tissues depending on the species. Phytohormones, together with many other internal and external stimuli, coordinate and guide every step of AR formation from the first event of cell reprogramming until emergence and outgrowth. In this review, we summarize recent advances in the molecular mechanisms controlling AR formation and highlight the main hormonal cross talk involved in its regulation under different conditions and in different model systems.


Introduction
Unlike most animals, plants have developed remarkable capacities of regeneration and propagation.
Inter alia, they can propagate both sexually and vegetatively due to their ability to develop adventitious roots (ARs) from aerial organs, which leads to the development of new genetically identical clonal plants. AR formation is a post-embryonic process, which is an intrinsic element of the normal development of monocots, while in both monocots and dicots it may occur in response to diverse environmental and physiological stimuli, such as darkness, flooding, mechanical wounding or nutrient deprivation (reviewed in Bellini et al. 2014, Steffens andRasmussen 2016). The process is a prime example of plants' evolution of sophisticated molecular machinery that precisely translates external cues and coordinates finely-tuned developmental responses that include de novo organogenesis (reviewed by Bellini et al. 2014, Xu 2018). This article is protected by copyright. All rights reserved. Vegetative or clonal propagation is exploited in horticulture and forest nurseries to produce large numbers of clones relatively quickly. However, some taxa (including economically important species) are difficult to root. For example, some Eucalyptus and Pinus species poorly develop AR without exogenous applications of phytohormones (Fett-Neto et al. 2001). Reasons for the wide variety in plants' capacity to form ARs are not yet clear.
It is now evident that the internal cues regulating AR initiation (ARI) are mainly integrated by the interactive effects of phytohormones (reviewed by Pacurar et al. 2014b). Auxin is the central player, but it acts with an array of other phytohormones through very complex crosstalk, modulating each other's levels and actions at every level: biosynthesis, metabolism, transport and signalling. In this review we summarize the most recent advances in our understanding of hormonal regulation of AR formation, particularly the initiation steps.

Adventitious root initiation and patterning
AR formation is a tightly controlled developmental program, which generally involves three steps.
First, cell specification and reprogramming, in which differentiated cells acquire new cell fate programs leading to the specification of AR founder cells. Second, initiation, in which the AR founder cells undergo successive cell divisions leading to the formation of primordia. Third, primordia emergence and outgrowth. ARI may occur in various cell types, with distinct anatomical and molecular identities, depending on both the species and environmental stimuli involved (reviewed by Geiss et al. 2009). In Arabidopsis thaliana (Arabidopsis), ARs may be initiated from xylem pole pericycle cells in both intact and de-rooted hypocotyls (Sorin et al. 2005, Sukumar et al. 2013).
Alternatively, they may be initiated from procambium, and probably from the adjacent parenchyma cells of leaf explants kept in the dark on hormone-free B5 medium (Liu et al. 2014). However, Bustillo-Avendaño et al. (2018) found that when a whole leaf including the petiole was used as a propagating system, ARs initiated from a pre-formed micro-callus. When the petiole is kept, vascular tissues including xylem and pericycle-like cells first undergo massive cell division, resulting in formation of a micro-callus, which subsequently undergoes a second step of reprogramming that specifies the AR founder cell (Bustillo-Avendaño et al. 2018). This process resembles the two-step mechanism during hormone-induced organogenesis (reviewed by Kareem et al. 2016 initiate from the peripheral vascular cylinder. In conclusion, whatever the tissue or organ, ARs are initiated from cells with vascular associations. Regardless of the cell type from which AR arise, phytohormones in coordination with environmental stimuli guide every step of AR formation. They usually act in complex interactions that provide the robust spatio-temporal cues required for AR development (Fig. 1), as outlined in the following sections.

Auxin: the master player and hub of any crosstalk
Indole 3-acetic acid (IAA), the most abundant natural auxin, is a weak organic acid that is mainly derived from L-tryptophan. IAA controls a plethora of developmental programs in plants, including AR formation. It has long been established that different types of auxins promote AR formation in different species (Bellini et al. 2014, Pacurar et al. 2014b). Indole-3-butyric acid (IBA) is the most frequently used natural auxin for clonal propagation in horticulture and forestry because of its stability and effectiveness in promoting AR from stem cuttings compared to other available auxins. IBA induces the auxin signalling machinery following conversion to IAA in planta (Strader et al. 2011).
Thus, the double mutant, ech2ibr10, in which IBA to IAA conversion is impaired, produces very few LRs and ARs compared to wild type counterparts (Strader et al. 2011, Fattorini et al. 2017. The conversion of IBA to IAA seems to be species and genotype dependent. This is significant, because (for example) the rooting capacity of elm (Ulmus americana) genotypes' stem cuttings is correlated with their ability to convert IBA to IAA (Kreiser et al. 2016). IBA-derived IAA also controls many other developmental processes (Frick and Strader 2017). For example, root cap-derived IBA plays a prominent role in lateral root pre-branch site establishment in Arabidopsis (Xuan et al. 2015).
In Arabidopsis, constitutively IAA-overproducing mutants such as superroot1 (sur1), superroot2 (sur2) and the activation tagged YUCCA1 (yuc1-D) spontaneously form AR in the hypocotyl (Boerjan et al. 1995, Delarue et al. 1998, Zhao et al. 2001). Very early steps in ARI and most organogenesis include formation of an IAA gradient and its accumulation in specific cell types, via processes that include polar auxin transport (PAT) and local auxin biosynthesis, conjugation and degradation.
IAA biosynthesis has been found to be crucial for the cell fate transition of procambium cells to AR founder cells in leaf explants grown on (hormone-free) B5 medium in darkness (Liu et al. 2014, Chen et al. 2016 (Pacurar et al. 2014a). Loss of function of any of these genes reduced numbers of ARs produced by sur2-1 mutants, confirming the pivotal role of IAA biosynthesis in AR formation (Pacurar et al. 2014a).
It is well known that polar auxin transport (PAT) plays a key role in IAA distribution and gradient establishment. Ahkami et al. (2013) found that free IAA levels increased quickly in the stem base of petunia (Petunia hybrida) cuttings in a biphasic manner, with IAA peaking 2 and 24 h after cutting.
The second peak ( Interestingly it has been demonstrated that JA is an inhibitor of ARI in Arabidopsis hypocotyls, and that auxin induces ARI by decreasing the pool of active JA through its conjugation to amino acids by GH3 enzymes (Gutierrez et al., 2012). Therefore, we propose that the peaks of IAA that occur in petunia cuttings may prevent inhibition of ARI by wounding-induced accumulation of JA.
This article is protected by copyright. All rights reserved. The ATP-BINDING CASSETTE B19 (ABCB19) and PIN-formed 1 (PIN1) seem to be the main IAA efflux carriers involved in auxin accumulation at the base of excised Arabidopsis hypocotyls and, thus, contributors to AR formation. Accordingly, Sukumar et al. (2013) found that de-rooted abcb19-1 and pin1-1 loss-of-function mutant seedlings produced fewer ARs than wild type counterparts, and the ABCB19 gene was rapidly induced upon excision, especially in the hypocotyl epidermis and vascular tissues. PIN6 is also involved in AR formation, as demonstrated by phenotypes of loss-or gain-offunction lines reported by Simon et al. (2016). The pin6 knock-out mutant produced more ARs, whereas the over-expressing line 35:PIN6 produced very few ARs compared to wild type counterparts. These findings confirm the importance of rootward IAA transport during ARI.
Many mutants with IAA perception or signalling impairments have been identified in the last decade and their functional analysis has illuminated some of the molecular mechanisms controlling ARI downstream of IAA. In intact Arabidopsis hypocotyls, IAA acts through transcription factors from the AUXIN RESPONSE FACTOR (ARF) family. ARF6 and ARF8 have been identified as positive regulators of ARI, and ARF17 as a negative regulator (Gutierrez et al. , 2012. Interestingly, ARF6 and ARF8 genes were found to be specifically induced in the rooting-competent phloem parenchyma cells in stem cuttings of black walnut (Juglans nigra L.), while ARF17 expression decreased during the same developmental stage (Stevens et al. 2018). This differential expression of the three ARF genes was observed before AR primordia formation (Stevens et al. 2018). Ruedell et al. (2015) found that expression of ARF6 and ARF8 was also induced in difficult-to-root Eucalyptus globulus donor plants treated with far-red enriched light, and after cutting. Moreover, this induction coincided with the promotion of AR formation under these conditions. These results suggest that ARF6 and ARF8, which were identified as positive regulators of AR formation in Arabidopsis, may be key elements of an ARI mechanism that has been conserved across diverse taxa.
ARF6 and ARF8 genes regulate expression of the well-known auxin responsive genes GH3.3, GH3.5 and GH3.6 and the triple null mutant gh3.3gh3.5gh3.6 reportedly produces fewer ARs than wild-type counterparts, highlighting a paradox (Gutierrez et al. 2012). The three corresponding GH3 proteins were initially described as auxin conjugating enzymes (Staswick et al. 2005), but knocking out gh3.3, gh3.5 and gh3.6 had no apparent effect on endogenous IAA contents, instead it resulted in higher production of free JA and its bioactive form (+)-7-iso-jasmonoyl-L-isoleucine (JA-Ile) (Gutierrez et al. 2012). These findings prompted the conclusion that auxin acted through the GH3 proteins by controlling levels of active JA, which was demonstrated to be a negative regulator of AR formation ( Fig. 1) (Gutierrez et al. 2012).
This article is protected by copyright. All rights reserved. Additional components of the IAA signalling machinery were identified in a genetic screen for suppressors of the AR phenotype of the auxin-overproducing mutant sur2-1 (Pacurar et al. 2014a). For example, the COP9 SIGNALOSOME SUBUNIT 4 (CSN4), which controls de-neddylation of CULLIN1 in the SCF complex, acts on the ARF6/8 regulatory module and likely at the crossroads of IAA, JA and/or light signalling pathways involved in the control of both AR and LR formation (Pacurar et al. 2017). Intriguingly, phenotypic characterization of available viable csn mutants indicated a possible differential role of the COP9 signalosome in AR and LR formation (Pacurar et al. 2017). Moreover, the gain-of-function mutants crane-2 and solitary root-1 (slr-1), which harbour mutations in domain II of IAA18 and IAA14, respectively, that confer resistance to degradation by the proteasome, were also shown to be affected in ARI ( mediates a conserved evolutionary mechanism of adventitious rooting in monocots and dicots, including woody species (Zhao et al. 2009, Xu et al. 2015.

Jasmonic acid: a positive or negative regulator of adventitious root initiation?
Jasmonic acid (  To obtain detailed insights into the key players involved in specific steps of AR formation in petunia stem cuttings, Druege et al. (2014) performed an extensive time course microarray analysis. This revealed that many ACC SYNTHASE (ACS) and ACC OXIDASE (ACO) genes were rapidly induced after cutting, suggesting that ET biosynthesis is required for ARI. In addition to ET biosynthesis genes, many APETALA2/ETHYLENE RESPONSE FACTOR (AP2/ERF) genes were differentially expressed at various stages during the process. Several AP2/ERF genes have been recently identified as potent players in regulation of tissue repair, de-novo organogenesis and regeneration (Heyman et al. 2017), indicating possible roles of the ET-induced AP2/ERF genes in AR formation.
In attempts to elucidate ET's role in ARI, Rasmussen et al. (2017) analysed several Arabidopsis ET biosynthesis and signalling mutants. Both ethylene overproducing 2-1 (eto2-1) and eto3-1 mutants, and constitutive triple response1-1 (ctr1-1) mutants, which have constitutive ET signalling, produced more ARs, while ethylene insensitive2-1 (ein2-1) and ein3-1 mutants produced fewer ARs, than wild type controls. To the same end, Rasmussen et al. (2017) showed that at very low concentration (0.01 µM) ACC induced slightly higher AR densities, but higher concentrations (0.1-1 µM) had no significant effect. These results clearly indicate that ET biosynthesis and signalling promote AR formation. The effect of ET on ARI is likely independent of strigolactone (Rasmussen et al. 2017).
However, contrasting results were obtained in another study using the same system and concentration of ACC by Veloccia et al. (2016), who found that treatment with 0.01 µM ACC had no significant effect and higher (0.1-1 µM) concentrations tended to suppress AR formation. These contrasting results might be due to differences in growth conditions and/or high concentrations of ACC affecting other physiological processes of the Arabidopsis seedlings, which might cause aberrant and pleiotropic developmental defects. Veloccia et al. (2016) assumed that ET inhibits AR formation in Arabidopsis by repressing the transcription of WEI2, WEI7 and YUC6 genes by an unknown mechanism. This would be surprising because of previous findings that ET induces expression of WEI2 and WEI7, which leads to increases in IAA biosynthesis and signalling in the Arabidopsis root tip (Stepanova et al. 2005). Because wei2 and wei7 mutations strongly suppressed the sur2-1 phenotype, Pacurar et al. (2014a) proposed that ET might have a positive role in early stages of AR formation, by inducing IAA biosynthesis, but a negative role in later stages.
This article is protected by copyright. All rights reserved. In conclusion, several studies have provided indications that ET is involved in AR formation, but its mode of action is still far from clear. Identifying its downstream targets and acquiring direct evidence about its specific spatiotemporal regulatory role would be helpful for elucidating its mode of action.

Cytokinins: the required inhibitors
Like IAA, cytokinins (CKs) have been extensively studied and characterized. They are adeninederived phytohormones that were called cytokinins because of their ability to promote cell division interestingly, the petioles that developed a micro-callus generated higher numbers of ARs.
Accordingly, downregulation of arr1arr10arr12, using oestradiol-inducible artificial microRNA, reportedly resulted in production of ARs rather than shoots by root explants incubated in CK-rich shoot induction medium (CIM) (Meng et al. 2017

Strigolactones: the newcomers to the club
Strigolactones (SL) are carotenoid-derived phytohormones, which participate in regulation of many developmental programs, including root architecture development (reviewed by Waters et al. 2017).
Their roles in the control of AR formation are not clearly understood, but they were recently shown to participate interactively in this process with other hormones in various species.
SLs promote crown root elongation in rice via effects on cell division in the meristematic zone (Arite et al. 2012). Recently, it has also been shown that SLs positively control crown root formation in rice.
In addition, Sun et al. (2015) found that the SL-deficient rice mutant dawarf10 (d10), in which the key gene of SL biosynthesis is impaired, produced fewer ARs than wild type counterparts. AR production was also reduced in the SL-insensitive mutant d3, in which the F-box protein (a key component of the SCF complex involved in SL signalling) is affected. However, treating plants with GR24, a synthetic SL analogue, complemented the d10 phenotype but not the d3 phenotype (Sun et al. 2015). SLs apparently promote AR formation in rice via the D4-dependent pathway, probably by modulating IAA transport (Sun et al. 2015). However, they seem to inhibit AR formation in tomato (Solanum lycopersicum), pea (Pisum sativum) and Arabidopsis (Kohlen et al. 2012, Rasmussen et al. 2012 between monocots (rice) and dicots (pea, tomato, Arabidopsis) raise intriguing questions that remain to be answered.

Abscisic acid (ABA), gibberellic acid (GA) and brassinosteroids (BRs): the least investigated hormones in adventitious rooting
Roles of ABA, GA and BRs in AR formation are largely unclear, but there are interesting indications that they participate. For example, GA stimulates AR formation in deepwater rice via a mechanism that requires the presence of ET (Steffens et al. 2006). Treatment with GA3 alone was reportedly ineffective, but in combination with ET it significantly stimulated AR emergence (Steffens et al. 2006 Populus with enhanced GA biosynthesis or perception produced fewer ARs than wild type counterparts. In the cited study, GA inhibited AR formation independently of SL biosynthesis and JA signalling, but probably interfered with establishment of an IAA gradient at the base of the stem cuttings by perturbing expression of IAA auxin efflux carrier genes. ABA, which is a stress-related hormone, has been shown to inhibit AR emergence in deepwater rice, possibly by interfering with ET and GA signalling pathways (Steffens et al. 2006  signalling acts synergistically with the ARF6/ARF8 signalling module identified by Gutierrez et al (2012) in ARI promotion.

Conclusion and perspectives
AR formation is a developmental program that is controlled by complex hormonal crosstalk with strongly non-linear effects (Fig. 1). Numerous studies with model species such as Arabidopsis, rice and petunia, have provided information about the complex interactions involved. However, despite remarkable advances in molecular-level understanding of the process in the model species, little is still known about it in woody species such as Populus, Pinus, and Norway spruce. Checking whether the mechanisms identified in the model species are evolutionarily conserved is important, but challenging.
Fortunately, the availability of genome sequences of increasing numbers of species and rapid development of genome editing techniques such as TALEN and CRISPR-Cas9 will greatly facilitate elucidation of AR formation in diverse taxa (particularly ecologically and economically important species) and the evolution of associated developmental machinery.
While efforts have focused on elucidating this process in single linear hormonal pathways, advances in modelling and integrative system biology approaches, including hormonomics, should be exploited to acquire more insights into the complex hormonal interactions involved and their final outcomes. It would also be interesting to discover how plants couple the intrinsic and extrinsic stimuli that guide and shape AR architecture. Further knowledge is required of how the external stimuli are perceived, transmitted and translated into tight internal signalling cues, as well as how specific subsets of cell types are targeted by these cues and transformed into AR founder cells.
• Author contributions • A.L. wrote the manuscript, A.L. and C.B. edited it. This article is protected by copyright. All rights reserved.

PPL_12823_F1.tif
This article is protected by copyright. All rights reserved.