MDF regulates both auxin-dependent and -independent pathways of adventitious root regeneration in Arabidopsis

Plants exhibit strong plasticity in growth and development, seen clearly in lateral and adventitious root development from differentiated tissues in response to environmental stresses. Previous studies have demonstrated the role of both auxin-dependent and auxin-independent signalling pathways in regulating the de novo formation of adventitious roots (ARs) from differentiated tissues, such as leaf petiole in Arabidopsis. One important question is how the auxin-dependent and -independent pathways are coordinated. To investigate this question, we used a combined approach of inducible gene expression, mutant, and signalling reporter gene analysis during AR regeneration in the Arabidopsis petiole to understand regulatory relationships. Auxin signalling components AXR1 and AXR3, and the PIN trafficking protein VAMP714, are each required for AR initiation, as is the ethylene signalling repressor POLARIS, but not EIN2. We identify the RNA splicing regulator MDF and the transcription factor RAP2.7 as new positive regulators of both the auxin-independent and auxin-dependent pathways, and show that MDF regulates RAP2.7, WOX5 and NAC1; while RAP2.7 regulates WOX5 but not NAC1 or YUC1. NAC1 is required for de novo root formation in a pathway independent of YUC1, WOX5 or RAP2.7. We propose a model in which MDF represents a point of molecular crosstalk between auxin-dependent and -independent regeneration processes.


Introduction
An important question in developmental biology is that of how new cell identities are acquired during organogenesis.In plants, regeneration of new organs following re-specification of cell identity is more common than in animals, but the mechanisms are incompletely understood (Xu and Huang, 2014;Jing et al., 2020).Adventitious roots (ARs) represent one form of regeneration from previously differentiated cells.They are typically initiated in response to changes in the environment such as flooding, drought, or wounding (Steffens and Rasmussen, 2016), and help plants to acclimate to stressful environments.
For example, ARs help plants absorb water and oxygen in flooding, and they can grow from wound sites as part of a regeneration process (Gonin et al., 2019).ARs that develop from a wound site represent de novo organogenesis.This kind of root developmental programme depends on endogenous hormones and is considered to arise from procambium or cambium cells, which contain stem cells located in the vascular tissues of aerial organs (Chen et al., 2014;Liu et al., 2014).
ARs can be grown from non-root tissue in isolation in response to hormonal signals, notably auxin.There are two promoting principles involved in damage-related formation of ARs (Chen et al., 2014).The first is the wound response that is induced during the isolation of the explant tissue from resource supply in the plant.The second is the signalling networks that affect organ regeneration postwounding.For example, detached leaf explants can activate various early signals that include both shortand long-range signals that lead to AR formation (Chen et al., 2014).AR formation from leaf explants has been considered to comprise three developmental phases: the early wound response, the creation of an auxin maximum in competent cells and the establishment of founder cells, leading to de novo root meristem formation and activity (Jing et al., 2020).Wound signals induce YUCCA1 (YUC1) and YUC4 expression near the wound region to help maintain auxin levels (Xu, 2018).YUC1 and YUC4 mediate auxin biosynthesis and they are important for cell fate transition during de novo organogenesis (Chen et al., 2016a, b).The YUC family also plays an important role in converting indole-3-pyruvic acid (IPA) to indole-3-acetic acid (IAA) (Sun et al., 2016).Cell fate transition is blocked if auxin signalling is inhibited (Liu et al., 2014;Xu, 2018).
Wound signals have many functions, including the activation of expression of the NAC1 gene (Xu, 2018).NAC1 (NAC DOMAIN-CONTAINING PROTEIN 1) is a transcription factor of the NAC family required for de novo root formation, and transgenic repression of NAC1 expression causes reduced development of adventitious roots (Chen et al., 2016c).The NAC pathway works independently from the auxin pathway and wounding plays an important role in inducing it in leaf explants of Arabidopsis.In addition, NAC1 may have a role in cell wall metabolism to promote AR development (Chen et al., 2016c).It is also suggested that NAC genes promote expression of KDEL-tailed Cys endopeptidase (CEP) genes (CEP1 and CEP2).CEP proteins have been found to contribute to programmed cell death, by their secretion to the cell wall to induce inactivation of EXTENSIN (EXT) proteins.EXT proteins are important for cell wall expansion and wound healing, and wound healing may antagonise AR emergence.Therefore, repression of EXT genes by NAC proteins may promote AR emergence from leaf explants (Chen et al., 2016c).NAC1 appears to be working independently of the auxin pathway and may have a role in cell wall metabolism to promote AR development.However, the regulation of NAC1 in the context of de novo root regeneration is not well understood.
The third phase of AR formation, namely the fate determination of founder cells and their subsequent development into a functional root meristem, can be described as involving four steps (Jing et al., 2020).The first step of cell fate determination ('Priming') involves auxin activation of the expression of WUSCHEL RELATED HOMEOBOX11 and 12 (WOX11/12) genes, involved in the transformation of competent cells to root founder cells in the leaf explants.The second step ('Initiation') involves cell division to create root primordium cell layers around 2-4 days after wounding (Xu, 2018).
The homeobox gene WOX5 is required to start this second step whereby root founder cells divide to become root primordium cells.In addition, transcriptional activation of WOX5 requires auxin, and repression of auxin causes reduced expression of WOX5 which in turn reduces de novo root formation (Hu and Xu, 2016).The third step ('Patterning') creates cellular pattern in the root apical meristem and is characterized by the loss of LATERAL ORGAN BOUNDARIES DOMAIN16 (LBD16) expression and restriction of WOX5/7 expression to the stem cell niche.Other regulatory genes such as SHORT ROOT (SHR), PLETHORA1/2 and SCARECROW (SCR) are also involved in meristem formation (Bustillo-Avenado et al., 2018;Kim et al., 2018).Finally, the fourth step ('Emergence') is characterized by root tip growth through the epidermis of the leaf explant.
The aim of the work described in this paper was to understand better the mechanistic basis and inter-relationship between the auxin-dependent and -independent control of de novo root regeneration.
We show a requirement for the auxin-regulated R-SNARE VESICLE-ASSOCIATED MEMBRANE PROTEIN 714 (VAMP714) in AR formation and growth, and identify the RNA splicing regulator MERISTEM-DEFECTIVE (MDF; Casson et al., 2009;Thompson et al., 2023) and the transcription factor RAP2.7 (RELATED TO AP2.7; Tair 2024) as new positive regulators of both the auxinindependent pathway and auxin-dependent pathways, and show that MDF regulates RAP2.7,WOX5 and NAC1; while RAP2.7 regulates WOX5 but not NAC1 or YUC1.NAC1 is required for de novo root formation in a pathway independent of YUC1, WOX5 or RAP2.7.We propose a model for the network of gene and hormone signalling interactions involved in adventitious root formation and suggest that MDF represents a point of molecular crosstalk between auxin-dependent and -independent regeneration processes.

Contrasting roles of AXR1 and AXR3 and ethylene signalling in de novo root initiation and growth
A simple method to induce the de novo development of roots in vitro has been described by Chen et al. (2014).This involves wounding the petiole of a 12 d-old leaf of Arabidopsis, and then incubating the leaf on B5 nutrient medium lacking both hormones and sucrose.New meristem initiation from the wild type petiole wound site is evident by ca. 5 days after wounding (DAW), associated with an auxin maximum as indicated by expression of the auxin reporter DR5::GUS in the new root primordium; and de novo roots emerging after 8-10 d DAW; Suppl.Fig. S1).To investigate the roles of signalling pathway regulators during de novo root meristem formation and growth, 12 d-old leaves were wounded by excision from the seedling of signalling mutant genotypes, and the regeneration responses compared to wild type responses were determined.Fig. 1A shows the mean numbers of root branches initiated on wounded petioles of wild type (Col-0) and mutants defective in auxin signalling (auxin-resistant axr1 and auxin-hypersensitive axr3; Lincoln et al., 1990;Leyser et al., 1996) and ethylene signalling (ethylene-insensitive ein2, ethylene hypersignalling pls; Casson et al., 2002, Chilley et al., 2006) after 12, 19 and 26 DAW.For wild type, the mean number was 1.4 at 12 DAW, rising to 36.8 at 26 DAW.axr1 was unable to regenerate any roots, while axr3 produced a mean of 2.1 at 12 DAW, rising to 13.7 (i.e.ca.37% that of wild type) at 26 DAW.The ein2 leaf produced a mean of 1.3 branches at 12 DAW, not significantly different to wild type either then or at 19 DAW, but produced a significantly lower number of branches at 26 DAW, with 23.5 (ca.64% of wild type; P <0.05).pls produced a mean of 0.9 root branches at 12 DAW, rising to 3.1 at 26 DAW (ca.12% of wild type).Fig. 1B shows the mean lengths of the roots initiated from wounded petioles of wild type and the auxin and ethylene signalling mutants.By 12 DAW the mean root length for Col-0 was 34.25 mm.
For axr1 no roots were detected at 12 DAW, for axr3 was mean length was 11.49 mm, for ein2 20.19 mm, and for pls 6.51 mm.By 19 DAW, the wild type mean root length increased to 52.2 mm, axr3 roots increased to a mean of 23.17 mm, pls increased by a much smaller amount to a mean of 7.75 mm, and ein2 roots were not significantly different to wild type at a mean of 43.5 mm.This general pattern was retained at 26 DAW (Col-0 122.41 mm, axr3 39.19 mm, ein2 117.61 mm, pls 8.30 mm, with no roots on axr1).
These results suggest AXR1-mediated auxin signalling, but not AXR3-mediated auxin or PLSor EIN2-mediated ethylene signalling, is essential for root initiation, given the number of roots produced by 12 DAW is similar in the axr3, pls and ein2 mutants.However, AXR3 and possibly also AXR1 (both required for correct auxin responses) and PLS (which represses ethylene responses) are required for wild type levels of root elongation.

Hormone signalling gene interactions in regeneration control
Previous work has shown a requirement for auxin biosynthesis for de novo root regeneration, via the YUCCA pathway (Chen et al., 2016a, b).Matosevich et al. (2020) suggested that, in contrast to earlier views (Sena et al., 2009), PIN-dependent polar auxin transport is not essential for root tip regeneration from wounded roots, based on analysis of pin mutants and naphthylphthalamic acid (NPA) treatments (NPA inhibits PIN protein activity; Abas et al., 2021).This is indicative of an auxin-independent pathway for de novo root regeneration, though there was suggested a possible role for short-distance (symplastic) auxin transport at the wound surface (Mellor et al., 2020).Nevertheless, PIN1, PIN3, PIN7 and auxin-inducible DR5::GUS expression is induced in the leaf petiole, in particular during regeneration (Suppl.Figs.S1-4), suggesting that the polar auxin transport pathway is activated.
Furthermore, the auxin-regulated VAMP714::GUS promoter-reporter is also expressed in a similar pattern to DR5::GUS (Suppl.Fig. S5), and the vamp714 loss-of-function mutant, defective in polar auxin transport, shows reduced adventitious root development compared to wild type (Fig. 2, A and B), with associated reduced expression of auxin-regulated genes IAA1 and IAA2 in both explanted leaf blade and petiole of the mutant (Fig. 2, C and D).VAMP714 is an R-SNARE vesicle-associated protein required for the trafficking of PIN proteins and polar auxin transport (Gu et al., 2021).In any case a new auxin maximum is established close to the wound site.The synthetic EBS::GUS ethylene reporter was expressed in the leaf blade up to 8 DAW, and in the petiole but not root primordium at 5-6 DAW (Suppl Fig. S6).However, after 8 DAW GUS expression occurred in the emerging de novo root, suggesting ethylene responses are not obviously important in early stages of primordium development; though the pls mutant data (Fig. 1A) indicate there may be required a regulated ethylene signalling response for adventitious root elongation.
To determine the relationship between hormone signalling and regulatory genes known to be involved in root regeneration, we first determined by RT-qPCR the expression of the genes NAC1, WOX5, YUC1 and YUC4 in the signalling mutants axr1, axr3, pls and ein2 compared to expression in wild type during a root regeneration time course of 0, 2, 3, 7 and 14 DAW.Fig. 3A shows a significant increase in NAC1 expression between 2 and 3 DAW, with no significant difference between wild type and mutants during this period.Interestingly, by 14 DAW NAC1 expression in wild type, ein2 and axr3 has continued to increase, but has dramatically reduced in both pls and axr1 mutants, associated with reduced (in pls) or no (in axr1) root initiation and growth in these mutants by this stage.It is notable however that during the period in which de novo root regeneration is initiated (5-8 DAW), NAC1 expression is unaffected in ethylene or auxin signalling mutants compared to expression levels in wild type.showed increased expression in wild type, ein2 and axr3 samples during the culture period, but very low levels of expression were seen in axr1 and pls from the beginning of the experiment, with no subsequent increase.For axr3, the expression of each gene was significantly lower than in wild type and ein2 from 3 DAW onwards, though significantly higher than in either axr1 or pls, which showed the lowest levels of expression of WOX5, YUC1 and YUC4.These results indicate that WOX5, YUC1 and YUC4 are dependent on both AXR1 and PLS for expression during de novo root formation.NAC1 expression was not dependent on the auxin signalling AXR1 and AXR3 genes, nor the ethylene signalling PLS and EIN2 genes during the regeneration process (4-8 DAW).

MDF regulates both auxin-dependent and -independent root regeneration
Next we investigated the possible regulatory role in de novo regeneration of three genes with roles in meristem function (MDF, RAP2.7 and NAC1).The MDF gene of Arabidopsis is required for root meristem organization and maintenance (Casson et al., 2009).It encodes a putative RS domain protein and has recently been demonstrated to be a splicing factor that regulates meristem function through both auxin-dependent and auxin-independent mechanisms, to maintain stemness (Thompson et al., 2023).
The mutant shows decreased levels of PIN family mRNAs and this is associated with a reduced auxin maximum in the basal region of the mdf embryo and seedling root meristem.Furthermore, seedling roots of mdf show reduced expression of SHORTROOT, SCARECROW, WOX5 and PLETHORA genes (Casson et al., 2009, Thompson et al., 2023).RNA-seq analysis shows slightly reduced VAMP714 expression in mdf (log2-fold 0.74, padj = 6.26E-10;Thompson et al., 2023), which may contribute to the observed defective PIN localization in the mdf mutant.The RAP2.7 gene is a member of the AP2/ERF family of transcription factors which are involved in regulating the process of flowering and innate immunity and may be involved in the control of meristematic activity (Tair, 2024).We recently found RAP2.7 to be a splicing target of MDF (it is mis-spliced in the mdf mutant; Thompson et al., 2023) and has also been identified as being expressed in the Arabidopsis root tip, similar to MDF (Birnbaum et al., 2003).Given that the NAC1 gene is known to be involved in de novo root formation (Chen et al., 2016c) and in lateral root development mediated by auxin signalling (Xie et al., 2000), we were interested in understanding its functional relationship, if any, with MDF and RAP2.7.
For analysis of MDF function, we utilized the previously described mdf-1 loss-of-function mutant, and also generated transgenic plants containing an estradiol-inducible XVE-35S::MDF gene construct for overexpression analysis.Three independent transformants were characterized, all showing ca.6-7-fold induction of MDF expression, and one representative line was chosen for further analysis (Suppl.Fig. S7).The results in Fig. 4, A and B shows the numbers and lengths of de novo roots formed from leaf of wild type, MDF transgenic overexpresser (MDF-OV) following estradiol induction, and the mdf-1 mutant at 12, 19 and 26 DAW.By 12 DAW the mean number of initiated roots was 1.66 for Col-0 and 2.56 for MDF-OV, which was not statistically significantly different.By 19 DAW the number of roots increased (mean of 24.2 roots for Col-0, 32.44 branches for MDF-OV) and there was a statistically significant difference between Col-0 (mean 40.1 roots initiated) and MDF-OV (mean 74.8 roots initiated; ca.187% of wild type) by 26 DAW.The mdf mutant failed to produce any detectable roots over the entire time course.This was associated with reduced expression of the auxin-responsive IAA1 gene over the 8 d regeneration time course (Suppl.Fig. S8).At 12 DAW, the roots produced from Col-0 and MDF-OV leaves showed no statistically significant difference in mean length; but at both 19 and 26 DAW the MDF overexpresser produced significantly longer roots than wild type (P < 0.01, n = 30; Fig. 4B).This shows a positive regulatory role for MDF in both initiation and growth of de novo adventitious roots in this system.
To determine whether MDF has a role in the regulation of the WOX5 gene, which is shown in Fig. 3B to be associated with adventitious root regeneration, RNA was extracted from cultured leaf of wild type, mdf-1 and the MDF-OV line, either untreated (-E) or treated with estradiol (+E) to induce MDF expression over an 8 d time course of culture, and transcriptional analysis was carried out.Fig. 4, C and D show the mean expression data from three biological replicate experiments, each with three technical replicates.In the absence of estradiol induction, there was no significant difference in WOX5 expression between wild type (Col-0) and MDF-OV samples, but there was a statistically significantly lower level of expression in the mdf-1 mutant by 3 DAW.When MDF expression was induced by estradiol, the MDF-OV samples showed significantly higher levels of WOX5 expression, showing that MDF is a positive regulator of WOX5.This correlated with an increase in both root length and root branch number in MDF-OV samples compared to wild type after longer periods of culture (Fig. 4, A   and B), and is consistent with an inductive effect of MDF on adventitious root meristem formation.
When MDF expression was induced by estradiol over an 8 d regeneration time course, the MDF-OV samples showed significantly higher levels of RAP2.7 expression (Fig. 4, E and F).In the absence of estradiol induction, there was no significant difference in RAP2.7 expression between wild type (Col-0) and MDF-OV samples, and there was no detectable expression in the mdf-1 mutant at any point during the experimental time course.NAC1 expression was undetectable in leaf at 0 DAW for any genotype, but was detectable by 3 DAW, and not significantly different at this timepoint in wild type and MDF-OV, but it was significantly reduced in the mdf-1 mutant (Fig. 3, G and H).By 5 and 8 DAW MDF overexpression led to a significant increase in NAC1 expression, while NAC1 expression was significantly reduced in the mdf-1 mutant, compared to levels in wild type (Fig. 3, I and J).This demonstrates a role for MDF in the positive transcriptional control of both RAP2.7 and NAC1 following wounding.

The RAP2.7 transcription factor regulates expression of WOX5 but not YUC1 or NAC1 in root regeneration
To investigate a possible regulatory role of the MDF target RAP2.7,transgenic overexpressers (pro35S::RAP2.7,RAP-OV) and two independent confirmed loss-of-function SALK mutants of RAP2.7 were characterized; one representative line of each was further analysed for adventitious root initiation and growth, and for the expression of YUC1, NAC1 and WOX5 genes at 12, 19 and 26 DAW.The RAP-OV line used showed a ca.7.4-fold increase in RAP2.7 expression compared to wild type (Fig. S9).At 12 DAW the mean number of roots initiated was not statistically significantly different between the three genotypes.By 19 DAW the number of roots initiated increased for both wild type and overexpresser, and there was no significant difference between them; however, the rap2.7 mutant produced significantly fewer roots than wild type and showed evidence of senescence (Fig. 4B).There were significantly more roots initiated in the RAP2.7 overexpresser, and significantly fewer in the rap2.7 mutant, than wild type by 26 DAW.At 12 DAW, there was no statistically significant difference in root length between Col-0 and RAP-OV genotypes, but the rap2.7 mutant produced significantly shorter roots than wildtype (Fig. 4, A and C).At both 19 and 26 DAW the RAP2.7 overexpresser produced significantly longer roots than wild type.These results show a role for RAP2.7 in both initiation and growth of de novo adventitious roots in this system.Expression levels of YUC1, NAC1 and WOX5 transcripts in the RAP2.7 overexpresser and mutant, compared to wild type, were determined during a 7 d culture period post wounding.The results in Fig. 4E show that, for all genotypes, the level of YUC1 expression increased during the experimental time course, but there was no statistically significant difference between mutant and overexpresser of RAP2.7 at any time point, compared to wild type.These data show that YUC1 expression is regulated by a pathway independent of RAP2.7.The same pattern was observed for NAC1 expression levels in the different genotypes (Fig. 4F), again indicating that NAC1 expression is independent of RAP2.7.However, WOX5 expression was significantly increased in the RAP2.7 overexpresser and decreased in the rap2.7 mutant by 5 DAW, showing that the WOX5 pathway is positively regulated by RAP2.7.

NAC1 regulates root regeneration independent of YUC1, WOX5 or RAP2.7
There is previous evidence of the NAC1 transcription factor being able to promote adventitious rooting in an auxin-independent manner (Chen et al., 2016c).To understand better the mechanistic basis of the regulatory role of NAC1, transgenic overexpressers (NAC-OV) and independent confirmed loss-of-function mutants of NAC1 were identified and one representative line of each was analysed for adventitious root formation and growth, and for the expression of YUC1, WOX5 and RAP2.7 genes during post-wounding culture of leaf.The NAC1-OV line used showed a ca.13-fold increase in RAP2.7 expression compared to wild type (Suppl.Fig. S10).Throughout the entire time course there was no statistically significant difference in number or length of roots formed between Col-0 and NAC-OV genotypes, but the nac1 mutant failed to produce any adventitious roots (Fig. 5, A-C).This confirms previous reports of the requirement of NAC1 for de novo root formation.Expression analysis showed that, for all genotypes, the level of YUC1, WOX5 and RAP2.7 expression increased during the experimental time course, but there was no statistically significant difference in expression levels for any of the genes between mutant and overexpressers of NAC1 compared to wild type (Fig. 5, D-F).
These results show that NAC1 does not regulate YUC1, WOX5 or RAP2.7 during de novo root regeneration, and its role in AR formation is by a pathway independent of these genes.

Discussion
Plants have a remarkable developmental plasticity and capacity for regeneration.They can create new organs from non-embryonic tissue (Hartmann et al., 2010;Chen et al., 2014;Liu et al., 2014) and readily repair damage that occurs upon wounding (Xu et al., 2006;Heyman et al., 2013).Wounding plays a major role in inducing the accumulation of the hormone auxin at high concentrations close to the cut surface (Chen et al., 2016a, b).Crosstalk between auxin, cytokinin and other hormones is an important regulatory mechanism of many aspects of plant development and regeneration (El-Showk et al., 2013;Moore et al., 2015Moore et al., , 2024;;Liu et al. 2017), including in de novo root formation.
There are considered to be two pathways that regulate de novo root regeneration from leaf explants, namely the auxin pathway and the NAC1 pathway; and past studies have shown that defects in either of these two pathways causes repression of de novo root development (Chen et al., 2016a-c).
During de novo root formation cell division occurs in competent cells in the petiole and transforms these cells to root founder cells, in a process involving the expression WOX11/12 genes (Liu et al., 2014).The root founder cells then divide to form root primordium cells, associated with expression of WOX5/7 genes, which activate WOX11/12 gene expression is required to activate WOX5/7 (Hu and Xu, 2016).
Finally, the root apical meristem continues to develop from the root primordium cells, leading to root emergence from the leaf explant.
MDF is a component of the plant spliceosome and is itself regulated transcriptionally in an auxin-independent manner (Casson et al., 2009;Thompson et al., 2023).It is required for the correct splicing of genes involved in meristem function, such as RAP2.7,RSZ33 and ACC1, and for the correct level of expression of other genes, downstream of spliced targets, that themselves regulate auxin transport and auxin-mediated meristem control (Thompson et al., 2023).Lack of MDF function leads to loss of an auxin maximum in the Arabidopsis root due to reduced PIN protein levels and reduced of expression of other meristem genes such as SHR and SCR in an auxin-independent manner, and illustrated by reduced IAA1 expression (Suppl.Fig. S8).Previously we have shown that MDF overexpression could induce ectopic shoot meristems (Casson et al., 2009).In the current paper we show that MDF is required for de novo root formation and the correct quantitative expression of RAP2.7,WOX5 and NAC1 (Fig. 4).MDF overexpression increases both the expression of these genes and also the number of de novo roots initiated compared to wild type.RAP2.7 itself can induce both WOX5 expression by 5 DAW and root initiation and growth by 12-19 DAW, but has no effect on the expression of either YUC1 or NAC1 (Fig. 5).While NAC1 is essential for de novo root regeneration, it is not required for YUC1, WOX5 or RAP2.7 expression, and is not a positive regulator of these genes (Fig. 6).
We show that the NAC1 pathway is not dependent on either auxin or ethylene signalling, whereby NAC1 expression was unaffected during de novo root development in mutants in either auxin signalling (axr1, axr3) or ethylene signalling (ein2, pls) (Fig. 3).The pls mutant is an ethylene hypersignalling mutant that also has low auxin concentrations in the root (Chilley et al., 2006), consistent with its observed low expression of YUC1 and 4, associated with defective de novo root initiation and growth (Figs. 1, 3).The EBS::GUS ethylene reporter was expressed in the petiole but at lower levels in the emerging root in the wild type background (Suppl.Fig. S6), consistent with low ethylene responses being associated with efficient root elongation.Root formation from leaf of the ethylene-insensitive ein2 genotype was similar to wild type, consistent with previous results suggesting that the ein3 mutant, also ethylene-insensitive, was able to form more de novo roots than wild type, and it was proposed that the ethylene hormone is a negative regulator of de novo root formation (Li et al., 2021).While there is good evidence that ethylene inhibits root growth (e.g.Casson et al., 2002;Lewis et al., 2011), the results here for ein2 are not completely in agreement with the data for ein3, as ein2 does not regenerate more or longer roots than wild type, showing that de novo root formation is not promoted by ethylene insensitivity; and indeed there may be fewer root branches from the ein2 leaf than for wild type at 26 d of culture (Fig. 1).This suggests some ethylene sensitivity/signalling activity is required for root initiation over a prolonged culture period, possibly through crosstalk with other hormones (Moore et al., 2024).
The ein2 genotype has no effect on YUC1 and YUC4 expression during leaf explant culture (Fig. 3).Previous results showed that the ethylene-stabilized transcription factor EIN3 directly represses the expression of both WOX11 and WOX5, which are key cell fate-determining genes, providing a possible mechanism for ethylene-mediated repression of de novo root formation (Li et al., 2021).Furthermore, ethylene signalling can reduce auxin accumulation to cause reduced root regeneration (Fig. 1) (Li et al., 2021).This possibility is consistent with our results which show WOX5, YUC1 and WUC4 expression is reduced in the ethylene hypersignalling and low auxin pls mutant; i.e. ethylene responses are repressed to allow root initiation (Fig. 3).Past studies have shown that there is crosstalk between ethylene and auxin in the root, whereby ethylene induces two Trp biosynthetic genes, WEI2/ASA1/TIR7 and WEI7/ASB1, which leads to increased auxin concentrations in the root tip and elongation zone, which is inhibitory to growth (Stepanova et al., 2007).It can be proposed that excess ethylene is inhibitory to the root regeneration process, but ethylene insensitivity has little adverse effect, and signalling needs to be maintained to allow efficient adventitious root formation to occur, likely in part at least through the requirement for ethylene for correct auxin and cytokinin patterning (Moore et al., 2024).
These results suggest a regulatory network in which MDF regulates both auxin-dependent (via RAP2.7 and WOX5, as well as e.g.PIN and PLT genes) and auxin-independent (via NAC1, as well as e.g.SHR and SHR; Casson et al., 2009, Thompson et al., 2023) pathways of de novo root regeneration (Fig. 7).AXR1 is essential for auxin responses and auxin biosynthesis via YUC1 and YUC4, required for root regeneration (Fig. 3), as a central component of the auxin response pathway (Kubalova et al., 2024).Interestingly, AXR3 appears to play a less critical role, as the gain-of-function axr3-1 mutant shows higher frequency root initiation than the axr1 mutant, with less severe negative effects on the expression of WOX5, YUC1 and YUC4 auxin pathway genes than axr1 (Fig. 3).Future studies will explore how MDF is itself regulated during development, to control diverse downstream pathways.

Plant material
Wild type (WT) Arabidopsis thaliana Columbia (Col-0) seeds were obtained from lab stocks, Department of Biosciences, Durham University.All mutant lines were also obtained from lab stocks -axr1, axr3, pls and ein2.GUS reporter lines and Salk mutant lines were obtained from the NASC website (http://arabidopsis.info).
Seedlings were grown in vitro on half strength Murashige and Skoog medium (1/2MS10) (Murashige & Skoog, 1962) containing 2.2 g/l half strength MS medium (Sigma M5519) with 10g/l of sucrose, adjusted to pH 5.7, and 8 g/l agar.For leaf culture, B5 medium (Sigma G5893) was used (3.2 g/l of B5 medium, adjusted to pH 5.7, solidified with 8 g/l agar).All the media were sterilized in the autoclave at 121°C, 1.1 bar for 20 min.
For regeneration experiments, seedlings or isolated leaves were cultured on solidified media and incubated in the growth room or growth cabinets under 16 h light: 8 h dark at 22°C (c.3000 lux).
Seeds were germinated and grown for 12 d on sterile square Petri dishes (size 10 x 10 cm) containing 50 ml of solid half MS medium.Then, after 12 d the first two leaves were removed with sterilized mini scissors and the leaves were transferred to B5 medium using sterilized forceps.For gene expression comparisons at different times, the leaves were collected on 0 h, 3 d, 5 d and 7 d (in the signalling experiment the leaves were collected on 0 h, 24 h, 48 h, 3 d, 5 d, 7 d,) and stored at -80°C prior to RNA extraction.

Gene cloning
An inducible MDF overexpressor transgenic line was designed for comparative studies with the lossof-function mdf-1 mutant and constructed using Gateway cloning (Invitrogen).All the primers were designed by SnapGene software and sequences are available in Suppl.Table S1.The pDONR207 vector was used for all the BP reactions, according to the manufacturer's instructions.
An MDF entry clone from lab stocks, containing the coding sequence of the MDF gene (Thompson et al., 2023), was used to produce the expression clone to overexpress MDF under the control of an estradiol inducible promoter XVE-35S in the vector pMDC7.This was expected to ensure the overexpression of the gene.The purified entry clone was used for LR reaction, cloned into the pMDC7 destination vector, and introduced into chemically competent DH5a cells according to the Gateway instructions.Positive colonies were confirmed by PCR.The correct XVE-35S::MDF expression clone was introduced into Agrobacterium tumefaciens strain GV3101 by the freeze-thaw method, and positive colonies were identified by their ability to grow on LB-agar plates with rifampicin, gentamycin, and kanamycin antibiotics.Arabidopsis was transformed by the method of Clough and Bent (1998).
Plants expressing MDF, which showed 3:1 segregation in the T2 generation (indicative of transgene insertion at a single locus), were screened for homozygous identification.Individual T2 plants were grown until seed development.Then, seeds were sown on hygromycin plates.Lines that showed 100% resistance to the antibiotic and able to overexpress MDF were identified as homozygous T3 seedlings.The line XVE-35S::MDF-H3 was selected as the overexpressor line, from now on will refer to this line as MDF-H3.RT-qPCR was performed to confirm the overexpression after induction with estradiol.
All the primers and entry cloning sequences for Gateway pDONR207 were designed using the Snapgene programme.In addition, different forward primers at different sites in the RAP2 and NAC genes were designed to confirm entry cloning following PCR cloning and sequencing.The list of primers is presented in Suppl.Table S1.Purified entry clones were used for LR reactions, cloned into the pMDC7 destination vector, and introduced into chemically competent DH5a cells according to the Gateway instructions.Positive colonies were confirmed by PCR and sequencing.The correct clones was introduced into Agrobacterium tumefaciens strain GV3101 by the freeze-thaw method, and positive colonies were identified by their ability to grow on LB-agar plates with rifampicin, gentamycin, and kanamycin antibiotics.Arabidopsis was transformed by the method of Clough and Bent (1998).
For leaf induction treatments, the hormone was a final concentration of 5 µM and was sprayed onto the leaf cultured on B5 medium.Then the leaf was collected after 24 h induction.For 0 h leaf the seedling on 11 d culture on half MS medium was sprayed with hormone; and after 24 h the leaf was collected after 12 d culture on half MS medium.

Quantitative RT-PCR (RT-qPCR)
RNA extraction and cDNA synthesis for RT-qPCR was carried out using wild type and mutant or transgenic seedlings essentially as described previously (Rowe et al., 2016), using three biological and three technical replicates.qPCRBIO SyGreen Mix Lo-ROX kit (PCR BIosystems) was used for qPCR analysis, and the reaction mixture and the programme conditions were set according to the supplier's instructions and run on a Rotor-Gene Q Machine (QIAGEN).UBIQUITIN C (UBC), UBIQUITIN10 (UBQ10) or ACTIN2 (ACT2) were used as reference genes.Expression analysis was conducted using the Rotorgene Q Series software v1.7.Relative normalised levels of transcript of each gene were calculated relative to the reference gene and analysed by comparative quantification using an assumption-free, linear regression analysis approach (Ramakers et al., 2003).Primer sequences are listed in Suppl.Table S1.

Histochemistry and microscopy
GUS histochemistry was carried out as described previously (Topping and Lindsey, 1997).Leaf samples were examined using a Zeiss Axioskop compound light microscopy (Carl Zeiss, Cambridge, UK), equipped with a QImaging Retiga-2000r camera (Photometrics, Marlow, UK) using a x10 0bjective.
Root lengths were measured on an Epson Expression 1680Pro flatbed scanner (Epson, UK) set at a resolution of 800 dpi.Measurements were quantified using ImageJ software (Schneider et al., 2012) with the 'smartroot' plugin (Shahzad et al., 2018).

Statistical analysis
Statistical significance was determined using either Student's t-test for independent samples compared to wild type values, with P-values <0.05 (*), P <0.01 (**), P-value < 0.001 (***); or one-way ANOVAs and correlation analysis were performed using SPSS 26.0 software, and Duncan 's new multiple range method was used for significance test (P < 0.05 ).Excel 2018 software was used to tabulate basic data, and Origin 2021 software was used to draw the single factor histogram.One-way ANOVA and correlation analysis was performed using SPSS 26.0 software, and Duncan 's new multiple range method was used for significance test ( p< 0.05 ).n in Fig. 2 showed that there was no significant difference between Col-0 and vamp714 during the period from 12 days edium.On the 21st day, the root length of wild type was mp714, while there was no significant difference in the the two.On the 25th day, the root length and branch antly higher than those of vamp714 ( p < 0.05 ).On the 25th mber of wild type were 12.4±5.1cmand 21±4, respectively.

Role of MDF in the regulation of RAP2.7 gene expression
To determine whether MDF has a role in the regulation of the RAP2.7 gene, RNA was extracted from cultured leaf of wild type, mdf-1 and MDF-OV either treated with estradiol or untreated over an 8 d time course of culture, and RT-PCR was carried out.Bands were quantified using ImageJ software.

Fig. 3 ,
Fig. 3, B-D show very similar patterns of expression of each of WOX5, YUC1 and YUC4 during the culture and regeneration period, in contrast with the pattern for NAC1.WOX5, YUC1 and YUC4

Figure legends Figure 1 .
Figure legends Figure 1.(A) Mean number of regenerated roots and (B) mean root lengths for wild type and mutants at 12, 19 and 26 d after wounding on B5 medium.Values are means of at least 30 samples + SEM.Statistical significance was determined using Student's t-test for independent samples compared to wild type values, with P-values <0.05 (*), P <0.01 (**), P-value < 0.001 (***).

Figure 2 .
Figure 2. (A) Mean number of regenerated roots and (B) mean root lengths for wild type (Col-0) and loss-of-function vamp714 mutant at 12 d, 17d, 21d, 25d after wounding on B5 medium.C, D: RT-qPCR analysis of (C) IAA1 and (D) IAA2 gene expression in wild type and mutant (vamp714) leaf blade and petiole at 0 d, 2 d, 4 d, 6 d and 8 d after wounding, using ACTIN2 as reference gene.Values represents means and error bars are SEM (n = three biological repeats with three technical repeats).

Figure 4 .
Figure 4. (A) Mean numbers of regenerated roots in wild type Col-0, estradiol-induced MDF overexpression (MDFov) and mdf-1 after estradiol treatment and wounding on B5 medium for 12, 19 and 26 d.Inset: (left) MDFov leaf with de novo root after 12 d; (centre) Col-0 leaf with de novo root after 12 d; (right) mdf-1 leaf after 12 d; (B) Mean de novo root length in wild type Col-0, MDFov and

Figure 6 .
Figure 6.(A) Mean numbers of regenerated roots and (B) mean root lengths for wild type, NAC1 overexpressers and nac1 mutants at 12, 19 and 26 d after wounding on B5 medium.Inset in (A):12 day old explants of (left) Col-0, (centre) NAC1 overexpresserand (right) nac1 mutant.Values are means of at least 30 samples + SEM.C-E: RT-qPCR analysis of (C) YUC1, (D) WOX5 and (E) RAP2.7 gene expression in wild type (Col-0), NAC1 overexpresser and nac1 mutant leaf during a culture time course of 0 -7 d after wounding, using UBC as reference gene.Values represents means and error bars are SEM (n = three biological repeats with three technical repeats).Statistical significance was determined using Student's t-test for independent samples compared to wild type values; no significant difference between samples was found.

Figure 7 .Figure S1 .
Figure 7. Summary network model for data integrated with known information on gene-hormone interactions.Wounding induces the AXR/auxin-dependent and MDF/auxin-independent pathways.These interact at the interface of PINs, which require MDF activity and regulate auxin transport produced by the YUC pathway.Auxin activates the WOX genes and MDF activates the independently regulated NAC1 and RAP2.7 pathways.PLS is positively regulated by auxin and is a negative regulator of both ethylene and cytokinin, to suppress the repressive effects of these hormones on root development.

Figure S2 .
Figure S2.PIN1 gene expression in Arabidopsis leaf and regenerating root.

Figure S3 .
Figure S3.PIN3 gene expression in Arabidopsis leaf and regenerating root.

Figure S4 .
Figure S4.PIN7 gene expression in Arabidopsis leaf and regenerating root.

Fig. 1 .
Fig. 1. (A) Mean numbers of initiated de novo root branches and (B) mean root lengths for wild type and mutants at 12, 19 and 26 d after wounding on B5 medium.Values are means of at least 30 samples + SEM.Statistical significance was determined using Student's t-test for independent samples compared to wild type values, with P-values <0.05 (*), P <0.01 (**), P-value < 0.001 (***).

Fig. 5
Fig. 5-3.Quantified RT-PCR data for WOX5 gene expression in wild type, MDF-OV and mdf-1 samples at 0 h, 3 d, 5d and 8 d after germination on B5 medium (A) with estradiol treatment (+E) and (B) without (-E) estradiol treatment.Gel bands were quantified using ImageJ.Values are means ±SEM for three biological repeats.Statistical significance was determined using Student's t-test for independent samples compared to wild type values, with P-values P< 0.05 (*) and P<0.01 (**).

Figure 5 -
Figure 5-4 shows representative gels following RT-PCR to determine RAP2.7 gene expression in wildtype, MDF-OV and mdf-1 samples at different times after germination on B5 medium with and without estradiol treatment, with three technical replicates shown for each

Figure
Figure 4. (A) Mean numbers of regenerated roots in wild type Col-0, estradiol-induced MDF overexpression (MDFov) and mdf-1 after estradiol treatment and wounding on B5 medium for 12, 19 and 26 d.Inset: (left) MDFov leaf with de novo root after 12 d; (centre) Col-0 leaf with de novo root after 12 d; (right) mdf-1 leaf after 12 d; (B) Mean de novo root length in wild type Col-0, MDFov and mdf-1 after estradiol treatment and culture on B5 medium for 12, 19 and 26 d.Values are averages of at least 30 leaves ± SEM.Statistical significance (A,B) was determined using Student's t-test for independent samples compared to wild type values, with P-values P<0.01 (**).Scale bar = 1 cm.(C-H).Relative expression of WOX5, RAP2.7 and NAC1 genes in wild type (Col-0), MDFov and mdf-1 samples at 0 h, 3 d, 5d and 8 d after germination on B5 medium either with estradiol treatment (+E) or without (-E) estradiol treatment to induce MDF overexpression (MDFov).Values are means ± SEM for three biological repeats.Statistical significance was determined using Student's t-test for independent samples compared to wild type values, with P-values P< 0.05 (*) and P<0.01 (**).

Figure 7 .
Figure7.Summary network model for data integrated with known information on gene-hormone interactions.Wounding induces the AXR/auxin-dependent (black text and arrows) and MDF/auxin-independent (blue text and arrows) pathways.These interact at the interface of PINs, which require MDF activity and regulate auxin transport produced by the YUC pathway.Auxin activates the WOX genes and MDF activates the independently regulated VAMP714, NAC1 and RAP2.7 pathways.PLS is positively regulated by auxin and is a negative regulator of both ethylene and cytokinin, to suppress the repressive effects of these hormones on root development.
4. (A) Mean numbers of regenerated roots in wild type Col-0, estradiol-induced MDF overexpression (MDFov) and mdf-1 after estradiol treatment and wounding on B5 medium for 12, 19 and 26 d.Inset: (left) MDFov leaf with de novo root after 12 d; (centre) Col-0 leaf with de novo root after 12 d; (right) mdf-1 leaf after 12 d; (B) Mean de novo root length in wild type Col-0, MDFov and mdf-1 after estradiol treatment and culture on B5 medium for 12, 19 and 26 d.Values are averages of at least 30 leaves ± SEM.Statistical significance (A,B) was determined using Student's t-test for independent samples compared to wild type values, with P-values P<0.01 (**).Scale bar = 1 cm.(C-H).Relative expression of WOX5, RAP2.7 and NAC1 genes in wild type (Col-0), MDFov and mdf-1 samples at 0 h, 3 d, 5d and 8 d after germination on B5 medium either with estradiol treatment (+E) or without (-E) estradiol treatment to induce MDF overexpression (MDFov).Values are means ± SEM for three biological repeats.Statistical significance was determined using Student's t-test for independent samples compared to wild type values, with P-values P< 0.05 (*) and P<0.01 (**).