A cell-wall-modifying gene-dependent CLE26 peptide signaling confers drought resistance in Arabidopsis

Abstract Plants respond to various environmental stimuli in sophisticated ways. Takahashi et al. (2018) revealed that CLAVATA3/EMBRYO SURROUNDING REIGON-related 25 (CLE25) peptide is produced in roots under drought stress and transported to shoots, where it induces abscisic acid biosynthesis, resulting in drought resistance in Arabidopsis. However, the drought-related function of the CLE26 peptide, which has the same amino acid sequence as CLE25 (except for one amino acid substitution), is still unknown. In this study, a phenotypic analysis of Arabidopsis plants under repetitive drought stress treatment indicates that CLE26 is associated with drought stress memory and promotes survival rate at the second dehydration event. Additionally, we find that a loss-of-function mutant of a cell-wall-modifying gene, XYLANASE1 (XYN1), exhibits improved resistance to drought, which is suppressed by the mutation of CLE26. XYN1 is down-regulated in response to drought in wild-type plants. A further analysis shows that the synthetic CLE26 peptide is well transported in both xyn1 and drought-pretreated wild-type plants but not in untreated wild-type plants. These results suggest a novel cell wall function in drought stress memory; short-term dehydration down-regulates XYN1 in xylem cells, leading to probable cell wall modification, which alters CLE26 peptide transport, resulting in drought resistance under subsequent long-term dehydration.


Introduction
Plants evolved the vascular system to expand their territory in terrestrial environments.Vascular plants transport water and nutrients from roots to above-ground organs through the xylem.However, land plants are often exposed to severe stresses, including drought, induced by drastic environmental changes.To adapt to drought stress, plants have evolved various response systems, such as dehydration tolerance by alleviating osmotic and/or oxidative damage, and dehydration avoidance by regulating water uptake and water loss, all of which determine drought resistance (1)(2)(3).
Xylem mediates not only the distribution of water but also the long-distance transport of signals.Drought induces the expression of CLAVATA3/EMBRYO SURROUNDING REGION-related 25 (CLE25) in roots, and the synthesized CLE25 peptide is transported to shoots through the xylem, where the peptide promotes the biosynthesis of abscisic acid (ABA), a stress hormone (4).When CLE2 and CLE3 are up-regulated in roots, specific metabolic and defense responses are induced in shoots, respectively (5,6).Okamoto et al. (7) identified a native CLE2 peptide in the xylem exudate, suggesting that these peptides function in specific cellular events as longdistance signals via xylem transport.Our previous study suggested the possibility that xylem transport is affected by plant cell walls that have been modified in response to various environmental stimuli (8).However, it remains unknown whether plant cell wall alterations are involved in drought response.We previously generated ∼50 types of transgenic Arabidopsis lines expressing different xylem cell-wall-related genes under the control of the xylem-specific Arabidopsis Tracheary Element Differentiation-related 4 promoter (9).These transgenic lines displayed alterations in xylem cell walls but did not exhibit any large changes in plant growth.Using a hypocotyl-to-leaf fluorescein transport assay that enables a rapid observation of xylem transport, we found that the xylem transport pattern of fluorescein was affected in a cargodependent manner in the transgenic lines (8).The fluorescence of 5 (6)-carboxyfluorescein (FAM)-labeled CLE25 peptide (CLE25F) was relatively weak in leaf veins, whereas that of CLE26F was strong.CLE26 had the same amino acid sequence as CLE25, except for one amino acid substitution, and did not induce drought response in wild-type (WT) plants (4).Interestingly, the fluorescence from FAM-labeled CLE peptides was altered in the T19 transgenic lines overexpressing Arabidopsis XYLANASE1 (XYN1), which encoded a glycosyl hydrolase family 10 protein with xylanase and/or xylan endotransglycosylase activity (8).This alteration was observed only in the T19 lines among all transgenic lines tested.These results suggest that cell-wall-related and/or cargodependent xylem transport might be involved in plant drought response.
Therefore, in this study, we investigated the relationship between XYN1 and drought stress, considering a possible involvement of xylem transport regulation of CLE25 and CLE26 peptides.The results revealed a stress memory, according to which short-term dehydration down-regulates XYN1 expression, which allows the activation of CLE26 peptide signaling, resulting in greater drought resistance under subsequent long-term dehydration.

XYN1-dependent drought resistance
To understand the relationship between XYN1 and drought stress, we first determined the expression pattern of XYN1 under normal conditions and then examined whether its expression is altered in response to drought stress.To perform this experiment, a pXYN1:: GUS construct was introduced into the WT.The XYN1::GUS signal was detected specifically in xylem cells of leaves and roots (Fig. 1a-c), as previously reported by Suzuki et al. (10).Dehydration diminished the pXYN1::GUS signal in both leaves and roots (Fig. 1d-f).A quantification of XYN1 mRNA levels confirmed a conspicuous decrease in XYN1 expression in response to dehydration (Fig. S2a).This result suggests that XYN1 is potentially involved in the drought stress response in plants.
Next, we examined the dehydration resistance of three Arabidopsis genotypes: loss-of-function xyn1 mutant; xyn1 complementation line carrying a genomic fragment of XYN1 (xyn1/ XYN1); and two T19 lines (T19-1 and T19-4) overexpressing XYN1 under the control of a xylem-specific promoter.Each genotype was treated with soil dehydration for 7 days.The xyn1 mutant, but not xyn1/XYN1, showed a higher survival rate than WT plants under dehydration stress (Figs.1g and S2b).On the contrary, both T19 lines exhibited no significant change in survival rate (Fig. 1h).This result indicates that the loss-of-function mutation of XYN1 increases drought resistance.
Plant responses to drought stress vary depending on many factors, such as the intensity and duration of the stress and plant genotype and growth phase (11).Interestingly, plants have stress memory, i.e. an imprint of the previous stress episodes (12)(13)(14)(15)(16).This imprint/ stress memory is defined as a collection of the structural, genetic, and biochemical modifications that occurred as a consequence of stress exposure, making the plant more resistant to future exposure to the same stress factor.To determine whether stress memory was established under our experimental conditions, WT plants were pretreated with a 2-day dehydration period, followed by 3 days of rehydration, and then subjected to >6 days of dehydration.Drought resistance increased in WT plants that had been pretreated (Figs.1i and S2c), suggesting the presence of stress memory.Next, we compared the pretreated WT plants with xyn1 mutant plants.The drought survival rate of the xyn1 mutant plants was similar to that of the pretreated WT plants (Fig. 1j).In contrast, the pretreatment did not enhance the drought resistance of T19-1 and T19-4 lines (Fig. 1k).Thus, together with the finding that drought stress reduces XYN1 expression, these results suggest that the down-regulation of XYN1 may be required for drought stress memory.

CLE26-mediated drought resistance
Takahashi et al. (4) reported that drought stress induces CLE25 expression in roots.The root-derived CLE25 peptide moves from roots to leaves, where it mediates the expression of osmotic stress-and ABA-inducible genes, resulting in drought resistance.In contrast, the CLE26 peptide, which shows only one amino acid substitution compared with CLE25 peptide and exhibits the same biological function in root growth as the CLE25 peptide (17,18), does not function like CLE25 in drought resistance (4).Our experiment using loss-of-function mutants of CLE25 and CLE26 (cle25-cr1 and cle26-cr1: Yamaguchi et al. (19)) confirmed that CLE25, but not CLE26, was required for drought resistance (Fig. 2a).Next, we examined the involvement of CLE25 and CLE26 in drought stress memory.After the drought pretreatment, the survival rate of cle25 plants was similar to that of WT plants, whereas that of cle26 plants was significantly reduced (Fig. 2b).This cle26 phenotype was complemented by the introduction of the CLE26 genomic fragment, suggesting that CLE26 is required for drought stress memory.CLE25 is transcriptionally upregulated in roots in response to dehydration (4).Therefore, we generated transgenic plants harboring the pCLE26::GUS construct, in which 3 kb upstream and 3 kb downstream sequences of CLE26 had been integrated, and investigated whether CLE26 expression is regulated by drought stress.The pCLE26::GUS signal was observed in vascular bundles, as previously reported (20).Drought stress, with or without the drought pretreatment, showed no clear change in its expression level or pattern in whole plants (Figs.2c-f and S3), implying that the potential function of CLE26 in drought response is not likely under transcriptional regulation.
To investigate the involvement of CLE26 in XYN1-dependent drought resistance, we generated the xyn1 cle26 double mutant.The enhanced survival rate of pretreated xyn1 plants was suppressed by the mutation of CLE26, and the survival rate of the xyn1 cle26 double mutant was similar to that of the cle26 single mutant and the WT (Fig. 2g).On the contrary, the mutation of CLE26 suppressed the pretreatment-dependent enhancement of drought resistance in both WT and xyn1 plants (Fig. 2h).These results suggest that CLE26 peptide signaling is involved in xyn1dependent drought resistance.
CLE25 peptide acts as a long-distance signal from roots to shoots, inducing drought resistance in Arabidopsis (4).ABA is primarily synthesized in leaf vascular tissues through the dehydration-inducible expression of NINE-CIS-EPOXYCAROTENOID DIOXYGENASE 3 (NCED3), which encodes a key ABA biosynthetic enzyme (21).Application of the synthesized CLE25 peptide, but not CLE26 peptide, to roots induces NCED3 expression in shoots (4).In this study, the application of the CLE25 peptide enhanced NCED3 expression in xyn1, xyn1/XYN1, and WT plants (Fig. 3a).The CLE26 peptide did not induce NCED3 expression in WT plants, as previously shown.However, in the xyn1 background, root treatment with the CLE26 peptide induced NCED3 expression in shoots (Fig. 3a).The xyn1/XYN1 complementation line did not show CLE26-peptide-dependent NCED3 induction.As aforementioned, the drought pretreatment enhances drought resistance in a CLE26-dependent manner.Therefore, we examined the inducibility of NCED3 in the shoots of drought-pretreated WT plants following root treatment with the CLE26 peptide.As expected, the CLE26 peptide strongly induced NCED3 expression in the pretreated plants (Fig. 3b).In addition, we observed that CLE26 peptide application to shoots was able to induce NCED3 expression in untreated WT plants (Fig. 3c).These results suggest that the drought pretreatment enhances the movement of the CLE26 peptide, acting in shoots.

XYN1-associated CLE26 peptide transport
To investigate the XYN1-dependent alteration of xylem transport in detail, we observed the accumulation of fluorescein in the leaf veins of xyn1 and T19 plants (Fig. S4) using the hypocotyl-to-leaf fluorescein assay (8).A fluorescent signal in leaf veins was more intense in T19 than in WT plants (Fig. S4a and b), confirming our previous result (8).In contrast, the fluorescein signal was weaker in xyn1 than in WT leaf veins (Fig. S4c and d).These results indicate that the opposite xylem transport phenotypes of T19 and xyn1 plants result from the up-and down-regulation of XYN1 expression, respectively.
Our previous study indicated a difference in fluorescence accumulation between FAM-labeled peptides, CLE25F and CLE26F, in leaf veins (8).Similarly, the application of CLE25F at hypocotyl cut ends produced a relatively weak vein fluorescence in WT as well as xyn1 plants (Fig. 4a and b).In contrast, CLE26F application resulted in a relatively intense vein fluorescence in WT plants but weak fluorescence in xyn1, which was restored in xyn1/XYN1 plants (Fig. 4c-e).Similarly, the drought pretreatment reduced CLE26F fluorescence in WT plants (Fig. 4f and g).These results clearly indicate a tight correlation between a weak CLE fluorescence signal in leaf veins and high drought resistance.This is also supported by shoot-specific NCED3 induction experiments, in which root-applied CLE26 peptide induced shoot-specific NCED3 expression in xyn1 and drought-pretreated WT plants (Fig. 3a and b).
We speculated that the weak CLE-FAM fluorescence signal in leaf veins implies active xylem transport beyond veins in leaves, and conversely, the high intensity of vein fluorescence reflects inactive transport from xylem in leaves, resulting in fluorescence within veins.To detect fluorescence signals spread beyond leaf veins, we used two different methods.First, we observed the fluorescein signal in detached leaves immediately after the transport assay in a high pH solution (Fig. S5) because the fluorescence of fluorescein drastically reduces in acidic pH; importantly, apoplastic pH is acidic in plants (22,23).The xyn1 and pretreated WT leaves showed higher CLE26F fluorescence than WT leaves (Fig. S5a-f).In contrast, T19 leaves showed lower CLE26F fluorescence (Fig. S5g-i).Second, we performed a diffusion assay to Endo and Fukuda | 3 visualize fluorescence over leaf veins by applying a tracer to young seedlings for 1 h.In this assay, CLE25F fluorescence spread and reached the trichomes in both WT and xyn1 plants (Fig. 5a, b, and e).On the contrary, the CLE26F signal did not spread to trichomes in WT plants, but it did in xyn1 (Fig. 5c, d,  and f), which is consistent with the NCED3-inducing activity of CLE25 and CLE26 peptides (Fig. 3a).These results strongly suggest that the XYN1-dependent regulation of CLE26 peptide signaling occurs at the transition of the CLE26 peptide from leaf veins to whole leaves.

XYN1-dependent transport of CLE26 peptide
Signaling molecules, as well as water and nutrients, are transported via xylem and released into the extracellular space.In this study, we revealed that under repetitive drought stress, the CLE26 peptide is transported and functions in inducing drought resistance.
Our previous study indicated that xylem-specific expression or suppression of some cell-wall-related genes alters cargo transport patterns in leaves (8).We here revealed that suppression of XYN1 enhances the release of CLE26 peptide from leaf veins, while the presence of XYN1 retains the CLE26 peptide inside the leaf veins.Plant cell walls can affect the diffusion and concentration of apoplastic molecules by functioning as a permeability barrier and an ion-exchange resin-like reservoir (24,25).Indeed, Zhu et al. (26) reported that the cell wall contributes to iron supply to alleviate iron-deficiency symptoms in plants.Plants have complicated cell wall structures, which are frequently modified in response to environmental cues and developmental signals by a large number of cell-wall-modifying genes, the biological roles of most of which remain unknown (27,28).In this study, we show that XYN1 plays an important role in drought resistance through the modification of xylem transport.Future studies on cell wall dynamics may provide keys to understanding the unique strategies employed by plants to respond to changing environments.

XYN1-dependent drought resistance
Maize WI5, encoding an endo-1,4-β-xylanase, is required for water transport through xylem (29).Furthermore, a loss-of-function mutant of OsXYN1, a rice homolog of W15, shows a defect in drought resistance (30).These facts suggest the importance of xylanaseassociated structural decoration of cell walls under the drought stress condition.We found that the xyn1 mutant exhibited increased drought resistance.On the contrary, XYN1 was immediately down-regulated when soil was dehydrated in pretreatment.XYN1, which encodes a protein exhibiting a xylanase and/or a xylan endotransglycosylase activity (10,31), is preferentially expressed in the xylem.Because XYN1 down-regulation continues at least for >3 days after rehydration, xylem cell walls should be formed without XYN1 during the rehydration period prior to the next drought.This modified cell wall formation may result in resistance to the second drought treatment after rehydration, which is similar to the resistance of xyn1 plants to drought without the pretreatment (Fig. 1).Indeed, downregulation of aspen PtxtXyn10A, the closest homolog of Arabidopsis XYN1, causes various changes in cell wall structure, including cellulose microfibril arrangement (31).On the contrary, xylem-specific XYN1 overexpression altered secondary cell wall structure, resulting in reduced cell wall recalcitrance in Arabidopsis (8).
Preexposure to stress often alters the subsequent responses of plants by producing faster and/or stronger reactions, implying that plants exercise a form of stress memory.Stress memory is involved in modifications at different levels, including morphological, physiological, transcriptional, translational, and epigenetic (12)(13)(14)(15)(16).During recurring dehydration stresses, Arabidopsis plants demonstrate transcriptional memory of the stress, as displayed by the increased rates of transcription and elevated transcript levels of a subset of stress-responsive genes.Plants also display drought stress memory on the Endo and Fukuda | 5 physiological and biochemical levels, which allow changes in photosynthetic rates, phytohormone levels, or plant biomass (16).This study revealed a novel drought stress memory system in which drought memory is written in cell walls as an alteration of cell wall structure, which allows the transport of a stress signal from leaf veins to whole leaves at the second drought stress event.

CLE26-mediated drought resistance
Two similar CLE peptides, CLE25 and CLE26, perform the same function in many plant developmental processes, including root growth (17), xylem development (18), and phloem development (32,33).In drought response, however, these two peptides behave differently (4).Drought stress induces only CLE25 in roots, and the CLE25 peptide is transported to shoots via xylem and induces ABA biosynthesis.In this study, we found that CLE26 also functions in drought resistance by drought stress memory, i.e. by enhancing drought stress in plants that had previously been exposed to short-term dehydration.Previous studies indicated that CLE26 does not show any significant transcriptional up-regulation in response to dehydration (34).Our pCLE26::GUS reporter analysis confirmed this result.Furthermore, when supplied exogenously to roots, the CLE25 peptide, but not CLE26 peptide, induced NCED3 expression in shoots.Once drought pretreated, however, the application of CLE26 peptide to roots induced NCED3 expression in shoots.Indeed, when XYN1 was down-regulated by the drought pretreatment, CLE26F, which was previously retained within leaf veins, diffused beyond leaf veins and even reached trichomes.These results strongly suggest that CLE26 peptide signaling functions as a component of drought memory in leaves.In conclusion, we investigated the involvement of XYN1 in the modification of the transport and function of CLE26 peptide in drought resistance, and successfully indicated a novel drought stress memory, in which short-term dehydration down-regulates XYN1 expression, which promotes CLE26 peptide transport in leaves, resulting in higher drought resistance at the subsequent long-term dehydration event.

Vector construction
Primers used for vector construction are shown in Table S1.To generate xyn1/XYN1 and cle26/CLE26 complementation lines, genomic DNA fragments of XYN1 and CLE26, respectively, spanning a 3-kb sequence upstream of the start codon, coding region, and 3-kb sequence downstream of the stop codon of the corresponding gene were cloned into pBG (35).The resultant plasmids were further converted into pXYN1::GUS and pCLE26::GUS by swapping the coding region of XYN1 and CLE26, respectively, with the GUS gene.Arabidopsis plants were transformed by floral dip with GV3101 (pMP90) containing one of the above-mentioned constructs.

Quantitative PCR
Total RNA was extracted from shoots or leaves using RNeasy Plant Mini (Qiagen), and reverse-transcribed into cDNA using SuperScript III (Invitrogen).Quantitative real-time PCR was performed on LightCycler 480 using TaqMan probe system (Roche).Primers used for real-time PCR are shown in Table S1.UBQ10 was used as a reference gene.

Drought treatment
To examine XYN1 expression level, plants that had been grown for 2 weeks on 1/2 MS-agar under the ShD condition were placed on water-soaked filter paper for 16 h.Plants were sampled before and after the 3-h dehydration on dried filter paper.The XYN1 expression level was analyzed by qPCR.
To perform the drought survival test, plants that had been grown until the four-leaf stage (cle25) or six-leaf stage (other genotypes) on vermiculite under the LoD condition were dehydrated on paper towels.After 1 day, RH was adjusted between 20 and 40%.Plants usually wilted in 7 or 8 days, while the comparison between WT plants with or without 8 or 9 days of pretreatment (Figs.1i and S2c).Then, plants were rehydrated and returned to their normal condition.The number of surviving seedlings was counted (Fig. S1).

Pretreatment
Plants at the four-leaf stage were dehydrated on paper towels for 1 day under the normal growth condition, and then transferred to the dry condition (RH = 20-40%) for 1 day.Subsequently, plants were rehydrated, and grown under normal conditions for 3 days.

NCED3 induction assay
Synthetic CLE peptides used for this assay are shown in Table S2.Plants were grown on 1/2 MS-agar under ShD at ∼90° to allow the roots to elongate on the surface of the agar medium and to keep shoots away from the surface.Then, 30 µL of 1 µM CLE25 peptide, 10 µM CLE26 peptide, or water was applied to the lower parts of roots (Fig. 3a) or leaves (Fig. 3c) of several plants of each genotype.Shoots were sampled after 6 h of incubation under the same growth conditions.To perform the assay with plants grown on vermiculite under LoD, 3 mL of 1 µM CLE26 peptide was added to each pot (Fig. 3b), when the plants had five leaves.Leaves were sampled after 2 days.NCED3 expression was quantified by qPCR.

Tracer assay
The FAM-labeled CLE peptides used for this assay are shown in Table S2.Mixtures of fluorescein, CLE25F, and CLE26F with rhodamine B (0.5 mM each) are used in Figs. 4 and S4.Diluted mixtures (0.1 mM each) and single CLE-FAM (0.2 mM) are used in Figs.S5  and 5, respectively.Plants were grown under the ShD-vermiculite condition (Figs. 4, S4, and S5) or the LoD-agar condition (Fig. 5), except the plants shown in Fig. 4f and g, which were grown and pretreated under the LoD-vermiculite condition.Tracers were applied to the cut ends of hypocotyls of plants at the six-leaf stage (Figs. 4, S4, and S5) or four-leaf stage (Fig. 5).During the 1 h of tracer application, plants were incubated at high RH.The fluorescence of fluorescein, CLE25F, or CLE26F was measured in the first, third, and fifth leaves (Figs. 4 and S4), the third and fourth leaves (Fig. S5), or the third leaves (Fig. 5).Images were taken using M205FA (Leica Microsystems).

Fig. 3 .
Fig. 3. CLE26-peptide-mediated drought response in Arabidopsis.Induction of NCED3 expression in WT, the xyn1 mutant, and xyn1/XYN1 complementation plants by the CLE26 peptide.Water-only, CLE25, or CLE26 synthetic peptide was applied to roots (a, b) or leaves (c) prior to the sampling of shoots.In (b, c), only WT plants were examined and the indicated ones had been drought pretreated (pretreated).Each replicate was normalized to an average of 0 and an SD of 1. Values are presented as the mean ± SD (n = 3 [a, b]; n = 4 [c], open circles).Asterisks and a dagger in (a, b) indicate significant and nearly significant differences compared with the WT with no peptide treatment (**P < 0.01, *P < 0.05, † P = 0.12; Dunnett's test).Asterisks in (c) indicate significant differences compared with the no-peptide treatment (**P < 0.01; Welch's t-test).

Fig. 4 .
Fig. 4. XYN1-associated distribution of the fluorescent-labeled CLE26 peptide in Arabidopsis at 10 min after application.a and b) CLE25F fluorescence in the WT and the xyn1 mutant plants.c-e) CLE26F fluorescence in WT, the xyn1 mutant, and xyn1/XYN1 complementation plants.f, g) CLE26F fluorescence in WT without the pretreatment (f) or with the pretreatment (g).Upper panels show fluorescence images of the indicated lines at 10 min after the application of FAM-labeled CLE25 (CLE25F) (a, b) and CLE26F (c-g) to the cut ends of hypocotyls.Images were taken from the top of plants with six leaves, as described in Ref. (8).Scale bars, 5 mm.Lower panels show the time-lapse analysis of CLE25F (a, b) and CLE26F (c-g) accumulation in leaves.Changes in fluorescence intensity (arbitrary unit) in the first (circles), third (squares), and fifth (triangles) leaves are shown at the indicated time points after peptide application.Leaves are numbered from adult to young.Values are presented as the mean of three observations ± SEM (n = 3).

Fig. 5 .
Fig. 5. XYN1-associated distribution of the fluorescent-labeled CLE peptides in Arabidopsis, at 1 h after application.a-d) Fluorescence images of the WT, the xyn1 mutant plants at 1 h after the application of CLE25F (a, b) or CLE26F (c, d) to the cut ends of hypocotyls.Scale bars, 5 mm.e, f) Changes in fluorescence intensity (arbitrary unit) in the third leaves of WT, the xyn1 mutant, and another xyn1 mutant (SALK_138469) plants at 1 h after application.Values are presented as the mean ± SD (n = 5 [e]; n = 6 or 7 [f], open circles).Asterisks and a dagger indicate significant and nearly significant differences compared with the WT (**P < 0.01, † P = 0.06; Dunnett's test).