The Progression of Xylem Vessel Cell Differentiation is Dependent on the Activity Level of VND7 in Arabidopsis thaliana

Xylem vessels are important for water conduction in vascular plants. The VASCULAR-RELATED NAC-DOMAIN (VND) family proteins, master regulators of xylem vessel cell differentiation in Arabidopsis thaliana, can upregulate a set of genes required for xylem vessel cell differentiation, including those involved in secondary cell wall (SCW) formation and programmed cell death (PCD); however, it is not fully understood how VND activity levels influence these processes. Here, we examined the Arabidopsis VND7-VP16-GR line, in which VND7 activity is post-translationally activated by treatments with different concentrations of dexamethasone (DEX), a synthetic glucocorticoid. Our observations showed that 1 nM DEX induced weak SCW deposition, but not PCD, whereas 10 or 100 nM DEX induced both SCW deposition and PCD. The decreased chlorophyll contents and SCW deposition were apparent after 24 h of 100 nM DEX treatment, but became evident only after 48 h of 10 nM DEX treatment. Moreover, the lower DEX concentrations delayed the upregulation of VND7 downstream genes, and decreased their induction levels. They collectively suggest that the regulation of VND activity is important not only to initiate xylem vessel cell differentiation, but also regulate the quality of the xylem vessels through VND-activity-dependent upregulation of the PCD- and SCW-related genes.


Introduction
Water conduction is crucial for terrestrial plant species living on dry land. To increase water conductive activity, vascular plants developed specialized conductive tissues called xylem vessels, which are characterized by secondary cell wall (SCW) deposition and programmed cell death (PCD) [1,2]. Because of the importance of xylem vessels, great efforts have been made to understand the molecular mechanisms underlying their formation, revealing, for example, that phytohormonal regulation involving auxin, cytokinins, and brassinosteroids is critical for inducing the differentiation of xylem vessel cells [3][4][5][6][7][8].
One of the remarkable achievements of molecular biological research into xylem vessel formation in Arabidopsis thaliana was the identification of the master transcriptional regulators of their differentiation, the VASCULAR-RELATED NAC-DOMAIN (VND) family proteins [9]. The Arabidopsis VND family contains seven members (VND1 to VND7), and overexpression analyses revealed that all VND genes possess the ability to induce xylem vessel cell differentiation [9][10][11][12][13]. A detailed analysis of the pathway functioning downstream of the VND family proteins demonstrated that they directly and/or indirectly upregulate the entire sets of genes required for SCW deposition and PCD [6,[14][15][16][17]. The upregulation of the SCW-related genes by the VND proteins is partially mediated by the secondary master regulators MYB46 and MYB83 [18][19][20][21], and multiple MYB transcription factors also function downstream in the SCW biosynthesis pathway [14,[22][23][24]. Functional analyses of the VND and MYB gene homologs suggested that the VND-MYB-based transcriptional network is well conserved among land plants [14,15], including moss (Physcomitrella patens) [25], pine (Pinus taeda) [26], poplar (Populus trichocarpa) [27,28], and rice (Oryza sativa) [29].
In addition, several other proteins have been reported to transcriptionally or post-transcriptionally regulate the function of the VND family proteins. The LOB-DOMAIN proteins LBD30, LBD18, and LBD15 form a positive transcriptional feedback loop with VND7: the expression of the LBD genes is regulated by VND7, but they, in turn, can upregulate VND7 expression [10,30,31]. VND-INTERACTING 2 (VNI2), a distinct NAC-domain protein, binds directly to VND7 to suppress its transactivation activity [32]. Moreover, a forward genetics approach revealed the importance of the post-translational regulation of VND7 activity to reflect cellular redox condition via the S-nitrosylation of specific cysteine residues in its C-terminal region [33,34]. In addition, light modulates VND1-3 activity to regulate the venation of cotyledons during seedling development [7]. These findings collectively suggest that xylem vessel cell differentiation is controlled by multiple layers of VND-related regulation, optimizing xylem vessel function in response to the environmental conditions [15]. Despite these discoveries, it is not yet clear how the levels of VND activity influence this transcriptional network and promote xylem vessel cell differentiation.
To elucidate the role of VND activity levels in xylem vessel differentiation, we utilized the VND7-VP16-GR system, an inducible VND7-activation line in which ectopic xylem vessel cell differentiation is induced by treatment with a glucocorticoid [12]. In this system, part of the rat glucocorticoid receptor (GR) protein is fused with VND7 and VP16, a strong transcriptional activator domain. The GR interacts with heat shock protein 90 (HSP90) to prevent its entry into the nucleus, but once a glucocorticoid molecule, such as the synthetic glucocorticoid dexamethasone (DEX), binds to the GR, the VND7-VP16-GR protein moves from the cytosol into the nucleus to function as a transcription factor ( Figure 1A) [12]. Previous studies showed glucocorticoid dose-dependent activation of GR-fused chimeric transcription factor proteins in plant cells [35,36], suggesting that the VND7-VP16-GR system would be useful for testing the responses of plant cells to different levels of VND7 activity. Here, we monitored the decreases in chlorophyll contents (an indicator of PCD) and the efficiency of SCW deposition after treatment of VND7-VP16-GR-expressing transgenic plants with different concentrations of DEX. Our data showed that the different DEX concentrations differentially induced xylem vessel differentiation through changes in the expression patterns of genes downstream of VND7, suggesting that the regulation of VND7 activity level is critical in xylem cell formation.

Lower DEX Concentrations Alter the Progression of PCD in VND7-VP16-GR
To elucidate the effects of differences in VND7 activity on the differentiation of the xylem vessel cells, we treated the VND7-VP16-GR seedlings with different concentrations of DEX (Figure 1), using the VP16-GR seedlings as the vector control. Previously, VND7-VP16-GR plants were treated with 10 μM DEX, which effectively induced the differentiation of the xylem vessel cells [12,33]. Here, we used DEX concentrations of 100, 10, 1, and 0 nM ( Figure 1B-I). In the vector control plants VP16-GR, no visible difference was detected by any conditions of DEX treatment ( Figure 1B-E). We observed that 100 nM DEX induced PCD in the 7-day-old VND7-VP16-GR seedlings, causing them to become completely bleached and die ( Figure 1I). By contrast, 10 nM DEX and 1 nM DEX only partially ( Figure 1H) and rarely ( Figure 1G) induced this bleaching of the leaves, respectively. Seedling growth was also differently affected by the DEX concentration; 1 nM DEX did not affect growth ( Figure 1G), but the seedlings treated with 10 nM DEX were smaller than the seedlings treated with 0 nM DEX ( Figure 1H). Thus, 1 nM DEX did not induce PCD in the 7-day-old VND7-VP16-GR seedlings.
We further checked the amount of chlorophyll in the seedling extracts to quantitatively explore the effects of the DEX concentration ( Figure 1J-M). In accordance with the observations described above, the chlorophyll contents of the VND7-VP16-GR seedlings treated with 0 and 1 nM DEX were comparable to the vector controls at all time points ( Figure 1J,K). By contrast, a decreased chlorophyll content was evident after 72 h of the 10 DEX treatment or 24 h of the 100 nM DEX treatment ( Figure 1L,M), although both of these treatments resulted in similar chlorophyll contents after 72 h. These results suggest that the lower the DEX concentration, the slower the progression of PCD. The schematic molecular mode of action of the VND7-VP16-GR system. The chimeric VND7-VP16-GR protein is localized in the cytosol in the absence of a glucocorticoid, since the glucocorticoid receptor (GR) interacts with heat shock protein 90 (HSP90) to prevent its entry into the nucleus. In the presence of glucocorticoid molecules, such as the synthetic glucocorticoid DEX, GR changes its structure to release HSP90, allowing the VND7-VP16-GR protein to localize to the nucleus. This resulted in the activation of the transcriptional activity of VND7 [12]. (B-I) Effects of different DEX concentrations on the growth and morphology of VP16-GR (vector control) and VND7-VP16-GR seedlings. The 7-day-old VP16-GR (B-E) and VND7-VP16-GR (F-I) seedlings were treated with 0, 1, 10, and 100 nM DEX for 3 d. Bar = 5 mm. (J-M) Changes in the chlorophyll amounts of VP16-GR and VND7-VP16-GR seedlings. The 7-day-old VP16-GR and VND7-VP16-GR seedlings were treated with 0, 1, 10, and 100 nM DEX, and sampled after 0, 24, 48, and 72 h. The data are shown as means ± SD (n = 4). Asterisks indicate statistically significant differences between VP16-GR and VND7-VP16-GR (Student's t-test, p < 0.05).

Lower DEX Concentrations Affect the Progression of SCW Deposition
Next, we examined the DEX-concentration dependency of ectopic SCW deposition in the VND7-VP16-GR cotyledons. The TOMEI-I method [37] was utilized for the visualization of SCW deposition ( Figures S2 and S3) [38]. Both methods demonstrated that the 1, 10, and 100 nM DEX treatments could induce the ectopic deposition of SCW after 72 h of treatment, although the thickness of the deposited SCW, as reflected by signal intensity, was lower in the samples treated with 1 nM DEX than those treated with 10 or 100 nM DEX ( Figure 2I,N; Supplementary Figure 2S). The time-course analysis showed that the relative proportion of SCW positive region was nearly reached its peak after 24 h of the 100 nM DEX treatment ( Figure 2Q,U), while the SCW-positive region was less than 50% of leaf area after a 24 h treatment with 10 nM DEX and peaked at 48 h ( Figure 2L,M,U). These observations clearly indicate that the DEX concentration affects not only the progression of PCD, but also that of SCW deposition.
The seedling growth and chlorophyll content suggested that 1 nM DEX did not induce PCD in VND7-VP16-GR ( Figure 1G,K); however, this treatment caused the ectopic deposition of a thin layer of SCW in the VND7-VP16-GR cotyledons, although not as much as was induced by the 10 nM or 100 nM DEX treatments ( Figure 2I,U). These findings could reflect a difference in the VND7 activity-dependency between PCD and SCW deposition. The genes encoding the PCD-related enzymes are directly activated by the VND proteins, whereas the SCW-related genes are under the control of not only the VND proteins [6,16,17], but also MYB46 and MYB83 [20,39]. It is therefore possible that the different effects of the 1 nM DEX treatment could be caused by the different transcriptional regulatory networks involved in PCD and SCW deposition.

Different DEX Treatments Altered the Expression Patterns of the Genes Downstream of VND7
To elucidate the VND7-activity-dependent activation of the genes downstream of this protein, we performed a quantitative RT-PCR analysis on selected genes known to function in this pathway (Figure 3; Supplementary Figure S4) [16,24,33,40]. We selected the PCD-related enzyme genes XYLEM CYSTEINE PEPTIDASE1 (XCP1) and METACASPASE9 (MC9) [40,41], the SCW-related transcription factor-encoding genes MYB46 and MYB63 [18][19][20]24,42], and the SCW biosynthetic genes, including CELLULOSE SYNTHASE7/IRREGULAR XYLEM3 (CESA7/IRX3) for cellulose biosynthesis, IRREGULAR XYLEM8 (IRX8) for xylan biosynthesis, and CAFFEOYL COENZYME A ESTER O-METHYLTRANSFERASE7 (CCoAOMT7) for lignin biosynthesis [43][44][45][46]. In addition, we determined the expression of LBD30 [31] and the endogenous VND7 gene [9]. Our time-course quantitative RT-PCR analysis revealed clear differences in the expression patterns of the genes downstream of VND7 following the different DEX treatments (Figure 3). The 10 and 100 nM DEX treatments led to a greater than ten-fold upregulation of these genes; however, 1 nM DEX did not strongly increase the gene expression levels (Figure 3; Supplementary Figure S4). Moreover, except for CCoAOMT7, the expression peaks of the genes downstream of VND7 were higher and reached more quickly in the plants treated with 100 nM DEX than in the plants treated with 10 nM DEX (Figure 3). These observations were consistent with the chlorophyll contents and SCW depositions observed following these treatments (Figures 1 and 2); therefore, the progression of PCD and SCW is likely directly linked to the VND7-regulated activation of these downstream genes.
Previous studies revealed that LBD30, MYB46, XCP1, and MC9 are direct targets of VND7 [6,16,17], while MYB63, CESA7, and IRX8 are direct targets of MYB46 [18,21]. In addition, the promoter region of CCoAOMT7 contains the sequence ACCAACC, which can be recognized by both MYB46 and MYB83 [18,21,42], suggesting that this gene might be directly regulated by these transcription factors. To test the effects of the DEX treatments on these transcriptional regulatory relationships, we performed a hierarchical clustering analysis on the gene expression patterns (Figure 4). The results clearly separated the tested genes into three groups based on their similarities in expression pattern: group 1 contained VND7, LBD30, XCP1, and MC9; group 2 comprised MYB46, CESA7, IRX8, and CCoAOMT7; and group 3 contained MYB63 alone (Figure 4), reflecting the transcriptional hierarchy of VND7, MYB46, and MYB63 [14,15,19,47]. The level of VND7 activity can therefore affect the induction of gene expression in a transcriptional regulatory hierarchical manner. We also observed that the MYB63 gene, encoding one of the third-layer transcription factors in the VND-MYB-based transcriptional network regulating lignin biosynthesis [24], showed an expression pattern distinct from that of the other genes at the transcriptional level (Figures 3 and 4). MYB63 expression was induced after 24 h and 48 h of the 100 nM and 10 nM DEX treatments, respectively, and continuously increased during 72 h of these treatments, whereas the other genes were transiently upregulated by these treatments (Figures 3 and 4). These observations support the hypothesis that the step-wise induction of the MYB transcription factors can be effective for regulating lignin biosynthesis temporally and flexibly to adapt the plant to its environmental conditions [15,23].

DEX Treatment
The 7-day-old seedlings were treated with 0, 1, 10, or 100 nM DEX (Sigma-Aldrich, St. Louis, MO, USA), as previously described [33,48]. Briefly, the DEX solution was poured directly onto the seedling plate to cover the seedlings, after which the plates were incubated under continuous light at 22 °C. The seedlings were then sampled at the indicated time points.

Chlorophyll Content Measurement
Chlorophyll was extracted from the DEX-treated seedlings using an overnight incubation in N,N-dimethylformamide at 4 °C. The chlorophyll contents were calculated as described by [49], using absorbance data obtained from a spectrometry analysis using a DU640 spectrophotometer (Beckman Coulter, Brea, CA, USA).

Microscope Observation
The DEX-treated seedlings were sampled and fixed in a mixture of 90% (w/v) ethanol and 10% (w/v) acetic acid. The samples were then washed three times with 90% (w/v) ethanol and stained with a 5 μg ml -1 propidium iodide (PI) solution. The PI-stained samples were cleared using TOMEI-I, as described by [37], and were then observed using a confocal microscope (FV-10i; Olympus, Tokyo, Japan).

Calculation of the Relative Proportion of SCW-Positive Cell Regions in the Cotyledons
To calculate the relative proportion of SCW deposition in cotyledons, the area displaying a PI signal was manually measured, or semi-automatically measured based on machine learning [38]. For the manual measurements, the confocal images were processed using ImageJ (https://imagej.nih.gov/ij/index.html) to obtain a maximum intensity projection image. The obtained images of partial cotyledon regions were then merged to generate an image of the whole cotyledon using the MosaicJ plugin [50]. The area of the merged image displaying the SCW signals was measured, and the relative proportion of the SCW-positive region as a proportion of the total area of each cotyledon was calculated.
For the semi-automatic detection of SCW deposition, the previously reported image analysis framework for object detection in transmission electron microscopic images was used [38]. First, square regions of interest (ROIs) (50 × 50 pixels) were tiled on a maximum intensity projection image of the PI-stained cotyledon (3040 × 3888 pixels) ( Figure S2A, top). Next, the ROIs with a maximum intensity of 254 or 255 were selected as candidates for SCW-positive regions ( Figure S2A, bottom). Cluster analyses of the candidate ROIs were performed using a self-organizing map (node number: 15 × 15) with statistical geometric features [51]. The feature set was optimized to detect the SCW based on interactive annotations and iterative clustering with a genetic algorithm ( Figure S2B) [38]. After the feature selection, the SCW-positive regions were automatically detected in the original PI images. The endogenous SCW-positive regions were eliminated using the mathematical morphological operators of erosion and dilation ( Figure S3A).

Quantitative RT-PCR Analysis
The quantitative RT-PCR analysis was performed as described by [16]. Briefly, the seedlings were collected at the indicated time after the DEX treatment, then ground in liquid nitrogen using a TissueLyser II (Qiagen, Hilden, Germany). Total RNA was isolated using an RNeasy Plant Mini Kit (Qiagen), according to the manufacturer's instructions. The cDNA synthesis was performed with 1 μg total RNA using a Transcriptor First-Strand cDNA Synthesis kit (Roche, Basel, Switzerland). The quantitative PCR was carried out using a LightCycler 480 DNA SYBR Green system (Roche). The expression levels of the genes of interest were normalized to the expression level of the internal control, UBIQUITIN10 (UBQ10). The primers used in this study were shown in Table S1 [ 16,24,33,40].

Hierarchical Clustering Analysis of the Expression of Genes Downstream of VND7
The quantitative RT-PCR data were normalized, then used for a hierarchical clustering analysis (Pearson correlation; average linkage clustering method) performed using the software MeV ver. 4.9.0 [52]. The heatmap visualization of the quantitative RT-PCR data was also created using MeV ver. 4.9.0.

Conclusions
Our data suggest that the level of VND activity can affect the progression of xylem vessel cell differentiation. The different levels of VND activity can differentially induce the PCD-and SCW-related genes (Figures 3 and 4), resulting in differences in the progression of PCD and SCW deposition (Figures 1 and 2). These findings also suggest that the regulation of VND activity is important not only to control the initiation of the molecular programs of xylem vessel cell differentiation, but also regulate the functional quality of the xylem vessels to change how PCD and SCW deposition progress. Previous works have revealed several post-translational regulatory mechanisms affecting VND7 activity, such as the physical interaction with the repressor protein VNI2 [32] and the S-nitrosylation of the VND7 protein [33,34]; however, the molecular evidence to link these regulatory mechanisms to xylem vessel quality is still not well studied. Further studies on the regulatory mechanisms of VND activity could provide new opportunities for biotechnologies to increase crop yields and plant biomass by influencing xylem vessel quality.
Supplementary Materials: The following are available online at www.mdpi.com/xxx/s1, Table S1. Primers used in this study; Figure S1. Secondary cell wall (SCW) deposition in VND7-VP16-GR following a 24 h treatment with 10 nM dexamethasone (DEX); Figure S2. Image processing workflow for the semi-automatic detection of the secondary cell wall (SCW)-positive regions; Figure S3. The relative proportion of secondary cell wall (SCW)-positive cells within VP16-GR and VND7-VP16-GR seedlings; Figure S4. Quantitative RT-PCR analysis of the expression of the genes downstream of VND7 in VND7-VP16-GR seedlings treated with 0 and 1 nM dexamethasone (DEX).