We additionally profiled the root transcriptomes of 1-month-old tomato plants under well-watered, waterlogged and water-deficit conditions. We hypothesized that genes directly involved in the biosynthesis and deposition of suberin will be highly correlated in both water-deficit and the introgression line population. By combining both introgression lines, waterlogging and water-deficit datasets in a weighted gene correlation network analysis3, we identified modules of co-expressed genes . A module containing 180 genes was significantly enriched in suberin-related genes . This was confirmed by intersection with a public dataset profiling gene expression in tomato DCRi lines . DCRi lines activate suberin-associated genes in the epidermal cells of fruit, which leads to suberization of the fruit surface. The ‘royalblue’ module contains several orthologues of well-known suberin biosynthetic gene families such as glycerol-3-phosphate acyltransferases , 3-ketoacyl-CoA synthases and feruloyl transferases . In addition, putative tomato orthologues of known transcriptional regulators of suberin biosynthesis: AtMYB41, AtMYB63 and AtMYB92 , among others, were found in this module.Although translatome profiles exist for the exodermis, these data do not provide resolution of the developmental gradient along which suberin is deposited. To refine the candidate suberin-associated gene set, we conducted single-cell transcriptome profiling of the tomato root. We used the 10X Genomics scRNA-seq platform to profile over 20,000 root cells. We collected tissue from 7-day-old primary roots of tomato seedlings up to 3 cm from the tip to include the region where suberin deposition is initially observed. Gene expression matrices were generated using cellranger and analysed in Seurat. Once the data were pre-processed and filtered for low-quality droplets,25 liter pot plastic the remaining high-quality transcriptomes of 22,207 cells were analysed.
After normalization, we used unsupervised clustering to identify distinct cell populations . These cell clusters were then assigned a cell type identity using the following approaches: We first quantified the overlap with existing cell type-enriched transcript sets from the tomato root and marker genes extracted from each of the clusters. An individual cluster was annotated as a specific cell type given the greatest overlap between the two sets and a significant adjusted P value . Then, to map gene expression dynamics across maturation, we examined cell-state progression by calculating pseudotime trajectories using a minimal spanning tree algorithm. The tree was rooted in the root meristematic zone , and clusters were grouped into 10 cell types to reflect existing biological knowledge on differentiation of the tomato root . Lastly, genes with previously validated expression patterns in tomato, transcriptional reporters and predicted cell type markers given their function in Arabidopsis, were overlaid on the clusters to refine annotation . Given the successful annotation of these cell types, we focused on the mapped developmental trajectories deriving from a presumedcortex–endodermal–exodermal initial population . Given the suggested link between suberin and drought tolerance, as well as the decreased suberin levels in both control and ABA conditions in our tomato mutants, we hypothesized that the slmyb92 and slasft lines would be more sensitive to water limitation compared with wild-type plants. We subjected 4-week-old well-watered plants to 10 days of water-deficit conditions . Suberin deposition and monomer levels were studied in the root system of slmyb92-1, slasft-1 and wild-type plants in both the water-sufficient and water-limited conditions. Under water-sufficient conditions, suberin deposition was only faintly observed in wild type, and exclusively in the exodermis, while being completely absent in both mutant lines . Consistent with this observation, very low levels of suberin monomers were detected, with no significant differences observed in the very long chain fatty acids, primary alcohols, ω-hydroxyacids, α-ω-dicarboxylic acids and aromatic components of suberin .
Under water limitation, however, deposition of exodermal suberin was increased, with both mutant lines having lower levels than wild type . The transcriptional regulator mutant slmyb92-1 showed a general reduction of most monomer groups compared with wild type. The slasft-1 mutant, in comparison, was primarily depleted in ferulic acid and its esterification substrates, as well as in individual primary alcohols and ω-hydroxyacids . Furthermore, stem water potential, stomatal conductance and transpiration rate were significantly decreased in response to water-limited conditions in both slmyb92-1 and slasft-1 relative to wild type, and leaf relative water content was also decreased in slmyb92-1 . When considering all physiological traits collectively using principal component analysis, slasft-1 showed a milder water-deficit response compared with wild-type plants, while slmyb92-1 was more extreme . These data demonstrate that decreased suberin levels in the tomato root exodermis directly perturb whole-plant performance under water-limited, but not under water-sufficient conditions. Furthermore, changes in specific suberin monomers and the lamellar structure that were observed between the two mutants in response to water-limited conditions may differently influence the extent of the physiological response.In the well-characterized Arabidopsis root endodermis, suberin is deposited as a hydrophobic layer between the plasma membrane and the primary cell wall5 . Developmentally, suberin biosynthesis and deposition occurs as a second step of endodermal differentiation, the first being the synthesis and deposition of the lignified Casparian strip. Suberin serves as an apoplastic barrier and a transcellular barrier, thus contributing to the regulation of the movement of water and solutes to the vascular cylinder.
Our collective observations demonstrate that, relative to Arabidopsis, the components of the pathway are conserved; their spatial localization is distinct; ASFT and MYB92 are critical regulators of suberin biosynthesis given their phenotypes as single loss-of-function mutant alleles, as opposed to their redundancy in Arabidopsis and exodermal suberin has equivalent function to endodermal suberin and can function in its absence . Spatially, in the tomato root exodermis, suberin lamellae are deposited between the exodermal primary cell wall and the plasma membrane all around the cell, similar to the Arabidopsis root endodermis and other suberized apoplastic barriers such as the potato periderm . In a temporally similar fashion to the Arabidopsis endodermis, there is a non-suberized zone at the root tip, a patchy suberized zone in the middle of the root and a continuous suberized zone nearer to the root– hypocotyl junction . We obtained clues to the underlying genes controlling exodermal suberin biosynthesis over developmental time by co-expression, single-cell transcriptome and genetic analyses. Conservation of the genes within the suberin biosynthetic pathway between Arabidopsis and tomato was evident from the functional genetic analysis of SlCYP86B, SlGPAT, SlLACS and SlASFT mutants. Despite the same genes controlling suberin biosynthesis, novelty in tomato is observed with respect to their tomato spatial expression and the critical contribution of SlASFT in primary cell wall attachment and inter-lamella adhesion of the suberin barrier. This phenotype has never been observed in Arabidopsis or potato asft mutant roots. In addition, members of the GPAT4 subclade have been regarded as exclusively involved in cutin biosynthesis, and here, SlGPAT4 was shown to participate in the formation of exodermal suberin . We focused on SlMYB92 as a candidate due to its expression at the end of the exodermal trajectory. Although the precise timing of these trajectories is largely predictive in nature,25 litre plant pot we note that the expression of the biosynthetic enzymes does not completely overlap that of SlMYB92 and suggests that SlMYB92 is not the sole transcriptional regulator of suberin gene expression. Our ability to obtain increasingly differentiated exodermal cells is probably limited by our ability to completely protoplast cells with secondary cell wall deposition. Therefore, the lack of SlMYB41/53/93 expression in the exodermal trajectory does not mean that these genes are not expressed in the exodermis. In Arabidopsis, single loss-of-function mutants of MYB41, MYB53 and MYB93 show no changes in suberin levels, while that of MYB92 shows a delay in suberization. By contrast, this extreme phenotype and compositional profiling in hairy roots and stable lines was observed in tomato when only MYB92 was mutated. The residual suberin levels found in the slmyb92 mutants could be regulated by other MYB transcription factors. Indeed, mutants in tomato orthologues of Arabidopsis MYB41 and MYB63 showed exodermal suberin phenotypes , suggesting that these genes may be expressed in later exodermal developmental stages. ABA-mediated regulation of tomato exodermal suberization is morphologically consistent with what is observed in the Arabidopsis root endodermis, with an increase in both the magnitude of suberin deposition and the proportion of the completely suberized zone, despite the distinct spatial localization.
At least in the case of the slmyb92 and slasft mutant alleles and the ABA assays, this transcription factor and biosynthetic enzyme influence both developmental and ABA-mediated suberin deposition patterns . Further analyses of mutant alleles of the tomato SlMYB41 and SlMYB62 transcription factors will determine whether a coordinated developmental and stress-inducible regulation of suberin biosynthesis is the norm for exodermal suberin. The degree to which this regulation is dependent on ABA signalling, as it is in Arabidopsis , also remains to be investigated. What remains to be identified, however, are the factors or regulatory elements that determine exodermal-specific regulation of these enzymes and transcriptional regulators, as well as how they are activated by ABA and why their activity is ABA-independent in S. pennellii. External application of ABA can be considered a proxy for both drought and salt-stress response. We tested the necessity of suberized exodermis for whole-plant performance under water-limited conditions in mature tomato plants . The strongly reduced response of slmyb92 and slasft plants to ABA was similarly observed upon drought stress. In both experiments, slmyb92-1 and slasft-1 failed to reach fluorol yellow signals and suberin levels equal to those of the control. Under control conditions , we detected overall low suberin levels, which were near the detection limit of 0.003 µg mg−1 and reduced our ability to identify significant differences between the lines. This was consistent with the lack of distinct fluorol yellow signal in mature root sections under water-sufficient conditions . The effect observed in chemical suberin quantification may have also been attenuated by the sample comprising whole root systems with highly branched lateral roots and including root areas with immature suberin. AtMYB92 is also known to regulate lateral root development in Arabidopsis together with its close orthologue AtMYB93, and differences in suberin within different root types are a possibility. Regardless, suberin monomeric levels were clearly decreased in the slmyb92-1 and slasft-1 mutants in a distinct and overlapping fashion in response to water-limited conditions. Consistent with its function, slasft-1 was primarily defective in accumulation of ferulate, primary alcohols and ω-hydroxyacids , while slmyb92-1 had defects in fatty acids and the predominant unsaturated C18:1 ω-hydroxyacids and dicarboxylic acids . The more extreme perturbation of physiological responses in response to water limitation in slmyb92-1 suggests that suberin composed of these fatty acid derivatives plays a role in controlling transcellular-mediated uptake of water . How the transcellular pathway operates in a root system where this apoplastic barrier is located four cell layers from the vascular cylinder remains an important and open question. The role of exodermal suberin as an apoplastic barrier to water flow has been studied in maize and rice, where it was determined as a barrier to water flow, although maize and rice also present a suberized endodermis. Thus, the role of exodermal suberin alone has never been studied with respect to its influence on plant responses to water limitation. The precise role of endodermal suberin, independent of the Casparian strip, has been studied in Arabidopsis, which lacks an exodermis. In 21-day-old, hydroponically grown Arabidopsis plants, the horst-1, horst-2, horst-1 ralph-1 and pCASP1:CDEF1 mutants with a functional Casparian strip but with reduced suberin were monitoredfor the importance of suberin in water relations. These mutants, except for horst-2, have higher Lpr and root aquaporin activity relative to wild type. One can extrapolate that the decreases in stem water potential, transpiration and stomatal conductance relative to wild type in water-limited conditions are a consequence of decreased suberin or perturbations in suberin composition . Assuming that our suberin-defective mutants have higher root hydraulic conductivity, our hypothesis to reconcile our observations with the higher Lpr would be that our mutants have compromised water-use efficiency under water limitation. This could lead to a delayed onset of the drought response such that the water loss is too great to recover by the time stomata are closed. The mechanisms by which this occurs need to be determined and could benefit from further exploration. The levels of lignin in the exodermis and endodermis were not altered in the mutants of the identified transcriptional regulators , and perturbations in endodermal lignin alone have no influence on root hydraulic conductivity in Arabidopsis, thus, lignin plays no role in our observations.