Analysis of the salt-stress response at cell-type resolution



    1. Temasek Lifesciences Laboratory, 1 Research Link, National University of Singapore 117604, Republic of Singapore and
    2. Department of Biological Sciences, National University of Singapore 14 Science Drive 4, 117543, Republic of Singapore
    Search for more papers by this author

J. R. Dinneny. e-mail:


Albert Einstein said that one should make everything as simple as possible, but not simpler. Likewise, biologists have often relied upon simplifying assumptions to study the fundamental properties of biological systems. To understand how complex multicellular plants respond to stresses such as high salinity, organ or organism-scale experiments have often been the standard and have led to important discoveries. However, recently, through the use of cell-type-specific analyses, it has become clear that most salt-stress regulation occurs at the scale of the cell or tissue type. Salt stress has been revealed to cause complex changes in growth, development and physiology that are dynamically regulated in both space and time. Thus, the next most important discoveries regarding how plants perceive, respond and adapt to this environmental stimulus will require approaches that enable high-resolution spatial and temporal observations to be made. In this review, we highlight studies taking both a genomic and sub-genomic approach to understand the salt response at high spatial resolution. These present and future studies will help lead to a more sophisticated understanding of the root as a complex system, which integrates information from different cell layers to generate synchronized changes necessary for the survival of the plant.


Perception, response and adaptation to changes in the environment are fundamental processes that occur in all organisms. From the simplest microbe to the most complex multicellular, multi-tissue plants and animals, evolution has selected for robust mechanisms to generate appropriate context-specific changes in physiology, growth and development (Lopez-Maury, Marguerat & Bahler 2008). Due to the fundamental nature of the interaction between an organism and its surroundings, environmental stimuli are powerful tools for exploring the core properties of biological systems. Nowhere is this more evident than in studies on single-celled organisms such as yeast and Escherichia coli, where complex genome-scale models for gene regulation, termed transcriptional networks, have been developed (Harbison et al. 2004; Seshasayee et al. 2006). These models provide tremendous insight into how a change in the external environment leads to a change in internal physiology. Similar studies have been performed in multicellular organisms (Zeller et al. 2009). Research into the mechanisms controlling the response of plants to abiotic and biotic stimuli have led to the identification of key regulatory pathways and have provided insight into signal transduction and the hormonal regulation of development (Wolters & Jurgens 2009). While organ- and organism-scale experiments are useful for understanding regulation that is organ wide or organism wide in scope, it is clear that, ultimately, all regulations must occur at the cellular level.

In single-celled organisms, no further reduction in experimental scale is necessary to understand cell-type-specific biology, although examining individual cells instead of populations of cells has proven powerful for exploring the role of variability in biological systems (Rosenfeld et al. 2005). Plants, however, are intricate compositions of cell and tissue types that enable more complex regulation. Many important biological functions are non-cell autonomous. For example, nutrient absorption from the soil requires specialized cell types in the epidermis, termed hair cells, to increase available surface area (Grierson & Schiefelbein 2002; Scheres, Benfey & Dolan 2002). Endodermal cells develop a specialized cell wall modification termed the casparian strip to seal off the apoplast to limit passive absorption of solutes by the central stele, and the stele contains highly modified and terminally differentiated vascular cells to enable transportation to other tissues and organs. In the shoot, where guard cells regulate gas exchange between the internal mesophyll cells and the external environment, there is a dependency upon subsidiary cells to exchange K+ ions to enable the dynamic regulation of turgor pressure necessary for opening and closing the stoma (Raschke & Fellows 1971). Thus, each cell type performs unique functions and these functions integrate with activities in other cell types to generate a coordinated and complex function. Analysis of the environmental regulation of these functions at the organ or organismal scale provides little insight into these complexities, but highlights biological processes that are common between cells. While understanding these processes is important, perhaps, the most interesting biology lies in the interactions between cells and in understanding how it is that one cell responds differently to a stimulus than its neighbour. In this review, we highlight research examining salt-stress regulation at the cell-type-specific level and include a description of work using both genomic and sub-genomic approaches.


A major roadblock in the past for understanding cell-type-specific regulation of salt responses has been the availability of methods for measuring biological information from individual cell types. In plants, cells cultured in isolation will de-differentiate with little resemblance to their progenitors (Takebe, Labib & Melchers 1971; Zhao et al. 2001). Nevertheless, important insights have been made using tools that enable cell-type-specific perturbation or by studying processes, such as gravitropism, which are known to be regulated in a cell-type-specific manner.

Cell-type-specific calcium oscillations are involved in abiotic stress response

Calcium is an important secondary messenger of environmental stimuli in plants. Complex changes in concentration have been linked to guard-cell movements with the frequency of oscillating high and low cytosolic Ca+ being linked to different responses (Allen et al. 2001). Ca+ also plays an important role in the salt-stress response, with the most mechanistic insight coming from studies on the SOS pathway. In this pathway, the SOS2 protein kinase is activated by a NaCl-mediated Ca+ increase via interaction with the SOS3 calcium sensor (Halfter, Ishitani & Zhu 2000; Liu et al. 2000; Xiong & Zhu 2002). Activated SOS2 then regulates the H+/Na+ anti-porter SOS1 to promote the export of sodium ions out of the cell (Shi et al. 2000; Quintero et al. 2002).

One of the first extensive studies directly aimed at determining whether cell identity influences environmental response was described by Kiegle et al. (Kiegle et al. 2000). In this work, the authors utilized the calcium-sensitive fluorophore, aequorin, to measure dynamic fluctuations in internal calcium levels. Previous studies by the research team had utilized a similar system to examine Ca+ dynamics in whole seedlings; however, it was unclear which cells were contributing to the calcium fluxes (Knight, Trewavas & Knight 1997). To circumvent these issues, Kiegle et al. expressed aequorin in individual cell types using the GAL4/UAS trans-activation system developed for Arabidopsis in the lab of Jim Haseloff at the University of Cambridge. This system has been extensively used in the characterization of developmental pathways in plants and animals and is effective for driving the expression of transgenes in tissue layers of interest (Brand & Perrimon 1993; Swarup et al. 2005).

Using this system, Kiegle et al. were able to monitor internal Ca+ fluctuations at high spatial and temporal resolution using a luminometer. With this technique, they found interesting variation in the responses of different cell types to diverse abiotic stresses. Cold stress, which presumably should affect each cell layer equally, led to peak Ca+ responses that were statistically indistinguishable between the cell layers. In contrast, drought stress led to the strongest response in the epidermis, while internal cell layers were less responsive. Intriguingly, salt stress resulted in high levels of response in both the epidermis and internal endodermal cell layers, but had a much more muted effect on the pericycle. Thus, each stimulus has a specific interaction with the various cell layers. The salt response, in particular, had an interesting effect on endodermal and pericycle Ca+ levels and was observed to lead to a series of secondary fluctuations after the first wave of elevated Ca+. In the endodermis, these oscillations were persistent, lasting for several minutes after the imposition of the stress. The endodermal cell layer plays an important role as a selective barrier for solute absorption by the root, and it is tempting to speculate that this unique response may be evidence for a special signalling role during salt stress.

Salt-stress inhibition of gravitropism: a possible avoidance mechanism

Roots are recognized as displaying certain tropisms. Gravitropism enables the root to grow along the gravity vector (Palme, Dovzhenko & Ditengou 2006), while hydrotropism (Eapen et al. 2005; Takahashi, Miyazawa & Fujii 2009), which is growth towards a region with relatively high moisture, can act in opposition to gravitropism. Recently, the possibility of an ‘anti-tropism’ or avoidance response has also been identified in Arabidopsis roots treated with high salinity (Sun et al. 2007; Dinneny et al. 2008). Gravitropism is regulated through the complex modulation of auxin transport and response in selected cell types; thus, salt stress is likely to affect this pathway through cell-type-specific regulation. Sun et al. and Dinneny et al. have shown that when roots are exposed to salt, the response to gravity is suppressed and root-tip bending is induced. Together, these responses may serve as an avoidance mechanism enabling the root to steer away from high-salt environments.

Sun et al. explored some of the potential mechanisms by which salt may inhibit gravitropism and found that amyloplast development was dynamically suppressed. Amyloplasts are starch-containing organelles in the columella root cap and are thought to play an important role in signalling the direction of the gravity vector through sedimentation (Leitz et al. 2009). Transportation of the plant hormone auxin also plays a key role in gravitropism. A select set of auxin-efflux transporters including PIN2 are necessary to generate gradients of auxin in the root that enable differential cell expansion in response to a change in the gravity vector (Muller et al. 1998). Sun et al. observed a significant decrease of both PIN2 transcript accumulation and protein levels within hours of salt induction. While PIN2 protein levels were stably reduced, PIN2 gene expression recovered after 24 h. PIN2 expression is also repressed by salt stress in the cortex cell layer in the analysis (described below) performed by Dinneny et al. Thus, regulation of PIN2 could potentially play an important role in gravity suppression, with changes in protein stability more likely being the primary cause of the long-term effects. Additional experiments testing whether stabilized PIN2 protein can prevent the salt-dependent suppression of gravitropism will be necessary to test this hypothesis, however.

Proof of Principle: cell-type-specific expression of HKT1;1 is necessary to confer salt tolerance

While it is intriguing to hypothesize that cell-type-specific regulation of the salt-stress response is important for controlling adaptive changes in physiology, this hypothesis is also largely untested. Experiments substantiating such a hypothesis would need to show that ubiquitous or ectopic activity of a salt tolerance mechanism puts the plant at a disadvantage. Several studies in the past have shown that constitutive expression of several genes can confer salt tolerance; however, none of these studies have compared the effect of constitutive expression with cell-type-specific expression (Apse et al. 1999; Shi et al. 2003). Recently, such experiments have been performed using the sodium transporter, HKT1;1 (Maser et al. 2002; Moller et al. 2009). Taking advantage of the GAL4/UAS trans-activation system (described above), Møller et al. tested whether tissue-specific expression of HKT1;1 could limit the amount of salt transported to the shoot. Reduced transportation of Na+ to the shoot is associated with salt tolerance (Møller & Tester 2007); salt tolerant species and plant strains often have reduced salt load in the shoot. In order to retrieve Na+ from the transpiration stream, enhancer trap lines were used that drive expression in the stele (vascular cylinder). Møller et al. confirmed that expression of HKT1;1 in cells of the stele increased their ability to take up Na+ by performing patch-clamp experiments on pericycle protoplasts isolated from roots in which HKT1;1 was expressed in this cell type.

Expression of HKT1;1, either in the pericycle or in cells of the vascular bundle (excluding the pericycle cells) resulted in increased Na+ uptake in the root, but reduced transport to the shoot. In turn, this led to healthier plants with increased total dry biomass under salt-stress conditions. Using X-ray microanalysis, expression of HKT1;1 in the pericycle was found to drive greater accumulation of Na+ in the pericycle and xylem parenchyma tissue layers, when measured in tissue-culture grown roots.

Importantly, Møller et al. also conducted an analysis of the effects of constitutive HKT1;1 expression on salt tolerance and found that this led to stunted growth, chlorosis and higher levels of Na+ in the shoot. These results are particularly interesting as the hkt1;1 loss-of-function mutation also causes salt hypersensitivity (Maser et al. 2002). Møller et al. hypothesize that constitutive expression may cause increased absorption of Na+ from the growth medium through the cortex and epidermis. It will be important in future studies to drive HKT1;1 expression in other tissue layers, beside the stele, to determine exactly which cells lead to salt hypersensitivity in 35S::HKT1;1 plants. Together, these results illustrate the complex effects that Na+ accumulation has on plant physiology and validates the hypothesis that cell-type-specific regulation is necessary in some processes to enable salt tolerance. It will be interesting to perform similar experiments with other genes that confer salt tolerance when over-expressed. We may find that each gene has a particular tissue in which expression is most effective and, perhaps, other tissues where the effect is detrimental.


The cell sorting methodology

The adaptation of fluorescence-activated cell sorting (FACS) technology for use in the root has revolutionized the study of biological processes at the cell-type-specific level in plants. The method begins with the generation of a reporter line expressing a fluorophore such as green fluorescent protein (GFP) in a cell type of interest (Harkins et al. 1990; Birnbaum et al. 2003, 2005). Next, thousands of roots expressing this reporter are grown in tissue culture, harvested and treated with an enzyme cocktail containing cellulases and pectinases, which act to digest the cell wall and release the protoplasts. FACS is then used to purify protoplasts based on their fluorescent properties. RNA is isolated, processed and hybridized to microarrays to analyse the transcriptome. In the earliest studies, the number of available reporter lines was limited, and therefore transcriptional profiles were obtained from overlapping tissue types. These data led to the development of additional GFP reporter lines, which more recently enabled the characterization of 14 different cell types (Brady et al. 2007). This high-resolution expression map can be used to infer the expression pattern of nearly 24 000 genes in the root and has revealed the complex spatial manner by which biological functions are controlled.

Several important questions remain to be addressed with these data sets, however. For example, how much of the cell-type-specific regulation observed is environmentally dependent? It is important to recognize that even standard growth conditions represent an environmental condition that may affect the cell-type-specific transcriptome. Thus, how much of the cell-type-specific transcriptome identified in these studies stably indicates the function of that cell-type? Also, how well does co-expression of a set of genes predict co-regulation? An important assumption of much of the analysis performed on these data sets is that genes that are expressed in a similar manner are also regulated by similar mechanisms. To address some of these issues, environmental stimuli can be useful for perturbing the system and observing the effects. If genes are co-regulated, they will likely show similar transcriptional changes under environmental stress. Furthermore, environmental stimuli can be useful for identifying genes that are robust markers for specific cell lineages and potential regulators of that fate.

Characterization of the cell-type-specific response to salt stress

To begin to address whether cell-type-specific regulation plays an important role in the salt-stress response, Ma and Bohnert utilized the tissue-specific data set generated by Birnbaum et al. to determine whether stress-regulated genes tend to have particular expression patterns under standard conditions (Ma & Bohnert 2007). Their analysis utilized a fuzzy K-means clustering algorithm (Gasch & Eisen 2002) to identify sets of co-expressed genes based on the AtGenExpress abiotic stress microarray data set (Zeller et al. 2009) or the Birnbaum et al. tissue-specific data set (Birnbaum et al. 2003). A comparison of cluster membership in the two data sets was then performed. While this analysis cannot be used to understand how stress affects the expression pattern of a gene, it can determine if there is any pre-existing transcriptional state that is common among stress-regulated genes. Interestingly, the authors found that genes that grouped together based on environmental responsiveness were usually distributed together in a small number of clusters based on the tissue-specific data set. For example, genes expressed strongly in the root cap and epidermis were frequently coordinately repressed by abiotic stress. Many of these genes encode ribosomal proteins or are associated with protein translation. These findings show that stress-regulated genes can control tissue-specific biological functions, at least under standard conditions, and suggests that there may be common pathways that regulate spatial expression and stress response. Interestingly, the observation that each cluster defined by the AtGenExpress data set subdivided into several smaller clusters according to the tissue-specific data set suggests that the common expression patterns observed under one set of conditions may not be a consistent predictor of co-regulation. Combining data sets, as employed by Ma and Bohnert, provides a useful approach for identifying high-confidence regulatory modules, which is a necessary first step in generating transcriptional network models for environmental regulation.

Generating a cell-type-specific response map for salt stress

To understand the role of tissue-specific regulation in the salt-stress response, it is ultimately necessary to analyse the biology of the root at cell-type resolution. Through the research of a few labs, FACS has now been used to study the effects of salt stress (Dinneny et al. 2008), iron deprivation and nitrogen stimulation in roots (Gifford et al. 2008). These data sets have led to the identification of many fold-more environmentally regulated genes than whole-organ analyses and to the identification of new biological functions involved in these responses. Interestingly, for all environmental stimuli examined to date, cell-type-specific regulation is the primary mechanism controlling most transcriptional changes.

The salt-stress microarray data sets described in Dinneny et al. were generated using three different sampling techniques. First, a time course analysis of the salt-stress response was performed using whole seedling roots to determine a preferred time frame for acquiring the two other data sets examining spatial regulation. The time course data set revealed that the salt response comprises of waves of transcriptional activity with very few genes being stably up- or down-regulated over the entire time course, which is in agreement with published microarray-based expression analyses in other species (Kawasaki et al. 2001; Wang et al. 2003). One of the first transcriptional events, observed 1 h after salt treatment, is the up-regulation of a large suite of transcription factors. Since one of the main areas of interest in the salt-stress field is the transcriptional regulation of the response, this time point was chosen as the focus for the remainder of the study. Other salt-regulated biological functions identified included those associated with photosynthesis. These surprising results were confirmed in salt-treated roots, which show a large accumulation of chloroplasts. It is unclear what the function of this response may be; however, these organelles are key sites for the metabolism of reactive oxygen species (ROS) during stress (Mittler et al. 2004). The accumulation of chloroplasts may be activated by the ROS and they may play an important role in removing these damaging molecules.

Salt-regulated gene expression was examined spatially along two orthogonal axes of the root. Along the longitudinal axis, cells undergo stereotypical cellular processes as they are generated and subsequently differentiated (Fig. 1) (Scheres et al. 2002). Cells closest to the root tip, where the stem-cell niche is located, compose the meristem and are in a state of active cell division. Cells then cease division and undergo rapid anisotropic expansion in the subsequent elongation zone. Finally, cells cease elongating and begin to adopt their final shapes and functions within what is termed as the maturation zone. The beginning of this zone is visually marked by the initiation of root hairs from the trichoblast cell lineage in the epidermis. Morphological markers and other optical properties were used to dissect roots, treated with standard media or supplemented with salt, into four longitudinal zones, which were then employed for microarray-based expression analysis. This data set revealed that most transcriptional changes are regulated in a longitudinal zone-specific manner. The second-most frequent transcriptional changes co-occurred in neighbouring longitudinal zones. Very few genes showed significant transcriptional changes in all four zones that were assayed. Genes associated with cell-wall loosening are up-regulated in the meristematic zone, which may be important for promoting growth during osmotic stress. The elongation zone is the transcriptionally most active zone, which may correspond to changes in cell shape, growth rates and root hair initiation that occur in response to salt stress (Burssens et al. 2000; West, Inze & Beemster 2004; Dinneny et al. 2008).

Figure 1.

Biological functions regulated along the longitudinal axis of the root. Diagram depicting the four longitudinal regions of the Arabidopsis root profiled in Dinneny et al. Biological functions associated with growth under standard conditions are shown on the left side, while salt-regulated biological functions are listed on the right.

The longitudinal zone data set revealed the important role played by spatial information in determining the nature of the salt response, and provided a strong justification for taking the next step to develop a cell-type-specific data set (Fig. 2). Using the methodology developed in Birnbaum et al., roots used for cell sorting were grown under standard conditions and then transferred to salt-supplemented media for 1 h prior to protoplasting. Based on control experiments, protoplasting affects the expression of 15.5% of salt-responsive genes. Both stresses affect expression in a similar manner, however, indicating that sorting is not likely to have a strong qualitative effect on the salt response observed. Six different GFP-marker lines were used in the study to capture expression profiles from all cell types present in the root. These data revealed a large number of salt-regulated genes with 3862 genes showing a statistically significant change of twofold or more across the different cell layers. A large majority of these genes responded to salt, either through up- or down-regulated gene expression, only in an individual cell layer. Very few genes showed significant expression level changes in all tissue layers. This is in contrast to the genes that were identified as differentially expressed in our whole-root time course analysis. In this data set, many cell-type-specific responding genes were not identified, and instead, enrichment was observed for genes regulated by salt in most tissue layers. This difference between data sets affected the biological functions that were uncovered with fewer and less-specific gene ontology categories (Ashburner et al. 2000) showing enrichment in the whole-root data set compared with the cell-sorted data set. Thus, cell sorting appears to provide a more detailed view of the salt-stress response.

Figure 2.

Biological functions regulated along the radial axis of the root. Diagram depicting several of the tissue layers profiled in Dinneny et al. Biological functions associated with growth under standard conditions are shown on the left side, while salt-regulated gene ontology categories are listed on the right.

Analysis of the biological functions regulated by salt stress revealed several interesting themes. First, genes associated with the cell wall or involved in cell-wall modification and biogenesis are frequently down-regulated by salt stress in the epidermis and cortex. These two layers undergo several structural changes in response to salt stress: the epidermis shows an immediate suppression of hair outgrowth, whereas the cortex undergoes radial expansion. In particular, genes previously identified as playing an important role in cell wall biogenesis, such as COBRA (Schindelman et al. 2001), RADIAL SWELLING3 (Burn et al. 2002) and KOBITO1 (Pagant et al. 2002), are down-regulated in these tissue layers. Previous loss-of-function analysis of these genes revealed changes in radial cell expansion similar to that observed under salt stress and suggests that the regulation of these genes may be important for controlling environmentally dependent cell shape.

Plant hormones are important secondary messengers in environmental responses (Wolters & Jurgens 2009). Based on transcriptional changes in the cell-sorted data set, several hormone-signalling pathways appear to be regulated by salt as a function of spatial location. Genes associated with auxin response are down-regulated specifically in the columella cell layer. This is particularly interesting since this tissue plays an important role in gravitropism, which is suppressed by salt treatment. In addition, PIN2 expression is suppressed in the cortex cell layer, consistent with the findings of Sun et al. Thus, gravitropic regulation may be affected through suppression of auxin transportation and response in specific tissue layers. Other hormone pathways were also regulated by salt stress. Genes associated with the ethylene-mediated signalling pathway were up-regulated in the epidermal layer. Previous studies have shown that ethylene production is elicited by salt treatment and ethylene signalling is necessary for the development of salt resistance (Achard et al. 2006). It will be intriguing to determine whether these transcriptional changes in ethylene signalling play a role in adaptation and whether this gaseous hormone can act in a cell-type-specific manner. In contrast to auxin and ethylene pathways, abscisic acid (ABA)-responsive genes are up-regulated by salt treatment in almost every tissue layer in the root suggesting that this hormone may be broadly active during stress response.

Regulatory mechanisms controlling salt responses: cis-elements

Several studies have identified short regulatory DNA sequences in promoters, termed cis-elements, which are important for mediating salt-dependent transcriptional changes. Specifically, the drought-responsive element (DRE) (Yamaguchi-Shinozaki & Shinozaki 1994) and the ABA-responsive element (ABRE) (Marcotte, Russell & Quatrano 1989; Mundy, Yamaguchi-Shinozaki & Chua 1990) are necessary for up-regulation of RD29A gene expression in salt-treated plants (Narusaka et al. 2003). Analysis of the promoter sequences of salt-regulated genes in the time course data set revealed that these elements are primarily enriched in genes that respond between 1 and 4 h after salt treatment. In the cell-type-specific data set, strong enrichment for the ABRE and DRE was observed in broadly responsive genes (responses occurring in three or more cell types), but was largely absent from cell-type-specific responding genes. Thus, these elements may play a primary role in controlling early non-cell-type-specific stress responses. This hypothesis is also consistent with the observation that marker genes (Nemhauser, Hong & Chory 2006) for the ABA response are up-regulated in all tissue layers of the root. In addition, the DREB2A transcription factor (Liu et al. 1998; Sakuma et al. 2006a), which is thought to promote gene expression through direct binding to the DRE element, is broadly expressed in the root under salt conditions. Furthermore, putative downstream targets of DREB2A are activated by salt stress in multiple cell layers more often than would be expected by chance (Sakuma et al. 2006b).

This analysis of cis-element enrichment highlights how sampling methods can affect one's understanding of a process. Both the ABRE and DRE cis-elements were identified through experiments utilizing whole organs (Marcotte et al. 1989; Mundy et al. 1990; Yamaguchi-Shinozaki & Shinozaki 1994). Consistently, we find that these elements are primarily important in regulating responses to salt that occur broadly across cell types in the root. The observation that these elements show little enrichment in the promoters of genes responding in a cell-type-specific manner suggests that these responses may not simply be controlled by the combination of stress regulatory elements with developmental regulatory elements in a single promoter. Alternatively, developmental and stress pathways may be integrated upstream of the promoter level. This model is based only on analyses of known cis-regulatory elements and could be modified as additional novel cis-elements important in these processes are identified.

Regulatory mechanisms controlling salt responses: developmental regulation

The Arabidopsis root has proven to be an excellent model for understanding the mechanisms controlling cell-type specification (Scheres et al. 2002). Studies on the patterning of the epidermis, ground tissue and stele have led to the identification of mutants that cause alterations in the cell types that comprise the root. These mutants represent a largely untapped resource for studies on environmental stimuli and can be used to ask how a cell type of interest contributes to a response.

In our study, we utilized mutants that lead to dramatic changes in the patterning of the epidermal cell layer, which is normally composed of hair and non-hair cells. These mutants were exposed to salt and transcriptionally profiled to identify genes whose salt regulation is dependent upon proper epidermal patterning. One of the most prominent groups of genes identified is expressed specifically in the hair-cell lineage and is salt repressed. Many of these genes have defined roles in hair-cell outgrowth and their expression in the time course data set revealed a highly dynamic mode of expression, with repression lasting only a few hours. These expression changes, in turn, precede dynamic changes in hair outgrowth; hair outgrowth is initially suppressed, and then recovers after 8 h of salt treatment. These data suggest that some level of adaptation may have occurred by this time point.

An unexpected finding of these studies is that perturbations in cell-type specification can also affect the response of other cell types to salt. Overall, 49% of the epidermal patterning-dependent genes identified normally respond to salt in non-epidermal cell layers. These findings suggest the presence of salt-stress dependent signalling between cell layers. This could potentially be mediated through hormone signalling or through the translocation of proteins between cell layers, as has been observed under normal conditions (Gallagher & Benfey 2005). Brassinosteroid signalling has also been shown to regulate shoot growth in a non-cell autonomous fashion (Savaldi-Goldstein, Peto & Chory 2007). Thus, these data provide support for the hypothesis that cell layers exert control over the transcriptional states of other layers in both developmentally and environmentally dependent contexts.

Identification of common stress regulatory modules

In addition to examining responses to salt stress, Dinneny et al. generated similar data sets exploring iron deprivation. These data provide insight into the spatial regulation of this stress and also create a useful basis for a comparative analysis with salt stress. Surprisingly, many cell-type-specific responsive genes are shared between the two stresses. This observation enabled the identification of genes that were consistently co-expressed across environmental conditions, so-called ‘super clusters’. The affinity-propagation clustering algorithm was run using gene expression data from each environmental condition. Similar to the approach taken in Ma and Bohnert (Ma & Bohnert 2007), genes were identified that consistently co-clustered across data sets. Interestingly, genes tended to exhibit nearly identical changes in expression pattern upon salt-stress or iron-deprivation treatment, suggesting that similar regulatory mechanisms may be at work. The largest super cluster exhibited strong expression under standard conditions in the epidermis, which was suppressed by stress. This cluster was highly enriched for genes involved in protein biosynthesis, a common stress-regulated biological process in many organisms (Mayer & Grummt 2006). Due to the large number of high-resolution experiments that group these genes together, it is more likely that they represent true regulatory modules, and it should be instructive to utilize these data to identify the regulatory elements involved.

The transcriptional state of a cell is composed of robust and environmentally sensitive components

The preservation of cell-type-specific expression across diverse environmental conditions suggests that the biological function encoded by the gene is constantly important in that cell type. In Dinneny et al., we utilized three data sets generated by cell sorting (standard conditions, salt stress and iron deprivation) to screen through the many genes that show cell-type-specific expression and identify those genes that consistently define each cell layer. This analysis led to the discovery that cell identity and the expression pattern of genes that regulate cell-fate decisions are robust to abiotic stress. In addition, biological functions were identified that consistently mark each cell layer. These findings suggest that, through the examination of the response of cell types to many diverse environmental stimuli, it may be possible to identify a core set of genes that are important for determining the essential features of a cell type. These data will simplify reverse genetic approaches aimed at defining the regulatory mechanisms controlling cell-fate decisions. Thus, studying classical physiological stimuli from a developmental perspective not only provides rich information for better understanding how roots respond to a changing environment, but should also benefit studies focused on understanding the process of cell-type specification.


In the study of development, a large effort has been made to understand the regulatory mechanisms that control cell-fate decisions. Often, cell-type-specific reporters or morphological and structural features that define certain cell types are used to track cell identity. Very rarely, however, is the response of a cell to a stimulus used to examine the outcome of a cell-fate decision. Utilizing environmental stimuli, we and others have characterized the transcriptional response of different cell layers and developmental stages of the root. These studies revealed that most transcriptional responses are cell-type specific, and that these transcriptional changes result in the differential regulation of specific biological functions in subsets of cell layers. Furthermore, mutants, which cause alterations in cell identify, also alter the response to these environmental changes. Thus, cell identity helps to define context-dependent transcriptional states in addition to determining structural features of a cell. The role of cell identity in mediating stress responses is likely to be generally applicable and it will be interesting to test this prediction by applying a similar approach to the study of other environmental stimuli.


The evolution of multicellularity enables organisms to perform more complex biological functions through the development of unique cell types that can perform specialized tasks. In the root, important functions associated with the salt-stress response are delegated to different cell layers. Several biological functions associated with cell shape changes, such as the suppression of root hair outgrowth and the swelling of the cortex cell layer, may act to prevent additional uptake of salt into the vasculature. Temporal analysis of the salt-stress response revealed a complex succession of waves of transcriptional activity, and it is likely that each cell type may continue to modulate the biological functions it performs as the root adjusts to long-term growth under high salinity. Thus, the complexity uncovered at 1 hour of salt exposure is likely to be the ‘tip of the iceberg’. It will be interesting for future studies to explore in greater detail the advantage to the organism of partitioning responses to external stimuli among the cell types. Are biological functions parsed out due to some incompatibility in regulating them in the same cell? Is there an advantage in regulating certain biological functions in the interior versus the exterior of the root? An understanding of the upstream pathways that control cell-type-specific responses may clarify some of these issues.


Funding is provided by a grant from the National Research Foundation of Singapore and the Temasek Lifesciences Laboratory.