The Art of Being Flexible: How to Escape from Shade, Salt, and Drought

Plants escape from stressful conditions through the plasticity of root and shoot development.

Abstract

Environmental stresses, such as shading of the shoot, drought, and soil salinity, threaten plant growth, yield, and survival. Plants can alleviate the impact of these stresses through various modes of phenotypic plasticity, such as shade avoidance and halotropism. Here, we review the current state of knowledge regarding the mechanisms that control plant developmental responses to shade, salt, and drought stress. We discuss plant hormones and cellular signaling pathways that control shoot branching and elongation responses to shade and root architecture modulation in response to drought and salinity. Because belowground stresses also result in aboveground changes and vice versa, we then outline how a wider palette of plant phenotypic traits is affected by the individual stresses. Consequently, we argue for a research agenda that integrates multiple plant organs, responses, and stresses. This will generate the scientific understanding needed for future crop improvement programs aiming at crops that can maintain yields under variable and suboptimal conditions.

A fundamental difference between plant and animal development is the plasticity in organ formation after germination. Whereas animals are born with a complete set of organs, a germinating seedling has just one embryonic root and one or two embryonic leaves, the cotyledons. All other organs are formed postembryonically, by the interplay of developmental programs and environmental conditions. So, although each plant has a basic body plan, its final size and shape are largely determined by the specific conditions that the plant experiences, and its growth can be adjusted to suit those conditions. This interplay is crucial in both natural and agricultural settings where plants forage for resources and often avoid/escape from stress.

Examples of how plants adjust to environmental conditions include phototropism (Darwin, 1880) to bring the photosynthesizing leaves into well-lit microsites such as canopy gaps and root proliferation toward moisture- or nutrient-rich areas to enhance water uptake and nutrient acquisition (Comas et al., 2013). Examples of stress escape include shoot elongation away from the shade of neighbor plants (shade avoidance; Pierik and de Wit, 2014), escape from submerged conditions to reach the air (Bailey-Serres and Voesenek, 2008), and root growth away from saline soil microsites (halotropism; Galvan-Ampudia et al., 2013). Although some of these responses are termed escape from stress (e.g. shade avoidance), others are considered as attraction to more favorable conditions (e.g. hydrotropism). In the case of directional growth responses, the most unifying way is probably to consider these as responses to gradients of stresses (e.g. salt) or resources (e.g. water).

The molecular, biochemical, and physiological pathways that underlie these responses have been intensively researched, and this has provided substantial knowledge on the regulatory mechanisms. However, relatively little research has been devoted to studying these modes of plasticity in combination. For example, dense plantings of crops growing on irrigated soils in arid conditions likely need to deal with drought, soil salinity, and shading by neighbor crops and weeds simultaneously. Above ground, plants use light cues, particularly enrichment of far-red light (FR) through reflection by nearby vegetation, to detect neighboring vegetation and respond with shade avoidance responses (Casal, 2013; Pierik and de Wit, 2014). Below ground, plants can sense neighbors and their abiotic environment through a variety of putative cues. Some of these result from selective changes made to the rhizospheres by root absorption of minerals and water and excretion of organic compounds. Plants respond to these cues in various ways, including growth toward or away from neighbors, nutrient hotspots, water, and more (Fang et al., 2013; Pierik et al., 2013).

Importantly, the global crop production chain is anticipating intensification of various abiotic stresses: increased temperatures, progressive salinization of highly water-limited production grounds, and more extreme situations of drought and flood (Tubiello et al., 2007; Bailey-Serres and Voesenek, 2008; Munns and Tester, 2008). At the same time, agricultural productivity must be increased to feed the ever-expanding global population, calling for high-density cropping systems with potentially severe mutual shading among plants. Therefore, it is of great importance to understand how plants respond to high-density and abiotic stress(es) simultaneously.

Here, we will review the current molecular and physiological understanding of both shoot developmental plasticity in response to high plant density-derived light signals (shade avoidance) and root developmental plasticity in response to the widely occurring abiotic stresses salt and drought. We will then implement this mechanistic knowledge to generate ideas about (1) how these different modes of plasticity may interact to modulate the known stress response phenotypes and (2) how responses to one stress may affect responses to a second. Addressing these ideas experimentally will generate the knowledge needed to guide crop improvement programs under suboptimal agricultural conditions.

SHADE AVOIDANCE

The shade avoidance syndrome (SAS; Fig. 1; Casal, 2013; Pierik and de Wit, 2014) refers to a suite of responses that enhance individual plant light capture in dense stands. The most striking SAS component is accelerated elongation of internodes and the resulting increased stem lengths (Fig. 1). This is accompanied by enhanced apical dominance, stimulating vertical growth of the main stem to escape the shade, rather than producing branches toward neighbors. In rosette species such as Arabidopsis (Arabidopsis thaliana), in which vegetative internode elongation is highly suppressed, leaf tips are positioned in more vertical canopy strata by hyponastic leaf growth (i.e. upward leaf movement through local cell expansion at the abaxial, proximal part of the petiole; Polko et al., 2012; Fig. 1). Being in a more vertical position, elongation growth of the petiole brings the leaves still higher in the stand. In seedlings, the hypocotyl typically shows a pronounced elongation response (Fig. 1), and this serves as a routine assay system (Keuskamp et al., 2011; Hornitschek et al., 2012; Li et al., 2012) for studies on shade avoidance. Similarly, the cotyledons show hyponastic movement and elongation in response to shade cues (Sessa et al., 2005).

Figure 1.

Shade avoidance responses enhance light capture at high plant densities. A, Arabidopsis Columbia-0 (Col-0; 30 d) plants grown at high plant density (2,000 plants m⁻²) display hyponastic leaves and accelerated petiole elongation. B, Compared with control plants (left), Arabidopsis Col-0 (30 d) plants grown individually in pots and exposed for 2 d to low R:FR also show hyponastic leaves and accelerated petiole elongation (right). C, Arabidopsis Col-0 seedlings (6 d) grown on Murashige and Skoog medium and exposed to low R:FR for 4 d have both elongated hypocotyls and petioles and reduced lamina size. D, Tomato seedlings (2 weeks) exposed to low R:FR (5 d) show strongly accelerated hypocotyl and internode elongation.

Since crops are typically grown at relatively high plant densities, they will experience FR-enriched conditions. This phenomenon is intensified by the presence of weeds that further enrich the crops with FR through FR reflection. Although many modern crops are semidwarf varieties, they still show FR- or high density-induced responses, such as stem elongation in maize (Zea mays; Dubois et al., 2010) and potato (Solanum tuberosum; Boccalandro et al., 2003), reduced tillering in wheat (Triticum aestivum; Ugarte et al., 2010), and reduced tuber growth in potato (Boccalandro et al., 2003). All these responses reduce crop yield.

Another cause of yield loss associated with shade avoidance may stem from the suppression of defenses against herbivores (Moreno et al., 2009) and both (hemi)biotrophic and necrotrophic pathogens (Cerrudo et al., 2012; de Wit et al., 2013), which may lead to increased severity of diseases and biomass loss. Although the exact mechanisms are unclear, it appears that FR enrichment leads to a decreased responsiveness of plants to jasmonic acid (JA) and salicylic acid (Moreno et al., 2009; Cerrudo et al., 2012; de Wit et al., 2013). Further details on this interaction are reviewed by Ballaré (2014).

Aboveground Plant Neighbor Detection: Multiple Signals

Plants use a variety of signals to sense neighboring foliage, and combinations of these signals represent different phases of competition intensity. Most aboveground neighbor detection cues are changes in the light spectral composition that occur because of selective absorption and reflection of particular wavebands. Blue light (B; 400–500 nm) and red light (R; 600–700 nm) are absorbed by the chlorophyll pigments to power the light reactions of photosynthesis. Green light (G; 500–580 nm) and particularly FR (700–800 nm), on the other hand, are mostly reflected by foliage. R and FR are perceived by phytochrome photoreceptors, which regulate downstream signaling pathways. B responses predominantly involve phototropins and cryptochromes (Casal, 2013). The relative occurrence and importance of these different light signals, and their interactions at different stages of stand development, were recently reviewed (Casal, 2013; Pierik and de Wit, 2014). Early neighbor detection occurs primarily through detection of the R:FR ratio and possibly G enrichment. True shading is detected through the same signals, combined with low fluence rates of both B and photosynthetically active radiation (PAR). In order for G and FR to be enriched prior to true shading, a vertical plant structure is required for horizontal light reflection toward neighbor plants. This is the case in most, but not all, stands of vegetation. It was recently shown that in horizontally growing stands of Arabidopsis rosettes, mechanical touching of leaves can precede FR signaling (de Wit et al., 2012). Here, we will address primarily phytochrome-mediated detection of the R:FR ratio, the predominant plant neighbor detection cue that induces shade avoidance.

Phytochrome Signaling of R:FR

Green plant tissues absorb R but not FR. Consequently, the R:FR in horizontally reflected light is reduced (Ballaré et al., 1990), and this condition is sensed by neighbors via the phytochrome family of photoreceptors. In Arabidopsis, there are five phytochromes: phyA to phyE. Of these, phyB is the dominant phytochrome involved in plant neighbor detection, with phyD and phyE having redundant roles. Phytochromes have two main domains: the N-terminal domain functions in photodetection and contains the bilin chromophore that is associated with photoconversion. The C-terminal domain transmits the perceived light signal (for review, see Chen and Chory, 2011). Phytochromes function as dimers and exist in two relatively stable, photoconvertible forms: the R-absorbing Pr form and the FR-absorbing Pfr form. The Pr form has an absorption maximum at λ = 660 nm (R) and is inactive. Upon absorbing R, Pr is photoconverted into the active Pfr form, which has an absorption maximum of λ = 730 nm (FR). Upon absorbing FR, Pfr is converted back into the inactive Pr form. Thus, a particular R:FR ratio is reflected in the Pfr:Pr ratio, thus determining the relative activity of phytochrome (for review, see Chen and Chory, 2011; Casal, 2013). Upon exposure to R, the cytosolic phyB Pr is photoconverted into active Pfr and migrates into the nucleus, where it directly interacts with phytochrome-interacting factors (PIFs; Chen and Chory, 2011; Leivar and Quail, 2011).

PIFs

PIFs are a subfamily of the basic helix-loop-helix (bHLH) family of transcription factors that physically interact with phytochrome. Upon binding to phytochrome, PIFs become phosphorylated, leading to their subsequent degradation through the ubiquitin-proteasome route (Leivar and Quail, 2011). Seven isoforms can physically interact with phyB: (PIF1/PIF3-LIKE5 [PIL5], PIF3, PIF4, PIF5/PIL6, PIF6/PIL2, PIF7, and PIF8). They interact via the conserved N-terminal active phyB-binding motif. PIFs bind to a G-box motif (CACGTG) in the promoter of target genes. Thus, it appears that Pfr binding to PIFs and subsequent PIF phosphorylation and degradation explain how phytochromes regulate transcription (Leivar and Quail, 2011; Zhang et al., 2013). PIF7 seems to be an exception to this rule, since, although it is phosphorylated, it does not seem to be degraded in the presence of Pfr, suggesting that PIF phosphorylation alone may suffice to regulate target genes (Li et al., 2012).

Once inactivated by FR, phyB does not interact with PIFs, allowing PIF proteins to accumulate again and regulate the transcription of target genes associated with shade avoidance and subsequently modulate SAS (summarized in Fig. 2). Accordingly, pif knockout mutants have altered shade avoidance responses to low R:FR. In particular, PIF4, PIF5, and PIF7, and to a lesser extend PIF3 and PIF1, are associated with shade avoidance responses. In double knockout pif4 pif5 mutants, hypocotyl elongation is reduced approximately 50% in response to low R:FR conditions (Lorrain et al., 2008). Accordingly, knockout of these two pif mutants partially suppresses the elongated hypocotyl phenotype of a phyb knockout mutant (Lorrain et al., 2008). The impact of these two PIFs, however, appeared to be even stronger in the hypocotyl elongation response to low PAR (Hornitschek et al., 2012). PAR is largely sensed through cryptochrome-mediated detection of B fluence rates, and indeed, low B-induced petiole elongation and hyponasty are also inhibited in the pif4 pif5 double knockout (Keller et al., 2011). Thus, although PIF4 and PIF5 are involved in shade avoidance responses to low R:FR, they are particularly important for cryptochrome-mediated responses to low B fluence rates. However, it is not known whether and how cryptochromes might interact with PIF4 and PIF5 and whether they affect PIF protein abundance or activity. It was shown recently that PIF7 is the key regulator of low R:FR-induced hypocotyl elongation: the pif7 knockout mutant shows an even more severely reduced hypocotyl elongation response to low R:FR compared with the pif4 pif5 mutant (Li et al., 2012).

Figure 2.

Plant neighbor detection through FR reflection inactivates phytochrome by photoconversion of the inactive FR-absorbing Pfr form into the active Pr form. This subsequently modulates a network of transcription factors, hormones, and cell wall-modifying proteins to induce shade avoidance responses. Brown boxes indicate hormones, brown letters (unboxed) indicate hormone biosynthesis enzymes, and purple border and no shading indicate transcription factors. SLs, Strigolactone; ACO, ACC OXIDASE.

In addition to regulating light responses, PIFs also regulate responses to other environmental signals (Leivar and Quail, 2011). Phytochrome-controlled shade avoidance responses are strongly affected by temperature (Halliday and Whitelam, 2003; Patel et al., 2013), and high temperature itself can also induce SAS-like responses. These responses rely on PIF4 (Koini et al., 2009) and PIF4-mediated auxin biosynthesis (Franklin et al., 2011). Another escape response that seems to involve PIFs is the shoot elongation response of flood-adapted terrestrial species, which allows them to escape stressful submergence (e.g. as a consequence of heavy rainfall). A transcriptome survey in the flood-adapted wild species Rumex palustris (marsh dock) identified that PIFs, as well as several other components of the transcriptional light signaling circuitry, are associated with enhanced petiole elongation in an ethylene-dependent manner (van Veen et al., 2013). Consistently, ethylene-induced hypocotyl elongation in light-grown Arabidopsis seedlings occurs through interaction of the ETHYLENE-INSENSITIVE3 protein with PIF3 (Zhong et al., 2012).

Transcriptional Control of Shade Avoidance

In addition to PIFs, a number of atypical bHLH members are induced, including LONG HYPOCOTYL IN FAR-RED1 (HFR1), PHYTOCHROME RAPIDLY REGULATED1 (PAR1), and PAR2. Unlike PIFs and other bHLH proteins, these proteins do not appear to bind DNA. Rather, HFR1 interacts with PIFs, probably through the DNA-binding motif of the PIF proteins, and by doing so regulates the transcription of PIF targets (Hornitschek et al., 2009; Galstyan et al., 2011). Since PIFs stimulate shade avoidance responses, HFR1 is likely to be a negative regulator, given its interaction with the DNA-binding motif of PIFs. Indeed, hfr1 knockouts display an enhanced hypocotyl elongation response to low R:FR conditions, whereas HFR1, PAR1, and PAR2 overexpressors have reduced hypocotyl elongation (Sessa et al., 2005; Roig-Villanova et al., 2007; Galstyan et al., 2011). Recently, two additional bHLH proteins, BRASSINOSTEROID-ENHANCED EXPRESSION1 (BEE1) and BRI1-EMS-SUPPRESSOR1 (BES1) INTERACTING MYC-LIKE (BIM), were shown to interact with PAR1 and positively regulated general plant growth as well as low R:FR-induced shade avoidance (Cifuentes-Esquivel et al., 2013).

In addition to bHLH proteins, several other transcription factors are regulated during phytochrome-mediated shade avoidance. The homeodomain Leu zipper proteins ATHB2 (Steindler et al., 1999) and ARABIDOPSIS THALIANA HOMEOBOX PROTEIN4 (ATHB4) (Sorin et al., 2009) are transcriptionally induced in response to low R:FR. ATHB2 stimulates longitudinal cell expansion in the Arabidopsis hypocotyl (Steindler et al., 1999), and ATHB4 may modulate hormone responses (Sorin et al., 2009). The B-box-containing (BBX) proteins BBX21 and BBX22 were shown to regulate shade avoidance downstream of the major photomorphogenesis regulator CONSTITUTIVELY PHOTOMORPHOGENIC1 (COP1), but in a manner that does not involve the major COP1-interacting protein LONG HYPOCOTYL5 (HY5; Crocco et al., 2010).

Physiological Targets

PIF proteins regulate the transcription of a wide variety of target genes, including cell wall-modifying enzymes and hormone biosynthetic genes (Fig. 2). PIF5 has been shown to bind directly to G-box motifs of DNA during shade avoidance (Hornitschek et al., 2012). Indeed, induction of the shade avoidance marker gene PIL1 by PIF5 relies on PIF5 interaction with the G-boxes in the PIL1 promoter (Hornitschek et al., 2012). This G-box dependence of target gene regulation was also observed for other PIFs in dark-grown seedlings (Zhang et al., 2013). G-box-containing PIF targets are many and include, but are certainly not limited to, various genes associated with biosynthesis and signaling of the plant hormone auxin (Hornitschek et al., 2012; Zhang et al., 2013).

Auxin Is a PIF Target at Various Levels

Endogenous auxin, indole acetic acid (IAA), is produced from the amino acid Trp, mainly by TRYPTOPHAN AMINOTRANSFERASE OF ARABIDOPSIS (TAA1)-mediated conversion of Trp into indole-3-pyruvate, which is subsequently converted into IAA by the YUCCA (YUC) family of flavin monooxygenases (Zhao, 2012). The main sites of auxin biosynthesis in plants are the young shoot tissues, and auxin is transported to other plant parts, including the root system, through polar auxin transport. This polar transport is facilitated through specific auxin influx and efflux facilitator proteins. AUXIN1 (AUX1) and LIKE-AUX1 (LAX) proteins (AUX1 and LAX1–LAX3 in Arabidopsis) are auxin influx carriers (Péret et al., 2012b), with partially overlapping and partially distinct roles in plant development.

PIN-FORMED (PIN) proteins, on the other hand, facilitate cellular auxin efflux and are a family of eight proteins in Arabidopsis (Krecek et al., 2009). By regulated intracellular (re)positioning of these auxin carriers, plants can direct auxin fluxes within and between organs. Auxin is sensed upon binding to its receptor, the F-box protein TRANSPORT INHIBITOR RESPONSE1 (TIR1), and its close homologs, the AUXIN SIGNALING F-BOX proteins (Calderón Villalobos et al., 2012). Binding of auxin promotes TIR1 receptor interaction with AUXIN/INDOLE ACETIC ACID (AUX/IAA) proteins. Being part of a Skp, Cullin, F-box containing (SCF) complex, TIR1 interaction with AUX/IAA proteins leads to polyubiquitination of AUX/IAAs, flagging them for degradation in the 26S proteasome (Teale et al., 2006). Degradation of AUX/IAAs subsequently releases AUXIN RESPONSE FACTOR (ARF) proteins from AUX/IAAs and degradation, allowing them to bind target promoter sequences and regulate (either positively or negatively) the transcription of target genes. In the absence of the auxin-TIR1 interaction, AUX/IAAs are not degraded and interact with ARFs, preventing ARFs from binding DNA (Teale et al., 2006).

In a search for PIF targets during shade avoidance, PIF4-HA and PIF5-HA chromatin immunoprecipitation experiments showed that both of these PIFs bind to YUC8 and IAA29 (Hornitschek et al., 2012). This indicates that PIFs can regulate both biosynthesis and signaling of auxin. Similar experiments using a PIF7-Flash construct confirmed PIF7 binding to YUC8 and YUC9 promoters (Li et al., 2012). In accordance, various auxin mutants have an altered shade avoidance response (Steindler et al., 1999; Pierik et al., 2009). In a screen for novel mutants with disturbed shade avoidance responses, the shade avoidance3-2 mutant was identified as a taa1 auxin biosynthesis mutant. Consistent with PIF-mediated expression of YUCs, whole-plant and shoot IAA levels increased in low R:FR conditions (Tao et al., 2008; Hornitschek et al., 2012; Li et al., 2012). Importantly, IAA levels in the elongating hypocotyl specifically are elevated in low R:FR (Keuskamp et al., 2010). This appears to be conditional upon PIN3-mediated auxin transport, because low R:FR-induced IAA accumulation in the hypocotyl was lost in a pin3 mutant. The polar, basal subcellular localization of PIN3 in the hypocotyl endodermis in white light facilitates rootward auxin transport. Upon low R:FR exposure, the level of PIN3 increases, and it becomes localized to the lateral sides of the endodermis cells. This likely facilitates auxin transport toward the epidermis, where elongation growth is controlled (Keuskamp et al., 2010). Pharmacological inhibition of auxin perception prevents this PIN3 response, suggesting that auxin signaling drives this PIN3 behavior. Exposure of etiolated Arabidopsis seedlings to unidirectional B also induces a lateral subcellular orientation of the PIN3 protein, facilitating hypocotyl bending toward the light (phototropism; Ding et al., 2011). In summary, several aspects of auxin biology are targets for PIF proteins and are key to shade-induced hypocotyl elongation.

Low R:FR Enhances GA Levels to Degrade DELLA Proteins

GAs have long been associated with plant growth in general and shade avoidance specifically. GA biosynthesis and response mutants typically are dwarfed, and this was instrumental to the development of semidwarf crop varieties during the Green Revolution of the 1960s (Hedden, 2003). Although many different GAs exist in plants, few are as bioactive as GA. GA is a diterpenoid produced from geranyl geranyl diphosphate, which is converted into ent-kaurene. Subsequent steps of GA biosynthesis involve GA 20-oxidase and GA 3-oxidase, which contribute to the production of bioactive GAs, typically GA₁ in plants. GA 2-oxidase catabolizes bioactive GAs into inactive forms (Hedden and Thomas, 2012). To induce shade avoidance, Arabidopsis plants were exposed to an end-of-day FR treatment (EODFR). This led to the induction of GA20ox expression (Hisamatsu et al., 2005), which is thought to stimulate GA levels in EODFR-exposed plants. Accordingly, a transcriptome study using Arabidopsis seedlings also found elevated GA20ox mRNA levels upon low R:FR exposure (Devlin et al., 2003). In a study in R. palustris, both GA20ox and GA3ox were transcriptionally induced upon exposure to low R:FR (Pierik et al., 2011).

It is unknown whether GA biosynthesis is regulated by PIFs during shade avoidance. DELLA proteins, an essential component in the GA signal transduction pathway, interact with PIF. DELLAs are growth-inhibiting nuclear proteins that indirectly control transcription by binding to PIF4. This disables the ability of PIF4 to regulate transcription (de Lucas et al., 2008), thus inhibiting shade avoidance responses. When GA binds to its receptor GIBBERELLIN INSENSITIVE DWARF1 (GID1), a family of three GA receptor proteins in Arabidopsis, the resulting complex promotes the binding of DELLAs to the F-box protein SLEEPY1 (SLY1). This leads to polyubiquitination of DELLAs by the SCF^SLY1 complex and subsequent degradation of DELLAs in the proteasome (Schwechheimer, 2008).

Using a GFP fusion of the REPRESSOR OF GA1-3 (RGA1) DELLA protein in Arabidopsis, it was shown that nuclear RGA1 disappears upon exposure to low R:FR conditions and concurrent with the induction of petiole elongation (Djakovic-Petrovic et al., 2007). RGA1 abundance was not affected by low R:FR when GA biosynthesis was disturbed with the GA biosynthesis inhibitor paclobutrazol, indicating that degradation is GA dependent and occurs through the GA-GID1-SLY1 pathway. In seedlings, similar results were found for hypocotyls and for both low R:FR and low B-mediated elongation (Djakovic-Petrovic et al., 2007; Pierik et al., 2009). The degradation of DELLAs prevents their inhibitory action with PIFs, thus allowing PIF-mediated transcription. As a result of this DELLA protein degradation, the PIF-mediated pathways toward shade avoidance are released.

Brassinosteroids

Brassinosteroids (BRs) are growth-promoting steroid hormones involved in a wide variety of developmental processes, including photomorphogenesis and pathogen resistance (Wang et al., 2012). The involvement of the BR response-associated proteins BEE1 and BIM in shade avoidance (Cifuentes-Esquivel et al., 2013) suggests that BR regulates shade avoidance. Furthermore, PIF4 was shown to interact in vitro and in vivo with the BR-activated transcription factor BRASSINAZOLE RESISTANT1 (BZR1; Oh et al., 2012), a protein that also interacts with DELLAs (Bai et al., 2012). EODFR-induced petiole elongation in Arabidopsis rosette plants is associated with the induction of BR-responsive genes (Kozuka et al., 2010). Furthermore, EODFR-induced petiole elongation was abolished in the rotundifolia3 (rot3) mutant, which lacks the ROT3 cytochrome P450 (CYP90C1) involved in BR biosynthesis. The rot3 mutant also suppresses the elongated petiole phenotype of the constitutively shade-avoiding phyb mutant in the phyb rot3 double mutant (Kozuka et al., 2010). This further underscores BR involvement in phytochrome-mediated shade avoidance. Furthermore, rot3 displays a reduced hypocotyl elongation response to B depletion (Keuskamp et al., 2011). Therefore, it appears that BR and auxin together drive most of the low B-induced hypocotyl elongation response in Arabidopsis, with partially shared and partially specific targets (Keuskamp et al., 2011), a frequently observed finding in studies on auxin and BR targets (Nemhauser et al., 2004).

Downstream Targets Include Cell Wall-Modifying Proteins

As described above, auxin and BR have partially overlapping physiological targets in regulating the shade avoidance response. For example, both hormones regulate the expression of XYLOGLUCAN ENDOTRANSGLUCOSYLASE/HYDROLASE (XTH) genes during low B-induced hypocotyl elongation (Keuskamp et al., 2011). In addition, PIF5 binds to the G-box-containing fragment of the XTH15/XTR7 (for XYLOGLUCAN ENDOTRANSGLYCOSYLASE7) promoter (Hornitschek et al., 2009). XTHs are cell wall-modifying proteins that act on the cellulose-hemicellulose (xyloglucan) network of the cell wall, which during cell expansion is the main tension-bearing structure of the cell wall (Rose et al., 2002). Its modification is key to turgor-driven cell expansion (Cosgrove, 2005). Another major class of cell wall-modifying proteins, expansins, also acts on the cellulose-hemicellulose network (for review, see Cosgrove, 2005). During shade avoidance, both XTH and EXPANSIN (EXP) genes are induced in Arabidopsis. Exposure to green shade, mimicking an overcast canopy, enhanced XTH physiological activity in Arabidopsis but did not affect EXP activity (Sasidharan et al., 2010). In Stellaria longipes, both XTH and EXP activities increased during shade avoidance (Sasidharan et al., 2008), suggesting that both groups of proteins contribute to shade avoidance in a species-specific manner.

Shade Avoidance and the Modification of Plant Architecture

The pronounced induction of auxin production and signaling in low R:FR conditions that is required for enhanced organ elongation also leads to induction of the cytokinin catabolism gene CKX6 in developing leaf primordia of Arabidopsis seedlings (Carabelli et al., 2007). This occurs especially in the preprovascular cells of these leaf primordia and implies that cytokinin levels are reduced. Accordingly, a ckx6 knockout does not display the low R:FR-induced inhibition of leaf primordial growth, whereas leaf primordia are clearly inhibited in wild-type plants (Carabelli et al., 2007).

Although cytokinins are major regulators of shoot branching, it is unknown whether CKX6 or other cytokinin regulators are also involved in phytochrome control of branching. However, other shoot-branching regulators have been associated with phytochrome signaling. The constitutively shade-avoiding phyB-1 mutant of sorghum (Sorghum bicolor) does not produce branches during vegetative development because axillary buds remain dormant (Kebrom et al., 2006). This bud dormancy is associated with high expression levels of not only the bud dormancy marker SbDRM1 but also the transcriptional regulator TEOSINTE BRANCHED1 (SbTB1) that inhibits bud outgrowth. In Arabidopsis, phytochrome inactivation during low R:FR exposure also leads to inhibited bud outgrowth, and this involves two homologs of TB1: BRANCHED1 (BRC1) and BRC2 (Finlayson et al., 2010). In addition, several hormonal factors mediate phytochrome-controlled axillary bud outgrowth: the nced3-2 (for nine-cis-epoxycarotenoid dioxygenase3-2), aba deficient2 (aba2)-1, auxin resistant1, and more axillary growth2 (max2) and max4 mutants had altered branching responses to low R:FR relative to the wild type (Finlayson et al., 2010; Reddy et al., 2013). This implies that abscisic acid (ABA), auxin, and strigolactones play parts in the phytochrome regulation of shoot branching. Interestingly, MAX2 also has been associated with various aspects of photomorphogenesis in seedlings, such as cotyledon expansion and hypocotyl elongation in response to R, FR, and B (Shen et al., 2007).

Resource investments into aboveground structures such as internodes and petioles occur at the expense of belowground investments in root systems of a variety of species, including maize and southern pea (Vigna unguiculata; Kasperbauer and Hunt, 1992; Page et al., 2009). In Arabidopsis seedlings, the number of lateral roots (LRs) is strongly reduced in phyB mutants and in wild-type plants exposed to low R:FR. This is at least partly due to the action of shoot-localized phytochrome on the root system via hormone action (Salisbury et al., 2007). Even stronger suppression of LR formation is observed when plants are exposed to darkness (Kircher and Schopfer, 2012; Sassi et al., 2012).

ROOT RESPONSES TO DROUGHT AND SALINITY

Root System Architecture

Roots provide support to anchor the plant in the soil and are critical for the uptake of nutrients and water. As with the shoot, development of the root is flexible and can be adjusted to the environment not only to optimize foraging for nutrients and water but also to reduce exposure to stress. The importance of the root system in supporting a new Green Revolution is receiving an increasing amount of attention (Den Herder et al., 2010; Lynch, 2013; Orman-Ligeza et al., 2013). Of concern is that, until now, breeding has not focused on the root system and its plasticity, and optimization of the hidden half of the plant’s architecture could play a significant role in increasing crop yield. Also important are the stress-induced signal transduction mechanisms initiated by the perception of abiotic stress either in the roots or leaves that lead to morphological responses in the root. These are considered the main targets for yield improvement in crop plants under drought and salinity stress (Marshall et al., 2012; Comas et al., 2013; Galvan-Ampudia et al., 2013).

Branching

Root system architecture (RSA) is defined as the spatial arrangement of all parts of the root system. It depends, first, on the activity of the primary root (PR) apical meristem, which drives PR growth. Subsequently, the root will branch through the formation of LRs in dicots. In monocots, in particular cereals, a large part of the root system consists of crown, seminal, and adventitious roots, originating from the stem (Orman-Ligeza et al., 2013). In both monocots and dicots, subsequent branching to form higher order LRs results in a complex root system. LRs are initiated from cell types deep within the root. In monocots, these are the phloem pericycle and endodermal cells, whereas in dicots, LR formation starts with cell divisions within the xylem pericycle. In both types of plants, the distinct phases of initiation, early cell divisions, emergence, and finally outgrowth of the root are all tightly regulated. These developmental programs have been reviewed extensively recently (De Smet, 2012; Orman-Ligeza et al., 2013). Here, we focus on how root development and growth can be modulated in response to specific environmental stresses, in particular drought and salinity, conditions that are highly relevant to agriculture. RSA is also affected by other environmental factors, most notably nutrient availability, including deficiencies for nitrogen, phosphorus, potassium, and sulfur and combinations of these deficiencies (Gruber et al., 2013; Kellermeier et al., 2014). Recently, a genetic basis for natural variation in RSA responses to several nutrient deficiencies was revealed using Arabidopsis accessions collected worldwide, indicating that root developmental plasticity is modulated by evolutionary adaptation (Kellermeier et al., 2013; Rosas et al., 2013).

Which Way to Grow?

Besides root elongation and branching, the direction of growth of the roots is important for an effective root system, especially under stressful or nonhomogenous conditions. PRs, like shoots, will normally follow the gravity axis, while LRs initially show more horizontal growth after emergence, before bending down. Interestingly, it was found only recently that emerging LRs integrate gravity sensing with a counteracting mechanism that offsets the angle of the emerging root (Rosquete et al., 2013; Roychoudhry et al., 2013). The resulting gravitropic set-point angle (GSA) of lateral organs directs LR growth away from the PR or from lower order LRs, which allows the root system to explore more space. The GSA can be modulated by stress, resulting in a deeper or shallower root system to match the nutrient and water conditions of the soil (Lynch, 2013). Also, the PR is able to redirect its direction of growth in response to water (hydrotropism; Takahashi et al., 2009; Iwata et al., 2013) or away from salt (halotropism; Galvan-Ampudia et al., 2013).

Plant Hormones Determine the Shape and Structure of the Root System

The plant hormone auxin plays a major role in virtually every aspect of RSA. It affects PR growth, development of adventitious roots, LR formation, and LR elongation as well as the direction of growth (Ottenschläger et al., 2003; Band et al., 2012; Lavenus et al., 2013). Auxin acts as an integrator of other endogenous hormonal signals as well as environmental cues, such as gravity, light, and salinity. The central action of auxin in root growth and development is modulated by other hormones, which are known to interfere at different steps (i.e. auxin homeostasis, transport, and signaling). Best described is the role of cytokinins, which antagonize auxin in many aspects of plant physiology, including RSA. Other hormones that not only affect root growth but also play a role in balancing PR growth versus branching, and are thus essential for RSA, are ABA, strigolactones, and ethylene. In addition, GA, JA, and BRs have been shown to play roles in root growth. In these ways, all plant hormones contribute to the final architecture of the root system (Fukaki and Tasaka, 2009; Lavenus et al., 2013).

Auxin Orchestrates Root Development

As detailed above, auxin distribution in the plant is determined in large part by the action of the AUX and PIN family proteins that facilitate auxin influx and efflux, respectively (Krecek et al., 2009; Grunewald and Friml, 2010; Péret et al., 2012b). In the root, AUX/PIN-mediated transport establishes the inverse fountain pattern of auxin flow (i.e. auxin is transported downward through the stele, while it is directed back up via the columella tissue and through the cortex/epidermis), resulting in an auxin maximum in the root tip and a gradient through the root that is fundamental to all aspects of RSA. Consistently, disruption of auxin transport blocks PR growth and branching (Lavenus et al., 2013) as well as gravitropic responses (Baldwin et al., 2013). In LR formation, redistribution of auxin is a decisive factor at every stage, from the initial priming of LR primordia to outgrowth of the LR. Auxin acts through distinct signaling modules at each stage, which have been reviewed by Lavenus et al. (2013).

Gravitropic growth of the PR of Arabidopsis was also shown to depend on PIN-mediated transport. When roots are gravitropically challenged by tilting them to a horizontal position, auxin is rapidly redistributed through the action of the PIN2 and PIN3 auxin efflux carriers (Baldwin et al., 2013). In addition, AUX1, localized to the LR cap and epidermal cells, is crucial for the basipital transport of auxin to the elongation zone (Swarup et al., 2005). Consistently, the aux1 mutant of Arabidopsis exhibits agravitropic growth of the root (Bennett et al., 1996; Swarup et al., 2005).

Only recently, it was found that the GSA of lateral organs is also dependent on differential auxin distribution (Rosquete et al., 2013; Roychoudhry et al., 2013). The LR angle after emergence is highly stable at approximately 60°, suggesting that it is genetically determined. While the initial angle at emergence is 90°, it is followed by a differential growth response. Using light sheet-based imaging of emerging LRs, an asymmetric onset of elongation in LRs was observed upon formation of the elongation zone, but differential auxin-induced gene expression was detected earlier (Rosquete et al., 2013). A crucial role for auxin in the regulation of the GSA was further shown by the manipulation of auxin levels; high auxin induced more vertical root systems, and low auxin induced more radial root systems. Consistent with this, tir1 and auxin binding protein1 receptor mutants show a shift to higher (more horizontal) GSA, consistent with the idea that auxin is required for and steers LR angle (Rosquete et al., 2013; Roychoudhry et al., 2013).

Cytokinins Inhibit Root Branching

Cytokinins in general antagonize auxin in plant developmental responses and have a profound impact on RSA because the balance between auxin and cytokinin levels largely determines the level of branching of the root. Whereas auxin decreases PR growth and stimulates branching, cytokinin inhibits LR formation. Although most cytokinin-auxin cross talk relies on transcriptional responses (Dello Ioio et al., 2008; Ruzicka et al., 2009), the effect of cytokinins on LR formation is mediated by its effect on PIN1 endocytic trafficking. Cytokinin-induced targeting of PIN1 to lytic vacuoles restricts auxin transport to developing LR meristems, thus inhibiting cell division in developing primordia and, ultimately, branching (Marhavý et al., 2011).

ABA Regulates Root Growth

The effect of ABA on root growth is complex. Consistent with its general role in the dormancy of tissues, ABA promotes the quiescence of both primary and LR primordia (De Smet et al., 2003) without loss of meristem function. As such, ABA is generally regarded as an inhibitor of root growth. On the other hand, using ABA-deficient mutants and inhibitors in maize, several studies have found that ABA also can be required to promote root growth (Saab et al., 1990) and acts by suppressing ethylene production (Spollen et al., 2000). Also, in Arabidopsis, ABA enhances both shoot and root growth under well-watered conditions (Barrero et al., 2005; Zhang et al., 2010). Recently, ABA was shown to promote the quiescence of the quiescent center and to suppress the differentiation of stem cells in the primary meristem of Arabidopsis roots, providing a cellular explanation for the complex effect of ABA on root growth and development (Zhang et al., 2010).

On the cellular level, ABA is perceived by receptors consisting of the PYRABACTIN RESISTANCE/PYRABACTIN RESISTANCE-LIKE/REGULATORY COMPONENT OF ABA RECEPTOR family in complex with Protein phosphatase 2C (PP2C) (Cutler et al., 2010). In the presence of ABA, this complex inhibits PP2C phosphatase activity, promoting the activity of the SNF1-RELATED PROTEIN KINASE2 (SnRK2) subclass 3 protein kinases that are required for downstream ABA responses. In particular, this involves the regulation of ABA-induced gene expression and ion transport in the root via K⁺ transporters and channels (Osakabe et al., 2013).

Other Phytohormones in RSA Control

Strigolactones are known to affect RSA by inhibiting adventitious root formation, increasing PR length and affecting LR formation depending on the nutrient conditions (Rasmussen et al., 2013a). Recently, they were shown to be important for water stress responses, as strigolactone-deficient max mutants were hypersensitive to drought and salinity, most likely because of their increased stomatal density, However, as the effect of root branching was not examined in great detail in that study, a role for the root system in this response cannot be excluded (Ha et al., 2014). As described above in the shade avoidance section, GA in general both promotes the growth of roots and shoots and antagonizes ABA action. Salt and osmotic stress reduce GA levels and lead to the stabilization of DELLAs (Achard et al., 2006). Consistent with this, GA production was disadvantageous for salt and drought tolerance of Arabidopsis plants and affected the growth of LRs more than PRs (Duan et al., 2013; Geng et al., 2013; Colebrook et al., 2014). At low concentrations, ethylene is known to affect RSA by promoting LR primordia formation through its interaction with auxin (Ivanchenko et al., 2008). Recently, ethylene-induced ROS formation was shown to promote salt tolerance by restricting Na⁺ influx into the root, leading to restricted Na⁺ transport to the shoot (Jiang et al., 2013). Similar to the action of strigolactones, the role of ethylene in the modulation of RSA in response to salinity remains to be established.

Root Plasticity under Stress

The developmental plasticity of root systems is key to the acclimation and survival of plants under adverse conditions such as nutrient shortage, drought and salinity, and ion toxicity. In general, severe osmotic stresses, including drought and salinity, decrease the overall root growth of seedlings. However, recent studies have shown that, in lower concentrations of salt or mild drought, specific adjustments in RSA can be observed, involving both branching and changes in the direction of growth. These responses are fine-tuned by other environmental factors such as nutrient availability, and they rely on the hormones discussed above, in particular auxin and ABA (Fukaki and Tasaka, 2009; Takahashi et al., 2009; Galvan-Ampudia and Testerink, 2011).

Drought Stress Affects Root Growth and Architecture: A Major Role for ABA

For optimal growth, development, and yield of crops, plants rely on sufficient water availability. Yet, in the case of drought, plants can modulate their root-shoot ratios as well as RSA to most efficiently get to water and maximize water use efficiency (Comas et al., 2013). ABA plays a key role in almost all responses to water stress. Drought and osmotic stress induce the formation of ABA both in roots and shoots (Claeys and Inzé, 2013). Above ground, ABA formation is essential for inducing stomatal closure. In the roots, drought leads to a decrease in both PR and LR growth (Hong et al., 2013).

As mentioned above, ABA can enhance the shoot and root growth of maize by suppressing ethylene production in well-watered conditions. During dehydration, ABA formation reduces shoot growth, but it promotes the growth of roots even under water-limiting conditions, leading to a dramatic stimulation of root-shoot ratio under water-limited conditions (Saab et al., 1990; Spollen et al., 2000; Leach et al., 2011). Consistently, mild drought stress also promotes root growth in Arabidopsis (Granier et al., 2006).

How ABA mediates RSA changes in response to drought has become increasingly clear recently. ABA inhibits LR formation in an auxin-independent manner (De Smet et al., 2003). Salinity-induced ABA formation restricts LR development and growth through endodermal signaling pathways (Duan et al., 2013), which are also likely to play a role in drought responses. In addition, LR emergence was recently found to depend on aquaporins, which regulate water transport into the developing PR primordium and allow the PR to break through the endodermal tissues. Spatial and temporal distribution of aquaporin-dependent water transport during LR emergence requires auxin (Péret et al., 2012a). Aquaporins represent a likely factor in RSA responses to drought or osmotic stress, which also regulate hydraulic conductance of the root (Li et al., 2014). In addition, cellular cycling of the plasma membrane aquaporins Plasma membrane intrinsic protein2;1 (PIP2;1) and PIP1;2, was found to be induced by salt treatment (Luu et al., 2012). Regulation of turgor and volume loss of the endodermis were shown to be essential not only for LR emergence but also for LR initiation (Vermeer et al., 2014). Turgor, LR development, and cell expansion in the root under salt and drought stress all seem to depend also on the maintenance of Na⁺/K⁺ ion homeostasis, which in turn appears to be regulated through the action of ethylene formation and K⁺ transport (Jiang et al., 2013; Osakabe et al., 2013).

Hydrotropism

Hydrotropism is the ability of the root to grow toward moisture (i.e. environments of relatively higher water potentials). This response is clearly different from gravitropism: mutants impaired in gravitropism were found to still be able to display a hydrotropic response (Jaffe et al., 1985). First described and investigated by Darwin (1880), the hydrotropic response of roots received little interest until quite recently. Thanks to the development of more robust assay systems for hydrotropism, including a simplified method for Arabidopsis seedlings (Takahashi et al., 2002; Eapen et al., 2003), its molecular basis is being addressed through the isolation of the mizu-kussei (miz) and no hydrotropic response mutants. MIZ1 encodes an unknown protein conserved in terrestrial plants (Kobayashi et al., 2007), whereas MIZ2 is a guanine-nucleotide exchange factor for ADP ribosylation-type G proteins and was isolated and described as GNOM (Miyazawa et al., 2009).

Hydrotropism is dependent on ABA and auxin synthesis (Takahashi et al., 2009) but surprisingly not on auxin distribution, with the latter conclusion based on inhibitor studies (Kaneyasu et al., 2007). LRs also exhibited a hydrotropic response, which was abolished in the miz1 mutant (Iwata et al., 2012). Recently, the physiological relevance of hydrotropism was addressed in an elegant approach using a soil system. LRs elongated in the direction of soil patches of higher water potential, and MIZ-overexpressing plants exhibited greater survival and increased shoot biomass in this system (Iwata et al., 2013).

Root Responses to Salinity

One of the major agricultural contaminants restricting crop growth is NaCl. Salinization is an increasing problem worldwide, occurring as a result of both human actions and natural contamination of soil. In particular, irrigation practices have resulted in a gradual buildup of Na⁺ ions in the root zone in some soils, rendering these soils increasingly less suitable for crop growth (Munns and Tester, 2008). Salinity stress, similar to drought, frost, and cold stresses, causes osmotic stress to plants and restricts water uptake. As a result, many of the same signaling pathways are induced, and similar effects on root growth are generally observed (Munns and Tester, 2008). Salt causes additional problems to the plant because of the toxicity of Na⁺ ions. As a consequence, besides the regulation of osmotic balance, plants also mount other responses to withstand NaCl, most notably the activation of ion channels to export and/or compartmentalize salt (Hasegawa, 2013; Maathuis, 2014).

Salinity induces both ABA-dependent and -independent pathways in order to stimulate the formation of compatible solutes that help protect proteins and membranes as well as maintain osmotic pressure for water uptake (Munns and Tester, 2008; Hasegawa, 2013). Another response to salt is modulation of RSA (Fig. 3A), which is complex due to the counteracting effects of the ionic and osmotic components of salinity stress (Galvan-Ampudia and Testerink, 2011). It has been suggested that, although osmotic stress in general reduces LR formation, ionic stress counteracts this effect by promoting LR formation (Zhao et al., 2011). Only recently has the effect of salinity on RSA been studied in more detail at the mechanistic and cellular levels. Salinity was found to exhibit a specific effect on LR formation by arresting LR development just prior to emergence (McLoughlin et al., 2012). In the same study, two salt stress-activated SnRK2 protein kinases were identified to function in the maintenance of PR growth and LR formation under salinity stress but not in control conditions.

Figure 3.

High salt levels change growth, branching, and the direction of growth of Arabidopsis and tomato roots. A, Transfer of 4-d-old Arabidopsis seedlings to homogenous NaCl-containing medium inhibits PR elongation and LR growth. Photographs show 12-d-old seedlings. B, In an NaCl gradient, PRs grow away from areas of higher salt (negative halotropism) as salt diffuses from the bottom part (below the indicated diagonal lines) into the top part (above the diagonal lines) of the agar medium. The sos1 mutant exhibits increased sensitivity to NaCl compared with wild-type Col-0 and mounts a negative halotropic response at lower NaCl concentrations. Five days after germination in control conditions, the bottom part of the gel marked by the white diagonal lines was replaced with the same medium supplemented with 200 mm NaCl. Black marks identify the daily positions of the PR tips during the course of the experiment (Galvan-Ampudia et al., 2013). C, Tomato seedlings in soil display salt-induced inhibition of shoot and root growth. Seedlings were germinated on agar plates, and 3-d-old seedlings were transferred to soil and grown for another 4 d before being irrigated for 6 d with water (control) or 75 mm NaCl. Photographs show 13-d-old plants.

Studies investigating the effect of salinity on PR and LR development showed that LRs are more sensitive to salinity and that endodermal ABA responses are key to the observed arrest in LR development and growth (Duan et al., 2013). Time-lapse imaging revealed a temporary cessation of root growth in response to salinity, which was followed by the resumption of growth (Geng et al., 2013). This process is mediated by the coordinated action of several tissue-specific regulatory programs and involves the action of GA, BR, and JA (Geng et al., 2013). Both GA- and BR-responsive gene expression showed early repression and later recovery in response to salinity. Consistently, both hormones were shown to be important for the growth recovery phase, whereas JA likely plays a role in growth reduction (Geng et al., 2013). Interestingly, GA was found to have a more pronounced effect on LR growth than on PR growth during salt stress (Duan et al., 2013). In addition, nutrient status is an important modulator of RSA. The observation that LR growth is more sensitive to salt than PR growth is highly dependent on nutrient conditions (Duan et al., 2013): lower nutrient availability reduced the difference in response between the two tissues. Therefore, nutrient status is likely to explain some of the apparent discrepancies between studies on the effect of salt on RSA (Galvan-Ampudia and Testerink, 2011).

Inhibiting ABA responses by targeted expression of a dominant-negative ABA insensitive1 (abi1-1) allele in the endodermis, but not other root tissues, abolished the inhibitory effect of salt on LR growth (Duan et al., 2013), consistent with the importance of the endodermis in the regulation of LR formation (Vermeer et al., 2014). However, in these studies, efforts were focused on NaCl (McLoughlin et al., 2012; Duan et al., 2013; Geng et al., 2013), and it would be interesting to investigate how the observed hormonal signaling pathways could also regulate drought-induced RSA responses.

Halotropism

When Arabidopsis seedlings are transferred to salt-containing medium, random and agravitropic root growth is observed. Salt modulates the direction of growth by interfering with gravitropism (Sun et al., 2008; Galvan-Ampudia and Testerink, 2011), and this could be a response to avoid high salinity. Recently, it was shown that roots indeed can avoid salinity by changing their direction of growth (Galvan-Ampudia et al., 2013). In a diagonal gradient (Fig. 3B), PRs of Arabidopsis and tomato (Solanum lycopersicum) exhibited a directional, rather than a random, response to a salt gradient and grew away from high salt concentrations. This response is known as halotropism. Salt gradients induced changes in PIN2-mediated auxin distribution, resulting in more auxin at the side having the lower salt concentration. This caused the root to bend and grow away from the higher salt. This process was shown to depend on phospholipase D (PLD)-dependent and clathrin-mediated endocytosis of PIN2 at the salt-exposed side.

Interestingly, the response occurred at much lower osmotic stress levels than those inducing hydrotropism (Takahashi et al., 2002), and it was specific to NaCl, as it did not occur in response to mannitol or KCl gradients (Galvan-Ampudia et al., 2013). Consistent with the Na⁺ specificity of the response, the overly salt-sensitive mutants sos1, sos2, and sos3, which accumulate excess Na⁺ (Hasegawa, 2013; Maathuis, 2014), were found to be hyperresponsive to salinity gradients and exhibit halotropic growth at much lower NaCl concentrations than wild-type Arabidopsis plants (Galvan-Ampudia et al., 2013; Fig. 3B). These results indicate that halotropic growth is specifically triggered by intracellular Na⁺ ions. An important question that remains to be answered is whether adequate salt avoidance by the root affects shoot growth, whole-plant performance, and reproduction.

The salt avoidance response was not only observed in agar plate assays but also for tomato and sorghum plants grown in soil (Galvan-Ampudia et al., 2013; Fig. 3C). As natural soils are not uniform in their salt concentrations (Rengasamy, 2006), the halotropic response is expected to be especially relevant under natural conditions. Delaying the effect of salt exposure by halotropic growth would allow more time for the plant to mount other salt tolerance responses, including ion channel regulation (Munns and Tester, 2008; Hasegawa, 2013). This would give the plant a competitive advantage over plants that cannot avoid initial salinity.

Cellular Signaling Pathways Involved

How the root senses salinity and drought is currently unknown, and genetic screens have failed to determine the identity of molecular receptors. Whereas drought-induced responses could rely on the detection of physical changes in the membrane by receptors or channels, the nature of the perception of the Na⁺ signal remains particularly elusive (Maathuis, 2014). On the other hand, significant progress has been made in elucidating the early signaling responses to salt and drought. These include an increase in the cytosolic level of Ca²⁺ (Tracy et al., 2008; Maathuis, 2014) and the enzymatic remodeling of membrane lipids, the latter leading to the release of lipid second messengers (Munnik and Testerink, 2009; Xue et al., 2009). In the case of salinity, Ca²⁺ was shown to be essential for ion homeostasis control (i.e. the maintenance of sodium/potassium balance) via the regulation of ion transport by CALCINEURIN B-LIKE (CBL) proteins/CBL-interacting protein kinases (Steinhorst and Kudla, 2013). Phospholipid signals and their downstream targets are emerging as signaling pathways that integrate stress with development (such as RSA) through ABA and auxin signaling. Other recently identified factors in both stress and root growth and development are the family of receptor-like kinases (RLKs; Marshall et al., 2012).

Cellular Membrane Trafficking

Polar auxin transport appears to be a major integrator through which environmental signals coordinate the development and growth of root and shoot. In particular, the localization and function of PIN2 is a target in signaling pathways induced by light, water availability, and salinity (Sun et al., 2008; Zhao et al., 2011; Sassi et al., 2012; Galvan-Ampudia et al., 2013). However, the upstream signal transduction pathways involved have only recently started to become clear.

Both salinity and drought induce the formation of lipid second messengers, including phosphatidic acid (PA). PA is formed both from the cleavage of membrane phospholipids such as phosphatidylcholine and phosphatidylethanolamine by PLD and through the phosphorylation of diacylglycerol by diacylglycerol kinase (Li et al., 2009; Testerink and Munnik, 2011). Mutants in individual or multiple PLD isoforms were shown to be impaired in root responses to salinity and osmotic stress. Mutants in the C2 domain-containing α1, α3, and δ isoforms showed reduced root growth and branching in salt, but not control, conditions (Bargmann et al., 2009; Hong et al., 2010). Also, the Phox-Pleckstrin Homology type PLDζ2 isoform was shown to function in both hydrotropic and halotropic responses (Taniguchi et al., 2010; Galvan-Ampudia et al., 2013), most likely through modulation of the clathrin-mediated endocytosis of PIN2 (Galvan-Ampudia et al., 2013). These results are supported by the identification of clathrin H chain and clathrin-mediated endocytosis AP180, amino-terminal, homology/epsin, amino-terminal homology domain-containing protein accessory proteins as PA-binding proteins that are recruited to cellular membranes upon treatment with salt (McLoughlin et al., 2013).

PA-Binding Protein Kinases and Phosphatases

PA accumulation activates specific downstream signaling pathways through the recruitment of protein targets to the membrane and direct modulation of the activity of some of these targets (Testerink and Munnik, 2011; McLoughlin and Testerink, 2013). Interestingly, several of these PA target proteins are protein kinases and phosphatases with regulatory functions in plant development and growth. As such, osmotic stress-induced PA formation is emerging as a signal to regulate downstream RSA responses (McLoughlin and Testerink, 2013; Fig. 4).

Figure 4.

Drought and salinity regulate a network of phospholipid signals, ion channels, protein kinases, and hormones. This results in the modification of RSA to increase salt and/or drought tolerance. Brown boxes indicate plant hormones, beige boxes indicate protein kinases, and light green shading indicates ion transport. The dashed black arrow indicates that aquaporin cycling is partially clathrin mediated but the increase in cycling observed by salt is not (Luu et al., 2012). Dashed red arrows indicate direct physical interaction of the protein or process with PA (see cellular signaling section).

Both the PINOID (PID) kinase and the ROOTS CURL IN NAPHTHYLPHTHALAMIC ACID1 (RCN1) phosphatase, which modulate PIN polarity in cells, have been shown to directly interact with PA (Testerink et al., 2004; Zegzouti et al., 2006; Gao et al., 2013). PID is an AGC protein kinase that affects PIN polar distribution through direct phosphorylation (Michniewicz et al., 2007). RCN1 is a regulatory subunit of the Protein phosphatase 2A (PP2A) complex, which functions antagonistically to PID through dephosphorylation of the central hydrophilic loop of PIN (Michniewicz et al., 2007). It is not known whether the observed effects of salt on PIN2 trafficking and halotropism (Galvan-Ampudia et al., 2013) are modulated by PID or RCN1.

ABI1 is one of the PP2Cs comprising the ABA receptor. Both ABI1 relocalization in response to ABA and ABI1 activity are inhibited by PA (Zhang et al., 2004). This pathway was shown to be relevant for stomatal closure (Mishra et al., 2006), but the recently discovered importance of ABA and ABI1 in LR formation (Duan et al., 2013) suggests that it could also play a role in osmotic stress-induced RSA responses (McLoughlin and Testerink, 2013). MPK6 is an osmotic stress-induced mitogen-activated protein kinase that is involved in root growth and might function by direct phosphorylation of SOS, which it phosphorylated in vitro (Yu et al., 2010).

The SnRK2 family subclass 1 protein kinases SnRK2.4 and SnRK2.10 are involved in maintaining PR and LR growth under salinity stress (McLoughlin et al., 2012), but unlike the subclass 3 described above, they are independent of ABA. Although it is possible that SnRK2.4 and SnRK2.10 are regulated by the same phosphatases that function in ABA signaling, they show different phosphorylation characteristics (Kulik et al., 2011). In addition, SnRK2 class 1 protein kinases (but not the ABA-dependent SnRK2.6; Julkowska et al., 2014) directly bind PA and relocalize to cellular membranes in response to salinity, suggesting alternative regulatory mechanisms that were recently summarized (McLoughlin et al., 2012; McLoughlin and Testerink, 2013). It will be important to identify the physiological phosphorylation targets of the subclass 1 SnRK2s to elucidate how these kinases regulate RSA.

RLKs

Several RLKs, including the BR receptor BRI1, play roles in drought and salt stress responses (Marshall et al., 2012). Phenotyping of a collection of 69 root-expressed Leu-rich repeat RLK knockout mutants revealed several that were involved in both the response to salt/osmotic stress and auxin and the inhibition of auxin transport in roots (ten Hove et al., 2011). Moreover, in Medicago truncatula, another Leu-rich repeat RLK, Slrk, was shown to be induced by salt and to mediate the reduction of root growth in response to salinity (de Lorenzo et al., 2009). The slrk mutants showed higher root growth under salinity stress and accumulated less sodium in their root systems (de Lorenzo et al., 2009). The exact molecular role of RLKs in the integration of stress and root development remains to be established.

Regulation of K⁺ Transport

Recently, K⁺ transporters were shown to be involved in osmotic stress- and ABA-induced changes in RSA. Knockout phenotypes of three K⁺ UPTAKE PERMEASE (KUP) K⁺ transporter family members suggested that they mediate K⁺ efflux in roots, together with the guard cell outward-rectifying K⁺ channel GORK (Osakabe et al., 2013). Mutants exhibited enhanced cell expansion, increased sensitivity to auxin, and reduced sensitivity to ABA, the latter both with respect to stomatal closure and inhibition of LR formation. Possibly, higher K⁺ uptake in the mutants explains their increased ability to maintain LR growth, even under salt stress conditions. Interestingly, KUP6 was found to be a direct phosphorylation target of the SnRK2 subclass 3 protein OPEN STOMATA1. Triple knockout mutants of all subclass 3 SnRK2s (srkdei) exhibited a similar phenotype to kup knockout mutants with respect to LR formation under salt stress and in response to ABA and auxin (Osakabe et al., 2013). Accordingly, in the more salt-resistant ethylene overproducer1 mutant, K⁺ levels in the root were elevated, and HIGH AFFINITY K⁺ TRANSPORTER5, another KUP family transporter, which mediates K⁺ influx, was up-regulated (Jiang et al., 2013).

What Is an Optimal RSA under Stress?

The effect of drought on the overall RSA of the plant is determined not only by the nature and severity of this stress but also by its duration and timing during the life cycle of the plant, and this effect is carefully balanced with shoot growth and reproduction (Claeys and Inzé, 2013). Although overall smaller, plants that experience drought invest relatively more mass in roots and reduce leaf surface to limit transpiration. In mechanistic studies aimed at the identification and characterization of genes that contribute to root plasticity, research has focused on seedlings, which are usually grown on agar plates. In this way, several highly relevant factors have been identified (e.g. by mutant screens and genetic and cell biological studies). In particular, the contribution of hormones and signaling pathways has been well characterized in this system (Fukaki and Tasaka, 2009).

On the other hand, most physiological and ecological studies have focused on root systems of adult plants that were grown either in soil or hydroponically, leaving a massive gap between mechanistic understanding and physiological and agricultural relevance. Closing this gap in our knowledge is essential to better understand how root plasticity under stress contributes to stress tolerance in the field or greenhouse. Recently, Uga et al. (2013) provided a compelling case for the relevance of RSA, in particular root growth angle, for the drought tolerance of rice. Cloning and characterization of the DEEPER ROOTING1 (DRO1) locus, which encodes for a protein with no known or predicted function, revealed its function in the downward direction of root growth, allowing for deeper rooting. Introduction of the DRO1 locus in shallow-rooting rice increased yield under drought conditions, with no apparent yield penalty under well-watered conditions (Uga et al., 2013).

FLEXIBLE SHOOTS AND ROOTS UNDER STRESS

Light Signaling Affects Root Development

It has been proposed that the reduced root investments upon low R:FR exposure are one reason why weeds early in the growing season can have lasting negative effects on crop yield at the end of the growth period. The presence of weeds leads to FR enrichment through reflection FR, thereby promoting shoot over root growth in the crops (Page et al., 2011), and these impoverished root systems may be less able to acquire belowground resources (Rajcan et al., 2004). Likewise, natural populations of Impatiens capensis become more sensitive to drought when expressing shade avoidance responses as compared with less shade-avoiding populations, presumably through reduced root investments in shade-avoiding plants (Huber et al., 2004). Although there have been studies showing yield suppression by low R:FR exposure in crops (Boccalandro et al., 2003; Ugarte et al., 2010), these have not considered light quality-induced suppression of root growth.

Understanding how light quality at high densities affects not just shoot development but also root development would facilitate crop improvement strategies aimed at optimizing root development at high-density light qualities. Salisbury et al. (2007) showed that LR formation is reduced in a phyb mutant in both soil and plate assays and upon low R:FR exposure in the wild type. However, it is unknown whether this is a consequence of an overall delayed rate of plant development or a selective change of the root system. The effect of varying R:FR conditions acted through signaling in the shoot, and it is proposed that this occurs through shoot-derived auxin (Salisbury et al., 2007). Indeed, low R:FR conditions in the shoot do affect auxin biosynthesis, transport, and signaling (see shade avoidance section), making this a feasible scenario for light quality-driven, shoot-controlled regulation of root development. As these R:FR manipulation studies were performed on agar plates in which both shoot and root were R:FR treated (Salisbury et al., 2007), future studies are needed that would treat shoots with varying ratios of R:FR while keeping the root system untreated and in darkness.

Another study comparing light- versus dark-grown seedlings also identified shoot-derived auxin as a key component in regulating PR elongation in a light-dependent manner (Sassi et al., 2012). It was shown that PIN1 transcription in the shoot was regulated by light in a COP1-dependent manner. As a consequence, PIN1 accumulates in the light, facilitating increased polar auxin transport toward the root system. In addition, PIN2 is involved in controlling root growth in a light-dependent manner, and COP1 regulates the light-dependent intracellular distribution of both PIN1 and PIN2 between the cytoplasm and plasma membrane in the root apical meristem. Although auxin is a strong candidate for shoot-mediated control of root growth in a light-dependent manner, shoot-derived sugars are an obvious alternative candidate. Indeed, Suc transported from the shoot to the root has also been shown to control root growth in a light-dependent manner (Kircher and Schopfer, 2012). In the early development of dense stands, FR enrichment is unlikely to have a major effect on rates of photosynthesis, making Suc a less likely candidate signal for root growth control. However, at later phases of development, shading between individual plants will most probably impact photosynthesis; therefore, altered auxin and Suc transport from the shoot to the root system may coregulate root development. In addition to light availability and quality, the length of the light period affects RSA and interacts with specific RSA responses to nutrient deficiencies (Kellermeier et al., 2014).

Does Stress Signaling in Roots Affect Shoot Plasticity?

Belowground stresses such as drought or salinity affect whole-plant performance, including shoot performance. Signaling intermediates involved in root system responses to abiotic stress may also affect shoot development directly. The accumulation of ABA in the shoot upon drought or salinity may affect shoot growth and plasticity to aboveground stresses. For example, escape from submergence is highly dependent on ABA depletion and is inhibited by high ABA levels (Chen et al., 2010). Low R:FR-induced shade avoidance, on the other hand, is not (Pierik et al., 2011), although low R:FR-mediated shoot branching in Arabidopsis does involve ABA (Reddy et al., 2013). In addition, belowground stresses that affect root system size potentially affect the accumulation of root-derived signaling intermediates such as cytokinin and strigolactones that may, in turn, impact shoot growth. Quantification and manipulation of the levels of hormones and signaling intermediates at the tissue- and cell-specific levels will be key in future research in this area.

In addition to signaling intermediates, water uptake and availability to the shoot will be crucial in driving growth and escape responses. Shade avoidance relies on turgor-driven cell expansion; therefore, water limitation due to drought or salinity is likely to suppress the shoot elongation response to R:FR.

FUTURE PERSPECTIVES

Because of global climate change and intensified use of agricultural fields, the control of root development by light quality will likely become increasingly relevant. The need for increased yield may also call for still higher plant densities, which will lead to even stronger FR reflection and subsequent root growth depression. However, impoverished root development under low R:FR conditions may have considerable negative impacts on plant growth. Not only will a smaller root system capture fewer nutrients, but also it will increase plant sensitivity to drought. Furthermore, the light-derived change in polar auxin transport toward the root system may affect the plasticity of the root system by affecting proliferation into nutrient-rich and/or water-rich patches, escaping salinity, and modifying the GSA of LRs in foraging for water. Shade avoidance responses are likely to further intensify with increasing global temperatures (Patel et al., 2013). It is unknown whether this will also further reduce root investments, but because light and temperature responses both occur through PIF proteins and PIF-mediated auxin control (Franklin et al., 2011; Li et al., 2012), this seems possible.

It is important to determine how multiple stresses are integrated at the levels of stress signaling, signal transduction, and phenotypic responses. It is possible that certain stress combinations cancel each other out, whereas combinations of others can be devastating for plant performance and yield. However, this cannot be predicted from knowledge of the single stresses alone: gene expression studies indicate that combining the stresses due to salt, mannitol, and heat treatments induces unique gene expression patterns that cannot be simply inferred from those induced by the individual stresses (Sewelam et al., 2014). Another study compared transcriptome changes in 10 Arabidopsis ecotypes in response to cold, heat, high light, salt, and flagellin treatments, as well as combinations of these stresses. Again, the majority of the transcriptome changes in response to double stresses were not predictable from the responses to single stress treatments (Rasmussen et al., 2013b).

Therefore, we argue for an integrative research agenda that transcends the single-stress and single-organ approaches. Although these highly focused approaches have been, and will continue to be, tremendously successful in elucidating pathways for stress resistance and plasticity, current global challenges call for studies on multiple stress responses that include root-shoot integration.

Glossary

FR

far-red light

SAS

shade avoidance syndrome

JA

jasmonic acid

B

blue light

R

red light

G

green light

PAR

photosynthetically active radiation

PIF

phytochrome-interacting factor

bHLH

basic helix-loop-helix

IAA

indole acetic acid

EODFR

end-of-day far-red light treatment

BR

brassinosteroid

ABA

abscisic acid

RSA

root system architecture

PR

primary root

LR

lateral root

GSA

gravitropic set-point angle

RLK

receptor-like kinase

PA

phosphatidic acid

PLD

phospholipase D

Col-0

Columbia-0