Introduction

Light, water and nutrients are the major limiting factors affecting plant growth and fitness in nature. To obtain sufficient energy and nutrients, plants as sessile organisms in nature alter the morphological and physiological traits in shoots and roots depending on resource availabilities in the environment. Among these traits, root development is crucial for plant growth, as it is strongly affected by the balance between resource uptake and investment. The expansion of the root system into the soil environment may increase the capacity for nutrient uptake, while the size of the root biomass can be limited by resource investment in photosynthesis. Given the significance of root development and proper functioning in nutrient acquisition, the effect of soil minerals on root morphology and underlying mechanisms have been intensively studied over the past few decades (López-Bucio et al. 2003; Krouk et al. 2011; Gruber et al. 2013; Giehl and von Wirén 2014).

Mineral nutrients are distributed unevenly in the soil environment. The top soil layer is generally more nutrient-rich than deep underground layers, and it contains relatively high concentrations of essential elements, such as phosphorus (Jobbágy and Jackson 2001; Lynch 2011). In contrast, nitrate and sulfate may be more abundant in deep underground layers because of their high mobility within the soil (Lynch 2013; Giehl and von Wirén 2014). In addition to mineral nutrients, water needs to be taken up by the root system, and a deeper root system is thought to be beneficial for efficient water uptake (Sanders and Markhart 1992; Uga et al. 2013). The development of widespread root systems leads to a consumption of large amounts of photoassimilates for respiration and needs to be balanced with carbon assimilation carried out in photosynthetic organs. However, organic acids produced in respiration can be exuded from roots and support the acquisition of certain nutrients, such as iron and phosphate. Thus, the root system needs to meet the demand for nutrients and water of the whole plant.

Plant root morphology changes in response to phytohormones including auxin as the major signal and determinant (Lavenus et al. 2013; Rodriguez-Villalon and Hardtke 2014; Satbhai et al. 2015). However, a number of studies presented over the last two decades indicate that root morphology is modulated also by the function of small signaling peptides (Murphy et al. 2012; Delay et al. 2013; Matsubayashi 2014; Endo et al. 2014; Murphy and De Smet 2014). Small peptides encoded by the CLAVATA3/EMBRYO SURROUNDING REGION (CLE) gene family have drawn particular attention as they are expressed in both shoots and roots and control the meristem functions. CLE peptides bind to leucine-rich repeat receptor-like kinases (LRR-RLKs) that initiate signal transduction pathways. The spatio-temporal functions of these peptide-receptor signaling modules are central to the control of stem cell division and differentiation. This review article will give an overview of the CLE peptide-driven CLAVATA signaling pathways and will further provide insights into their potential interactions with N signaling in roots.

CLAVATA signaling pathways regulate cell proliferation in the shoot apical meristem

Documentation of developmental traits associated with CLAVATA1 (CLV1) dates back to the 1960s in literature reporting lumpy and club-shaped siliques and early flowering phenotypes of Arabidopsis mutants (McKelvie 1961). However, detailed descriptions of CLV1 gene functions awaited later investigations which characterized the clv1 mutant phenotypes particularly in relation to mechanisms controlling the size of the shoot apical meristem (SAM) and the number of floral organs besides changes in silique shapes (Clark et al. 1993, 1997). CLV1 encodes a receptor kinase from the LRR-RLK family and is expressed in the SAM and floral meristems (Clark et al. 1997). Further genetic studies indicate that other CLV genes (CLV2 and CLV3) act in the same signaling pathway (Clark et al. 1995; Kayes and Clark 1998; Fletcher et al. 1999; Jeong et al. 1999), while they have distinct roles. CLV2 encodes a leucine-rich repeat (LRR) protein anchored on the plasma membrane, while CLV3 encodes a tri-arabinosylated 13-amino acid small signaling peptide expressed in the SAM (Ogawa et al. 2008; Ohyama et al. 2009). The clv2 mutant shows a weak phenotype (Kayes and Clark 1998), and the function of the LRR protein encoded by CLV2 is dependent on the kinase domain protein CORYNE (CRN) (Müller et al. 2008; Bleckmann et al. 2010). CLV2 and CRN interact and assemble into an LRR-kinase complex requiring both of the interacting components for the intra-cellular trafficking from the endoplasmic reticulum to the plasma membrane (Bleckmann et al. 2010). The CLV3 peptides are excreted to the apoplast of the central zone, and the binding of CLV3 to CLV1 initiates a signaling pathway that controls the expression of WUSCHEL (WUS) in the organizing center (OC) in SAM (Laux et al. 1996; Mayer et al. 1998; Schoof et al. 2000; Rojo et al. 2002). WUS encodes a homeobox transcription factor, whose expression is also regulated by cytokinins, independent of the CLV3-CLV1 signal (Mayer et al. 1998; Leibfried et al. 2005). WUS can diffuse from the OC to other cells via plasmodesmata to reach the surface of the SAM (Yadav et al. 2011). The diffusion of WUS from OC to SAM surface enables WUS to induce the expression of CLV3 (Schoof et al. 2000). WUS represses cell division of OC, and thereby the feedback loop of the CLV3-CLV1-WUS pathway maintains the size of the SAM.

Another set of receptor-like kinases affecting the size of the SAM are LRR-RLKs encoded by BARELY ANY MERISTEM 1 (BAM1), BAM2 and BAM3 (DeYoung et al. 2006). The bam single mutants do not show clear phenotypes, while the bam1 bam2 double mutant exhibits significant defects, such as a SAM with reduced size, asymmetric leaf formation, decrease in floral organ numbers and abortion of anther development (DeYoung et al. 2006; Hord et al. 2006). BAM1 and BAM2 are expressed broadly in the SAM and other plant organs and have partially overlapping roles with CLV1. Introducing bam1 and bam2 mutations can partially rescue the phenotype of the clv3 mutant, suggesting that BAM1 and BAM2 sequester CLE peptides from the CLV1-mediated signaling pathway (DeYoung and Clark 2008). Several lines of genetic evidence corroborate the idea of functional redundancy and interactions among BAMs and CLV1. The null recessive mutation of CLV1 (clv1-null) causes weak phenotypes, while the clv1-null bam1 bam2 triple mutant presents stronger phenotypes (DeYoung and Clark 2008) comparable to those observed in the clv1 mutants with dominant negative alleles (Diévart et al. 2003). The clv1-null bam1 bam2 triple mutant shows stronger phenotypes compared to the clv3 mutant, suggesting involvement of other CLE peptides in the CLV signaling pathway (DeYoung and Clark 2008). The clv1-null mutant shows phenotypes even weaker than the clv3 mutant, suggesting partially redundant roles of BAM1 and BAM2 in the CLV signaling pathways (Nimchuk et al. 2015).

LRR-RLKs assemble into homo- or hetero-multimeric protein complexes that appear to modulate their roles in controlling SAM differentiation based on their potential functional redundancies. CLV1 can make homo-multimer protein complexes, while it also binds to CLV2/CRN, BAM1, BAM2 and RECEPTOR-LIKE PROTEIN KINASE 2 (RPK2) (Bleckmann et al. 2010; Zhu et al. 2010; Guo et al. 2010; Betsuyaku et al. 2011). CLV2 itself cannot be directly associated with CLV1, however CRN can interact with CLV1 and probably contributes to the assembly of the CLV1-CRN-CLV2 complex (Zhu et al. 2010). BAM1 and BAM2 interact with CLV1, CLV2/CRN and RPK2 (Guo et al. 2010; Shimizu et al. 2015). RPK2 can also form a homodimer (Kinoshita et al. 2010).

CLAVATA signaling pathways control root development

The CLV-like signaling cascade is also found in the root apical meristem (RAM). Arabidopsis CRINKLY4 (ACR4) (Gifford et al. 2003; Watanabe et al. 2004) is a gene identified to encode a receptor-like kinase homologous to CRINKLY4 (CR4) in maize (Becraft et al. 1996; Tanaka et al. 2002). The maize mutant for CR4 shows defects in the development of the epidermis and aleurone structure (Becraft et al. 1996; Jin et al. 2000; Becraft et al. 2001). The Arabidopsis acr4 mutant shows disorganized epidermal cell differentiation, sepal margin, and defects in ovule development (Gifford et al. 2003; Watanabe et al. 2004). The acr4 mutant and the acr4 ccr double and triple combinatorial mutants with mutations introduced in ACR4 and its homologs (CCRs) exhibit significant morphological phenotypes in roots, such as frequent formation of lateral root primordia, led by enhanced cell divisions in the pericycle of the vasculature and development of multiple disorganized columella stem cell layers in the RAM (De Smet et al. 2008). Consistent with these phenotypic observations, ACR4 gene expression occurs predominantly in lateral root initials and columella stem cells (Gifford et al. 2003; De Smet et al. 2008).

In this signaling pathway, CLE40 is likely a ligand peptide bound to ACR4 (Hobe et al. 2003; Stahl et al. 2009, 2013). Downstream of the CLE40-ACR4 signaling module, WUSCHEL RELATED HOMEOBOX5 (WOX5) is expressed in the quiescent center (QC) for maintenance of stem cell activity in the RAM (Haecker et al. 2004; Stahl et al. 2009; Forzani et al. 2014; Ji et al. 2015). WOX5 can move from the QC to columella stem cells located on the distal side of QC, and repress cell division by suppressing the activity of cyclin D (Sarkar et al. 2007; Forzani et al. 2014). However, when CLE40 peptide (CLE40p) is externally applied to roots, the ACR4 expression domain can be expanded over the QC, and this in turn alters the site for WOX5 expression to be shifted towards the proximal side of the QC (Stahl et al. 2009). The application of CLE40p leads to consumption of columella stem cells, while the cle40 mutation causes an opposite effect, i.e. columella stem cells that divide into multiple cell layers as in the acr4 mutant (Stahl et al. 2009, 2013). Accumulation of auxin in the RAM also controls WOX5 gene expression to be confined in the QC, as auxin efflux carriers (PIN3, PIN4 and PIN7) distribute auxins to columella cells, where IAA17 and ARF10/16 take actions to repress WOX5 expression (Ding and Friml 2010). Furthermore, WOX5 induces the expression of PLETHOLA (PLT) (Ding and Friml 2010), which allows stem cells at the proximity of QC to be maintained at an undifferentiated state (Aida et al. 2004; Galinha et al. 2007). At the proximal side of the QC, REPRESSOR OF WUSCHEL1 (ROW1) is expressed in the proximal meristem and controls WOX5 gene expression to be confined in the QC, as ROW1 binds to trimethylated histone H3 lysine 4 in the WOX5 promoter region (Zhang et al. 2015).

The CLE40-ACR4 signaling module appears to play a central role in controlling differentiation of the RAM in Arabidopsis. However, there seems to be a room for other pathways to be involved. With regard to CLE peptide perception, a degree of functional redundancy may exist between ACR4 and other LRR-RLKs. ACR4 interacts with its homologs (i.e. CRRs) to form receptor complexes (Meyer et al. 2011, 2015). Furthermore, ACR4 is required for CLV1 expression and it physically interacts with CLV1 to control stem cell differentiation in the RAM possibly by modulating the mobility of stemness signals through plasmodesmata (Stahl et al. 2013). Other LRR-RLKs and WOX5-independent mechanisms that activate CLE peptide signaling pathways may also be present in the RAM, as the wox5 mutant phenotypes are enhanced by CLE40p application. Application of CLV3p and ectopic overexpression of CLV3 disrupt the RAM (Hobe et al. 2003; Fiers et al. 2005). Other CLE peptides (CLE19, CLE25, CLE26) have similar effects on the RAM (Casamitjana-Martínez et al. 2003; Kinoshita et al. 2007; Miwa et al. 2008; Czyzewicz et al. 2015). Therefore, in addition to CLE40, other CLE peptides expressed in RAM may be participating in WOX5-dependent and independent signaling pathways. However, in contrast to the mechanisms under control of ACR4, CLV1 and related LRR-RLKs, the CLV2-CRN LRR-kinase complex seems to be important only for perception of CLE peptide signals to control proximal cell division (Pallakies and Simon 2014).

Systemic integration of nitrogen and CLAVATA signaling pathways in roots

The CLE peptides in the group of CLE1 to CLE7 and their receptor CLV1 comprise a signaling module that regulates lateral root emergence in Arabidopsis in a N-dependent manner (Araya et al. 2014a) (Fig. 1). In particular the expression of CLE1/3/4/7 is induced in roots under low-nitrate conditions, and the corresponding CLE peptides are produced predominantly in pericycle cells. Lateral root growth inversely correlates with CLE3 expression levels in wild type but not in clv1-4 mutant plants, suggesting that CLV1, localized in companion cells of phloem, is the receptor (Araya et al. 2014a). The mechanisms that control lateral root development downstream of CLV1 appear to involve yet unknown long-distance signals transferred through the phloem sieve elements, because the CLE receptor, CLV1, is expressed in the phloem companion cells (Araya et al. 2014b). Thus, the mode of action of this CLV1 signaling pathway appears to be different from the ACR4-mediated pathway controlling cell division of lateral root initials (De Smet et al. 2008). Available transcriptome data on N responses in Arabidopsis (Scheible et al. 2004; Patterson et al. 2010; Ruffel et al. 2011) further suggest that CLE3 expression is also enhanced by a systemic nitrate demand signal and following application of ammonium to N-starved plant seedlings (Araya et al. 2014a) (Fig. 1). Despite the lines of evidence suggesting the roles of these CLE peptides (CLE1–CLE7) in regulation of lateral root development, overexpression and peptide application of none of them affected primary root development (Strabala et al. 2006; Kinoshita et al. 2007; Araya et al. 2014a).

Fig. 1
figure 1

Interaction of CLE and N signaling pathways in root and nodule development. As represented by the CLE3-CLV1 pathway in Arabidopsis (Araya et al. 2014a), CLE peptides are produced under low-N conditions or in response to systemic nitrate (NO −3 ) demand signals (left). CLE3 gene expression can also be induced in response to ammonium (NH +4 ). CLE peptides bind to CLV1. Signals downstream of CLV1 repress lateral root emergence. In addition to N signals controlling CLV1-dependent pathways, NH +4 locally stimulates lateral root initiation (Lima et al. 2010; Araya et al. 2016) and NO −3 demand signals promote root elongation (Ruffel et al. 2011). In legume plant species, rhizobia-induced CLE (RIC) and nitrate-induced CLE (NIC) peptides bind to leucine-rich repeat receptor-like kinases (LRR-RLK) (right). Both long-distance and local signals repress nodule development (Reid et al. 2011; Okamoto et al. 2013)

Full size image

The model proposed for these signaling pathways involves: (1) CLE3 being expressed in roots under low nitrate supply or de-repressed by systemic N demand signals, (2) CLV1 in the phloem companion cells for perception of CLE peptides, and (3) CLV1-dependent downstream signals that are systemically translocated through the phloem to control the outgrowth of lateral root primordia at a distant location (Araya et al. 2014a, b). Furthermore, the signals downstream of CLV1 feedback regulate the levels of CLE peptides (CLE2/3/4/7) in the pericycle. Therefore, it might be possible that the absence of CLV1 in the clv1 mutant would release the feedback mechanism and allow these CLE peptides to accumulate and interfere with other CLEs (Czyzewicz et al. 2015) or to affect ACR4 and related LRR-RLK-mediated pathways controlling lateral root development (De Smet et al. 2008). However, this seems to be an unlikely scenario, because the frequencies of lateral root formation are found to be fairly consistent between the wild type and clv1 mutants (Araya et al. 2014a). The CLE3-CLV1 module rather controls lateral root emergence from the primary root. Various combinations of CLE peptides and LRR-RLKs including CLV1 may be redundantly functional in this regulatory mechanism, although the actual contributions of these homologous components remain unclear. The function of CLV1 appears to be robust for this N-dependent mechanism, as suggested by the potential inhibitory action of CLE3 on lateral root development, being almost completely prevented by the loss of its receptor CLV1 in clv1 mutants (Araya et al. 2014a).

In addition to the CLV1-mediated signaling pathway, several other mechanisms control lateral root development in N-dependent manners (Forde 2014; Giehl and von Wirén 2014). Increasing the number of lateral roots appears to be an effective strategy for NO −3 foraging, as it is modulated in two ways by NO −3 signals integrated with auxin response mechanisms. NO −3 acts as an effector for enhancing gene expression of auxin receptor AFB3 and for activation of its downstream network, while biosynthesis of N metabolites that follows NO −3 uptake results in induced expression of miR393 for post-transcriptional gene silencing of AFB3 (Vidal et al. 2010, 2013). Cell divisions of lateral root initials allow formation of lateral root primordia that grow and emerge from primary roots. Along this developmental process, CLV1-mediated signaling pathway inhibits pre-emergent growth of lateral root primordia (Araya et al. 2014a). NO −3 availabilities also alter auxin levels in lateral root primordia. When NO −3 supply is limited, NO −3 transporter NPF6.3/NRT1.1 mediates basipetal auxin redistribution and limits the growth of lateral root primordia (Krouk et al. 2010). The NO −3 transceptor function of NPF6.3/NRT1.1 appears to be a key element required for inducing gene expression of high-affinity NO −3 transporter NRT2.1 (Ho et al. 2009). NPF6.3/NRT1.1 also acts on other downstream NO −3 responses including NO −3 -dependent lateral root development (Remans et al. 2006; Krouk et al. 2010; Bouguyon et al. 2015). In contrast, the expression of the auxin biosynthetic enzyme TAR2 is induced during the early stages of lateral root development in response to N deprivation (Ma et al. 2014). These lines of evidence suggest that auxin remains active under N deficiency, while its positive effect may be counterbalanced by inhibitory mechanisms that control lateral root development to prevent extension of root systems into the N-poor environment. It is likely that these signaling mechanisms are affecting each other to optimize the growth and development of lateral roots in response to N availabilities.

In legumes, CLV1-like LRR-RLKs are essential for systemic autoregulation of nodulation (AON), a long-distance mechanism controlled by N availability (Djordjevic et al. 2015). HYPERNODULATION ABERRANT ROOT FORMATION 1 (HAR1) in Lotus japonicus, NODULE AUTOREGULATION RECEPTOR KINASE (NARK) in soybean, and SUPER NUMERIC NODULES (SUNN) in Medicago truncatula are the CLV1-like LRR-RLKs mediating CLE-signaling pathways for the AON (Krusell et al. 2002; Nishimura et al. 2002; Searle et al. 2003; Schnabel et al. 2005; Miyahara et al. 2008). The CLE peptides involved in AON are expressed in response to supply of inorganic N sources or following infection of Rhizobium species that establish symbiotic relationships with legume plants for N2 fixation. The inhibition of nodule formation through this mechanism requires the function of CLV1-like LRR-RLKs (Okamoto et al. 2009; Mortier et al. 2010; Reid et al. 2011; Saur et al. 2011; Mortier et al. 2012; Okamoto et al. 2013). It has been suggested that nitrate-induced CLEs (NICs) and rhizobia-induced CLEs (RICs) play distinct roles in AON in soybean by interacting with the CLE receptor NARK (Reid et al. 2011) (Fig. 1). NICs bind locally to the receptor in roots, while RICs translocate from roots to shoots to meet the receptor that subsequently transfers shoot-derived inhibitory signals back to roots to feedback regulate nodule development (Reid et al. 2011). Thus, both of these signaling pathways operate for AON but in slightly different manners, requiring distinct members of CLE peptides representing nitrate- or rhizobia-induced signals (Okamoto et al. 2009; Mortier et al. 2010; Reid et al. 2011; Saur et al. 2011). CLE-RS2 from L. japonicus is functional when it is arabinosylated, and this mature form [Ara3]CLE-RS2 binds to HAR1 but not the har1-4 mutant form which carries a missense mutation in the LRR domain (Okamoto et al. 2013). This mature arabinosylated form of a CLE peptide has been found in the xylem sap collected from the aerial part of L. japonicus plants, providing direct evidence that RIC is transported as a root-derived signal to bind to the LRR-RLK in shoots (Okamoto et al. 2013). The shoot-derived inhibitory signals that complete the circuit of AON mechanism have yet to be identified (Fig. 1).

Future perspectives

The integrations of CLAVATA signaling pathways into N-responsive and rhizobia-dependent signaling pathways imply that they may play important roles in optimizing root shape and nodule numbers in response to changes in N supply in the environment. In their mature forms CLE peptides involved in these regulatory processes are glycosylated. The arabinosylation of CLE peptides appears to be a conserved mechanism to retain the functionalities of CLV3 and CLE2 in Arabidopsis (Ogawa et al. 2008; Ohyama et al. 2009) and CLE-RS2 in L. japonicus (Okamoto et al. 2013), however, for distinct purposes in a different biological context. RICs are not found in Arabidopsis, suggesting that they would have been developed evolutionarily in legumes specifically for the regulation of nodule development. Studying downstream signals of CLAVATA signaling pathways in non-nodulating and nodulating plant species may provide novel insights into relationships between lateral root and nodule development that are both relevant to N acquisition.

A few fundamental questions related to the physiological relevance of these CLAVATA-dependent signaling pathways in the communication between roots and shoots remain to be investigated. The shoot-derived inhibitory signals downstream of CLV1-like LRR-RLKs for AON in legumes are yet unidentified, although they are known to interact with cytokinin signaling pathways in roots (Reid et al. 2011; Saur et al. 2011; Mortier et al. 2012). In Arabidopsis shoots, the roles of CLV1 and LRR-RLKs in the perception of root-derived N-responsive CLEs and the identity of subsequent systemic signals are still not well defined (Araya et al. 2014a). Thus, the mobility of N-responsive CLEs in Arabidopsis needs further testing if these CLEs are functionally equivalent to the NICs from legumes. Interaction between CLV1 and LRR-RLKs and cell-type specific expression of these receptors may be checkpoints allowing to distinguish the effect of CLE signals on specific downstream pathways. With regard to signals modulating CLE expression, transcriptome data suggest that CLE3 expression can be enhanced in Arabidopsis roots by a systemic nitrate demand and by local ammonium supply (Scheible et al. 2004; Patterson et al. 2010; Ruffel et al. 2011). The actual contribution of the CLE3-CLV1 regulatory module in lateral root phenotypes may require a more detailed analysis, because the systemic nitrate demand and ammonium signals have strong but distinct effects on root morphology, such that an increased demand for nitrate generally enhances lateral root elongation (Ruffel et al. 2011) while ammonium stimulates lateral root branching (Lima et al. 2010). Statistical modeling approaches may distinguish the effect of nitrate, ammonium and CLE peptides on root growth at specific checkpoints in root development (e.g. lateral root initiation, emergence or elongation), as it has been successfully employed to quantitatively characterize lateral root development and branching phenotypes in Arabidopsis (Araya et al. 2016).

Redundancies among the CLE peptides and the LRR-RLKs allow numerous peptide-receptor combinations, while a subset of signaling modules are formed to have specialized functions in specific cell types and conditions. The cell-type specificities of these functionally relevant peptide-receptor signaling modules primarily depend on a locally restricted transcriptional regulation of the receptor and the peptide-encoding genes. Modification of CLE peptides (e.g. by glycosylation) and protein–protein interactions of LRR-RLKs may also contribute to determining the specificities of peptide-receptor signaling pathways. Cell-to-cell movement of CLEs and related mobile signals through plasmodesmata would allow these signaling modules to be integrated with cell-specific signaling pathways controlling the stem cell fate and differentiation. Long-distance translocation of these signals through vascular connections may control the growth behavior of organs that are distantly located from domains where intrinsic cues or environmental signals can be sensed. Plants could have developed such higher dimensional signaling cascades to acclimate to changes in the environment. As highlighted here, the small peptide signals and their mobility and interaction with nutritional signals are emerging research areas to be explored for identifying the molecular mechanisms underlying the developmental plasticity of plant root systems.