Communication between the extracellular matrix and the cell interior is essential for all organisms as intrinsic and extrinsic cues have to be integrated to co-ordinate development, growth, and behaviour. This applies in particular to plants, the growth and shape of which is governed by deposition and remodelling of the cell wall, a rigid, yet dynamic, extracellular network. It is thus generally assumed that cell wall surveillance pathways exist to monitor the state of the wall and, if needed, elicit compensatory responses such as altered expression of cell wall remodelling and biosynthesis genes. Here, I highlight recent advances in the field of cell wall signalling in plants, with emphasis on the role of plasma membrane receptor-like kinase complexes. In addition, possible roles for cell wall-mediated signalling beyond the maintenance of cell wall integrity are discussed.
The plant cell wall
Plant cell walls form a major part of the planetary biomass, control plant growth and morphogenesis, and have recently attracted a lot of attention as a possible alternative to fossil fuels. However, our knowledge about biosynthesis, deposition, and modification of the plant's extracellular matrix (ECM) is still limited. The principles of plant architecture differ fundamentally from those of animals in that individual cells serve as ‘osmotic bricks’ characterized by the equilibrium between the internal turgor pressure and countering mechanical stresses in the surrounding cell wall . As each plant cell is tightly connected to its neighbours through shared walls, plant tissues form a continuum of cells, eliminating cell migration as a means to drive morphogenesis. Cell elongation requires co-ordinated increase in cell wall extensibility and is accomplished by controlled wall loosening, which triggers stress relaxation and the reduction in turgor pressure, and enables the influx of water. This last event, in turn, results in cell expansion, restoration of wall stress, and increase in turgor pressure until a new equilibrium is reached. Thus, plant growth is mechanistically controlled by the properties of the cell wall .
During growth, the characteristic anisotropy of plant cell expansion is believed to be conferred by the alignment of cellulose microfibrils, which restricts growth to the axis perpendicular to the net cellulose orientation. Cellulose is synthetized by hexameric cellulose synthase (CESA) complexes travelling in the plasma membrane propelled by their own activity, which consists of adding glucose residues to the cellulose chains. The trajectory of CESA movement, and thus the deposition of the microfibrils, is guided by the arrangement of cortical microtubules (MTs) underneath the plasma membrane [2,3]. The microfibrils are embedded in a hydrated network formed by the ‘matrix’ polysaccharide classes of pectins and hemicelluloses, which are both synthesized in the Golgi apparatus and reach the cell wall through secretory vesicles. These three main polysaccharide classes (cellulose, hemicelluloses, and pectins) are interconnected through a variety of different linkages, the full extent and function of which are just beginning to emerge . The cell wall also contains structural proteins forming molecular ties with the polysaccharides [5–8].
In recent years, the model of plant cell wall architecture had to be revised, as genetic studies and novel approaches in the biophysical analysis of cell wall properties have challenged in particular the role of hemicelluloses in the primary, growing cell wall . In addition, it became apparent that pectin distribution and modification is much more relevant to the mechanical properties of cell walls and plant development than previously thought. Homogalacturonan (HG), a polymer of galacturonic acid and the major type of pectin, is synthesized in the Golgi apparatus in a highly methyl esterified state. After delivery to the extracellular space, HG is de-methyl esterified by pectin methylesterase (PME), yielding pectate with a free carboxylic acid group, methanol, and a proton [10,11]. PME activity, in turn, is affected by interaction with PME inhibitor proteins . Depending on the pattern of carboxylic acid and methylester groups created by PME, the mechanical properties of the wall are altered, the mechanism of which is not well understood [10,13,14]. Importantly, the position of almost all plant cells is fixed relative to neighbouring cells by their shared middle lamella, which is particularly rich in pectin. This suggests that directional growth has to be co-ordinated between cells through cell-to-cell communication, and a role for pectin de-methyl esterification in this process can be expected.
In light of these and other recent results, a new model of the cell wall has been proposed that postulates only a limited number of load-bearing contact sites between hemicelluloses and cellulose, which are the likely targets for cell wall loosening agents such as expansins [4,15]. In contrast with earlier views, in this model, pectins are believed to engage in a variety of linkages with the other polymers (as well as with proteins ), although their relative contribution as a load-bearing component is still unclear [4,16].
In addition to their role in growth control, cell wall features are frequently a hallmark of cell differentiation. Cell wall composition and, by extrapolation, biophysical properties vary greatly between individual cell types as demonstrated, for example, through the use of cell wall epitope-specific antibodies on tissue sections . The most fundamental difference in composition and functionality, however, is observed between primary cell walls, defined as walls surrounding cells still capable of growing, and secondary cell walls, which can fortify cells such as xylem elements after cessation of growth. While both primary and secondary walls contain pectin, cellulose, and hemicelluloses, the relative abundance and type of the individual polysaccharide classes can be very different. For example, pectin is the most abundant polysaccharide in most dicot primary walls, but is present only in minor amounts in secondary walls . On the other hand, secondary walls, the acquisition of which is one of the last steps in vascular differentiation before programmed cell death, are typically rich in lignin, a polyphenolic substance coating the cellulose microfibrils, which leads to increased stiffness and resistance against negative pressure in the water-conducting xylem elements . In essence, the lignified secondary cell wall of xylem cells is therefore the feature that allows plants to grow to the impressive heights observed in trees, and forms a major portion of the planetary biomass .
ECM signalling across the tree of life
As briefly outlined above, it is generally accepted that cell expansion in plants is mediated by regulated and selective loosening of the cell walls, which comprises breaking of load-bearing bonds and subsequent displacement of cell wall polymers. After cessation of growth, wall rigidity is consolidated and cell wall extensibility is reduced through the formation of new load-bearing connections and bonds between the polymers, as well as delivery of additional cell wall material [1,21]. Therefore, cell wall integrity is challenged by growth itself and tight control must be exerted on this process to prevent cell swelling or bursting. To ensure cell wall integrity and homoeostasis during growth, as well as to enable cell wall fortifications in response to extrinsic challenges, cell wall surveillance pathways must exist to perceive changes in cell wall state and elicit intracellular responses [22–24]. In addition, new cell walls are inevitably created through cell divisions; thus, adjacent cells within a tissue are tightly linked and intrinsic differences in growth rate between cells will lead to tension and compression, respectively [25,26]. Hence, the expansion of neighbouring cells needs to be co-ordinated to allow for controlled directional tissue growth. Conversely, the cell walls on each face of the cell are the result of separate division events that occurred at separate points in time. Yet, cell walls at opposite sides of the cell have to elongate co-ordinately to allow directional growth. Thus, plant growth requires both mechanical symmetry breaking (between walls that expand and those that do not) and the maintenance of symmetry (between parallel walls). It therefore appears intuitively logical to assume that plant cells gather information about the state of their ECM. However, in contrast with the situation with other models, our knowledge about signalling in response to cues derived from the plant cell wall is relatively scarce . This should probably not come as a surprise, given that we are far from understanding the structure and biochemical complexity of the cell wall itself.
Despite the aforementioned limitations, researchers have tried to identify signalling pathways that connect self-surveillance and intracellular responses, and considerable progress has been made. Traditionally, these pathways have been subsumed under the term ‘cell wall integrity signalling’ , possibly influenced by research in yeast where weakening of the cell wall through external drug application forced single yeast cells to reinforce their wall to prevent bursting. However, focusing on the ‘integrity’ aspect probably falls short of spanning all the possible complexity of ECM-triggered signalling in multicellular plants. In contrast with yeast, individual cells are not only glued to their neighbours through the wall they share, but also have to find ways to co-ordinate and control their characteristically anisotropic growth within a tissue. In addition, and in contrast with single-celled wall-enclosed organisms, structural integrity of most plant cells is expendable due to the vast redundancy and enormous developmental plasticity typical of multicellular plants. On the other hand, cell wall damage is often one of the first signs of pathogen attack, and perception of plant cell wall breakdown products as damage-associated patterns (DAMPs) , recognition of pathogenic cell wall molecules, and triggering of defence responses by pathogen-derived cell wall degrading enzymes are important mechanisms ensuring plant survival. These processes are studied extensively in the field of plant immunity; however, it is not always possible (or helpful) to uphold the distinction between immunity and development-related wall signalling, as considerable cross-talk between growth control, self-perception, and defence signalling has been demonstrated (see below). Arguably, not only yeast, but also animal systems, can provide useful analogies demonstrating the complexity and potential of ECM signalling. For instance, it has been revealed that biochemical signals present in the animal ECM, as well as its mechanical properties, govern cell fate decisions [29,30], and we should strive to assess whether similar fundamental developmental regulation is mediated by wall-related signalling in plants, even though little molecular homology is to be expected.
To account for all this potential complexity, we use the term ‘cell wall signalling’, defined as any signalling event that occurs in response to cues from the ECM. Notably, under this definition, the signals that are perceived are not necessarily confined to cell wall molecules, but could also include derived properties such as mechanical stress and membrane tension. In addition, in analogy to the situation in animal ECM signalling, changes in cell wall properties might reveal previously masked binding epitopes or change mobility and/or availability of cell wall-associated protein ligands such as peptide hormones and growth factors.
Cell wall signalling in plants
Even though we know of only a limited number of wall-related signalling components, there is ample evidence for their activities, as cell wall signalling is presumably at the basis of the astonishing plasticity of plant cell wall composition. Early on, researchers studying impairment of cell wall biosynthesis or modification have noticed (compensatory) changes in components and pathways not directly affected by the mutation under study [31–38]. It can be safely assumed that signalling and cell wall surveillance are required to orchestrate these responses. Especially insightful in that regard have been approaches which genetically or pharmacologically interfered with the production of cellulose, revealing increased pectin biosynthesis and de-methyl esterification, ectopic lignification, enhanced callose deposition, and the induction of hormone pathways, up-regulation of defence-associated transcripts, and production of reactive oxygen species (ROS) . In addition, presumably signalling-mediated compensatory responses have been observed in response to genetic impairment of xyloglucan biosynthesis  and in lignin biosynthesis mutants [40–42]. One conclusion from these observations is that it is not straightforward to interpret mutant phenotypes of cell wall-related genes without having a complete picture of potential compensatory responses. A prime example comes from one of the candidates for a cell wall signalling sensor, THESEUS1 (THE1, described below) . As noted above, interference with cellulose biosynthesis results in many biochemical and metabolic phenotypes but also often in pronounced dwarfism, for example in mutants of CESAs (cesa mutants). Together with the observed and expected reduction in cellulose levels in those mutants, the dwarfisms could easily be interpreted as evidence for the growth-limiting effect of the reduced cellulose. However, loss of function of THE1 results in a striking recovery of the growth phenotype without rescuing cellulose production, suggesting that the reduction in growth of cesa mutants is a secondary effect, a syndrome induced by a compensatory response . Owing to the remarkable plasticity of cell wall composition and the fact that cell wall properties can be viewed as being both ‘upstream’ and ‘downstream’ of cell wall signalling pathways if those result in a response that maintains or alters wall properties, it is technically challenging to identify signalling components under physiological conditions. Instead, the cell wall often has to be severely altered to drive robust homoeostasis mechanisms out of balance and reveal phenotypes that can be scored and used to identify pathway components through genetic screens [43,45,46]. While this can be regarded as artificial, it allows the discovery of the pathway as such, which then has to be followed up upon with deciphering its physiological role. Again, the yeast cell wall integrity signalling field offers a useful model demonstrating the feasibility of this approach .
Cell wall signalling can function as a means to maintain cell wall integrity in the face of extrinsic challenges, such as changing osmolarity and physical damage [28,48,49]. In addition, growth itself can be viewed as a challenge to cell wall integrity and stress relaxation and polymer creep need to occur in a controlled manner to maintain the primary cell wall's unique ability to reconcile strength with extensibility . However, other potential roles of cell wall signalling are conceivable, such as the maintenance of cell wall homoeostasis, control of cell differentiation, and cell identity patterning. An exciting recent example that builds on earlier studies [51,52] is provided by the observation that oligosaccharides produced by an endo-β-mannanase suppressed cell wall thickening in the xylem of Populus, controlling the switch from primary to secondary cell wall deposition . Thus, cell wall-derived signals affect stress responses, but can also regulate developmental processes.
Irrespective of their function, all cell wall signalling pathways share the requirement that information has to be transmitted from the outside across the plasma membrane to the inside of the cell, where a response can be orchestrated. Numerous possibilities to transduce signals are conceivable (Figure 1), such as gated ion channels, arabinogalactan proteins (AGPs), proteins that connect the cell wall and the cytoskeleton, mechanosensors analogous to yeast mating pheromone-induced death 1, and osmosensors [6,54–58]. For instance, it has been demonstrated recently that MECHANOSENSITIVE CHANNEL OF SMALL CONDUCTANCE-LIKE 8, a protein from Arabidopsis thaliana (hereafter referred to as Arabidopsis) homologous to the mechanosensitive channel of small conductance from Escherichia coli, senses and responds to changes in osmotic potential in the pollen . However, many of the known cell wall signalling components seem to operate in pathways featuring a member of the large family of receptor-like kinases (RLKs) to transmit signals from the outside to the inside of the cell. Plant RLKs form expanded gene families with >600 and 1000 members in Arabidopsis and rice, respectively [60–62], the majority of which are, to date, orphan receptors without a known ligand or downstream intracellular targets. Generally, RLKs contain an N-terminal signal peptide, a variable extracellular domain (ECD), a single-pass transmembrane domain, and a cytosolic protein kinase domain that is related to animal Pelle/IRAK-4 kinases [60,61]. The kinase domains serve as a platform for interaction with regulatory receptor-like kinases, inhibitory proteins, and accessory factors, as well as downstream signalling components that are phosphorylated after ligand perception. Thus, signalling strength is, in part, determined by the composition of the heteromeric receptor complex .
Overview over selected potential cell wall-related signalling receptors and signals.
The largest subgroup among RLKs is formed by proteins with a leucine-rich repeat extracellular domain (LRR-RLKs), and individual members of this group have essential roles in many aspects of plant development and immunity [64–70]. Many ligand receptor pairs have been studied in detail combining genetic, biochemical, and structural analyses. These LRR-RLKs generally seem to form heterodimers with members of the SOMATIC EMBRYOGENESIS RECEPTOR KINASEs (SERKs) family, facilitated by binding of the ligand to the ECDs [63,71,72]. This, in turn, juxtaposes the two kinase domains that are then activated through auto- and trans-phosphorylation at multiple residues. Recent evidence suggests that plant LRR-RLKs are not restricted to the phosphorylation of serine and threonine residues, but are dual-specificity kinases that can act, in addition, on tyrosine residues [73–77]. An important aspect of RLK signalling is thus to decipher how receptor phosphorylation at multiple sites affects signalling dynamics and specificity . Once activated, heteromeric RLK complexes recruit and phosphorylate immediate downstream targets, which, as a common theme for many pathways, often belong to the large family of receptor-like cytoplasmic kinases (RLCKs) [63,79]. RLCKs initiate a signal transduction cascade which in some cases involves MITOGEN-ACTIVATED PROTEIN (MAP) kinases and/or calcium-dependent protein kinases, and eventually results in altered activity of transcription factors and thus ligand-responsive gene expression changes. However, only for very few RLK pathways, such as brassinosteroid (BR) signalling (see below), a comprehensive picture of potential signalling components and transcriptional outputs has emerged. In addition to these transcriptional outputs, the activity of RLKs can produce post-translational cellular responses such as the generation of ROS or alter transport process and ion fluxes across membranes, adding another level of signalling complexity and dynamics (Figure 1) [63,80].
Possible control mechanisms to fine-tune signalling strength and temporally restrict the activity of RLK pathways have been observed at all steps of the signalling cascade. At the level of the receptor complex, these can include reversible association of inhibitory proteins with the kinase domain [76,81,82], the presence of pseudokinases in the receptor complex [83,84], the activity of phosphatases to remove activating or inhibitory phosphorylation from the kinase domain [85–88], modulation of receptor abundance and/or signalling capacity through ubiquitination-mediated endocytosis and entry into vacuolar degradation pathways [89–91], and modulation of receptor activity through other post-translational modifications such as glycosylation, acetylation, dithiol–disulfide interconversion, and S-thiolation [92,93].
The various RLK subgroups differ most notably in their extracellular domains that, in some instances, contain putative carbohydrate-binding motifs, rendering them interesting candidates for sensing cell wall-related signals (Figure 1). However, compared with LRR-RLKs (see above), less is known regarding the intracellular signalling following activation of these receptors.
Given the privileged position of RLKs to potentially sense and transduce cell wall-related signals, the following section will concentrate on highlighting the role of various classes of receptor-like kinases in cell wall sensing and signalling.
Catharanthus roseus RLK1-like kinases
Members of the Catharanthus roseus receptor-like kinase1-like (CrRLK1L) subfamily have received by far the most attention of any group of receptor-like kinases in the context of cell wall signalling [94–97]. In part, this is based on the existence of putative carbohydrate-binding domains within their cell wall-facing ectodomain. However, there is also convincing genetic evidence linking CrRLK1L receptors to cell wall homoeostasis, mechanoperception, cell wall integrity maintenance, and growth control [94,97–100]. The CrRLK1L gene family comprises 17 members in Arabidopsis, to seven of which a function has been ascribed to . The general domain structure consists of a signal peptide, a divergent ectodomain of ∼400 amino acids, a juxtamembrane region required for binding of the co-receptors LORELEI/LORELEI-LIKE-GPI-ANCHORED PROTEIN 1, a transmembrane domain, and a highly conserved cytoplasmic serine/threonine kinase domain with a more divergent C-terminal tail of unknown function. Within the ectodomain, one or two regions can be found that share limited homology with malectin , a protein originally described in Xenopus laevis  that is involved in ER quality control . Interestingly, malectin is capable of interacting with di-glucose motifs of N-linked oligoglycans and is structurally similar to carbohydrate-binding modules of glycosyl transferases , prompting speculation that the homologous CrRLK1L domain might mediate cell wall binding in plants . However, this remains to be experimentally verified by plasmolysis of plant cells expressing the ECD of CrRLK1Ls or by binding assays using glycan arrays  or similar approaches. Notably, the presence of malectin-like domains is not unique to CrRLK1Ls in plants; among others, members of the LRR I family of RLKs contain malectin-like domains in the N-terminal part of the ectodomain followed by a cleavage motif and the LRRs [44,104]. In Lotus SYMBIOSIS RECEPTOR-LIKE KINASE, the malectin-like domain prevents the interaction with other signalling components before being shed . It is thus not yet known whether malectin domains in plants, which share only limited homology with animal malectin, indeed interact with glycans or have acquired a different function.
THE1 was identified in a suppressor screen of the CESA mutant cesa6prc1-1 . THE1 loss-of-function alleles were capable of rescuing the dwarf phenotype and ectopic lignification of many cesa mutants, notably without alleviating cellulose deficiency. This demonstrates that the reduced cellulose content is not growth-limiting in those mutants, and that THE1 orchestrates a response to changes in the wall, which comprises, among others, oxidative burst mediated by the NADPH oxidase respiratory burst oxidase homologue (RBOH) D and a negative effect on growth . In line with this, a sizeable portion of transcripts differentially regulated in cesa6prc1-1, were THE1-dependent. Among THE1-dependent transcripts, those associated with ROS detoxification, defence, glucosinolate biosynthesis, and cell wall cross-linking were especially prominent . This could indicate that the defective cell wall in cesa6prc1-1 is interpreted as being indicative of a pathogen attack or a general danger signal. Notably, the1 mutants do not show obvious phenotypes in the absence of cellulose biosynthesis inhibition . Thus, it remains to be demonstrated in which developmental or stress-related processes the THE1 pathway is involved. It will be interesting to see whether THE1 interacts with wall-derived carbohydrates through its malectin domains and whether it binds additional ligands as demonstrated for the best-characterized CrRLK1L, FERONIA (FER).
FER was originally described as essential for fertilization, as loss of FER in the female gametophyte leads to the accumulation of supernumerary pollen tubes that failed to discharge [106,107]. FER localizes to the filiform apparatus of the two synergid cells that form the entry point in the female gametophyte for the pollen tube . fer mutants also show defective root hair growth, and various mutant alleles show a confusing diversity of growth-related phenotypes, some of which can be expected to be an indirect consequence of the response to the primary defect [94,95,97,109]. In addition, pleiotropic phenotypes can presumably result from the fact that FER intersects with several hormone pathways such as abscisic acid [88,110], auxin , ethylene [111,112], and BR signalling [112,113], as well as defence [114,115] and mechanical signalling .
Recently, the FER ectodomain was shown to bind to the secreted peptide RAPID ALKALINIZATION FACTOR 1 (RALF1) . Application of RALF1 leads to an increase in apoplastic pH, a transient increase in cytoplasmic calcium, a concomitant cessation of growth, and the differential phosphorylation of many proteins, including FER itself [88,117–119]. Insensitivity of fer mutants towards RALF1, together with binding of the peptide to the ECD, strongly suggests that FER is the receptor for RALF1 , and, potentially, for other RALF-like peptides including those of pathogens that might target FER to suppress immune responses . This finding is somewhat at odds with the proposed carbohydrate-binding function of the malectin-like domain in FER and other CrRLK1Ls. While, in theory, glycan decorations of protein ligands would be consistent with a carbohydrate-binding activity of the malectin domain, RALF1 does not show a predicted glycosylation site in the mature, processed part. In addition, RALF1 expressed in E. coli, and thus devoid of glycosylation, was capable of interacting with FER . However, it can, of course, not be excluded that RALF1 binds to another region of the relatively large FER ectodomain or that malectin-like domains adopted another function in plants. Notably, peptide signalling-mediated regulation of apoplastic pH is yet another way in which FER is connected to the control of growth through cell wall properties, and it has been proposed that this signalling module is part of a feedback loop that connects cell wall modification through PMEs, pH oscillations, and growth . In accordance with this, the absence of FER leads to disturbed growth co-ordination with random patterns of cell expansion in the root tip .
Interestingly, FER is also required for parts of the response towards mechanical stress. Compressive strain in root epidermis cells leads to biphasic calcium transients mirrored by extracellular alkalinisation. The second peak of this biphasic response to mechanical load is clearly FER-dependent, in line with the reduced ability of fer mutants to adapt their growth to mechanical impediments. At present, it is unclear whether FER-mediated mechano-responses depend on the availability/release of RALF peptides, which could be affected by mechanical changes to the cell wall, reminiscent of the complex interactions between ligands of cell surface receptors and the animal ECM. In addition, one potential caveat is that fer mutants have altered cell wall composition, and thus altered mechanical properties , which could affect the response of those mutants to mechanical triggers. Nevertheless, the data are consistent with FER being required to sense both extrinsic and intrinsic mechanical cues to co-ordinate growth responses [116,122] and thus strongly suggest that mechanosensing is required for supracellular growth co-ordination. Interestingly, the erratic expansion pattern observed in fer mutant roots, and potentially also in anx1 anx2 mutant pollen (see below), might point towards a mechanical feedback loop co-ordinating morphogenesis, not unlike what has been proposed for the shoot apical meristem [26,123].
Contrasting with FER's growth-co-ordinating role, its first described function, pollen tube discharge, is a controlled loss of cell wall integrity. Recent evidence points towards an essential role of FER-dependent ROS and complex calcium dynamics, respectively, in both synergids and pollen tube [124–128], suggesting that FER mediates communication between the male and female gametophytes. Thus, calcium transients mediated by FER have been observed during fertilization, root hair growth, RALF1 response, and mechanosensing, suggesting that at least some downstream signalling events are shared between FER's many different functions (see below). Another common theme is the essential role of the GPI-anchored protein LORELEI or it's close paralogue LORELEI-LIKE-GPI-ANCHORED PROTEIN 1 in FER-mediated signalling, which is required for correct FER targeting and can be pulled down in complex with FER by immobilized RALF1 .
The closest homologues of FER, ANXUR1, and ANXUR2 (ANX1, 2) are primarily expressed in the male gametophyte and are essential for pollen tube cell wall integrity [130,131]. Interference with both ANX1 and ANX2 resulted in strongly reduced fertility and premature pollen tube burst, whereas overexpression of ANX1 or 2 inhibited pollen tube growth and led to thick, pectate-rich cell wall accumulation and plasma membrane invagination. Enhanced cell wall thickening preceded membrane invagination, which would be consistent either with oversecretion of cell wall material, with cell wall properties that do not allow elongation growth, or with a combination thereof . As previously described for FER and THE1, ANX1, 2 seem to promote the generation of ROS, in this case through the pollen-expressed NADPH oxidases RBOHH and RBOHJ. NADPH oxidase-mediated ROS, in turn, are required to maintain functional calcium dynamics during pollen tube growth. Interestingly, a marked increase in cytoplasmic calcium preceded the premature pollen tube burst in anx1 anx2 , similar to the calcium spike induced by exogenous ROS before rupture . Thus, both elevated and reduced levels of ROS lead to pollen tube burst, and in both cases, the underlying mechanism remains to be determined. Potentially, ROS could act as signalling molecules, cell wall-modifying agents, or both. Recently, screening for fer-like pollen tube overgrowth phenotypes revealed that TURAN, a uridine diphosphate-glycosyltransferase involved in N-linked protein glycosylation, is required for export of ANXUR1 from the ER, suggesting that ANX1 is subject to glycosylation-dependent quality control. Accordingly, turan loss of function showed a pollen tube phenotype reminiscent of anx1 anx2 .
Signalling downstream from CrRLK1L
Regarding the signalling downstream from CrRLK1Ls, the available data for FER, ANX1,2, and THE1 seem to suggest that similar immediate downstream signalling components such as ROP (Rho of plants)-GEF1 (guanidine nucleotide exchange factor), ROP GTPases, and NADPH oxidases are used . More evidence for a common downstream pathway has been provided by domain swap experiments. Whereas the FER extracellular domain cannot be replaced by that of its closest relative ANX1, the cytoplasmic kinase domain is interchangeable between FER, ANX1, and HERCULES1 . This strongly suggests that different ligands are recognized by the various CrRLK1L ECDs, but ligand binding then triggers downstream signalling pathways that are at least partially shared . Corroborating this notion, FER, ANX1/2, and THE1 all promote activation of NADPH oxidases [48,109,132]. Interestingly, constructs that introduced mutations in the catalytic site and activation loop of the FER kinase domain were still capable of rescuing the fer phenotype in the female gametophyte, suggesting that kinase activity could be dispensable for FER function in this context . It will be interesting to see whether kinase activity is dispensable for other functions of FER and, if so, whether the cytoplasmic domain rather acts as a scaffold for protein complex assembly, possibly recruiting active kinases. Notably, RALF1 treatment induced the phosphorylation of the C-terminal tail of FER that was hypothesized to be the result of autophosphorylation, consistent with what is observed in other RLKs . Interrogating the phosphorylation status of the presumed kinase-dead FER version after RALF1 treatment would thus be an interesting first step to answer the aforementioned questions. If FER phosphorylation still occurs in this mutant, transphosphorylation would have to be assumed, possibly by a so-far unknown co-receptor. In the well-characterized LRR-RLK BRASSINOSTEROID-INSENSITIVE 1 (BRI1), autophosphorylated residues, including those at the C-terminal tail region, largely overlap with those transphosphorylated by the co-receptor BRI1-ASSOCIATED KINASE 1/SOMATIC EMBRYOGENESIS RECEPTOR KINASE 3 (BAK1/SERK3) [135,136]. It should be noted that most, if not all, RLKs studied to date form heterodimers with co-receptors, facilitating transphosphorylation and activation . Consistent with this, the kinase domain of the LRR-RLK HAESA and that of its co-receptor SERK1 were both capable of using the kinase-dead version of the respective partner as a substrate . In many cases, RLKs acting as a co-receptor are redundant and act in multiple pathways, and could thus be missed in screens such as those leading to the identification of fer and other mutants with similar, fer-like phenotypes [106,107,133,138]. Notably, FER has been recently identified in a complex that included several other RLKs and cytoplasmic receptor-like kinases . In light of the exchangeability of the FER kinase domain, an obvious question is whether THE1 also activates signalling by RAC/ROP GTPases through interaction with ROPGEFs.
Recently, another signalling component downstream from CrRLK1Ls was identified in a screen for suppressors of the almost infertile anx1anx2 mutant. A hypermorphic allele of MARIS (MRI), an RLCK, was capable of rescuing fertility and pollen germination of anx1 anx2 plants, while it had detrimental effects in the wild-type background . Conversely, mri loss-of-function mutants show strongly reduced transmission through the male gametophyte and displayed pollen tube bursting, suggesting that MRI is essential for pollen tube function. Genetic interaction studies showed that MRI acts downstream from NADPH oxidases and requires the activity of the latter to be active, as expression of a gain of function but not of the wild-type version could rescue fertility of the rbohH rbohJ mutant. Interestingly, mri loss-of-function mutants showed root hair defect reminiscent of fer, and the MRI gain-of-function mutant under control of its own promoter was capable of rescuing the fer defects in root hairs, where MRI is also expressed . This suggests that MRI acts downstream from FER in root hairs and, by extrapolation, that RLCKs might be a general component of CrRLK1L-mediated signalling, in line with the fact that at least FER, ANX1/2, and THE1 all regulate NADPH oxidases. Recently, this hypothesis was corroborated through the identification of a member of the RLCK-VII subfamily that is genetically required for the response to RALF1 and is recruited to the FER complex in response to the peptide, triggering mutual phosphorylation of FER and the RLCK .
In pollen tubes, calcium and ROS dynamics are intimately linked to the control of growth rate and the deposition of cell wall material at the tip [132,142–144]. Whereas cell wall (mainly pectate) deposition precedes the growth burst, calcium increase in the cytosol happens concomitantly with the peak growth rate . Interestingly, rbohH rbohJ double mutants show increased growth oscillations compared with the more regularly elongating wild-type tubes. Taken together, these results suggest that calcium-mediated regulation of NADPH oxidase-derived ROS is required for balancing tube elongation and exocytosis. Notably, ANX1/2-mediated maintenance of pollen tube integrity and FER-mediated pollen tube burst act through controlling calcium and ROS levels, consistent with CrRLK1 signalling as the regulatory element in a feedback loop that involves the cell wall to control growth (see above). What remains to be shown is whether this feedback loop also involves cell wall sensing, either directly or indirectly, by CrRLK1Ls. Moreover, from a wall-centered view, it will be interesting to see whether wall modification, possibly downstream from CrRLK1L signalling, is involved in mediating pollen tube bursting in addition to ion flux-mediated changes in osmolarity [145,146] and the potential direct role of ROS in affecting wall mechanics [147–150]. Here, pectin modification is a prime candidate as (i) pectin is the prevalent component in pollen tube walls, (ii) interference with pectin modification compromises pollen tub integrity, and (iii) modification of existing wall components, rather than synthesis, seems a conceivable regulatory step in the hectic life (and death) of a pollen tube [151–156].
Members of the family of wall-associated kinases (WAKs) have long been discussed as potential cell wall receptors. Although redundancy and clustered orientation of the five WAK genes in a 30 kb chromosomal region have rendered the family genetically intractable until the advent of genome editing, antisense approaches and dominant-negative alleles have provided evidence for an involvement of WAKs in development [157–159]. In addition, the extracellular domains of WAKs, which contain epidermal growth factor (EGF)-like repeats, tightly bind to pectate. Concerning the mechanism of wall interaction, it seems that the N-terminal part of the WAK extracellular domain can bind to de-esterified HG and polygalacturonic acid, whereas the EGF-like domains do not seem to be involved in binding [160–162]. Thus, WAKs are the only receptor class implicated in cell wall signalling for which binding to wall components has been demonstrated. The WAK extracellular domain showed a preference for cross-linked configuration of pectate mediated by calcium ions (the so-called egg box pectin). Interaction also occurs with shorter HG fragments and oligomers of galacturonic acid, but not with monomeric galacturonic acid, and is strongly reduced with the other pectin classes rhamnogalacturonan I and rhamnogalacturonan II, and by HG methyl esterification . Interestingly, interaction with polygalacturonic acid seems to be charge-based mediated by patches of hydrophobic amino acids, similar to what was observed with pectate binding of the extracellular LRR protein POLYGALACTURONASE-INHIBITING PROTEIN [160,163], and suggesting that this could be a general mechanism of interaction with the negatively charged groups of pectate. Notably, this is mechanistically different from the interaction between sugars and carbohydrate-binding modules, which is often mediated by hydrophobic stacking between aromatic residues and the sugar ring [164,165].
It was suggested that WAKs are receptors for pectin and/or oligogalacturonides (OGs) [162,166], small breakdown products of pectin that are generated during pathogen attack, but potentially also play a role during development by antagonizing auxin functions [167–174]. The plant's response to the application or in vivo generation of OGs is overlapping with, but distinct from the response to pathogen-associated molecular patterns and genetically requires a MAP kinase phosphorylation cascade and the activity of calcium-dependent protein kinases, among others [175–177]. In addition, exposure to OGs leads to rapid changes in protein phosphorylation patterns . Implication of WAK1 as an OG receptor was based on protein chimeras in which the intracellular serine/threonine kinase domain of WAK1 was exchanged for the kinase domain EF-TU RECEPTOR . These chimeras produced an EF-TU RECEPTOR signalling-like response after application of OGs. In addition, WAK1 overexpression showed enhanced OG responses such as ROS production and callose deposition, as well as increased resistance towards pathogens [166,178], whereas GLYCINE-RICH PROTEIN-3 AND KINASE-ASSOCIATED PROTEIN PHOSPHATASE, two proteins that interact with WAK1, have a negative effect on those outputs. The use of gene-editing tools should now allow assessment of whether WAK1 loss-of-function mutants are less sensitive to OGs and more susceptible to pathogen attack. Until then, we are left with the puzzling conundrum posed by the fact that WAKs could be receptors for pectate in a cell wall signalling role, OG receptors involved in the response to pathogens, or both . Remarkably, short pectate fragments can efficiently outcompete longer pectins from binding of the WAK1 extracellular domain, and there is evidence for a competition between pectate and OGs in vivo . An interesting question is thus whether newly synthesized pectin can compete with WAK1-bound pectate after de-methyl esterification, which could be a way to create dynamic signals to be read by WAKs, as opposed to stationary anchoring of the protein to the walls.
An exciting new study uncovered an interesting twist related to cell wall signalling and cell adhesion. Mutants of QUASIMODO1/GAUT8 and QUASIMODO2/TUMOROUS SHOOT DEVELOPMENT 2, a putative GT8 family galacturonosyltransferase and putative pectin methyltransferase, respectively, show a 50% reduction in HG content and a clear cell adhesion defect with frequent detachment of cells [181–184]. Unsurprisingly, the reduction in HG, the prevalent polysaccharide in the middle lamella, was assumed to be causal for the loss of cell adhesion in those mutants. However, a genetic suppressor screen revealed that mutations in ESMERALDA1 (ESMD1) could rescue cell adhesion in both mutants without affecting HG content, suggesting a signalling-related cause for the phenotype in QUASIMODO mutants . In line with this, altered gene expression in quasimodo2-1 was also reversed in quasimodo2 esmd1 double mutants. Interestingly, ESMD1 encodes a putative protein O-fucosyltransferase predicted to act on proteins with EGF-like repeats, similar to the gene affected in the previously identified friable1 (frb1) mutant, which also showed cell adhesion defects . Genetically, ESM1 and FRB1 act antagonistically in the same pathway that is affected by mutations in QUASIMODO 1 and 2 . Taken together, these results suggest that altered pectin in the qua mutants is perceived by a pectin-responsive signalling pathway, the activation of which leads to changes in gene expression and cell wall properties and, ultimately, to the cell detachment phenotype. ESMD1 and FRB1 affect the activity of this pathway positively and negatively, respectively, perhaps by adding fucose residues to the same EGF-like repeat-containing protein. While this protein remains to be identified, it is tempting to speculate that one or several of the WAKs could be the target(s) for ESMD1 and FRB1, based on the presence of EGF-like repeats in the WAK ECD and the well-documented role of WAKs in pectin-related signalling.
Co-option of the BR pathway for cell wall homoeostasis
The BR signalling module is a central regulator of plant growth and one of the best-characterized signalling pathways in plants, with known components for each step of the likely signal transduction cascade from ligands and plasma membrane receptors to transcription factors mediating BR-dependent gene expression [187,188]. In brief, binding of the BR ligand to the ECD of BRI1 creates a docking platform for the shape-complementary ECD of SERK co-receptors such as BAK1 [189,190]. Interaction of the ECDs brings the two intracellular kinase domains in close proximity once the inhibitory protein BRI1 KINASE INHIBITOR 1 is phosphorylated by BRI1 and dissociates [76,81,191]. Activation of the heteromeric receptor complex is achieved by sequential trans- and auto-phosphorylation of the kinase domains in various regions, including the activation loop [73,192–195], whereas other autophosphorylation events have an inhibitory effect and restrict signalling [73,193]. To initiate signal transduction, the activated BR receptor complex phosphorylates and activates RLCKs such as BR SIGNALLING KINASES and CONSTITUTIVE DIFFERENTIAL GROWTH 1, the activity of which leads to activation of the phosphatase BRI1 SUPPRESSOR 1 (BSU1) [196–200]. BSU1, in turn, removes phosphorylation from the negative regulator BRASSINOSTEROID-INSENSITIVE 2 (BIN2), a cytoplasmic kinase related to animal glycogen synthase kinase (GSK) 3. As a result of BSU1-mediated inactivation of BIN2, the equilibrium between the phosphorylated and unphosphorylated forms of the transcription factors and BIN2 targets BRASSINAZOLE-RESISTANT 1 (BZR1), and BRI1-EMS-SUPPRESSOR 1 (BES1) is shifted towards the unphosphorylated form through the help of PROTEIN PHOSPHATASE 2A (PP2A). Consequently, the unphosphorylated forms of the transcription factors are no longer sequestered by 14-3-3 proteins in the cytosol and are free to migrate to the nucleus and mediate BR-responsive transcriptional changes (Figure 2). Among BR target genes, cell wall-related genes are overrepresented [201,202], in line with the fundamental role of BR signalling in cell elongation and growth control. Recently, it was demonstrated that cell wall modification is not only a downstream consequence of BR signalling, but can also modulate BR signalling strength to mediate cell wall homoeostasis . Plants challenged with a reduction in PME activity through overexpression of a PME inhibitor protein (PMEIox) react with a compensatory response that maintains cell integrity, but causes a wide range of secondary phenotypes through the activation of BR signalling. This can be partially suppressed by mutations in the BR receptor BRI1, suggesting that information about the state of the cell wall must somehow be conveyed to BR signalling. Cell wall-mediated activation of BR signalling depends on the LRR receptor-like protein (RLP)44, which is capable of interacting with the BR co-receptor BAK1. Consistent with a role of the RLP44-mediated feedback signalling in development, loss of RLP44 function results in impaired growth and stress responses . Interestingly, RLP44 alone is sufficient to boost BR signalling, very similar to overexpression of BRI1 , but without altering sensitivity to the hormone ligand . Conversely, loss of RLP44 does not affect the response to altered levels of BRs either, suggesting that this receptor is not part of the BR signalling pathway itself. Taken together, these results suggest that RLP44 mediates the integration of cell wall and BR signalling by interacting with the BR receptor complex (Figure 2). Given that the role of RLP44 was discovered in pectate-limited conditions  and that BRs promote both pectin biosynthesis and PME activity , it seems possible that RLP44 is at the centre of a feedback mechanism that fuels BR signalling in conditions during which pectate becomes restricted, such as elongation growth or ion-mediated interference with pectate configuration [50,205,206]. Alternatively, the RLP44–BR module might be involved in fine-tuning of signalling strength in other realms of BR function such as the control of differentiation, cell proliferation, and the integration of growth with external cues [207–214], potentially integrating feedback information from cell wall surveillance. Extensive co-option of BR signalling components by other pathways has been observed before, but the cross-talk ‘hotspots’ are located at the level of the GSK3 kinase BIN2 and the BZR1/BES1 transcription factors [187,207,215–217]. RLP44-mediated signalling integration, on the other hand, seems to occur at the level of the receptor complex. RLPs in Arabidopsis form a gene family of 57 members, some of which have been shown to act in defence signalling, whereas others are well-known regulators of development such as CLAVATA 2 (RLP10) and TOO MANY MOUTHS (RLP17) [218,219]. Members of this quite heterogeneous gene family are characterized by an LRR ectodomain of variable length, a single pass transmembrane domain, and a short cytoplasmic tail, which requires RLPs to interact with proteins that contain cytoplasmic kinase domains such as CORYNE in the case of CLAVATA 2 [220,221] or the RLK SUPPRESSOR OF BIR1 1 in the case of several other RLPs [218,222]. RLP44 interacts with the BR co-receptor BAK1/SERK3 and is phosphorylated in a cell wall-responsive manner, but the role of this phosphorylation for RLP44 function, as well as whether it is mediated by the BR receptor complex or other kinases, remains to be demonstrated. In addition, it remains to be shown whether RLP44 or components of the same pathway senses cell wall status and, if so, by which mechanism. One attractive, testable hypothesis is that RLP44 itself might associate with a pectin epitope, as shown for the LRR protein POLYGALACTURONASE-INHIBITING PROTEIN . In this very simple, speculative scenario, RLP44 would be sequestered by pectate when the latter is available in sufficient quantities near the plasma membrane and released to activate BR signalling when pectate levels or biosynthesis rates drop, and thus ensure pectate homoeostasis.
Model of signalling integration between RLP44-mediated cell wall and BR signalling.
Schengen: controlling border walls?
A process that could fall under the broad definition of cell wall signalling spelled out above is a phenomenon observed in mutants defective in Casparian strip (CS) formation . The CS is a specialized cell wall modification of the root endodermis that ensures selectivity of uptake into the root stele, which in many aspects resembles polarized epithelia of animals. In addition, the CS is thought to be required for the build-up of root pressure to enable vascular transport . Functionality of the CS as a supracellular apoplastic diffusion barrier depends on oxidative cross-linking of lignin monomers through the concerted action of NADPH oxidase RBOHF and the PEROXIDASE 64 , and also requires the presence of the extracellular dirigent domain-containing protein ENHANCED SUBERIN 1 . Spatial precision of lignin formation is dependent on small transmembrane proteins called CASPARIAN STRIP MEMBRANE DOMAIN PROTEINS (CASPs) that define a continuous specialized membrane domain underlying the median section of anticlinal endodermal walls called the Casparian strip domain (CSD) . In differentiating endodermal cells, CASPs recruit RBOHF and PEROXIDASE 64 to this domain, resulting in extremely localized generation of ROS and lignin polymerization, respectively. Later in development, endodermal cells are covered by suberin (cork-like) lamellae . In mutants lacking a functional CS, cell wall material is ectopically deposited to eventually constitute a functional diffusion barrier, despite the absence of a continuous CSD in the plasma membrane. Moreover, suberin biosynthesis is enhanced and occurs earlier in these mutants, indicating a response compensating for the initial defect . Interestingly, an LRR-RLK, SCHENGEN3 (SGN3), is required for both continuous CS formation and the presumed compensatory response to a non-functional diffusion barrier, as both ectopic cell wall deposition and enhanced suberin production were absent from CSD mutants crossed with sgn3. In line with a lack of signalling-mediated compensation, sgn3 mutants are sensitive to suboptimal growth conditions . The CSD and its components, such as CASP1, are present in sgn3, but fail to form a continuous domain, suggesting that SGN3 does not control expression of the involved genes. Although speculative at this point, these results would allow for the hypothesis that SGN3 is directly or indirectly involved in sensing and controlling the CS assembly, consistent with its localization flanking the CSD. Remarkably, mutants of the transcription factor MYB36, a positive regulator of CASPs, ENHANCED SUBERIN 1, and PER64, show a defective CS pattern in addition to an enhanced deposition of suberin as observed in enhanced suberin 1 and casp1 casp3 [229,230]. SGN3 is up-regulated in myb36 mutants, but it is at present unclear whether it is required for the ‘alternative’ CS-like cell wall modification. Taken together, the available data strongly suggest that SGN3 is either directly involved in sensing, controlling, and correcting CS assembly or, alternatively, regulates a pathway that performs these tasks. It remains to be seen whether the SGN3 ligand is really dependent on cell wall cues or is rather a direct readout for the state of the CSD in the membrane, for example, by analyzing suberin deposition in mutants that have a defective CS but unaffected CSD, such as rbohF .
Yet another pair of LRR-RLKs that have been implicated in cell wall-related signalling are FEI1 and FEI2, which act together with SOS5, an AGP, to maintain cellulose biosynthesis and anisotropic growth under high-sucrose and high-salinity conditions [231,232]. In addition, the FEIs interact with enzymes responsible for the biosynthesis of 1-aminocyclopropane-1-carboxylic acid (ACC), the precursor molecule of the gaseous phytohormone ethylene. However, ACC seems to act independently of ethylene perception and signalling in the FEI pathway and mediates at least some of the response to the inhibition of cellulose biosynthesis through isoxaben treatment . Whereas ACC clearly is involved in the response to at least impaired cellulose biosynthesis, the FEI pathway seems to act upstream of cellulose biosynthesis, indicating a more complex relationship between cell wall biosynthesis, surveillance, and response. Interestingly, a suppressor screen of the fei1 fei2 double mutant revealed auxin biosynthesis to be involved in mediating some of the phenotypes displayed by these plants. Impaired auxin biosynthesis also seems to be capable of suppressing the phenotypes caused by pharmacologically (isoxaben) and genetically (cesa6prc1, cobra) reduced cellulose biosynthesis . Consistent with the long-known effect of auxin signalling on cell wall properties and cell wall-related genes, these observations establish auxin signalling as a component of the compensatory response towards cell wall perturbations .
Other carbohydrate-binding receptors
Identification of true cell wall signalling receptors is mainly hampered by two independent bottlenecks: in those instances, where it has been possible to genetically establish signalling events in response to cell wall perturbations [43,45], a cell wall ligand has yet to be identified; on the other hand, several putative receptor proteins have been reported to bind cell wall components, but their potential signalling function in response to wall perception remains elusive. An example of the latter is provided by PROLINE-RICH EXTENSIN-LIKE RECEPTOR KINASE 4, which is involved in the response to abscisic acid during germination and root growth and requires pectinase treatment to be released in the soluble fraction of cell extracts . Other RLKs from the pathogenesis-related gene 5-like receptor kinase/thaumatin and LysM families, respectively, bind polysaccharides, but their physiological roles seem specific for plant–microbe interactions [236–242]. A relatively small number of the large group of plant receptor-like kinases are classified as lectin-type RLKs (LeRLKs). While these proteins generally seem to be involved in perception of and defence against pathogens as well , one L-type lectin RLK (LecRK1.9) mediates plasma membrane–cell wall attachment through binding of RGD (arginine–glycine–aspartic acid) motifs [244,245], analogous to animal cell adhesion mediated by the integrin–fibronectin network. Homology to integrin, on the other hand, has been observed in NON-RACE-SPECIFIC DISEASE RESISTANCE 1, which also mediates cell wall-plasma membrane attachment . FORMIN HOMOLOGY 1 (FH1) is another protein that mediates contact with the cell wall, in this case through a proline-rich extracellular domain . Intracellularly, the FH domain is capable of organizing actin cables. Thus, FH1 would be an ideal candidate for transducing mechanical signals into cytoskeletal rearrangements. For cortical MTs, mechanically triggered realignment has been shown to play important roles in organ and cell patterning [26,248–250]; however, it is at present unclear whether MT responses require transmembrane protein components such as, for instance, the CESA complex .
Finally, AGPs and extensin domain-containing LRR-RLKs are important cell wall-related plasma membrane proteins that have been implicated in signalling and are known or expected to be cross-linked with cell wall components, but it is not yet shown whether cell wall association actually triggers or modulates signalling [56,252–256].
Cell wall signalling and plant immunity
The genome of many plant pathogens and saprophytes encodes a suite of enzymes tailored for piercing the mechanical barrier that is the cell wall. Some breakdown products, however, can act as DAMPs  and reveal the presence of the pathogen to the plant. Apart from DAMPs, which give away the presence of cell wall-degrading activity, plants can also perceive the enzyme itself. For example, an RLP from tomato recognizes a fungal xylanase and triggers defence mechanisms , whereas the Arabidopsis RLP RESPONSIVENESS TO BOTRYTIS POLYGALACTURONASE 1/RLP42 detects the presence of fungal endopolygalacturonases . As indicated above, OGs created from HG by the activity of plant or pathogen-derived polygalacturonases can elicit defence responses, presumably through WAK receptors , and WAK-like proteins are required for resistance against Fusarium in Arabidopsis , rice blast disease , and head smut and corn leaf blight in maize [261,262], respectively. Therefore, cell wall signalling and immunity signalling overlap at least to some extent. This is also supported by obvious similarities between the response to pathogens and the response to the inhibition of primary cell wall cellulose biosynthesis, which includes ROS production, defence gene expression, and the accumulation of jasmonic acid and salicylic acid . Notably, altered cell wall composition and properties often lead to increased pathogen resistance . Future work will show the role of cell wall components in the perception of a damaged self and reveal the extent of cross-talk between defence and growth-related signalling.
Conclusion and outlook
It is a fundamental question of plant biology how plant cells control cell elongation through cell wall loosening and remodelling while maintaining structural integrity. This, but also potentially many other processes, such as feedback control of cell differentiation, cell identity patterning and maintenance, and the perception of a damaged self, require communication between the extracellular and intracellular space through cell wall signalling. Above, I highlighted some of the recent advances regarding the role of RLKs in cell wall signalling. This area of research faces several challenges that need to be addressed in the future. First, the signals created in the cell wall are in most cases not well characterized. Possible signals range from purely mechanical cues to carbohydrate ligands, the identification of which is intrinsically difficult due to the heterogeneity of the cell wall and the technical challenges of carbohydrate chemistry. A potential breakthrough is provided by the development of glycan arrays that can be probed with the recombinant ECDs of cell wall receptor candidates. It should be noted that cell wall binding as such does not equal cell wall sensing and to unravel the mechanism of this sensing is one of the most profound challenges for the field in the future. Second, the identification of novel cell wall signalling pathways is often limited to serendipitous findings or requires brute force approaches such as transgenic manipulation of wall properties to induce a phenotype that can be used as a readout for genetic screens. Deciphering how signalling from the cell wall is integrated with the regulatory framework of plant development, defence, and stress responses thus will require a truly multidisciplinary effort combining genetics, biochemistry, (phospho)proteomics, carbohydrate chemistry, mechanics, and computational modelling to cope with the complexity and non-intuitive effects of (multiple) feedback loops. Despite these challenges, more and more components of cell wall signalling are being discovered, and hopefully the near future will show how these feedback signalling pathways are involved in the orchestration of growth, defence, differentiation, and development.
BRI1-ASSOCIATED KINASE 1
BRI1 SUPPRESSOR 1
CASPARIAN STRIP MEMBRANE DOMAIN PROTEIN 1
Catharanthus roseus receptor-like kinase1-like
Casparian strip domain
damage-associated molecular patterns
epidermal growth factor
FORMIN HOMOLOGY 1
guanidine nucleotide exchange factor
pectin methylesterase inhibitor
PROTEIN PHOSPHATASE 2A
RAPID ALKALINIZATION FACTOR 1
respiratory burst oxidase homologue
receptor-like cytoplasmic kinase
Rho of plants
reactive oxygen species
SOMATIC EMBRYOGENESIS RECEPTOR KINASE
WALL-ASSOCIATED KINASE 1.
Research in the author's laboratory is funded by the German Research Foundation (Deutsche Forschungsgemeinschaft, DFG) through the Emmy Noether programme [grant number WO 1660/2-1].
The Author declares that there are no competing interests associated with this manuscript.