Plasmodesmata at a glance

Plasmodesmata (singular plasmodesma) mediate direct cell-to-cell communication between neighboring cells; this facilitates the transport of cellular contents that vary in size and chemical form. They are one of the key cellular structures that distinguish plants from the animal system. Adjacent animal cells can form gap junction-mediated cytoplasmic pores to enable the exchange of small diffusible molecules across their plasma membranes (PMs). In plants, the cellulose walls of individual plant cells keep them from making direct contact with neighboring cells at their PM; instead, plasmodesmata serve as the PM-lined cytoplasmic bridges across the cell walls. Plasmodesmata are considered essential to support the diversification of cell and tissue types in multicellular plant bodies; however, despite their importance, their molecular composition, as well as their underlying mechanisms of transport and biogenesis, remain largely enigmatic. In this Cell Science at a Glance article and the accompanying poster, we will highlight the formation and morphology of plasmodesmata, along with signaling molecules that are localized to or pass through them, to demonstrate their dynamic and adaptable nature. Furthermore, we will summarize methods to dissect whether a protein is associated with or is mobile across plasmodesmata, as well as techniques to monitor and assess the state of plasmodesmal permeability.

Plasmodesmata bridge adjacent cells across cell wall matrices as membrane-lined cytoplasmic channels. In land plants, primary plasmodesmata form during cell division when the endoplasmic reticulum (ER) is incorporated into the developing cell plate, which creates the cytoplasmic pore – termed the cytoplasmic sleeve – between the PM and ER membrane linings (see poster). The ER within the plasmodesma (the desmotubule, or appressed ER) is continuous with the cortical ER, but appressed tightly, leaving no luminal space (Robards and Lucas, 1990). Interestingly, certain membrane microdomains within plasmodesmata have recently been found to be enriched in sterols and sphingolipids (Grison et al., 2015), which may attract proteins that carry the glycosylphosphatidylinositol (GPI) moiety, or other lipid-based post-translational modifications [as is the case for PLASMODESMATA CALLOSE BINDING PROTEIN (PDCB), β-1,3-glucanase (PdBG2), remorin, etc.]. Actin filaments and myosin-like proteins are localized within the cytoplasmic sleeve (Radford and White, 1998; White et al., 1994), and proteinaceous spoke-like projections that are regularly positioned within the cytoplasmic sleeve are thought to create nanochannels of varying size (Fisher, 1999; Terry and Robards, 1987). Callose, a β-1,3-glucan polymer, appears to serve as both a structural and a regulatory element of plasmodesmata. During cytokinesis, callose is deposited at the cell plate, but after cell division is completed, it is degraded from the new cell wall, except in the portions of cell walls that are adjacent to plasmodesmata (Northcote et al., 1989). Moreover, callose accumulates around plasmodesmata at various levels, which provides a mechanism to regulate their permeability (see poster).

Small (<1 kDa) soluble molecules move freely and extensively between cells through the cytoplasmic sleeve. The possibility exists that the appressed ER may also provide a membrane surface as a route for molecular transport within the plasmodesmata (Grabski et al., 1993; Guenoune-Gelbart et al., 2008), but this aspect remains to be further explored. Large diffusible molecules can also move between cells if their three-dimensional shape allows passage through the cytoplasmic nanochannels (e.g. free GFP, which has a molecular mass of ∼27 kDa, likely diffuses between cells because it forms a cylindrical shape that can fit in the 3–4 nm-wide microchannels; see poster). When plasmodesmal permeability is significantly increased though dilation or active gating, or through structural remodeling of widening the pore size (see poster), various macromolecules, including transcription factors, signaling molecules and various types of RNA–protein complexes, may traffic across specific cellular boundaries or domains (Lucas and Lee, 2004). Exploiting this endogenous trafficking mechanism, plant viruses can spread infectious materials between host cells (Heinlein, 2015; Schoelz et al., 2011) (see Box 1).

In land plants, plasmodesmata are, by default, formed during cell division as strands of cortical ER membrane are caught within the developing cell plate (see poster). Although the process of appressed ER formation is not fully understood, it is likely to be mediated by the combined effects of pressure from the growing cell wall and the interaction between ER and PM proteins, such as RETICULON-LIKE PROTEIN SUBFAMILY B 3 and 6 (RTNLB3/6), SYNAPTOTAGMIN 1 (SYT1) and VESICLE-ASSOCIATED MEMBRANE PROTEIN-ASSOCIATED PROTEIN 27 (VAP27), which stabilize this process (Knox et al., 2015; Kriechbaumer et al., 2015). Primary plasmodesmata are typically grouped in clusters, called pit fields, that exist in thinner areas of the cell walls (see poster).

As cells grow and expand in their cell wall surfaces, the density of primary plasmodesmata naturally decreases if no secondary plasmodesmata are produced across existing cell walls. This reduced plasmodesmal density is common within cells that have undergone substantial and rapid expansion, such as the cell walls that separate the epidermal layer from the underlying cortex in mature roots, or phloem sieve elements from parenchyma cells (Kollmann and Glockmann, 1999). Moreover, in some cells, plasmodesmata are actively disintegrated as part of a required cell differentiation process – for example, during guard cell maturation or germ cell development (Sager and Lee, 2014). By gradually reducing the density, or active degradation of plasmodesmata, individual cells, or groups of cells, can become symplasmically isolated or lose cytoplasmic connectivity from their neighbors.

In contrast, the number of plasmodesmata can also increase in existing walls independently of cell division (see poster). These secondary plasmodesmata often form during cell expansion to maintain or increase plasmodesmal density in the growing walls (Burch-Smith et al., 2011). The mechanisms that underlie secondary plasmodesmata formation, including the cell wall-degrading enzymes and the proteins that are involved in ER membrane insertion, remain to be uncovered. However, various cellular signals have been implicated as inducers of this process: plants that are treated with phytohormones, such as salicylic acid (Fitzgibbon et al., 2013), cytokinin (Ormenese et al., 2006) or flower-inducing day length (Ormenese et al., 2000), have been shown to increase secondary plasmodesmal formation. New plasmodesmal connections are created between the fusion walls of graft partners during wound healing (Kollmann and Glockmann, 1991). Furthermore, parasitic plants appear to exploit the mechanism of secondary plasmodesmata formation to develop new plasmodesmata between themselves and their hosts, which allows them to siphon off host nutrients (Kim et al., 2014). These examples demonstrate a remarkable flexibility of plant cells to ensure the presence of the right number of plasmodesmata for their developmental or cellular context.

Plasmodesmal morphology varies greatly during the growth and differentiation of a cell, or, depending on the intercellular transport requirements, between different cell–cell junctions (see poster). A simple plasmodesma consists of a single channel between cells, but this can become modified into a branched or complex plasmodesma, with multiple channels that merge, split and may share the same opening (Faulkner et al., 2008). Simple and complex plasmodesmata regulate intercellular transport differently, in that complex plasmodesmata seem to act somewhat like a filter, increasing the specificity of which molecules can cross between cells. For example, as tobacco leaf plasmodesmata undergo changes from simple to branched forms during the sink-source transition (the time when the increase in daily carbohydrate exchange rate exceeds the daily increase in leaf carbohydrate content), cell-to-cell diffusion of GFP that is expressed under the control of a phloem-specific promoter is greatly reduced (Oparka et al., 1999; Roberts et al., 2001).

Plasmodesmata can take on a highly specialized morphology to control movement between certain tissues. For instance, in phloem, asymmetric plasmodesmata develop between companion cells and sieve elements, which give rise to a unique complex called the plasmodesmata-pore unit (see poster). Also in phloem, plasmodesmata at the sieve plate (where the ends of sieve elements are joined) become substantially enlarged through callose-mediated cell wall restructuring, which creates expanded cytoplasmic pores through which the mass flow of phloem sap can occur (Esau and Thorsch, 1985). A recent study on the unloading of root phloem revealed unique ‘funnel’-shaped plasmodesmata in the cell wall junctions between phloem sieve elements and the phloem pole pericycle cells, with a much wider opening on the phloem sieve element side compared to the phloem pole pericycle side (Ross-Elliott et al., 2017). The funnel shape thereby helps to direct the majority of molecules into the phloem pole pericycle during unloading, prior to transport into the outer root. Another study has visualized the development of plasmodesmata in root tips, and revealed that the PM and ER membranes in newly formed plasmodesmata are in such close contact that they completely lack the cytoplasmic sleeve (Nicolas et al., 2017). Intriguingly, even without a clearly discernable cytoplasmic sleeve, these new plasmodesmata can still facilitate cell-to-cell movement of GFP. It is likely that, as imaging methodologies for studying plasmodesmal ultrastructure advance, using tools such as electron tomography (Nicolas et al., 2017) and 3D scanning electron microscopy (Kremer et al., 2015), we will learn more about the unique morphologies that plasmodesmata may have.

Genetic mutant screening using Arabidopsis has identified a number of genes that impact cell-to-cell movement of molecules and/or plasmodesmal structure. These genes include INCREASED SIZE EXCLUSION LIMIT (ISE) 1 and ISE2, which encode a mitochondrial DEAD-box RNA helicase and chloroplast DEVH-type RNA helicase, respectively (Kobayashi et al., 2007; Stonebloom et al., 2009); DECREASED SIZE EXCLUSION LIMIT 1 (DSE1), encoding a putative WD-repeat protein (Xu et al., 2012); GFP ARRESTED TRAFFICKING 1 (GAT1), encoding a chloroplast thioredoxin (Benitez-Alfonso et al., 2009); and the CHOLINE TRANSPORTER-LIKE 1 (CHER1) (Dettmer et al., 2014; Kraner et al., 2017). Mutations in these genes affect plasmodesmata in different ways. In ise1/2 loss-of-function mutant embryos, branched plasmodesmata increase in frequency (Burch-Smith and Zambryski, 2010), whereas their frequency decreases in dse1 embryos (Xu et al., 2012). In gat1 loss-of-function mutants, GFP unloading from the phloem is restricted in the root, and plasmodesmata are often occluded with heavily deposited callose (Benitez-Alfonso et al., 2009). ISE1 likely regulates the redox states of mitochondria and plastids, whereas ISE2 and GAT1 regulate those of plastids, altering organellar redox states to bring about changes in plasmodesmal permeability (Benitez-Alfonso et al., 2009; Stonebloom et al., 2012). Loss-of-function mutations in CHER1 lead to alterations in choline metabolites and defects in phloem transport, which correlates with malformation of sieve pores and sieve plates (Dettmer et al., 2014). Additionally, cher1 mutants develop fewer plasmodesmata per unit cell wall length, and the plasmodesmata that do form remain abnormal and consist of fewer branches (Kraner et al., 2017). The effects of these plasmodesmal mutants on its morphology are likely indirect, since the proteins encoded by their wild-type alleles associate with subcellular compartments other than plasmodesmata. However, these mutants illustrate the complexities that are associated with plasmodesmal form and regulation well.

Currently, there are only two different structural elements of plasmodesmata that are known to regulate cell-to-cell movement: the actin cytoskeleton and callose (see poster). It is speculated that F-actin and unconventional myosin motors gate and regulate the plasmodesmal size exclusion limit (SEL), but this is based solely on their localization to plasmodesmata, and the detrimental effects on cell-to-cell movement when actin filaments are disrupted (Tilsner et al., 2016). For callose, however, a considerable amount of physiological, cellular and genetic data is available, and this data supports the idea that an increase in plasmodesmal callose levels reduces molecular movement between cells. Whereas callose is considered a normal structural constituent of plasmodesmata, it usually accumulates only transiently, for example, at the cell plate during cell division or at wound sites (Jacobs et al., 2003; Nishimura et al., 2003; Northcote et al., 1989). At plasmodesmata, callose specifically enriches in the extracellular space between the cell wall and the PM lining, and it may span the entire length of the plasmodesmal channel, or be more focused around the neck regions at each orifice (Radford et al., 1998; Vatén et al., 2011) (see poster). Plasmodesmal morphology may even be altered by callose occlusion, for example as observed in mutants for SUCROSE-EXPORT-DEFECTIVE (SXD-1) (Botha et al., 2000).

Functionally, callose accumulation effectively reduces plasmodesmal permeability, which can eventually lead to plasmodesmal closure, by restricting passage through the cytoplasmic sleeve. As callose accumulation increases on the inflexible cell wall outside of the PM lining, it pushes the PM inward, and is therefore assumed to physically squeeze the cytoplasmic sleeve to narrow the plasmodesmal aperture (Radford et al., 1998). In contrast, callose degradation would reverse this spatial constraint, increasing plasmodesmal permeability. The levels of callose accumulation at plasmodesmata are determined by the balance between the activities of two types of enzymes, callose (or β-1,3-glucan) synthases and glucanases (Zavaliev et al., 2011).

Key genes that regulate callose levels specifically at plasmodesmata include members of the A. thaliana β-1,3-glucanase, AtBG_ppap (Levy et al., 2007) and PdBGs, as well as the A. thaliana callose synthase (CalS) family, encoded by CalS10/GSL8/CHORUS, CalS3/GSL12, CalS1/GSL6 or CalS8/GSL4. Mutations in the PdBG1 or PdBG2 genes increase callose accumulation and perturb normal lateral root formation (Benitez-Alfonso et al., 2013). In cals10 mutants, callose deposition at the cell plate is abolished, causing seedling lethality and pleiotropic phenotypes that range from defects in cell division and guard cell patterning, to alterations in phototropic response and plasmodesmal permeability (Guseman et al., 2010; Han et al., 2014). Gain-of-function mutations in CalS3 increase callose levels and alter root development (Vatén et al., 2011). Mutations in CalS1 abolish the hyperaccumulation of callose induced by salicylic acid or pathogenic bacterial infection, while those in CalS8 eliminate the callose deposition induced by mechanical wounding or reactive oxygen species. Furthermore, cals8 mutants show an increase in basal plasmodesmal permeability compared to that found in wild-type plants (Cui and Lee, 2016). These data, together with the observations on additional proteins that are found at plasmodesmata (see below, Box 2 and poster), suggest that a network of callose-regulating factors control the permeability of plasmodesmal channels.

A number of plasmodesmata-localized proteins, unrelated to any callose synthase or hydrolase family, have been found to regulate plasmodesmal permeability through callose-dependent or independent modes (see poster). Arabidopsis PDCB1 encodes an extracellular protein that contains a GPI-anchor within its X8 domain (Simpson et al., 2009). PDCB1 binds to callose in vitro, and its ectopic overexpression increases callose deposition at plasmodesmata. A chitin-binding receptor-like protein, LYSIN MOTIF DOMAIN-CONTAINING GPI-ANCHORED PROTEIN 2 (LYM2), is a PM protein that also accumulates at plasmodesmata (Faulkner et al., 2013). In lym2 mutants, the normal chitin-induced reduction in cell-to-cell movement is abolished, and chitin-triggered defenses in Arabidopsis are also compromised. Whether LYM2 regulates plasmodesmal permeability through a callose-dependent mechanism is unknown.

The PLASMODESMATA-LOCATED PROTEIN (PDLP) family encodes eight members of the receptor-like integral membrane proteins in Arabidopsis; they consist of a functionally unknown extracellular domain and a short cytoplasmic tail across a transmembrane helix (Thomas et al., 2008) (see poster). The pdlp1 pdlp3 and pdlp2 pdlp3 double mutants exhibit increased GFP movement between cells, whereas overexpression of PDLP1 decreases GFP trafficking (Thomas et al., 2008). The pdlp5 mutant leads to an increased basal plasmodesmal permeability, whereas overexpression of PDLP5 induces the opposite phenotype (Lee et al., 2011). PDLP5 regulates cell-to-cell movement by stimulating callose deposition at plasmodesmata through CalS1 and CalS8 (Cui and Lee, 2016) (see poster). PDLP5 is also required for a proper innate immune response to bacterial pathogens, as well as the normal salicylic acid-induced reduction in plasmodesmal permeability (Wang et al., 2013). Finally, both PDLP1 and PDLP5 are necessary for systemic acquired resistance in Arabidopsis (Lim et al., 2016) (see Box 2). It remains unknown whether PDLP1 and other PDLP members regulate plasmodesmal permeability in a callose-dependent manner, similar to PDLP5.

In the following sections, we summarize select methodologies for analyzing protein association with and impact on the permeability of plasmodesmata at the molecular and cellular level.

If a newly identified protein localizes to plasmodesmata, it should be present at the cell periphery as discrete puncta when tagged with a fluorescent protein (FP) and viewed with confocal microscopy. Certain proteins may move out of the cells in which they are originally produced and into other cells, either locally or systemically (Gallagher et al., 2014). The mobility of these non-cell autonomous proteins can be examined after biolistic DNA delivery or by comparing the overall accumulation pattern of the candidate protein in tissues to its native transcriptional expression domain (see poster). It is important to validate that the punctate fluorescent signals only associate with junctional walls by performing a series of optical z-sections partially or completely through the depth of the cell that expresses the fluorescently tagged candidate protein (Lee et al., 2011). This method can be combined with techniques that cause plasmolysis in the cell; if the punctate fluorescent signals remain at the walls when the protoplast is pulled away, it is more likely that they are associated with plasmodesmata (see poster). In addition, plasmodesmal association can be confirmed by testing for colocalization of the candidate protein with cell wall callose by using a fluorescent stain such as Aniline Blue (Zavaliev and Epel, 2015). Alternatively, a known plasmodesmata-associated protein can be used as a marker for colocalization (see poster). Popular plasmodesmal markers include viral movement proteins, such as MP30, encoded by the Tobacco mosaic virus, or MP17, which is encoded by the Potato leaf roll virus (Hofius et al., 2001). Endogenous plasmodesmal proteins can also be used as markers, including all eight members of the PDLP family (Lee et al., 2011; Thomas et al., 2008), PDCB1/2 (Simpson et al., 2009) and PdBG1/2 (Benitez-Alfonso et al., 2013). Correlative light and electron microscopy (CLEM) (Lee et al., 2011; Modla et al., 2015) (see poster), and super resolution three-dimensional structured illumination microscopy (Fitzgibbon et al., 2010), are also suitable to address whether a protein locates within or around the plasmodesmata, as an alternative to immunogold labeling.

In response to varying exogenous and endogenous factors, plants constantly modify the extent of molecular trafficking between cells by fine-tuning plasmodesmal permeability (Sager and Lee, 2014). One method that is used to detect such changes in plasmodesmal permeability quantifies the extent to which free GFP (or other FPs in monomeric forms) diffuse from cell to cell (Stadler et al., 2005). The plasmodesmal SEL may differ in particular cell types, or change under specific conditions, and these changes can be monitored by using concatenated GFP (2× or 3× GFP) (Kim et al., 2005). Similar to the protein movement assays described earlier, free GFP as a symplasmic reporter is expressed either through DNA bombardment in individual epidermal cells or by transgenic expression in conjunction with a cell- and/or tissue-specific promoter. These FP-based cell-to-cell movement assays have been the most broadly used approaches in the field. Microinjection techniques using various fluorescent tracers of different molecular mass have been extensively employed to evaluate the SEL, as well as the permeability of plasmodesmata in different cell types of varying plant species (Kragler, 2015). Alternatively, membrane-permeable dyes could be applied to intact plants in a non-invasive manner to measure plasmodesmal permeability at the tissue level. For example, the Drop-ANd-See (DANS) dye-loading assay utilizes the membrane-permeable non-fluorescent dye carboxyfluorescein diacetate, which enters epidermal cells once loaded onto an intact plant surface and is immediately converted into the membrane-impermeable fluorescent tracer carboxyfluorescein (Cui and Lee, 2016; Cui et al., 2015; Gui et al., 2014; Lee et al., 2011; Lim et al., 2016; Wang et al., 2013) (see poster). This method is an ideal choice when addressing questions that require real-time, non-invasive analyses. It represents a good approach to determine whether or not plasmodesmal permeability may have been affected owing to genetic mutations or specific exogenous treatments (see poster). In addition, quantitative analysis of plasmodesmal callose levels using Aniline Blue stain can serve as an indirect measure for fluctuations in plasmodesmal gating (Zavaliev and Epel, 2015), based on the negative correlation between plasmodesmal permeability and callose accumulation. However, since plasmodesmal permeability might also be altered in a manner independent of callose, it is advisable to perform the callose measurement in conjunction with additional assays that directly report changes in cell-to-cell movement. Measuring changes in plasmodesmal permeability can be quite challenging, because plasmodesmal gating is a highly dynamic cellular event and the impact of a particular gene and/or protein may be specific to certain cell or tissue types, developmental stages or environmental conditions. Hence, multiple approaches may be necessary to address whether a novel gene and/or protein is related to plasmodesmata-mediated cell-to-cell communication.

Plasmodesmata are essential for the life of multicellular plants, yet we still lack a great deal of knowledge regarding their molecular anatomy, composition and regulatory mechanisms. For example, many of the molecules that constitute the plasmodesmal channel itself have yet to be discovered. Furthermore, the genetic factors that control primary and secondary plasmodesmal formation, as well as the proteins required for structural modification, remain to be identified. Finally, how specific proteins associate with plasmodesmata and transport through plasmodesmata, and how cells integrate specific intracellular signals with plasmodesmal components to execute intercellular responses, and vice versa, remain to be elucidated. However, those gaps in our understanding of plasmodesmata will be narrowed through genomic and proteomic data mining, and the development of new methodologies and technologies enabling real-time monitoring of cell-to-cell movement and structural analysis of developing plasmodesmata. In particular, current advances in imaging acquisition, resolution and analysis tools seem to hold much promise to unfold morphological details and variations in plasmodesmal ultrastructure. Since accumulating evidence now supports that animal cells and even bacteria form plasmodesmata-like intercellular channels (see Box 1), understanding the molecular details of plasmodesmata could provide significant insights into universal mechanisms that orchestrate communication between cells.

We apologize to the colleagues whose great works could not be directly or indirectly cited in this review article because of the limitations in the scope and the space.