Advances in Plant ER Architecture and Dynamics^,

Abstract

Recent advances highlight mechanisms that enable the morphological integrity of the plant ER in relation to the other organelles and the cytoskeleton.

The endoplasmic reticulum (ER) is a dynamic subcellular compartment that is essential to eukaryotic life because it contributes significantly to the synthesis of fundamental building blocks of the cell, including proteins and lipids, and it acts as an important architectural scaffold to maintain a well-organized spatial distribution of the other endomembrane organelles. Recent analyses with live cell imaging coupled with genetics studies have brought to light the incredible dynamism of this organelle and the underlying drivers as well as the impact of the ER organization on the general cellular homeostasis and plant growth. In this review, we highlight the most recent advances in the understanding of the mechanisms that enable the morphological integrity of the plant ER in relation to the other organelles and the cytoskeleton.

The endomembrane system comprises endocytic and biosynthetic cellular processes that are closely integrated. At the core of the endomembrane system lies the ER, an essential and largely pleiotropic organelle. With its network of interconnected tubules and flattened cisternae, the ER represents the organelle with the largest membrane surface area and can be considered as the gatekeeper of the secretory pathway that controls multiple checkpoints in protein biosynthesis: folding, quality control, signaling, and degradation. In addition to proteins such as receptors, ion channels, and enzymes, the ER synthesizes a wide variety of cargo molecules that control a large spectrum of physiological and essential processes, and are eventually shipped from the ER or retained in this organelle (Aridor and Hannan, 2000; Kim and Brandizzi, 2016; Brandizzi, 2017). Furthermore, with its function in controlling protein synthesis and folding, the ER has an important role in abiotic and biotic stress resistance through the unfolded protein response signaling (Angelos et al., 2017). The ER also is an important cellular compartment for calcium storage and carbohydrate metabolism (Vitale and Denecke, 1999; Vitale and Galili, 2001).

At a submicron level, the ER network is organized in domains that are morphologically distinct and that assume specific functions (Staehelin, 1997). These characteristics make the ER a morphologically continuous cellular compartment that is nonuniform at the functional and structural level. In humans, alterations in ER-mediated processes cause disease phenotypes that have been classified into three groups: (1) cargo retention and degradation, (2) cargo accumulation and ER stress, and (3) ER transport machinery diseases (Aridor and Hannan, 2000). Also in plant cells, defects in ER functionality lead to various developmental defects (Tamura et al., 2005; Conger et al., 2011; Stefano et al., 2012; Renna et al., 2013), supporting a critical role of the ER for organism biology at large.

Morphologically, the ER is able to reorganize, enlarge, and contract its highly dynamic polygonal tubular network both spatially and temporally (Ridge et al., 1999; Sparkes et al., 2009a; Stefano et al., 2014). The integrity of the ER network structure is important to maintain an efficient unfolded protein response (UPR), as demonstrated by the evidence that loss of ER-shaping proteins leads to attenuation of the UPR signaling in conditions of accrual of unfolded secretory proteins in the ER (Lai et al., 2014). Therefore, there is a strong connection between the morphology and the functional integrity of the ER. The plant ER also entertains strong functional connections with other organelles, including the Golgi apparatus and the vacuole to which newly synthesized proteins can be exported (Brandizzi et al., 2002; Shimada et al., 2003; Brandizzi, 2017), but also with chloroplasts with which the ER synthesizes essential lipids (Hurlock et al., 2014; Block and Jouhet, 2015). Possibly via direct connections, the ER movement influences the movement of other organelles, supporting an emerging model in cell biology that the ER interaction with other organelles is important not only for the exchange of constituents with other cellular compartments but also for their spatial distribution and function (Stefano et al., 2014, 2015) In this review, we highlight some of the recent and exciting literature addressing the fundamental questions on how the ER morphology and dynamics are controlled and how the ER interacts with the cytoskeleton and other organelles in plant cells.

THE ER IS A PLEOMORPHIC ORGANELLE WHOSE MORPHOLOGY AND DYNAMICS CHANGE DURING THE LIFE OF THE CELL

In general, the plant ER assumes the shape of a membrane network resembling the arrangement of a spider web with interconnected tubules and cisternae within the cell. Although the bulk of the ER is restricted at the cell cortex where it is sandwiched between the tonoplast and the plasma membrane (PM), long ER tubular strands characterized by a high streaming velocity cross along the central vacuole (Ueda et al., 2010; Sparkes et al., 2011). Electron microscopy studies have revealed the presence of smooth ER, rough ER, and nuclear envelope regions (Hawes et al., 1981; Craig and Staehelin, 1988; Staehelin, 1997). The rough and smooth ER regions are subdomains with associated ribosomes or ribosome-free regions, respectively. The nuclear envelope is enwrapped by the ER, resulting in a double membrane delimiting the nucleus. Additionally, the ER passes through the plasmodesmata, which are tiny channels that protrude into the cell wall and interconnect the cytoplasm of neighboring cells (Carr, 1976; Wright and Oparka, 2006). These structures are unique to plants and are crossed by a narrow tube-like structure, named desmotubule, which is derived from the ER (Quader and Zachariadis, 2006; Knox et al., 2015; Nicolas et al., 2017). As a result, the ER of each cell is interconnected to the neighboring cells through these channels, forming a virtually unique organelle whose extension is not delimited by the cell’s boundaries. The ER also is attached to the PM through ER-PM contact sites (EPCSs), which are largely immotile subdomains of the ER that underlie the PM and whose number and density at the cell cortex diminish as cells expand (McFarlane et al., 2017). Fusion profiles of the ER membrane with the PM at the EPCSs have not been observed; however, proteins such as VAP27 proteins and Synaptotagmin1 (SYT1) have been shown to accumulate at the EPCSs (Wang et al., 2014; Levy et al., 2015; Pérez-Sancho et al., 2015; Siao et al., 2016; McFarlane et al., 2017). The VAP proteins are conserved across kingdoms and possess three main regions, a C-terminal transmembrane domain, an N-terminal major sperm domain, and a coiled-coil domain (Wang et al., 2014). The VAP subfamily of VAP27 proteins may contribute to ER anchoring to the cell surface via the plasmodesmata as well as bridging the ER with the PM, through yet unknown mechanisms that depend on cell wall integrity (Wang P. et al., 2016). The plant SYT1 is a close ortholog of the yeast tricalbins and metazoan synaptotagmins that serve as ER-PM anchors in these nonplant cell systems. Similar to VAP27, SYT1 is localized to the bulk ER, and the EPCSs and SYT1 has been localized at EPCSs demarcated by VAP27 (Pérez-Sancho et al., 2015). Based on the ability of VAP27 to interact with the plant-specific actin-binding protein NET3c and microtubules (MTs), it appears that the EPCSs marked by VAP27 may serve as ER-PM hubs where the two major cytoskeletal components of plant cells converge (Wang et al., 2014). While roles for nonplant EPCSs are emerging, including lipid homeostasis and Ca²⁺ influx (van der Kant and Neefjes, 2014; Wakana et al., 2015), a physiological role for the plant EPCSs has yet to be defined. It has been demonstrated that SYT1 interacts with phospholipids possibly through electrostatic interactions and may therefore bridge the ER membrane with the PM. A loss of SYT1 reduces the ability of cells to withstand mechanical pressure (Pérez-Sancho et al., 2015), which may underlie a function of these sites in the plant adaptation to abiotic stresses present in the natural environment. The molecular mechanisms underlying such a role have yet to be explored, but the evidence that the EPCSs demarcated by SYT1 largely overlap with VAP27-EPCSs suggests a functional connection between mechanosensing and cytoskeletal organization at these enigmatic sites. The identification of SYT1 and VAP27 as components of the EPCS proteome opens up the exciting opportunity to further define the EPCS constituents, which may provide additional tools to understand the cellular role(s) of plant EPCSs.

With electron microscopy analyses, in addition to the EPCSs and plasmodesmata, the plant ER also has been found in contact with other membranes, including Golgi, mitochondria, vacuole, and plastids (Juniper et al., 1982; Staehelin, 1997). Indeed, using laser trap technology, it has been shown that the ER physically contacts the chloroplasts obtained from ruptured protoplasts expressing a fluorescent ER marker. More specifically, it was shown that the released chloroplasts remained attached to ER fragments that could be stretched out by optical tweezers. The applied force of 400 pN, which is a magnitude compatible with protein-protein interactions, could not drag a chloroplast free from its attached ER (Andersson et al., 2007). Using confocal imaging of a fluorescent fusion, the Brassica napus CLIP1 lipase/acylhydrolase (BnCLIP1) has been detected recently in transient expression in tobacco (Nicotiana tabacum) at the ER-chloroplast contact sites (Tan et al., 2011), also known as PLAMs. BnCLIP1 enzyme exhibited a discrete localization on the outer envelope membrane at the junction between the ER and plastids. The subcellular distribution of a protein such as BnCLIP1 at the PLAM is consistent with a potential role of these sites in the coordinated synthesis of lipids between the ER and plastids (Tan et al., 2011). Yet, it is unknown whether proteins of this kind have only a biosynthetic role at the PLAMs, such as lipid synthesis and transport, or have a scaffolding role to tether the two organelles together.

Using an optical trapping and tweezer system, physical contacts of the ER with the Golgi also have been established. In plant cells, the Golgi apparatus is dispersed into polarized mini-stacks that are motile (Boevink et al., 1998; Nebenführ et al., 1999). By trapping and pulling Golgi stacks in cells coexpressing fluorescent reporters for the ER and for the Golgi, it was shown that the pulling of Golgi stacks in cells where the movement of the ER and the Golgi was chemically inhibited was followed by a movement of an ER tubule in association with the Golgi (Sparkes et al., 2009b). Recently, using a similar approach, it was shown that overexpression of a truncated membrane-anchored Golgi matrix protein AtCASP lacking the coiled-coil domain in the cytosolic region could weaken the ER-Golgi connections, supporting the presence of proteinaceous scaffolding that tethers the ER and the Golgi together (Osterrieder et al., 2017). Although the identification of AtCASP as a putative ER-Golgi tether is a landmark in plant cell biology, it will be important to pursue further the identification of the proteins responsible for the tethering of different membranes with the plant ER not only at the Golgi but also with the PM and the other organelles in the plant cell. This is a field that is considerably lagging behind compared with nonplant cell systems (Rocha et al., 2009; Eden et al., 2010; Kornmann et al., 2011; Stefan et al., 2011; Michel and Kornmann, 2012; Murley et al., 2015).

THE ER CHANGES SHAPE DURING CELL DEVELOPMENT

One of the most noticeable features visible through microscopy analyses of live cells expressing bioreporters of the ER is the high dynamicity of this organelle; indeed, the ER tubules and cisternae continually move and rearrange, evolving the overall architecture during time and cell development. Curiously, in plant cells, the ER does not have only one form. Indeed, during the life cycle of a plant cell, the ER undergoes considerable reorganization of the morphology and dynamics (Fig. 1; Ridge et al., 1999; Stefano et al., 2014; McFarlane et al., 2017). Initial reports that the ER assumes different shapes were provided more than 2 decades ago in analyses using a fluorescent probe for lipids (DiOC6) as well as a fluorescent protein targeted to the bulk ER (Hepler et al., 1990; Ridge et al., 1999). The DiOC6 dye was used to explore the structure of the ER in moss during bud formation. Noticeable changes in ER architecture were established, starting from a dense meshwork of membranes that was reorganized into an open reticular network as the cell underwent growth. Through analyses performed using a fluorescent protein tagged to the ER on root and hypocotyl cells of Arabidopsis (Arabidopsis thaliana), it was possible to distinguish an ER characterized by lamellar sheets in early phases of cell expansion followed by a change toward a reticulate tubular structure, which is typical of the ER of fully expanded cells (Ridge et al., 1999). It also was noticed that in fully expanded root epidermal cells that give origin to root hairs, the reticular ER network was condensed at the sites where the root hairs are formed (Ridge et al., 1999). The organization of the ER morphology becomes even more puzzling in dividing Pinus root cells in which the ER at preprophase and prophase is spatially rearranged to overlay the MTs (Quader and Zachariadis, 2006). Indeed, treatment with oryzalin, a MT inhibitor, affects the formation of the tubular ER (tER)-preprophase band, tER-metaphase spindle, and tER phragmoplast, suggesting that at least in Pinus root cells the ER network organization may depend on MTs during mitosis and cytokinesis. Although together these results support that the ER assumes different morphology during cell growth, it is yet to be shown whether there may be a functional link between the shape of the ER and the ER function at specific stages of cell growth and development. It has been shown recently that mutations in the ER-shaping protein Root Hair Defective3 (RHD3) compromise not only the overall organization of the ER and the transition from extensive sheet-like form to a reticulated pattern, but also cell elongation (Stefano et al., 2014). These results therefore support the existence of a close connection between the ER shape and cell elongation, although the underlying mechanisms have yet to be defined. Based on the evidence that the loss of RHD3 alters the distribution of auxin in roots (Stefano et al., 2015), increases the cell’s phospholipid content, and leads to an attenuation of the UPR of the ER (Lai et al., 2014; Maneta-Peyret et al., 2014), the relationship between ER morphology and cell growth may be the sum of several processes that are affected by the alteration of ER architecture.

Figure 1.

In plant cells, ER architecture is correlated to cell expansion. Shown are confocal images of wild-type Arabidopsis Col-0 cotyledon epidermal cells expressing an ER lumen marker ERYK (Nelson et al., 2007) at different phases of cell expansion. Note the change in the morphology of the ER network, which from a most cisternal appearance in cells 3 d after germination (DAG) assumes progressively a more reticulated organization as cells expand. Scale bar = 5 μm.

ER DYNAMICS DEPEND ON THE CYTOSKELETON AND ER-SHAPING PROTEINS

In fully expanded cells, the ER network is highly pleiotropic. Despite an anchoring to stable EPCSs, the network undergoes profound rearrangements through tubule interconversion into other tubules as well as fusion of tubules with cisternae. While processes of fission/fragmentation have been verified in nonplant cells systems, such as sea urchin, starfish, lacrimal cells, and nonneuronal cell lines (Terasaki and Jaffe, 1991; Jaffe and Terasaki, 1993; Subramanian and Meyer, 1997; Dayel et al., 1999; Ribeiro et al., 2000; Harmer et al., 2002; Kucharz et al., 2009), the common profiles for rearrangement of the plant ER are homotypic membrane fusion and tubule emergence from other tubules. These rearrangements are guided primarily by the actin cytoskeleton (Sparkes et al., 2009a), with the MTs offering a minor yet significant contribution. Specifically, MTs guide ER tubule extension with an almost 20-fold slower rate compared with actin-based extension, and MTs appear to provide anchoring for the formation of multiway junctions (Hamada et al., 2014). While the identity of the proteins connecting the plant ER to the MTs are yet unknown, recently two proteins have been shown to link ER to actin. The first protein identified, SYP73, belongs to the three-member SYP7 family of plant-specific SNAREs (Sanderfoot et al., 2000). These proteins, named SYP71, SYP72, and SYP73, share a high degree of sequence similarity (Sanderfoot et al., 2000; Cao et al., 2016).

SYP73 binds actin directly and possesses a short luminal domain, a putative transmembrane domain, and a cytosolic region, which contains an actin-binding motif (Cao et al., 2016). SYP73 is primarily localized to the ER. In conditions of overexpression of SYP73, the architecture of the ER changes dramatically with a reduction of the cisternal profiles and a network pattern that overlaps the actin filaments. Conversely, a loss of SYP73 causes enlargement of the ER, reduces its streaming, and compromises cell elongation (Cao et al., 2016). These results pose that SYP73 is an ER-actin linker. SYP73 is only one of the components of the SYP family; therefore, other SYP proteins might share similar functions that may be important in different developmental stages. The other protein implicated in ER-actin connection is NET3B (for NETWORKED 3B), which is a plant-specific protein belonging to a superfamily containing 13 members (Hawkins et al., 2014). All NET proteins contain two important domains. In the C-terminal region, a predicted coiled-coil domain may be important for protein-protein interactions, and the second important domain that characterizes this family is the NET actin-binding domain, which may act as an adapter to link membranes to the actin cables. In particular, NET3B has been shown to associate in vivo with the ER. When the protein fused with a fluorescent marker is overexpressed, similarly to conditions of SYP73 overexpression, the architecture of the ER rearranges into resembling the actin cytoskeleton (Wang and Hussey, 2017). Although differently from SYP73, a direct interaction of NET3B with actin has not been established yet, these results suggest that NET3B may function as a linker between the ER and actin. Analysis of a NET3B knockout did not show any significant defects when compared with the wild type (Wang and Hussey, 2017), suggesting that there may be functional redundancy among the NET proteins.

While the cytoskeleton provides a dynamic framework for the overall ER architecture, ER proteins such as reticulons and RHD3 have an important role in the shaping of the network. Reticulons are membrane integral proteins that assume a wedge-like topology with their transmembrane regions (Voeltz et al., 2006; Nziengui et al., 2007; Sparkes et al., 2010). By inserting into the membrane, the reticulons form low-mobility oligomers and induce high curvature of the ER membrane, which results in the formation and stabilization of tubules (Shibata et al., 2008; Hu et al., 2011). Consistent with this function, overexpression of Arabidopsis RTN13, a reticulon that localizes at the ER tubules and the edges of ER cisternae, causes constrictions of the ER lumen and reduces the diffusion of lumen markers in the ER (Tolley et al., 2008, 2010). The function of RTN13 depends on a small conserved domain at the C-terminal region that contains a putative amphipathic helix (APH; Breeze et al., 2016). Deletion of APH did not impair oligomer formation but disrupted the membrane-shaping function of RTN13 in vivo (Breeze et al., 2016). These results are important as they support that the membrane-shaping function of reticulons may not be linked to their ability to oligomerize but to the presence of the APH domain. The plant family of reticulons contains 21 members (Nziengui et al., 2007; Sparkes et al., 2010). Given the large size of this family and the possibility that reticulons may share overlapping functions, no phenotype of the ER network in knock-outs has been reported. Nonetheless, it would be interesting to carry out complementation tests with reticulons lacking the APH to pinpoint the molecular role of this domain in the context of ER shaping.

In addition to the reticulons, RHD3 has important shaping activity at the ER. RHD3 is an ER integral protein with two putative transmembrane domains. A defective allele of RHD3 was first identified in a screen for root hair defects (Schiefelbein and Somerville, 1990). Loss of function of RHD3 leads to a reduced elongation of cells of the primary root (Stefano et al., 2012). At a subcellular level, the loss of RHD3 also reduces the formation of three-way junctions owing to the capacity of the protein to fuse membranes in a GTP-dependent fashion (Ueda et al., 2016). In this context, RHD3 functions analogously to similar membrane-associated dynamin-like GTPases, such as metazoan atlastin and yeast Sey1p (Chen et al., 2011; Zhang et al., 2013; Yan et al., 2015). Nonetheless, it has been shown that the functional regulation of RHD3 may depend on plant-unique features. In particular, overexpression of the C-terminal region of RHD3 (RHD3 amino acids 677–802) disrupts the ER network integrity; conversely, overexpression of the analogous Sey1p region (Sey1p amino acids 682–776) does not (Stefano and Brandizzi, 2014). Intriguingly, the C-terminal domain of RHD3 is phosphorylated, and it has been shown that kinase treatment of RHD3 induces oligomerization of this protein, which in turn may modulate its ER-shaping function (Ueda et al., 2016).

One puzzling question about RHD3 in relation to ER shape in general concerns the physiological role of ER membrane fusion. RHD3 belongs to a three-member family of proteins composed of RHD3, RHD3-like 1, and RHD3-like 2. The evidence that a double deletion of RHD3 and RHD3-like 1 is lethal and that the combined loss of RHD3 and RHD3-like 2 causes pollen lethality (Zhang et al., 2013) suggests that the formation of the tubular ER network is extremely important for the cell. Nonetheless, the evidence that in very young cells the ER does not have a reticulated form raises the question on how the shape of the ER may influence the physiology of cells at certain stages of growth compared with others.

In addition to reticulons and RHD3 proteins, there may be other proteins involved in ER shape. For example, in nonplant cells, the DP1/YOP1 proteins work as ER shapers in synergy with the reticulons. In Arabidopsis, five Yop1 homologs have been identified in the HVA22 family of proteins (Brands and Ho, 2002). One HVA22 protein, fused with a fluorescent protein, is localized at the ER with RHD3 (Chen et al., 2011). A functional characterization of HVA22 proteins in the context of ER shape is lacking, and it cannot be excluded that additional proteins may be involved in plant ER architecture. In metazoans and yeast, it has been shown that lunapark (Lnp1), a two-transmembrane domain protein, is required for ER shaping. In particular, Lnp1 has been implicated in contributing to the tubule-to-sheet conversion, most likely by stabilizing the three-way ER junctions (Chen et al., 2015; Wang S. et al., 2016). A functional ortholog of Lnp1 has yet to be identified in Arabidopsis, and it cannot be yet excluded that other proteins may have analogous functions to Lnp1 or that the stabilization of the plant ER junctions depends on different mechanisms compared with nonplant cell systems.

ER MOVEMENT AND CYTOPLASMIC STREAMING: MORE THAN MOLECULAR TRACKS AND MOTORS?

While in animal cells the rearrangement of the ER network as well as transport of vesicles largely depend on MTs, in plant cells this role is served primarily by actin and myosin motor proteins, like the plant-specific myosin XI-K (Li and Nebenführ, 2007; Prokhnevsky et al., 2008; Peremyslov et al., 2010; Ueda et al., 2010). Indeed, the loss of myosin compromises the movement of ER, Golgi stacks, peroxisomes, and mitochondria (Peremyslov et al., 2008; Ueda et al., 2010), which collectively is called cytoplasmic streaming (Woodhouse and Goldstein, 2013; Stefano et al., 2014). The biological role of cytoplasmic streaming has not been established at a molecular level, but it is plausible to hypothesize that, owing to the presence of a large central vacuole that can occupy up to 90% of the total cell volume, the movement of the cytoplasmic content may facilitate the delivery of nutrients as well as communication between distal sites in the cells. The ER network is a pervasive organelle that contacts heterotypic membranes (Andersson et al., 2007; Mehrshahi et al., 2013; Stefano et al., 2015), and it therefore may have a bearing on their movement. Indeed, as plant cells expand, concomitantly with the changes of the ER architecture from cisternal to tubular morphology, the velocity of ER streaming as well as overall cytoplasmic streaming increases (Stefano et al., 2014). Based on this evidence and the findings that the loss of RHD3 compromises the spatial distribution as well as the streaming not only of the ER but also of other organelles, such as the Golgi, peroxisomes, mitochondria, and endosomes, during cell expansion, it has been suggested that the ER movement may contribute to the general cytoplasmic streaming through the physical connections that it establishes with heterotypic membranes (Stefano et al., 2014, 2015). Intriguingly, a disruption of endocytosis has been verified in connection with the loss of ER streaming and disruption of ER morphology in an RHD3 mutant (Stefano et al., 2015). These results support the hypothesis that the dynamic positioning of organelles such as endosomes is important to ensure their function and maintain the overall cell homeostasis.

CONCLUDING REMARKS

We are witnessing an exciting era for the understanding of the plant ER. Although in recent years enormous advances have been made toward the understanding of how the structure and dynamic architecture of the ER are maintained, it is yet unknown how other fundamental aspects of the ER are established, including how ER subdomains attain and maintain their identity and what their cellular role may be. Some ER subdomains have been investigated recently, and the existence of an association between the ER and endosomes and the identity of the molecular players involved in the association of the ER with the PM or actin have been established. The proteins and regulatory processes underlying homotypic tubule fusion also are emerging (Fig. 2). One of the challenges for upcoming years will be to gain higher resolution of the three-dimensional architecture of the ER in relation to other organelles and the cytoskeleton. This likely will be achievable with electron tomography, which allows visualizing structures at a high resolution (6–8 nm; Donohoe et al., 2006). An example of the power of this method is provided by a recent study using cryo-electron tomography showing at molecular resolution the three-dimensional architecture of EPCSs in nonplant cells (Fernández-Busnadiego et al., 2015). This approach is likely to lead to more insights on the recently discovered linkage of the ER with the cytoskeleton and other components of the endomembrane system in plant cells.

Figure 2.

Diagram showing the association between the ER and other organelles in a plant cell. Abbreviations not defined in the text: CW, cell wall; PD, plasmodesma; V, vacuole; N, nucleus; Px, peroxisome; Mt, mitochondrion; GA, Golgi stack; En, endosome. Black numbered square regions indicate the ER-organelle associations or ER rearrangement properties identified in plant cells.