SEARCH

SEARCH BY CITATION

Keywords:

  • RHD3;
  • GTPase;
  • Arabidopsis thaliana;
  • endoplasmic reticulum;
  • Golgi movement

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

The mechanisms underlying the organization and dynamics of plant endomembranes are largely unknown. Arabidopsis RHD3, a distant member of the dynamin superfamily, has recently been implicated in plant ER morphology and Golgi movement through analyses of dominant-negative mutants of the putative GTPase domain in a heterologous system. Whether RHD3 is indispensable for ER architecture and what role regions other than the putative GTPase domain play in RHD3 function are unanswered questions. Here we characterized an EMS mutant, gom8, with disrupted Golgi movement and positioning and compromised ER shape and dynamics. gom8 mapped to a missense mutation in the RHD3 hairpin loop domain, causing accumulation of the mutant protein into large structures, a markedly different distribution compared with wild-type RHD3 over the ER network. Despite the GOM8 distribution, tubules fused in the peripheral ER of the gom8 mutant. These data imply that integrity of the hairpin region is important for the subcellular distribution of RHD3, and that reduced availability of RHD3 over the ER can cause ER morphology defects, but does not prevent peripheral fusion between tubules. This was confirmed by evidence that gom8 was phenocopied in an RHD3 null background. Furthermore, we established that the region encompassing the RHD3 hairpin domain and the C-terminal cytosolic domain is necessary for RHD3 function. We conclude that RHD3 is important in ER morphology, but is dispensable for peripheral ER tubulation in an endogenous context, and that its activity relies on the C-terminal region in addition to the GTPase domain.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

How organelles establish and maintain their morphological and functional identity is a fundamental question in cell biology. The secretory pathway is an excellent system by which to address this question because it comprises numerous organelles that are morphologically and functionally distinct. In most eukaryotes, the first secretory organelle, the endoplasmic reticulum (ER), assumes a characteristic network-like morphology (Baumann and Walz, 2001; Sparkes et al., 2009a). In plants, the Golgi apparatus, which receives secretory materials from the ER for processing and sorting, is composed of dispersed and polarized stacks (Matheson et al., 2006), which move by alternating wiggly and directional motion at a relatively high velocity (Boevink et al., 1998; Nebenfuhr et al., 1999). In highly vacuolated cells, Golgi stacks appear to be attached to the ER surface (Sparkes et al., 2009b). Little is known about how the complex architecture of the ER is formed and maintained, or what role the overall structure plays in ER function. In plant cells, the ER can be split in two interconnected regions: a peripheral network (cortical ER) and an underlying fast-streaming one (inner ER) (Ueda et al., 2011). Live-cell imaging has revealed that the ER is a highly dynamic organelle showing continuous growth, retraction, sliding and fusion of tubules to form strands, cisternae and polygons (Sparkes et al., 2011). ER tubules are pulled and extended from a membrane reservoir by cytoskeletal elements, like microtubules in mammalian cells and characean intermodal cells (Vedrenne and Hauri, 2006; Foissner et al., 2009), or as actin filaments in plant and yeast cells polymerize (Prinz et al., 2000; Brandizzi et al., 2003b). These high-curvature tubules are then stabilized by cytoskeleton-independent mechanisms. Members of a family of membrane curvature-inducing proteins, named reticulons/DP1/Yop1p proteins, have been implicated in such mechanisms (Voeltz et al., 2006; Tolley et al., 2008; Sparkes et al., 2010). Although reticulons and/or DP1/Yop1p appear to be the minimal components required for ER tubule formation in vitro, there may be additional uncharacterized factors that determine the shape of the ER network in vivo (Hu et al., 2009). This is particularly relevant for the dynamic rearrangement of the tubules that follows their stabilization. Dynamin-like proteins such as the metazoan atlastins and the yeast functional ortholog, SEY1, have been implicated in remodeling of the ER, specifically in ER homotypic fusion, as demonstrated by depletion of atlastins and over-expression of dominant-negative mutants that inhibit the formation of tubule interconnections (Hu et al., 2009; Orso et al., 2009). In vitro experiments have shown that Drosophila atlastin drives membrane fusion by bringing the ER membranes into contact through head-to-head interaction of the large GTPase domains and consequent conformational changes of the two juxtaposed atlastin proteins (Moss et al., 2011). It has been suggested that Arabidopsis RHD3 may be a functional ortholog of atlastin and SEY1 (Hu et al., 2009), but several crucial questions remain unanswered. It has been reported that over-expression of RHD3 proteins with mutations in the putative GTPase domain causes formation of unbranched ER and reduced Golgi movement in tobacco leaf epidermal cells (Chen et al., 2011). Because these results were obtained through over-expression of dominant-negative mutants in a heterologous system, it is not yet known to what the extent RHD3 is involved in organization of the ER and Golgi movement in plant cells. Several other important additional questions are also unanswered. For example, the role played by other RHD3 domains in addition to the putative GTPase domain, which is located in the N-terminal region of the protein, is unknown. Also, RHD3 has not been tested to determine whether it is a bona fide GTPase.

Here we report on a mutant, gom8, that was identified through a forward genetics screen of the plant Golgi (Boulaflous et al., 2008). We show that gom8 has reduced Golgi motility and defects in Golgi distribution and ER organization compared to wild-type. The gom8 mutant possesses a proline to serine substitution at position 701 in the anchor domain of the C-terminal region of RHD3. By characterization of gom8 and an RHD3 null mutant, we show that RHD3 plays an important role in organization of the ER, but that between-tubule fusion still occurs in the peripheral ER when the ER is depleted of RHD3, highlighting a dispensable role for RHD3 in tubule fusion. These data also show that integrity of the C-terminal region has a fundamental role in RHD3 function. We further support this conclusion by demonstrating not only that the C-terminal region of RHD3 is essential for ER anchoring, but also that replacement of the RHD3 C-terminal region by the ER anchor domain of PVA12, a type II ER protein, causes loss of RHD3 function.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Isolation of a Golgi movement and positioning mutant

To identify mutants with altered features of the plant Golgi, we performed a forward genetics screen based on confocal microscopy imaging of Arabidopsis thaliana Col-0 cotyledon pavement cells of EMS-mutagenized 12-day-old seedlings expressing ST–GFP (Boulaflous et al., 2008). Based on 25 000 seedlings from approximately 415 independent F2 EMS lines, we isolated a mutant (named gom8, for Golgi movement 8) in which Golgi stacks appeared to move less compared to the control (Movie S1). Quantification of the gom8 Golgi velocity in the cortical region of the cells (i.e. 3.0–5.0 μm from the cell surface) showed reduced velocity values in gom8 compared to wild-type (Figure 1a,b). Further subcellular analyses of ST–GFP in gom8 pavement cells showed large structures, usually one per cell, that included aggregates of Golgi stacks and fewer ST–GFP-positive membranes in the cortical region. The abnormal aggregations of Golgi stacks were located towards the inner region of the cell, generally below the region where the Golgi velocity measurements were acquired (Figure 1c,d). Individual Golgi stacks appeared to enter and leave the aggregates over time (Movie S1). Aggregates of Golgi stacks were also visible in root epidermal cells and root hairs (Figure S1). Similar aggregates were also visible in gom8 seedlings stably expressing the cis- and medial-Golgi marker GA–YK (Nelson et al., 2007), indicating that aggregation affects the entire Golgi stack, rather than isolated cisternae (Figure S2).

image

Figure 1.  Identification of gom8. (a, b) Maximal and mean velocity (μm/sec) of Golgi stacks in non-mutagenized ST–GFP plants (Col-0, control) and the gom8 mutant measured by time-lapse confocal microscopy of cotyledon pavement cells. Error bars represent standard deviation (SD). Asterisks indicate that data are significantly different from the gom8 mutant at < 0.05. (c) Reconstruction of confocal sequential optical sections in the Z-dimension (0–18 μm) of epidermis of ST-GFP and gom8 cotyledons. (d) As (c), but viewed as depth code images [blue, 0 μm (upper level); red, 18 μm (lowest depth)]. In (c) and (d), arrowheads indicate Golgi stacks; arrows show aggregates of Golgi stacks. Scale bar = 20 μm.

Download figure to PowerPoint

The ER morphology and dynamics are compromised in the gom8 mutant

Because of the close association of the Golgi and ER in vacuolated cells (Sparkes et al., 2009b) we determined the morphology of the ER in the gom8 mutant. To do so, we analyzed wild-type and gom8 stably expressing ER–YK, a soluble marker based on secretory YFP targeted to the ER by virtue of the signal peptide of AtWAK2 and retained in the ER by the tetrapeptide HDEL (Nelson et al., 2007). In wild-type cotyledons, the ER displayed the typical organization described previously for Arabidopsis cotyledon petioles (Ueda et al., 2011), i.e. a relatively static peripheral ER composed of a tubular and cisternal network (Figure 2a and Movie S2). The tubular network undergoes remodeling due to extension, docking and fusion of tubules with other tubules, as well as lateral fusion of the tubules (Movie S2). Observations at a more internal focal plane, generally approximately 2.0 μm deeper than the peripheral ER, revealed rapidly moving thick ER strands, which were also described previously in cotyledon petioles (Ueda et al., 2011). These ER strands had high YFP fluorescent intensity and were connected to the peripheral ER by tubules that were very dynamic, continuously forming and dissolving over time (Movie S2). Compared to the wild-type, the gom8 mutant showed enlarged and unbranched strands in the inner ER that were connected to the polygonal tubular ER in the cell cortex (Figure 2a and Movie S3; compare with Movie S2 for wild-type). Due to the close association of the Golgi and ER in highly vacuolated cells, these ER strands probably produce the dynamic Golgi aggregates described in Figure 1 (Figure S3). Compared to wild-type, the gom8 peripheral ER showed smaller ER polygons formed by the ER tubules (Figure 2a,b). Furthermore, the ER streaming velocity of the gom8 ER was significantly reduced compared to wild-type (Figure 2c,d).

image

Figure 2.  Comparison of ER organization in wild-type and the gom8 mutant. (a) Confocal images of the ER network in pavement cells of either Col-0 stably expressing ER–YK (control) or gom8/ST–GFP/ER–YK show a different architecture of the cortical ER with smaller polygons in the gom8 mutant. For clarity, only the YFP channel is shown in the gom8 panel. Scale bar = 5 μm. (b) Area measurements of polygons (μm2) and error bars (SD). (c) Maximal and mean velocities for ER streaming in ER–YK transformants (control) and the gom8/ST–GFP/ER–YK mutant. Data were measured in the YFP channel and normalized to wild-type. Error bars represent SD. Asterisks indicate that data are significantly different from wild-type at < 0.05

Download figure to PowerPoint

The gom8 mutant is linked to a missense mutation in RHD3

To establish the identity of the mutation responsible for gom8 phenotype, we generated a mapping population by crossing gom8 with Ler, and rough-mapped the mutation using a GeneChip Arabidopsis ATH1 genome array (Borevitz, 2006). The mutation was rough-mapped to chromosome 3 between 4.5 and 6.2 Mb. To identify the mutation, we sequenced gom8 genomic DNA using the Illumina Genome Analyzer II (Solexa), as described previously (Marti et al., 2010). Three homozygous single nucleotide polymorphisms (SNPs) were identified within the rough-mapped region, but only one SNP, in the At3g13870 locus, resulted in a typical G/C[RIGHTWARDS ARROW]A/T EMS transition (Maple and Moller, 2007). The CCT[RIGHTWARDS ARROW]TCT transition caused a missense mutation of a proline residue into a serine residue (P701S, Figure 3a,b). The locus represents RHD3 (root hair defective 3), encoding a putative GTPase originally identified in a screen for Arabidopsis mutants with root hair elongation defects (Wang et al., 1997). Consistent with the original phenotypic description of RHD3 alleles (Wang et al., 1997), gom8 shows defects in root hair elongation, as well as stunted growth of the primary root (Figure 3c). RHD3 has a putative hairpin domain, causing both the N- and C-termini of the protein to be located in the cytosol (Hu et al., 2009). Biochemical characterization of the cytosol-exposed N-terminal domain showed that RHD3 is an active GTPase (Figures S4 and S5). The P701S mutation occurs in the last amino acid residue before the second RHD3 transmembrane domain in the short loop of the hairpin region (Figure 3b). To confirm that the P701S mutation is responsible for the gom8 phenotype, we stably expressed wild-type RHD3 in the gom8 background, and found that the primary root and root hair length were complemented by wild-type RHD3 (Figure 3c,d), confirming that gom8 is an allele of RHD3. Finally, analyses of the distribution of a soluble secretory marker and fluorescence recovery after photobleaching (FRAP) analyses of ST–GFP in the Golgi did not reveal obvious differences between gom8 and wild-type (Figure S6), confirming that growth defects in rhd3 are unlikely to be linked to ER export defects (Chen et al., 2011).

image

Figure 3.  RHD3 complements the gom8 mutant. (a) Schematic diagram of the open reading frame of the At3g13870 locus (filled boxes, exons; lines, introns). The position of the T-DNA insertion in rhd3-7 is indicated by an arrowhead and was confirmed by sequencing. (b) Domain structure of the RHD3 protein. The GTPase domain is located in the cytosolic N-terminal region, followed by a middle domain, two transmembrane domains (black boxes) in the hairpin region, and a C-terminal cytosolic domain. The gom8 mutation is located at position 701 (P[RIGHTWARDS ARROW]S). (c) Root and root hairs of gom8/35S::RHD3, gom8 and Col-0 plants. The short root hair phenotype in gom8 plants was rescued in gom8/35S::RHD3 transformants. The insets show magnified regions of roots with root hairs. (d) RT-PCR in gom8/35S::RHD3, gom8 and Col-0 plants, showing higher levels of transcripts in the gom8/35S::RHD3 lines compared to the controls, as expected for transformed lines. UBIQUITIN10 (At4g05320) was used as the PCR control.

Download figure to PowerPoint

RHD3 is dispensable for between-tubule fusion

It has been proposed that atlastins, the animal homologs of RHD3, may be involved in ER architecture by facilitating homotypic fusion of the ER tubules (Moss et al., 2011). Analyses of the peripheral ER network in the gom8 mutant revealed that tubular fusion did occur (Figure 4a,b), suggesting that RHD3 may be important for ER morphology but not indispensable for tubule fusion, at least in the peripheral ER. To confirm this, we isolated a T-DNA insertion allele of RHD3 (SALK_106309; rhd3-7) (Figure S7), which showed no amplification of full-length RHD3 transcript and had the defects in root hair and root hair elongation typical of other RHD3 mutants (Wang et al., 1997; Zheng et al., 2004; Chen et al., 2011). We stably transformed rhd3-7 with GFP–CX, which is a GFP fusion to a sporamin signal peptide for ER insertion at the N-terminus of the fluorescent protein, and to the transmembrane domain and cytosolic tail of an Arabidopsis calnexin, an ER-membrane associated lectin (Wada et al., 1991), at the C-terminus. Confocal microscopy analyses of cotyledon epidermal cells showed that the ER in rhd3-7 phenocopied that of gom8, in that the inner ER formed enlarged strands (Figure S8), but tubular fusion events still occurred in the peripheral ER network (Figure 4c). To test the possibility that either the gom8 or rhd3-7 mutation led to an increase in transcription of two RHD3 isoforms, RHD3-Like 1 (At1g72960) and RHD3-Like 2 (At5g45160), we performed RT-PCR and real-time quantitative RT-PCR using gene-specific primers and found no appreciable differences in transcript levels (Figure S9). These data exclude the possibility of a compensatory effect of other RHD3 isoforms on the morphology of the ER of the RHD3 mutants. Furthermore, not only do these data indicate that gom8 is due to loss of function of RHD3 because it was phenocopied in a null allele, they also establish that RHD3 plays an important role in ER morphology but it is dispensable for between-tubule fusion in the peripheral ER.

image

Figure 4.  ER tubule fusion occurs in gom8 and rhd3-7 mutants. Images from a time-lapse sequence of pavement cells in an ER–YK transformant (a), gom8/ST–GFP/ER–YK (b) and rhd3-7/CX–GFP (c). For clarity, only the YFP channel is shown in the gom8 panel. The sequence shows tubular fusion events in the peripheral ER in wild-type and mutants. The frame acquisition time (top of panels) is identical for all samples. Arrowheads indicate ER tubules that have fused with other tubules. Scale bar = 5 μm.

Download figure to PowerPoint

The gom8 phenotype is linked to an aberrant distribution of RHD3 in the ER

To obtain further insight into the causes of the gom8 phenotype, we performed subcellular localization of YFP–RHD3(P701S) in comparison with wild-type RHD3 in stably transformed gom8 cotyledons. As the rhd3-7 mutant phenocopied gom8, we expected that the gom8 phenotype would be linked with reduced availability of RHD3 in the ER network compared to wild-type RHD3. Subcellular localization analyses of a YFP fusion to RHD3 showed a clear ER distribution of YFP–RHD3 (Figure 5a), as previously reported (Chen et al., 2011). YFP–RHD3 successfully rescued the subcellular gom8 Golgi phenotype, indicating that the fusion is functional. However, expression of YFP–RHD3(P701S) did not complement the gom8 phenotype (Figure 5b), as expected. Also, in contrast to the ER network-wide distribution of YFP–RHD3, YFP–RHD3(P701S) was mainly localized into bright structures (Figure 5b). Biomolecular fluorescence complementation (BiFC) analyses (Zamyatnin et al., 2006; Waadt and Kudla, 2008) using a cytosolic C-terminal half YFP and a complementary N-terminal half YFP fused to RHD3(P701S) at its N-terminus showed reconstitution of fluorescence (Figure S10), indicating that the N-terminal region of RHD3(P701S) is located in the cytosol, similarly to that of RHD3 (Figure S10). However, co-localization analyses of YFP–RHD3(P701S) with the ER marker ssGFP–HDEL (Brandizzi et al., 2003a) in stably transformed Arabidopsis cotyledons showed that the YFP–RHD3(P701S) structures are adjacent to the ER tubules (Figure S11), suggesting that a large proportion of the protein may not be associated with the ER. Although the mechanisms that lead to the aberrant distribution of RHD3(P701S) are not investigated here, our data indicate that RHD3(P701S) has a markedly different localization compared to the ER network distribution of wild-type RHD3.

image

Figure 5.  The P701S mutation affects the functionality and distribution of RHD3. Confocal analyses of pavement cells of cotyledons of gom8/ST–GFP transformed with either YFP–RHD3 (a) or YFP–RHD3(P701S) (b), showing that YFP–RHD3 is localized at the ER but the mutant protein is present in large structures (arrow and inset). In (a), the arrows indicate a Golgi stack and the arrowheads indicate one of the ER punctae in which RHD3 accumulates, as described previously (Chen et al., 2011). Scale bars = 5 μm.

Download figure to PowerPoint

The RHD3 C-terminal region has a functional role in ER morphology

Our finding that a point mutation in the RHD3 C-terminal domain affected RHD3 subcellular distribution and function prompted us to further explore the involvement of this region (residues 677–802) in the role of RHD3 in ER morphology by testing whether such a domain acts merely as an anchor domain region. To do so, we stably expressed a YFP fusion of the RHD3 region encompassing the N-terminal cytosolic domain [residues 1–676; YFP–RHD3(1–676)] in gom8, as well as transiently in tobacco leaf epidermal cells. We found that the protein was localized to the cytosol and Golgi stacks (Figure S12). Complementation of the gom8 phenotype was not observed (Figure S12A). These data indicate that the C-terminal region of RHD3 [RHD3(677–802)] is responsible for the ER association of RHD3. We then transformed gom8 with a YFP fusion of the RHD3 cytosolic domain [YFP–RHD3(1–676)] and the transmembrane domain and luminal domain of PVA12 [YFP–RHD3(1–676)PVA12(657–720)], an ER-associated type II membrane protein (Saravanan et al., 2009). The N-terminal domain of PVA12 is exposed in the cytosol (Saravanan et al., 2009). We therefore predicted that the N-terminal region of RHD3 in YFP–RHD3(1–676)PVA12(657–720) would also be localized in the cytosol, and confirmed this through BiFC experiments (Figure S10). We reasoned that if the C-terminal region of RHD3 has roles additional to that of a membrane anchor, the YFP–RHD3(1–676)PVA12(657–720) protein would not rescue the gom8 phenotype. Consistent with our hypothesis, we found that the gom8 phenotype was not rescued by YFP–RHD3(1–676)PVA12(657–720), although this fusion was localized over the ER network (Figure 6 and Figure S10), in marked contrast to the distribution of YFP–RHD3(P701S) (Figure S11). These data support the conclusion that the RHD3 C-terminal region has an important role in ER morphology in addition to that of a membrane anchor region.

image

Figure 6.  YFP–RHD3(1–676)PVA12(657–720) does not complement the gom8 phenotype. Confocal images of gom8 cotyledon cells co-expressing YFP–RHD3(1–676)PVA12(657–720) show a lack of complementation of the gom8 phenotype. Scale bar = 5 μm.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Here we describe identification and characterization of a recessive RHD3 allele (gom8) that allowed us to address several important questions that could not be answered in a recent study based on over-expression of RHD3 dominant-negative mutants in an heterologous system (Chen et al., 2011). By characterization of the gom8 mutant and an RHD3 null mutant, we established that reduced availability of RHD3 from the ER network leads to formation of unbranched ER strands in the inner ER; however, the peripheral ER appears to be less affected, with the occurrence of between-tubule fusion events. These data advance our knowledge on plant ER shape by showing that RHD3 is important for ER morphology but is not essential for ER tubule fusion in the peripheral ER. The gom8 mutation, which occurs in the membrane anchor region of RHD3, prompted us to investigate the role of the C-terminal region in the function of RHD3. We have established that this region is important for the role of RHD3 in ER morphology, in addition to its function as a membrane anchor. These data further advance our understanding of RHD3 involvement in ER shape by showing that regions other than the N-terminal domain have a bearing on RHD3 function in ER morphology.

RHD3 is an important but dispensable GTPase for control of the plant ER morphology

Here we describe the isolation and characterization of an RHD3 mutant with a serine in place of a proline residue in a short loop connecting two transmembrane domains in the hairpin region of RHD3. Prolines are key residues in hairpin loop integrity (Krieger et al., 2005), and it is possible that the P701S mutation in GOM8 may change the stereochemistry of the interhelix. This may alter not only the type of turn (β), but also affect nucleation of the succeeding helix. The change in turn type would arguably affect the position and orientation of the second putative transmembrane helix (Creighton, 2011). Changing the nucleation behavior of the second transmembrane helix could affect integration of the helical bundle into the membrane, perhaps leading to aberrant folding of RHD3 and/or reduced interaction with other proteins or lipids that may be necessary for stabilization of RHD3 in the ER membrane. Fluorescence reconstitution of cytosolic half YFP with the complementary half YFP fused to RHD3(P701S) confirms that the N-terminal region of the mutated RHD3 is exposed to the cytosol similarly to RHD3. However, the finding that YFP–RHD3(P701S) is confined mainly to globular structures suggests that the P701S mutation alters the spatial distribution of RHD3 in or on the ER, perhaps by inducing at least partial aggregation of the RHD3 protein. Independently of the effect of the P701S mutation on RHD3 at a molecular level, our results show that the wild-type distribution of RHD3 in the ER network is lost in the gom8 mutant, and that the gom8 ER and plant phenotypes are phenocopied in an RHD3 null mutant. Based on these data, we suggest that availability of RHD3 in the ER network is an important factor controlling ER morphology. In addition, we found that loss of RHD3 function yields enlarged unbranched tubular profiles in the inner ER and dense peripheral ER, compared to wild-type. We have also shown that ER tubule formation can occur in the peripheral ER in the gom8 mutant, similarly to an rhd3-7 mutant. Although we have not compared the rate of tubule formation in the mutant versus wild-type, ER polygons, which are the product of ER tubule fusion, as well as tubule fusion events, are clearly visible in the gom8 and rhd3-7 mutants. These data indicate that RHD3 may be dispensable for tubule fusion, at least in the peripheral ER. Recent evidence has shown that over-expression of RHD3-Like 2 can compensate for the root growth phenotype of an rhd3-defective allele (Chen et al., 2011). Although it is possible that RHD3-Like 2 is partially compensating for RHD3 loss of function in gom8 and rhd3-7, RHD3-Like 2 transcript levels did not increase in the gom8 mutant. Additionally, RHD3-Like 1, which is generally expressed in stamens, was not expressed in cotyledons. It is possible that RHD3 and RHD3-Like 2 may control tubule fusion in different domains of the ER. In this model, RHD3 could control tubule fusion in the inner ER, while RHD3-Like 2 controls tubule fusion in the peripheral ER. In the gom8 and rhd3-7 mutants, the strands of the inner ER would be unbranched because of lack of membrane fusion with the peripheral ER tubules, but tubule fusion would occur normally at the cell periphery due to the presence of a functional RHD3-Like 2 protein.

The RHD3 C-terminal region is an ER anchor and influences RHD3 function

It has been shown that over-expression of mutants of the conserved GTPase domain in the N-terminal region of RHD3 exert a dominant-negative effect on ER tubulation (Chen et al., 2011). However, it was not known whether RHD3 is a GTPase. We have biochemically established that RHD3 is a functional GTPase, which supports findings that specific mutations in the N-terminal region of RHD3 can affect its function in the ER (Chen et al., 2011). However, the influence of other domains on the role of RHD3 was not explored. Here we have found that the ER anchor domain of RHD3 is located in the C-terminal region (residues 677–802), as indicated by the finding that expression of the N-terminal region alone (residues 1-676) results in distribution in the cytosol and Golgi stacks in both the gom8 mutant and a heterologous system. These results are consistent with the presence of two putative transmembrane domains in the RHD3 C-terminal region. However, replacement of the RHD3 C-terminal region by a membrane-anchor region of a different protein, in this case PVA12, led to loss of function of RHD3, despite the fact that the N-terminal region of the protein chimera remained in the cytoplasm and the protein was distributed over the ER network, similarly to RHD3. Atlastins and SEY1 are known to interact with a number of ER-shaping proteins through their transmembrane domains (Hu et al., 2009; Park et al., 2010). It is possible that replacement of RHD3 transmembrane domains by that of PVA12 may displace RHD3-interacting proteins from a functional complex, leading to loss of function of RHD3, and, in turn, to lack of complementation of the gom8 phenotype. However, we cannot exclude the possibility that the C-terminal cytosolic domain of RHD3 downstream of the two transmembrane domains may also influence membrane architecture. The C-terminal region of atlastins has been proposed to form a membrane-active amphipathic helix (Moss et al., 2011), which may have a bearing on membrane structure. The absence of this region in the YFP–RHD3(1–676)PVA12(657–720) protein suggests a function of the C-terminal cytosolic domain of RHD3 in ER shaping. In either case, our results highlight a role for the C-terminal region of RHD3 in ER shaping, in addition to its function as an ER membrane anchor. These data also suggest the presence of other functional domains of RHD3 in addition to the N-terminal region.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Molecular cloning

Wild-type and mutant RHD3 cDNA were amplified from cDNA of wild-type and gom8 plants, respectively. The cDNAs were sub-cloned into binary vector pFGC5941 under the control of the CaMV 35S promoter (McGinnis et al., 2005). pFGC5941 was obtained from the Arabidopsis Biological Resource Center and modified by excision of the CHSA intron that is generally necessary for generation of RNAi constructs. We used the multiple cloning site to sub-clone the cDNA. Constructs were confirmed by sequencing. The sequences of primers used in this work are available upon request.

RNA extraction and PCR analysis

RNA extraction was performed using an RNeasy plant mini kit (Qiagen, http://www.qiagen.com/). Reverse transcription experiments were performed using a Superscript III first-strand synthesis kit (Invitrogen, http://www.invitrogen.com/). PCR experiments were performed as described previously (Marti et al., 2010). Real-time quantitative RT-PCR was performed by SYBR Green detection in triplicate using an Applied Biosystems 7500 real-time PCR system (http://www.appliedbiosystems.com/), as described previously (Chen and Brandizzi, 2011). Data were analyzed by the inline image method. The transcript level was normalized to that of isopentenyl pyrophosphate (IPP2) for each sample. For RHD3 and RHD3-Like 2, the transcript level is expressed as the fold change (mean ± SD) in gom8 relative to wild-type (set to a value of 1).

Plant materials and growth conditions

We used wild-type Arabidopsis thaliana (ecotypes Col-0 and Ler) and a homozygous transgenic Arabidopsis line (ecotype Col-0) expressing ST–GFP. The Ler ecotype was used exclusively for mapping. With the exception of the plants used for the BiFC experiments and those illustrated in Figure S12, in which we used tobacco leaf epidermal cells, the plants used in this work were all stable Arabidopsis transformants at 12 days post-germination obtained either through floral crosses or by the floral-dip method (Clough and Bent, 1998), and subsequent selection on MS medium supplemented with Gamborg’s B5 vitamins, 1% w/v sucrose, the appropriate antibiotics and 0.8% w/v agar. Seeds were surface-sterilized and germinated at 21°C under 16 h light/8 h dark conditions. Transient expression in tobacco leaf epidermis was performed as described previously (Batoko et al., 2000) using Agrobacterium tumefaciens (OD600 = 0.05) containing the binary vectors.

Isolation of the gom8 mutant and genetic analyses

M1 and M2 ST–GFP seeds were prepared as described previously (Faso et al., 2009). Sixty seeds from each M2 line were analyzed using confocal microscopy to identify mutants with apparent reduced velocity of the Golgi stacks. The gom8 mutant was crossed with Ler to generate a mapping population. Mapping was performed using a GeneChip Arabidopsis ATH1 genome array using 80 F2 individuals with the aberrant phenotype and 80 F2 plants with a wild-type phenotype. From each plant, a leaf disc (0.60 mm diameter) was collected using a hole punch. The samples were processed for genomic DNA extraction in groups of five seedlings each. The genomic DNA obtained from each sample was quantified according to procedures described previously (Borevitz, 2006) Identification of SNPs in the mapped region was achieved by deep genome-wide sequencing, as described previously (Marti et al., 2010), using an Illumina Genome Analyzer II (Bentley et al., 2008).

Confocal laser scanning microscopy

An inverted laser scanning confocal microscope (LSM510 META; Zeiss, http://www.zeiss.com/) was used for confocal analyses. The fluorescent proteins used in this work were GFP5 (Haseloff et al., 1997), eYFP (Clontech, http://www.clontech.com/), Venus (Nagai et al., 2002) (for BiFC experiments only) and monomeric RFP (Campbell et al., 2002). Imaging settings for single fluorochromes or combinations were as described previously (Brandizzi et al., 2002a; Hanton et al., 2007; Faso et al., 2009). Images shown for co-localization microscopy experiments are representative of at least five independent experiments. To analyze the velocity of the Golgi stacks, 50 time-lapse sequences with 90 frames for each sequence in the cortical region (3.0–5.0 μm depth) of cotyledon pavement cells were recorded. Time-lapse frames at a 512 × 512 pixel resolution were captured using a 2 μm pinhole at low laser power (i.e. 10% of an argon 488 nm laser line) to avoid photobleaching, and 3 digital zoom using an EC Plan-Neofluar 40 ×/1.30 objective. Velocity was calculated in post-acquisition using plug-ins available through Image J version 1.45k (http://rsbweb.nih.gov/) using the manual tracking plug-in, and data were confirmed using Image Pro-Plus 6 (Media Cybernetics). Velocity values were calculated by averaging the velocity of all the Golgi stacks in the 90 512 × 512 frames in each time-lapse sequence; then mean values were calculated as the mean of the velocities estimated in the 50 time-lapse sequences. Maximal velocity values were estimated by averaging maximal data values of each time-lapse sequence for each sample. Over 3000 Golgi stacks in the gom8 mutant and the control were used in the experiment presented in Figure 1. To evaluate ER streaming, 25 time-lapse sequences at 100 frames per sequence in the cortical ER were taken for each sample at a 512 × 512 pixel resolution, using a 2 μm pinhole at low laser power (i.e. 5–10% of an argon 514 nm laser line) and 3 digital zoom using a Plan-Apochromat 63 x/1.40 objective. ER streaming in ER–YK (control) and the gom8/ER–YK mutant were analyzed using the KbiFlow plug-in available through Image J. Mean values for ER streaming were then calculated as the mean for ER streaming estimated in 25 time-lapse sequences. Maximal velocity values for ER streaming were estimated by averaging maximal data values of each time-lapse sequence for each sample. To measure polygonal areas circumscribed by ER tubules, 23 cells with a total of 1400 areas were measured using Image J. Statistical analysis was based on a two-tailed Student’s t test assuming equal variances for FRAP data, as described previously (Brandizzi et al., 2002b). Photoshop (http://www.adobe.com/) was used for further image handling.

Recombinant protein production

Escherichia coli BL21(DE3) cells transformed with His-RHD3(1–676)-pET28b were grown in Luria broth + kanamycin at 37°C. Cells from a pre-culture grown overnight in superbroth (SB) + kanamycin were used to seed 4 liters of SB + kanamycin at an initial OD of approximately 0.4. The cells were grown in a baffled flask on an orbital shaker (200 rpm) in a 37°C incubator (New Brunswick). When the culture reached an OD of approximately 0.8, isopropyl β-d-1-thiogalactopyranoside (IPTG) was added to a final concentration of 0.5 mm to induce protein expression. After 4 h of growth under expression conditions. The cells were washed with cold buffer (25 mm HEPES, 200 mm KCl) and centrifuged at 17 500 g at 4°C for 5 min. The washed cells were broken in 40 ml of breaking buffer [25 mm HEPES/KOH. pH 7.4, 400 mm KCl, 10% glycerol, 2 mmβ-mercaptoethanol and complete EDTA-free proteinase inhibitor (Roche, http://www.roche.com)] by passage through an Emulsiflex C-3 high-pressure homogenizer (Avestin) twice at 15 000–20 000 psi. The cell extract was cleared by centrifugation at 186 000 g at 4°C for 1 h, and then passed through a 0.45 μm filter (Whatman). The His-tagged protein was purified from the cell extract by Ni-affinity chromatography using a 1 ml HiTrap chelating HP column on an ÄKTAprime protein purification system (GE Healthcare, http://www.gehealthcare.com) in a 4°C environmental chamber. The protein on the column was washed with 20 column volumes of wash buffer (25 mm HEPES/KOH, pH 7.4, 100 mm KCl, 10% glycerol, 2 mmβ-mercaptoethanol and 20 mm imidazole), after which 1 ml elution fractions were collected over a 20–500mm imidazole gradient. Elution fractions were analyzed by SDS–PAGE, and peak fractions were pooled and stored at −80°C. Protein concentrations were determined by amido black protein assay.

GTPase activity assay

GTPase activity measurements were performed using an EnzChek® phosphate assay kit (Molecular Probes) to measure the rate of release of inorganic phosphate during GTP hydrolysis. Recombinant cytosolic domains of Drosophila atlastin or Arabidopsis RHD3 [RHD3(1–676)] were mixed in a 100 μl reaction volume with 1 U ml−1 purine nucleoside phosphorylase, 200 μm 2-amino-6-mercapto-7-methylpurine riboside and 0.5 mm GTP in a transparent 96-well half-area plate (Greiner). The plates were warmed to 37°C in an Infinite M200 microplate reader (Tecan), and 5 mm Mg2+ was added to start the reaction. In the presence of inorganic phosphate, purine nucleoside phosphorylase catalyzes the conversion of 2-amino-6-mercapto-7-methylpurine riboside to ribose 1-phosphate and 2-amino-6-mercapto-7-methylpurine, resulting in a spectrophotometric shift in absorbance from 330 to 360 nm. The absorbance increase at 360 nm was measured in real-time approximately every 20 sec over a 20 min period. Absorbance was normalized to phosphate standards, and initial rates were calculated.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

We thank Eileen Morey for editing the manuscript. We thank Andrew Goulet, Alyssa Burkhardt and Frederique Breuer for technical assistance. We are grateful to Professor A. Nebenführ (University of Tennessee) for the gift of the GA–YK and ER–YK seeds, and the Arabidopsis Biological Resource Center for the rhd3-7 line. We acknowledge support by the Chemical Sciences, Geosciences and Biosciences Division, Office of Basic Energy Sciences, Office of Science, US Department of Energy (award number DE-FG02-91ER20021) and the US National Science Foundation (MCB 0948584) (to F.B.).

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Figure S1. Golgi aggregates are visible in roots and root hairs.

Figure S2. The gom8 mutation affects Golgi stacks.

Figure S3. gom8 ER strands underlie Golgi aggregates.

Figure S4. GTPase activity measurements for RHD3.

Figure S5. GTPase activity of RHD3(1–676) in the presence of GTP, GDP and ATP.

Figure S6. FRAP results on Golgi bodies in non-mutagenized Col-0 and the gom8 expressing ST–GFP.

Figure S7. Characterization of rhd3-7.

Figure S8. Aberrant ER phenotype in rhd3-7 transformed with GFP–CX.

Figure S9. RT-PCR on independent gom8/ST–GFP/ER–YK lines.

Figure S10. BiFC assay in tobacco leaf epidermal cells expressing YFP fused to various RHD3 proteins.

Figure S11. Localization analyses of YFP–RHD3(P701S).

Figure S12. Subcellular localization of RHD3 cytosolic region RHD3(1–676).

Movie S1. Golgi move in and out the gom8 Golgi aggregates.

Movie S2. The wild-type ER is intact at the cell periphery and strands, which are formed dynamically in the inner ER.

Movie S3. The gom8 ER is characterized by enlarged and unbranched strands in the inner ER.

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer-reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

FilenameFormatSizeDescription
TPJ_4846_sm_FigS1.tif24678KSupporting info item
TPJ_4846_sm_FigS2.tif24682KSupporting info item
TPJ_4846_sm_FigS3.tif24682KSupporting info item
TPJ_4846_sm_FigS4.tif24678KSupporting info item
TPJ_4846_sm_FigS5.tif24679KSupporting info item
TPJ_4846_sm_FigS6.tif24683KSupporting info item
TPJ_4846_sm_FigS7.tif24695KSupporting info item
TPJ_4846_sm_FigS8.tif24685KSupporting info item
TPJ_4846_sm_FigS9.tif24681KSupporting info item
TPJ_4846_sm_FigS10.tif24684KSupporting info item
TPJ_4846_sm_FigS11.tif24685KSupporting info item
TPJ_4846_sm_FigS12.tif24684KSupporting info item
TPJ_4846_sm_MovieS1.mov4640KSupporting info item
TPJ_4846_sm_MovieS2.mov2463KSupporting info item
TPJ_4846_sm_MovieS3.mov14957KSupporting info item
TPJ_4846_sm_SupportingInformationLegends.doc39KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.