Dynamic organization of COPII coat proteins at endoplasmic reticulum export sites in plant cells

Authors

  • Sally L. Hanton,

    1. Department of Biology, 112 Science Place, University of Saskatchewan, Saskatoon, SK S7N 5E2, Canada
    Search for more papers by this author
  • Loren A. Matheson,

    1. Department of Biology, 112 Science Place, University of Saskatchewan, Saskatoon, SK S7N 5E2, Canada
    Search for more papers by this author
  • Laurent Chatre,

    1. Department of Biology, 112 Science Place, University of Saskatchewan, Saskatoon, SK S7N 5E2, Canada
    Search for more papers by this author
  • Federica Brandizzi

    Corresponding author
    1. Department of Biology, 112 Science Place, University of Saskatchewan, Saskatoon, SK S7N 5E2, Canada
    2. Plant Research Laboratory, Department of Energy, Michigan State University, East Lansing, MI 48824, USA
      (fax 1 517 353 9168; e-mail brandizz@msu.edu).
    Search for more papers by this author

(fax 1 517 353 9168; e-mail brandizz@msu.edu).

Summary

Protein export from the endoplasmic reticulum (ER) is mediated by the accumulation of COPII proteins such as Sar1, Sec23/24 and Sec13/31 at specialized ER export sites (ERES). Although the distribution of COPII components in mammalian and yeast systems is established, a unified model of ERES dynamics has yet to be presented in plants. To investigate this, we have followed the dynamics of fluorescent fusions to inner and outer components of the coat, AtSec24 and AtSec13, in three different plant model systems: tobacco and Arabidopsis leaf epidermis, as well as tobacco BY-2 suspension cells. In leaves, AtSec24 accumulated at Golgi-associated ERES, whereas AtSec13 showed higher levels of cytosolic staining compared with AtSec24. However, in BY-2 cells, both AtSec13 and AtSec24 labelled Golgi-associated ERES, along with AtSec24. To correlate the distribution of the COPII coat with the dynamics of organelle movement, quantitative live-cell imaging analyses demonstrated that AtSec24 and AtSec13 maintained a constant association with Golgi-associated ERES, irrespective of their velocity. However, recruitment of AtSec24 and AtSec13 to ERES, as well as the number of ERES marked by these proteins, was influenced by export of membrane cargo proteins from the ER to the Golgi. Additionally, the increased availability of AtSec24 affected the distribution of AtSec13, inducing recruitment of this outer COPII coat component to ERES. These results provide a model that, in plants, protein export from the ER occurs via sequential recruitment of inner and outer COPII components to form transport intermediates at mobile, Golgi-associated ERES.

Introduction

Secretory proteins leave the endoplasmic reticulum (ER) from specialized subdomains called ER export sites (ERES). The essential components of the COPII machinery that mediates ER export are highly conserved across kingdoms (reviewed by Hanton et al., 2005); however, recent evidence suggests that fundamental differences exist in the subcellular distributions of ERES (Hanton et al., 2006; Ward and Brandizzi, 2004), which may reflect evolutionary patterns of cell differentiation. For example, the Golgi apparatus of budding yeast is dispersed, and soluble COPII proteins are present throughout the cytoplasm, whereas Sec12 is distributed throughout the ER (Rossanese et al., 1999). In contrast, the Golgi stacks in the yeast Pichia pastoris are immobile, coherent and adjacent to discrete ERES containing COPII coat proteins and Sec12 (Rossanese et al., 1999). In mammals, ERES are relatively immobile punctae labelled by COPII coat components (Ward et al., 2001). They are separated from the Golgi apparatus by the ER-Golgi intermediate compartment, which has no known counterpart in plant cells.

In plants, although there is a general consensus that Sec12 is distributed throughout the ER (daSilva et al., 2004; Yang et al., 2005), it is still unclear where the soluble COPII components are localized (Robinson et al., 2007; Staehelin and Kang, 2008). Plant ERES were initially visualized in tobacco leaf epidermal cells by live-cell imaging of a yellow fluorescent protein fusion to a tobacco Sar1 isoform (NtSar1-YFP, daSilva et al., 2004). When co-expressed with fluorescent Golgi membrane markers, NtSar1-YFP was found to label motile regions of the ER that moved in synchrony with Golgi bodies. NtSar1-YFP cycled on and off these regions, whereas Golgi cargo molecules cycled between the ER and the Golgi. Both processes appeared to occur continuously, forming the basis for the ‘secretory unit’ model, whereby cargo proteins continually travel between the ER and motile Golgi stacks (daSilva et al., 2004). Further support for this model was provided by the visualization of fluorescent fusions of Sec23, Sec24 and two isoforms of Sar1 from Arabidopsis thaliana, in the leaf epidermis of tobacco and Arabidopsis (Hanton et al., 2008; Matheson et al., 2006; Stefano et al., 2006). These markers labelled ERES in the absence of Golgi cargo; however, the ERES labelled with NtSar1-YFP, AtSar1-YFP, YFP-AtSec24 or YFP-AtSec23 were all found in continuous association with Golgi bodies, when Golgi membrane markers were co-expressed (Hanton et al., 2008; daSilva et al., 2004; Stefano et al., 2006). A study in tobacco BY-2 suspension cells presented an alternative scenario (Yang et al., 2005). Using a fluorescent fusion of an outer COPII coat protein, Sec13 from tomato (LeSec13-GFP), it was suggested that ERES in BY-2 cells largely outnumbered the Golgi bodies, and that Golgi bodies were not continuously associated with structures labelled by LeSec13-GFP (Yang et al., 2005). This led to the proposal of the ‘kiss-and-run’ model, which postulated that protein transport from the ER to the Golgi occurs in a discontinuous manner (Yang et al., 2005). Further variations between the two systems were highlighted by the finding that the distributions of NtSar1 and AtSec23/24 were sensitive to inhibition of ER/Golgi protein traffic mediated by the drug brefeldin A (BFA) in tobacco leaf epidermal cells (daSilva et al., 2004; Stefano et al., 2006). In contrast, the distribution of LeSec13 in BY-2 cells was largely unaffected by BFA (Yang et al., 2005).

The reported variation in the behaviour of the COPII components in plant cells raises fundamental biological questions. It is possible that the distributions of ERES in BY-2 cells and leaf tissue differ noticeably, suggesting the existence of essential differences in the organization of ER subdomains between systems. Alternatively, the inner COPII components may have different dynamics from those that assemble at the outside of the coat. Finally, it is conceivable that plant Sec13 isoforms perform other cellular roles in addition to ER export. To explore these possibilities, we considered that a direct comparison between the systems and marker proteins was necessary. Establishing the distribution of the inner and outer COPII components would also provide us with tools that could be used to follow the dynamics of the coat at ERES in living plant cells. We therefore compared the distributions of fluorescent fusions with Sec24 and Sec13 from A. thaliana in tobacco and Arabidopsis leaf epidermis, as well as in BY-2 suspension cells. We found that in all three systems ERES labelled with YFP-AtSec24 were in continuous association with Golgi bodies, suggesting that the distribution of this marker is common across plant systems. In contrast, AtSec13-YFP had a more pronounced cytosolic distribution in leaf tissue, although in BY-2 cells it labelled well-defined punctate structures that corresponded to those labelled by YFP-AtSec24 at the peri-Golgi area. Treatment with BFA disrupted the punctate labelling patterns of both YFP-AtSec24 and AtSec13-YFP. These observations indicate a considerable level of conservation in the distribution and dynamics of the COPII coat at ERES in different plant expression systems, although AtSec24 appears to be a more consistent ERES marker than AtSec13. Finally, quantitative live-cell imaging analyses of COPII assembly in BY-2 cells demonstrated that the recruitment of both AtSec13 and AtSec24 to Golgi-associated ERES is influenced by the export of membrane cargo proteins to the Golgi, and that increased availability of AtSec24 led to the recruitment of AtSec13 to Golgi-associated ERES.

These results not only support the model that, in plants, ERES are subdomains of the ER at the peri-Golgi area, where the COPII coat machinery is assembled in response to the need for export of proteins to the Golgi apparatus, but also provide new in vivo evidence of the formation of COPII carriers by the sequential assembly of COPII coat components at the ERES.

Results

YFP-AtSec24 labels Golgi-associated ERES in tobacco and Arabidopsis leaf epidermal cells, as well as in tobacco BY-2 suspension cells

A yellow fluorescent protein (YFP) fusion to AtSec24 (YFP-AtSec24; Stefano et al., 2006) was chosen as an inner COPII marker, as it retains the ability to bind AtSec23, and localizes at ERES without affecting protein export (Hanton et al., 2007; Stefano et al., 2006). YFP-AtSec24 was found to localize in the cytosol and on punctate structures in tobacco (Figure 1a) and Arabidopsis leaf epidermal cells (Figure 1e). Co-expression of YFP-AtSec24 with ERD2-GFP, an established ER and Golgi marker in leaves (Boevink et al., 1998), confirmed that these structures were associated with the Golgi area in both systems (Figure 1b–d, f–h), corresponding to previously reported distributions (Hanton et al., 2007; Matheson et al., 2006; Stefano et al., 2006).

Figure 1.

 YFP-AtSec24 labels endoplasmic reticulum export sites (ERES) in tobacco and Arabidopsis leaf epidermal cells, and in tobacco BY-2 suspension cells.
(a) Arabidopsis YFP-Sec24 (Stefano et al., 2006) labels the cytosol and punctate structures (arrows and inset) in tobacco leaf cells.
(b–d) The punctate structures labelled by YFP-AtSec24 (b) localize at the peri-Golgi area labelled with ERD2-GFP (Boevink et al., 1998; c, arrows and insets). (d) Merged image of (b) and (c).
(e) A similar labelling pattern was observed on the expression of YFP-AtSec24 in Arabidopsis leaf cells. Punctate structures labelled by YFP-AtSec24 are indicated by arrows, and an enlarged view is shown in the inset.
(f–h) Co-expression of YFP-AtSec24 (f) with ERD2-GFP (g) shows that YFP-AtSec24 accumulates at the peri-Golgi area in Arabidopsis leaf cells (arrows and insets). (h) Merged image of (f) and (g).
(i) BY-2 suspension cells transformed with YFP-AtSec24 show fluorescence distributed at punctate structures (arrows and inset), and in the cytosol.
(j–l) Co-transformation of YFP-AtSec24 (j) with GmMan1-GFP (k, Nebenführ et al., 1999) demonstrates that the punctate structures labelled by YFP-AtSec24 are found at the peri-Golgi area (arrows and insets). (l) Merged image of (j) and (k).
Scale bars: 5 μm.

To determine whether the YFP-AtSec24 distribution at ERES is specific to leaves, irrespective of the species used, or whether the marker behaves in the same manner in other plant systems, we generated stably transformed BY-2 cells with YFP-AtSec24, either with or without α-1,2-mannosidase I-GFP (GmMan1-GFP), an established Golgi marker in BY-2 cells (Nebenführ et al., 1999). In cells expressing YFP-AtSec24 alone, the marker had a similar distribution to that in leaf tissue, labelling both punctate structures and cytosol (Figure 1i). In cells transformed with both YFP-AtSec24 and GmMan1-GFP, the punctate structures labelled by YFP-AtSec24 localized in the vicinity of the Golgi (Figure 1j–l). Comparable results were obtained in BY-2 cells stably co-expressing YFP-Sec24 and ERD2-GFP (Figure S1a–c). Similar to tobacco and Arabidopsis leaf epidermal cells (Stefano et al., 2006; Figure S1g–i), YFP-AtSec24-labelled structures that were not associated with the Golgi area were observed very infrequently (Figure S1d–f). These structures occurred only in a very small proportion of the cells imaged, unlike those reported by Yang et al. (2005).

These data show that in the three plant expression systems analysed, YFP-AtSec24 is consistently distributed at ERES that are in association with Golgi bodies.

Arabidopsis Sec13 shows a predominantly cytosolic distribution in tobacco and Arabidopsis leaf epidermis

Having established that AtSec24 labels punctate, Golgi-associated structures in all three systems tested, we wanted to investigate the distribution of the outer COPII component Sec13. To this end, we used a YFP fusion to AtSec13 (AtSec13-YFP) as a marker. We first transformed tobacco leaf epidermal cells with AtSec13-YFP either alone or with ERD2-GFP (Figure 2a–d). When expressed alone, AtSec13-YFP appeared to be mainly cytosolic (Figure 2a); some punctate structures were present (Figure 2a, arrows), but these were not clearly defined. Despite a low signal to noise ratio, these punctate structures appeared to be in the vicinity of the Golgi (Figure 2b–d, arrows).

Figure 2.

 The AtSec13-YFP distribution exhibits variations in distribution between expression systems.
(a) AtSec13-YFP is predominantly cytosolic when expressed in tobacco leaf epidermis, although some diffuse punctate structures are visible (arrows).
(b–d) The diffuse punctate structures labelled by AtSec13-YFP (b) mostly localize in the peri-Golgi area (arrows), labelled by ERD2-GFP (c). (d) Merged image of (b) and (c).
(e) Punctate structures labelled by AtSec13-YFP (arrows) are more apparent in Arabidopsis leaf epidermis than in tobacco leaves.
(f–h) Co-expression of AtSec13-YFP (f) with ERD2-GFP (g) demonstrates that the punctate structures labelled by AtSec13-YFP (h, arrows, insets) are found at the peri-Golgi area. (h) Merged image of (f) and (g).
(i) AtSec13-YFP labels defined punctate structures (arrows) in BY-2 cells. Fluorescence was also observed in the nucleoplasm (arrowhead), as is common for small cytosolic fusions to fluorescent proteins (Brandizzi et al., 2003; Haseloff et al., 1997; daSilva et al., 2004).
(j–l) Co-expression of AtSec13-YFP (j) with GmMan1-GFP (k) demonstrates that the punctate structures labelled by AtSec13-YFP co-localize with the Golgi marker (l, merged image, arrowheads and insets).
Scale bars: 5 μm.

To observe the behaviour of AtSec13-YFP in an endogenous context, we next wanted to establish its distribution in Arabidopsis cells. Figure 2e shows that the fluorescence pattern of AtSec13-YFP in Arabidopsis leaf epidermal cells is similar to that in tobacco leaves. Upon co-expression with ERD2-GFP, more defined AtSec13-YFP-labelled punctate structures were visible in the area of the Golgi (Figure 2f–h, arrows), although these were less distinct than those labelled by YFP-AtSec24 in the same expression system (Figure 1e–h).

To compare the distribution patterns of AtSec13-YFP in an additional cell system, we generated stably transformed BY-2 cell lines, with or without GmMan1-GFP (Figure 2i–l). Expression of AtSec13-YFP in BY-2 suspension cells gave a more distinct pattern of fluorescence than those observed in leaf tissue. Clearly defined punctate structures were visible (Figure 2i, arrows) in addition to cytosolic fluorescence. On co-expression with GmMan1-GFP (Figure 2j–l), it was clear that the punctate structures labelled by AtSec13-YFP were found in the Golgi area (Figure 2j–l, arrows), similar to those labelled by YFP-AtSec24 (Figure 1j–l). Co-expression of fluorescent fusions of AtSec24 and AtSec13 in the same cells further verified these data (Figure S2), confirming that the cytosolic structures labelled by each marker have overlapping distributions.

Taken together, our data indicate that the punctate appearance of AtSec13 varies, to a certain extent, depending on the expression system used. However, the fluorescent protein fusion accumulated at the peri-Golgi area in both tobacco BY-2 cells and in the endogenous Arabidopsis system, in a more defined manner than was observed in tobacco leaf epidermal cells. The finding that the YFP-AtSec24 distribution was comparable across the three systems (Figure 1 and Figures S1 and S2) suggests that the observed behaviour of AtSec13-YFP (Figure 2 and Figure S2) may depend on specific attributes of AtSec13.

ERES in BY-2 cells expressing either YFP-AtSec24 or AtSec13-YFP are sensitive to the inhibition of secretion mediated by BFA

Having established the distributions of YFP-AtSec24 and AtSec13-YFP, we next wanted to test whether the localization of COPII components at ERES in BY-2 cells would be affected by the disruption of the retrograde route for protein transport from the Golgi to the ER. This has been shown to be the case for YFP-AtSec24-, YFP-AtSec23- and NtSar1-YFP-labelled ERES in tobacco leaves (daSilva et al., 2004; Stefano et al., 2006), but not for LeSec13-GFP in BY-2 cells (Yang et al., 2005). We used BY-2 suspension cells as an expression system, as the distribution of AtSec13-YFP is more distinct than that in leaves. We treated cells with the drug BFA, which inhibits COPI-mediated retrograde transport and ultimately disassembles the Golgi apparatus (Driouich et al., 1993; Saint-Jore et al., 2002). We used cells expressing either YFP-AtSec24 or AtSec13-YFP, with GmMan1-GFP as an internal control for the effect of BFA on the Golgi. To gain further insights into the behaviour of ERES during the BFA treatment, we examined the distributions of the two COPII markers in cells at different time points throughout the treatment, after 10–20, 25–35 and 40–50 min of incubation with the drug (Figure 3).

Figure 3.

 Treatment of BY-2 cells with brefeldin A (BFA) causes the redistribution of endoplasmic reticulum export sites (ERES) markers to the cytosol.
Images of BY-2 cells expressing GmMan1-GFP with either YFP-AtSec24 (a) or AtSec13-YFP (b) at different time points during incubation with 100 μg ml−1 BFA (Stefano et al., 2006). In both cases, as the Golgi marker is redistributed into the ER, the YFP fluorescence of the ERES markers shows a concomitant release from ERES to the cytosol.
Scale bars: 5 μm.

Treatment of BY-2 cells with BFA initially resulted in the clumping of Golgi bodies and associated ERES (Figure 3, 10–20 min, arrows). At the 25–35-min time point, a partial redistribution of the Golgi marker to the ER, and ERES markers to the cytosol, was evident, although Golgi and associated ERES clumps remained visible. After 40–50 min, the majority of the Golgi bodies had disappeared. Similarly, both AtSec13-YFP and YFP-AtSec24 appeared to be released into the cytosol within 40–50 min of treatment.

These data not only confirm that BFA-induced disruption of retrograde transport from the Golgi to the ER affects the integrity of the Golgi membranes in BY-2 cells (Ritzenthaler et al., 2002; Saint-Jore et al., 2002), but also indicate that the distributions of both YFP-AtSec24 and AtSec13-YFP at ERES depend on active ER–Golgi protein exchange in this system.

Association of COPII proteins with ERES is independent of the velocity of the ERES-Golgi units

In tobacco leaf epidermal cells, fluorescent protein fusions to NtSar1, AtSar1, AtSec23 and AtSec24 label mobile ERES that associate continuously with the Golgi (Hanton et al., 2007, 2008; Robinson et al., 2007; daSilva et al., 2004; Stefano et al., 2006). However, it has been suggested that the speed of movement of Golgi bodies might influence the level of association of ERES with the Golgi, as shown in BY-2 cells co-expressing LeSec13-GFP and GmMan1-RFP (Yang et al., 2005). We wanted to determine whether the association of YFP-AtSec24 and AtSec13-YFP with the peri-Golgi area would, like that of LeSec13-GFP, vary depending on the speed of Golgi movement in BY-2 cells. We therefore measured the velocities of Golgi bodies in cells expressing either YFP-AtSec24 or AtSec13-YFP with GmMan1-GFP. We then quantified the level of YFP fluorescence associated with each Golgi body by measuring the fluorescence intensity within a circle of diameter 2.5 μm centred on each Golgi stack at each step of the analysis, as described by Yang et al. (2005). This allowed us to compare the distributions of the two ERES markers, YFP-AtSec24 and AtSec13-YFP, with respect to both moving and relatively stationary Golgi bodies. The velocity (in μm sec−1) of each Golgi body analysed was plotted against the fluorescence intensity (in arbitrary units) of the ERES marker fusions for each time point (Figure 4). In cells expressing either of the ERES markers, we found that a single Golgi body could travel at several different velocities, but that the level of YFP fluorescence intensity associated with that Golgi body remained constant, independent of the speed of movement (Figure 4). These data verify that, in BY-2 suspension cultures, the movement of YFP-AtSec24- or AtSec13-YFP-labelled structures is synchronized with that of Golgi bodies.

Figure 4.

 The YFP fluorescence intensity in the peri-Golgi area is unaffected by the velocity of movement.
The velocities (in μm sec−1) of five independent Golgi bodies for either YFP-AtSec24 (a) or AtSec13-YFP (b) were measured and plotted against the fluorescence intensities (in arbitrary units) within a circle of diameter 2.5 μm, centred on the Golgi stack of interest, at each step of the analysis. Each different symbol type represents a different Golgi body.

Sec24 induces recruitment of Sec13 to ERES

We next aimed to determine whether a spatially-regulated sequence of recruitment events leads to the formation of ERES and to the subsequent export of cargo from them. Our hypothesis is that ER export begins at the ER membrane with the activation of Sar1 by Sec12, leading to the recruitment of the structural components Sec24, and ultimately Sec13, to ERES. Sec12 is found throughout the ER network (daSilva et al., 2004), whereas Sar1b labels the ER membranes as well as ERES (Hanton et al., 2008). Sec24 and Sec13 accumulate at ERES and in the cytosol (Stefano et al., 2006; this work), although the ERES labelling of Sec24 is clearer than that of Sec13 (Figures 1 and 2). YFP-Sec24 appeared partially redistributed to the ER when co-expressed with Sar1b (Figure 5), most likely as a result of high levels of expression of the GTPase. These data suggested that elements of the COPII coat may influence other COPII components for the assembly of the coat. To verify this hypothesis with functional data, we next aimed to test whether the recruitment of Sec13 to ERES was dependent on Sec24. Based on the sequential recruitment of COPII coat components established in vitro (Barlowe, 2002; Bickford et al., 2004), we hypothesized that expression of YFP-AtSec24, as an inner component of the COPII coat, would increase recruitment of AtSec13-CFP, an outer component, to complete the coat. To test this we used BY-2 cells, where the punctate labelling pattern of AtSec13-CFP was more clearly visualized than in the leaves. Following the quantification protocol described by Hanton et al. (2007), we measured the fluorescence intensity in a circle centred on an ERES structure and in an identical circle in the cytosol adjacent to the ERES, and then calculated the ratio of fluorescence intensity at ERES to that in the cytosol (see also Experimental procedures). Similar measurements and calculations were carried out on cells expressing each marker alone or in combination (Figure 6a). We observed no change in the fluorescence intensity ratio for YFP-AtSec24 on co-expression of AtSec13-CFP (= 0.975). However, a significant increase was observed in the fluorescence intensity ratio for AtSec13-CFP when co-expressed with YFP-AtSec24 (= 2.45 × 10−12). These data confirm our prediction that the increased availability of AtSec24 induces recruitment of AtSec13 to ERES. The specificity of this effect is shown by the unchanged fluorescence intensity of YFP-AtSec24 in cells co-expressing AtSec13-CFP (Figure 6a).

Figure 5.

 AtSar1b labels endoplasmic reticulum export sites (ERES) as well as ER membranes, and its overexpression induces a partial redistribution of Sec24 to the ER.
(a–c) Co-expression of AtSar1b-CFP (a) with YFP-AtSec24 (b) confirms that the punctate structures (arrows) labelled by AtSar1b correspond to those labelled by AtSec24. YFP-AtSec24 shows a closer association with the ER membranes (arrowhead) than when expressed alone (compare with Figure 1a). (c) Merged image of (a) and (b).
Scale bars: 5 μm.

Figure 6.

 Support for the sequential recruitment of COPII.
(a) Ratio of CFP or YFP fluorescence intensity at endoplasmic reticulum export sites (ERES) to that in the cytosol in cells expressing AtSec13-CFP or YFP-AtSec24 alone (white bars), or in combination with the other ERES marker (grey bars). Error bars represent the SEMs for 200 samples.
(b) Quantification of the number of ERES per 100 μm3 labelled by AtSec13-CFP or YFP-AtSec24 in the absence (white bars) or presence (grey bars) of the other ERES marker. Error bars represent the SEMs for 20 samples.
(c) Ratio of YFP fluorescence intensity at ERES to that in the cytosol in cells expressing YFP-AtSec24 or AtSec13-YFP alone (white bars), or in combination with the Golgi membrane cargo protein GmMan1-GFP (grey bars). Error bars represent the SEMs for 200 samples.
(d) Quantification of the number of ERES per 100 μm3 labelled by YFP-AtSec24 or AtSec13-YFP in the absence (white bars) or presence (grey bars) of the Golgi membrane cargo protein GmMan1-GFP. Error bars represent the SEMs for 20 samples.

As co-expression of Golgi and ERES markers has been shown to increase the number of ERES in plant cells (Hanton et al., 2007), we wanted to investigate whether co-expression of two ERES markers would affect the de novo formation of ERES. We therefore quantified the number of ERES per 100 μm3 for YFP-AtSec24 and AtSec13-CFP in cells expressing each marker alone or in combination with the other (Figure 6b). No significant difference was observed for either marker (= 0.999 for YFP-AtSec24; = 0.984 for AtSec13-CFP), and the numbers of ERES were comparable between the two markers. These results suggest that the induction of ERES formation is an event that is specific to increased availability of membrane cargo.

The differentiation of ERES in BY-2 cells is stimulated by the presence of membrane cargo proteins

To test the sequential recruitment model further, we aimed to define the behaviour of AtSec13-YFP in cells co-expressing Golgi cargo. We postulated that if AtSec13-YFP is incorporated into functional COPII coats, its distribution would be altered in cells co-expressing it with Golgi cargo, compared with those expressing AtSec13-YFP alone. This hypothesis was based on the recruitment of YFP-AtSec24 to ERES by export-competent Golgi membrane cargo (Hanton et al., 2007). The recruitment of Sec24 to ERES would form an incomplete COPII coat that requires Sec13 for completion and budding. Therefore, we hypothesized that Sec13 would be recruited to ERES in the presence of Golgi membrane cargo proteins. For this experiment, we chose to use both AtSec13-YFP and YFP-AtSec24 to label ERES. The use of YFP-AtSec24 would serve as a control to ensure that increased recruitment of this marker to ERES occurred in BY-2 cells in the same way as in tobacco leaf epidermal cells (Hanton et al., 2007), having the added advantage of providing more information regarding the similarities and differences between BY-2 cells and tobacco leaf epidermal cells. We therefore quantified the recruitment of AtSec13-YFP and YFP-AtSec24 to ERES in BY-2 cells expressing either of the two ERES markers, with or without GmMan1-GFP (Figure 6c). This quantification was possible because we used confocal microscope settings that avoid signal crosstalk between the two fluorophores (Figure S3; see also Experimental procedures). We observed that there was a slight but significant increase in fluorescence intensity for both ERES markers upon co-expression of GmMan1-GFP (Figure 6c; P = 8.10 × 10−5 for YFP-AtSec24; = 2.57 × 10−4 for AtSec13-YFP). These data indicate that the number of ERES labelled with either marker increased to a similar extent in response to the expression of the Golgi marker. We quantified the number of ERES per 100 μm3 in BY-2 cells expressing each marker either alone or in combination with GmMan1-GFP as a Golgi membrane cargo protein, using a similar protocol to that described by Hanton et al. (2007; see also Experimental procedures). For both YFP-AtSec24 and AtSec13-YFP there was a significant increase in the number of ERES on co-expression of GmMan1-GFP (Figure 6d; = 5.0 × 10−4 for YFP-AtSec24; = 9.0 × 10−4 for AtSec13-YFP). There was no significant difference in the numbers of ERES for YFP-AtSec24 compared with AtSec13-YFP in the absence (= 0.734) or presence (= 0.862) of GmMan1-GFP (Figure 6d).

The increased recruitment of AtSec13 from the cytosol to ERES caused by expression of ER export-competent Golgi cargo suggests that AtSec13 accumulates at ERES as part of the COPII coat complex, in response to an increased need for the secretion of membrane cargo proteins. Because of its position on the coat, Sec13 is unlikely to interact directly with cargo: the increased recruitment of Sec13 to ERES on co-expression of Golgi-destined cargo is therefore most likely secondary to the assembly of the inner COPII coat.

Discussion

COPII coat components are associated with ERES in the vicinity of the Golgi in different plant cell systems

Understanding the distribution of ERES in plant cells addresses an important biological question, examining both the kinetics of protein movement between organelles and the dynamics of organelle movement, as the plant Golgi apparatus consists of discrete stacks that are motile on the ER network (Boevink et al., 1998; Hawes and Satiat-Jeunemaitre, 2005). In this study we have compared the distributions of COPII markers in different expression systems, and tested the effects of cargo loading and availability of coat proteins on the distributions of these markers. Our major findings can be summarized as follows:

  • (i) Golgi-associated ERES labelled by YFP-AtSec24 behave in a similar manner independently of the expression system used. These findings demonstrate a considerable level of conservation in COPII-mediated protein transport from the ER to the Golgi in plant cells.
  • (ii) These results have been strengthened further by the finding that a fluorescent protein fusion of AtSec13, a component of the outer layer of the COPII coat, clearly co-localizes with YFP-AtSec24 in BY-2 cells.
  • (iii) In a similar manner to YFP-AtSec24, AtSec13-YFP is concentrated at the peri-Golgi area independently of the velocity of movement of Golgi bodies, and its punctate distribution is sensitive to the disruption of protein transport between the Golgi and ER induced by BFA.
  • (iv) Increased availability of Sec24, as well as of cargo, influences the distribution of Sec13, providing in vivo support for a model in which the inner COPII coat layer assembly affects that of the outer layer in plant cells.

Live-cell imaging demonstrates the sequential recruitment of COPII coat markers at ERES

Export from the ER is initiated by the activation of the small GTPase Sar1 by the membrane-associated GDP/GTP exchange factor (GEF), Sec12 (d’Enfert et al., 1991). Activated Sar1 recruits the heterodimer Sec23/24, and the Sec23/24-Sar1 complex is then thought to select cargo for export (Bickford et al., 2004; Miller et al., 2003; Mossessova et al., 2003; Roberg et al., 1999). The Sec13/31 unit, the most external heterodimer of the coat proteins, polymerizes into an octahedral cage and deforms the membrane into a bud (Barlowe et al., 1994; Bickford et al., 2004; Fath et al., 2007; Stagg et al., 2008). How the COPII coat assembles specifically at defined areas of the ER is unclear, but evidence indicates the involvement of membrane cargo molecules destined for the Golgi (Aridor et al., 1999; daSilva et al., 2004; Farhan et al., 2008; Guo and Linstedt, 2006; Hanton et al., 2007). Our data support this model, as we have shown that accumulation of AtSec24 and AtSec13 at Golgi-associated ERES is influenced by protein export from the ER: it is diminished during the BFA-mediated inhibition of protein export, and enhanced by increased availability of Golgi cargo proteins.

The coincident punctate distribution of AtSec13 and AtSec24 in BY-2 cells provided us with a unique opportunity to explore whether the accumulation of AtSec13 at ERES was directly influenced by Sec24, rather than by ER export competent cargo. The observed increase in AtSec13-CFP accumulation at ERES in cells co-expressing YFP-AtSec24 compared with those expressing AtSec13-CFP alone indicated that the increased availability of the inner COPII component YFP-AtSec24 induced a higher level of accumulation of AtSec13-CFP at ERES. Our data do not exclude the possibility that the endogenous Sec24 participates in the recruitment of Sec13-YFP, nor that it may be an indirect effect resulting from recruitment of other factors by Sec24. However, increased availability of Sec13 did not affect YFP-Sec24 distribution, suggesting that the inner coat assembly influences the outer coat assembly, and not vice versa. These data support the in vitro evidence that, when Sec23/24 is readily available, Sec13/31 and Sec23/24 co-assemble into structures larger than those formed in the absence of Sec23/24 (Stagg et al., 2008).

The localizations of proteins involved in the formation and function of COPII carriers also correspond to the sequential recruitment model. Sec12, the GEF that catalyses the activation of Sar1 and leads to the recruitment of other COPII components, is distributed throughout the ER in plants (daSilva et al., 2004). Sar1b from Arabidopsis associates with the ER membranes, as well as accumulating at ERES in the absence of Golgi-destined cargo (Hanton et al., 2008). This suggests that AtSar1b may interact with AtSec12 throughout the ER network, but that it accumulates at ERES for the specific purpose of cargo export. It may be that this accumulation is mediated by the presence of Golgi-destined cargo (Giraudo and Maccioni, 2003), or that the lipid composition at ERES is such that activated Sar1 is able to insert its N terminus into the membrane more easily than at other sites. Sec23/24 is thought to accumulate at ERES as a result of recruitment by active Sar1 (Aridor et al., 2001), but also through the interaction of Sec24 with export signals on cargo proteins (Hanton et al., 2007; Miller et al., 2003). Finally, the Sec13/31 complex associates with the rest of the COPII components. Interestingly, when AtSar1b-CFP was co-expressed with YFP-AtSec24, YFP-AtSec24 appeared to associate more with the ER membranes than when it was expressed alone. This may result from the increased recruitment of AtSec24 caused by overexpression of AtSar1b, similar to the effect of co-expressing AtSec24 with AtSec13 in BY-2 cells. This provides further in vivo evidence for the sequential recruitment of COPII to ERES in plant cells, as previously indicated in other systems (Barlowe et al., 1994; Bickford et al., 2004). The accumulation of export-competent membrane cargo causes the recruitment of AtSec24 to ERES (Hanton et al., 2007), although the interaction between the Sec23/24 complex and Sar1 may also play a role in recruiting structural components to the ER, followed by the subsequent quantitative recruitment of Sec13 to the same sites.

The distribution and behaviour of ERES in BY-2 cells as visualized with different markers

Our investigations have contributed to the understanding of the behaviour of ERES in BY-2 cells. A recent study showed that increased expression of membrane cargo proteins destined for the Golgi causes differentiation of new ERES, and increased recruitment of YFP-AtSec24 to ERES in tobacco leaf epidermis (Hanton et al., 2007). Our data indicate that similar effects occur in BY-2 cells, although the extent of the increase in number or fluorescence intensity is not the same as in tobacco leaves. This may result from differences between the systems, or from the use of GmMan1-GFP as a Golgi marker; Hanton et al. (2007) demonstrated that different Golgi marker proteins have different effects on the recruitment of YFP-AtSec24 to ERES. The level of fluorescence of ERES labelled by YFP-Sec24 in the absence of the Golgi marker also varies between expression systems (Hanton et al., 2007): this may result in part from the higher level of secretory activity of BY-2 cells compared with tobacco leaf epidermal cells (Yang et al., 2005).

It appears that the ERES punctae are sensitive to the inhibition of protein transport at the ER–Golgi interface, as demonstrated by treatment with BFA. BFA causes a redistribution of YFP-AtSec24 to the cytoplasm in tobacco leaves (Stefano et al., 2006), most likely through the indirect inhibition of the formation of anterograde carriers following the disruption of cycling of membrane components from the Golgi apparatus to the ER. We found that BFA has a similar effect in BY-2 cells, whereby the Golgi marker GmMan1-GFP was redistributed to the ER (see also Ritzenthaler et al., 2002), and the ERES-associated YFP-AtSec24 or AtSec13-YFP became cytosolic. These data provide further evidence for the similarity of behaviour of ERES labelled by YFP-AtSec24 in tobacco BY-2 cells and leaves.

The data we have obtained from BY-2 cells differ from those presented by Yang et al. (2005), who used LeSec13 as a COPII marker. The reason for these apparent discrepancies is not clear, although it may result from the use of different isoforms of Sec13. Evidence from yeast indicates that Sec13 plays cellular roles other than ER–Golgi protein transport (Siniossoglou et al., 2000). The multiple functions of Sec13 in yeast appear to be mirrored in plants, as demonstrated by recent findings that another isoform of AtSec13 (AtSeh1) was localized to multiple locations, including the nucleus, the prevacuolar compartment and the Golgi complex (Lee et al., 2006). AtSeh1 has also been implicated in regulating the cycling of a dynamin-related protein (AtDRP2A) between membrane-bound and soluble forms, playing a role in protein trafficking from the trans-Golgi network to the central vacuole (Lee et al., 2006). As the structures labelled by LeSec13-GFP in BY-2 cells were not able to be defined unequivocally as ERES (Yang et al., 2005), it is possible that the LeSec13 isoform may perform functions other than those involved with COPII transport. Regardless, our results emphasize the need for more than one marker to define the localization of ERES in plants. Further investigation will also be required to determine the ultrastructural nature of ERES in plants: at this stage it is not clear whether the structures labelled by the various COPII coat proteins used in this study represent regions of a general ER membrane, or whether they are more differentiated, forming an intermediate compartment at the ER–Golgi interface. Such analyses may aid in the identification of ER budding profiles (Robinson et al., 2007), although support from immunocytochemistry with specific antibodies for the COPII coat components will be required.

The use of AtSec13 as an ERES marker

We have shown that the clarity of labelling of the fluorescent fusions to AtSec13 appears to vary under different experimental conditions. This may be linked to the transient or stable expression of the marker, as well as the different cell types. The distribution of Sec13 may also be linked to the levels of activity of Sec31, which exists in a heterodimeric state with Sec13. Sec31 has been shown to stimulate the GAP activity of Sec23, leading to the destabilization of the COPII coat and to the uncoating of the carriers (Bi et al., 2007). In this case, the association of Sec13 with ERES would be of a limited duration compared with that of Sec24. It may also be the case that the predominantly expressed Sec31 isoforms in the three systems investigated have different levels of activity, potentially explaining the variation in labelling between these systems.

The evidence that Sec13 may have multiple functions in a cell, coupled with the variability of the distribution of the fluorescent marker in the different expression systems presented in this work, suggests that AtSec24 may be a more consistent ERES marker in plant cells when compared with the isoforms of Sec13 and experimental conditions tested to date.

Experimental procedures

Molecular cloning

Standard molecular techniques were used, as described by Sambrook et al. (1989). Fusions of AtSec13 (AGI: At2g30050) to either EYFP or ECFP (Clontech, http://www.clontech.com) were created as follows: AtSec13 cDNA (ABRC) was amplified by PCR and subcloned upstream of EYFP or ECFP using the unique XbaI and SalI sites of the binary vector pVKH18En6. The primer sequences used for cloning are available upon request.

Expression in plant cells

Nicotiana tabacum (cv. Petit Havana) glasshouse plants were transiently transformed with Agrobacterium tumefaciens, as described previously (Batoko et al., 2000). The bacterial optical density (OD600) used for plant transformation was 0.05 for Sec24 and Sec13 fusions, and 0.2 for ERD2-GFP. For transient expression in Arabidopsis leaves we used a biolistic procedure adapted from Seki et al. (1998). BY-2 cells were stably transformed by co-cultivation with A. tumefaciens (strain C58) for 2 days at 25°C with shaking, followed by selection on media containing hygromycin (YFP-AtSec24, AtSec13-YFP, ERD2-GFP or combinations of these markers), kanamycin (GmMan1-GFP) or both (co-transformation of either ERES marker with GmMan1-GFP).

Imaging

Imaging of leaf tissue was performed 48 h post-transformation, whereas 3–4-day-old BY-2 cells in liquid culture were used for microscopy. Imaging of the transformed plant material was performed using an inverted Zeiss laser scanning confocal microscope (LSM510 META; Zeiss, http://www.zeiss.com), and a 63× water immersion objective. For imaging expression of GFP constructs, YFP constructs or both in leaf tissue, we used imaging settings as described in Brandizzi et al. (2002b) with a 3-μm pinhole diameter. The spectral properties of mGFP5 (in ERD2-GFP) allow efficient spectral separation from YFP (Brandizzi et al., 2002a). For imaging expression of GFP constructs, YFP constructs or both in BY-2 cells, we used excitation lines of the argon ion laser at 458 nm for GFP, and 514 nm for YFP. A 458/514-nm main dichroic beam splitter, 515-nm secondary dichroic beam splitter, and 475–525 and 560–615-nm bandpass filters were used for GFP and YFP, respectively. Appropriate controls were used to exclude the possibility of energy transfer between fluorochromes and crosstalk (see Figure S3). For imaging expression of CFP constructs with or without co-expressed YFP constructs in all systems, we used imaging settings as described in Brandizzi et al. (2002a).

Brefeldin A (BFA; Sigma-Aldrich, http://www.sigmaaldrich.com) was used at a concentration of 100 μg ml−1 to maintain the experimental conditions used by Stefano et al. (2006) in tobacco leaf epidermal cells. Post-acquisition image processing was carried out using CorelDraw (Corel, http://www.corel.com).

Quantification methods

For quantification of YFP-AtSec24 or AtSec13-YFP fluorescence and ERES number, we collected imaging frames from cells expressing YFP-AtSec24 or AtSec13-YFP either alone or with GmMan1-GFP using identical imaging conditions, with a 3-μm optical section. In contrast with a smaller pinhole, this allows maximization of the fluorescent signal from the whole ERES–Golgi unit. The volume of each image was calculated as the optical slice multiplied by the area of the cell visible within the frame (measured for each image using the quantification tools of the LSM510 browser). Visible ERES within these images were counted, and the average number of ERES per 100 μm3 was calculated. We measured the fluorescence pixel intensity levels of visible ERES that were in focus within manually selected regions of interest of equal dimensions in the planar axes (2 μm2) for the ERES in 20 cells (measuring the fluorescence of 10 ERES for each cell, i.e. 200 ERES per marker) for each combination of markers. To normalize the fluorescence intensity values of ERES, we used an identical procedure and sampling size to measure the fluorescence intensity from the cytosol adjacent to the ERES. Measurements of the YFP fluorescence levels at the ERES and cytosol were made within a 2-μm2 circle using ImageJ v1.37 in post-acquisition analysis. Identical methods were used to measure AtSec13-CFP fluorescence at ERES and in the cytosol for experiments involving the co-expression of AtSec13-CFP with YFP-AtSec24. Statistical analyses used the Student’s two-tailed t-test assuming equal variance, and data with P < 0.05 were considered significant. Post-acquisition image processing was carried out using CorelDraw.

Acknowledgements

This work was supported by the Canada Foundation for Innovation, Canada Research Chair Program, Natural Science and Engineering Research Council of Canada and the Department of Energy, Michigan State University. We thank L. Renna (University of Saskatchewan) for subcloning the AtSec24 gene into the plant vector pVKH18En6, A. Nebenführ (University of Tennessee, Knoxville, TN) for α-1,2-mannosidase I-GFP DNA, and P. Fobert (NRC Plant Biotechnology Institute, Saskatoon, Saskatchewan) and R. Degenhardt (University of Saskatchewan) for assistance with Arabidopsis transformation.

Accession number for AtSec13: At2g30050.

Ancillary