These authors contributed equally to this work.
The Syntaxins SYP31 and SYP81 Control ER–Golgi Trafficking in the Plant Secretory Pathway
Article first published online: 7 AUG 2008
© 2008 The Authors. Journal compilation © 2008 Blackwell Munksgaard
Volume 9, Issue 10, pages 1629–1652, October 2008
How to Cite
Bubeck, J., Scheuring, D., Hummel, E., Langhans, M., Viotti, C., Foresti, O., Denecke, J., Banfield, D. K. and Robinson, D. G. (2008), The Syntaxins SYP31 and SYP81 Control ER–Golgi Trafficking in the Plant Secretory Pathway. Traffic, 9: 1629–1652. doi: 10.1111/j.1600-0854.2008.00803.x
- Issue published online: 15 SEP 2008
- Article first published online: 7 AUG 2008
- Received 9 April 2008, revised and accepted for publication 21 July 2008, uncorrected manuscript published online 7 August 2008, published online 26 August 2008
- BFA resistance;
- SNARE overexpression;
- Top of page
- Materials and Methods
- Supporting Information
Overexpression of the Golgi and endoplasmic reticulum (ER) syntaxins SYP31 and SYP81 strongly inhibits constitutive secretion. By comparing the secreted reporter α-amylase with the ER-retained reporter α-amylase-HDEL, it was concluded that SYP81 overexpression inhibits both retrograde and anterograde transport, while SYP31 overexpression mainly affected anterograde transport. Of the other interacting SNAREs investigated, only the overexpression of MEMB11 led to an inhibition of protein secretion. Although the position of a fluorescent tag does not influence the correct localization of the fusion protein, only N-terminal-tagged SYP31 retained the ability of the untagged SNARE to inhibit transport. C-terminal-tagged SYP31 failed to exhibit this effect. Overexpression of both wild-type and N-terminal-tagged syntaxins caused standard Golgi marker proteins to redistribute into the ER. Nevertheless, green fluorescent protein (GFP)–SYP31 was still visible as fluorescent punctae, which, unlike SYP31–GFP, were resistant to brefeldin A treatment. Immunogold electron microscopy showed that endogenous SYP81 is not only present at the ER but also in the cis Golgi, indicating that this syntaxin cycles between these two organelles. However, when expressed at non-inhibitory levels, YFP–SYP81 was seen to locate principally to subdomains of the ER. These punctate structures were physically separated from the Golgi, suggesting that they might possibly reflect the position of ER import sites.
Successful vesicle-mediated protein trafficking in eukaryotic cells is dependent upon the co-ordinated function of several different molecular machineries. These include small guanosine triphosphatases (GTPases), which first recruit cytosolic coat proteins for vesicle formation and subsequently catalyze their disassembly after vesicle budding (1,2). Kinesin- and dynein-type molecular motors have been implicated in vesicle movement along cytoskeletal elements (3,4). Specific tethering of vesicles to defined target membranes is believed to be orchestrated by peripheral membrane proteins that can be recruited by small GTPases of the rab/ypt family (5,6). Finally, a group of membrane-anchored proteins called SNAREs control the ultimate fusion event between a vesicle and its target organelle (7–9). The combination of long- and short-range capturing systems is thought to ensure that vesicles deliver their cargo in an efficient and highly specific manner (7,10,11).
While the intermediate steps between vesicle budding and fusion are poorly understood and much remains to be discovered about the mechanism of organelle movement and the numerous rab effectors, the process of vesicle fusion with the target membrane is well established (7,8,12). The SNARE hypothesis predicts that the docking of transport vesicles is a consequence of the assembly of a tetrameric trans SNARE complex formed from a single vesicle (v)-SNARE molecule (members of the synaptobrevin or VAMP family of proteins) and a trimeric target membrane (t)-SNARE complex (members of the syntaxin and SNAP-25 families) (13,14). Only the correct pairing of cognate v- and t-SNAREs releases sufficient energy to overcome the activation energy barrier for bilayer merger (8,15,16). After fusion has taken place, the resulting cis SNARE complexes are then disassembled by the combined action of n-ethylmaleimide-sensitive factor, an adenosine triphosphatase, and its cofactor α-SNAP, allowing the SNARE molecules to be recycled. How SNAREs are recycled and how they accumulate to specific steady-state levels in their preferred membrane of residence are still poorly understood.
SNAREs usually have a long cytosolic N-terminal domain and are anchored in the membrane mostly through a C-terminal transmembrane domain (with or without a short lumenal C-terminus) or sometimes a prenyl group (7–9). The cytosolic domain contains one (sometimes two) helical stretch of around 65 residues – the so-called SNARE motifs, which allow v- and t-SNARES to form the stable four-helix bundle constituting the trans SNARE complex (17). SNAREs have also been grouped into R- or Q-SNAREs according to the presence of an arginine or a glutamine residue centrally located (the ‘zero layer’) within the SNARE motif (18). As a result of this structural classification, five subfamilies of SNAREs have emerged: R-SNAREs carry an arginine in the zero layer and are similar to VAMP1 or synaptobrevin1. Three Q-SNAREs can be distinguished (carrying glutamine in the most conserved zero layer): Qa-SNAREs are usually the largest polypeptide chains and are related to syntaxin, Qb-SNAREs are related to the N-terminal helix of SNAP25 and Qc-SNARES are similar to the C-terminal helix of SNAP25. Finally, the unique SNAP25 family carries two SNARE motifs within the same polypeptide and can be regarded as a combined Qb–Qc SNARE.
The yeast (Saccharomyces cerevisiae) genome encodes for 23 SNAREs (19), of which several are critical for transport between the endoplasmic reticulum (ER) and the Golgi apparatus. The heavy chain Qa syntaxin Sed5p can form a t-SNARE complex with seven other SNARE proteins (Bet1p, Bos1p, Sec22p, Gos1p, Ykt6p, Sft1p and Vti1p) (20,21). This allows it to function in both anterograde traffic between the ER and the Golgi and retrograde traffic within the Golgi apparatus (22–28). In contrast, the heavy chain Qa syntaxin Ufe1p interacts with Sec20p, Sec22p and the unusual SNARE Use1p and appears to regulate only retrograde traffic between the Golgi and the ER (29–31). Under steady-state conditions, Sed5p and Bet1p are principally found in the Golgi apparatus (22,32), whereas Ufe1p and Sec22p are mainly localized to the ER (33,34).
Because most SNAREs are expected to enter the secretory pathway at the ER (35), they must be transported (and recycled) by vesicles, and many SNAREs cycle continuously between the ER and the Golgi apparatus to various extents (32,33,36–40). Therefore, a range of regulatory mechanisms must exist to prevent inappropriate SNARE interactions for molecules in transit to their final destination. It was recently demonstrated that cycles of phosphorylation and dephosphorylation are crucial for the normal functioning of Sed5p (41). Such regulatory controls may explain why overproduction of syntaxins can perturb protein transport towards their organelle of residency, as shown for Sed5p in yeast where it compensates for a deletion of ERD2 (22,42,43) and for PEP12 in plants that causes secretion of vacuolar cargo (44). Moreover, the discovery of inhibitory SNARE (iSNARE) activity of certain SNAREs (45) even suggests that the secretory pathway can utilize some of these undesired interactions to increase polarity and improve the fidelity of the various interconnected transport routes.
Over 50 SNARE genes have been identified in Arabidopsis, leading to a plant-based nomenclature for them (9,46). It has been suggested that this large number reflects a greater diversity in plant-specific cellular functions (47). According to Uemura et al. (48) and Moreau et al. (49), 15 SNAREs locate to the early secretory pathway of plant cells of which 9 locate principally to the Golgi apparatus and 6 to the ER. Among these are homologs of the SNAREs that regulate ER–Golgi transport in yeast: BS14 (Bet1p), MEMB11 (Bos1p), SEC22 (Sec22p), SYP31 (Sed5p) and SYP81 (Ufe1p). To date, only one paper has been published that deals with the functionality of these SNAREs in ER–Golgi protein trafficking in plants. In a transient expression study with tobacco leaf epidermal cells, Chatre et al. (50) were able to localize fluorescent protein-tagged BS14, MEMB11 and SYP31 to Golgi stacks and SEC22 to both Golgi and ER. Overexpression of all of these SNAREs led to the retention of a secretory form of the yellow fluorescent protein (YFP) in the ER, pointing to their importance in intracellular protein transport. In fact, as monitored by fluorescent ER-deficient mutant 2 (ERD2) and sialyl transferase (ST) fusions (cis and trans Golgi markers, respectively), the overexpression of MEMB11 and SEC22 was seen to cause the Golgi apparatus to collapse into the ER, a phenotype similar to the action of brefeldin A (BFA).
In this study, we show that the syntaxins SYP31 and SYP81 may be used as tools to inhibit both anterograde and retrograde transport between the ER and the Golgi apparatus, whereby SYP81 overexpression appeared to inhibit retrograde traffic to a greater extent than SYP31. Of the other interacting SNAREs investigated, only the overexpression of MEMB11 resulted in an inhibition of α-amylase secretion in tobacco leaf protoplasts. The position of a fluorescent tag is crucial to the functionality but not necessarily for the subcellular location of the expressed SNARE molecule. Thus, only when tagged at the N-terminus of SYP31 can a perturbatory effect on secretion be measured. Whereas SYP31 clearly located to Golgi stacks and SYP81 was present on tubular subdomains of the ER network [possibly reflecting ER import sites (ERIS) domains], overexpression of both N-terminal-tagged syntaxins caused standard Golgi marker proteins to redistribute into the ER. Interestingly, and in direct contrast to C-terminal-tagged SYP31, the enlarged fluorescent signals for N-terminal-tagged SYP31 remained punctate and were BFA resistant.
- Top of page
- Materials and Methods
- Supporting Information
Overexpression of the Golgi syntaxin SYP31 inhibits constitutive secretion
When genes are essential for the survival of the cell, overexpression analysis can shed light on their mode of action by titration of interacting endogenous components (51). Experiments with a plasma membrane syntaxin (NtSYP121) have revealed that overexpression of its cytosolic domain interferes with endogenous components and results in defective secretion, whereas coexpression of the wild-type (WT) syntaxin alleviates the inhibition (52). However, the opposite results were obtained with the prevacuolar compartment syntaxin SYP21 (PEP12), where overexpression of the WT molecule inhibited vacuolar sorting, while expression of the cytosolic SP2-like fragment revealed no measurable effect on traffic (44). Further conflicting results have been recently presented in which the cytosolic SP2 fragment of PEP12 inhibited transport of membrane-spanning cargo to the tonoplast (53). Therefore, to obtain further information on syntaxin-mediated protein sorting, we tested the behaviour of the major Golgi syntaxin SYP31 (SED5). To this end, several chimeric constructs were created to express either full-length SYP31, a soluble SP2-like fragment, or an N-terminally truncated form lacking the unstructured region that interacts with the SM (Sec1/Munc18) family protein SLY1 (54) (Figure 1A).
Using the robust tobacco mesophyll protoplast expression system, increasing plasmid DNA concentrations for each of these constructs were cotransfected with constant plasmid concentrations of the secretory cargo barley α-amylase. As in other papers from our laboratories (51,55,56), the secreted and cellular amounts of α-amylase were determined after 24 h of incubation and the secretion index calculated (Figure 1B–F). Full-length SYP31 was highly effective in inhibiting the secretion of α-amylase in a dose-dependent manner when expressed in either the WT untagged form (Figure 1B) or the N-terminally tagged with the large YFP (Figure 1C) or a small haemagglutinin (HA) epitope (Figure 1D). Although we could only monitor the expression of the tagged versions, the effect was clearly dosage dependent in all cases. At the highest level of SYP31 overexpression, corresponding to 20 μg of plasmid DNA, the secretion index was reduced to roughly 25% of control levels.
To determine the domain of SYP31 that is responsible for the inhibition of secretion when overexpressed, we performed experiments with two truncated versions of this SNARE. Whereas deletion of the N-terminal 40 residues did not diminish the inhibitory effect of this SNARE (Figure 1E), overexpression of the whole of the cytosolic domain of HA–SYP31 lacking the transmembrane domain was without effect on secretion (Figure 1F). This points to the importance of the transmembrane domain in mediating the inhibitory effects of SYP31, a feature shared with SYP21 (44) but not with plasma membrane syntaxins such as NtSYP121 (52).
Employing plasmid DNA concentrations of HA–SYP31 that caused a 50% reduction in the secretory index, we next investigated the kinetics of the inhibitory action of SYP31 overexpression on α-amylase synthesis and secretion. As seen in Figure S1A, secretion is both rapidly and continuously affected by SYP31 expression. Whereas in the short term this effect seems to be more a consequence of the prevention of α-amylase release from the protoplasts, after about 4 h of SYP31 overexpression, the synthesis of α-amylase itself is reduced (Figure S1B,C).
The results obtained with the quantitative reporter α-amylase were also visually confirmed in the confocal microscope (Figure 2). Overexpression of either the full-length YFP–SYP31 or the YFP–SYP31ΔNT caused the Golgi marker Man1–red fluorescent protein (RFP) to locate principally to the ER (Figure 2A–F) rather than being present as the usual punctate structures of 1-μm diameter that represent individual Golgi bodies in plants (Figure 2H,I, but see also Figure S2A,B). In contrast, overexpression of YFP–SYP31ΔTM produced a cytosolic signal for this truncated SNARE that had no effect on the typical punctate Golgi signal (Figure 2G–I), which behaved essentially as in the control protoplasts expressing Man1–RFP alone. In these protoplast expression experiments, the fluorescent signals for both full-length YFP–SYP31 and YFP–SYP31ΔNT remained punctate, although the structures were larger and present in smaller numbers than the typical Golgi signal.
iSNARE activity but not Golgi localization of SYP31 is compromised by C-terminal tagging
Because Chatre et al. (50) purposely chose to place a fluorescent tag at the C-terminus in their study of ER–Golgi SNAREs in tobacco leaves, we decided to check whether the position of the tag was of importance. Surprisingly, in contrast to green fluorescent protein (GFP)–SYP31, the C-terminal-tagged variant SYP31–GFP exhibited hardly any effect on the secretion index of α-amylase even when 30 μg of plasmid was transfected (Figure 3A). Analysis of the individual activities in cells and medium revealed that N-terminal-tagged SYP31 reduces the total yield of α-amylase, whereas the C-terminal-tagged SNARE had much less detrimental influence on the synthesis of this cargo molecule (Figure 3B).
Visual confirmation of lack of function for C-terminal-tagged fluorescent SYP31 was obtained by co-electroporating standard Golgi (ST–RFP) and plasma membrane (RFP–TMD23) markers (Figure 4). We first, however, checked whether the expression period had any effect on the Golgi localization of SYP31–GFP and did not observe any significant difference in the size, number or distribution of the punctate signals with time (Figure 4A–C). For the coexpression experiments, we therefore decided upon an intermediate expression period (15 h). At this time-point, ST–RFP and SYP31–GFP signals showed high colocalization (Figure 4D–F) and RFP–TMD23 targeted successfully to the plasma membrane (Figure 4G–I). In comparison, expression of the N-terminally tagged version of SYP31 caused the Golgi-located SYP31 signals to aggregate with time (Figure 5A–C) and elicited serious mislocalizations of the two standard markers. The Golgi marker ST–GFP was partially distributed in the ER (compare Figure 5D–F with Figure 5G–I), and RFP–TMD23 localized to the ER rather than to the plasma membrane (Figure 5J–L).
Punctate N-terminal-tagged SYP31 structures lack Golgi markers are located on ER tubules and are BFA resistant
At low expression levels in leaf infiltration experiments, (X)FP–SYP31 colocalizes with the Golgi marker ST–cyan fluorescent protein (CFP) in mobile punctae (Figure 6A–C,I; Figure S3). However, at higher magnification (Figure 6D–H; see also Figure S3), the signals are actually closely appressed on one another; consistent with the concept that SYP31 is expected to reside on the cis Golgi cisternae, whereas ST–CFP is known to be a trans Golgi marker (57). As with protoplasts (Figure 5), at higher levels of expression, the Golgi marker ST–CFP is also redistributed into the ER in infiltrated epidermal cells although the (X)FP–SYP31 signal remains punctate (Figure 6J). When (X)FP–SYP31 is overexpressed in leaves of tobacco plants stably overexpressing GFP–HDEL, it is clear that the SYP31 punctae are located on the ER network and these are also mobile (Figure 6K; Figure S4). This situation is unchanged when the epidermal cells are exposed to BFA (Figure 6L), a well-known inhibitor of the secretory and endocytic pathways (58,59). We have confirmed the data obtained by leaf infiltration by electroporating protoplasts using Man1–(X)FP as a cis Golgi marker (Figure S5). Whereas BFA causes this marker to be completely redistributed into the ER (Figure S5A,D), (X)FP–SYP31 signals remain punctate, after both 7 h (Figure S5B,C) and 24 h (Figure S5E,F) expression periods. This contrasts with a partial redistribution of SYP31–GFP into the ER upon BFA treatment (Figure S5G–I).
The effect of N-terminally tagged overexpressed SNAREs that act in conjunction with SYP31
Having established that C-terminal tagging can perturb the biological function of a syntaxin, we decided to analyse the other SNAREs that act in conjunction with SYP31 at the level of the Golgi apparatus. These included the v-SNARE BS14a and the light chains SEC22 and MEMB11 that have been shown to contribute to the regulation of ER–Golgi traffic in conjunction with SYP31 (45,60). N-terminally YFP-tagged derivatives of these SNAREs were cotransfected at increasing plasmid DNA concentrations in tobacco mesophyll protoplasts with constant amounts of plasmid encoding secretory cargo α-amylase. As practiced previously, the secreted and cellular amounts of α-amylase were determined after 24-h incubation and the secretion index calculated (Figure 1). In addition, we tested also the untagged WT versions of the three SNAREs. Of these, only MEMB11-derived constructs led to a consistent dose-dependent inhibition of secretion (Figure 7). Similar results were obtained with HA-tagged derivatives (Figure S6).
In parallel, we investigated the effects of overexpressed YFP-tagged SNAREs on the location of coexpressed standard Golgi (Man1–RFP) and ER (GFP–HDEL) markers (Figure 8). Two of the SNAREs: BS14a and MEMB11 colocalized with the Golgi marker (Figure 8A–F), whereas SEC22 localized to both the Golgi stacks and the ER (Figure 8G–L). In contrast to the other SNAREs, MEMB11 expression led to the formation of Golgi aggregates and also the redistribution of some of the Man1–RFP into the ER (Figure 8: compare panels D–F with A–C and G–I). These results are in partial agreement with those of Chatre et al. (50) who also recorded that BS14a and MEMB11 located to the Golgi apparatus and SEC22 had a dual location at the ER and Golgi apparatus. However, unlike our finding that the overexpression of MEMB11 led to the formation of Golgi aggregates, Chatre et al. (50) observed a redistribution of the two Golgi markers ERD2 and ST–YFP into the ER. Taken together, these results suggest that, in addition to the syntaxin SYP31, expression levels of MEMB11 are critical to support normal protein trafficking in the early secretory pathway in plants. Both SNAREs act in an inhibitory manner when overexpressed, but this is not a general property of SNAREs operating in the early secretory pathway as this feature is not shared by the v-SNARE BS14a and the light chain SEC22.
Overexpression of the ER syntaxin SYP81 inhibits both retrograde and anterograde transport between the ER and the Golgi apparatus
While the Golgi-located syntaxin SYP31 is thought to mediate Golgi import of ER-derived material, the syntaxin SYP81 is believed to reside on the ER (48) and mediates the fusion of retrograde Golgi-derived coat protein I (COPI) vesicles with the ER membrane. Interference with retrograde transport could therefore shed further light on the regulatory mechanisms that control traffic between these two organelles. For this reason, we examined the effects of overexpression of WT and C/N-terminal-tagged SYP81 on α-amylase secretion in tobacco mesophyll protoplasts. We also used the retrograde cargo molecule α-amylase-HDEL (51) to monitor for potential inhibition of the retrograde transport route.
The results show that overexpression of SYP81 or YFP–SYP81 caused a dosage-dependent inhibition of α-amylase secretion (Figure 9A,C, grey bars). In contrast to SYP31, this inhibition was independent of the position of the fluorescent tag (Figure 9A,B). However, unlike α-amylase, α-amylase-HDEL follows both anterograde and retrograde transport routes (51,61). If SYP81 overexpression interferes with both, it should have opposing, partially compensating, effects on the secretion index of α-amylase-HDEL, leading to lower values. The observed effect of this SNARE on the retrograde cargo was indeed significantly lower (Figure 9C, white bars). To confirm this hypothesis, we compared the effect of SYP31 overexpression on α-amylase with its effect on α-amylase-HDEL. In contrast to the ER syntaxin SYP81, the cis Golgi syntaxin SYP31 affected both cargo molecules in a similar manner (Figure 9D). These results indicate that SYP31 overexpression inhibits mainly anterograde transport, whereas SYP81 overexpression compromises both transport routes and thus shows a more neutral effect on the cargo α-amylase-HDEL, which follows both these transport routes.
SYP81 localizes to subdomains of the ER
Like SYP31, overexpression of SYP81 led to mistargeting of Golgi markers (Figure 10). Both ST–GFP and Man1–GFP were no longer detected as punctate signals but instead assumed an ER-like distribution (Figure 10A–F). Interestingly, the YFP–SYP81 fusion did not show a reticular ER pattern but instead highlighted mobile, punctate structures that were generally smaller and outnumbered the typical number of Golgi stacks roughly threefold (Table 1). When coexpressed with ER markers (GFP–HDEL and calnexin–GFP), the YFP–SYP81 punctae were seen to lie on the ER (Figure S7).
|Cell 1||Cell 2||Cell 1||Cell 2|
|Number of labelled compartments||648||353||1131||1373|
|Diameter of protoplasts (μm)||55||35||48||50|
Because of the unusual nature of these findings, we decided to investigate further the effects of SYP81 overexpression by leaf infiltration, which permits lower expression levels because of stable integration of limited T-DNA copies (62). When coexpressed with the Golgi marker ST–CFP in tobacco leaf epidermis cells, a dosage-dependent inhibition of ER export became apparent. To achieve very low YFP–SYP81 expression levels close to the detection limit, short incubation periods of 30 h after infiltration and selection of weak expressing cells were necessary. Under these conditions, the Golgi marker ST–CFP was present mostly as punctate structures (Figure 10G–I). These did not colocalize at all with YFP–SYP81 punctae, which were physically separated, much smaller and more numerous, as deduced from the earlier experiments with protoplasts (Table 1). At intermediate expression levels, ST–CFP was partially detected in a tubular ER network and in fewer remaining punctae (Figure 10J–L). Higher expression levels of YFP–SYP81 were found in cells after longer incubation periods after infiltration. To avoid saturation, laser output detector gain was lowered for the YFP channel. Under these conditions, ST–CFP was no longer present as punctate structures representing the Golgi apparatus but was exclusively localized in the ER, which showed an increased tendency to form cisternae in addition to tubules (Figure 10M–O). These results can be interpreted as an increased tendency of SYP81 to block ER export and to progressively modify ER morphology because of the accumulation of trapped proteins that failed to exit.
To further characterize YFP–SYP81-labelled structures, the fusion protein was coexpressed with the ER markers GFP–HDEL and GFP–calnexin. GFP–HDEL expression normally highlights an ER network that is typically of tubular nature with very few or no cisternae (Figure 11A). When GFP–HDEL was coexpressed with YFP–SYP81, the typical tubular ER network highlighted by GFP–HDEL was replaced by a modified ER showing a much higher proportion of cisternae in addition to the tubular network (Figure 11B–D). However, it was clear that YFP–SYP81 punctae were always associated with the tubular ER network rather than the cisternae and often highlighted extended ‘fingers’ protruding from other connected tubular networks of the ER.
To confirm the observation that SYP81 is restricted to tubular rather than cisternal domains of the ER, ER cisternae were induced by expressing the type I membrane-spanning ER marker GFP–calnexin (55,63). Even without YFP–SYP81, the ER showed a much higher proportion of cisternae (Figure 11E) compared with cells expressing solely the lumenal ER marker GFP–HDEL. However, upon coexpression with YFP–SYP81, the proportion of cisternal ER was increased further. Nevertheless, punctate labelling by the SNARE fusion was mostly restricted to the few remaining tubular parts of the ER, while the cisternal sheets were mostly devoid of this marker (Figure 12F–H).
In an attempt to gain more information on the location of SYP81, we performed immunogold electron microscopy (immunogold EM) with a polyclonal antiserum generated against the cytosolic domain of SYP81 expressed as a recombinant protein in Escherichia coli. Because the antigen was prepared from an Arabidopsis clone, we opted for Arabidopsis root cells as the most appropriate system to investigate. In addition, the visualization of the ER in high-pressure frozen samples is much better in these cells than in leaf mesophyll cells, which are difficult to cryofix and only have a thin parietal layer of cytoplasm. In addition to being found at the ER (Figure 12A,B), endogenous SYP81 was also detected at the cis Golgi (Figure 12C), especially at the vesiculating periphery of cisternae (Figure 12C). This suggests that, as is the case for the Golgi syntaxin SYP31, the ER syntaxin SYP81 also cycles between the ER and the Golgi apparatus. The distribution of gold label over the ER in Arabidopsis root cells was not uniform, but it was not possible in our sections to distinguish with certainty tubular from cisternal ER.
Taken together, our results show that SYP81 locates principally to the ER and that, when overexpressed, YFP–SYP81 labels a discrete tubular subdomain of the ER. Increased expression of this ER syntaxin inhibits anterograde as well as retrograde traffic between the ER and the Golgi and leads to morphological changes of the ER reminiscent of the increased accumulation of membrane-spanning proteins in the ER.
SYP31 and SYP81 do not complement mutations in their presumptive yeast orthologs
To establish whether SYP31 and SYP81 were orthologs of Sed5p and Ufe1p, respectively, we expressed the plant proteins in yeast strains carrying chromosomal deletions of SED5 or UFE1 balanced by a counter-selectable copy of the corresponding WT gene on a plasmid (see Materials and Methods). When expressed from a relatively strong promoter and on a low-copy-number plasmid, SYP31 was unable to support the growth of sed5Δcells. Expression of SYP31 in the sed5-1 mutant had a modest inhibitory effect on cell growth, and SYP31 was unable to suppress the temperature-sensitive defect of this strain at 37°C. Unlike SYP31, expression of SYP81 was strongly inhibitory to yeast cell growth. Although transformants did arise following an extended incubation period (5 days at 30°C), they were unable to grow on media that selected against Ufe1p (data not shown).
- Top of page
- Materials and Methods
- Supporting Information
Although SNARES are essential for the correct functioning of a number of different trafficking pathways in plants including secretion to the cell surface, vacuolar protein transport and cytokinesis (47,64,65), only two papers give any details on SNAREs involved in ER–Golgi transport in plants (48,50), and only the latter paper contains functional data. Interestingly, these data arose from overexpression studies rather than gene knock out. Because key regulators of the secretory pathway are encoded by essential housekeeping genes, the consequences of their absence are usually lethal and can therefore yield little information about their mode of action. This is indeed the case for the null mutants of Sed5p and Ufe1p in yeast (22,32,66–68).
For our investigation, we have chosen the same set of SNAREs as Chatre et al. (50), including the syntaxin SYP31, the two light chains SEC22 and MEMB11 and the v-SNARE BS14a, but have added the t-SNARE SYP81 because it is believed to regulate retrograde Golgi-to-ER transport. In this context, one should be aware that published results on syntaxin function of the late secretory pathway have yielded conflicting results (44,52), making it important to extend these studies to syntaxins of the early secretory pathway. To improve the reliability of our expression results in tobacco, we used tomato clones because the differences in homology between the two plant types were in some cases significant (Tables S1 and S2). In contrast, the SYP81 studies were conducted with an Arabidopsis clone because we planned to carry out immunogold EM, and for this, Arabidopsis roots were the tissue of choice for this type of localization.
iSNARE activity is not a general property of all SNAREs
In agreement with Uemura et al. (48) as well as Chatre et al. (50), we confirm that the v-SNARE BS14a and the light chain MEMB11 locate to the Golgi apparatus together with the syntaxin SYP31. We also concur that the light chain SEC22 has a dual location, being found both on Golgi stacks and on the ER. However, in terms of function, our data are in many aspects contradictory to those presented by Chatre et al. (50). In our studies, the only SNARE that was inhibitory in the α-amylase secretory assay and in addition elicited a structural effect in the early secretory pathway was YFP–MEMB11. Chatre et al. (50) also recorded a negative effect of overexpression of MEMB11–YFP on secretion, but unlike our finding that the expression of YFP–MEMB11 led to the formation of Golgi aggregates, they observed a redistribution of the two Golgi markers ERD2 and ST–YFP into the ER. Also, in contrast to the results of Chatre et al. (50), we observed neither a physiological nor a structural effect on tobacco protoplasts through the overexpression of SEC22.
Overexpression of the syntaxins SYP31 and SYP81 perturb protein trafficking in the early secretory pathway
In yeast, Sed5p (SYP31) localizes to the cis Golgi and plays a central role in mediating both anterograde traffic from the ER to the Golgi and retrograde traffic within the Golgi (69,70). In tobacco, SYP31 also localizes to the cis Golgi (Figure 6A–H), consistent with a role in the correct functioning of protein trafficking in the early secretory pathway in plants. Because loss of function of Sed5p is lethal in yeast, we have therefore chosen to explore the possibility that overexpression of WT SYP31 or truncated derivatives may interfere with endogenous components and thus shed light on its possible mode of action. Unlike the plasma membrane syntaxin (52), our results show no measurable interference when soluble cytosolic domains are overexpressed. Instead, overexpression of the full-length SNARE severely impaired anterograde ER-to-Golgi transport (Figures 1B,C and 3A). Thus, SYP31 shows the same behaviour as SYP21 (44). Furthermore, iSNARE activity was independent of the N-terminal domain thought to interact with SLY1 (54) (Figure 1E).
Our results with SYP31 are also fundamentally different to those of Chatre et al. (50) who because of the use of C-terminally tagged SYP31 were unable to observe the redistribution of Golgi marker enzymes into the ER. We have specifically tested for this and demonstrated that C-terminally tagged SYP31 no longer interfered with constitutive secretion, as determined using the quantitative α-amylase assay (Figure 3). Chatre et al. (50) purposely opted for C-terminal tagging as a precautionary measure to avoid possible interference with the regulatory N-terminal SNARE domains, although as seen in a number of earlier studies on yeast and mammalian cells (32,71–74), there is no basis for this. Because N-terminally tagged SYP31 (using either YFP or a smaller HA tag) behaved in essentially the same way as untagged WT SYP31, we conclude that N-terminal tagging is the method of choice for this particular syntaxin and in this respect note that N-terminally tagged syntaxins have repeatedly complemented yeast mutants. It is possible that the presence of a fluorescent protein at the C-terminus of SYP31 somehow prevents this syntaxin from entering into cis or trans SNARE configurations with other SNAREs in the Golgi apparatus. However, the perturbatory effects of untagged or N-terminal-tagged SNARE overexpression might be explained by the excessive formation of non-fusogenic SNARE complexes that as a consequence reduces the availability of fusogenic SNAREs.
Strong overexpression of Sed5p in yeast has been shown to be toxic (75), but overexpression of full-length SYP31 had no discernable effect on growth of yeast cells (data not shown). However, expression of SYP31 in yeast was unable to support the growth of cells in the absence of Sed5p. The inability of SYP31 to functionally substitute for Sed5p in yeast is perhaps not surprising, given the number of essential binding partners for Sed5p. Despite significant amino acid similarity between SYP31 and Sed5p, the failure of SYP31 to bind to a single essential Sed5p-interacting partner would render SYP31 non-functional in yeast cells.
Ufe1p is regarded as the t-SNARE for retrograde traffic from the Golgi to the ER in yeast (30,33). In agreement with Uemura et al. (48), the plant ortholog SYP81 also localizes to the ER but is also found at the cis Golgi, reflecting its recycling (see below). Even if SYP81 has an exclusive role in controlling retrograde Golgi-to-ER traffic, it is not possible to demonstrate clearly the induced secretion of retrograde cargo after introducing non-stoichiometric inhibitory levels of SYP81. This is because retrograde and anterograde transport between the ER and the Golgi are partially interdependent, and inhibition of the recycling step will thus have an indirect effect on anterograde transport. The fact that the Arabidopsis maigo2 mutant shows impaired ER exit of seed storage proteins highlights this dilemma. MAIGO2 encodes the plant homolog of yeast TIP20 (75), a component of the Dsl1 complex, the presumed tether on the ER to receive COPI vesicles. MAIGO2 interacts with the ER SNAREs Sec20p and Ufe1p in vitro and is thus a central key element of the ER import machinery, yet its absence causes ER retention of vacuolar proteins.
Using the combination of the reporter cargo molecules α-amylase (for anterograde transport) and α-amylase-HDEL (for anterograde and retrograde transport), we have overcome this problem and provided evidence that SYP81 overexpression interferes with both anterograde and retrograde transport pathways. α-amylase-HDEL behaved relatively neutral to the increased dosage of SYP81, whereas α-amylase secretion was strongly inhibited. Thus, SYP81 overexpression, as for Ufe1p in yeast (76) and syntaxin 18 in mammals (77), is a potent inhibitor for protein trafficking in the early secretory pathway. Overexpression of SYP81 and syntaxin 18 leads to the disappearance of the Golgi apparatus, but, unlike syntaxin 18, SYP81 expression did not elicit the production of special aggregates of ER membrane (77). Our findings on the effects of SYP81 expression in yeast cells contrast with those of SYP31. Expression of SYP81 from a low-copy-number plasmid had a strong inhibitory effect on the growth of WT yeast cells, and SYP81 expression was incapable of supporting the growth of yeast strains lacking UFE1. This phenotype was unique to SYP81 as WT cells overexpressing Ufe1p were indistinguishable from the corresponding empty vector controls (data not shown).
Overexpressed SYP31 accumulates in BFA-resistant punctate structures
Both N-terminal- and C-terminal-tagged SYP31 localized to the Golgi apparatus. However, only N-terminally tagged SYP31 resulted in a clear-cut mistargeting of Golgi markers to the ER because of impaired ER export followed by turnover of Golgi-resident markers or increased retrograde transport or both. Thus, only the C-terminal-tagged version is appropriate as a neutral marker for this organelle. However, an interesting effect of overexpression of GFP–SYP31 is the conferment of BFA resistance to the SYP31-positive structures (Figure 6; Figure S5). Because, as a result of the overexpression of SYP31, these structures lack standard cis and trans Golgi enzymes, the immediate question is what are they? We can rule out the possibility that the SYP31-positive punctae represent unspecific artifacts of expressed protein for the following reasons. First, such aggregates are not formed in protoplasts electroporated with a truncated cytosolic version of SYP31, and this expression does not cause a Golgi marker to relocate to the ER (Figure 2G–I). Second, as shown by our leaf infiltration experiments, BFA-resistant (X)FP–SYP31 aggregates are mobile and remain associated with the ER network (Figure 6K,L; Figure S4).
Studies on mammalian cells have shown that while BFA causes Golgi enzymes to redistribute into the ER, a number of structural proteins, principally the Golgins, remain intact in the cytosol (78). These provide a scaffold for ER-derived Golgi elements to assemble onto either after BFA wash out or during telophase (79). BFA treatment of mammalian cells is therefore considered to mimic the onset of mitosis because this involves a drastic shut down of ADP ribosylation factor (Arf)1 activity [for literature, see Altan-Bonnet et al. (80)] and the molecular target for BFA is an Arf-GTP exchange factor (81) Nevertheless, recent research into Golgi inheritance during mitosis in mammalian cells indicates that in prometaphase, Golgi fragments accumulate near to ER exit sites (ERES), before being absorbed completely into the ER during metaphase, which itself then fragments (82).
Although many Golgin homologs are present in plants (83), the Golgi apparatus in higher plants does not fragment during mitosis. It is however possible that SYP31 overexpression leads to the formation of stable structures analogous to prometaphase fragments, which are also formed in mammalian cells after BFA treatment. In this regard, it is interesting to note that Hardwick and Pelham (22) also showed that the overexpression of Sed5p in yeast led to an accumulation of Golgi-derived vesicles.
Exactly what the BFA-resistant GFP–SYP31 structures are must await the availability of stable cell lines expressing this SNARE construct under the control of an inducible promoter because an EM analysis of transiently expressing protoplasts is not possible because of low transformation levels and technical difficulties in processing protoplasts for immunogold EM with cryomethods.
Consequences of syntaxin localization(s)
As in yeast and mammalian cells, we have shown that in tobacco mesophyll cells, the principal location of SYP31 is the cis Golgi. However, Sed5p, the orthologue of SYP31 in yeast, is known to cycle between the Golgi and the ER (32). This should be a consequence of the presence of both ER retrieval and ER export motifs in the cytosolic domain of Sed5p. KKXX at the carboxy terminus and variously located arginine motifs are well-known ER retrieval motifs and interact with COPI subunits (84–86). Sed5p has two KK motifs, but these are internally located (aa 18 + 19, 56 + 57), but no arginine motifs. However, an ‘A-Site’ (NSNPF, aa 203–208) and a ‘B-Site’ (LxxL/ME, aa 237–242) for COPII interaction (87) are present. SYP31 also has two internal KK motifs (aa 20 + 21, 49 + 50), no arginine signals, a likely A-Site motif (SSNPF, aa 192–196) but no clear B-Site motif. Thus, the molecular basis for Sed5/SYP31 cycling between ER and Golgi membranes is not immediately apparent. Nevertheless, we must assume that our inability to visualize this syntaxin in the ER is a consequence of low concentrations and large membrane surface area.
With respect to SYP81, the situation is even less clear. There is no direct observation in the literature that Ufe1p in yeast or syntaxin 18 in mammals cycles between the ER and the Golgi. Ufe1p possesses two KK motifs (aa 103 + 104, 263 + 264), but again, these are internally located and the immediately following amino acid residues tend to disqualify them as retrieval motifs anyway (86). Arginine motifs are also missing. However, a B-Site-type motif (LQVLE, aa 244–248) and a DxE-type ER export motif (aa 149–151) are present in SYP81. It is therefore likely that SYP81 can at least exit the ER. In contrast, although, a possible semi-conserved B-Site motif (LLDDE, aa 239–242) for ER export does exist. SYP81 has a potential arginine motif RRKPKR (aa 178–183), even though KK motifs are lacking. In regard to the presence of SYP81 in Golgi membranes, the situation is the reverse to that with SYP31, providing a greater chance of detection, and indeed, our immunogold labelling data does suggest that SYP81 cycles between the ER and the Golgi apparatus. However, our fluorescence data give no indication of the presence of this syntaxin in the Golgi. This apparent contradiction may lie in inherent differences in the distribution of SYP81 in roots versus leaves or indeed between tobacco and Arabidopsis. It could also reflect the difficulties in recognizing immunoreactive portions of the ER in the narrow ER–Golgi interface. When these are tubular, as seems to be the case for leaf epidermal cells (88), and are not decorated with ribosomes, it is possible that they cannot be adequately distinguished from fenestrated membranes at the cis Golgi and thus giving the false impression of Golgi localization.
At first sight, the punctate, mobile fluorescent signal for SYP81 on the ER surface is reminiscent of the punctate distribution of SAR1, the GTPase that recruits COPII (89). These ERES are closely linked with Golgi stacks in tobacco epidermal cells with only few isolated ERES being discernible (90). In cells exhibiting higher secretory activities such as tobacco BY-2 cells, the frequency of Golgi-independent ERES is much higher and these may represent de novo formations of ERES (56). Several observations made on tobacco epidermal cells suggest that ERES are restricted to tubular rather than cisternal ER. First, Golgi stacks are restricted to tubular ER, except when the actin cytoskeleton is depolymerized (88). Second, Golgi stacks and ERES move together (55). Third, when the proportion of cisternal ER is increased in relation to tubular ER by overexpressing calnexin, Golgi stacks are still only found over tubular ER (63).
In contrast to ERES, very little or no colocalization of SYP81 with Golgi markers was detected in our studies. In the tobacco epidermis and in protoplasts, transiently expressed SYP81 located to punctate structures that were approximately threefold more frequent than Golgi stacks (Table 1). When SYP81 expression was low enough to avoid interference with the Golgi marker ST–CFP, the punctate signals of SYP81 did not form units with Golgi stacks. These discrete SYP81 sites are thus separate from ERES, and we postulate that they may represent ERIS. Thus, although ERES and ERIS represent separate domains of the ER, both appear to reside on tubular rather than cisternal ER.
While it is clear that only a combination of techniques including EM, fluorescence microscopy and biochemical transport assays can shed light on the true complexity of vesicle trafficking at the ER–Golgi interface, further attempts to positively identify and characterize ERIS are unfortunately fraught with problems. First, overexpression of SYP81 severely disrupts anterograde and retrograde transport between the ER and the Golgi apparatus. Unless expression is extremely low, YFP–SYP81 is therefore likely to disrupt ERIS and consequentially ERES too. Second, as we have seen, immunogold EM can confirm the presence of SYP81 at the ER but gives little information in addition. In section, tubular ER can be cross-sectioned where it will look like a vesicle, and it is purely a matter of luck that a portion of the tubule is cross-sectioned where SYP81 lies. Tubular ER can also be sectioned parallel to its long axis and will then be indistinguishable from ER cisternae, which do not bear SYP81. Therefore, many of the ER profiles that are visible in section probably would not be immunoreactive. Most importantly, and in contrast to vesicle budding, it is not yet possible to capture vesicle fusion events by cryotechniques. This is well known from studies on pollen tubes where huge numbers of vesicles collect at the growing tip, but fusion profiles are not visible (91). At best, ER syntaxins can only act as indicators for fusion events that have occurred before specimen preparation.
Materials and Methods
- Top of page
- Materials and Methods
- Supporting Information
Recombinant plasmid production
All DNA manipulations were performed according to established procedures. The highly competent E. coli MC1061 strain (92) was used for the routine amplification of all plasmids. Previously established plasmids were used, encoding for α-amylase (93), ST–YFP (from pVKH18En6::ST-YFP) (94), Man1–RFP (from pBP30-GmMan1:RFP) (56), Man1–GFP (from pBP30) (95), calnexin–GFP (96), GFP–HDEL (97) and RFP–TMD23 (94).
To create a cyan fluorescent Golgi marker, ST–YFP clone was reconstructed using an amplified cerulean CFP SalI–BamHI fragment, yielding the pUC18-derived plasmid pOF64. A ClaI–BamHI fragment carrying the resulting ST–CFP-coding region and the 3′-untranslated end of nopaline synthases carrying an effective polyadenylation signal (3′nos) was further subcloned into a pDE1001-derived plant vector (98) under the transcriptional control of the CaMV 35S promoter, yielding plasmid (pTOF64). The latter was transformed into Agrobacterium strain C58C1Rif as described previously (44). Transgenic plants were produced using the standard leaf disc method as described earlier (55), and taking full advantage of the position effect accompanied with Agrobacterium-mediated stable plant transformation, transgenic lines were screened for low-to-moderate expression levels, permitting ideal imaging of Golgi bodies with barely detectable signals of ST–CFP in natural transit through the ER, where it is produced initially.
Tomato (Lycopersicon esculentum) EST clones closely homologous to AtSYP31, AtMEMB11, AtBS14 and AtSEC22 were obtained from Clemson University Genomics Institute (Clemson, SC, USA). The full-length coding region of LeSYP31 (accession number AW222275) was amplified using ClaI and BamHI as restriction sites for insertion into pLL4 (97) between the CaMV 35S promoter and the 3′-untranslated end of the nopaline synthase gene (3′nos). This created the chimerical expression plasmid pJB26 that was used to produce untagged natural WT LeSYP31 in plant cells.
The full-length coding region of SYP81 (accession number At1g51740, NP564597) was amplified from Arabidopsis thaliana complementary DNA (cDNA), and the fragment was inserted in the bacterial expression plasmid pAmy (93) through NcoI and BamHI restriction sites. The new plasmid was named pDS01.
Plasmids expressing other SNARE chimerical constructs, as LeMEMB11 (accession number BG130394), LeSEC22 (accession number BG123258) and LeBS14a (accession number AW033981), were based on pJB26, cloned between ClaI and BamHI sites and named pJB23, pJB24 and pJB25, respectively.
All SNARE constructs except SYP81 were N-terminally tagged with an HA peptide. The full-length coding region of HA–LeSYP31 and a truncated fragment lacking the C-terminal transmembrane domain (HA–LeSYP31ΔTMD) were cloned as ClaI–BamHI fragments into the expression plasmid pLL4 to obtain pJB01 (HA–LeSYP31) and pJB02 (HA–LeSYP31ΔTMD), respectively.
Based on pLL4, the expression plasmid pJB00 was generated expressing the HA tag within a multiple cloning site (ClaI–XhoI–HA tag–SacI–BamHI). By polymerase chain reaction (PCR) assembly, a ClaI–BamHI fragment was generated and cloned into pLL4.
Another truncated fragment of LeSYP31 lacking the N-terminal peptide, HA–LeSYP31ΔNT, was generated by PCR and cloned as a SacI–BamHI fragment into the expression plasmid pJB00. This yielded plasmid pJB03 expressing HA–LeSYP31ΔNT.
The construction of N-terminally HA-tagged LeBS14a, LeSEC22 and LeMEMB11 was performed in the same way, yielding pJB05 (HA–LeBS14a), pJB07 (HA–LeSEC22) and pJB10 (HA–LeMEMB11), respectively.
To engineer a LeSYP31 fusion to the C-terminal end of GFP (GFP–LeSYP31), an EcoRI–BglII fragment carrying the CaMV 35S promoter followed by the GFP-coding region was cut out of the GFP–SYP21 expression plasmid pOF22 described previously (44) and ligated together with an annealed BglII–ClaI linker into its plasmid pJB01 for fusion to LeSYP31. The linker was created by annealing the sense and antisense oligos, yielding a fragment encoding the symmetric linker peptide SerAlaGlyGlyAlaSer for optimal positioning and functional maintenance of the LeSYP31 function within the context of a GFP fusion. pJB01 was cut with EcoRI and ClaI, followed by dephosphorylation and subsequent ligation to both fragments. This yielded pJB12 expressing GFP–HA–LeSYP31.
To generate LeSYP31 N-terminal fused to GFP (LeSYP31–GFP), we amplified the LeSYP31-coding region and introduced a tetrapeptide (GlyGlyGlyAla) between the natural C-terminal Ala codon of LeSYP31 and the N-terminal Met codon of the GFP-coding region. The XhoI–NcoI fragment was ligated into the expression plasmid pCK(X/S)LTEV-EGFP (99) to obtain pJB13 expressing LeSYP31–GFP.
To generate the YFP–LeSYP31 fusion, the ClaI–BamHI fragment carrying the LeSYP31-coding region was cut out of the expression plasmid pJB01 and ligated into pOF21 (44) that was cut with ClaI and BamHI, followed by dephosphorylation, to replace the SYP21-coding region by that of LeSYP31. This yielded plasmid pJB21 (YFP–HA–LeSYP31).
The pJB21 plasmid was used to clone an N-terminal-fused YFP–SYP81. To achieve this, the plasmid was NdeI and ClaI cleaved, and the fragment containing the CaMV 35S promoter and the YFP sequence was isolated. The annealed sense and antisense oligos were used as a linker to create a new NcoI restriction site at the fragment, replacing ClaI and render further cloning. To complete the construct, pDS01 was cleaved with NcoI and NdeI and then ligated with the NdeI–ClaI fragment of pJB21 and the linker, obtaining an YFP–SYP81 construct (pDS02).
All the other YFP-tagged constructs were also based on pJB21 but cloned between ClaI and BamHI sites. ClaI–BamHI fragments carrying the LeMEMB11-, LeSEC22- and/or LeBS14a-coding region were cut out of their expression plasmids pJB23, pJB25 and/or pJB24 to replace the LeSYP21-coding region by that of LeMEMB11, LeSEC22 and/or LeBS14a, respectively, thus, creating the expression plasmids pJB31 (YFP–LeMEMB11), pJB32 (YFP–LeSEC22) and pJB33 (YFP–LeBS14a).
For Agrobacterium tumefaciens transient expression of YFP–LeSYP31, the 35S-YFP-HA-LeSYP31-3′nos chimeric gene of pJB21 was inserted as an EcoRI–HindIII fragment directly into the polylinker of the binary plant vector pDE1001 (98) cut with EcoRI and HindIII, followed by dephosphorylation. The resulting plasmid pJB22 works in a 2- to 3-day infiltration assay without selection.
The new plasmid (pJB22) was used to insert SYP81 as a BamHI–ClaI fragment, getting a new binary plant vector (pDS04) containing the chimeric YFP–SYP81 gene.
Plant material and transient protoplast expression procedure
Plants of Nicotiana tabacum cv., Petit Havana (101), were grown from surface-sterilized seeds in Murashige and Skoog medium (102) and 2% sucrose in a controlled room at 22°C with a 16-h daylength at a light irradiance of 200 mE/m2/second. Preparation of tobacco leaf protoplasts was done as described previously (44). Overnight digestion of yellow–green leaves, previously perforated using a needle bed (51), was followed by filtering the suspension through a 80-μm nylon mesh and brief washing of the cell debris with electroporation buffer [0.4 m sucrose (13.7%), 2.4 g/L HEPES, 6 g/L KCl and 600 mg/L CaCl2, brought to pH 7.2 with KOH] to release further protoplasts from the tissue remnants. The protoplast suspensions were then centrifuged in Falcon tubes (50 mL) for 15 min at 80 × gat room temperature in a swing-out rotor to prevent resuspension of the floating protoplast band. The pellet with dead cells and the underlying medium were removed and discarded using a peristaltic pump and a sterile Pasteur pipette until the band of floating (living) protoplasts reached the bottom. After resuspending the cells in 25 mL of electroporation buffer and a further centrifugation at 80 × gfor 10 min, the pellet and the underlying medium were removed again. This procedure was repeated twice by reducing the volume of electroporation buffer. In a last step, protoplasts were resuspended in electroporation buffer at a concentration of 5 × 106 protoplasts/mL. A total volume of 500 μL of the obtained protoplast mix was pipetted into a disposable 1-mL plastic cuvette and mixed with an appropriate amount of plasmid DNA or mixtures of plasmids previously dissolved in 100 μL of electroporation buffer. The protoplasts were electroporated with stainless steel electrodes at a distance of 3.5 mm using a complete exponential discharge of a 1000 μF capacitor charged at 160 V. After 30 min of absolute rest, electroporated protoplasts were removed from the cuvettes and transferred to 5-cm Petri dishes with 2 mL of transient expression (TEX) buffer (44). Protoplasts were then incubated for 24 h at 25°C in a dark chamber. Plasmid concentrations used are given in the figure legends. All incubations were performed for 24 h unless otherwise indicated.
Harvesting cells and medium from cell suspension were performed as described previously (44).
α-Amylase assay and secretory index determinations
Preparation of fractions and determination of extracellular (secreted) and intracellular α-amylase activities were performed as described before (44,93). The secretory index is defined as the ratio of extracellular to intracellular activities, whereby the total activity represents the sum of the α-amylase activities in the medium and the protoplasts (103).
Tobacco leaf infiltration procedure
Soil-grown N. tabacum cv. Petit Havana (101) or transgenic plants producing ST–CFP (unpublished data) were infiltrated with A. tumefaciens cultures (at optical density 0.1) as described previously (104) and analysed after various times indicated in the Results section and legends to figures 6, 10 and 11.
Production of SYP81 antibodies
The cytosolic domain of AtSYP81 (without the SNARE motif to prevent cross reactions) was amplified with the following oligonucleotides: BamHI sense: 5′-CGCGGATCCATGCGAGATTCAGAGAC-3′, SalI antisense: 5′-GGATCCGTCGACTTACGGCATAGCTCTGTT-3′. The PCR product was then ligated into the pGEX-4T3 vector previously cut with BamHI/SalI. The cleaved vector was dephosphorylated prior to ligation. One millimolar IPTG was used to induce the expression of the glutathione S-transferase (GST) fusion protein in BL21 E. coli for 3 h. The inclusion bodies containing GST–SYP81 were solubilized using an established protocol with N-lauroylsarcosine (105).
The recombinant protein was then affinity purified with GST–Sepharose. SDS–PAGE was performed to separate the protein band of interest, which was then excised, the gel fragments electroeluted and the fusion protein finally dialyzed. Three hundred milligrams of lyophilized GST–SYP81 was used for commercial immunization in rabbits (Eurogentec).
Protein extraction and gel blot analysis
Protein analysis of both culture medium and cell pellet obtained from 2.5 mL was performed as described previously (44) to obtain a 10-fold higher concentration compared with the original volume of cell suspension. Protein gel blots and immunodetection were performed as described previously (106) using rabbit polyclonal antiserum raised against GFP (1:2000 dilution; Molecular Probes) and mouse monoclonal antiserum raised against the HA tag (1:5000 dilution; Covance).
Infiltrated tobacco leaf squares (0.5 × 0.5 cm) were mounted in tap water with the lower epidermis facing the cover glass (22 × 50 mm; no. 0). Hundred microlitres of protoplast suspension was pipetted in an area (10 × 15 mm) bordered with a frame of 100-μm thick plastics isolating tape on a slide to protect the protoplasts of pressure. The area was covered with a cover slip (24 × 32 mm). Cells were observed under a Zeiss Axiovert LSM510 Meta microscope using a C-Apochromat ×63/1.2 W corr water immersion objective. Special settings were designed for observing single expression and coexpression with different XFP constructs. For imaging the coexpression of CFP/YFP or GFP/YFP, excitation lines of an argon ion laser of 458 nm for CFP/GFP and 514 nm for YFP were used alternately with line switching on the multitrack facility of the microscope. Fluorescence was detected using a 545-nm dichroic beam splitter and a 480- to 520-nm bandpass filter for GFP and a 560- to 615-nm bandpass filter for YFP. For imaging of the coexpression of YFP and GFP constructs, excitation lines of an argon ion laser of 458 nm for GFP and 514 nm for YFP were used alternately with line switching on the multitrack facility of the microscope. Fluorescence was detected using a 545-nm dichroic beam splitter and a 475- to 525-nm bandpass filter for GFP and a 560- to 615-nm bandpass filter for YFP. For imaging of the coexpression of YFP/RFP and GFP/RFP, fluorescence was detected by the metadetector using main beam splitter 405/514 and 488/543. Fluorophores were excited by line switching in the multitracking mode of the microscope. Double detection of YFP and RFP was performed by excitation of YFP at 514 nm and emission at 550–571 nm and by excitation of RFP at 543 nm and emission at 603–646 nm. Double detection of GFP and RFP was performed at excitation at 488/543 nm and an emission at 496–529 and 596–646 nm. Pinholes were adjusted to 1 Airy Unit for each wavelength (108, 119 and 138 μm). In all cases, post-acquisition image processing was performed with the LSM 5 image Browser (Zeiss version 188.8.131.526).
Immunogold electron microscopy
Six-day-old Arabidopsis root tips were excised from entire plants and submerged in 140 mm sucrose, 7 mm trehalose and 7 mm Tris buffer (pH 6.6) (for 10 min). Four to five submerged root tips were then mounted onto planchettes and frozen in a high-pressure freezer (HPF010; Bal-Tec). Freeze substitution was then performed in a Leica AFS freeze substitution unit (Leica) in dry acetone supplemented with 10% methanol and 0.3% uranyl acetate at −85°C for 20 h before gradually warming up to −50°C over an 19-h period. After washing twice in 100% ethanol for 60 min, the roots were infiltrated and embedded in Lowicryl HM20 at −35°C and polymerized for 3 days at the same temperature with ultraviolet (UV) light in the freeze substitution apparatus (107). To increase sectioning quality, the blocks were then hardened with UV light for another 4 h at room temperature. Ultrathin sections were cut on a Leica Ultracut S (Leica) and incubated with antibodies against SYP81 at a dilution of 1:70, followed by incubation with 10-nm gold-coupled secondary antibodies (BioCell GAR10; BioCell) at a dilution of 1:50 in PBS supplemented with 1% BSA. Sections were post-stained with aqueous uranyl acetate/lead citrate and examined in a Philips CM10 transmission electron microscope operating at 80 kV. Negatives were scanned (Epson Perfection 4990 Photo) and adjusted in size, contrast and brightness using Photoshop (Adobe Systems) to improve visibility of gold particles.
Yeast cells were grown at 30°C in yeast peptone dextrose, yeast peptone galactose medium, synthetic dextrose (SD) or synthetic galactose (SG) media lacking the appropriate amino acids. Yeast transformations were performed using lithium acetate (108).
The DNA sequences corresponding to the open reading frames of SYP31 and SYP81 were amplified with oligo primers and the polymerase chain reaction using LeSYP31 and AtSYP81 cDNAs as templates. The DNA fragment corresponding to the SYP31-coding sequence was amplified using the sense primer 5′-ACCAGAATTCATGCCTGTGAAAGTAGCAAGTG-3′ and the antisense primer 5′-ACCAGGATCCCTATGCCACGAAAAATAG-3′ and cloned as an EcoRI–BamHI fragment into EcoRI–BamHI-digested pRS414 (TRP1 and CEN6) containing the triosephosphate isomerase (TPI) promoter and iso-1 form of cytochrome c (CYC1) transcription terminator sequence derived from pYES2 (Invitrogen), generating pTPISYP31. The DNA fragment corresponding to the SYP81-coding sequence was amplified using the sense primer 5′-ACCACAATTGATGGCATCGAGATTCAGAG-3′ and the antisense primer 5′-ACCAGGATCCTTAACTGTACCAATCC-3′ and cloned as an MfeI–BamHI fragment into EcoRI–BamHI-digested pRS415 (LEU2 and CEN6) containing the TPI promoter and the CYC1 transcription terminator sequence, generating pTPISYP81. The authenticity of pTPISYP31 and pTPISYP81 was verified by DNA sequence determination.
pTPISYP31 was transformed into SFY1 (MATa his3-Δ200 leu2-3,-112 ura3-52 trp1-1 lys2-801 sed5::LEU2 containing SED5, CEN6 and URA3) and SFY2 (MATa his3-Δ200 leu2-3,-112 ura3-52 trp1-1 sed5::LEU2 containing pGAL1SED5 and pRS316) cells. pTPISYP81 was transformed into BY4741 (MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0) and UFY1 [MATa his3-Δ200 leu2-3,-112 ura3-52 trp1-1 ufe1::TRP1 containing pUFE1 (UFE1, CEN6 and URA3)] cells. Transformants were selected on SD or SG media lacking the appropriate amino acids. The results of yeast cell transformation were assessed after 3- to 5-day incubation at 30°C. Transformants derived from SFY1 and UFY1 cells were patched onto plates containing 1 mg/mL 5-flouroorotic acid (Sigma) and assessed for growth following 3–5 days of incubation at 25°C.
Accession numbers and sequence alignments
Sequence data from this article can be found in the NCBI nucleotide sequence databases (includes GenBank) under accession numbers AW033981 (LeBS14a), BG123258 (LeSEC22), BG130394 (LeMEMB11), AW222275 (LeSYP31) and At1g51740 (AtSYP81). The protein sequence analysis and alignment of tomato and Arabidopsis SNARE proteins used in this study against yeast and humans were performed with BioEdit(109). For multiple sequence alignments, ‘clustal-w’ was used (available at http://www2.ebi.ac.uk/clustalw/help.html).
- Top of page
- Materials and Methods
- Supporting Information
Financial support from the German Research Council (DFG Ro 440/11-3) to D. G. R. is gratefully acknowledged. D. K. B. gratefully acknowledges support from the Research Grants Council (RGC) of Hong Kong. Grants awarded to J. D. from the Biotechnology and Biological Sciences Research Council (BBSRC) and the European Union (Training network HPRN-CT-2002-00262: BioInteractions and FP6-PHARMA-PLANTA consortium) have supported E. H., J. B. and D. S. while working in Leeds. O. F. is indebted to the BBSRC and the EU for a PhD studentship. An EU Marie Curie Training grant also contributed to the bench exchanges of E. H. and J. B. in Leeds. We thank Goretti Virgili Lopez and Andreas Nebenführ for providing the plasmids expressing RFP–TMD23 and Man1–GFP, respectively.
- Top of page
- Materials and Methods
- Supporting Information
- 104The green fluorescent protein (GFP) as a reporter in plant cells. In: HawesC and Satiat-JeunemaitreB, editors. Plant Cell Biology. Oxford, UK: Oxford University Press; 2001, pp. 127–142., .
- 109BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nul Acids Symp Ser 1999;41:95–98..
- Top of page
- Materials and Methods
- Supporting Information
Additional Supporting Information may be found in the online version ofthis article:
Figure S1: Overexpression of SYP31 has both an immediate and a long-term inhibitory effect on secretion. A–C) Co-electroporations of tobacco mesophyll protoplasts with plasmid DNAs encoding for α-amylase (5 μg) and HA-tagged SYP31 (5 μg) as well as control electroporations with α-amylase alone were performed over a 24-h expression period with samples being removed at the times indicated. In addition to the calculation of the secretory index (A), the total activity of α-amylase (B), the results of protein gel blotting for secreted α-amylase and the expression of the SNARE are given (C).
Figure S2: Control organelle marker signals in tobacco protoplasts. A–E) Electroporations of tobacco mesophyll protoplasts were performed as in Figure 2 but with standard marker proteins for the Golgi apparatus (A), the ER (C and E) and the plasma membrane (D). In addition, a protoplast from leaves stably expressing the Golgi marker ST–GFP is presented (B).
Figure S3: Movie showing Golgi localization of YFP–SYP31 and ST–CFP in tobacco leaf epidermis.
Figure S4: Movie showing the mobility of BFA-resistant YFP–SYP31 structures in tobacco leaf epidermis.
Figure S5: BFA resistance of SYP31 structures is independent of duration of SYP31 expression. A–E) Tobacco leaf protoplasts were electroporated with 5 μg plasmid DNA encoding for Man1–GFP (A, D and G), GFP–SYP31 (B, C, E and F) or SYP31–GFP (H and I) and observed in the CLSM after 7-h (A–D) or 24-h (E–I) expression. BFA (end concentration10 μg/mL) was added to the medium after 6-h (A–D) and 23-h (E–I) expression. Whereas the Golgi marker protein Man1–GFP is redistributed into the ER (D), the punctate fluorescent signals for GFP–SYP31 remain intact and become larger with increased expression (compare D with B and E). In the case of SYP31–GFP, BFA causes only part of the signal to move into the ER.
Figure S6: Secretion index measurements for HA-tagged SNAREs (BS14a, SEC22 and MEMB11). In agreement with the secretory index measurements for WT and YFP-tagged SNAREs, only overexpressed MEMB11 is effective in inhibiting secretion.
Figure S7: YFP–SYP81 localizes to the ER in tobacco mesophyll protoplasts. YFP–SYP81 was co-electroporated with ER markers (GFP–HDEL and calnexin–GFP). Although the YFP–SYP81 fluorescent signal was in the main punctate, the punctae were always distributed on the ER.
Table S1: SNARE sequence similarities (Tomato, Arabidopsis, Yeast and Human)
Table S2: SNARE sequence similarities (Arabidopsis, Yeast and Human)
Table S3: List of constructs used in this study together with sense and antisense oligonucleotides
Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.
|TRA_803_sm_FigureS1.jpg||726K||Supporting info item|
|TRA_803_sm_FigureS2.jpg||432K||Supporting info item|
|TRA_803_sm_FigureS3.jpg||361K||Supporting info item|
|TRA_803_sm_FigureS4.jpg||247K||Supporting info item|
|TRA_803_sm_FigureS5.jpg||768K||Supporting info item|
|TRA_803_sm_FigureS6.jpg||681K||Supporting info item|
|TRA_803_sm_FigureS7.jpg||637K||Supporting info item|
|TRA_803_sm_TableS1-3.doc||71K||Supporting 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.