Cytokinesis and cell polarity are supported by membrane trafficking from the trans-Golgi network (TGN), but the molecular mechanisms that promote membrane trafficking from the TGN are poorly defined in plant cells. Here we show that TRAPPII in Arabidopsis regulates the post-Golgi trafficking that is crucial for assembly of the cell plate and cell polarity. Disruptions of AtTRS120 or AtTRS130, two genes encoding two key subunits of TRAPPII, result in defective cytokinesis and cell polarity in embryogenesis and seedling development. In attrs120 and attrs130, the organization and trafficking in the endoplasmic reticulum (ER)–Golgi interface are normal. However, post-Golgi trafficking to the cell plate and to the cell wall, but not to the vacuole, is impaired. Furthermore, TRAPPII is required for the selective transport of PIN2, but not PIN1, to the plasma membrane. We revealed that AtTRS130 is co-localized with RAB-A1c. Expression of constitutively active RAB-A1c partially rescues attrs130. RAB-A1c, which resides at the TGN, is delocalized to the cytosol in attrs130. We propose that TRAPPII in Arabidopsis acts upstream of Rab-A GTPases in post-Golgi membrane trafficking in plant cells.
Rab proteins, a family of small monomeric GTPases, have well-established roles in regulating tethering/docking of vesicles to a specific target membrane (Zerial and McBride, 2001). In Arabidopsis, members of several Rab-A subclasses (the closest homologs of Rab11) have been shown to perform important regulatory roles in exocytic trafficking (Preuss et al., 2004, 2006; de Graaf et al., 2005; Blanco et al., 2009; Szumlanski and Nielsen, 2009) and the targeted delivery of materials to the growing cell plate (Chow et al., 2008; Boutte et al., 2010). However, the regulation of the function of Rab-A proteins in plant membrane trafficking remains elusive. In yeast, trafficking between the trans-Golgi, early endosome and plasma membrane also requires a pair of Rab11 proteins: the Ypt31/32 pair (Benli et al., 1996; Chen et al., 2005). In animal cells, Rab11 acts at the recycling endosome in the delivery of receptors and transporters to cell surfaces (Ren et al., 1998; Emery et al., 2005; Gonzalez et al., 2007), and performs important functions in the abscission of daughter cells in the late stages of cytokinesis (Wilson et al., 2005; Prekeris and Gould, 2008). Genetic evidence suggested that in yeast, TRAPPII, a modified version of TRAPPI (a seven-subunit transport protein particle: Trs20, Trs23, Trs31, Trs33, Trs85, Bet3 and Bet5), with the addition of three specific subunits Trs65, Trs120 and Trs130, acts upstream of Ypt31/32 (Yamamoto and Jigami, 2002; Sciorra et al., 2005). In yeast cells, TRAPPII can serve as a guanine nucleotide exchange factor (GEF) for Ypt31/32 proteins (Jones et al., 2000; Morozova et al., 2006; Sacher et al., 2008). In higher eukaryotic cells, Trs65 is missing, but the two TRAPPII-specific subunits, Trs120 and Trs130, are conserved in all sequenced eukaryotic genomes (Cox et al., 2007). In Drosophila, the Trs120 homolog Bru is required for the proper localization of Rab11, and genetically interacts with Rab11 and Rab11 effector PI4Kβ1 during cleavage furrow ingression in dividing male meiotic cells (Robinett et al., 2009). However, recent data suggest that, in mammalian cells, Trs130 co-localizes with early Golgi markers, and can serve as a GEF for Rab1 in early Golgi trafficking (Yamasaki et al., 2009).
The role of TRAPPII in plant membrane trafficking has not been investigated. In Arabidopsis, there are single homologs of yeast Trs120 and Trs130, At5g11040 and At5g54440, respectively. In this paper, we refer to them as AtTRS120 and AtTRS130, respectively. Using a combined approach of genetics, molecular biology, and optical and transmission electron microscopy, we provide evidence that TRAPPII in Arabidopsis acts upstream of Rab-A GTPases. The function of TRAPPII is required for the proper transport of proteins in post-Golgi trafficking pathways to the growing cell plate in mitotic active cells, and for the selective transport of proteins to the plasma membrane in polarized plant cells, but not for ER–Golgi as well as biosynthetic and endocytic vacuolar transport. The function of TRAPPII is crucial for normal plant development.
Disruption of either AtTRS120 or AtTRS130 leads to similar defects in embryogenesis and post-embryonic organ development
Based on a survey on the Expression Atlas of Arabidopsis Development (Schmid et al., 2005) and our RT-PCR analyses, we noted that AtTRS120 and AtTRS130 were expressed ubiquitously in almost all plant tissues (Figure S1). We therefore identified four T-DNA insertion mutant lines for AtTRS120 (designated as attrs120-1, attrs120-2, attrs120-3 and attrs120-4) and one for AtTRS130 (designated as attrs130) (Figure 1a) for further functional analyses. Semi-quantitative RT-PCR analyses indicated that attrs120-1, attrs120-2, attrs120-3 and attrs130 are transcriptional knock-out mutants in the respective genes, and that attrs120-4 is a transcriptional knock-down mutant (Figure 1b, only attrs120-2, attrs130 and attrs120-4 are shown).
We noted that knock-out alleles of attrs120 and attrs130 exhibited very similar developmental defects. Embryos that were homozygous for these alleles were able to survive early embryogenesis, but in the late stage of embryogenesis, mutant hypocotyls did not elongate properly and bent abnormally. This was paired with distorted cotyledons (Figure 1c, only attrs120-2 and attrs130 are shown), which often resulted in a curled mature seed (Figure 1c). These mutant embryos were able to germinate, but with no differentiation of true leaves, and root growth was arrested at the very early seedling stage (Figure 1d). In siliques (n = 42) of heterozygous attrs120-2 or attrs130, no aborted ovules or aborted embryos were found. However, only 5–10% of embryos developed waved hypocotyls, distorted cotyledons and were seedling lethal (Figure 1c,d). These phenotypes co-segregated with the T-DNA insertions in AtTRS120 or AtTRS130. The result indicates that the observed defects in embryogenesis and seedling development are genetically linked to the T-DNA insertions in the respective genes. Male transmission is likely to be defective in the mutants.
The weak allele attrs120-4 was able to complete embryogenesis with no detectable defect. However, the growth of the primary root was reduced in attrs120-4 (Figure 1e). Under our growth conditions, the average root length of attrs120-4 10 days after germination was approximately two-thirds of the length of wild-type roots (2.7 ± 0.25 cm for attrs120-4 roots versus 4.0 ± 0.3 cm for wild-type roots). Adult attrs120-4 plants were also dwarfed (Figure 1f).
When AtTRS120-YFP was transformed back to attrs120-2 and attrs120-4, and AtTRS130-YFP was transformed back to attrs130, the developmental phenotypes of attrs120-2 (Figure 1g, compare seedling #3 with #2), attrs130 (Figure 1h, compare seedling #3 with #2) and attrs120-4 were complemented, confirming that the phenotypes were caused by the loss of the respective proteins. The molecular complementation was further supported by genetic complementation tests between attrs120 alleles. When attrs120-1, attrs120-2 and attrs120-3 were crossed with each other, and when attrs130 was crossed with attrs120-2, F1 plants heterozygous for either allele of attrs120 still presented typical mutant phenotypes (Figure S2b,d), F1 plants heterozygous for attrs120-2 and attrs130 had a wild-type phenotype (Figure S2c). We therefore used attrs130, attrs120-2 and attrs120-4 for subsequent analyses.
Mutations in AtTRS120 and AtTRS130 affect cytokinesis and cell polarity
To explore the cellular basis of developmental defects revealed in attrs120-2 and attrs130, we analyzed the cellular morphology of hypocotyls and cotyledons in late embryos using propidium iodide. Strikingly, in both attrs120-2 and attrs130, a considerable number of cells had unfinished cell plates (Figure 2a,b arrows) and were multinucleated (Figures 2b arrowheads; Video Clips S1–S3). This observation was further confirmed by toluidine blue staining of thin sections of cotyledons (Figure S3). In wild-type cotyledons, pavement cells had many interlocked, jigsaw puzzle-shaped lobe structures (Figure 2a,b), but in both attrs120-2 and attrs130, pavement cells were largely rectangular or round, with very few concave bends (Figure 2a,b). We also analyzed the cellular morphology of root-tip cells of young seedlings after germination. In root-tip cells of attrs120-2 and attrs130, incomplete cell plates were also observed (Figure 2c, red arrows). Furthermore, compared with well-patterned cell files in wild-type root tips (Figure 2d; Video Clip S4), the cell file alignment in root tips of attrs120-2 and attrs130 was disordered, and many cells were abnormally shaped (Figure 2d; Video Clips S5,S6). Together, these results indicate that cytokinesis as well as the subsequent cell shaping/growth that are crucial for generating a basic body plan (e.g. cotyledons) and post-embryonic organ formation (e.g. roots), are compromised in both attrs120-2 and attrs130.
Abnormal vesicular-tubular structures accumulate in the cytoplasm of attrs120-2 and attrs130
In plants, both cell division and cell polarity are supported by membrane trafficking from the TGN to deliver membrane and cell wall materials to the cell plate in mitotic active cells (Chow et al., 2008; Boutte et al., 2010), and to selected regions of the cell wall/plasma membrane in polarized plant cells (Preuss et al., 2004, 2006; Lee et al., 2008; Szumlanski and Nielsen, 2009). To understand the subcellular basis of the cytokinesis and cell polarity defects observed in attrs120-2 and attrs130, we examined the endomembrane system of seedling cells of attrs120-2 and attrs130 at the ultrastructural level by transmission electron microscopy (TEM). We found that the structure of the endoplasmic reticulum (ER) [compare the ER in Figure 3b (attrs130) with that in 3a (wild type); also see Figure 4a,b for the overall ER morphology] and Golgi stacks [compare Golgi bodies in Figure 3b,c (attrs130) with those in 3a (wild type); also see Figure 4c,d for the overall Golgi distribution] appeared normal in both attrs120-2 and attrs130 (here and hereafter we will only show data from either attrs120-2 or attrs130, as identical results were always obtained for attrs120-2 and attrs130 in our analysis). However, TEM revealed that there was abnormal accumulation of vesicles in the cytoplasm of the mutant cells (compare Figure 3b–d with 3a, arrows). In wild-type cells, roughly 1.7 vesicles μm−2 (Figure 3e) were found in the cytoplasmic regions on transmission electron micrographs (a total of 66 μm2 cytoplasm not occupied by organelles was examined), but in attrs130 mutant cells, more than twice the number of vesicles (approximately 3.7 vesicles μm−2; Figure 3e) were observed (a total of 147 μm2 of cytoplasm was examined). There was no significant enlargement of vesicles in the mutants, as vesicles in both the wild type (Figure 3a) and the mutants (Figure 3b–d) had diameters ranging from 60 to 110 nm. However, we found that in wild-type cells, most of the vesicles (approximately 85%; Figure 3f) were within a 0.5-μm radius from the center of the Golgi apparatus, and many of them were located at the trans- side of the Golgi (Figure 3a), probably representing vesicles that have just budded off from the TGN. In attrs130 mutant cells, however, the majority of the vesicles (approximately 80% of vesicles; Figure 3f) was located away from the Golgi, and accumulated elsewhere in the cytoplasm (Figure 3b–d, arrows). Some vesicles were found to be associated with tubular structures (Figure 3b–d, arrowheads), but structures resembling typical TGN (Figure 3a) were rarely seen in mutant cells (Figure 3b–d).
Secretion of secretory GFP, but not vacuolar transport is affected in attrs120-2 and attrs130
TRAPPII in yeast cells acts in post-Golgi trafficking pathways (Jones et al., 2000; Morozova et al., 2006), but in mammalian cells TRAPPII acts in early Golgi trafficking (Yamasaki et al., 2009). The TEM results described above motivated us to use an in vivo imaging approach to investigate which step of the vesicle trafficking pathways was impaired in mutant cells. When the ER marker GFP-HDEL and the trans-Golgi marker ST-YFP were crossed into both attrs130 and attrs120-2, we noted that the morphology of the ER and Golgi was not affected in the mutants. GFP-HDEL revealed a polygonal ER network in the wild type (Figure 4a) and attrs120-2 (Figure 4b). ST-YFP was still targeted to the Golgi with no retention in the ER in attrs120-2, resulting in punctate structures (compare Figure 4d with 4c). The motility of Golgi in the mutants also appeared to be normal [compare Video Clip S8 (attrs120-2) with S7 (wild type)]. Furthermore, the biosynthetic vacuolar transport of secN-Rm-2A, an RFP-based vacuolar transport marker (Samalova et al., 2006), was not inhibited in attrs130 either (Figure 4e–h). FM4-64 is a lipophilic styryl fluorescent dye used to track endocytic transport from the plasma membrane to the tonoplast. In root tip cells of Arabidopsis, FM4-64 (5 μm) is first internalized into the TGN/EE positive for Rab-A2/A3 and VHA-a1 (a subunit of the vacuolar V-ATPase complex). Roughly 1 h after internalization, the dye is transported to the late endosome (LE) positive for RAB-F GTPases. The dye reaches the tonoplast in roughly 2 h (Chow et al., 2008; Dettmer et al., 2006; Figure 4i). Interestingly, the internalization and transport of FM4-64 to tonoplast membranes in attrs120-2 or attrs130 were not delayed (Figure 4i,j). This indicates that, similar to biosynthetic vacuolar transport of secN-Rm-2A, endocytic transport of FM4-64 from the plasma membrane to the tonoplast was not impaired in the mutants.
secGFP is a secretory variant of GFP that is synthesized in the ER, but is transported via the Golgi to the cell wall, where the condition is suboptimal for GFP fluorescence (Zheng et al., 2004, 2005a). Fluorescence of secGFP in wild-type Arabidopsis seedlings was dim, but could sometimes be observed in the cell wall (Figure 5a), and a few cells (approximately 4%, n = 323) showed an intracellular secGFP signal (Figure 5c). However, when secGFP was introgressed into attrs130 and attrs120-2, and imaged using identical confocal microscope settings, a relatively strong GFP signal was detected in mutant cells of attrs120-2 (Figure 5d) and attrs130, with approximately 24% of cells (n = 358) exhibiting an intracellular secGFP signal. Higher magnification confocal analysis revealed that the intracellular secGFP marked punctate organelles (Figure 5b). When attrs120-2 expressing secGFP was stained with FM4-64, we noted that most secGFP punctae (approximately 95%, n = 132) were stained by early FM4-64 (Figure 5e, white arrows) within 60 min of administering the dye. The secGFP marked compartment was also partially co-localized and/or closely associated with the VHA-a1-mRFP compartment (Figure 5f, white arrows). The results indicate that secGFP in the mutants is retained in TGN/EE. Based on these data we conclude that in attrs120-2 and attrs130, ER–Golgi as well as biosynthetic and endocytic vacuolar transport is not affected, but Golgi transport to the cell wall is inhibited at the level of TGN.
Overexpression of the GTP-locked form of RAB-A1c partially suppresses attrs130
In Arabidopsis, RAB-A2a is localized to a population of TGN/EE (Chow et al., 2008). In our functional analysis of RAB-A1c, we found that GFP-RAB-A1c was co-localized (Figure S4a), and moved together with YFP-RAB-A2a (Video Clip S9). Furthermore, GFP-RAB-A1c was also extensively [93% of GFP-RAB-A1c (n = 475)] co-localized with FM4-64 within 1 h of staining (Figure S4b). Remarkably, AtTRS130-YFP, a functional AtTRS130 fusion (Figure 1h) was extensively co-localized with GFP-RAB-A1c (Figure 6a, white arrows). Furthermore, when constitutively active RAB-A1c(Q72L) was expressed under the dexamethasone (dex)-inducible synthetic promoter pOP6 (Craft et al., 2005) in attrs130, we found that RAB-A1c(Q72L) was able to partially rescue attrs130 (Figure 6b–j). The roots of attrs130::RAB-A1c(Q72L) were three or four times longer (Figure 6e,j) than those of un-induced attrs130::RAB-A1c(Q72L) (Figure 6d,j), un-induced (Figure 6b,j) or induced (Figure 6c,j) non-transformed attrs130 seedlings, but was still shorter than those of the wild type (Figure 6f,g,j). Expression of RAB-A1c(Q72L) in wild-type seedlings did not enhance the root growth of wild-type plants (Figure 6h–j). Chow et al. (2008) reported that an inactive form of RAB-A2a is largely delocalized to the cytoplasm. In wild-type cells (n = 114), GFP-RAB-A1c was mainly seen in TGN as numerous small punctate structures (Figure 7a), but when GFP-RAB-A1c in attrs130 cells (n = 83) was imaged with the same confocal microscope settings [40×/1.3 oil objective, pinhole of 66 μm in diameter, excitation with 488 nm (13.8% of argon laser power)], GFP-RAB-A1c showed a diffuse localization in the cytoplasm in attrs130 (Figure 7b; Video Clip S10) cells, suggesting that RAB-A1c was not activated properly in the mutant. Taken together, these results suggest that TRAPPII in Arabidopsis acts upstream of Rab-A1c, probably as a GEF.
In attrs130, GFP-RAB-A1c-positive punctate organelles, some of which appeared to be larger or agglomerated (Figure 7b, arrow), were still visible. We also noted in our FM4-64 endocytic transport study that, in root epidermal cells of attrs120-2 and attrs130 at early time points, structures labeled by FM4-64 appeared larger than those in wild-type cells (compare Figure 8f with 8e). When attrs130 expressing GFP-RAB-A1c was stained with FM4-64, we noted an extensive co-localization (approximately 92%, n = 156) between remnant punctate organelles marked by GFP-RAB-A1c and features labeled by internalized FM4-64 within 60 min (Figure 7c, white arrows). This suggests that the remnant organelle marked by GFP-RAB-A1c in the mutant is TGN/EE in nature, and that TGN/EE is disorganized in the mutants.
GFP-RAB-A1c as well as FM4-64 accumulates abnormally in the mitotic cells of attrs120 and attrs130
It has been increasingly clear that, in plants, Rab-A proteins play important roles not only in post-Golgi trafficking to the cell wall, but also to the cell plate in mitotic active cells (Chow et al., 2008; Boutte et al., 2010). Similar to YFP-RAB-A2a and FM4-64, GFP-RAB-A1c was also relocated to a disc-like flat structure resembling the cell plate in mitotic active cells in root tips (Figure S4c) and shoot meristems (Figure S4d). Therefore, we examined the assembly of the cell plate using GFP-RAB-A1c and FM4-64 in root-tip cells of attrs120-4 and attrs130. In wild-type root tips (n = 73 seedlings), disc-like flat structures that resemble growing cell plates were frequently observed in mitotic active cells with GFP-RAB-A1c (Figure 9a,c arrow) and FM4-64 (Figure 9e arrow). In the root tips of attrs120-4 seedlings (n = 89 seedlings), cell plates were often patchy (Figure 9b, arrow). In attrs130 (n = 67 seedlings), irregular aggregation of GFP-RAB-A1c, as well as FM4-64 punctates (Figure 9d,f, arrowheads) was often seen around a cell plate-like structure. Our results were consistent with those of Thellmann et al. (2010), who showed that in attrs120 mutants, the assembly of KNOLLE, a cytokinesis-specific SNARE into the cell plate is defective. These results indicate that, in mitotically active root-tip cells of TRAPPII mutants, materials destined to cell plates were abnormally accumulated, and the assembly of the cell plate was affected.
The establishment or maintenance of polar localization of PIN2, but not PIN1, is significantly affected in attrs130 and attrs120-2
In Arabidopsis root tips, PIN1, an auxin transport carrier, is localized to the basal plasma membrane in root stele cells; PIN2, another auxin transport carrier, is localized to the apical plasma membrane in root epidermal cells (Grieneisen et al., 2007). Recent data suggest that PIN1 and PIN2 may take different pathways to achieve a polar distribution (Jaillais et al., 2007; Teh and Moore, 2007; Robert et al., 2008). To examine if the TRAPPII-regulated trafficking pathway plays a role in PIN1 and PIN2 trafficking, we investigated the localization of PIN1-GFP and PIN2-GFP in root-tip cells of the attrs130 and attrs120-2 mutants. In the wild type, PIN1 was mainly localized on the basal part of the plasma membrane in root stele cells (Figure 8a), and PIN2 was mainly located on the apical part of the plasma membrane in root epidermal cells (Figure 8c, arrow). In both attrs120-2 (Figure 8b) and attrs130, although PIN1-GFP highlighted a disorganized cell file, the fusion protein was targeted to the basal part of the plasma membrane in root stele cells. On the other hand, PIN2-GFP was distributed with less polarity in the root epidermal cells of the mutants, as a strong signal was found at other plasma membrane domains (Figure 8d, arrowheads), in addition to the apical plasma membrane (Figure 8d arrow). Furthermore, intracellular punctate structures (Figure 8d inset, arrowhead), larger than those occasionally seen in wild-type cells (Figure 8c), were also observed in our mutants. When attrs120-2 expressing PIN2-GFP was stained with FM4-64, we found that in root epidermal cells intracellular punctae of PIN2-GFP (91%, n = 342) were marked by early internalized FM4-64 (Figure 8f, white arrows). Furthermore, the intracellular PIN2 compartment was also partially co-localized and/or closely associated with VHA-a1-mRFP (Figure 8g, white arrows), suggesting that PIN2-GFP is retained in TGN/EE in the mutants. According to Robert et al. (2008), trafficking of PIN2 is sensitive to Endosidin1, a drug that affects the morphology of a population of TGN. We observed that in cells treated with Endosidin1, but not with Wortmannin [an inhibitor that impairs endocytosis and also induces swelling of multivesicular bodies (MVBs); Emans et al., 2002; Wang et al., 2009), AtTRS130 tended to aggregate (Figure 10)].
In plants, auxin flows downwards from the shoot apical region to the root tip, and a stable auxin maximum is established and maintained at the root tip (Grieneisen et al., 2007). The expression of GFP under the control of the auxin-responsive DR5 promoter has been used to monitor the auxin maxima (Grieneisen et al., 2007). When DR5:GFP was expressed in attrs120-2 or attrs130, we found that the auxin maximum displayed an altered pattern in terms of the cells involved. In the wild type, the auxin response peak was seen only in several layers of root-cap cells in front of the quiescent center (Figure S5a,e), whereas in root tips of the mutants, the GFP signal was observed in a wider region of the root tip (Figure S5b,f), indicating that the auxin distribution is impaired in the mutant.
Role of TRAPPII in post-Golgi trafficking
The function of TRAPPII and its mode of action in membrane trafficking have been debated in yeast, Drosophila and mammalian cells (Jones et al., 2000; Morozova et al., 2006; Sacher et al., 2008; Robinett et al., 2009; Yamasaki et al., 2009). In this study, we provide evidence that TRAPPII in Arabidopsis acts in post-Golgi trafficking pathways. It appears that the function of TRAPPII is not required for the biosynthetic vacuolar transport of SecN-Rm-2A and endocytic transport of FM4-64 to the tonoplast, but is required for the secretion of secGFP, polar targeting of PIN2 to the plasma membrane and the assembly of the cell plate. The rescue of attrs130 by RAB-A1c(Q72L) and delocalization of GFP-RAB-A1c in the mutants suggest that TRAPPII in Arabidopsis acts upstream of Rab-A GTPases, potentially as a GEF. In mammalian cells, the Golgi is linked to the ER with a highly mobile ER–Golgi intermediate compartment (ERGIC; Appenzeller-Herzog and Hauri, 2006). However, in yeast, plants and Drosophila, Golgi stacks exhibit a close spatial association with transitional ER sites (Brandizzi et al., 2002; Kondylis and Rabouille, 2003; Matsuura-Tokita et al., 2006). Yamasaki et al. (2009) proposed that the difference in the architecture of the ER–Golgi pathway may be the underlying reason for the different action of TRAPPII in budding yeast and mammalian cells, and that the ERGIC in mammalian cells may be equivalent to the Golgi in yeast. If this is true, the action of TRAPPII in post-Golgi trafficking in yeast (Jones et al., 2000; Morozova et al., 2006), plant (this study) and Drosophila (Robinett et al., 2009) may be expected.
We noted that when GTP-locked RAB-A1c(Q72L) was overexpressed in attrs130, the developmental phenotype of attrs130 could only be partially rescued. One possible explanation is that TRAPPII might have an additional function, e.g. serve as a tethering factor for secretory vesicles (Sacher et al., 2008), that cannot be suppressed by the overexpression of RAB-A1c(Q72L). In addition, unlike their animal counterparts, Rab-A proteins in plants proliferated spectacularly during the evolution of the plant kingdom (Woollard and Moore, 2008). It is therefore likely that, in the absence of TRAPPII, multiple members of the Rab-A subfamily are improperly activated, so RAB-A1c(Q72L) could only partially complement the defect of attrs130. In yeast, the expression of either Ypt31 or Ypt32 suppresses a growth defect of trs130Δ mutant cells, but the extent of the suppression is different between Ypt31 and Ypt32, as the two proteins could act at distinct but overlapping pathways (Yamamoto and Jigami, 2002). It will be interesting to test whether members of other Rab-A subclasses, especially Rab-A2/A3, which represents the ancestral Rab-A in the plant lineage (Woollard and Moore, 2008), can rescue attrs130 at the gross and cellular levels. Recently, Cai et al. (2008) revealed that TRAPPII in yeast cells may also act as a GEF for Ypt1. In Arabidopsis, Rab-D proteins, homologs of Ypt1, have been shown to localize to a population of TGN/EE (Pinheiro et al., 2009), but the functional significance of this localization is not yet known. It will be interesting to examine if TRAPPII is also functionally linked to Rab-D proteins in post-Golgi trafficking in Arabidopsis.
Role of TRAPPII in plant cytokinesis
In plant cells, cytokinesis is accompanied by the formation of a cell plate in the center of the phragmoplast (Verma and Gu, 1996). The assembly of a cell plate generally consists of the following stages: (i) transport and fusion of Golgi-derived vesicles, (ii) formation of a tubular vesicular network, (iii) formation of an interwoven tubular network, and (iv) maturation of a new cell plate (Samuels et al., 1995; Segui-Simarro et al., 2004). However, little is known about the molecular regulation of these processes. KNOLLE, a syntaxin-related protein has been implicated in vesicle fusion in cytokinesis (Lauber et al., 1997). Fragmoplastin, a cell plate-localized dynamin-like protein (Gu and Verma, 1996), is proposed to be involved in the fusion or squeezing of the vesicles into the tubular structure at the cell plate (Verma and Gu, 1996). In attrs120 and attrs130 mutants, although the structure of the ER–Golgi interface is not altered, there is an abnormal accumulation of vesicles and tubular vesicular structures in the cytoplasm of the mutants. We suspect that in the mutants, the fusion of Golgi-derived vesicles necessary for the formation of the tubular vesicular network and/or an interwoven tubular network (Samuels et al., 1995; Segui-Simarro et al., 2004) is defective. Such a defect may arise from the improper activity of Rab-A proteins, which are proposed to regulate vesicle fusion in cytokinesis (Chow et al., 2008).
In the absence of TRAPPII, Arabidopsis plants can still survive early embryogenesis, but later embryogenesis is affected. Our interpretation of this is that TRAPPII has no detectable effect in cell division in early embryogenesis, but may have a pronounced effect on actively dividing cells in the later embryogenesis. In yeast, Ypt31/32 is important for cell viability (Benli et al., 1996). Rab11 in the model animal systems, is crucial for embryogenesis (Cao et al., 2008; Zhang et al., 2008). If the function of Rab-A proteins in plants is indispensable, it is possible that TRAPPII is not the sole GEF for plant Rab-A proteins. In this regard, we note that dominant-negative RAB-A2a (Chow et al., 2008) and RAB-A1c (X. Qi and H. Zheng, unpublished data), which presumably stabilize interactions with their GEFs, do not produce developmental defects to the same extent as described in this study for attrs120-2 or attrs130. Rab-A proteins may preferentially interact with and titrate different GEFs. A candidate protein that may serve as a GEF for Rab-A proteins is SCD1, a DENN domain protein that is involved in cytokinesis and polarized cell expansion (Falbel et al., 2003).
Role of TRAPPII in the polar targeting of PIN2
It has recently been reported that, to achieve a polar distribution, PIN proteins are first evenly distributed to the plasma membrane, and then undergo endocytosis, before being redirected to the proper domains of the plasma membrane in a polar distribution via endosomal recycling mechanisms (Dhonukshe et al., 2008). The subcellular detail of the transport and recycling of different PIN proteins, however, remains elusive. It is interesting to note that the polar targeting of PIN2, but not PIN1, was significantly affected in both attrs130 and attrs120-2. It seems that the TRAPPII-Rab-A mediated trafficking pathway acts selectively in the polar transport of PIN2. At the moment, we do not know how polar targeting of PIN2 is affected in the absence of TRAPPII. It is possible that there is delayed traffic of the newly synthesized PIN2 molecules to the plasma membrane. PIN2 is seen in the plasma membrane domains other than the apical plasma membrane in attrs120-2 or attrs130 mutants. Furthermore, the endocytic transport of FM4-64 appears normal in the absence of TRAPPII. Thus it is possible that polar recycling of PIN2 may not operate properly. Different from PIN1 (Jaillais et al., 2007), PIN2 is largely internalized and recycled through a GNL1-positive pathway to a population of endosomes that is sensitive to Endosidin1 (Teh and Moore, 2007; Robert et al., 2008). We show here that the distribution of AtTRS130 is sensitive to Endosidin1. Therefore, it is possible that TRAPPII-Rab-A1/A2/A3 defines a population of endosomes that are identical to the endosome defined by Endosidin1 (Robert et al., 2008).
In addition to the aberrant targeting of PIN2, we noted that the auxin distribution also altered in a wider region in root tips of either attrs130 or attrs120-2. In pin2, the change in DR5:GFP is relatively minor (Shin et al., 2005). It is therefore likely that in attrs120-2 or attrs130, the polar distribution of other auxin transporters is also affected. Indeed, the observed DR5:GFP pattern in attrs120-2 mutants is somewhat similar to what is seen in the pin2 pgp1/abcb1 double mutant (Blakeslee et al., 2007). It will be interesting to test whether the transport of ABCB1 is impaired in our mutants.
Plant materials and growth conditions
Salk T-DNA insertional mutant lines were obtained from the Arabidopsis Biological Resource Center (http://www.arabidopsis.org). attrs120-2, attrs120-4 and attrs130 expressing GFP-RAB-A1c, secGFP, GFP-HDEL, ST-YFP (Zheng et al., 2004), SecN-Rm-2A-GH (Samalova et al., 2006), PIN1-GFP, PIN2-GFP (Grieneisen et al., 2007) and DR5:GFP (ABRC stock CS9361) were made by crossing transgenic Arabidopsis lines (pollen donors) with heterozygous attrs120-2 and attrs130 and homozygous attrs120-4 (egg donors). Arabidopsis plants expressing GFP-RAB-A1c and YFP-RAB-A2a were made by crossing the transgenic YFP-RAB-A2a line (Chow et al., 2008) with transgenic Arabidopsis expressing GFP-RAB-A1c. To co-localize VHA-a1-mRFP with secGFP or PIN2-GFP in attrs120-2 and attrs130, the transgenic VHA-a1-mRFP line (Dettmer et al., 2006) was crossed with heterozygous attrs120-2 and attrs130 expressing secGFP or PIN2-GFP, respectively. For the dexamethasone-inducible expression of RAB-A1c(Q72L) in attrs120-2 and attrs130, heterozygous attrs120-2 and attrs130 (egg donor) was crossed with the LhGR driver line 4C-S5/7 (Craft et al., 2005) expressing RAB-A1c(Q72L). Seedlings were germinated and grown on hygromycin and kanamycin AT (Arabidopsis thaliana) plates (Haughn and Somerville, 1986) supplemented with dexamethasone solution (20 μm) diluted from a 100 mm DMSO stock. Plants on AT plates or on soil (Sunshine#5; SunGro, http://www.sungro.com) were grown at 22–24°C under continuous light (80–100 μE m−1 s−1 photosynthetically active radiation).
Molecular biology and generation of constructions
To examine the expression of AtTRS120 and AtTRS130, total RNAs were extracted from roots, stems, leaves and flowers of 3-week-old wild-type Columbia Arabidopsis plants and siliques of mature plants, as described by Zheng et al. (2004). To compare the expression of AtTRS120 and AtTRS130 in the wild type and mutants, total mRNAs were extracted from seedlings of the wild type and mutants. Reverse transcriptase-PCR was then performed using the Invitrogen SuperScript III System (Invitrogen, http://www.invitrogen.com).
To rescue attrs120-2 and attrs120-4, the cDNA of AtTRS120 was amplified using primers AtTRS120 forward (5′-CCGTCGACATGGAACCTGACGTC-3′) and AtTRS120 reverse (5′-CAGTGCACCTCCAGCTACACAG-3′). The amplified cDNA was first subcloned into pCR8/GW/TOPO according to the manufacturer’s instructions (Invitrogen) and sequenced. The sequenced clone was then subcloned into the pEarleyGate 101 vector (ABRC stock CD3-683) to make AtTRS120-YFP. To rescue attrs130, the cDNA of AtTRS130 was amplified using primers AtTRS130 forward (5′-CCGGATCCATGGCGAACTACTTG-3′) and AtTRS130 reverse (5′-CTTGACAGGTAAGCAGTAGGAAG-3′), and cloned and sequenced in pCR8/GW/TOPO. The sequenced clone was then subcloned into pEarleyGate101 to generate AtTRS130-YFP. The cDNA of RAB-A1c was amplified using primers RAB-A1c forward (5′-GGGAATTCGTCGACATGGCGGGTTACAGAGC-3′) and RAB-A1c reverse (5′-GCGGATCCGAGCTCTTAGTTCGAGCAGCATCC-3′). The amplified cDNA was first cloned into pBluescript KS as the EcoRI-BamHI fragment and then sequenced. The Q72L substitution was effected into RAB-A1c by overlapping PCR using primers QL forward (5′-GGGATACTGCTGGTCTAGAAAGGTACCGAGCC-3′) and QL reverse (5′-GGCTCGGTACCTTTCTAGACCAGCAGTATCCC-3′). The sequenced wild-type RAB-A1c was subcloned into pVKH-GFPN (Zheng et al., 2005b) as the SalI-SacI fragment to generate GFP-RAB-A1c, which was transformed into wild-type Arabidopsis ecotype Col-0 and crossed with attrs130. The modified RAB-A1c(Q72L) was subcloned into the pV-TOP vector (Craft et al., 2005) and transformed into the LhGR driver line 4C-S5/7 (Craft et al., 2005), and crossed with attrs130.
Phenotyping and light microscopy
For embryo development, developing embryos of heterozygous attrs120 and attrs130 mutants were removed and cleared in Hoyer’s solution, as described by Liu and Meinke (1998). They were analyzed with a Leica DMI6000B microscope (Leica, http://www.leica.com), and images were recorded with a QImaging Retiga EXi digital color camera (QImaging, http://www.qimaging.com). Images of seeds and seedlings were taken using a QImaging Micropublisher3.3 digital CCD color camera installed on a Leica MZ16F stereomicroscope. Adult plants were imaged with a Nikon D80 digital camera (Nikon, http://www.nikon.com).
Seedlings (7-days old) of attrs130 expressing AtTRS130-YFP were incubated in 1 ml AT liquid medium containing 20 μm Wortmannin (prepared from 1 mm stock solution dissolved in DMSO) or 33 μm Endosidin1 (prepared from 1.67 mm stock solution dissolved in DMSO) at room temperature (23–25°C) for 2 h. Control treatments were performed with equal volumes of DMSO.
Propidium iodide (PI) and FM4-64 staining, fluorescence microscopy and confocal microscopy
To visualize cell morphology, embryos and seedlings were mounted in 1 mg ml−1 propidium iodide (P3566; Invitrogen) on slides for 1 min. To study the localization and dynamics of FM4-64 as well as to visualize early endosomes and cell plates in wild-type and mutant cells, seedlings were stained with 5 μm FM4-64 (T13320, diluted from a 5 mm stock solution in water; Invitrogen) on microscropy slides for 5, 10, 20, 40, 60, 80, 100, 120 and 140 min. At each given time point, between four and six seedlings were stained. Each experiment was repeated at least three times. Visualization of FM4-64 was performed in root epidermal cells. For fluorescence microscopy of GFP, YFP, PI and FM4-64, seedlings were analyzed with a Leica DMI6000B microscope. Images were recorded with a QImaging Retiga EXi digital color camera. Confocal microscopy was carried out with an inverted Zeiss LSM 510 Meta confocal laser scanning microscope (Zeiss, http://www.zeiss.com). Single-color images of GFP, PI and FM4-64, and multicolor images of GFP/YFP, GFP/PI, GFP/FM4-64 and YFP/FM4-64 were acquired as described by Zheng et al. (2005a). The Zeiss LSM image browser, volocity (PerkinElmer, http://www.perkinelmer.com) and photoshopcs2 (Adobe, http://www.adobe.com) were used for post-acquisition image processing.
High-pressure freezing (HPF) and freeze substitution
Cotyledons and root tips of seedlings were high-pressure frozen in 1-hexadecene using an EM PACT2 (Leica Microsystems). Frozen samples were transferred to a frozen freeze-substitution medium containing 2% osumium tetroxide and 8% 2,2-dimethoxypropane in glass-distilled acetone in cryovials. The vials with frozen samples were transferred to a pre-cooled (–95°C) FreasySub temperature-control chamber (Cryotech, http://www.cryotech.com). The freeze-substitution starts at −90°C. The samples were kept at −90°C for 48 h, and then warmed to −60°C (8 h), then to −20°C (8 h) and finally to 4°C (2 h). The rate of temperature change was 5°C per h. The samples were left at room temperature for 2 h, rinsed in pure acetone and then processed for embedding in Spurr resin. The resin was gradually added over 4 days. Polymerization was performed at 65°C for 12 h.
For light microscopy, 300-nm sections were mounted on glass slides and stained with toluidine blue. Observation was performed with a Zeiss light microscope AXIO. Images were captured using an AxioCam MRm digital black/white camera (Zeiss). For electron microscopy, sections of between 50- and 70-nm thick were mounted on formvar-coated grids, stained with uranyl acetate for 30 min and lead citrate for 15 min. Observation, quantification and imaging was performed with an FEI Tecnai electron microscope (FEI, http://www.fei.com) operated at 120 kV.
Sequence data from this article can be found in the Arabidopsis Genome Initiative database under the following accession numbers: AtTRS120 (At5g11040); AtTRS130 (At5g54440) and RAB-A1c (At5g45750).
We thank Zhengyun Xu (McGill University, Montreal, Canada) for the identification of attrs130 and attrs120 alleles; Ian Moore (University of Oxford) for the LhGR/pOP6 system, and the YFP-RAB-A2a and SecN-Rm-2A-GH lines; Jian Xu and Ben Scheres (Universiteit Utrecht, Utrecht, the Netherlands) for providing seeds of ProPIN1:PIN1-GFP and ProPIN2:PIN2-GFP; Karin Schumacher (University of Heidelberg) for the VHA-a1-mRFP line; Glenn Hicks (University of California, Riverside) for Endosidin1; the SALK Institute (http://www.salk.edu) for T-DNA insertion lines. We also thank Tamara Western (McGill University, Montreal, Canada) for her critical reading of the manuscript. This work was supported by grants from the Natural Sciences and Engineering Research Council of Canada to HZ and AG. XQ was supported by a Quebec Merit award (2009–2011) for international students, and MK by a grant from the Government of Canada for postdoctoral studies.