Overexpression of Arabidopsis Sorting Nexin AtSNX2b Inhibits Endocytic Trafficking to the Vacuole

Molecular Plant, Nov 2008

Sorting nexins are conserved proteins that function in vesicular trafficking and contain a characteristic phox homology (PX) domain. Here, we characterize the ubiquitously expressed Arabidopsis thaliana sorting nexin AtSNX2b. Sub-cellular fractionation studies indicate that AtSNX2b is peripherally associated with membranes. The AtSNX2b PX domain binds to phosphatidylinositol 3-phosphate in vitro and this association is required for the localization of GFP–AtSNX2b to punctate structures in vivo, identified as the trans-Golgi network, prevacuolar compartment and endosomes. Overexpression of GFP-tagged AtSNX2b produces enlarged GFP-labeled compartments that can also be labeled by the endocytic tracer FM4-64. Endocytic trafficking of FM4-64 to the vacuole is arrested in these GFP–AtSNX2b compartments, and similar FM4-64-accumulating compartments are seen upon overexpression of untagged AtSNX2b. This suggests that exit of membrane components from these enlarged or aggregated endosomes is inhibited. Vacuolar proteins containing an N-terminal propeptide, but not those with a C-terminal propeptide, are also present in these enlarged compartments. We hypothesize that AtSNX2b is involved in vesicular trafficking from endosomes to the vacuole.

A PDF file should load here. If you do not see its contents the file may be temporarily unavailable at the journal website or you do not have a PDF plug-in installed and enabled in your browser.

Alternatively, you can download the file locally and open with any standalone PDF reader:


Overexpression of Arabidopsis Sorting Nexin AtSNX2b Inhibits Endocytic Trafficking to the Vacuole

Nguyen Q. Phan 0 1 Sang-Jin Kim 0 1 Diane C. Bassham bassham@iastate 0 1 0 edu , fax 515-294-1337, tel. 515-294-7461. a The Author 2008. Published by the Molecular Plant Shanghai Editorial Office in association with Oxford University Press on behalf of CSPP and IPPE , SIBS, CAS. doi: 10.1093/mp/ssn057, Advance Access publication 3 October 2008 Received 12 June 2008; accepted 18 August 2008 1 a Department of Genetics, Development and Cell Biology, Iowa State University , Ames, IA 50011 , USA b Interdepartmental Genetics Program, Iowa State University , Ames, IA 50011 , USA c Plant Sciences Institute, Iowa State University , Ames, IA 50011 , USA Sorting nexins are conserved proteins that function in vesicular trafficking and contain a characteristic phox homology (PX) domain. Here, we characterize the ubiquitously expressed Arabidopsis thaliana sorting nexin AtSNX2b. Sub-cellular fractionation studies indicate that AtSNX2b is peripherally associated with membranes. The AtSNX2b PX domain binds to phosphatidylinositol 3-phosphate in vitro and this association is required for the localization of GFPAtSNX2b to punctate structures in vivo, identified as the trans-Golgi network, prevacuolar compartment and endosomes. Overexpression of GFP-tagged AtSNX2b produces enlarged GFP-labeled compartments that can also be labeled by the endocytic tracer FM4-64. Endocytic trafficking of FM4-64 to the vacuole is arrested in these GFP-AtSNX2b compartments, and similar FM4-64-accumulating compartments are seen upon overexpression of untagged AtSNX2b. This suggests that exit of membrane components from these enlarged or aggregated endosomes is inhibited. Vacuolar proteins containing an N-terminal propeptide, but not those with a C-terminal propeptide, are also present in these enlarged compartments. We hypothesize that AtSNX2b is involved in vesicular trafficking from endosomes to the vacuole. - INTRODUCTION The structure and function of cellular organelles are maintained by a network of pathways for protein synthesis, sorting, recycling, regulation, and degradation. In the endomembrane system, new proteins are synthesized on membrane-bound ribosomes at the endoplasmic reticulum (ER). These proteins are folded and glycosylated before being transported to the Golgi, where they can be further modified as they move from the cis to the trans cisternae. At the trans face of the Golgi resides the trans-Golgi network (TGN), where critical sorting events take place to target proteins either to the cell surface or to the vacuole. Transport to the plasma membrane may occur via a default pathway (Batoko et al., 2000; Mitsuhashi et al., 2000), while vacuolar proteins bind to specific receptors for sorting to their final destinations (Kirsch et al., 1994; Ahmed et al., 2000; Shimada et al., 2003; Park et al., 2005). However, some secreted material, such as cell wall components, is deposited in a polarized manner (Freshour et al., 2003; Lee et al., 2005), and some proteins show a polarized distribution in the plasma membrane (Dhonukshe et al., 2005), suggesting that the default secretion hypothesis is an oversimplification. In plants, the vacuole is a multifunctional organelle that is essential for plant survival. Vacuolar functions include storage, degradation, sequestration of xenobiotics, maintenance of turgor and protoplasmic homeostasis (Marty, 1999). Whereas some cell types appear to contain a single central vacuole that performs all of these functions, other cell types may contain multiple functionally distinct vacuoles that can be distinguished by their unique protein composition (Paris et al., 1996; Di Sansebastiano et al., 2001; Olbrich et al., 2007). Delivery of proteins to the vacuole in plants involves two general pathways: (1) the biosynthetic route via ER and Golgi compartments, and (2) the endocytic route via endocytosis from the cell surface (Chrispeels and Raikhel, 1992; Marty, 1999; Matsuoka and Neuhaus, 1999; Hanton and Brandizzi, 2006). Three types of signal have been identified for targeting proteins to the plant vacuole via the biosynthetic routea C-terminal propeptide (CTPP), a sequence-specific vacuolar sorting signal, usually in the form of an N-terminal propeptide (NTPP), and an internal signal (Marty, 1999). These signals are recognized at the TGN by vacuolar sorting receptors for correct targeting to the lytic or storage vacuole (Ahmed et al., 2000; Paris and Neuhaus, 2002; Shimada et al., 2003; Jolliffe et al., 2005). Distinct pathways, and possibly different receptors, exist for transport to different vacuole types, and some storage proteins are even transported directly from the ER to the vacuole, bypassing the Golgi apparatus entirely (Hillmer et al., 2001; Shimada et al., 2003; Park et al., 2005). Endocytic trafficking to the plant vacuole begins at the plasma membrane, where membrane and extracellular cargo are internalized either by clathrin-coated vesicles or via a clathrin-independent pathway (Grebe et al., 2003; Baluska et al., 2004; Holstein and Oliviusson, 2005). These vesicles probably fuse with endosomes, organelles that function in sorting, recycling and further transport of cargo and membrane (Samaj et al., 2005; Muller et al., 2007). They can be classified into several distinct types based both on their protein content and by functional criteria (Ueda et al., 2004; Yamada et al., 2005; Dettmer et al., 2006; Haas et al., 2007; Lam et al., 2007; Jaillais et al., 2008). The biosynthetic and endocytic transport pathways probably merge at late endosomes, which may be equivalent to the prevacuolar compartment (PVC), from which point both pathways utilize similar trafficking components to aid routing of cargos to the vacuole (Samaj et al., 2005; Yamada et al., 2005; Muller et al., 2007; Jaillais et al., 2008). Increasing evidence indicates a role for the sorting nexin family of lipid-binding proteins in protein trafficking and sorting. Sorting nexins (SNXs) are membrane-associated proteins that are involved in endocytosis and protein trafficking through phosphatidylinositol lipid-containing organelles and are conserved in eukaryotes (Worby and Dixon, 2002; Carlton et al., 2005). SNXs are part of a large family of proteins that are defined by the presence of a phox homology domain (Teasdale et al., 2001), a 100130 amino acid domain that binds phosphatidylinositol (PI) phosphates and targets the proteins to specific cellular membranes enriched in that phospholipid. In mammals, SNX functions include recycling or degradation of receptors (Kurten et al., 1996), control of endosome morphology (Barr et al., 2000), and regulation of endosomal function (Xu et al., 2001; Rojas et al., 2007). Several lines of evidence have implicated SNXs in an endosome-to-Golgi retrograde trafficking pathway. Genetic screens in yeast have identified a pentameric retromer complex required for the transport of proteins, including the sorting receptor Vps10p, from the PVC to the Golgi complex (Seaman et al., 1997, 1998). Vps5p and Vps17p are SNXs that dimerize (Horazdovsky et al., 1997) to mechanically aid in vesicle budding, while Vps35p and Vps29p act as cargo selectors (Seaman et al., 1998; Reddy and Seaman, 2001). Lastly, Vps26p binds the complex together (Reddy and Seaman, 2001). Recent work from several laboratories has provided evidence of retromers in other model organisms and identified homologs of the yeast genes VPS29, VPS30, and VPS35 in Arabidopsis and mammals (Haft et al., 2000; Oliviusson et al., 2006; Shimada et al., 2006; Jaillais et al., 2007; Yamazaki et al., 2008). These homologs show similarity to the yeast counterparts and assemble into a macromolecular complex that is presumed to similarly function in endosometo-Golgi transport/recycling. Among the better characterized SNXs, mammalian SNX1 and 2, orthologs of Vps5p, have been shown to play a role in the mammalian retromer complex (Haft et al., 2000; Rojas et al., 2007). Other mammalian SNXs are involved in endocytic trafficking and endocytosis of plasma membrane receptors (Parks et al., 2001; Xu et al., 2001; Leprince et al., 2003; Lundmark and Carlsson, 2003; Merino-Trigo et al., 2004). Initial evidence for a function of SNXs in plants came from a yeast two-hybrid assay demonstrating that the Brassica oleracea S locus receptor kinase intracellular domain interacts with a SNX (Brassica oleracea SNX1, BoSNX1) during the self-incompatibility response in pollen recognition (Vanoosthuyse et al., 2003). The closest BoSNX1 homolog in Arabidopsis is the gene At5g06140 (named AtSNX1). AtSNX1 is involved in trafficking of the auxin transport component PIN2 through an AtSNX1containing PVC via a novel transport pathway (Jaillais et al., 2006, 2008) and is probably a component of the Arabidopsis retromer complex (Jaillais et al., 2007). Two additional SNX genes are present in the Arabidopsis genome, designated AtSNX2a (At5g58440) and AtSNX2b (At5g07120) (Vanoosthuyse et al., 2003). We demonstrate here that the AtSNX2b protein can bind to phosphatidylinositol 3-phosphate (PI3P) and that this interaction is required for its localization to endosomes. Overexpression of AtSNX2b leads to enlargement or aggregation of the AtSNX2b-containing endosomes and interferes with efficient transport to the vacuole. These data implicate AtSNX2b in trafficking of cargo and membrane through an endosomal compartment. Three putative SNXs have been identified in the model plant Arabidopsis (Vanoosthuyse et al., 2003). Here, we focus on AtSNX2b, a potential SNX based on amino acid similarity (4050%) with yeast and human SNXs, with the majority of the similarity residing in the region of the PX domain. Of the Arabidopsis sorting nexins, AtSNX2b is most similar to AtSNX2a (86% amino acid similarity) and shows much lower similarity to AtSNX1 (47% amino acid similarity). AtSNX2b has two major domains: (1) an N-terminal conserved PX domain which defines a SNX (Worby and Dixon, 2002; Carlton et al., 2004, 2005), and (2) a C-terminal coiled-coil region potentially important for proteinprotein interactions (Zhong et al., 2002; Leprince et al., 2003; Merino-Trigo et al., 2004; Carlton et al., 2005; Gallop and McMahon, 2005; Figure 1A). AtSNX2b Is a Ubiquitously Expressed Membrane Associated Protein To determine the expression pattern of AtSNX2b, RNA was extracted from different Arabidopsis plant organs and RT PCR was performed using AtSNX2b gene-specific primers. Figure 1B shows that AtSNX2b expression can be detected in all of the plant organs tested, including roots, rosette leaves, cauline leaves, stem, flowers, and siliques, suggesting that its function Figure 1. The Arabidopsis SNX2 family. (A) Structural features of AtSNX2b. The AtSNX2b protein contains a PX domain near the N-terminus and a C-terminal coiled-coil region. (B) Expression of AtSNX2b mRNA throughout the Arabidopsis plant. RTPCR analysis using AtSNX2b-specific primers shows AtSNX2b is ubiquitously expressed in roots (R), rosette leaves (RL), cauline leaves (CL), inflorescence stem (St), flowers (F), siliques (Si), and senescing leaves (SL). 18S RNA is present as a loading control. (C) Expression of AtSNX2a mRNA throughout the Arabidopsis plant. RTPCR analysis was performed as in (B), except using primers specific to AtSNX2a. 18S RNA is used as a loading control. is important throughout the plant. For comparison, the expression pattern of AtSNX2a was also determined (Figure 1C). AtSNX2a mRNA was also detected throughout the plant, although at lower levels in flowers, siliques, and senescing leaves. The overlapping expression patterns of AtSNX2b and AtSNX2a raises the possibility that the two genes may also have overlapping functions. To analyze the AtSNX2b protein, antibodies were raised against recombinant AtSNX2b. Full-length AtSNX2b protein was synthesized in Escherichia coli as a His-tagged fusion protein and purified by affinity chromatography. The protein was injected into rabbits and the generated antibodies were affinity-purified against the recombinant protein prior to use. Immunoreactivity of the affinity-purified antibodies (Figure 2A) was compared with that of the crude serum (Figure 2C) and preimmune serum (Figure 2B) by immunoblotting against the recombinant protein antigen and a protein preparation from Arabidopsis. The purified anti-AtSNX2b antibodies recognized the recombinant protein and a band of similar molecular weight (approximately 67 kDa) in a total protein preparation from Arabidopsis (Figure 2A), which was not recognized by the pre-immune serum (Figure 2B). Cross-reacting bands were present in the crude serum blots (Figure 2C) that were mostly absent after affinity purification. It was observed that in addition to a prominent band of the expected molecular mass, a second, weak band of slightly lower mobility on SDSPAGE (approximately 70 kDa) was sometimes recognized by the AtSNX2b antibodies. We hypothesized that this band may either correspond to a modified form of AtSNX2b, or, because of the sequence similarity between AtSNX2a and AtSNX2b, may correspond to AtSNX2a. To investigate this further, an Arabidopsis knockout mutant was isolated from the GABI-Kat flanking sequence tag database (Rosso et al., 2003) in which the AtSNX2b gene was disrupted by a T-DNA insertion (Figure 2D). Loss of gene expression was confirmed by RTPCR using gene-specific primers. No phenotype has yet been observed for the Atsnx2b mutant, either at a morphological level after examination throughout its lifecycle, in protein trafficking pathways to the vacuole, or in hormone-related responses such as gravitropism (data not shown). As the AtSNX2a protein is closely related in sequence and AtSNX2a and AtSNX2b have overlapping expression patterns (Figure 1B and 1C), we hypothesize that the two genes may perform redundant functions. Comparison of protein extracts from the Atsnx2b mutant with those from wild-type plants by immunoblotting demonstrated that the major, lower band of the doublet recognized by the AtSNX2b antibodies was absent in the mutant, confirming that it corresponds to AtSNX2b itself. The higher, much weaker band was still present in the mutant, suggesting that this is most likely AtSNX2a, rather than a modified version of AtSNX2b. The distribution of AtSNX2b protein in aerial organs of Arabidopsis mature plants was analyzed (Figure 2E) and compared with the RTPCR analysis of mRNA level (Figure 1B). High protein levels were seen in flowers, inflorescence stems, and cauline leaves, and lower but detectable levels in siliques and rosette leaves (Figure 2E), as well as a significant amount in roots and young seedlings (see also Figure 6). These results are consistent with the RTPCR analysis, although much greater variation is seen in protein level than mRNA level, possibly suggesting that post-translational regulation may occur. Because SNXs are typically associated with membranes, a total protein extract was separated into membrane and soluble fractions by centrifugation at 125 000 g and analyzed by immunoblotting using AtSNX2b antibodies. AtSNX2b was detected in both the pellet and soluble fractions, indicating that AtSNX2b is partially membrane-associated (Figure 2A). To analyze further this membrane association, differential centrifugation was performed at 12 000, 39 000 and 125 000 g and fractions were probed for the presence of AtSNX2b. AtSNX2b was detected in all three membrane fractions and the soluble fraction, confirming that AtSNX2b is partially membrane-associated (Figure 2F). The upper weak band, potentially AtSNX2a, also appeared to be membrane-associated, although the weak and variable cross-reactivity made it difficult to draw definitive conclusions about this protein. To determine how AtSNX2b protein associates with the membrane, total membrane pellets were re-suspended in either extraction buffer alone, or extraction buffer containing 2 M NaCl, 0.1 M Na2CO3, 2 M urea or 1% (v/v) triton X-100. After incubation for 2 h, membranes were repelleted and pellet and supernatant fractions analyzed by immunoblotting with AtSNX2b antibodies (Figure 2G). Each of the treatments was able to extract AtSNX2b from the membrane, indicating that AtSNX2b is peripherally associated with membranes. As a control, SYP41, an integral membrane protein (Bassham et al., 2000), was only extracted from the membrane by the detergent triton X-100 (Figure 2G). AtSNX2b Can Bind PI3P SNXs are defined by the presence of a PX domain and its ability to bind phosphoinositol lipids (Worby and Dixon, 2002).To determine if the AtSNX2b PX domain can associate with PI lipids, or phospholipids in general, the PX domain fragment of AtSNX2b was fused with GST (glutathione-S-transferase) to generate GSTPX. As a control, point mutations were introduced into the PX domain to create an amino acid 233RR/LG change that has been shown previously to prevent PI binding in human SNXs (Zhong et al., 2002). GST fusions with the wildtype or mutant PX domains were synthesized in E.coli, purified over a glutathione resin (Figure 3A) and allowed to bind to PIP strips (Echelon Inc.) containing various phospholipids. Binding was detected using antibodies against GST. The PX domain of AtSNX2b (GSTPX) specifically bound to PI3P while the PX domain mutant (GSTPX-1) and GST alone were not able to bind to any lipid (Figure 3B). Localization of AtSNX2b To gain insight into AtSNX2b function, its sub-cellular localization was examined in vivo. A GFP fusion was generated with full-length AtSNX2b (GFPAtSNX2b) and transiently expressed in Arabidopsis protoplasts derived from suspension cultured cells. In addition, immunofluorescence using anti-AtSNX2b antibodies was used to assess AtSNX2b sub-cellular localization in Arabidopsis protoplasts. Both immunofluorescence and GFP-fusion localization show that AtSNX2b localizes to punctate spots in the cytoplasm (Figure 4A and 4B). To verify that a full-length GFPAtSNX2b fusion is produced and correctly associates with membranes, membrane and soluble fractions were prepared from protoplasts expressing GFP AtSNX2b, or GFP as a control. The proteins were expressed for 20 h to allow the proteins to accumulate to high enough levels for detection by immunoblotting using GFP antibodies. A GFPAtSNX2b fusion of the expected size was detected predominantly in the membrane fraction (Figure 4D). To analyze the role of the PX domain in localization of AtSNX2b to these structures, a fusion between the full-length AtSNX2b protein containing the PX domain mutation described above and GFP was generated (GFPAtSNX2b-1). In contrast to the wild-type protein, the GFPAtSNX2b-1 mutant did not localize to discrete structures but rather showed a diffuse fluorescence pattern throughout the cell (Figure 4C), suggesting that the PX domain is required for correct localization in vivo. To determine whether the PX domain alone is sufficient for localization to punctate spots, GFP was fused with the PX domain or mutant PX domain alone and the localization analyzed by fluorescence microscopy. GFPPX and GFP PX-1 fusions showed diffuse GFP patterns similar to the GFPAtSNX2b-1 mutant (Figure 4E). To confirm this result, an additional PX domain mutant was generated (PX-2; 211PP/AA). Unfortunately, the full-length GFPAtSNX2b-2 fusion protein was not expressed in protoplasts, based on GFP fluorescence. The fusion of GFP with the PX-2 mutant PX domain also showed diffuse cytoplasmic localization, as for the wild-type and PX-1 fusions. Our results suggest that the PX domain is necessary but not sufficient for the localization of AtSNX2b to punctate compartments. As an initial approach to determine the identity of the AtSNX2b-labeled structures, the distribution of AtSNX2b and several known intracellular markers in a sucrose density gradient (1355%) was examined. Fractions from the gradient were analyzed by immunoblotting using antibodies against aleurain (ALEU; vacuolar soluble protein), cTIP (vacuolar membrane protein), fumarase (FUM1; mitochondria), SYP21 (PVC tSNARE), VTI12 (TGN-localized v-SNARE), and AtSNX2b. Both the AtSNX2b band (67 kDa) and the cross-reacting 70-kDa band were visible in these fractions. Figure 5 shows that AtSNX2b has a bipartite distribution with a small part soluble (fractions 25) and the majority membrane-bound (fractions 712), as predicted from the differential centrifugation (Figure 2F). The 70-kDa band was present in the same fractions as AtSNX2b, but showed a greater percentage present in the soluble fractions compared with membrane-bound. The distribution of the membrane-associated portion of AtSNX2b on the sucrose gradient overlapped extensively with that of VTI12 and with the upper band of SYP21. The SYP21 antibodies recognize a triplet of proteins, all of which correspond to the SYP21 protein (Conceicao et al., 1997); the nature of these three isoforms is unclear but may be due to post-translational modifications. These results suggest a possible localization to an organelle with similar density to the TGN and/or PVC, and distinct from the vacuole and mitochondria. To determine more precisely the localization of AtSNX2b, roots of transgenic Arabidopsis lines expressing markers for the TGN (T7-SYP42; Bassham et al., 2000), PVC (T7-SYP21 and T7-SYP22; Sanderfoot et al., 1999), late endosomes (YFP-Rha1; Preuss et al., 2004), or Golgi apparatus (sialyl transferase (ST)-GFP; Wee et al., 1998) were analyzed by double immunofluorescence-labeling using AtSNX2b and T7 antibodies or comparison with YFP or GFP fluorescence as appropriate. As in protoplasts, AtSNX2b antibodies recognized punctate structures in root cells that were not labeled with preimmune serum (Figure 6A). To assess the specificity of AtSNX2b labeling, immunofluorescence was also performed under identical conditions on roots from the Atsnx2b knockout mutant. No specific signal was seen, indicating that the immunofluorescence staining observed corresponds only to AtSNX2b. White spots in merged images show that some of the AtSNX2b-labeled organelles also contained T7-SYP42, T7-SYP21 (data not shown), T7-SYP22, or YFP-Rha1 (Figure 6B). By contrast, no overlap was seen between ST-GFP and AtSNX2b, suggesting that AtSNX2b does not reside in the Golgi apparatus. Percent co-localization was determined by counting the number of AtSNX2b-labeled structures that co-localized with each of the markers. Partial co-localization of AtSNX2b with Figure 4. Transient Expression of GFP-Tagged AtSNX2b and PX Domains. (A) Immunofluorescence labeling of endogenous AtSNX2b in Arabidopsis protoplasts. (B) Expression of GFP-tagged AtSNX2b in Arabidopsis protoplasts. Punctate GFP structures are seen, similar to the endogenous protein labeling. (C) Expression of the GFPAtSNX2b-1 PX domain mutant in Arabidopsis protoplasts shows diffuse cytoplasmic GFP labeling. (D) Immunoblot of GFPAtSNX2b expression in protoplasts. Extracts from protoplasts transiently expressing GFPAtSNX2b were fractionated by centrifugation at 125 000 g yielding soluble (S) and membrane pellet (P) fractions. GFP-fused AtSNX2b is membrane-associated while GFP is mostly soluble. Molecular mass markers are shown at left (kDa). (E) Fluorescence images of GFP-tagged PX domain (GFPPX) and GFP-tagged PX domain mutants (GFPPX-1 and GFPPX-2) showing diffuse cytoplasmic GFP labeling. Insets show brightfield images. Scale bars for all figures are 10 lm. T7-SYP42, T7-SYP21, T7-SYP22, and YFP-Rha1 was observed (Figure 6B and 6C). The PVC and late endosomal markers T7-SYP21, T7-SYP22, and YFP-Rha1 show extensive overlap in their localization (Lee et al., 2004), leading to the conclusion that a large portion of AtSNX2b (6080%) does not co-localize with any of the markers tested. As our antibodies are specific for AtSNX2b under these experimental conditions (Figure 6A), this localization pattern suggests that AtSNX2b might localize to or cycle between PVC/late endosomal compartments, the TGN, which has been suggested also to be an early endosome (Dettmer et al., 2006; Lam et al., 2007), and possibly an additional unidentified compartment. Figure 5. Sucrose Gradient Profiles of AtSNX2b and Various SubCellular Markers. An Arabidopsis protein extract was fractionated on a 1355% sucrose density gradient and analyzed by immunoblotting with antibodies against aleurain (ALEU), fumarase (FUM1), c-tonoplast intrinsic protein (cTIP), the t-SNARE SYP21, the v-SNARE VTI12 and AtSNX2b. Overexpression of AtSNX2b Affects Trafficking It is known that Arabidopsis cells contain multiple endosome types (Ueda et al., 2004), and we therefore hypothesized that the unidentified compartment with which AtSNX2b associates could be an additional type of endosome. To test this hypothesis, we analyzed the localization of GFPAtSNX2b in transiently transformed protoplasts compared with the fluorescent marker FM4-64. FM4-64 is widely used as an endocytic tracer in live cells (Vida and Emr, 1995; Betz et al., 1996; Bolte et al., 2004). The cell cultures used contain cells of varying sizes; no differences were seen between large and small cells in any experiment. FM4-64 binds to the plasma membrane, is internalized by endocytosis, traffics through the endosomal system and reaches the vacuolar membrane after 34 h in Arabidopsis (Bolte et al., 2004). Arabidopsis protoplasts were transiently transformed with GFPAtSNX2b and incubated for 12 h to allow expression of the protein. They were then labeled with FM4-64 and uptake of the dye was observed over a time course of up to 4 h (Figure 7). After 0.5 h of uptake, FM4-64 staining was seen mainly at the plasma membrane, with a few puncta in the cytoplasm; no co-localization of FM4-64 with GFPAtSNX2b was seen at this early time point. Between 1 and 3 h of uptake, FM4-64 was found in punctate structures that have been shown previously to correspond to Golgi/TGN, PVC, and endosomes (Betz et al., 1996; Bolte et al., 2004; Dettmer et al., 2006). Almost complete co-localization of FM4-64 and GFPAtSNX2b was seen at these times, suggesting that AtSNX2b is predominantly localized to organelles on the endocytic pathway. By the 4-h time point, FM4-64 reached the vacuolar membrane in control protoplasts transformed with GFP alone, while, in GFPAtSNX2b transformed protoplasts, FM4-64 was not present on the vacuolar membrane, instead being trapped in cytoplasmic structures containing GFPAtSNX2b. FM4-64 was not able to exit the GFP-labeled compartments even after 12 h. In addition, the appearance of the GFPAtSNX2b structures varied over time. At early time points, the GFPAtSNX2b localized to structures similar in appearance, although apparently somewhat larger than the structures in which endogenous AtSNX2b resides; the size of these organelles is difficult to assess by fluorescence microscopy. At later time points, these structures became enlarged and/or aggregated, and, by 4 h of FM4-64 labeling, the protoplasts contained just a few large GFP AtSNX2b-labeled structures (Figure 7). This is not related to the presence of FM4-64, as similar effects are seen in the absence of FM4-64 staining, with a gradual increase in the size of the labeled structures over time. The GFPAtSNX2b in the enlarged structures is most likely membrane-associated, as it pellets with a membrane fraction after lysis of protoplasts (see Figure 4D) and it co-localizes with FM4-64, which is a membrane-bound dye. These results suggest that overexpression of AtSNX2b causes inhibition of FM4-64 trafficking to the vacuole, possibly by blocking the exit of material from endosomes. To confirm that the inhibition of transport of FM4-64 to the vacuole was caused by an effect on trafficking, rather than a loss of cell viability, protoplasts were transiently transformed with the GFPAtSNX2b construct and expression allowed to proceed for up to 24 h. Fluorescence microscopy confirmed the formation of enlarged structures as above, and protoplasts were stained with the vital stain fluorescein diacetate to assay for cell viability. No difference was seen between untransformed and transformed protoplasts, and, in both cases, almost all protoplasts survived, indicating very little loss of viability (Figure 8). To verify that overexpression of AtSNX2b alone affects FM464 trafficking to the vacuolar membrane, rather than the presence of the GFP tag, an untagged AtSNX2b overexpression construct was introduced into protoplasts, followed by FM464 labeling as above. Similar to the effect of GFPAtSNX2b expression, enlarged FM4-64 structures were observed in the cytoplasm, and most of the FM4-64 failed to reach the vacuolar membrane (Figure 7), even at later time points. This confirms that overexpression of AtSNX2b affects trafficking along the endocytic pathway to the vacuole. Biosynthetic protein trafficking to the plant vacuole occurs through at least two major pathways, and markers are available for each pathway consisting of GFP fused to an N-terminal vacuolar sorting signal (NTPPGFP; Ahmed et al., 2000) or a C-terminal vacuolar sorting signal (GFPCTPP; Fluckiger et al., 2003; Sanmartin et al., 2007). To determine whether the overexpression of AtSNX2b inhibits either of these biosynthetic pathways in addition to endocytic trafficking, AtSNX2b was co-expressed in protoplasts with either NTPPGFP or GFPCTPP. At 15 h after transformation, punctate motile GFP spots were seen (Figure 9A, arrows) along with a clear vacuolar NTPPGFP or GFPCTPP signal. By 30 h after transformation, motile GFP spots diminished and vacuolar labeling of GFP became very dominant (Figure 9A). In protoplasts transformed with NTPPGFP or GFPCTPP alone, only vacuolar GFP signal was observed after 30 h. In double transformants expressing NTPPGFP plus 35S::AtSNX2b, non-motile GFP spots accumulated in addition to vacuolar GFP labeling after 30 h (Figure 9A, arrowheads); in contrast, double transformed GFPCTPP protoplasts did not accumulate GFP spots and only vacuolar GFP labeling was evident at the 30-h time point. These results indicate that overexpression of AtSNX2b causes accumulation of NTPPGFP in punctate structures in addition to the vacuole, whereas no effect is seen on GFPCTPP trafficking. To determine whether the structures accumulating NTPP GFP upon overexpression of AtSNX2b are the same structures in which FM4-64 accumulates, protoplasts overexpressing AtSNX2b and NTPPGFP were labeled with FM4-64. At 3 h after FM4-64 uptake, FM4-64 labeling co-localized with non-vacuolar NTPPGFP signal; these co-localized structures persisted up to 12 h after FM4-64 uptake (Figure 9B). This demonstrates that the NTPPGFP-containing bodies that are the result of overexpression of AtSNX2b are endosomes. While most of the NTPPGFP still reaches the vacuole upon AtSNX2b overexpression, these results suggest that overexpression of AtSNX2b partially interferes with the normal trafficking of NTPPGFP (Figure 9B). Endogenous AtSNX2b co-localizes with TGN and endosomal markers (Figures 6 and 7). To determine whether these markers are also present in the enlarged structures produced upon AtSNX2b overexpression, protoplasts overexpressing GFP AtSNX2b were probed with antibodies against the TGN marker SYP41 (Bassham et al., 2000) or the PVC marker SYP21 (Conceicao et al., 1997) followed by immunofluorescence confocal microscopy. Both SYP41 and SYP21 were present in the enlarged, GFPAtSNX2b-containing structures and the typical punctate organelles labeled by these antibodies in wild-type cells (Ueda Figure 8. Viability of Arabidopsis Protoplasts. Arabidopsis protoplast viability was determined in control (untransformed) and GFPAtSNX2b-transformed Arabidopsis protoplasts using fluorescein diacetate, 24 h after transformation. et al., 2004; Dettmer et al., 2006; Tamura et al., 2007) are largely absent. This indicates that these enlarged structures are likely to be aberrant membrane structures or aggregates containing markers proteins from multiple organelles (Figure 9C and 9D). DISCUSSION The Sorting Nexin Family in Arabidopsis Arabidopsis has three sorting nexins named AtSNX1, AtSNX2a, and AtSNX2b (Vanoosthuyse et al., 2003; Jaillais et al., 2006). AtSNX1 is most similar to Brassica oleracea SNX1 and yeast Vps5p and has been suggested to function as Vps5p in the plant retromer complex (Jaillais et al., 2007). Vps17p, an additional SNX that partners with Vps5p in the retromer complex, does not have an easily identifiable homolog in Arabidopsis but one of the other SNXs may perform this function (Vanoosthuyse et al., 2003; Oliviusson et al., 2006). AtSNX1 functions in the trafficking of the auxin transport component PIN2 through a novel pathway independent from that of PIN1/GNOM (Jaillais et al., 2006). In addition, it has been implicated as a component of the retromer complex in Arabidopsis (Jaillais et al., 2007; Oliviusson et al., 2006). AtSNX2a and AtSNX2b are highly similar and have been suggested to be encoded by duplicate genes with redundant functions (Vanoosthuyse et al., 2003; Jaillais et al., 2006). This study presents the characterization of AtSNX2b as a SNX involved in trafficking in the Arabidopsis endosomal system. Typical of SNXs, AtSNX2b is present in a soluble and membrane-bound state, with the membrane-bound form most likely to be the active form in trafficking. Consistent with common motifs in SNXs, AtSNX2b has a PX domain in the N-terminal region of the protein and a C-terminal coiled-coil region that is likely to be a BAR domain. BAR domains form curved structures that can sense membrane curvature (Habermann, 2004; Peter et al., 2004; Ren et al., 2006) and, together with the PX domain, they target the SNX to specific phosphatidylinositol-lipid-rich organelles. The characteristic PX domain of AtSNX2b selectively binds to PI3P in vitro, although it is not sufficient for correct localization of GFP in vivo, suggesting that other factors or regions of the protein are also required for membrane targeting. PI3P lipids are reported in various species to be most abundant in endosomes (Gillooly et al., 2000; Gruenberg, 2003; Lemmon, 2003; Vermeer et al., 2006), Golgi (Gillooly et al., 2001; Vermeer et al., 2006), TGN (Kim et al., 2001), PVC (Vermeer et al., 2006), and the vacuole (Kim et al., 2001; Vermeer et al., 2006). Using a GFPfused endosomal binding domain in Arabidopsis, Kim et al. (2001) localized PI3P to the TGN, PVC, tonoplast, and vesicles. In addition, they proposed PI3Ps to be synthesized at the TGN and transported from the TGN through the PVC to the central vacuole, presumably for degradation by vacuolar hydrolases. Consistent with these results, AtSNX2b is localized to endosomes, TGN, and the PVC, as determined by co-localization with markers for these compartments and with the fluorescent endocytic marker FM4-64. This localization to multiple compartments suggests that AtSNX2b may cycle between the TGN, PVC, and endosomes, and may function in trafficking between or through these organelles. To further analyze the function of AtSNX2b, we have isolated a T-DNA knockout mutant in the AtSNX2b gene. Expression of the AtSNX2b mRNA and protein is lost in this mutant, demonstrating that it is a null mutant and is expected to have completely lost AtSNX2b function. As the AtSNX2a protein shows a high degree of sequence similarity to AtSNX2b (86% amino acid similarity), and both genes are expressed ubiquitously throughout the plant (Figure 1 and www.genevestigator.ethz.ch; Zimmermann et al., 2004), we hypothesize that these two proteins may perform redundant functions. One possibility is that AtSNX2a and/or AtSNX2b may function as the Vps17 constituent in Arabidopsis retromer, as a component of the sub-complex also containing AtSNX1. However, no obvious phenotype is evident for the Atsnx2b mutant, either at a morphological level, in protein trafficking pathways, or in auxin-related responses such as gravitropism (data not shown). In addition, we have been unable to detect an interaction between AtSNX2b and AtSNX1 in pull-down assays after transient expression. An Atsnx2a/Atsnx2b double mutant will help to clarify this issue. Overexpression of AtSNX2b Inhibits Vesicle Trafficking Overexpression of trafficking components can lead to inhibition of trafficking by disruption of the dynamics of the trafficking process as a result of sequestration of receptors or other factors involved in trafficking (Barr et al., 2000). Expression of high levels of GFP-tagged AtSNX2b leads to the formation of large, GFPAtSNX2b-containing compartments in the cell. These compartments are most likely enlarged or aggregated endosomes as they accumulate FM4-64, which becomes trapped in these compartments and can no longer reach the vacuolar membrane. They may also contain membrane and cargo derived from multiple sources, as TGN and PVC markers accumulate within them. Overexpression of untagged AtSNX2b gave a similar phenotype in that FM4-64-labeled enlarged compartments are present and FM4-64 in these compartments does not reach the vacuolar membrane. AtSNX2b overexpression also resulted in partial accumulation of the vacuolar marker NTPPGFP in endosomes, although most of the NTPPGFP still reached the vacuole. In a small but consistent number of cells, a portion of the NTPPGFP was arrested in the enlarged FM4-64-labeled endosomes. This relatively minor effect on biosynthetic cargo compared with endocytic cargo may suggest that the effect on NTPPGFP is a secondary effect of disrupting the structure of the endomembrane system, rather than indicating a direct role for AtSNX2b in vacuolar trafficking. Based on the localization of AtSNX2b on endosomes, the TGN, and PVC, and the overexpression phenotype of enlarged endosomes or endosomal aggregates and inhibition of transport through these endosomes, we hypothesize that AtSNX2b is involved in exit from endosomes. However, there are several possible explanations for the phenotype caused by AtSNX2b overexpression. First, the overexpression of AtSNX2b may interfere with the function of endogenous AtSNX2b. Second, overexpression could interfere with functions of other SNXs by sequestering components common to multiple pathways. In this case, the phenotype observed could be a result of the disruption of several trafficking pathways involving different SNXs. Finally, sequestration of PI3P due to AtSNX2b binding may prevent recognition of PI3P-rich lipid membranes by other proteins. The phenotype would therefore be a result of blocking multiple PI3P-dependent transport pathways. In this study, we have shown that AtSNX2b overexpression disrupts vacuolar protein trafficking through biosynthetic and endocytic pathways, suggesting that the AtSNX2b sorting nexin may be involved in protein trafficking. Additional experiments are now underway to define a more precise function of AtSNX2b in vesicle trafficking pathways. Plant Materials and Growth Conditions Arabidopsis thaliana seeds were surface-sterilized in 33% (v/v) bleach and 0.1% (v/v) triton X-100 solution for 20 min followed by cold treatment of at least 2 d at 4 C. Plants were grown on soil or MS solid medium (MurashigeSkoog Vitamin and Salt Mixture (Caisson Lab, Inc., North Logan, UT), 1% (w/v) sucrose (Sigma-Aldrich, St Louis, MO), 2.4 mM 2-morphinolino-ethanesulfonic acid (MES; Sigma-Aldrich, St Louis, MO) and 0.8% (w/v) phytoblend agar (Caisson Lab, Inc., North Logan, UT)) under long-day conditions with ambient light (16 h light; 100 lmol m 2 s 1) at 22 C. Arabidopsis thaliana suspension cell cultures were maintained as described by Contento et al. (2005). A homozygous Atsnx2b knockout (GABI_105E07) mutant line was received from the GABI-Kat mutant collection at the Max-Planck-Institute for Plant Breeding Research (Rosso et al., 2003). The T-DNA insertion site was verified by GABI-Kat personnel by sequencing the junction between the T-DNA left border and the AtSNX2b first exon. Additional verification of the mutant allele was done by analysis of segregation of the sulfadiazine resistance marker encoded by the T-DNA insertion by growing seedlings on MS medium containing 12 mg L 1 (4-amino-N-[2-pyrimidinyl]benzenesulfonamide-Na). RTPCR Analysis of AtSNX2b and AtSNX2a Total RNA was extracted from Arabidopsis organs (3-week-old roots grown in liquid culture, and 6-week-old mature plants grown in soil for aerial organs) using the TRIzol RNA isolation method (http://www.arabidopsis.org/portals/masc/AFGC/ RevisedAFGC/site2RnaL.htm#isolation) with DNase I treatment. cDNAs were generated using Superscript III reverse transcriptase (Invitrogen, Carlsbad, CA, USA) using an oligo dT primer. Gene-specific primers were used to amplify a 1.7-kb fragment containing the complete open reading frame of AtSNX2b (At5g07120); (forward) 5#-GGATCCAAAGAAGAGATGGAGAAAC-3 and (reverse) 5#-GGATCCAATTACACTGTGCTCTCATG-3 or 1.8-kb fragment containing the complete open reading frame of AtSNX2a (At5g58440); (forward) 5#ACTCCAGAGAAGTCGAAATG-3 and (reverse) 5#-CCAAAATGAACATCGTTCAC-3#. Introduced restriction sites are underlined. For each sample, 0.25 lg of cDNA was amplified for 29 (AtSNX2b) or 31 (AtSNX2a) cycles, with an annealing temperature of 54 C. Plasmid Construction A GFPAtSNX2b fusion was constructed using a pJ4GFPXB vector (Igarashi et al., 2001) with modifications as described in Contento et al. (2005). A BamHIBamHI fragment consisting of the coding region of the AtSNX2b cDNA (At5g07120) was amplified as above. The BamHI-digested fragment was then sub-cloned into the modified pJ4GFPXB digested with BglII. Similarly, a GFPAtSNX2b-1 mutant fusion construct was made using a three-step PCR mutagenesis (Li and Shapiro, 1993) amplifying a 233RR/LG amino acid mutation using mutation primers 5#-GTGGAGCAGCTAGGAGTTGCATTGG-3 (forward) and 5#-CCAATGCAACTCCTAGCTGCTCCAC-3 (reverse) in the first two steps followed by amplification of the full AtSNX2b-1 mutant using AtSNX2b-specific primers (above) and finally sub-cloning into the modified pJ4GFPXB at the same BglIIdigested site. GFPPX and GFPPX mutant fusion constructs were made using a similar procedure. A digested BamHIBamHI fragment of the PX domain of AtSNX2b (amino acids 142257) was subcloned into a BglII-digested pJ4GFPXB to generate GFPPX construct. The mutant PX domain (PX-1, amino acid 233RR/LG) was generated using the mutation primers (above) using the three-step PCR mutagenesis. A second PX domain mutant (PX-2, amino acid 211PP/AA) was generated through the same three-step PCR mutagenesis using primers 5#-CTGCATTGCAGCGAGGCCAGATAA-3 (forward) and 5#-CTTATCTGGCCTCGCTGCCAATGCAG-3 (reverse). PX primers used were 5#-GGATCCCCGAATTCCCGGGTCGA-3 (forward) and 5#GGATCCTCAGTCACTCAAAGCGGTAACTTCCC-3 (reverse) for PX, PX-1, and PX-2 constructs. Restriction sites are underlined. The digested BamHIBamHI PX fragment was then cloned into BglII-digested pJ4GFPXB. Antibody Production and Purification The AtSNX2b coding region flanked by SalINotI restriction sites was generated by RTPCR from total Arabidopsis RNA using the following primers: (forward) 5#-GTCGACTCCCATCTCCACTCATCC-3 and (reverse) 5#-GCGGCCGCATTACACTGTGCTCTC-3 and sub-cloned into pET28b (Novagen, Madison WI) to produce HISAtSNX2b fusion plasmid. The fusion protein was synthesized in E. coli according to the Novagen protocol, where it accumulated in inclusion bodies. HIS-fusions were purified following the manufacturers protocol using HIS bind resin (Novagen, CAT# 70666). Cells were broken by sonication, and insoluble material was pelleted at 24 000 g. The pellet was re-suspended in 6 M urea in HISbinding buffer for 1 h at 4 C and pelleted again by centrifugation at 24 000 g. After centrifugation, the supernatant was incubated with HIS bind resin for 30 min and eluted with 100 mM imidazole elution buffer. The eluted protein was separated by SDSPAGE (200 lg), cut from the gel and used to immunize rabbits. For affinity purification of antibodies, purified HIS AtSNX2b protein was separated by SDSPAGE, transferred to nitrocellulose, and the strip containing the fusion protein was cut out after staining with Ponceau S. After blocking in 3% (w/v) dried non-fat milk in PBS, serum was incubated with the strip for 2 h at 4 C to allow binding of the antibodies. The strip was washed with PBS and specific antibodies were eluted using 100 mM glycine, pH 2.5. The eluate was adjusted to pH 7.0 using 2 M Tris-HCl, pH 8. These affinity-purified antibodies were used in all further experiments. Lipid Overlay Assay Wild-type PX domain and the mutant PX domain PX-1 were generated by PCR using the primers 5#-GTCGACCGCTCTGATTACATCAAGATC-3 (forward) and 5#-GCGGCCGCTCAAAGCGGTAACTTCCCTTGCG-3 (reverse) from either GFPPX or GFPPX1. SalINotI fragments of PX and PX-1 were sub-cloned into pGEX5X-1 (Amersham Bioscience, Piscataway, NJ) to yield GSTPX and GSTPX-1. GSTPX and GSTPX-1 fusion proteins were synthesized in E. coli according to the Novagen protocol. GST-fusions were purified using GST bind resin (Novagen, CAT# 70541) following the manufacturers protocol. Lipid overlay assays were performed according to Dowler et al. (2002). In brief, purified GST-fusions were allowed to bind to PIP lipid strips (Echelon Biosciences Inc., Salt Lake, UT) followed by immunoblotting using GST antibodies (Invitrogen, Carlsbad, CA) to detect bound GST-fusions. Transient Transformation of Arabidopsis Protoplasts For suspension cells, protoplasts were prepared and transformed according to Contento et al. (2005). For leaf tissue, protoplasts were prepared and transformed according to Sheen (2002). Protoplasts were transformed with 30 lg of DNA per transformation. Images were obtained using fluorescence and confocal laser microscopy. Immunofluorescence Three-day-old Arabidopsis seedlings grown on MS solid medium were fixed following Sivaguru et al. (1999). Plants were transferred into 5 mL of MTSB buffer (50 mM PIPES-KOH (pH 6.9), 5 mM EGTA, and 5 mM MgSO4) containing 5% (v/v) dimethyl sulfoxide for 15 min at room temperature. Afterwards, they were fixed with 4% (w/v) paraformaldehyde in the above buffer containing 10% (v/v) dimethyl sulfoxide for 60 min at 20 C, with the initial 10 min under vacuum. Plants were then washed with MTSB prior to immunostaining. Immunofluorescence staining of treated Arabidopsis seedlings was performed according to M uller et al. (1998). Treated seedlings were incubated with primary antibodies in a humid chamber for 1518 h at 4 C, washed three times for 5 min in MTSB and further incubated for 45 min2 h at room temperature with conjugated secondary anti-rabbit or anti-mouse IgG antibodies in 3% (w/v) bovine serum albumin in MTSB. Seedlings were washed five times with MTSB and mounted with a coverslip in 50% (v/v) glycerol in phosphate-buffered saline (PBS, pH 7.4). Primary antibodies used were T7 tag monoclonal antibodies (Novagen/EMD Biosciences, Inc., La Jolla, CA; 1:100), anti-AtSNX2b antibodies (1:200) and preimmune antibodies (for AtSNX2b; 1:200). Secondary antibodies used were Alexa Fluor 594-conjugated goat antirabbit IgG, Alexa Fluor 488-conjugated goat anti-mouse IgG and Alexa Fluor 488-conjugated goat anti-rabbit IgG or Alexa Fluor 594-conjugated goat anti-rabbit IgG (Molecular Probes, Eugene, OR, USA; 1:250). For protoplasts, transformed protoplasts were fixed in 4% (w/v) paraformaldehyde in MTSB buffer for 20 min followed by three washes with MTSBS (MTSB containing 0.4 M sorbitol) prior to immunostaining. Protoplasts were permeabilized using permeabilization solution (3% (v/v) Triton X-100, 10% (v/v) dimethyl sulfoxide in MTSB) for 20 min. Immunofluorescence staining of fixed protoplasts was performed according to Kang et al. (2001). Cells were washed five times with MTSB and mounted with a coverslip in 50% (v/v) glycerol in PBS. Antibodies used were anti-AtSNX2b antibodies (1:200), anti-SYP41 antibodies (1:200; Bassham et al., 2000), and anti-SYP21 antibodies (1:200; Conceicao et al., 1997) for primary antibodies and Alexa Fluor 488-conjugated goat anti-rabbit IgG or Alexa Fluor 594-conjugated goat anti-rabbit IgG as secondary antibodies. Fluorescent signal detection and documentation was performed using a confocal laser scanning microscope (Leica TCS/NT, Leica Microsystems, Exton, PA, USA). The confocal laser microscope utilizes a Krypton 568-nm and Argon 488-nm laser for excitation. Filters for emission were RST588 BP525625 (FITC-specific detection) and LP590 (TRITC-specific detection). Images were further processed for graphic presentation using Adobe Photoshop (Adobe Systems, Mountain View, CA, USA). For all experiments, controls were performed consisting of omission of both primary antibodies (to control for nonspecific staining), omission of only one primary antibody (to confirm that no fluorescence bleed-through between filters was visible in double labeling), and omission of secondary antibodies or all immunochemicals (to control for fixativeinduced autofluorescence). The controls confirmed the absence of non-specific fluorescence. All experiments were carried out at least three times with cells from independent preparations. Immunoblot Analysis Arabidopsis plants or suspension cells were homogenized in PBS, 1 mM EDTA, 0.1 mM phenylmethylsulfonyl fluoride. The homogenate was centrifuged at 1000 g for 10 min at 4 C to remove cell debris and large organelles. The pellet was discarded and the supernatant was incubated in SDS-reducing sample buffer (Biorad, Hercules, CA) for 5 min at 65 C, and separated by electrophoresis on 10% SDSPAGE gels. Proteins were electrotransferred to nitrocellulose membranes; blots were blocked with PBS/4% low-fat milk powder for at least 1 h and incubated with anti-AtSNX2b antibodies for 1518 h at 4 C. Signal detection was achieved using peroxidaseconjugated secondary antibodies and chemiluminescence reaction followed by X-ray film exposure. For differential centrifugation, the 1000-g supernatant from suspension cells was centrifuged sequentially to produce 12 000, 39 000, and 125 000 g pellets and a 125 000-g supernatant that were analyzed by immunoblot with AtSNX2b antibodies or GFP antibodies (Invitrogen, Carlsbad, CA). Extraction of AtSNX2b from Membranes A 1000-g supernatant from suspension cells was centrifuged at 125 000 g to produce a total membrane pellet (P125). The supernatant fraction was discarded and membrane pellets were re-suspended in 200 lL of extraction buffer (PBS, 1 mM EDTA, 0.1 mM phenylmethylsulfonyl fluoride) or extraction buffer containing 0.1 M Na2CO3, 1 M NaCl, 2 M urea, or 1% (v/v) triton X-100, and incubated for 2 h on ice. Insoluble material was pelleted at 125 000 g and pellets were re-suspended in SDS sample buffer. Supernatants were precipitated using TCA, and protein pellets were washed in acetone and re-suspended in SDS sample buffer. Samples were analyzed by SDSPAGE and immunoblotting using AtSNX2b antibodies, or SYP41 antibodies as a control (Bassham et al., 2000). Sucrose Gradients Five-day-old Arabidopsis suspension cultures were homogenized in HKE buffer (50 mM Hepes-KOH, pH 7.5, 10 mM potassium acetate, and 1 mM EDTA) containing 400 mM Suc, 1 mM dithiothreitol, and 0.1 mM phenylmethylsulfonyl fluoride, and centrifuged at 1000 g for 5 min to generate a post-nuclear supernatant. The supernatant was loaded onto a sucrose step gradient as described in Sanderfoot et al. (1998). Gradients were centrifuged at 150 000 g in a swinging-bucket rotor at 4 C for 18 h. Fractions (1 mL) were collected from the top of the gradient. Protein in each fraction was analyzed by SDSPAGE and immunoblotting. Blots were probed using antibodies against aleurain (1:2000; Ahmed et al., 2000), cTIP (1:500; Hicks et al., 2004), FUM1 (fumarase; 1:500; Behal and Oliver, 1997), SYP21 (1:1000; Sanderfoot et al., 2001), AtSNX2b (1:1500), and VTI12 (1:500; Bassham et al., 2000) followed by secondary antibodies conjugated to horseradish peroxidase. FM4-64 Staining of Protoplasts Protoplasts were stained with FM4-64 according to Ueda et al. (2001) by incubation for 10 min at 4 C with 50 lM FM4-64 in MS medium containing 0.4 M mannitol. They were washed three times with the same medium, followed by incubation at room temperature for 30 min to 12 h. Confocal microscopy was performed with a Leica TCS/NT confocal microscope (Leica Microsystems, Exton, PA, USA) as described above. Viability Assays Protoplasts were incubated in 50 lg mL 1 fluorescein diacetate for 30 min followed by visualization by fluorescence microscopy using a FITC filter. Counts for viable (fluorescent) and nonviable (unstained) cells were performed and recorded. Four replicates of at least 700 cells per treatment were analyzed. This work was supported by the National Science Foundation (grant number IOB-0515998 to D.C.B.) and the Iowa State University Plant Sciences Institute (grant to D.C.B.). ACKNOWLEDGMENTS We thank Drs Chris Hawes, Erik Nielsen, David Oliver, Natasha Raikhel, and Tony Sanderfoot for antibodies, constructs, and transgenic lines, Margie Carter (ISU Confocal Microscopy and Image Analysis Facility), and Tracey Pepper (ISU Microscopy and Nanoimaging Facility) for valuable assistance and expertise in microscopy and Tony Contento for helpful comments on the manuscript. No conflict of interest declared.

This is a preview of a remote PDF: https://mplant.oxfordjournals.org/content/1/6/961.full.pdf

Nguyen Q. Phan, Sang-Jin Kim, Diane C. Bassham. Overexpression of Arabidopsis Sorting Nexin AtSNX2b Inhibits Endocytic Trafficking to the Vacuole, Molecular Plant, 2008, 961-976, DOI: 10.1093/mp/ssn057