The exocyst complex contributes to PIN auxin efflux carrier recycling and polar auxin transport in Arabidopsis


  • Edita Janková Drdová,

    1. Institute of Experimental Botany, Academy of Sciences of the Czech Republic, Prague 6, Czech Republic
    Search for more papers by this author
  • Lukáš Synek,

    1. Institute of Experimental Botany, Academy of Sciences of the Czech Republic, Prague 6, Czech Republic
    Search for more papers by this author
  • Tamara Pečenková,

    1. Institute of Experimental Botany, Academy of Sciences of the Czech Republic, Prague 6, Czech Republic
    2. Department of Plant Physiology, Faculty of Science, Charles University, Prague 2, Czech Republic
    Search for more papers by this author
  • Michal Hála,

    1. Institute of Experimental Botany, Academy of Sciences of the Czech Republic, Prague 6, Czech Republic
    2. Department of Plant Physiology, Faculty of Science, Charles University, Prague 2, Czech Republic
    Search for more papers by this author
  • Ivan Kulich,

    1. Department of Plant Physiology, Faculty of Science, Charles University, Prague 2, Czech Republic
    Search for more papers by this author
  • John E. Fowler,

    1. Center for Genome Research and Biocomputing, Department of Botany and Plant Pathology, Oregon State University, Corvallis, OR, USA
    Search for more papers by this author
  • Angus S. Murphy,

    1. Department of Horticulture & Landscape Architecture, Purdue University, West Lafayette, IN, USA
    Search for more papers by this author
  • Viktor Žárský

    Corresponding author
    1. Department of Plant Physiology, Faculty of Science, Charles University, Prague 2, Czech Republic
    • Institute of Experimental Botany, Academy of Sciences of the Czech Republic, Prague 6, Czech Republic
    Search for more papers by this author

  • We declare that no authors have any conflicts of interest.

For correspondence (e-mail


In land plants polar auxin transport is one of the substantial processes guiding whole plant polarity and morphogenesis. Directional auxin fluxes are mediated by PIN auxin efflux carriers, polarly localized at the plasma membrane. The polarization of exocytosis in yeast and animals is assisted by the exocyst: an octameric vesicle-tethering complex and an effector of Rab and Rho GTPases. Here we show that rootward polar auxin transport is compromised in roots of Arabidopsis thaliana loss-of-function mutants in the EXO70A1 exocyst subunit. The recycling of PIN1 and PIN2 proteins from brefeldin–A compartments is delayed after the brefeldin-A washout in exo70A1 and sec8 exocyst mutants. Relocalization of PIN1 and PIN2 proteins after prolonged brefeldin-A treatment is largely impaired in these mutants. At the same time, however, plasma membrane localization of GFP:EXO70A1, and the other exocyst subunits studied (GFP:SEC8 and YFP:SEC10), is resistant to brefeldin-A treatment. In root cells of the exo70A1 mutant, a portion of PIN2 is internalized and retained in specific, abnormally enlarged, endomembrane compartments that are distinct from VHA-a1-labelled early endosomes or the trans-Golgi network, but are RAB-A5d positive. We conclude that the exocyst is involved in PIN1 and PIN2 recycling, and thus in polar auxin transport regulation.


Auxin functions as a key regulator of polarity and morphogenesis in land plants. Although most plant cells can synthesize and metabolize auxin, the regulation of plant morphogenesis is largely dependent on long-range polarized transport of the compound from the shoot to the root. This transport relies on an array of membrane transporters: influx carriers (AUX1/LAX), efflux carriers of the PIN family and pumps of the ABCB family (reviewed in Zažímalová et al., 2010).

A crucial factor in the generation of developmentally relevant polarized auxin flows is the localization of PIN auxin efflux carriers in polarized plasma-membrane domains in auxin-conducting cells (Gälweiler et al., 1998; Wiśniewska et al., 2006). The polar localization of PIN proteins is associated with their constitutive cycling between the plasma membrane (PM) and endosomal compartments – a process sensitive to the ARF-GEF inhibitor brefeldin A (BFA), which induces aggregations of the trans-Golgi network and endosomal vesicles into so-called BFA compartments (Geldner et al., 2001, 2003; Dhonukshe et al., 2007, 2008). PIN proteins synthesized de novo are initially delivered to the PM in a non-polar manner, and their polar localization in PM domains is established by subsequent endocytic recycling, which depends on clathrin-mediated endocytosis (Dhonukshe et al., 2007, 2008; Kitakura et al., 2011). Only a few regulators of PIN polar localization have been described so far: Ser/Thr kinase PINOID (PID; Friml et al., 2004), ENHANCER OF PINOID (ENP; Furutani et al., 2007, 2011), protein phosphatase 2A (Michniewicz et al., 2007), ARF guanine nucleotide exchange factor GNOM (Steinmann et al., 1999; Geldner et al., 2001, 2003) and the ICR1 protein mediating the ROP–SEC3 interaction (Lavy et al., 2007).

The PM of eukaryotic cells is a highly dynamic system that continuously undergoes exocytosis and endocytosis. These recycling processes are tightly regulated on many different levels. Polarized exocytosis is controlled by the coordinated action of small GTPases (mainly from Rab and Rho families), the vesicle-tethering complex exocyst (as their effector), SNARE proteins and regulatory syntaxin-binding proteins (Koumandou et al., 2007; reviewed in Žárský et al., 2009).

The exocyst is an evolutionarily conserved complex that consists of eight subunits (Sec3, Sec5, Sec6, Sec8, Sec10, Sec15, Exo70 and Exo84). In yeast and mammals, the exocyst acts at specific domains of the PM that exhibit extensive fusion of exocytic vesicles. Research in yeast suggests that Exo70 and Sec3 subunits serve as spatial PM landmarks for the assembly of the exocyst complex that interacts with, and might also be activated by, Rho GTPases to coordinate the tethering of incoming secretory vesicles before SNARE complex formation (Boyd et al., 2004; He et al., 2007; Zhang et al., 2008; Wu et al., 2010).

Phylogenetic analyses indicate that all exocyst subunits are present in land plants (Cvrčková et al., 2001; Eliáš et al., 2003; Koumandou et al., 2007), and that the EXO70 gene has undergone a dramatic evolutionary expansion: e.g. in Angiosperms, with 22 presumably functional genes present in Arabidopsis thaliana (Eliáš et al., 2003; Synek et al., 2006). By contrast, only a single copy of EXO70 is present in yeast and animals, suggesting an elaboration of functional or tissue-specific specialization in plants (Žárský et al., 2009). Phenotypic analysis of exo70 mutants and expression patterns of EXO70s in Arabidopsis indicate that the EXO70A1 isoform plays a major role in the sporophyte (Synek et al., 2006). The plant exocyst complex has been shown to function in the cell morphogenesis of growing pollen tubes, root hairs and etiolated hypocotyls (Cole et al., 2005; Wen et al., 2005; Synek et al., 2006; Hála et al., 2008), as well as in cytokinesis and seed coat formation (Fendrych et al., 2010; Kulich et al., 2010).

As Arabidopsis exo70A1 mutants exhibit phenotypes indicative of altered auxin transport and distribution, such as loss of apical dominance, retarded root growth, delayed lateral root initiation or decreased growth of root hairs (Synek et al., 2006), we hypothesized that EXO70A1 (possibly as a part of the exocyst complex) may affect auxin transport/distribution and polar targeting of PIN auxin transporters to the PM. This possibility is supported by a recent report that the Arabidopsis ICR1 adaptor protein, which links activated ROP GTPases and the SEC3 exocyst subunit, also regulates polar auxin transport (PAT; Lavy et al., 2007; Hazak et al., 2010). Based on our results, we conclude that the exocyst complex is involved in PIN1 and PIN2 recycling and localization, with a direct impact on polar auxin transport processes.


Rootward auxin transport is compromised in the exo70A1 mutant

We previously described symptoms of the A. thaliana exo70A1 mutant pointing to possible changes in auxin homeostasis or signalling (Synek et al., 2006). Further analysis revealed new deviations supporting such a notion. We observed a loss of root waving and a skewing to the opposite side, compared with the wild type (WT), in plants cultivated on agar plates inclined at 50° (Figure 1a). Also, distinctly larger root caps caused by accumulation of border-like cell layers were obvious in 7-day-old exo70A1 seedlings (Figure 1b,c). Microscopic inspection of the root columella cells revealed a reduced level of statolite starch, as evaluated by image analysis after Lugol's staining (exo70A1, 2063 ± 1660 pixels, n = 6; WT, 7501 ± 2915 pixels, n = 8; Student's t-test P = 0.0004; Figure 1b). However, root gravitropism was unaffected, as the only root gravitropic defect observed in exo70A1 plants was transient primary root disorientation during seed germination. Whereas only 3% of WT primary roots started to grow in a direction differing by more than 45° from the gravity vector (n = 370), as many as 25% of exo70A1 roots showed such disorientation (n = 117). Normal root growth direction, however, was established in those seedlings in the 3 days after germination.

Figure 1.

Root phenotypes and auxin transport in the Arabidopsis exo70A1 mutant. (a) Wild-type (WT) seedlings grown on inclined agar surfaces (at a 50° angle from the horizontal) exhibit the characteristic waving and skewing to the left. Waving is suppressed and skewing to the opposite side occurs in exo70A1 mutants. (b) Starch staining in root tips of 7-day-old seedlings revealed the reduced level of statolite starch in exo70A1 root tips. The significantly enlarged root cap is also apparent in exo70A1. Scale bar: 100 μm. (c) Visualization of auxin-response distribution in root tips of 4-day-old seedlings expressing the ProDR5:GUS reporter showed a reduced auxin level in exo70A1 root tips compared with the WT. Scale bars: 100 μm. (d) Nanoscale rootward auxin transport was measured in roots after [3H]IAA was applied to the root–shoot junction. Gravacin, 1-NOA or no auxin transport inhibitors were used. After 12 h, the level of [3H]IAA was determined in the roots, and revealed a significant reduction in auxin transport in exo70A1. DPM, disintegrations per minute, ±SDs are indicated; n = 10 for each column. (e) Rootward auxin transport was examined by the application of IAA to the root–shoot junction (marked by arrows) of vertically grown ProDR5:GUS-expressing seedlings. Staining was performed 0, 3, 6 and 12 h after the application of IAA. An auxin canalization defect in exo70A1 is illustrated by images of whole seedlings, by detailed images of root–shoot junctions and by sections from the middle part of roots (the last column). Scale bars: 100 μm.

The ProDR5:GUS auxin response reporter (Ulmasov et al., 1997) was used to reveal defects of exo70A1 roots in auxin metabolism, transport or signalling. The GUS signal was weaker in the root tips of 4-day-old light-grown exo70A1 seedlings than in the WT (exo70A1, 27.5 ± 4.5, n = 10; WT, 40.1 ± 5.3, n = 10; Student's t-test P = 0.0003; Figure 1c). Furthermore, rootward (acropetal) transport of [3H]indole-3-acetic acid (IAA) from the shoot–root transition zone was assayed, and was found to be decreased by ~40% in exo70A1-2 (disintegrations per minute, DPM = 829 ± 91; n = 10) compared with the WT (DPM = 1321 ± 69; n = 10; Student's t-test P < 0.0001; Figure 1d). A comparable decrease in auxin transport was also detected for exo70A1-1, another exo70A1 mutant allele, but not for GFP:EXO70A1 complemented exo70A1-2 plants (Figure S1). Subsequent analyses of ProDR5:GUS expression after external application of IAA locally on the shoot–root transition zone confirmed our previous observation (Figure 1e). In WT plants, IAA was transported towards the root tip and induced ProDR5:GUS expression along the vascular tissue within 3 h, thereby reporting efficient IAA canalization into the stele and transport to the root tip (Figure 1e). In exo70A1, on the contrary, IAA was accumulated around the region of application, was improperly canalized and the GUS signal along most of the vascular tissue length was still absent 3 h after application (Figure 1e; quantitative analysis in Figure S2), indicating that the rootward auxin transport was retarded.

The observed decrease in rootward auxin transport is more consistent with what has been reported for loss of PIN function (25–35%) rather than the greater decreases (60–80%) seen in abcb mutants (Blakeslee et al., 2007). Consistent with an impact primarily on PIN-mediated auxin transport, the ABCB19 inhibitor gravacin (Rojas-Pierce et al., 2007) inhibited the rootward transport in exo70A1 only 15% more than in the WT, and the AUX1/LAX auxin uptake inhibitor 1-NOA had no less effect on transport in exo70A1 when applied to seedlings 1 h after auxin application (Figure 1d). We conclude that rootward PAT is severely affected in Arabidopsis exo70A1 mutants, possibly in a PIN-dependent way.

The recycling of plasma membrane proteins is disturbed in exo70A1 mutant plants

The exo70A1 mutant accumulates a portion of PIN2:GFP in abnormally expanded endomembrane compartments that can be visualized with FM4-64 dye 30 min after application, and that are absent in the WT (Figure 2a). These compartments were observed in all exo70A1-2 seedlings analysed with an average incidence of 7.0% of root cells per plant (12 plants, total cell number = 330). We did not detect such compartments in any WT plant analysed (n = 12, total cell number = 290). In the case of exo70A1-1 (Figure S3), we detected the abnormal compartments with an average incidence of 6.5% in all mutant seedlings studied (10 plants, total cell number = 353). Again, no such compartment was detected in WT plants (n = 9, total cell number = 338). The difference in the incidence of these abnormal compartments between WT and mutant plants was tested using the generalized linear models with binomial distribution of errors (GLMbinom). The weight of each individual plant in the model was proportional to the number of cells analysed. The significance of GLMbinom was tested through the χ2 test (Faraway, 2006) and found to be highly significant (P < 0.0001).

Figure 2.

Root epidermal cells of exo70A1 retain PIN2:GFP in abnormally expanded endomembrane compartments. Seedlings expressing PIN2:GFP (a), YFP:RAB-A5d (b) or VHA-a1:GFP (c), respectively, were stained by FM4-64 dye 30 min before imaging. An accumulation of PIN2:GFP in abnormally expanded endomembrane compartments was observed in exo70A1 root cells (marked by arrowheads). These structures (marked by arrowheads) colocalized with the FM4-64 dye (a) and were YFP:RAB-A5d positive (b), but were VHA-a1:GFP negative (c). Scale bars: 10 μm.

These abnormal endomembrane compartments are positive for YFP:RAB-A5d, a marker of recycling endosomes (Geldner et al., 2009) (Figure 2b), but are clearly distinct from the early endosome/trans-Golgi network marked by the VHA-a1:GFP marker (Dettmer et al., 2006) (Figure 2c).

A significant difference between WT and exo70A1 mutant cells was also observed in the re-establishment of the localization of PIN1 and PIN2 to the PM after BFA-induced internalization and subsequent BFA removal (GLMbinom P < 0.0001; the numbers of analysed seedlings and cells are indicated in Table S1). A 2-h BFA treatment resulted in the internalization of PIN1:GFP and PIN2:GFP in both WT and exo70A1 plants (Figure 3a,b). Washout of BFA resulted in the re-establishment of PM localization of both PIN1:GFP and PIN2:GFP after 30 min in the WT. In contrast, BFA compartments containing GFP-tagged PINs were still present in cells even 90 min after the washout in exo70A1 seedlings (Figure 3a,b). To exclude that the disaggregation of BFA compartments after BFA washout was affected by different accumulation of BFA compartments after 2 h of BFA treatment, we compared the proportion of affected cells after 2 h of BFA treatment in both WT and exo70A1 expressing PIN2:GFP. We found no difference between both genotypes (GLMbinom P > 0.05; Table S1).

Figure 3.

Both exo70A1 and sec8 mutants exhibit impaired recycling of PIN1 and PIN2 from BFA compartments to plasma membrane. Seedlings expressing PIN1:GFP (a) or PIN2:GFP (b) were imaged before and after 2 h of treatment with brefeldin A (BFA), 30 and 90 min after BFA washout, and after a prolonged 12-h treatment with BFA. The localization of PIN proteins in the plasma membrane was re-established by 30 min after washout in the wild type (WT). However, a delay in the recovery of PIN1:GFP and PIN2:GFP was observed in exo70A1 and sec8 plants even 90 min after BFA washout. The relocalization of PIN1:GFP and PIN2:GFP from BFA compartments to the PM after 12 h of BFA treatment was impaired in both exo70A1 and sec8 seedlings. Arrowheads point to sites of strong PIN2:GFP super-apicalization. Scale bars: 10 μm.

We observed the relocalization of both PIN1:GFP and PIN2:GFP proteins from BFA compartments to the apical PM domain after a prolonged 12 h of BFA treatment in WT cells, which is consistent with a previous report by Kleine-Vehn et al. (2008). In contrast, such relocalization was significantly impaired in exo70A1 mutants, where PIN proteins remained accumulated mainly in BFA compartments (GLMbinom, < 0.0001; Table S1; Figure 3a,b), pointing strongly to the EXO70A1 function in the recycling of PINs between endomembrane compartments and the PM. Although we observed the partial relocalization of PIN1:GFP from basal to apical PM, referred to as transcytosis in the WT, it was not so pronounced in exo70A1 cells (Figure 3a). Despite the fact that BFA compartments were still present after prolonged BFA treatment, partial PIN2:GFP super-apicalization, i.e. the relocalization of PINs to the middle of the apical membrane, was documented in exo70A1 epidermal cells (Figure 3b).

To confirm results from both transient and prolonged chronic BFA treatment, and to eliminate possible artefacts caused by GFP fusions, we also visualized the effect of BFA using anti-PIN1 and anti-PIN2 antibodies (Figure S4) as well as FM4-64 (Figure S5). These experiments provided data identical to results obtained with GFP-tagged PIN1 and PIN2, and fully proved our observations.

Similar defects in PIN1:GFP and PIN2:GFP recycling, after both transient and prolonged BFA treatment, were also observed in the sec8-m1 mutant partially complemented by ProLAT52:SEC8 (Cole et al., 2005) (Figure 3a,b). Again, the difference between WT and mutant plants was highly significant (GLMbinom, < 0.0001; Table S1).

Loss of EXO70A1 function also affected the trafficking of the brassinosteroid receptor BRI1, a transmembrane protein that constitutively cycles between endomembrane compartments and the PM through BFA-sensitive trafficking (Geldner et al., 2007). BFA treatment of WT and exo70A1 seedlings expressing BRI1:GFP resulted in the formation of BRI1-positive BFA compartments (Figure 4); recycling of BRI1:GFP back to the PM was significantly delayed after the washout, and relocalization of BRI1:GFP from BFA compartments to the PM after prolonged BFA treatment was blocked in exo70A1 mutants (Figure 4). In both cases, the difference between WT and mutant plants was highly significant (GLMbinom, < 0.0001; Table S1). These results suggest that the EXO70A1 exocyst subunit is required for a more general recycling of PM proteins.

Figure 4.

BRI1 recycling is disturbed in the exo70A1 mutant. Seedlings imaged before treatment with brefeldin A (BFA) showed comparable localization of BRI1:GFP to the plasma membrane in exo70A1 and wild-type (WT) seedlings. After 2 h of BFA treatment, exo70A1 and WT seedlings exhibited BRI1:GFP internalized in BFA compartments. Whereas BRI1:GFP was sufficiently relocalized to the plasma membrane by 30 min after BFA washout in the WT, it persisted as late as 90 min after washout in exo70A1 seedlings. The localization of BRI1:GFP to the plasma membrane was also restored during a prolonged 12-h treatment with BFA in the WT, whereas BRI1:GFP remained in the BFA compartments in exo70A1 seedlings. Scale bars: 10 μm.

The subcellular localization of exocyst subunits is unaffected by treatment with brefeldin A

After transformation of heterozygous exo70A1 plants with Pro35S:GFP:EXO70A1, the restoration of WT phenotypes was observed in exo70A1 homozygotes (Figure S6). GFP:EXO70A1 was localized evenly at the PM, partially in the cytoplasm and at post-cytokinetic cross walls in both transformed mutant plants and WT (Figure 5a; Fendrych et al., 2010). We also recorded a similar localization for other exocyst subunits ProSEC8:GFP:SEC8 and Pro35S:YFP:SEC10 (Figure 5a). The PM localization of all exocyst subunits tested was resistant to treatment with 50 μm BFA for 2 h (Figure 5b).

Figure 5.

Exocyst subunits localize to the plasma membrane, and their localization is brefeldin-A resistant. (a) GFP:EXO70A1, GFP:SEC8 and YFP:SEC10 exhibit cytosolic and non-polar localization to the plasma membrane (upper row). Labelling by FM4-64 is displayed in the lower row. Scale bars: 10 μm. (b) Whereas treatment with BFA for 2 h caused the typical internalization of PIN2:GFP into BFA compartments, the localization of fluorescently labelled exocyst subunits remain unchanged (upper row). Labelling by FM4-64 is displayed in the lower row. Scale bars: 10 μm.


In this study, we documented that exocyst components participate in the recycling of PIN1 and PIN2 auxin transporters. Such a function is consistent with the observed exo70A1 mutant phenotypes, including compromised apical dominance, retarded root hair elongation and delayed lateral root initiation: processes dependent on auxin level (e.g. reviewed in Benjamins and Scheres, 2008). Strikingly similar phenotypic deviations in comparison to exo70A1 are exhibited by weak gnom alleles in Arabidopsis: gnomR5 and gnomB/E. Whereas loss-of-function alleles of the GNOM gene lead to severe defects in the early embryonic development, characterized by the perturbed establishment of an embryonic axis, lack of the embryonic root and fused or missing cotyledons (Mayer et al., 1993), the weak alleles are nearly indistinguishable from the WT up to 5 days after germination. Weak gnom alleles are defective in lateral root formation (Geldner et al., 2004), similar to exo70A1 and sec8-m1 mutants. After treatment with synthetic auxin (NAA), strongly increased expression of the auxin-response reporter DR5::GUS throughout the whole root tissue of the gnomR5 mutant was observed, whereas WT effectively canalized and transported auxin (Geldner et al., 2004). This observation strongly resembles the defect in exo70A1 plants and points to the auxin transport defect. We did not observe any apicalization of PIN1 or PIN2 in exo70A1 mutants, as described for weak gnom alleles (Kleine-Vehn et al., 2008; Ikeda et al., 2009), because the exocyst probably functions in different steps of recycling downstream of GNOM. We conclude that EXO70A1 and SEC8 (most probably as a part of the entire exocyst complex) also participate in PIN PM recycling.

The loss of EXO70A1 function resulted in a completely perturbed root-waving phenotype in the mutants. Waving has been proposed to reflect circumnutation, gravitropism and negative thigmotropism, and it has been supposed that auxin transport and signalling is also involved in this process (Migliaccio and Piconese, 2001; Santner and Watson, 2006). Recently, WAG1 and WAG2 proteins, being suppressors of waving (Santner and Watson, 2006), were shown to act redundantly in PINOID function (Dhonukshe et al., 2010). Unlike exo70A1, both single and double mutants in WAG1 and WAG2 display an exaggerated waving phenotype (Santner and Watson, 2006). It was proven that WAG1 and WAG2 as well as PINOID are ACG kinases that instruct basal-to-apical PIN polarity shifts (Friml et al., 2004; Dhonukshe et al., 2010). These findings corroborate a possibility that changes in the root-waving phenotype and root tip auxin maxima reflect changes in auxin distribution in exo70A1 caused by compromised PIN1 and PIN2 recycling.

We have observed only slight temporary defects in the root gravitropic response during seed germination of exo70A1 mutants, despite the fact that a compromised gravitropic response is a common feature of auxin transport and signalling mutants (Bennett et al., 1996; Chen et al., 1998). Thus, the perturbation of PIN1 and PIN2 cycling is not so severe to alter gravitropic root responses in exo70A1 mutants, suggesting functional redundancy between EXO70 isoforms in the PIN-dependent lateral auxin redirection involved in response to gravity (Friml et al., 2002). Functional redundancy between EXO70 isoforms is highly probable, and might be very complex: we previously reported that at least 16 out of 23 EXO70s are expressed in Arabidopsis root cells, with EXO70A1 being the most abundantly expressed in both roots and shoots (Synek et al., 2006; Genevestigator database). Also, in the case of PIN proteins, all five root-expressed PINs were documented to partially compensate each other (Blilou et al., 2005; Vieten et al., 2005).

We also described here a defect in the separation of root cap border-like cells in exo70A1 mutants. The root cap architecture is known to be regulated by the secretion of cell wall-macerating enzymes (Durand et al., 2009), as well as by auxin distribution (Sabatini et al., 1999; Ponce et al., 2005; Jiang et al., 2006). The mechanism of the malformed root cap formation in exo70A1 will be a subject of our further study.

In root tissues of the exo70A1 mutant, we observed significantly retarded rootward auxin transport and perturbed proper canalization of externally applied IAA. At the cellular level, exo70A1 accumulates a portion of PIN2:GFP in abnormally enlarged compartments that do not colocalize with the early endosome/trans-Golgi network marker VHA-a1:GFP, but overlap with the YFP:RAB-A5d recycling endosome marker. Other Arabidopsis exocyst mutants, exo84b-1 and exo84b-2, exhibit pronounced secretory defects, demonstrated by the accumulation of unspecified enlarged vesicles (Fendrych et al., 2010). The formation of aberrant endomembrane compartments in plant exocyst mutants is in accordance with studies in animal and yeast cells. In Drosophila, for example, mutations in any of exocyst subunits Sec5, Sec6 or Sec15 provoke the accumulation of DE-Cadherin in enlarged Rab11-recycling endosomes (Langevin et al., 2005). In Saccharomyces cerevisiae, an accumulation of vesicles was also observed in mutants in exocyst subunits (Novick et al., 1980; Guo et al., 1999).

A pronounced delay in PIN1:GFP and PIN2:GFP recycling to the PM was observed after transient BFA treatment in exo70A1 and sec8 mutants. To confirm that the observed defect in recycling is not an artefact of using GFP-based markers, we performed the immunofluorescence labelling of PIN1 and PIN2 and FM4-64 dye staining in the same experimental design. Both experiments confirmed the longer persistence of BFA compartments in exo70A1 mutants compared with WT plants. As similar recycling defects were recorded for BRI1:GFP in exo70A1 plants, we conclude that the EXO70A1 exocyst subunit is possibly involved in the recycling of at least some PM proteins, similarly to mammalian Exo70, as mentioned above. In mammals and yeast, indeed, the exocyst complex participates in the targeting of transporters and receptors to the PM. In mouse adipocytes, the tethering of secretory vesicles carrying the GLUT4 glucose transporter is regulated by the interaction of the Exo70 exocyst subunit with the Rho GTPase TC10 (Inoue et al., 2003). The involvement of exocyst subunits in cell polarization was also described in developing neurons, demonstrating the role of the rat TC10-Exo70 complex in IGF-1 receptor polarization and membrane expansion (Dupraz et al., 2009). In Drosophila embryo polarization, epithelial polarity is established by the delivery of vesicles carrying the Crumbs protein, a key determinant of epithelial identity, to the apical PM in a process regulated by the EXO84 exocyst subunit (Blankenship et al., 2007).

Prolonged BFA treatment typically induces PIN transcytosis in Arabidopsis, i.e. the relocalization of PINs from the basal to the apical PM domain (Kleine-Vehn et al., 2008). However, the transcytotic relocalization of PINs also occurs under physiological conditions, as shown by a massive redirection of PIN2 from the apical and basal cell sides of a newly formed cell plate to strictly apical or basal sides (Kleine-Vehn et al., 2008). The considerably impaired relocalization of PIN proteins demonstrated by the persistence of both PIN1:GFP and PIN2:GFP in BFA compartments during prolonged BFA treatment in exo70A1 and sec8 mutants was the most striking phenotype observed on the cellular level. A transcytotic defect was further confirmed using PIN1 and PIN2 antibodies and FM4-64 dye staining, which showed that GFP-tagged PINs behave the same way as the endogenous PINs, and that the recycling defect observed in our study does not originate from a methodological artefact of GFP tagging.

A similar recycling defect was also observed for the BRI1 receptor in exo70A1, indicating that the recycling of some transporters and receptors is an exocyst-dependent process. In animals, the exocyst was also revealed to participate in transcytosis. Oztan et al. (2007) observed localization of the exocyst to the apical and basolateral endocytic compartments, and described the involvement of Sec15A in basolateral to apical transcytosis. Recently, we described the involvement of the exocyst in the maturation of the cell plate (Fendrych et al., 2010), indicating a possibility of exocyst-dependent PIN repolarization on these PM domains after the cytokinesis.

In summary, our results indicate that the exocyst, an ancient eukaryotic module for polarized exocytosis, functions in polar auxin transport processes that contribute to the establishment of polarity, which underlies development in land plants.

Experimental Procedures

Plant material and growth conditions

Mutant Arabidopsis lines exo70A1 (Synek et al., 2006) and sec8-m1 (Cole et al., 2005) have been described previously, as well as lines expressing ProDR5:GUS (Ulmasov et al., 1997), VHA-a1:GFP (Dettmer et al., 2006), YFP:RAB-A5d (Geldner et al., 2009), PIN1:GFP (Benková et al., 2003), PIN2:GFP (Xu and Scheres, 2005) and BRI1:GFP (Russinova et al., 2004). The preparation of Col-0 lines expressing GFP:EXO70A1 and GFP:SEC8 was described in Fendrych et al. (2010). The SEC10 CDS was cloned into a modified pVKH vector using SpeI and SalI sites to yield the YFP:SEC10 fusion, and the construct was used for Agrobacterium-mediated transformation of Col-0 plants.

Arabidopsis seeds were surface sterilized (10 min in 20% commercial bleach and rinsed three times with sterile distilled water; Bochemie, and dispersed onto agar plates with growing medium: half-strength MS salts (Sigma-Aldrich, supplemented with 1% (w/v) sucrose (Fluka,, vitamins, 1.6% (w/v) plant agar (Duchefa,, buffered to pH 5.7. Stratification was performed at 4°C for 3 days. Seedlings were either grown vertically or grown on an inclined agar surface (at a 50° angle from the horizontal) in a climate chamber held at 22°C under long-day conditions (16 h of light per day).

Auxin transport assay and GUS staining

Rootward auxin transport was measured in 7-day-old seedlings, as described in Blakeslee et al. (2007), with the following modifications: a 20-nl droplet of 1 mm IAA ([3H]IAA 25 Ci mmol−1; GE Healthcare, was applied to the root–shoot transition zone. The 1-NOA inhibitor (20 μm; Sigma-Aldrich) was applied 1 h after the initiation of assays by pipetting 1-μl droplets in 0.3% agarose at the root–shoot transition zone. Gravacin (20 μm; Chembridge, was applied after 1 h as 5-μl droplets on the surface of each primary root, with the solution restricted to the chromatography paper strip supporting this section of the root. As roots of light-grown exo70A1 seedlings are on average about 30% shorter than WT roots (Synek et al., 2006), only exo70A1 and WT seedlings with comparable root length were used for the measurements.

Auxin transport was visualized in 7-day-old seedlings expressing ProDR5:GUS similarly as described in Rashotte et al. (2001). Plantlets were treated with an application of 10 μm IAA (Sigma-Aldrich) in a 1% agar cylinder on the root–shoot junction for 3, 6 or 12 h.

For GUS staining, seedlings were vacuum infiltrated with staining solution [50 mm sodium phosphate buffer, pH 7.2, 250 μm K3Fe(CN)6, 250 μm K4Fe(CN)6, 2% Triton-X and 1 mm 5-bromo-4-chloro-3-indolyl-β-d-glucuronic acid (X-GlcA); Duchefa, for 1 h and then incubated at 37°C until proper colour development. Tissue was stepwise cleared (by 30 min) in 20% ethanol, 40% ethanol, a mixture of ethanol : acetic acid (3:1), 40% ethanol and 20% ethanol, and then mounted in 50% glycerol for imaging. To evaluate the staining intensity, we counted the number of pixels for which the signal in the blue channel was stronger than a threshold value. To test for significant differences between WT and mutant plants, we first used the Shapiro–Wilk test for normality and subsequently applied the Student's t-test.

Starch staining

Five-day-old seedlings were stained for 15 min with Lugol's solution, followed by a 10-min wash. Images of root tips were analysed with imagej: the intensity of the signal was evaluated as the number of pixels of brightness darker than 45 on the 256-degree grey scale. To test for a significant difference between WT and mutant plants, we first used the Shapiro–Wilk test for normality and subsequently applied the Student's t-test.

Drug treatments

Seedlings were transferred from agar plates into six-well cell culture plates containing half-strength MS liquid medium supplemented with 50 μm BFA (50 mm stock in DMSO; Sigma-Aldrich) and incubated under light conditions in a climate chamber for 2 h. DMSO in the same final concentration (0.1%) was added to negative controls. The washout of BFA was performed by the transfer of seedlings to medium without the drug. Seedlings were labelled by FM4-64 (Invitrogen, at a final concentration of 5 μm 30 min before observation.


After fixing 4-day-old seedlings in 4% paraformaldehyde for 2.5 h under low pressure we followed the procedure described by Friml et al. (2003), employing an InsituPro roboter, except we used an MTSB buffer (50 mm PIPES, 2 mm EGTA, 2 mm MgSO4, pH 6.9) for the whole procedure instead of PBS. Both the goat anti-PIN1 antibody (Santa Cruz Biotechnology, and the chicken anti-PIN2 antibody (Agrisera, were diluted 1:300. Secondary antibodies anti-goat Alexa Fluor 488 (Invitrogen) and anti-chicken Alexa Fluor 555 (Invitrogen) were diluted 1:600.


Seedlings expressing ProDR5:GUS or stained with Lugol's solution (Sigma-Aldrich) were documented using an Olympus BX-51 microscope with an Olympus DP50 camera attached (Olympus,

Seedlings expressing fluorescently labelled proteins and immunolabelled seedlings were observed using a Zeiss LSM 5 DUO confocal laser scanning microscope equipped with a Zeiss C-Apochromat 40 × /1.2 water-corrected objective (Zeiss, For live imaging, we used a chambered coverglass Lab-Tek II (Thermo Scientific, in which seedlings in half-strength MS liquid medium were fixed under a block of agar.


We gratefully acknowledge: Jiří Friml and Steffen Vanneste (Flanders Institute for Biotechnology, Ghent), for their help with immunolocalization and functional interaction of EXO70A1 with PINs; Markus Geisler, for valuable discussions on both the data and the article; Rex Cole, for the preparation of the sec8 mutant expressing PIN1:GFP; and Marek Eliáš (Charles University, Prague) for the preparation of the YFP:SEC10 construct. This work was supported by grants from the Czech Science Foundation (GACR P501/11/P853 and P305/11/1629), the Ministry of Education, Youth and Sports of the Czech Republic (KONTAKT ME10033, MSMT LC06034, MSM0021620858), the Grant Agency of the Academy of Sciences of the Czech Republic (KJB600380802), and from the US National Science Foundation to A.S.M. (IOS 0820648).