Arabidopsis R-SNARE Proteins VAMP721 and VAMP722 Are Required for Cell Plate Formation

vamp721vamp722 mutations result in seedling lethality

No obvious differences in plant growth were observed among single vamp721, vamp722 mutants, heterozygous double mutants and wild-type control plants (Figure S1). Interestingly, we identified vamp721vamp722 double homozygous mutant seedlings in the progeny of self-fertilized heterozygous double mutant plants based on genotyping and RT-PCR. (Figure 1A–1C and Figure S1). As shown in Figure 1C and 1F, the vamp721vamp722 seedlings arrested 2 d after germination and grew extreme rudimentary roots, hypocotyls, and cotyledons, thus leading to seedling death 10 days later. Moreover, compared with wild-type seedlings, vamp721vamp722 mutant roots exhibited disorganized root tips including abnormal meristematic cells and root caps (Figure 1D and 1E). The incorporation of GFP-VAMP721 translational fusion gene under control of the native promoter into vamp721^-/- vamp722^+/- plants fully rescued the seedling lethality phenotype in the double homozygous mutant, confirming that the phenotypic alterations were due to the T-DNA insertions in the two genes (Figure S1).

vamp721vamp722 mutant seedlings exhibit cell wall stubs and incomplete cytokinesis

To gain insights into VAMP721 and VAMP722 function, the root longitudinal sections were prepared and compared. As a result, vamp721vamp722 mutant seedlings exhibited disordered cell file alignment with cell wall stubs or gaps in comparison with wild-type plants (Figure 2A and 2B), implying that the cell division patterns were severely affected. When root tips were stained with Calcofluor and propidium iodide for cell walls and nuclei respectively, 97.8% of root cells in wild-type seedlings displayed normal cytokinesis characterized by one nucleus within a single cell and complete cell walls (Figure 2C and Table S2). However, the root cells of vamp721vamp722 seedlings displayed a high incidence of binucleate cells and ruptured cell walls with the frequencies of 34.2% and 19.0%, as opposed to 1.3% and 0.9% in the wild-type plants (Figure 2D and Table S2). Similar to the wild type, 96.5% of root cells in complemented double mutant showed normal cytokinesis (Table S2). Furthermore, cell wall stubs were frequently observed in the cotyledon epidermal cells of vamp721vamp722 mutants compared to wild type seedlings (Figure 2E–2G).

To determine whether loss function of VAMP721 and VAMP722 affected cell plate expansion, we traced the process of cell plate formation using time-lapse analysis by staining root tips of wild-type and vamp721vamp722 mutant seedlings with FM4-64. As shown in Figure 2H, membrane structures in root cells stained with FM4-64 surrounded the cell-division plane prior to cell plate emergence in wild-type seedlings. After 6 min, a cell plate labeled with FM4-64 was apparently observed at the division plane. During the membrane fusion process, two edges of the cell plate continually expanded outward and finally anchored to parental membrane at 18 min. In contrast, root cells of vamp721vamp722 seedlings exhibited asymmetric initiation of the cell plate at the beginning. Cell plate expansion was not detected after increasing the FM4-64 staining duration. Even after 25 min, cell plates in vamp721vamp722 root cells slightly or rarely expanded (Figure 2I).

VAMP721 and VAMP722 are localized to the cell plate during cytokinesis

To examine the subcellular localization of VAMP721 and VAMP722, we created transgenic plants in which N-terminal GFP or mCherry tagged VAMP721 and VAMP722 translational fusions were expressed respectively under the control of their native promoters in wild type (Col-0) background. Under the confocal microscope, we found that both GFP-VAMP721 and GFP-VAMP722 obviously appeared at expanding cell plates and postcytokinetic walls, which were merged with FM4-64-labeled membranes in root meristematic cells (Figure 3A and 3C). We also found a similar labeling pattern in cotyledon epidermal cells (Figure S2). To get deep insight into the GFP-VAMP721 and GFP-VAMP722 cytokinetic localization, the root meristematic cells of 3- to 5-d old Arabidopsis seedlings were monitored to visualize the process of cell plate formation. During the onset of cytokinesis, the GFP-VAMP721 signal was strongly associated with the newly formed cell plate at the middle of the division plane, and the organelles labeled with GFP-VAMP721 surrounded the cell plate (Figure 3B). Along with the process of cytokinesis, GFP-VAMP721 substantially accumulated at the entire cell plate and two edges of the cell plate expanded symmetrically outward before anchoring with the mother wall. After complete fusion of both cell plate edges with the parental PM, GFP-VAMP721 still remained associated with the postcytokinetic wall, and its intensity gradually decreased to the level of the surrounding cells after 30 min (Figure 3B).

As expected, GFP-tagged VAMP722 displayed a similar cytokinetic localization pattern to GFP-VAMP721. As shown in Figure 3D, the GFP-VAMP722 signal was apparently associated with the emerging cell plate at the division plane soon after the onset of cytokinesis. Membrane organelles labeled with GFP-VAMP722 were distributed in the vicinity of the cell plate during the process of cytokinesis. Two edges of the cell plate attached simultaneously to the parental PM at 16 min. Soon after that, the cell plate completed fusion with the plasma membrane, and the GFP signal decreased and appeared evenly along the entire cell plate (Figure 3D).

The complemented double mutant reestablishes the proper cytokinesis

We further quantified the phenotypes of cell plate formation in cytokinetic root tip cells of control, vamp721vamp722 and complemented double mutant seedlings by the fluorescent labeling (Figure 4 and Table S3). As expected, colocalization at the cell plates and postcytokinetic walls was observed in the root cytokinetic cells co-expressing cytokinesis-specific marker GFP-KNOLLE and mCherry-VAMP721 (Figure S3). Therefore, we used the GFP-KNOLLE transgenic lines as the control. The results showed that 96.9% of cell plates in control cells exhibited symmetric assembly or complete formation. However, abnormal and asymmetric cell plate assemblies were scored highly in 41.8% and 10.9% of vamp721vamp722 cytokinetic root cells labeled with GFP-KNOLLE, consistent with the observation that the double mutant frequently showed disordered cell file alignment and cell wall stubs. Similar to the control, 94.3% of the mitotic root cells in complemented double mutant showed normal cell plate formation monitored by GFP-VAMP721.

VAMP721 and VAMP722 define the trans-Golgi/early endosomal compartments

Apart from the cell plate localization, we also found that GFP-VAMP721 and GFP-VAMP722 were frequently distributed to the cytoplasmic endosomes (Figure S4). To define the endosomes labeled with GFP-VAMP721 and GFP-VAMP722, we first use the trafficking inhibitor Brefeldin A (BFA), a fungal macrocyclic lactone, which targets (guanine nucleotide exchange factors for ARF GTPases) ARF-GEFs, thus inhibiting the function of ARF GTPases and induces heterogeneous aggregations (so called BFA compartments) of early endosomal membranes in plants [19], [20], [21]. When root tips were treated with 50 µM BFA plus FM4-64, the endosomes labeled with GFP-VAMP721 were induced to form aggregates that colocalized with positive BFA compartments labeled with endocytic FM4-64 (Figure 5A). Similar results were obtained when GFP-VAMP722-labeled root tip cells were treated with BFA (Figure 5B). We further found that BFA treatment induced membrane aggregates labeled with VHA-a1-GFP, a TGN/early endosome marker (Figure 5C). However, the Golgi marker N-ST-YFP did not enter the BFA compartments, but instead surrounded FM4-64-enriched BFA compartments (Figure 5D).

To determine whether the endosomes labeled with GFP-VAMP721 and GFP-VAMP722 belong to the prevacuolar compartment (PVC)/late endosome (LE), we further used wortmannin, an inhibitor of phosphatidylinositol-3 kinase, which targets the PVC/LE that then dilates and blocks the traffic to vacuole in plants [22], [23]. After treatment with wortmannin at 33 µm for 60 min in the transgenic root tip cells, we observed that the PVC marker GFP-RabF2b was induced to form small vacuoles, which are the representative wortmannin treatment structures (Figure 5E). In contrast, wortmannin treatment did not cause visible changes of the organelles labeled with GFP-VAMP721 and GFP-VAMP722 in size or number, similar to the results in DMSO control (Figure 5F and 5G). In the root cells co-expressing mCherry-tagged VAMP721 and fluorescence marker of Golgi, mCherry-VAMP721-labeled organelles were in physical proximity with the Golgi marked with N-ST-YFP (Figure 5H). For root cells co-labeled with mCherry-VAMP721 and the PVC marker GFP-RabF2b, we found that the organelles labeled with mCherry-VAMP721 were often transiently close to, but distinct from the PVC (Figure 5I). Similarly, the organelles labeled with mCherry-VAMP722 were distinct from the Golgi apparatus and PVC markers (Figure 5J and 5K).

In Arabidopsis, lipophilic styryl dye FM4-64 is internalized from plasma membrane to lytic vacuole within 1 to 2 h via passing through a variety of endosomes along the endocytosis [24], [25], [26]. When FM4-64 was applied to root tips expressing the early endosome marker VHA-a1-GFP, extensive colocalization was observed in epidermal cells after uptake for 6 min (Figure 6A). In contrast, the internalized FM4-64 did not colocalize with PVC labeled with GFP-RabF2b after 6 min, although they were adjacent to each other (Figure 6B). Even after 15 min, GFP-RabF2b-labeled PVC showed very limited colocalization with the internalized FM4-64 (Figure 6C). However, internalized FM4-64 colocalized mostly with transgenes-labeled endosomes after 6 min in cells expressing GFP-VAMP721 or GFP-VAMP722, similar to the labeling pattern of the VHA-a1-GFP compartment (Figure 6D and 6E). To unveil the spatial relationship between the VAMP721/VAMP722 and VHA-a1 compartments, we crossed plants expressing mCherry-VAMP721 or mCherry-VAMP722 with the VHA-a1-GFP lines. Under the confocal microscope, we observed that mCherry-VAMP721 and VHA-a1-GFP exhibited overlapping membrane distributions (Figure 6F). Similarly, fluorescence signals from mCherry-VAMP722 were colocalized with those of VHA-a1-GFP (Figure 6G).

Inhibiting traffic at the trans-Golgi network affects cell plate formation

ConcanamycinA (ConcA) is a membrane-permeable macrolide antibiotics that binds to the V-ATPase subunits c and inhibits proton transport [27]. It has been demonstrated that ConcA blocks trafficking at the TGN, which affects cell plate formation in Arabidopsis [21], [28], [29]. When ConcA was applied to root tip cells, the organelles labeled with GFP-KNOLLE were clearly trapped in cytosolic aggregates (Figure 7B and Figure S5). As expected, the ConcA treatment retarded the expansion of the cell plate labeled with GFP-KNOLLE, leading to cell wall stub, as shown in Figure 7B. Similarly, we found that ConcA treatment induced a massive intracellular accumulation of the endosomes labeled with GFP-VAMP721 and GFP-VAMP722 (Figure 7D, 7F and Figure S5). Furthermore, cell wall stubs with irregular direction and thickness surrounded with swollen structures were observed in ConcA-treated root dividing cells expressing GFP-VAMP721 and GFP-VAMP722 (Figure 7D and 7F). However, these fluorescent signals were localized to the normal expanding cell plates in untreated dividing cells (Figure 7A, 7C and 7E).

VAMP721 and VAMP722 are required for secretory trafficking to the plasma membrane

The role of de novo secretory trafficking in plant cytokinesis has been emphasized [28]. The occurrence of incomplete cell plate in vamp721vamp722 seedlings and their plasma membrane localization imply that VAMP721 and VAMP722 probably are involved in the secretory trafficking to the plasma membrae. In vamp721vamp722 mutant seedlings expressing plasma membrane marker protein GFP-Lti6a, we observed an abnormal accumulation of GFP signals in the cytoplasm of root epidermal cells (Figure 8B). Even at a higher resolution, we did not detect colocalization between GFP-Lti6a and FM4-64 staining at the plasma membrane (Figure 8B). Similarly, we observed GFP signals inside aberrant intracellular compartments in vamp721vamp722 roots expressing another PM marker, PIP2A-GFP (Figure 8D). In contrast, GFP-Lti6a and PIP2A-GFP showed clear PM localization overlapping with the FM4-64 staining at low or high magnification in control plants (Figure 8A and 8C). However, the cells expressing TIP1;1-GFP in vamp721vamp722 mutant displayed similar tonoplast labeling patterns to that of controls in roots and hypocotyls (Figure 8E–8H).

Plant materials and growth conditions

All plants used for experiments were Arabidopsis Col-0.The homozygous vamp721 (At1g04750), vamp722 (At2g33120) mutants, SALK_037273, SALK_119149 and vamp721^+/-vamp722^-/-, vamp721^-/- vamp722^+/- heterozygous double mutants were kindly provided by Paul Schulze-Lefert [30]. vamp721vamp722 double mutants were isolated from progeny of heterozygous double mutants. Primers used for mutants genotyping are listed in Table S1. Plants expressing N-ST-YFP, VHA-a1-GFP, GFP-RABF2b, GFP-KNOLLE and GFP-LTi6a have been described previously [20], [21], [28], [36], [44]. For vamp721vamp722 mutant seedlings expressing GFP-LTi6a, GFP-KNOLLE, F2 lines derived from crosses between GFP-LTi6a or GFP-KNOLLE lines and vamp721^+/-vamp722^-/- plants were genotyped to isolate the heterozygous double mutants containing the marker protein. For vamp721vamp722 mutant seedlings expressing PIP2A-GFP and TIP1;1-GFP, the plasmids PIP2A-GFP(CD3-1003) and TIP1;1-GFP (CD3-971) [45] were used to transform the heterozygous double mutants. Wild-type looking seedlings from the mother plants were used as the controls.

For growing seedlings on agar-containing plates, Arabidopsis seeds were pretreated in 70% ethanol for 1 min, surface-sterilized in 2.5% bleach for 10 min, and washed with distillated water at least five times. Seeds were planted on 1% agar-containing 0.5x Murashige and Skoog Salts, 1% sucrose, pH 5.8, stratified and placed at 4°C in the dark for 2 d before germination. Growth conditions were at 23°C with a 16-h-light/8-h-dark cycle, either in soil or on MS plates.

Constructs and plant transformation

For fluorescent fusion protein constructions, 1.8kb VAMP721 promoter and 2.0kb VAMP722 promoter before the start codon of each gene were amplified from genomic DNA of wild-type Arabidopsis thaliana ecotype Columbia plants and cloned into the pCAMBIA1300 binary expression vector with HindIII and SalI respectively. To create the translational fusions of VAMP721 or VAMP722 tagged with GFP or mCherry, a cloning vector pUC18/pCAMBIA1300-GFP-AtFim1ABD2 [46] was used. GFP sequence was replaced with cDNA encoding mCherry. The genomic sequences of VAMP721 and VAMP722 were PCR amplified and subcloned into the cloning vectors with SpeI and NotI replacing the ABD2 fragment. Then the resulted cloning vectors were digested with SalI and EcoRI and the SalI-EcoRI fragments including GFP/mCherry-VAMP721/VAMP722 were cloned into the pCAMBIA1300 under VAMP721 or VAMP722 promoter. Primer sequences used for the constructs are listed in Table S1. All The sequences cloned above were checked by sequencing. The binary vectors were transformed into A. tumefaciens strain GV3101. Transformation of Arabidopsis plants was performed by floral dipping using Agrobacterium tumefaciens [47]. The selection of transgenic lines was performed on 1/2 MS solid medium containing 3% sucrose with 25ug/ml hygromycin. For complementation the resulting A.tumefaciens transformant pVAMP721::GFP-VAMP721 was used to transform vamp721^-/- vamp722^+/- plants. The homozygous identity of T-DNA insertion of the rescued plants was confirmed by PCR assay in the T2 plants.

RT-PCR analysis

Total RNA was extracted from seedlings of four-day-old wild-type and vamp721vamp722 double mutant using SV Total RNA Isolation Kit (promega), and cDNA was synthesized with SuperScript III reverse transcriptase (Invitrogen). PCR conditions were as follows: 95°C (4 min); 28 cycles of 95°C (30 s), 57°C (30 s), 72°C (1 min), and 72°C (10 min). Primers used for RT-PCT are listed in Table S1.

Cross-section analysis of wild-type and vamp721vamp722 mutant roots

The root tissue from 4-day-old seedling was cut and immediately vacuum infiltrated and fixed with 2.5% (v/v) glutaraldehyde and 2% (v/v) paraformaldehyde in 0.1 M phosphate-buffered saline (pH 7.2) at room temperature for 4 h, followed by postfixation in 1%OsO₄ buffer at 4°C overnight. Samples were subsequently rinsed with 0.1 M phosphate buffer and dehydrated through a graded ethanol series (30%–100%). Then, samples were embedded in LR White (EMS) and polymerized at 60°C for 24 h. Semithin (1 µm) sections were cut using an Ultracut microtome (EM UC6; Leica). Semthin sections were stained with toluidine blue O before observation.

Fluorescent dye and inhibitor treatments

For FM4-64 staining, seedlings were incubated in half-strength MS liquid containing 5 µM FM4-64 (Invitrogen, diluted from a 5 mM stock in water) for a specified time at room temperature. To stain cell walls and nuclei simultaneously, Calcofluor and propidium iodide were used as described [25]. For drug treatments, three- to five-day-old seedlings were incubated in 1ml of liquid medium (half-strength MS medium) containing 50 µM brefeldin A (BFA), 33 µM wortmannin or 2 µM concanamycin A. The seedlings were incubated with inhibitors at room temperature for the indicated times before observation. Control treatments were performed with equal amounts of the DMSO. The following stock solutions were used: 50 mM BFA (Sigma-Aldrich) in DMSO; 33 mM wortmannin (Sigma-Aldrich) in DMSO; 2 mM concanamycin A (Sigma-Aldrich) in DMSO. Each treatment for confocal imaging was repeated at least three times with similar results.

Confocal microscopy

For confocal analysis, seedlings mounted in half-strength MS liquid were analyzed with an upright Zeiss LSM 510 laser scanning microscope equipped with a META device. GFP or YPF was visualized by excitation with an Argon laser at 488 nm and detected with a 505- to 550-nm emission filter. For imaging of GFP/FM4-64, YFP/FM4-64, or GFP/chlorophyll, the signals were excited with an Argon laser at 488nm and detected with a spectral detector set BP 500-550 IR for green signal and LP 560 for red signal. Co-localization analyses were performed on F1 or F2 hybrid seedlings co-expressing GFP- and mCherry-tagged proteins under Zeiss LSM 5 LIVE. GFP and mCherry were excited with a 488-nm and 561-nm laser respectively (multitrack mode). The fluorescence emission was detected with spectral detector set BP 520-555 (GFP) and LP 575 (mCherry). To image propidium iodide and Calcofluor simultaneously by Zeiss LSM 5 LIVE, the parameter set was used as described [25]. Images were edited using ImageJ software (http://rsb.info.nih.gov/ij/) and Adobe Photoshop CS2.