Introduction

The shuttling of vesicles within different membrane-delimited compartments is a common feature in eukaryotes, which maintains the endomembrane system and delivers proteins to their sites of action in the cell. Generally, the vesicle transport can be divided into four essential steps including vesicle budding, transport, tethering, and fusion (Cai et al. 2007). These steps are tightly regulated by many critical proteins, such as coat proteins, motor proteins, and tethering factors (Cai et al. 2007). The specifity of vesicle targeting and correction of cargo delivery are required for the various physiologic processes during plant growth and development (Pratelli et al. 2004). An understanding of the importance of vesicle traffic in plants is rapidly developing, but its molecular mechanism is mostly unknown.

The fusion of vesicles with specific target compartment, a process mediated by the SNARE (soluble N-ethylmaleimide-sensitive factor attachment protein receptor) complex, is the final step in vesicle-mediated transport (Jahn and Scheller 2006). The SNARE protein family is highly conserved in eukaryotic cells, and all SNARE proteins are characterized by the coiled-coil helices, named SNARE motif, adjacent to a transmembrane domain (Sollner et al. 1993). Based on the conserved amino acids in SNARE motif, SNARE proteins can be classified into four groups: (1) Qa-SNAREs, (2) Qb and Qc-SNAREs, (3) SNAP25-like containing Qb and Qc tandem SNARE motifs and (4) R-SNAREs (Fasshauer et al. 1998). Q-SNAREs, also called t-SNAREs, are localized to target organelle membranes, while R-SNARE, also called v-SNAREs, is localized to vesicle membranes. One set of Qa-, Qb-, Qc- and R-SNAREs constitutes a functional SNARE complex with a four-helical bundle assembled by four SNARE motifs (Hong 2005). The correct combinations of cognate SNAREs drive specific membrane fusions.

R-SNAREs can be further subdivided into short VAMPs (vesicle-associated membrane proteins) or brevins and long VAMPs or longins within the animal and fungi lineage (Rossi et al. 2004). In plants, only longin type R-SNAREs exist (Sanderfoot 2007). The Arabidopsis genome encodes 15 R-SNAREs including two Sec22-like, two Ykt6-like and 11 VAMP7-like longin R-SNAREs (Lipka et al. 2007). The VAMP7-like proteins in higher plants consist of two major groups: VAMP71 and VAMP72 groups (Sanderfoot 2007). The VAMP72 group appears to be specific to the green plant lineage and likely represents the R-SNARE components for secretion (Sanderfoot 2007). Recently, the significance of R-SNAREs in plant growth has received considerable attention.

Previously it has been reported that Arabidopsis R-SNAREs VAMP721 and VAMP722 are involved in immune responses to pathogens (Kwon et al. 2008). Our previous work showed that loss function of VAMP721 and VAMP722 resulted in a dwarf seedling phenotype characterized with rudimentary roots, cotyledons and hypocotyls (Zhang et al. 2011). Recently, Yi et al. (2013) found that ABA treatment enhanced growth inhibition and early depletion of the amount of VAMP721/722 protein level in haploinsufficient VAMP721 +/− VAMP722 −/− and VAMP721 −/− VAMP722 +/− plants (Yi et al. 2013). In this study, we further investigated the function of VAMP721 and VAMP722 in plant growth by analyzing mutant phenotypes. Our results suggest that VAMP721 and VAMP722 are essential for embryo development and normal cell growth in seedling roots.

Materials and methods

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 described previously (Kwon et al. 2008). According to the aberrant seedling morphology, vamp721vamp722 double mutant seedlings were isolated from progeny of either heterozygous double mutants (Zhang et al. 2011). Plants expressing GFP-VAMP721 and GFP-VAMP722 (Zhang et al. 2011) were used for confocal analysis. For growing seedlings on agar-containing plates, Arabidopsis seeds were pretreated with 70 % ethanol for 1 min, surface-sterilized in 2.5 % bleach for 10 min, and washed with distilled water. The seeds were planted on 1/2 MS medium (Sigma) supplemented with 1 % (w/v) sucrose, 1 % (w/v) agar (pH 5.8), and placed at 4 °C in the dark for 48 h before germination. Growth conditions were at 23 °C with a 16-h-light/8-h-dark cycle.

Progeny analysis and genotyping

For the progeny analysis of self-pollinated heterozygous double mutant plants, the Arabidopsis seeds were planted horizontally on 1/2 MS solid medium. The vamp721vamp722 seedlings were counted based on its obviously abnormal morphology within 7 days after germination. The remaining seedlings were identified after 25-days growth in medium. For the progeny analysis of reciprocal crosses, 25-days-old seedlings cultured on 1/2 MS solid medium were used for the genotyping. The transmission efficiency (TE) = (observed number of mutant alleles/observed number of wild-type alleles) × 100, as described previously (Ebel et al. 2004). The genotypes of T-DNA insertion lines were determined by a PCR-based method. The following primers were used: LBb1 (5′-GCGTGGACCGCTTGCTGCAACTCTC-3′), LP1 (5′-CCCCCGTCCATTAAGAATTAAG-3′), RP1 (5′-ATTGAGGACAGAAAGGGTCAGATTC-3′) for vamp721, and LBb1 (5′-GCGTGGACCGCTTGCTGCAACTCTC-3′), LP2 (5′-CTCTGAGATCGGTCCCGTAAAATCGG-3′), RP2 (5′-AACTATGCCCATGAATCATAGAC-3′) for vamp722.

Complementation of the vamp721vamp722 double mutant

For rescuing the vamp721vamp722 mutant, 1.8 kb VAMP721 promoter and 2.0 kb 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. The genomic sequence of VAMP721 plus 1.0 kb 3′UTR was PCR amplified and cloned into the pCAMBIA1300 with SalI and KpnI under VAMP721 promoter. The genomic sequence of VAMP722 plus 1.0 kb 3′UTR was PCR amplified and cloned into the pCAMBIA1300 with BamHI and EcoRI under VAMP722 promoter. All The sequences cloned above were checked by sequencing. The rescue constructs were transformed into Arabidopsis tumefaciens strain GV3101. The resulting A. tumefaciens transformants were used to transform vamp721 −/− vamp722 +/− and vamp721 +/− vamp722 −/− plants, respectively using the floral dip method (Clough and Bent 1998). The selection of transgenic lines was performed on 1/2 MS solid medium containing 3 % sucrose with 25 μg/ml hygromycin. The homozygous identity of T-DNA insertion of the rescued plants was confirmed by PCR assay in the T2 plants.

Clearing of Arabidopsis embryos

For clearing of Arabidopsis embryos, growing siliques were harvested from soil-grown plants and dissected under a stereo-microscope. Ovules from individual siliques were collected and fixed for 1–4 h in ethanol/acetic acid (6:1) at room temperature. Then, ovules were washed three times for 5 min in 100 % ethanol and one time in 70 % ethanol. In turn, the ovules were incubated in a clearing solution (chloralhydrate/glycerol/water 8:1:2 v/v) for 24 h, mounted on slide with 30 % glycerol, and observed by Olymplus BX51 with Nomarsky Differential Interference Contrast (DIC) optics.

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 mm) sections were cut using an ultracut microtome (EM UC6; Leica). Semthin sections were stained with toluidine blue O before observation.

Confocal microscopy

For confocal analysis, seedlings mounted in half-strength MS liquid were analyzed with Leica SP5 confocal microscope. For imaging of GFP, the signals were visualized by excitation with an Argon laser at 488 nm and detected with a 500- to 550-nm emission filter. The images were edited using Image J software and Adobe Photoshop CS2.

Statistical analysis

The segregation ratios of progeny genotypes recovered from self-pollinated heterozygous double mutant plants were analyzed with χ2 test using SPSS software.

Results

vamp721 shows synthetic lethality with vamp722

We previously identified the vamp721vamp722 homozygous double mutant seedlings from the progeny of heterozygous double mutants cultured on the 1/2 MS solid medium (Zhang et al. 2011). In this study, we analyzed the progeny segregation in detail. The results showed that the percentage of vamp721vamp722 seedlings segregated from VAMP721 −/− VAMP722 +/− mutants was 17 %, less than the expected 25 %. Similar results were obtained from VAMP721 +/− VAMP722 −/− mutants (Table 1). Moreover, the segregation ratios of progeny genotypes recovered from the two parental lines VAMP721 −/− VAMP722 +/− and VAMP721 +/− VAMP722 −/− indicated a deviation from the expected 1:2 segregation, but this was statistically significant only for the latter (Table S1), suggesting that the gametophytic activity was disturbed at some extent.

Table 1 Progeny analysis of self-pollinated Arabidopsis seedlings segregating the vamp721vamp722 mutants
Full size table

The analysis of gametophyte transmission

To determine whether VAMP721 and VAMP722 play a role in gametophyte development, we analyzed the gametophyte transmission between heterozygous double mutant and wild-type plants (Table 2). The results showed that 86.3 and 87.8 % of both vamp721 and vamp722 alleles were transmitted via the female gametophytes of VAMP721 −/− VAMP722 +/− and VAMP721 +/− VAMP722 −/−, respectively. When double heterozygous mutant plants were used as the pollen donors, 89.9 % of vamp721vamp722 gametes were inherited from the male parental line VAMP721 −/− VAMP722 +/−. While, 53.9 % of the double mutant gametes were transmitted for parent VAMP721 +/− VAMP722 −/−, indicating that the TE of both mutant alleles was reduced through male gametophytes of VAMP721 +/− VAMP722 −/−.

Table 2 Transmission efficiency (TE) of vamp721vamp722 allele in reciprocal crosses
Full size table

The vamp721vamp722 mutant shows embryo development defects

The segregation ratio of vamp721vamp722 seedlings derived from the progenies of self-pollinated VAMP721 −/− VAMP722 +/− plants suggested that the double mutant was partially embryonic lethal; thus, we examined the premature seeds in the seedpods of VAMP721 −/− VAMP722 +/− plants. The results showed that some of the seeds from VAMP721 −/− VAMP722 +/− plants exhibited a yellowish appearance (green:yellow = 629:167); however, almost all the seeds from single mutant vamp721, vamp722, or wild-type plants showed normal appearance (Fig. S1). We also observed the abnormal appearance of seeds from VAMP721 +/− VAMP722 −/− plants (Fig. S2).

We then examined the embryogenesis in more detail, observing embryos in cleared seedpods from the VAMP721 −/− VAMP722 +/− mutant plants at different development stages. We observed the development of control embryos including globular, heart, torpedo, early cotyledon, and late cotyledon stage (Fig. 1a–e). The examination of yellowish seeds in mutant siliques showed that abnormal embryos were observed from late heart to cotyledon stage. As shown in Fig. 1f, one embryo at late heart stage grew two asymmetric developing cotyledons. Similar appearance was observed at the torpedo stage embryo which also exhibited increased lateral proliferation of root compared with the elongated root in control (Fig. 1g). Figure 2h showed that one torpedo stage embryo of mutant plants has three developing cotyledons instead of two in the control. Moreover, cotyledon-stage embryos from mutant plants developed rudimentary cotyledons and roots, and did not fold, or only partially folded, compared to the fully folded embryonic root and cotyledons in the control (Fig. 1i, j). Interestingly, the development of part embryos was arrested at globular-like stage (Fig. 1k), while the control embryos from the same seedpods were at cotyledon stage. We verified that these abnormal embryos were the vamp721vamp722 double homozygous mutants by PCR-based genotyping (Fig. S3).

Fig. 1
figure 1

Embryo development of vamp721vamp722 double mutant. (a–e) The wild type-looking (control) embryos at globular (a), heart (b), torpedo (c), and cotyledon stage (d, e) dissected from seedpods of VAMP721 +/− VAMP722 −/− plants. (f–j) vamp721vamp722 double mutant embryos exhibited abnormal morphology at late heart (f arrowheads indicate asymmetric developing cotyledons), torpedo (g, h arrowheads indicate asymmetric and three developing cotyledons, respectively), and cotyledon stage (i, j arrowheads indicate rudimentary cotyledons and roots). k Embryogenesis of part double mutant was arrested at globular-like stage in the mature seedpods. Bars = 50 μm

Full size image
Fig. 2
figure 2

Phenotypic analysis of vamp721vamp722 double mutant roots. a 4-days-old vamp721vamp722 seedlings displayed a dwarf stature compared with wild-type seedlings. b, c The meristem zone in vamp721vamp722 roots (b) was much smaller than in wild type (c). The red arrow indicates the uppermost cell in meristem zone. d, e The longitudinal section in wild-type (d) and double mutant (e) roots. f, g The cross section in wild-type (d) and double mutant (e) roots. Bars = 1 mm in (a); 50 μm in (b–g). (Color figure online)

Full size image

The incorporation of a genomic fragment containing the VAMP721 gene into VAMP721 −/− VAMP722 +/− plants fully rescued the defective embryo phenotype in the double homozygous mutant (Fig. S1). Similarly, the incorporation of a genomic fragment containing the VAMP722 gene into VAMP721 +/− VAMP722 −/− plants also rescued the defective embryo phenotype (Fig. S2).

vamp721vamp722 mutations cause defects in root development

The analysis of embryo development suggested that vamp721vamp722 mutations likely affect postembryonic root growth. In order to confirm the function of VAMP721 and VAMP722 in root development, we analyzed and compared the root cell morphology in wild-type and double mutant seedlings. In contrast to wild-type seedlings, 4-days-old vamp721vamp722 seedlings showed severely retarded growth characterized with stunted cotyledon, hypocotyl and root (Fig. 2a) . The meristem zone in vamp721vamp722 roots was smaller than that in wild-type siblings. Moreover, vamp721vamp722 roots exhibited disorganized columella cells and no obvious transition from elongation to differentiation zone compared with control roots (Fig. 2b, c). In comparison with wild-type roots, the root longitudinal sections of vamp721vamp722 showed disordered cell layer pattern for epidermis, cortex, and stele, which displayed abnormal cell files, such as expanded cells in epidermis and cortex. Particularly, vamp721vamp722 roots failed to specify the quiescent center (QC) cells compared with that in wild-type roots (Fig. 2d, e). The disordered cell pattern and cell file alignment were also observed in vamp721vamp722 root cross sections compared with wild-type roots (Fig. 2f, g).

Early data have shown that VAMP721 and VAMP722 were localized at plasma membrane and unknown organelles in Arabidopsis protoplasts by transient assays (Uemura et al. 2004, 2005). Our previous work confirmed the plasma membrane localization of VAMP721 and VAMP722 in transgenic Arabidopsis seedings and showed that VAMP721 and VAMP722 were localized to trans-Golgi network (TGN) in root cells (Zhang et al. 2011). However, the expression pattern of VAMP721 and VAMP722 in root tissue is still unclear. Thus, we examined the expression of VAMP721 and VAMP722 in root cells using seedlings expressing GFP-VAMP721 or GFP-VAMP722 under control of their native promoters. The results showed that GFP-VAMP721 was expressed in root cap, columella cells, stele cells, epidermis and cortex cells in meristem zone, and cells in elongation zone (Fig. 3a, b). In the differentiation zone, VAMP721 showed strong expression in xylem, root hairs, epidermis and cortex cells (Fig. 3c). The signals of GFP-VAMP722 were also detected in whole root, including meristem, elongation and differentiation zone (Fig. 3d–f), similar to the expression pattern of GFP-VAMP721.

Fig. 3
figure 3

The expression of VAMP721 and VAMP722 in roots. a–c GFP-VAMP721 were expressed in whole root, including root cap, meristem zone (a), elongation zone (b), and differentiation zone (c). d–f The expression of GFP-VAMP722 in root cells. Bars = 20 μm

Full size image

Discussion

Recent studies indicate that SNARE proteins are not restricted to the housekeeping function in membrane fusion, but play important roles in plant shoot morphogenesis (Ohtomo et al. 2005), vascular network formation (Shirakawa et al. 2009), plant defense (Assaad et al. 2004), vacuole biogenesis (Ebine et al. 2008), and abiotic stress responses (Leshem et al. 2006). However, the knowledge about the function of R-SNAREs in plant development is still limited. Based on the sequence information from available Arabidopsis genome assemblies, VAMP721 and VAMP722 are classified into R-SNARE members of the VAMP72 group (Sanderfoot 2007). The T-DNA insertion lines of VAMP721 and VAMP722 and their heterozygous double mutants were indistinguishable from wild-type plants. Moreover, no homozygous double mutant plants can be isolated from the progeny of heterozygous double mutants (Kwon et al. 2008), suggesting that VAMP721 and VAMP722 have redundant as well as essential functions in plant development. We previously identified the vamp721vamp722 double mutant seedlings in 1/2 MS medium (Zhang et al. 2011). In the present study, we analyzed the progeny segregation of heterozygous double mutants and found that VAMP721 −/− VAMP722 +/− and VAMP721 +/− VAMP722 −/− plants exhibited a non-mendelian segregation ratio, but this was statistically significant only for the latter, consistent with other reports (Kwon et al. 2008). Particularly, the percentage of vamp721vamp722 seedlings segregated from either heterozygous double mutant was less than the expected ratio. These results suggest that loss function of VAMP721 and VAMP722 leads to a gametophytic defect and/or portion of embryo lethality.

Assuming random segregation during meiosis and the absence of post-meiotic selection, the genotypes of the gametes should be equally distributed (Howden et al. 1998). In our study, we assessed the TE of double mutant alleles through male and female gametes. Reciprocal crosses between VAMP721 −/− VAMP722 +/− and wild-type plants revealed that vamp721vamp722 gametes could be inherited from the male or female parent, but at a slightly reduced frequency. Moreover, the transmission of vamp721vamp722 allele was partially reduced through the pollen, not the female gametophyte of VAMP721 +/− VAMP722 −/−, suggesting that VAMP721 +/− VAMP722 −/− plants produce partial either non-viable or non-functional pollen. However, the reduced male gametophyte had no obvious effects on the fertilization of ovules. Recent studies indicated that Arabidopsis R-SNARE SEC22 plays an essential role in vesicle trafficking between ER and Golgi, and sec22 mutation highly reduced male and female gametophyte transmission (El-Kasmi et al. 2011), suggesting that early secretory pathway is required for the gametophyte development. Arabidopsis Qa-SNAREs SYP21 and SYP22 mediate the protein trafficking between pre-vacuolar compartments (PVC) and vacuoles and are essential for male gametophytic viability (Shirakawa et al. 2011). In addition, TGN-localized Qa-SNAREs SYP41 and SYP42, which regulate the secretory and vacuolar transport pathways in the post-Golgi network, also play important roles in pollen function (Sanderfoot et al. 2001; Uemura et al. 2012). We previously found that VAMP721 and VAMP722 were localized at TGN and mediated the secretory trafficking to plasma membrane (Zhang et al. 2011). Based on our results together with recent publications, we speculate that the protein traffic mediated by VAMP721 and VAMP722 is required for the gametophyte activity in Arabidopsis.

The segregation ratio of vamp721vamp722 seedlings implied the possibility of partial embryo development defects. The examination of premature seeds revealed that VAMP721 −/− VAMP722 +/− or VAMP721 +/− VAMP722 −/− plants developed part of seeds which displayed yellowish appearance. The clearing of embryos from these abnormal seeds showed defective embryo development from late heart to cotyledon stage, such as embryos with asymmetric developing cotyledons, expanded roots, three developing cotyledons, and unfolded cotyledons and roots. In addition, the arrested embryo development at globular-like stage was frequently observed from the abnormal seeds, which likely accounts for the reduced segregation ratio of vamp721vamp722 seedlings. We confirmed that these abnormal embryos contained both vamp721 and vamp722 T-DNA insertions by PCR-based genotyping. The complementation of the vamp721vamp722 double mutant rescued the defective embryo phenotype, confirming that the embryo alterations were due to the T-DNA insertions in the VAMP721 and VAMP722 genes. The microarray data from Arabidopsis eFP Browser show that VAMP721 and VAMP722 genes are expressed throughout embryo development, and VAMP721 exhibits a higher expression level than that of VAMP722 (Fig. S4) (Winter et al. 2007), this might suggest that VAMP721 and VAMP722 play independent functions in embryogenesis. It was reported that Arabidopsis v-SNAREs VTI11 and VTI12, two members of VTI1 family, regulated vesicle trafficking to lytic and storage vacuoles, respectively, and loss function of VTI11 and VTI12 caused embryo lethality (Surpin et al. 2003; Sanmartin et al. 2007), indicating that the vacuolar trafficking mediated by SNARE proteins is required for embryo viability. The Arabidopsis cytokinesis-specific Qa-SNARE KNOLLE (KN) is also essential for embryo development, and mutations in the KN gene disrupt the regular pattern of embryogenesis by altering the rate and plane of cell division as well as cell morphology (Lukowitz et al. 1996). Recent studies indicated two distinct types of KNOLLE-containing SNARE complexes appear to jointly medi-ate membrane fusion in Arabidopsis cytokinesis and either complex includes VAMP721 and VAMP722 as the R-SNARE components (El Kasmi et al. 2013). Given that VAMP721 and VAMP722 are involved in the cell plate formation during cytokinesis (Zhang et al. 2011), our results suggest that the membrane fusion mediated by VAMP721 and VAMP722 is required for the cell division and cell morphology in embryo development.

vamp721vamp722 seedlings exhibited a dwarf stature characterized with much smaller organs (Zhang et al. 2011). In this study, we focus on the roles of VAMP721 and VAMP722 in root development. Our results showed that vamp721vamp722 mutations led to severe reduction of meristem zone and mixed elongation and differentiation zone in roots. In addition, disordered cell layer pattern and cell file alignment were observed in root longitudinal and cross sections. Particularly, vamp721vamp722 mutations impaired the QC patterning. Furthermore, VAMP721 and VAMP722 were expressed in all root cells, confirming their essential function in root growth. Our previous work showed that vamp721vamp722 mutant root cells frequently exhibited cell wall stubs or gaps and incomplete cytokinesis characterized with a high incidence of binucleate cells. Moreover, abnormal and asymmetric cell plate assemblies were scored highly in vamp721vamp722 cytokinetic root cells labeled with cell plate-specific marker GFP-KNOLLE (Zhang et al. 2011). Concerning the significant roles of VAMP721 and VAMP722 in cytokinesis (Zhang et al. 2011), it is reasonable to deduce that the rudimentary roots of vamp721vamp722 mutants are likely due to the disruption of cell division activity. Furthermore, VAMP721 and VAMP722 were shown to localize at plasma membrane and TGN compartment in non-dividing cells and were required for the secretory trafficking to plasma membrane, since the GFP-tagged PM proteins aquaporin PIP2a and low temperature responsive protein LTI6A were severely accumulated in the cytoplasm of root cells (Zhang et al. 2011). Our new findings indicate that VAMP721 and VAMP722 are involved in the polar PM localization of auxin transporters (our unpublished data). Aquaporins are ubiquitous channel proteins that facilitate root water uptake (Javot et al. 2003). Auxin transporters are essential for the directional auxin transport and cellular auxin homeostasis which plays important roles in root development, such as root meristem patterning and tropic growth (Rahman et al. 2010; Cazzonelli et al. 2013). Based on the analysis above, we speculate that VAMP721 and VAMP722 probably play a general role in the secretory trafficking of PM proteins, such as channel proteins and transporters which are essential for root growth. Taken together, our results suggest that the protein trafficking mediated by VAMP721 and VAMP722 is essential for root development.

In conclusion, R-SNARE proteins play essential roles in plant growth and development in addition to vesicle trafficking. Our results indicated that R-SNAREs VAMP721 and VAMP722 are required for the gametophyte activity in Arabidopsis. Furthermore, vamp721vamp722 mutations resulted in abnormal embryo and seedling root development. Based on the function of SNARE protein in membrane traffic, we propose that VAMP721 and VAMP722 are involved in plant development through the regulation of protein trafficking related to specific physiologic process.