Divergent functions of VTI12 and VTI11 in trafficking to storage and lytic vacuoles in Arabidopsis

VTI12 Is a SNARE Involved in Trafficking of VAC2.

SNARE proteins are essential to drive membrane fusion in exocytotic and vacuolar trafficking pathways in eukaryotes (21). Several SNARES have been localized to the late secretory pathway in Arabidopsis, and are candidates to regulate membrane fusion events in vacuolar trafficking pathways, and may therefore participate in transport of VAC2. Among these, VTI11, SYP22/SYP21, and SYP51 are members of a SNARE complex (SNvti11complex) found at the prevacuolar compartment (PVC) and the tonoplast that has been proposed to be involved in trafficking to the lytic vacuole (22, 23). VTI12, SYP41/SYP42, and SYP61 are members of another discrete SNARE complex (SNvti12complex) at the TGN (22, 23). No definitive function has been assigned to this complex, although some evidence implicates VTI12 in autophagy (24). VTI11 and VTI12 are homologous to a single yeast gene, Vti1, which is involved in trafficking to vacuoles through all pathways that have been described in that organism (25), supporting a role for their plant counterparts in vacuolar trafficking.

We have previously shown that SYP21, SYP22, SYP41, and SYP42 are essential for pollen development and that homozygous mutants cannot be recovered (16). In contrast, null mutants of VTI11 and VTI12 are viable, so their function in VAC2 trafficking can be tested. The vti11 mutants are agravitropic and show defects in leaf morphology and central lytic vacuole formation (26), whereas vti12 plants show no obvious developmental phenotype under normal growth conditions (24). We crossed single null mutants of vti11 and vti12, and the corresponding wild type (Wt) background ecotype (Col-0), into a marker line expressing VAC2. We used the L1 marker line [see Supporting Information (SI)] that is homozygous for a single insertion of VAC2 in the intergenic region between At3g46110 and At3g46120. This line expresses amounts of VAC2 that are close to the empirically determined saturation levels (SI and data not shown). Therefore, we anticipated that even weak inhibition of trafficking in the mutants would lead to secretion of VAC2 and consequent early termination of meristems. The F1 plants from the three crosses were identical and featured normal meristems. In the F2 populations, we found plants with terminated shoot and floral meristems in the cross with vti12 mutants but not in the crosses with vti11 mutants or Wt plants. Genotyping by PCR demonstrated that all of the plants showing determinate shoot meristem and early termination of floral meristems were homozygous vti12 mutants expressing VAC2 (Figs. 1A and 2). No effect of VAC2 on the shoot or floral meristem was observed in Wt or in vti11 (Figs. 1A and 2). We analyzed the expression of VAC2 to ensure that the phenotype in vti12 plants was not due to enhanced VAC2 levels. We could not detect the VAC2 protein by regular Western blot analysis, most likely because of the in planta processing of CLV3 (27), so we analyzed VAC2 expression by Northern blot. VAC2 expression was similar in all genotypes (SI), suggesting that secretion of VAC2 in vti12 was most likely due to defects in its transport.

To exclude the possibility that trafficking of the chimeric VAC2 construct was not representative of the trafficking of natural storage proteins, we analyzed transport of full-length BL. We crossed vti11 and vti12 plants to a BL-expressing line (6) and compared the accumulation of BL in apoplasmic protein samples from leaves of Wt, vti11, and vti12 plants. As shown in Fig. 1B, vti12 mutants accumulated BL in the extracellular space, whereas no BL was detected in Wt plants, indicating that VTI12 is involved in vacuolar trafficking of both BL and VAC2. We also detected weaker BL secretion in vti11 plants, which is consistent with the observation that in vti11 VTI12 assumes VTI11 functions (27) and thus may be partially depleted from its role in transport of BL.

VTI11 Conditionally Substitutes for VTI12 in VAC2 Trafficking.

Although VTI11 and VTI12 serve independent functions in Wt plants, they can replace each other under certain conditions (24): in vti11 mutants, VTI12 is transferred to the SNvti11 complex and partially substitutes for VTI11 function in the gravitropic response pathway. Conversely, in vti12 null mutants, VTI11 is transferred to the SNvti12 complex. Double vti11/vti12 mutants are embryo lethal, indicating that one or both of their functions are essential for viability of the plant and that they have some redundant roles (24). We predicted that in plants devoid of VTI12, VTI11 would take over its function in transport of VAC2. To test this hypothesis we created an allelic series of plants with various combinations of VTI11 and VTI12 mutant alleles. In plants homozygous for VAC2 there were increasing degrees of meristem termination that correlated with increasing loss of function of the SNvti12 complex. We found four classes of phenotypes that were associated with specific genotypes (Fig. 2): plants that showed no effect on the meristem had two functional copies of VTI12 or had one functional copy of VTI12 and were not null for VTI11 (Class 1); the weakest meristem phenotype, consisting of termination of several flower meristems before the production of carpels (arrowheads, Fig. 2), was observed in vti11 plants that were hemizygous for VTI12 (designated enhanced vti11 plants; Class 2); an intermediate effect was observed in vti12 plants with two functional copies of VTI11(Class 3). In these plants the primary meristems were terminated (Fig. 1A), but the secondary meristems gave rise to inflorescences where only some lateral meristems were terminated (arrows, Fig. 2), and <50% of the flowers had carpels and developed into siliques (arrowheads). The strongest phenotype was observed in vti12 plants hemizygous for VTI11 (designated enhanced vti12 plants), in which most meristems were terminated before elongation, and only a few lateral meristems gave rise to elongating stems with meristems terminated after the emergence of cauline leaves (Class 4). Flowers were seldom produced, and the few that were lacked carpels. Even after 5 months, these plants did not produce fertile flowers (Fig. 2 Inset). These experiments show that VTI12 was involved in VAC2 trafficking, whereas VTI11 was required only when the amount of VTI12 was limiting and VTI11 replaced its functions by transferring to the SNvti12 complex. We conclude from these results that trafficking of VAC2 required the TGN-localized SNvti12 complex, and was independent of the PVC-localized SNvti11, which may instead be involved in central lytic vacuole formation (26).

The Trafficking of Storage and Lytic Vacuole Markers Is Differentially Affected in vti12 and vti11.

To determine whether the results with VAC2 and BL could be generalized to other vacuolar proteins, we studied the localization of tobacco chitinase and barley aleurain, which are markers of PSVs and LVs, respectively. Tobacco chitinase contains a VSS that, when fused to GFP (GFP-CHI), targets the fusion protein to PSVs, whereas the VSS from barley aleurain targets GFP (ALEU-GFP) to LVs (28, 29). These fluorescent reporters can be used to monitor in vivo trafficking through storage and lytic vacuole pathways. We analyzed the localization of the two markers by transient expression in leaf epidermal cells. Over 90% of Wt cells transformed with GFP-CHI showed a diffuse fluorescence throughout most of their cellular volume (Fig. 3), the expected pattern for a protein localized to the vacuole lumen in Arabidopsis epidermal cells, and a weak fluorescence in the ER. Nine percent of transformed Wt cells showed a predominant localization of GFP-CHI in the ER (data not shown). This additional localization of GFP-CHI in the ER has been reported for this construct (28, 30) and most likely reflects proteins in transit to the vacuole. In vti11 and enhanced vti11 mutants, >90% of the transformed cells showed diffuse vacuolar fluorescence, although in many cells (marked by arrows in Fig. 3) at a lower level than in Wt plants. In those cells showing lower vacuolar fluorescence, we also observed spherical compartments that may correspond to the small vacuole-like structures seen by electron microscopy in vti11 plants (26) and that are probably visualized by the GFP-CHI present in the ER network surrounding these vacuoles. Importantly, in vti12 mutants, most transformed cells (81% in vti12 and 92% in enhanced vti12 plants) showed very weak diffuse fluorescence, and, instead, GFP-CHI was observed predominantly in the ER (Fig. 3). These results indicate that GFP-CHI trafficking to the vacuole was hindered in vti12 mutants, and the underlying ER-localized GFP-CHI was revealed. Similarly, in Wt cells incubated in the light, the vacuole-localized GFP-CHI is degraded, and then the underlying ER-network pattern becomes evident (30). We did not observe accumulation of GFP-CHI in the apoplasm in vti12 mutants, possibly because the protein diffuses or is degraded in the extracellular space and does not accumulate to a significant concentration.

When transformed with ALEU-GFP, fluorescence was observed in the vacuole lumen in all transformed cells from Wt, vti12, and enhanced vti12 plants, indicating that ALEU-GFP transport to the vacuole does not require VTI12. In contrast, bright granular staining, in addition to the diffuse vacuolar fluorescence, was observed in vti11 and enhanced vti11 cells transformed with ALEU-GFP (arrowhead, Fig. 3). These punctate structures may correspond to Golgi or post-Golgi/prevacuolar compartments that have been shown to accumulate ALEU-GFP when transport is blocked by a dominant negative mutant version of m-Rab_mc (4). VTI11 and m-Rab_mc are both localized in the prevacuolar compartment and the Golgi apparatus, and their mutations may cause the retention of ALEU-GFP in these compartments. In contrast, in vti12 cells, we observed accumulation of GFP-CHI in the ER but not in punctate structures, even though VTI12 is also localized in the Golgi. Vacuolar storage proteins are sorted earlier in the secretory pathway than lytic vacuole proteins (8), in some cases, even in the ER (1, 10). Moreover, storage proteins may be transported to the PSV through a direct pathway from the ER that bypasses the Golgi (1, 10). The accumulation of GFP-CHI in the ER in vti12 cells may reflect the specialized role of this compartment in the sorting and transport of vacuolar storage cargo in plants.

VTI12 Is Involved in Transport of Endogenous Vacuolar Proteins.

Several classes of VSSs have been described in plants (13). Among these, the COOH-terminal class of VSSs (ctVSS) has only been found in storage vacuole proteins. In Arabidopsis, no proteins containing ctVSSs have been characterized (31), so the analysis of this class of vacuolar proteins in this model plant has relied on the expression of exogenous markers, such as BL and tobacco chitinase. Recently, we described a subgroup of Arabidopsis peroxidases (VacPerox) that contain putative ctVSSs and were found in enriched vacuolar fractions (32). We raised antibodies against an internal peptide that is specific to this subclass of peroxidases. Using immunoelectron microscopy, we localized VacPerox to the vacuole of Wt plants (SI), confirming that they are true vacuolar residents. Some labeling was also observed in the cytosol but not in the apoplasm. Accordingly, we did not observe substantial accumulation of VacPerox in apoplastic fluids isolated from Wt plants, but neither did we observe VacPerox in the apoplasm of vti mutants (data not shown), possibly because of redundancy between VTI11 and VTI12. We reasoned that defects in VacPerox trafficking might be detected in conditions closer to saturation, i.e., in plants overexpressing other vacuolar cargo, so we analyzed plants expressing BL and VAC2. We observed secretion of VacPerox in vti12 plants expressing BL, but importantly, not in Wt, nor in vti11 plants expressing BL (Fig. 4A). We then examined plants expressing VAC2 and detected secretion of VacPerox in enhanced vti12 plants, but not in Wt, vti11, vti12, or enhanced vti11 plants (Fig. 4B and data not shown). Two bands were occasionally observed reacting with the anti-VacPerox antibodies (Fig. 4B) and may represent different isoforms of the subgroup of vacuolar peroxidases. Importantly, both bands were observed in the apoplasm fraction of vti12 mutants. Moreover, there was specificity in the secretion of vacuolar cargo, because no secretion of AtCPY and AtFruct4, markers of LVs, was observed in vti12 or in enhanced vti12 plants expressing either VAC2 or BL (Fig. 4B and data not shown). These data show that the specific role of VTI12 was in transport of endogenous vacuolar proteins containing putative ctVSS signals and not in the transport of LV markers. In addition, the enhanced VacPerox secretion observed in BL- and VAC2-expressing plants indicates that all these cargo proteins use a common transport pathway, and, thus, VacPerox may serve as endogenous storage vacuole markers in Arabidopsis vegetative tissues.

Seed Storage Protein Deposition Is Altered in vti12 Mutants.

12S globulins are a major class of storage proteins in Arabidopsis seeds. It has been shown that their normal deposition requires the putative vacuolar sorting receptor VSR1, which interacts with a C-terminal peptide from 12S globulins (12). Seeds of vsr1–1 mutants secrete a fraction of 12S globulins to the apoplasm and accumulate their precursors, possibly because the secreted pool is not processed. As shown in Fig. 4C, we also observed accumulation of 12S globulin precursors in siliques from vti12 plants, suggesting that partial secretion occurred. Precursor accumulation was higher in siliques from enhanced vti12 plants, even though, because of low transmission of the vti11 allele, only 45% of the embryos in those siliques were actually enhanced vti12 (as determined by genotyping the progeny, n = 132). However, even considering this, the levels of precursor accumulation are lower than those observed in vsr1–1 mutants, indicating that redundancy from VTI11 is more prominent in seeds than in vegetative tissues. Consistent with a higher contribution from VTI11 in seed storage protein transport, precursors of 12S globulins were observed in vti11 and enhanced vti11 mutants (Fig. 4C), albeit at lower levels than in enhanced vti12 siliques. Furthermore, three of the four members of the Arabidopsis VTI family, VTI11, VTI12, and VTI14, are induced during seed maturation (33), coinciding with the induced expression of seed storage genes, and may provide added redundancy to the trafficking step regulated by VTI12. An alternative model to explain the relatively low levels of precursor accumulation in enhanced vti12 mutants is that they may cosecrete both the storage proteins and their processing peptidases, leading to maturation of 12S globulins in the apoplasm.

To gain further evidence of the possible role of VTI12 in storage protein deposition in seeds, we analyzed the morphology of PSVs, which was altered in the vsr1–1 trafficking mutant (12). In enhanced vti12 embryos, we observed PSVs that were smaller than in Wt (Fig. 4D). Furthermore, the cellular space in enhanced vti12 embryos was not completely occupied by PSVs, which retained spherical shapes and clear separations between them. In contrast, Wt cells were tightly packed with PSVs, which had irregular shapes and narrower intervening spaces between them. Therefore, PSVs occupy a smaller fraction of cellular volume in enhanced vti12 embryos, suggesting that deposition of storage proteins is decreased.

Taken together, our results lend strong support to the role of VTI12 in mediating transport to PSVs in both vegetative and seed tissues, although in these latter tissues redundancy from VTI11 and VTI14 may make the system less dependent on VTI12. The identification of VTI12 provides important leads to elucidate the possible role of other components of the trafficking machinery to PSVs in plants, such as SYP41, SYP42, SYP61, YKT61, YKT62, and AtVPS45, all of which have been shown to form complexes with VTI12 (22, 23, 34).

Plant Materials.

The Arabidopsis vti11 (zig1) and vti12 mutants were described (24) and are in the Col-0 background. The Arabidopsis vsr1–1 allele is in the Wassilewskija (Ws) background (12). Genotyping of the VTI11 and VTI12 loci was done as described (24).

Production of Antiserum Against VacPerox.

Antibodies were raised against a peptide (SFRTEKDAFGNANSARG) conserved in vacuolar peroxidases. Antibodies from an immunized rabbit (Invitrogen, San Diego, CA) were affinity-purified by transfer of the peptide to nitrocellulose membranes and subsequent immunoabsortion and elution of the antibody as described (6).

Isolation of Apoplastic Fluids.

Apoplastic fluid was obtained from 5-week-old plants. One gram of leaves was vacuum infiltrated in 50 ml of 50 mM Na-phosphate buffer, pH 7.0, and 150 mM NaCl for 15 min, breaking the vacuum every 3 min. After infiltration, leaves were blotted dry and the apoplastic fluid was collected by low-speed centrifugation (900 × g). Total proteins were extracted from leaves after apoplastic fluid collection. Aliquots from apoplastic and total protein fractions obtained from different plants were analyzed by SDS/PAGE and immunoblotting.

Transient Expression and Fluorescence Microscopy.

Five micrograms of plasmid DNA were used to coat gold particles for bombardment using a helium-driven particle accelerator (PDS-1000/He; Bio-Rad, Hercules, CA). Bombarded leaves were incubated for 24 h in the dark on solid media containing MS salts and 1% sucrose. Leaves were then mounted in water on glass slides with the adaxial epidermis facing the cover glass. GFP fluorescence was viewed with an Axiovert 200 confocal microscope (Zeiss, Thornwood, NY) coupled to the Bio-Rad Radiance 2100 laser scanning confocal imaging system with LaserSharp version 5 imaging software.

Storage Vacuole Visualization.

Seed covers were removed, and dissected embryos were mounted in water. PSV autofluorescence was excited at 488 nm by using an argon ion laser and subsequently detected through a 500LP nm emission filter.