The

Pla

nt C

ell

The Plant Cell, Vol. 15, 2885–2899, December 2003, www.plantcell.org © 2003 American Society of Plant Biologists

The VTI Family of SNARE Proteins Is Necessary for Plant Viability and Mediates Different Protein Transport Pathways

Marci Surpin,

a,1

Haiyan Zheng,

a,1,2

Miyo T. Morita,

b,1

Cheiko Saito,

b

Emily Avila,

a

Joshua J. Blakeslee,

c

Anindita Bandyopadhyay,

c

Valentina Kovaleva,

a

David Carter,

a

Angus Murphy,

c

Masao Tasaka,

b

and Natasha Raikhel

a,3

a

Center for Plant Cell Biology and Department of Botany and Plant Sciences, University of California, Riverside, California 92521

b

Nara Institute of Science and Technology, Graduate School of Biological Sciences, Ikoma Nara 630-0101, Japan

c

Horticulture Department, Purdue University, West Lafayette, Indiana 47907-1165

The Arabidopsis genome contains a family of v-SNAREs:

VTI11

,

VTI12

, and

VTI13

. Only

VTI11

and

VTI12

are expressed atappreciable levels. Although these two proteins are 60% identical, they complement different transport pathways when ex-pressed in the yeast

vti1

mutant.

VTI11

was identified recently as the mutated gene in the shoot gravitropic mutant

zig

.Here, we show that the

vti11 zig

mutant has defects in vascular patterning and auxin transport. An Arabidopsis T-DNA in-sertion mutant,

vti12

, had a normal phenotype under nutrient-rich growth conditions. However, under nutrient-poor condi-tions,

vti12

showed an accelerated senescence phenotype, suggesting that VTI12 may play a role in the plant autophagypathway. VTI11 and VTI12 also were able to substitute for each other in their respective SNARE complexes, and a double-mutant cross between

zig

and

vti12

was embryo lethal. These results suggest that some VTI1 protein was necessary forplant viability and that the two proteins were partially functionally redundant.

INTRODUCTION

The plant endomembrane system plays a variety of rolesthroughout plant development. It is composed of a series ofmembrane-bound organelles that include the endoplasmicreticulum, the Golgi apparatus, the

trans

-Golgi network (TGN),the prevacuolar compartment (PVC), the vacuole, and endo-somes. Vesicle transport is the major means by which proteinstravel to and from these organelles. The correct targeting ofproteins to the vacuole or other organelles of the endomem-brane system is dependent on the action of SNAREs (solubleNSF attachment protein receptors). The majority of SNAREsare type-II membrane proteins with coiled-coil domains that areimportant for interacting with other SNAREs. SNAREs are di-vided into two groups—vesicle-SNAREs (v-SNAREs) and tar-get-SNAREs (t-SNAREs), which are located on the target mem-brane. Generally, three t-SNAREs form a

cis

-SNARE complex,which is recognized by a v-SNARE that aligns its coiled-coil re-gion to form a four-helix bundle. The formation of this complexdrives the fusion of the target and vesicle membranes, allowingthe delivery of cargo protein to the target compartment. Oncedelivery is complete, the complex is dissociated. Dissociationof the four-helix bundle requires soluble NSF and

�

-SNAP pro-teins. After the SNARE complex is dissociated in an energy-

requiring process, the v-SNARE is recycled back to its originalcompartment.

Although the basic machinery of vesicular transport is con-served among all eukaryotic cells, the plant endomembranesystem also has unique components. In mammals and yeast,each cell has only one type of vacuole, whereas in plants, dif-ferent types of vacuoles—lytic and protein storage—frequentlycoexist in one cell (Vitale and Raikhel, 1999). It is possible thatthis complex vacuolar system requires more sophisticatedtransport machinery in plants than in yeast. Based on genomeanalysis, there are 21 SNAREs in

Saccharomyces cerevisiae

. InArabidopsis, there are 55 SNAREs, and these can be sub-grouped into several families (Sanderfoot et al., 2000). Thesemultigene families could reflect a more complex endomem-brane system in plants.

The VTI1-type v-SNARE has multiple functions. In yeast, al-though encoded by a single gene, it can form SNARE com-plexes with five syntaxins (Abeliovich et al., 1999; Fischer vonMollard and Stevens, 1999). Accordingly, it has been shown tofunction in multiple pathways to the vacuole: the CPY pathwayvia the PVC, the ALP pathway via the Golgi apparatus, and theCvt pathway, which transports proteins such as API from thecytoplasm to the vacuole. In mouse, MmVTI1a is localized tothe Golgi and involved in intra-Golgi trafficking. MmVTI1b isfound on the Golgi, the TGN, and possibly the endosome (Xu etal., 1998).

In Arabidopsis, the VTI1 family is composed of three closelyrelated members:

VTI11

,

VTI12

, and

VTI13

.

VTI11

and

VTI12

are highly expressed in all organs.

VTI13

does not produce anymRNA that is detectable by reverse transcriptase–mediated(RT) PCR, although Basic Local Alignment Search Tool (BLAST)searches show that there is one

VTI13

EST clone in the data-

1

These authors contributed equally to this report.

2

Current address: Core Technology Area, Discovery Research, NovartisPharmaceuticals Corp., 556 Morris Avenue, Summit, NJ 07901.

3

To whom correspondence should be addressed. E-mail nraikhel@citrus.ucr.edu; fax 909-787-2155.

Online version contains Web-only data.Article, publication date, and citation information can be found atwww.plantcell.org/cgi/doi/10.1105/tpc.016121.

The

Pla

nt C

ell

2886 The Plant Cell

base, compared with

�

12 each for

VTI11

and

VTI12

(M. Surpinand N. Raikhel, unpublished data). VTI11 and VTI12 are 60%identical. They are

�

25% identical to yeast Vti1p and are closerto mammalian VTI1a than to mammalian VTI1b (Sanderfoot etal., 2000). Previously, we described the characterization ofVTI11 (referred to as AtVTI1a by Zheng et al. [1999]). VTI11 islocalized on the TGN with the NTPP cargo receptor ELP, and itcolocalizes with the SYP2 and SYP5 groups of syntaxins on thePVC (Zheng et al., 1999; Bassham et al., 2000; Sanderfoot etal., 2001b). Recently, a defect in the

VTI11

gene was identifiedas the

zig

mutation in a screen for mutants impaired in theshoot gravitropic response (Kato et al., 2002). In the

zig

mutant,amyloplasts do not sediment to the bottom of endodermalcells, and abnormal vesicular structures are found in severaltissues. Endodermis-specific expression of wild-type

VTI11

inthe

zig

mutant complements the shoot gravitropism defect.These results suggest that vacuole formation is important forthe shoot gravitropic response (Morita et al., 2002).

There is evidence that VTI11 and VTI12 function in differentvesicle transport pathways. First, VTI11 and VTI12 complementdifferent yeast

vti1

alleles that block different vacuolar proteintransport pathways (Zheng et al., 1999). VTI11 expressed inyeast

vti

temperature-sensitive and null mutants suppressesthe CPY mistargeting and growth-defect phenotypes, whereasVTI12 restores API transport via the Cvt pathway as well as thetransport of vacuolar ALP. Arabidopsis VTI12 also rescues theyeast

vti1p

null mutant, but the cells do not grow as quickly asthose complemented with Arabidopsis VTI11 (Zheng et al.,1999). Second, in plant cells, VTI12 forms a SNARE complexwith the SYP4- and SYP6-type SNAREs (Bassham et al., 2000;Sanderfoot et al., 2001b). These results indicate that VTI12plays a role in vesicle trafficking different from that of VTI11.

We took a genetic approach to determine the respectivefunctions of VTI1 family members. Here, we describe the Arabi-dopsis mutant

vti12

, further characterize the

zig

mutant, andpresent the results of a genetic cross between these two lines.We show that VTI11 and VTI12 mediate different protein trans-port pathways, and although they do not fully complementeach other at a functional level, they can substitute for eachother in their respective SNARE complexes. We also show thata double-mutant cross between

zig

and

vti12

is embryo lethal,indicating that at least some VTI1 family protein must bepresent to ensure viability.

RESULTS

Characterization of

vti12

A reverse-genetics approach was taken to address the functionof VTI12. To identify a T-DNA insertion mutant of

VTI12

, poolsof Arabidopsis Col-

gl

seeds (see Methods) mutagenized withT-DNA were screened using gene-specific primers and T-DNAborder primers as described by Sanderfoot et al. (2001a). Onemutant plant with a T-DNA inserted at

�

76 nucleotides up-stream of the transcription start site (Figure 1A) was identified.This allele was named

vti12

for its mutation in

VTI12

. TotalRNAs prepared from wild-type plants, heterozygous plants,and homozygous

vti12

plants were used as templates for RT-

PCR. As shown in Figure 1B, RT-PCR of homozygous

vti12

plants detected no mRNA for

VTI12

. RT-PCR of an unrelatedgene (

NPSN12

) using the same RNA samples was used as acontrol. Wild-type plants and plants with homozygous inser-tions were characterized further by protein gel blot analysis. Asshown in Figure 1C, the VTI12 antibody recognized a protein inwild-type plants at

�

28 to 30 kD that was missing from the ho-mozygous

vti12

plants. By contrast, when VTI11 antibodieswere used for protein gel blot analysis, there was no significantdifference between proteins from wild-type and

vti12

plants.Under normal growth conditions,

vti12

plants had the samemorphology and growth characteristics as wild-type plants.

VTI11 and VTI12 Functional Redundancy

Previously, we showed that VTI1 family proteins, althoughclosely related, form SNARE complexes with different sets ofsyntaxins (Bassham et al., 2000; Sanderfoot et al., 2001b). Fur-

Figure 1. vti12 Is a Null Mutant for VTI12.

(A) The T-DNA insertion in vti12. T-DNA is inserted at �76 nucleotidesupstream from the transcription start of VTI12 on chromosome I.(B) RT-PCR of VTI12. Total RNAs from seedlings of the wild type (�/�)and heterozygous (�/�) and homozygous (�/�) insertion mutant plantswere used as templates. Molecular weights are indicated at left. RT-PCR of NPSN12 using the same templates with gene-specific primersfor NPSN12 mRNA is shown at bottom.(C) Protein gel blots of total protein extracted from wild-type (�/�) andvti12 (�/�) seedlings. The proteins were separated by SDS-PAGE, andimmunoblots were used to visualize VTI11 or VTI12. Arrows indicate theVTI12 protein band that was missing in vti12 plants.

The

Pla

nt C

ell

Arabidopsis VTI1 v-SNARE Family 2887

thermore, although

vti12

appeared the same as the wild typeunder normal growth conditions, plants that are mutated at the

zig

locus have a distinctive visible phenotype (Kato et al., 2002;Morita et al., 2002). Thus, VTI11 and VTI12 have functional dif-ferences as well. In addition, the double homozygous mutant

vti11 vti12

was embryo lethal (see below), suggesting that (1)some VTI family protein was essential for viability and (2) bothexpressed VTI proteins were able to substitute functionally foreach other. To evaluate the second possibility, we analyzed theVTI11- and VTI12-containing SNARE complexes in wild-type,

vti12

, and

zig-1

plants. As shown in Figure 2, when VTI11 anti-bodies were used for immunoprecipitation, SYP2 and SYP5families of syntaxins were coprecipitated at similar levels inboth wild-type and

vti12

plants. However, there were signifi-cantly more SYP4- and SYP6-type SNAREs coprecipitatedwith VTI11 from

vti12

than from wild-type plants (Figure 2).Thus, we concluded that VTI11 forms a SNARE complex withSYP4- and SYP6-type SNAREs and possibly performs thefunction of VTI12 in the

vti12

mutant.When we performed coimmunoprecipitations on

zig-1

ex-tracts from both roots and shoots, we detected VTI12 in a com-plex with SYP5- and SYP2-type SNAREs (Figure 2). The exper-imental procedure was the same as that used with

vti12

andwas performed using purified anti-VTI12 antibodies (see Meth-ods). Equal amounts of VTI12 were immunoprecipitated fromwild-type and

zig-1

plants. SYP6 was coprecipitated from bothwild-type and

zig-1

plants at similar levels (Figure 2). We ob-served no significant differences in the amounts of coprecipi-tated SYP4-type SNAREs (data not shown). However, moreSYP2- and SYP5-type SNAREs were immunoprecipitated from

zig-1

than from wild-type plants, likely as a result of their inter-action with VTI12 in

zig-1

.We examined whether VTI12 can substitute functionally for

VTI11 in plants. First, we doubled the copy number of

VTI12

inthe

zig-1

mutant by introducing a genomic fragment containing

VTI12

with expression driven by its own promoter. There wasno effect on the

zig-1

phenotype in these transgenic lines (datanot shown). We then overexpressed

VTI12

in

zig-1

plants using

the 35S promoter. We analyzed a number of independent linesof

zig-1

transgenic plants carrying the 35S-

VTI12

cDNA con-struct. Although we observed some slight variation amongthese lines, the majority exhibited the phenotype shown in Fig-ure 3. Plants homozygous for the transgenic construct had nor-mally shaped leaves and stems, although the stems wereslightly smaller and thinner than those of the wild type (Figures3A and 3D). Plants heterozygous for the transgenic constructhad an intermediate phenotype for stem morphology but werewild type with respect to leaf shape (Figures 3B and 3E, left).The zygosity of the transgene also affected the gravitropic re-sponse.

zig-1

plants that were homozygous for the 35S-

VTI12

construct had a normal gravitropic response. Plants that wereeither heterozygous or did not have any copies of the trans-gene showed reduced or no gravitropic response, respectively.In addition, we examined plants for relative amounts of VTI12.The amount of VTI12 correlated with the zygosity of the trans-gene.

zig-1

plants that were homozygous for the 35S-

VTI12

construct contained excessive amounts of VTI12 (Figure 3F),and plants with one or no copies contained intermediate or nor-mal levels of VTI12, respectively (Figure 3F). These results sug-gested that an excess of VTI12 was able to form a SNAREcomplex with SYP2 and SYP5 in the absence of VTI11 (Figure2), and the complex was functional to the extent that it wasable to support embryonic development (see below). However,normal amounts or a fewfold difference in the quantity of VTI12was not enough to complement the leaf and stem morphogen-esis defects in zig-1.

Based on the data described above, we propose that bothVTI11 and VTI12 can substitute for each other in their respec-tive SNARE complexes at the molecular level. We also believethat these complexes are, to some degree, functional.

VTI12 Functions in the Plant Autophagy Pathway

The vti12 plants were checked for defects in NTPP vacuolarprotein transport. It has been shown that Arabidopsis aleurainhas a typical NTPP signal and possibly is transported through a

Figure 2. Immunoprecipitation of VTI12 from Wild-Type and zig Plants and of VTI11 from Wild-Type and vti12 Plants.

Total membrane protein samples (T) and eluates (EL) from anti-VTI11 and anti-VTI12 columns were separated by SDS-PAGE. Potential interactingSNARE partners were visualized by immunoblotting. In the zig-1 (vti11) mutant, VTI12 forms a SNARE complex with SYP2- and SYP5-type proteins.In the vti12 mutant, VTI11 coimmunoprecipitates with SYP4- and SYP6-type SNAREs. IP, immunoprecipitate; WT, wild type.

The

Pla

nt C

ell

2888 The Plant Cell

Figure 3. Excess VTI12 Can Complement Defects in the Atvti11/zig-1 Mutant.

(A) to (C) Rosettes of 3-week-old plants.(D) and (E) Four-week-old plants.T3 siblings of zig-1 transformed with 35S-VTI12 showed plants homozygous for the transgenic construct ([A] and [D]), heterozygous for the construct([B] and [E], left), and with no copies of the construct ([C] and [E], right). The gravitropic response was tested by placing plants horizontally at 23�C inthe dark for 90 min. The gravitropic responses are indicated as follows: ��, �90�, same as in the wild type; �, �45 to 60�; and �, no response, sameas in zig-1.(F) VTI12 protein content. The solubilized membrane fractions from stems with 1% Triton X-100 were analyzed by SDS-PAGE and protein gel blotanalysis. Top gel: lane 1, zig-1; lane 2, T3 siblings homozygous for the transgenic construct; lane 3, T3 siblings heterozygous for the transgenic con-struct; lane 4, zig-1 plants with no transgene; lane 5, vti12 (negative control). The bottom gel is a Coomassie blue–stained gel showing that equalamounts of protein extracts were loaded.

The

Pla

nt C

ell

Arabidopsis VTI1 v-SNARE Family 2889

pathway common for NTPP proteins (Ahmed et al., 2000). In to-tal protein extracts, no difference in the total amount of maturealeurain was found between wild-type and vti12 plants (datanot shown), nor was there any obvious abnormal accumulationof higher molecular variants of aleurain in the vti12 plants. Weconcluded that the VTI12 v-SNARE does not act in any of thepreviously characterized NTPP cargo transport pathways.

We showed previously (Zheng et al., 1999) that the VTI12(AtVTI1b) cDNA complements the cytoplasm-to-vacuole (Cvt)pathway in yeast vti1p mutants. In yeast, the Cvt pathway sharesmany components with the autophagy pathway (Reggiori andKlionsky, 2002). Thus, we determined whether vti12 mutantsshow a similar phenotype under the same conditions as seen inknown Arabidopsis autophagy mutants. Seeds were germinatedon Gamborg’s B-5 medium with 2% sucrose. At 2 weeks, seed-lings were transferred to plates containing no nitrogen/no car-bon (No N/C) medium for varying amounts of time. The plantswere transferred back to Gamborg’s B-5 medium, and thenumber of plants recovered (scored by new green leaves) wascounted. Figure 4A shows the recovery curve of wild-type,vti12, and zig-1 plants that were on No N/C medium for 32days. vti12 plants clearly recovered from starvation more slowlythan either wild-type or zig-1 plants, and by this time point, asmall proportion of the vti12 plants did not recover at all.

Once recovered, vti12 plants were smaller than identicallytreated wild-type plants (Figure 4B). Seedlings were grown onGamborg’s B-5 medium for 2 weeks, transferred to No N/Cmedium for 30 days, and then transferred back to Gamborg’sB-5 medium. At 13 days after transfer, wild-type plants were

significantly larger and more robust than vti12 plants. Further-more, vti12 plants on No N/C medium developed their primaryinflorescence (bolted) earlier than wild-type plants under thesame conditions (see supplemental data online).

Previously characterized Arabidopsis autophagy mutantshave shown accelerated leaf senescence (Doelling et al., 2002;Hanaoka et al., 2002). We performed detached leaf assays todetermine whether vti12 mutants also have this phenotype.First and second true leaves were excised from 2-week-oldseedlings and were either placed on filter paper soaked in Mesbuffer or floated on water. In either condition, the leaves wereplaced in the dark and retrieved at various time points. Figure4C shows that vti12 leaves became chlorotic �1 day before ei-ther wild-type or zig-1 leaves. We also examined the inductionof the SEN1 and YSL4 genes, which are markers for leaf senes-cence in Arabidopsis, by semiquantitative RT-PCR (Hanaoka etal., 2002) (Figure 4D). In wild-type plants, there was initially a lowlevel of SEN1 transcript and then the level of transcript in-creased between 24 and 48 h. By contrast, the leaves of the vti12mutant accumulated a large amount of SEN1 transcript between12 and 24 h, and this high level of accumulation was main-tained through the 48-h time point. In wild-type plants, YSL4showed a large increase in the accumulation of transcript be-tween 24 and 48 h. In vti12, YSL4 mRNA accumulation appearedto be increased already at time 0. It increased significantly at�12 h and decreased subsequently at 24 and 48 h. Together,these results indicated that vti12 plants not only entered senes-cence prematurely after leaf detachment but also may have ex-isted already in a senescent state before leaf excision.

Figure 4. The vti12 Mutant Has a Defect in the Autophagy Pathway.

(A) vit12 plants do not recover as quickly from nutrient deprivation as wild-type or zig-1 plants.(B) Recovery of plants grown on No N/C medium. Recovery at 3 days (top) and the same plates at 13 days (bottom).(C) Detached leaf assay. First and second true leaves were excised from seedlings and placed in the dark for the times indicated. The vti12 leavesshowed signs of chlorosis at least 1 day before the wild-type and zig-1 leaves.(D) RT-PCR of RNA isolated from detached leaves. Leaves were collected at the times indicated, and RT-PCR of the senescence markers SEN1 andYSL4 was performed, with Act2 as a loading control.

The

Pla

nt C

ell

2890 The Plant Cell

Electron Microscopic Examination of Wild-Type and vti12 Plants Grown under Starvation Conditions

The results from the starvation and detached-leaf assays sug-gested a role for VTI12 in the autophagy pathway. We micro-scopically examined wild-type and mutant cells under condi-tions that are known to induce autophagy. Plants were grownon rich medium for 2 weeks and then transferred to plates con-taining No N/C medium. Ultrathin sections were made from theleaves of wild-type and vti12 plants at various times, and usingelectron microscopy, we examined cells from the two lines.We also made corresponding thick sections to orient ourselvesto the gross morphology of the leaf sections (data not shown).We noted that similar effects of starvation occurred in bothpalisade and spongy parenchyma cells. Figure 5 shows elec-tron micrographs of wild-type and vti12 cells at different timepoints. Cells from time 0 for both wild-type and vti12 plants(Figures 5A and 5B) showed similar morphologies. The vacuoledid not take up the entire cellular space, and there was a perim-eter of cytoplasm that contained a number of organelles. After6 days on No N/C medium, the vacuole from wild-type plantswas expanded and appeared to have engulfed the organelles,whereas the cytoplasmic perimeter disappeared in wild-typecells (Figure 5C). In vti12 cells, however, the cytoplasmic perim-eter persisted and the vacuole did not expand to the same ex-tent as in wild-type cells (Figure 5D). After 12 days on No N/Cmedium, the organelles from wild-type plants appeared to havebeen fully engulfed and were in the process of being digested,whereas in vti12 plants, the organelles still appeared to be in-tact and remained outside the vacuole (Figures 5E and 5F, re-spectively). By 19 days on No N/C medium (Figures 5G and5H), the wild-type cells had vacuolar inclusions that containedremnants of organelles with some intact, but disorganized,membrane structure. By contrast, vti12 plants had vacuolar in-clusions that were bubble-like in appearance (Figure 5H) andappeared to have been composed of many small, empty vesi-cles. We observed a number of these unique multivesicularstructures in vti12 cells but not in any wild-type cells. Clearly,the progression of the autophagic process was impeded invti12 cells. Besides being slower, there appeared to be a defectin autophagosome formation and/or fusion.

Double-Mutant Studies with vti12 and zig-1

We took a genetic approach to better understand the individualand possibly overlapping functions of the VTI family of proteins.Toward this end, we created a double-mutant cross, but wewere unable to detect any homozygous double mutants in thecourse of our genotyping experiments. We examined siliquesfrom F1 plants to test whether the double homozygous mutantcombination was embryo lethal. Whereas seeds from wild-typeplants rarely aborted (Figure 6A), self-pollinated vti12/vti12 ZIG-1/zig-1 and VTI12/vti12 zig-1/zig-1 siliques frequently had nonvia-ble progeny (Figures 6Ai to 6Aviii). As shown in Figures 6Aiii to

Figure 5. Electron Microscopic Examination of Autophagy in Wild-Typeand vti12 Cells.

Plants were grown on Gamborg’s B-5 medium for 2 weeks and trans-ferred to No N/C medium, tissue was collected, and ultrathin sectionswere made at the times indicated below.(A), (C), (E), and (G) Wild-type cells.(B), (D), (F), and (H) vti12 cells.(A) and (B) Spongy parenchyma cells at time 0.(C) and (D) Cells at 6 days on No N/C medium. The vacuoles in wild-type plants are fully expanded, whereas those in vti12 cells are only par-tially expanded (arrows). A number of autophagosomes are visible at theperiphery of the wild-type cells (C).(E) and (F) Cells at 12 days on No N/C medium. Organelles were fullyengulfed in wild-type cells, whereas a number of organelles still wereoutside the vacuole in vti12 cells (arrows).

(G) and (H) Vacuolar inclusions from cells grown for 19 days on No N/Cmedium.

The

Pla

nt C

ell

Arabidopsis VTI1 v-SNARE Family 2891

Figure 6. The Genetic Cross of vti12 and zig (vti11) Mutants.

(A) Effect of the vti11 or vti12 mutation on seeds and embryo development.(i) to (v) Whole-mount images of the mature silique cleared by chloral hydrate:glycerol:water solution.(i) Wild-type.(ii) vti12/vti12.(iii) zig-1/zig-1 (vti11/vti11).(iv) vti12/vti12 ZIG-1/zig-1.(v) VTI12/vti12 zig-1/zig-1.(vi) to (viii) Light microscopic images of mature and abnormal seeds.(vi) Seeds maturing normally.(vii) Shriveled seeds.(viii) Aborted seeds.Arrowheads indicate unnatural intervals in which neither ovules nor funiculi were found. Abnormal seeds (shriveled and aborted seeds) or their traces(unnatural intervals) were observed frequently in zig-1/zig-1, vti12/vti12 ZIG-1/zig-1, and VTI12/vti12 zig-1/zig-1. Bars � 1 mm.(B) Visible phenotypes of 5-week old plants.(i) Plants that are genotyped as VTI12/vti12 zig-1/zig-1 (two bottom plants) have a phenotype that is more severe than that of zig-1 alone. They aresmaller, and their leaves are more wrinkled. These plants are severely developmentally delayed and have an altered phyllotaxy (not shown).(ii) Close-up of a 5-week-old VTI12/vti12 zig-1/zig-1 plant.(iii) A VTI12/vti12 zig-1/zig-1 10-week-old plant, at right, has a fasciated phenotype. The wild-type plant is at left.

The

Pla

nt C

ell

2892 The Plant Cell

6Av, there were many shriveled seeds and vacant spaces inthese siliques. Upon close examination, we detected the tracesof ovaries and funiculi in some vacancies. Thus, we categorizedthe nonviable progeny into three classes: shriveled seed (Figure6Avii), aborted seed (Figure 6Aviii), and unnatural interval (Fig-ure 6Av, arrowheads). We calculated the percentage of nonvia-ble progeny from vti12, zig-1, vti12/vti12 ZIG-1/zig-1, andVTI12/vti12 zig-1/zig-1. The results are summarized in Table 1.In both single mutants, a percentage of progeny did not de-velop into normal seeds (vti12, 13%; zig-1, 32%). In the vti12/vti12 ZIG-1/zig-1 and VTI12/vti12 zig-1/zig-1 plants, 31 and58% of the self-fertilized progeny were dead, respectively,whereas 25% of the self-progeny of vti12/vti12 ZIG-1/zig-1 andVTI12/vti12 zig-1/zig-1 would have been expected to be lethal.The actual lethality scores in the progeny of vti12/vti12 ZIG-1/zig-1 or VTI12/vti12 zig-1/zig-1 were higher than 25%, suggest-ing that the single mutations and the double homozygous mu-tation affect viability independently. In accordance with thishypothesis, the theoretical lethality (expected lethality) was cal-culated as shown in Table 1. Actual lethality scores (lethality ofembryo) in vti12/vti12 ZIG-1/zig-1 and VTI12/vti12 zig-1/zig-1were very similar to the theoretical values, in agreement withour statistical expectations.

We observed that �2 of every 15 plants had a zig-1 pheno-type and that �1 of every 20 plants had a novel phenotype bestdescribed as an enhanced zig phenotype. These plants weresmaller than zig-1 plants, had extremely wrinkled leaves, andexhibited a severe developmental delay (Figure 6Bi). The lifespan of these plants was correspondingly and significantlylonger. Mature plants also exhibited varying degrees of fascia-tion at the shoot apical meristem (Figure 6Biii) and an alteredphyllotaxy (data not shown). Despite their rather severe pheno-type, however, the plants that reached adulthood were fertileand yielded viable progeny. Genotyping of a large number of F2progeny showed that all of the plants with this enhanced zigphenotype were VTI12/vti12 zig-1/zig-1. Based on the principleof independent assortment, we would have expected 2 of every

15 plants to have an enhanced zig phenotype. A ratio of 1:20suggests that as a group these plants were not entirely viable.Plants with the inverse genotype, vti12/vti12 ZIG/zig-1, had thesame phenotype as their vti12 parents (i.e., they had no appar-ent abnormal phenotype under normal growth conditions).

Histological Studies of Mutant Lines

VTI11 was shown recently to encode ZIG, which, when mu-tated, causes a shoot gravitropism defect (Kato et al., 2002).The endodermal cells of inflorescence stems contain starch-filled amyloplasts that sediment to the bottom of the cell andact as statoliths (Fukaki et al., 1998). The endodermal cell layerof the zig mutant does not completely sediment starch-contain-ing amyloplasts, as in wild-type plants (Morita et al., 2002). Weextended this study to plants with the enhanced zig phenotypeby making longitudinal thin sections from wild-type, vti12, zig-1,and VTI12/vti12 zig-1/zig-1 inflorescence stems (Figure 7A). Asexpected, amyloplasts sedimented almost exclusively to thebottom of cells in the endodermal layer in wild-type and vti12plants. Amyloplasts could be found in approximately equal num-bers at the tops and bottoms of endodermal cells from zig-1 in-florescence stems. In the VTI12/vti12 zig-1/zig-1 line, the cortexand endodermal cell layers were highly disorganized and theindividual cell types were not readily distinguished. Cells thatdid contained amyloplasts featured them on the tops, bottoms,and occasionally the sides of these cells.

In wild-type and vti12 plants, a single epidermal layer en-closes approximately three to seven cortex cell layers and oneendodermal cell layer (Figures 7Ai, 7Aii, 7Bi, and 7Bii). The cellshapes and sizes in all of these layers were highly uniform andregular. These layers were easily distinguishable in zig-1, al-though the cell shapes and sizes were slightly irregular. In theVTI12/vti12 zig-1/zig-1 line, the epidermal, cortex, and endo-dermal cell layers were highly disorganized and often indistin-guishable. We made thick cross-sections in all four lines, andthe results are shown in Figure 7B. Both the wild-type and vti12

Table 1. Summary of the Lethality of vti1 Mutant Progeny

Genotype No.

Percent Normal Seedsa

Percent Shriveled Seedsb

Percent Aborted Seedsc

Percent Unnatural Intervalsd

Percent Lethality of Embryose

Percent Expected Lethality

Wild type (Co-0) 487 98 0.4 0.6 0.8 1.8 —vti12/vti12 630 96 1.0 1.7 2.1 3.8 1.8zig-1/zig-1 515 68 26 3.3 2.7 32 1.8vti12/vti12 ZIG-1/zig-1 358 69 6.7 21 3.1 31 29f

VTI12/vti12 zig-1/zig-1 184 52 17 30 1.6 48 50g

a Phenotype shown in Figure 6Avi.b Phenotype shown in Figure 6Avii.c Phenotype shown in Figure 6Aviii.d Phenotype shown in Figure 6Av, arrowheads.e Percentage of shriveled seeds � aborted embryos � unnatural intervals.f Calculated from the following formula: 1 � {Vcol � (1 � Ldm) � Vvti12}, where Vcol � 0.98 (viability in Columbia), Ldm � 0.25 (theoretical frequencyof zig-1/zig-1 vti12/vti12 progeny), and Vvti12 � 0.96 (viability in vti12/vti12 progeny).g Calculated from the following formula: 1 � {Vcol � (1 � Ldm) � Vvti11}, where Vcol � 0.98 (viability in Columbia; shown in this table), Ldm � 0.25(theoretical frequency of zig-1/zig-1 vti12/vti12 progeny), and Vvti11 � 0.68 (viability in zig-1/zig-1 progeny).

The

Pla

nt C

ell

Arabidopsis VTI1 v-SNARE Family 2893

lines showed similar patterns of tissue organization (Figures 7Biand 7Bii), although in the vti12 section there appeared to besome regions where there were two layers of endoderm cells.The zig-1 cross-section had recognizable vascular bundles,and the xylem cells appeared to be of normal size, shape, andnumber. However, the phloem cells were spread out over alarger area than in wild-type or vti12 cross-sections (Figure7Biii). In the enhanced zig phenotype cross-section (Figure7Biv), the epidermal layer appeared to be intact and composedof regularly sized cells, but the cortex and endodermal layersshowed irregularly shaped cells and there did not appear to beany distinction between the cell types. The vascular bundleswere not as prominent as in the wild-type and vti12 sections.The xylem cells were small with almost no sclerotic cells, andthe phloem appeared to form an almost continuous ring aroundthe cross-section, with correspondingly fewer interfascicularcells. The inner portion of the stele had cells that were smallerand had a more regular shape than in the wild-type, vti12, andzig-1 lines. There also were no gaps between the stele cells inthe enhanced zig line.

It appears, then, that the absence of VTI11 caused tissueidentity defects. The cross-sections all were obtained from tis-sue approximately halfway down the floral inflorescence stemat �1 week after bolting. It may be possible, however, thateven though we waited until comparable times after bolting, theages of the tissue were not strictly equivalent, because boththe zig-1 and VTI12/vti12 zig-1/zig-1 lines have significantly ex-panded life spans compared with the wild-type and vti12 lines.Nonetheless, it is clear that VTI11 plays some role in tissueidentity and organization.

Plants Mutated in VTI11 Have Auxin Transport Defects

Endodermal expression of VTI11/ZIG is necessary for the grav-itropic response (Morita et al., 2002). Other studies have shownthat vesicular transport is necessary for some aspect of theshoot gravitropic response. Arabidopsis pin3 mutants, whichcontain defects in an auxin efflux carrier, have a shoot-agravi-tropism phenotype (Friml et al., 2002). Recent studies haveshown that Arabidopsis seedlings treated with the drug brefel-din A, which inhibits vesicle transport, have mislocalized PINproteins (Geldner et al., 2001; Friml et al., 2002). Therefore, wedecided to determine whether the vesicle-transport defect inzig plants in turn causes a defect in the localization of auxin ef-flux carriers. First, we measured whole-plant auxin transportactivity in vti12 and zig-1 seedling hypocotyls. The zig-1 mutantshowed a significant reduction (�35%) in basipetal auxin trans-port through the vascular system from the shoot apical tip tothe hypocotyl-root transition zone (Figure 7C). Thus, we exam-ined the localization of PIN efflux carrier proteins in zig seed-lings. The localization in vascular tissues of PIN1 and PIN3 didnot appear to be altered dramatically in the zig mutants.

DISCUSSION

The purpose of this study was to determine the functions of thetwo expressed VTI1 family proteins, VTI11 and VTI12. Our re-

sults show that VTI12 functions in the plant autophagy pathwayand that VTI11 plays a role in basipetal auxin transport and celltype differentiation and specificity in addition to NTPP-medi-ated vacuolar transport. Furthermore, we showed that bothVTI1 family proteins substitute for each other at the molecularlevel and, to some extent, at the functional level. A double-mutant cross experiment showed that the double homozygousvti11 vti12 mutant was embryo lethal, thus demonstrating thatsome VTI1 family protein is required for plant viability.

Some VTI1 Protein Is Necessary for Plant Viability, and VTI11 and VTI12 Can Substitute for Each Other in Their Respective SNARE Complexes

The shoot gravitropism mutant zig has a lesion in the VTI11gene and has a severe defect in the gravitropism of the inflores-cence stem and hypocotyl (Kato et al., 2002). The vti12 mutantexhibited no visible phenotype under normal conditions. Wecreated a double-mutant cross between the vti12 and zig mu-tants. The double homozygous mutant was embryo lethal. Fur-thermore, examination of progeny siliques from genotypedplants clearly demonstrated a reduction in embryo viability. Thecomplete loss of VTI1 family protein appeared to affect the em-bryos quite early in development, whereas the loss of one orthe other protein increased the number of deflated seeds. Atthis time, we do not know precisely at which stage the vti11vti12 double mutant aborts. The vacuole biogenesis mutantvacuoless1 (vcl1) also is embryo lethal and terminates some-time near the torpedo stage (Rojo et al., 2001). The VCL1 geneencodes the Arabidopsis homolog of yeast Vps16p, a compo-nent of the class-C VPS complex. In yeast, the class-C VPScomplex is required for Golgi-to-vacuole protein transport andhas been shown to mediate the trans-SNARE pairing of theVti1p-Vam7-Vam3 complex (Sato et al., 2000). In Arabidopsis,the VCL1 complex has been shown to contain SYP2-type syn-taxins (Rojo et al., 2003), which interact with VTI11 (Bassham etal., 2000; Sanderfoot et al., 2001b). Therefore, we speculatethat the Arabidopsis vti11 vti12 mutant perhaps no longer me-diates vesicle fusion events involving the VCL1 complex andconsequently may be embryo lethal at the same stage as thevcl1 mutant.

Our results from coimmunoprecipitation experiments usingextracts from the vti12 and zig mutants suggested that both VTIfamily SNARE proteins were able to substitute for each other atthe molecular level. We also complemented the zig-1 pheno-type with a 35S-VTI12 construct. zig-1 plants that were ho-mozygous for the transgenic construct showed a wild-typegravitropic response, whereas mutant plants that were het-erozygous for the transgene showed an intermediate response.Thus, it appears that the mutant complexes are functional butprobably are less efficient in their cargo delivery. Therefore,cargo is delivered, but perhaps not in a timely manner, and thisaffects the pathway defined by VTI11 much more than that de-fined by VTI12.

VTI1-type SNAREs are mobile proteins. Immunocytochemi-cal studies have localized both VTI11 and VTI12 on the TGNand the PVC, although their distribution ratios between the

The

Pla

nt C

ell

2894 The Plant Cell

Figure 7. Histological Analysis of the Wild Type, vti12, zig-1 (vti11), and VTI12/vti12 zig-1/zig-1.

(A) Longitudinal stem sections of each line (wild type [i], vti12 [ii], zig-1[iii], and VTI12/vti12 zig-1/zig-1 [iv]) were stained with toluidine blue and exam-ined for amyloplast localization and tissue organization. The VTI12/vti12 zig-1/zig-1 plants showed more haphazard placement of amyloplasts andmore extreme disorganization of tissue layers than the zig-1 single mutant line.(B) Cross-sections of wild type (i), vti12 (ii), zig-1(iii), and VTI12/vti12 zig-1/zig-1 (iv) plants. Cross-sections were taken from midway down the primaryfloral inflorescence at �1 week after bolting. Wild-type and vti12 cross-sections have a normal appearance with clearly differentiated vascular tissues.

The

Pla

nt C

ell

Arabidopsis VTI1 v-SNARE Family 2895

TGN and the PVC may be different, as reflected by their densitygradient fractionation pattern difference (Zheng et al., 1999; H.Zheng and N. Raikhel, unpublished results). VTI11 and VTI12both may shuttle between the TGN and the PVC. When one ismissing, the other has the chance to form a complex with itsSNARE partner and fulfill its functions. Thus, although VTI1-type proteins may have critical functions in the plant, the loss ofone or the other is tolerated.

VTI12 Is Involved in the Plant Autophagy Pathway

Thus arises the challenge of identifying the particular functionof each VTI1 family protein. Reverse genetics is a powerfulmeans by which to address questions regarding the functionsof VTI12 and its SNARE complex. The vti12 mutant was viable,and the loss of active VTI12 protein caused no obvious defectsin mutant plants grown under normal conditions. When ex-pressed in yeast, VTI12 functions in vacuolar membrane pro-tein transport and the Cvt pathway (Zheng et al., 1999). Be-cause the yeast Cvt and autophagy pathways share manycomponents, we examined whether Arabidopsis VTI12 was in-volved in the plant autophagy pathway. There have been elec-tron microscopic descriptions of plant macroautophagy (Aubertet al., 1996), and recently, two groups have begun to addressthe molecular mechanisms of the plant autophagy pathway(Doelling et al., 2002; Hanaoka et al., 2002). Reports from bothof these groups have demonstrated that the Arabidopsis ge-nome contains orthologs for 12 of the 15 described yeast au-tophagy (or APG) genes. Mutations in two of these genes,APG7 and APG9, have strikingly similar phenotypes. Whengrown under normal conditions, both show accelerated senes-cence after bolting, and apg9 plants are reported to bolt earlierthan wild-type plants. We did not detect any significant differ-ences between wild-type and vti12 plants that were grown un-der nutrient-rich conditions. Both apg7 and apg9 are hyper-sensitive to nutrient-limiting conditions, showing acceleratedchlorosis and premature leaf senescence compared with wild-type plants (Doelling et al., 2002; Hanaoka et al., 2002). Ourown experiments clearly showed that vti12 plants grown undernutrient-limited conditions have a phenotype similar to that ofthe other known Arabidopsis autophagy mutants (Doelling etal., 2002; Hanaoka et al., 2002). The milder vti12 phenotypecompared with apg7 and apg9 can be attributed to the fact thatboth APG7 and APG9 are single-copy genes, whereas there isboth molecular and genetic evidence that VTI11 can compen-sate partially for VTI12 in the vti12 mutant. Vti1p has beenshown to be involved in autophagosome docking and fusion inyeast (Fischer von Mollard and Stevens, 1999; Ishihara et al.,

2001), and it is reasonable to speculate that VTI12 may play asimilar role.

The zig-1 Mutant Is Affected in Auxin Transport

The zig-1 mutant has a distinctive visible phenotype, and wehave some knowledge about VTI11 function at the molecularlevel. VTI11 has been shown, via subcellular fractionation andimmunoelectron microscopy, to colocalize with the vacuolarsorting receptor ELP on the TGN and the PVC (Zheng et al.,1999) and to colocalize with NTPP proteins on the Golgi(Ahmed et al., 2000). Most likely, VTI11 functions as a v-SNAREthat targets ELP and NTPP cargo-containing vesicles from theTGN to the PVC. We extended previous histological studies ofthe zig mutant to include the VTI12/vti12 zig-1/zig-1 line. It iswell established that endodermal cells are the gravity-sensingcells in Arabidopsis (Masson et al., 2002). Starch-filled amylo-plasts sediment to the bottom of the shoot endodermal cells inresponse to the gravity vector and act as statoliths. Expressionof wild-type ZIG driven by the endoderm-specific SCR pro-moter in zig-1 plants complements the abnormal gravitropic re-sponse of the mutant and restores the normal sedimentation ofamyloplasts (Morita et al., 2002). Histological analysis of longi-tudinal thick sections stained with toluidine blue show thatamyloplasts do not sediment in the zig-1 line, as expected(Kato et al., 2002). The individual tissue layers (epidermis, cor-tex, endoderm, and stele) still are recognizable, if slightly lessorganized. In addition to the erratic distribution of amyloplasts,there also was both extreme disorganization of tissue layersand loss of recognizable cell identity in the VTI12/vti12 zig-1/zig-1 line. This loss of tissue identity suggests a specific role forthe VTI11 protein, because we observed no obvious abnormalphenotype in the reciprocal genotype (i.e., vti12/vti12 ZIG/zig-1plants).

Based on the gravitropic phenotype in zig-1 and recent re-ports that vesicle cycling is involved in the asymmetric distribu-tion of auxin efflux carrier proteins (Friml et al., 2003), we hy-pothesized that VTI11 was involved in auxin efflux carriercycling. Geldner et al. (2001) and Friml et al. (2002) have shownthat PIN1 and PIN3 continuously cycle in membrane vesiclesthat travel along the actin cytoskeleton between the plasmamembrane and the endosome. Functional PIN3 is required forgravitropism, and disruption of PIN3 cycling via auxin efflux orvesicle trafficking inhibitors or actin depolymerization agentsresults in a gravitropism defect. zig plants had a significant de-fect in auxin transport, whereas PIN1 localization in the hypo-cotyls and roots of the vti11 and vti12 mutants was not signifi-cantly different from that in the wild type in the vascular tissue.

zig-1 has recognizable vascular bundles, but the phloem is slightly more dispersed than in wild-type or vti12 plants. The enhanced zig line (VTI12/vti12zig-1/zig-1) has clear vascular defects, with an almost continuous phloem ring, fewer interfascicular cells, and small, nonsclerotic xylem cells.(C) Comparison of auxin transport in vti12 and zig-1 mutants. The mean wild-type value was 4583.4 � 145.1 dpm, and transport in roots was not sig-nificantly different from that in the wild type. Values shown (vti12 � 106 � 26.6, zig-1 � 64 � 5.4) are means � percent standard deviations of two in-dependent experiments. The difference for the zig-1 line is significant by Student’s t test (t � 0.050:0.996, P 0.01).

Figure 7. (continued).

The

Pla

nt C

ell

2896 The Plant Cell

The 35% decrease in auxin transport observed in zig-1 may beattributable to the partial mislocalization of PIN1 in cells adja-cent to the vascular tissue or to a deficiency in auxin transportin the improperly developed phloem cells. Evidence that PIN1localization occurs in the cells adjacent to the vascular tissue inyoung seedlings was described recently (Noh et al., 2003).Given the fact that zig plants have tissue identity and organiza-tion phenotypes, perhaps basipetal transport is impeded be-cause the vascular tissue is not developed correctly. Indeed,with respect to vascular patterning, the zig and enhanced-zigphenotypes described in this study are not entirely unlike thosedescribed for wild-type plants treated with the auxin transportinhibitor NPA (Zhong and Ye, 2001). A defect in basipetal auxintransport could, in turn, decrease the absolute amount of auxinmoving through the system (Cambridge and Morris, 1996), butit may not have any effect on the auxin carrier proteins per se.Alternatively, it is possible that VTI11 plays a role in properauxin efflux carrier localization, but because VTI12 can substi-tute for VTI11 in the zig mutant, the delivery of vesicle cargo ismerely slowed and not impeded entirely. Further study will benecessary to resolve this issue. Therefore, VTI11 defines acargo transport pathway that is a major player in the plant’smorphology and development. Although the VTI12-SYP21/22-SYP51/52 SNARE complex rescues the mutant plants from le-thality, it is clear from the distinctive zig phenotype that thiscomplex is not capable of delivering pathway cargo in a com-pletely efficient or timely matter.

The phenomenon that two functionally different SNARE pro-teins can substitute for each other in vivo has been observed inyeast (Liu and Barlowe, 2002), and noncognate SNAREs havebeen shown to form complexes in arbitrary combinations invitro (Fasshauer et al., 1999). It is not difficult to imagine thatthe more closely related VTI11 and VTI12 could be exchangedin their SNARE complexes in vitro. In vivo, the possibility ofsubstitution also exists because both VTI11 and VTI12 are lo-calized on the TGN and the PVC membrane (Zheng et al.,1999). There are a number of reports of mutated genes in Ara-bidopsis causing no apparent abnormal phenotype. Normally,this lack of effect is attributed to gene redundancy. However,when we start to understand the machinery of the plant cellfunction in more detail, “redundant” genes may be found tohave functional differences. The mechanism for structurally re-lated but not fully redundant proteins to substitute for eachother represents just one facet of the flexibility of plant cells.

METHODS

Plant Material and Growth Conditions

Arabidopsis thaliana plants grown on plates were incubated under long-day conditions (16 h of light/8 h of dark) at 22�C. Seeds were surface-sterilized in either 50% bleach and 0.1% Triton X-100 solution or 70%ethanol and 0.05% Triton X-100 solution. Plates contained either 1�

Murashige and Skoog (1962) Minimal Organics medium (Invitrogen,Carlsbad, CA) or, for the rich-growth phase of the autophagy experi-ments, Gamborg’s B-5 medium with 2% sucrose. For the poor-mediumphase of the autophagy experiments, seedlings germinated on Gam-borg’s B-5 medium with 2% sucrose were transferred to a no nitrogen/

no carbon (No N/C) medium as described by Doelling et al. (2002). Plantsgrown on soil were cultivated in a growth facility that features long-dayconditions and maintained at 22�C. Plants were watered with an auto-matic watering system every 3 days.

Wild type in this study is the Columbia-0 (Col-0) ecotype. The vti12mutation is in a Columbia background that also is homozygous for theglabrous (gl) mutation (see below). The Arabidopsis VTI11 mutant, zig/sgr4, is in a Columbia background and was described previously (Katoet al., 2002; Morita et al., 2002). The construct was transformed intoAgrobacterium tumefaciens strain MP90 and then introduced into theAtvti11/zig-1 plants. T1 plants were selected based on resistance to kan-amycin. The genotypes of the T3 siblings used for the experiment wereconfirmed by checking the segregation ratio of kanamycin resistance tokanamycin susceptibility in the T4 generation.

Isolation of the vti12 T-DNA Insertion Mutant

The isolation and identification of vti12 were performed essentially asdescribed by Sanderfoot et al. (2001a). The following primers were usedto screen T-DNA inserted into the VTI12 gene: 5 end gene-specificprimer, 5-TATTTCCTGGACGAGTAATCTTGGTTCTGC-3; 3 end gene-specific primer, 5-TCTGACGTGACAGTGGGTCTCCTGCCTGCG-3;T-DNA left border, 5-CTCATCTAAGCCCCCATTTGGACGTGAATG-3;T-DNA left border nested, 5-TTGCTTTCGCCTATAAATTACGACGGA-TCG-3; T-DNA right border, 5-TGGGAAAACCTGGCGTTACCCAAC-TTAAT-3. The PCR conditions for amplification of the genomic DNA werestandard conditions described in the manufacturer’s recommendations(Invitrogen). For reverse transcriptase–mediated (RT) PCR, total RNAwas extracted from wild-type Col-0 seedlings and heterozygous and ho-mozygous seedlings as described by Zheng et al. (1999). Reverse tran-scription reactions were performed using the Superscript II reverse tran-scriptase system according to recommendations from the manufacturer(Invitrogen). The following gene-specific primers were used to generatefirst-strand cDNAs: for VTI12, 5-GAGCCACGATTACCGATGT-3; forNPSN12, 5-AGTGTAATATGCACCAAACC-3. The forward primers wereas follows: for VTI12, 5-GAAAATGTCACTCTGCATCG-3; for NPSP12,5-GAGCCTGAAATAATCCGGCAGAT-3.

Immunoprecipitation

Antibody characterization is described in the supplemental data online.Twenty grams of 21-day-old liquid-cultured Arabidopsis roots was ho-mogenized on ice in extraction buffer (50 mM Hepes-KOH, pH 6.5, 10 mMpotassium acetate, 100 mM NaCl, 5 mM EDTA, and 0.4 M sucrose) withComplete protease inhibitor tablets (Roche, Mannheim, Germany). Toprepare total membranes, the homogenate was centrifuged at 1,000g for15 min, and the supernatant was subjected to 100,000g ultracentrifuga-tion for 3 h. The pellet was homogenized in TBS (Tris-balanced buffer;0.14 M NaCl, 2.7 mM KCl, and 25 mM Tris, pH 8.0) with miniCompletetablets (Roche). The total membranes were solubilized by adding TritonX-100 to 1%. The solubilized membranes were further cleared by ultra-centrifugation at 100,000g for 30 min. The supernatant was incubatedwith antibodies or preimmune columns for 2 h. After washes with TBST(TBS � 1% Triton X-100), the bound protein was eluted with 0.1 M gly-cine, pH 2.5. The eluted protein was precipitated by adding 10% trichlo-roacetic acid. After two acetone washes, the protein pellet was solubi-lized in 2� Laemmli buffer (125 mM Tris, pH 6.8, 20% glycerol, 2% SDS,2% 2-mercaptoethanol). The protein was separated by SDS-PAGE, anddifferent proteins were detected by protein gel blot analysis.

For immunoprecipitation of VTI12 protein from the wild type and thezig-1 mutant, total membranes were prepared from 10 g of liquid-cul-

The

Pla

nt C

ell

Arabidopsis VTI1 v-SNARE Family 2897

tured wild-type or zig-1 roots. After solubilization with 1% Triton X-100,the membrane preparation was cleared by passing it through a VTI11antibody column to subtract any VTI11 in the solubilized membrane.VTI12 then was immunoprecipitated using purified anti-VTI12 antibodiesas described above.

VTI12 Functional Substitution for VTI11

To create the zig-1 line containing the VTI12 transgenic construct, VTI12cDNA was inserted into a binary vector derived from pBI121 that con-tains the 35S constitutive promoter of Cauliflower mosaic virus. The con-struct was transformed into A. tumefaciens strain MP90 and then intro-duced into the Atvti11/zig-1 plants. T1 plants were selected based onresistance to kanamycin. The genotypes of the T3 siblings used for theexperiment were confirmed by checking the segregation ratio of kana-mycin resistance to kanamycin susceptibility in the T4 generation. Grav-itropic responses were measured as described previously by Fukaki etal. (1996). For immunoprecipitation analysis, membrane fractions weresolubilized with 1% Triton X-100 and then used for SDS-PAGE and pro-tein gel blot analysis.

Autophagy Experiments

All autophagy experiments were performed as described previously byDoelling et al. (2002) and Hanaoka et al. (2002). For the detached-leaf as-say, first and second true leaves were excised from 2-week-old plantsand either placed on filter paper saturated with 3 mM Mes, pH 5.7, orfloated on water in the wells of 12-well tissue culture dishes. Detachedleaves then were placed in the dark at room temperature for the times in-dicated in Results. Detached leaves were used for semiquantitative RT-PCR of senescence-related genes.

For the recovery experiments, seeds were germinated on Gamborg’sB-5 (rich) medium (Invitrogen) with 2% sucrose and then transferred toNo N/C medium when they were 2 weeks old. At the times indicated inResults, plants were transferred back to Gamborg’s B-5 medium andscored as recovered upon the appearance of new green leaves.

For the SEN1 and YSL4 RT-PCR, total RNA was isolated from Col-0,vti12, and zig-1 leaves and stems using the RNEasy kit (Qiagen, Chatsworth,CA) with DNase I treatment. The primers for the SEN1 and YSL4 geneswere the same as those described by Hanaoka et al. (2002). RT-PCRwas performed using the Qiagen One-Step RT-PCR kit. The cycling con-ditions were identical to those described for semiquantitative RT-PCRby Hanaoka et al. (2002), except that 500 ng of starting material wasused, there was an initial 30-min incubation at 50�C for the reverse tran-scription reaction, and a 15-min PCR activation step at 95�C and 26 cy-cles for both genes was used.

Double-Mutant Studies

For the double-mutant studies, plants were grown at 23�C under con-stant white light as described by Fukaki et al. (1996). Arabidopsisecotype Col-0 was used as the wild-type reference. To genotype the F2progeny of the vti12 � zig-1 cross, we isolated genomic DNA samplesusing the quick DNA prep for PCR described by Weigel and Glazebrook(2002). The following primers were used for F2 genotyping of VTI12: 5

VTI12, 5-TTCCGAGAGCTGAAAAAGTGA-3; 3 VTI12, 5-TCGTCT-TCTAATCCAAGCAATG-3; and T-DNA LB (to identify vti12), 5-TTG-CTTTCGCCTATAAATACGACGGATCG-3. The following primers wereused for F2 genotyping of zig-1: ZTAILF, 5-CAAGACTTGCACGGTCAG-AGACAG-3; 3 ZIG, 5-GTTCATCCTCCTCGTCATG-3; and MPO12.P7R,5-GAGCAGGTGCAAGAAGGTCC-3. Seven-week-old plants were used

for wild-type, zig-1/zig-1, and vti12/vti12, and 10- to 12-week-old plantswere used for VTI12/vti12 zig-1/zig-1 and vti12/vti12 ZIG-1/zig-1. Si-liques were collected from the main shoot, and the first four siliques werenot used because embryo lethality was high even in wild-type plants.Clearing of siliques was performed as described by Aida et al. (1997). Si-liques were fixed overnight in ethanol:acetic acid (9:1) solution at roomtemperature. After rehydration in a graded ethanol series (90, 70, 50, and30%) for 20 min each, siliques were cleared with chloral hydrate:glyc-erol:water solution. Siliques were mounted with chloral hydrate:glycerol:water solution on a glass slide, covered with a cover slip, and observedwith a stereomicroscope (SMZ-U; Nikon, Tokyo, Japan) and photo-graphed (COOLPIX; Nikon). Then, the lethality of the seeds was reexam-ined closely and quantitated using a microscope with Nomarski optics(ECLIPSE E800; Nikon).

Histological and Electron Microscopic Analysis

We used the same method of fixation for both histological and electronmicroscopic analyses. Small pieces (�1 cm) were cut from the uprightstems of Arabidopsis plants (wild-type and mutant lines) and fixed undervacuum in a buffer containing 4% formaldehyde, 0.5% glutaraldehyde,0.1 M sucrose, and 0.05 M sodium phosphate. The sections were fixedfor 6 h with a fixative change every hour. We used 0.2-mL tubes andmaintained the growth orientation throughout the fixing process. Afterfixation, the specimens were rinsed in the same buffer and postfixed inOsO4 for 1 h. The dehydrated specimens were embedded in LondonResin White (Electron Microscopy Sciences, Fort Washington, PA).

Thick (1-�m) and thin (80-nm) sections were made with an Ultracut UCTmicrotome (Leica, Solms, Austria) equipped with a diamond knife. For his-tological analysis, sections were stained with toluidine blue O. For electronmicroscopy, sections were stained with 2% uranyl acetate and lead citrate.

Optical images were taken on an MCID Elite workstation (ImagingResearch, Saint Catherines, Ontario, Canada) equipped with a motorizedAxiovert 100S microscope using a Planapo 40�/0.85NA objective (CarlZeiss, Jena, Germany) and a CoolSnap fx monochrome camera (RoperScientific, Trenton, NJ). Tiled-field mapping with automatic focus wasused to automatically visit, capture, and align up to 16 fields of view pertissue section.

Auxin Transport Analysis

Auxin transport assays were conducted on intact light-grown seedlingsas described previously (Noh et al., 2001) with the following exceptions:before assay, 10 seedlings were transferred to vertically discontinuousfilter paper strips saturated in one-quarter-strength Murashige andSkoog (1962) medium and allowed to equilibrate for 2 h. Auxin solutionsused to measure transport were made up in 0.25% agarose containing2% DMSO and 25 mM Mes, pH 5.2. A 0.1-�L microdroplet containing500 nM unlabeled indoleacetic acid and 500 nM 3H-indoleacetic acid(specific activity, 25 Ci/mmol; American Radiochemical, St. Louis, MO)was placed on the apical tips of seedlings using a modified microliterHamilton syringe. Injections were performed using a Narishige microma-nipulator (Narishige Scientific Instrument Lab, Tokyo, Japan) mountedon a Florod LFA microscope-guided x-y laser cutter (Florod Corp., Gar-dena, CA). Seedlings then were incubated in the dark for 5 h. After incu-bation, the hypocotyls and cotyledons were removed, and a 2-mm sec-tion centered on the root-shoot transition zone was harvested, alongwith a 1- to 2-mm basal section of each root. Data shown are means �percent standard deviations for two independent experiments.

Upon request, materials integral to the findings presented in this pub-lication will be made available in a timely manner to all investigators on

The

Pla

nt C

ell

2898 The Plant Cell

similar terms for noncommercial research purposes. To obtain materials,please contact Natasha Raikhel, nraikhel@citrus.ucr.edu.

ACKNOWLEDGMENTS

We thank Klaus Palme for kindly providing PIN1 and PIN3 antibodiesand Anton Sanderfoot, Diane Bassham, and members of the Raikhel lab-oratory for helpful discussions. Jocelyn Brimo prepared the figuresand the manuscript. We are grateful for help from Francis Marty andRoland Douce in interpreting the electron micrographs. Optical imagingand sample preparation were performed at the Center for Plant CellBiology Microscopy Core Facility. Electron microscopy was performedat the Central Facility of Advanced Microscopy and Microanalysis (Uni-versity of California, Riverside). This study was funded by: the NationalScience Foundation (MCB 0296080) to N.R.; the United States-IsraelBinational Science Foundation (1999355) to N.R.; a Grant-in-Aid for Sci-entific Research on Priority Areas from the Ministry of Education, Science,and Culture of Japan (14036222) to M.T.; a Grant-in-Aid for ScientificResearch from the Ministry of Education, Science, and Culture of Japan(13440241) to M.T.; the Japan Society for the Promotion of Science(15570035) to M.T.M.; the U.S. Department of Agriculture National Re-search Initiative Grant 2002-35304-12290 to A.M.

Received August 18, 2003; accepted October 8, 2003.

REFERENCES

Abeliovich, H., Darsow, T., and Emr, S.D. (1999). Cytoplasm to vacu-ole trafficking of aminopeptidase I requires a t-SNARE-Sec1p com-plex composed of Tlg2p and Vps45p. EMBO J. 18, 6005–6016.

Ahmed, S.U., Rojo, E., Kovaleva, V., Venkataraman, S., Dombrowski,J.E., Matsuoka, K., and Raikhel, N.V. (2000). The plant vacuolarsorting receptor AtELP is involved in transport of NH2-terminalpropeptide-containing vacuolar proteins in Arabidopsis thaliana. J.Cell Biol. 149, 1335–1344.

Aida, M., Ishida, T., Fukaki, H., Fujisawa, H., and Tasaka, M. (1997).Genes involved in organ separation in Arabidopsis: An analysis of thecup-shaped cotyledon mutant. Plant Cell 6, 841–857.

Aubert, S., Gout, E., Bligny, R., Marty-Mazars, D., Barrieu, F.,Alabouvette, J., Marty, F., and Douce, R. (1996). Ultrastructural andbiochemical characterization of autophagy in higher plant cells sub-jected to carbon deprivation: Control by the supply of mitochondriawith respiratory substrates. J. Cell Biol. 133, 1251–1263.

Bassham, D.C., Sanderfoot, A.A., Kovaleva, V., Zheng, H., andRaikhel, N.V. (2000). AtVPS45 complex formation at the trans-Golginetwork. Mol. Biol. Cell 11, 2251–2265.

Cambridge, A.P., and Morris, D.A. (1996). Transfer of exogenous auxinfrom the phloem to the polar auxin transport pathway in pea (Pisumsativum L.). Planta 199, 583–588.

Doelling, J.H., Walker, J.M., Friedman, E.M., Thompson, A.R., andVierstra, R.D. (2002). The APG8/12-activating enzyme APG7 is re-quired for proper nutrient recycling and senescence in Arabidopsisthaliana. J. Biol. Chem. 277, 33105–33114.

Fasshauer, D., Antonion, W., Margittai, M., Pabst, S., and Jahn, R.(1999). Mixed and non-cognate SNARE complexes. J. Biol. Chem.274, 15440–15446.

Fischer von Mollard, G., and Stevens, T.H. (1999). The Saccharomy-ces cerevisiae v-SNARE Vti1p is required for multiple membranetransport pathways to the vacuole. Mol. Biol. Cell 10, 1719–1732.

Friml, J., Wisniewska, J., Benkoa, E., Mendgen, K., and Palme, K.

(2002). Lateral relocation of auxin efflux regulator PIN3 mediates tro-pism in Arabidopsis. Nature 415, 806–809.

Fukaki, H., Fujisawa, H., and Tasaka, M. (1996). SGR1, SGR2, SGR3:Novel genetic loci involved in shoot gravitropism in Arabidopsisthaliana. Plant Physiol. 110, 933–943.

Fukaki, H., Wysocka-Diller, J., Kato, T., Fujisawa, H., Benfey, P., andTasaka, M. (1998). Genetic evidence that the endodermis is essentialfor shoot gravitropism in Arabidopsis thaliana. Plant J. 14, 425–430.

Geldner, N., Friml, J., York-Dieter, S., Jürgens, G., and Palme, K.(2001). Auxin transport inhibitors block PIN1 cycling and vesicle traf-ficking. Nature 413, 425–428.

Hanaoka, H., Noda, T., Shirano, Y., Kato, T., Hayashi, H., Shibata, D.,Tabata, S., and Ohsumi, Y. (2002). Leaf senescence and starvation-induced chlorosis are accelerated by the disruption of an Arabidopsisautophagy gene. Plant Physiol. 129, 1181–1193.

Ishihara, N., Hamasaki, M., Yokota, S., Suzuki, K., Kamada, Y., Kihara,A., Yoshimori, T., Noda, T., and Ohsumi, Y. (2001). Autophagosomerequires specific early Sec proteins for its formation and NSF/SNAREfor vacuolar fusion. Mol. Biol. Cell 12, 3690–3702.

Kato, T., Morita, M.I., Fukaki, H., Yoshiro, Y., Uehara, M., Nihama,M., and Tasaka, M. (2002). SGR2, a phospholipase-like protein, andZIG/SGR4, a SNARE, are involved in the shoot gravitropism of Arabi-dopsis. Plant Cell 14, 33–46.

Liu, Y., and Barlowe, C. (2002). Analysis of Sec22p in endoplasmicreticulum/Golgi transport reveals cellular redundancy in SNARE pro-tein function. Mol. Biol. Cell 13, 3314–3324.

Masson, P.H., Tasaka, M., Morita, M.T., Guan, C., Chen, R., andBoonsirichai, K. (2002). Arabidopsis thaliana: A model for the studyof root and shoot gravitropism. In The Arabidopsis Book, C.R. Somervilleand E.M. Meyerowitz, eds (Rockville, MD: American Society of PlantBiologists), doi/10.1199/tab.0043, http://www.aspb.org/publications/arabidopsis/.

Morita, M.T., Kato, T., Nagafusa, K., Saito, C., Ueda, T., Nakano, A.,and Tasaka, M. (2002). Involvement of the vacuoles of the endoder-mis in the early process of shoot gravitropism in Arabidopsis. PlantCell 14, 47–56.

Murashige, T., and Skoog, F. (1962). A revised medium for rapid growthand bioassays with tobacco tissue culture. Physiol. Plant. 15, 473–497.

Noh, B., Bandyopadhyay, A., Peer, W.A., Spalding, E.P., and Murphy,A.S. (2003). Enhanced gravi- and phototropism in plant mdr mutantsmis-localizing the auxin efflux protein PIN1. Nature 423, 999–1002.

Noh, B., Murphy, A.S., and Spalding, E.P. (2001). Multidrug resistance-like genes of Arabidopsis required for auxin transport and auxin-medi-ated development. Plant Cell 13, 2441–2454.

Reggiori, F., and Klionsky, D.J. (2002). Autophagy in the eukaryoticcell. Eukaryot. Cell 1, 11–21.

Rojo, E., Gillmor, C.S., Kovaleva, V., Somerville, C.R., and Raikhel,N.V. (2001). VACUOLESS1 is an essential gene required for vacuoleformation and morphogenesis in Arabidopsis. Dev. Cell 1, 1–20.

Rojo, E., Zouhar, J., Kovaleva, V., Hong, S., and Raikhel, N.V.(2003). The AtC-VPS protein complex is localized to the tonoplastand the prevacuolar compartment in Arabidopsis. Mol. Biol. Cell 14,361–369.

Sanderfoot, A.A., Assad, F.F., and Raikhel, N.V. (2000). The Arabidop-sis genome: An abundance of soluble N-ethylmaleimide-sensitivefactor adaptor protein receptors. Plant Physiol. 124, 1558–1569.

Sanderfoot, A.A., Pilgrim, M., Adam, L., and Raikhel, N.V. (2001a).Disruption of individual members of Arabidopsis syntaxin gene fami-lies indicates each has essential functions. Plant Cell 12, 659–666.

Sanderfoot, A.A., Kovaleva, V., Bassham, D.C., and Raikhel, N.V.(2001b). Interactions between syntaxins identify at least five SNAREcomplexes within the Golgi/prevacuolar system of the Arabidopsiscell. Mol. Biol. Cell 12, 3733–3743.

The

Pla

nt C

ell

Arabidopsis VTI1 v-SNARE Family 2899

Sato, T.K., Rehling, P., Peterson, M.R., and Emr, S.D. (2000). Class CVps protein complex regulates vacuolar SNARE pairing and is re-quired for vesicle docking/fusion. Mol. Cell 6, 661–671.

Vitale, A., and Raikhel, N.V. (1999). What do proteins need to reach dif-ferent vacuoles? Trends Plant Sci. 4, 149–155.

Weigel, D., and Glazebrook, J. (2002). Arabidopsis: A Laboratory Man-ual. (Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press).

Xu, Y., Wong, S.H., Tang, B.L., Subramaniam, V.N., Xhang, T., andHong, W. (1998). A 29-kilodalton Golgi soluble N-ethylmaleimide-

sensitive factor attachment protein receptor (Vti1-rp2) implicated inprotein trafficking in the secretory pathway. J. Biol. Chem. 273,21783–21789.

Zheng, H., Fischer von Mollard, G., Kovaleva, V., Stevens, T.H., andRaikhel, N.V. (1999). The plant vesicle-associated SNARE AtVTI1alikely mediates vesicle transport from the trans-Golgi network to theprevacuolar compartment. Mol. Biol. Cell 10, 2251–2264.

Zhong, R., and Ye, Z.-H. (2001). Alteration in polar auxin transport inthe Arabidopsis ifl1 mutants. Plant Physiol. 126, 549–563.

DOI 10.1105/tpc.016121; originally published online November 13, 2003; 2003;15;2885-2899Plant Cell

RaikhelBandyopadhyay, Valentina Kovaleva, David Carter, Angus Murphy, Masao Tasaka and Natasha Marci Surpin, Haiyan Zheng, Miyo T. Morita, Cheiko Saito, Emily Avila, Joshua J. Blakeslee, Anindita

Protein Transport PathwaysThe VTI Family of SNARE Proteins Is Necessary for Plant Viability and Mediates Different

 This information is current as of December 14, 2020

 

Supplemental Data /content/suppl/2003/12/01/15.12.2885.DC1.html

References /content/15/12/2885.full.html#ref-list-1

This article cites 32 articles, 21 of which can be accessed free at:

Permissions https://www.copyright.com/ccc/openurl.do?sid=pd_hw1532298X&issn=1532298X&WT.mc_id=pd_hw1532298X

eTOCs http://www.plantcell.org/cgi/alerts/ctmain

Sign up for eTOCs at:

CiteTrack Alerts http://www.plantcell.org/cgi/alerts/ctmain

Sign up for CiteTrack Alerts at:

Subscription Information http://www.aspb.org/publications/subscriptions.cfm

is available at:Plant Physiology and The Plant CellSubscription Information for

ADVANCING THE SCIENCE OF PLANT BIOLOGY © American Society of Plant Biologists