RESEARCH ARTICLE 3001 Cell cycle-independent expression of the Arabidopsis cytokinesis-specific syntaxin KNOLLE results in mistargeting to the plasma membrane and is not sufficient for cytokinesis

Axel Völker1, York-Dieter Stierhof2 and Gerd Jürgens1,* 1Zentrum für Molekularbiologie der Pflanzen, Entwicklungsgenetik, Universität Tübingen, Auf der Morgenstelle 1, D-72076 Tübingen, Germany 2Zentrum für Molekularbiologie der Pflanzen, Mikroskopie, Universität Tübingen, Auf der Morgenstelle 1, D-72076 Tübingen, Germany *Author for correspondence (e-mail: [email protected])

Accepted 21 May 2001 Journal of Cell Science 114, 3001-3012 (2001) © The Company of Biologists Ltd

SUMMARY

The Arabidopsis KNOLLE gene encodes a cytokinesis- cells, whereas no mislocalisation was observed in specific syntaxin that localises to the plane of division and proliferating cells. By comparative in situ hybridisation to mediates cell-plate formation. KNOLLE mRNA and embryo sections, the 35S promoter yielded, relative to the protein expression is tightly regulated during the cell cycle. endogenous KNOLLE promoter, low levels of KNOLLE To explore the significance of this regulation, we expressed mRNA accumulation in proliferating cells that were KNOLLE protein under the control of two constitutive insufficient to rescue cytokinesis-defective knolle mutant promoters, the flower-specific AP3 and the cauliflower embryos. Our results suggest that in wild type, strong mosaic virus 35S promoter. The transgenic plants expression of KNOLLE protein during M phase is developed normally, although KNOLLE mRNA and protein necessary to ensure efficient during accumulated to high levels in non-proliferating cells cytokinesis. and protein was incorporated into membranes. Immunolocalisation studies in transgenic seedling roots revealed mistargeting of KNOLLE protein to the plasma Key words: Arabidopsis, Cytokinesis, Syntaxin, KNOLLE, membrane in tip-growing root hairs and in expanding root Expression

INTRODUCTION KNOLLE and KEULE, result in cytokinesis defects, such as enlarged cells with incomplete cell walls and more than one Cytokinesis partitions the cytoplasm of the dividing cell, which nucleus (Lukowitz et al., 1996; Assaad et al., 1996; Nacry et requires targeting of membrane vesicles to the plane of al., 2000). KNOLLE encodes a cytokinesis-specific syntaxin division. In yeast and animal cells, a contractile actomyosin (Lukowitz et al., 1996; Lauber et al., 1997). KEULE is a ring supports ingrowth of the existing plasma membrane, and member of the Sec1 family of syntaxin-binding proteins that this cleavage furrow is expanded by the fusion of membrane interacts with KNOLLE in vitro and in vivo, and mutations in vesicles that are delivered along furrow microtubule arrays both genes result in the accumulation of unfused cytokinetic (reviewed by Robinson and Spudich, 2000; Straight and Field, vesicles (Assaad et al., 2000; Lauber et al., 1997; Waizenegger 2000). By contrast, plant cells form the partitioning plasma et al., 2000). Whereas the KEULE gene appears to be expressed membrane de novo from the centre to the periphery of the cell in both proliferating and non-proliferating cells (Assaad et al., (reviewed by Staehelin and Hepler, 1996; Heese et al., 1998). 2000), the expression of KNOLLE is tightly regulated during Golgi-derived vesicles are transported along the microtubules the cell cycle. KNOLLE mRNA accumulates transiently in of a plant-specific cytoskeletal array, the phragmoplast, to the proliferating cells, giving a patchy pattern that reflects plane of cell division where they fuse with one another to form asynchrony of cell division in the embryo (Lukowitz et al., a transient membrane-bounded compartment, the cell plate, 1996). KNOLLE protein accumulates only during M phase, which matures into a cell wall with flanking plasma initially in patches presumed to represent Golgi stacks, then membranes. The lateral expansion of the cell plate is mediated localises to the forming cell plate during telophase and by the transformation of the phragmoplast from a compact disappears at the end of cytokinesis (Lauber et al., 1997). The array into a widening hollow cylindrical structure that delivers tight regulation of KNOLLE expression is reminiscent of the additional vesicles to the growing edge of the cell plate until synthesis and degradation of mitotic cyclins (Ito, 2000). the latter fuses with the parental plasma membrane (Samuels KNOLLE syntaxin appears to be involved in all sporophytic et al., 1995). Thus, plant cytokinesis is a special case of vesicle cell divisions as well as in endosperm cellularisation (Lauber trafficking and fusion. et al., 1997). Mutations in several genes of Arabidopsis, including Syntaxins are components of SNARE complexes that play 3002 JOURNAL OF CELL SCIENCE 114 (16) an important role in membrane fusion events (reviewed by MATERIALS AND METHODS Jahn and Südhof, 1999). The SNARE core complex consists of three or four proteins that form a four-helix bundle: a Plant material, growth conditions and in vitro culture bipartite t-SNARE on the target membrane, which consists of Arabidopsis thaliana ecotypes Wassilewskija (WS), Landsberg/ a syntaxin and a SNAP25 protein or two t-SNARE light-chain Niederzenz (Ler/Nd) heterozygous for the knolle mutation X37-2 proteins, interacts with the v-SNARE on the (Lukowitz et al., 1996) and the EMS-induced knolle allele UU1319 vesicle membrane (Clague and Herrmann, 2000). There are (kindly provided by U. Mayer) were grown on soil at 18°C, as numerous members of each SNARE protein family in yeast, described previously (Mayer et al., 1991). The Arabidopsis cell suspension culture (Fuerst et al., 1996) was a gift from the John Innes animals and plants that have been implicated in diverse vesicle Centre (Norwich, UK). The in vitro culture was done in petri dishes trafficking pathways between membrane compartments (for containing 1% Select Agar (Gibco BRL, Karlsruhe, Germany) and reviews on plant SNAREs, see Blatt et al., 1999; Sanderfoot 0.5× or 1× Murashige and Skoog (MS) salts (Ducheva, Haarlem, The et al., 2000). In general, syntaxins and Netherlands) at 18°C under constant illumination. To induce root hair involved in a particular pathway appear more closely related formation, 1% sucrose was added. To determine the organisation of to functional counterparts in different organisms than to vacuoles in root hairs, seedlings were grown on 0.5× MS medium family members involved in a different pathway within the containing no, 1% or 3% sucrose. Callus induction of knolle-X37-2 same organism. The original SNARE hypothesis postulated mutant seedlings carrying the 35S::KN transgene was carried out that specific pairs of cognate syntaxins and synaptobrevins with modified callus-inducing (1 mg/l 2.4D, 0.25 mg/l kinetin) or provide specificity to vesicle trafficking (Söllner et al., 1993a; shoot-inducing (0.5 mg/l NAA, 0.25 mg/l kinetin) media (Soni et al., 1995). Söllner et al., 1993b). This idea was challenged in recent in vitro interaction studies that provided evidence for Plant transformation and selection of transgenic plants promiscuity among interacting SNARE partners (Fasshauer et WS and knolle-X37-2 (Ler/Nd) heterozygous plants were transformed al., 1999). However, thorough analyses of yeast SNARE by a modified transformation protocol (Bechtold and Pelletier, 1998; interactions in liposome assays have indicated a high degree Clough and Bent, 1998), using a combination of vacuum infiltration of specificity of interaction between syntaxins and and additional dip transformation one week later. One hundred to 150 synaptobrevins (Fukuda et al., 2000; McNew et al., 2000; plants were transformed with an Agrobacterium GV3101 culture bearing the desired transgene. Infiltration medium consisted of 0.5× Parlati et al., 2000). × µ KNOLLE is a distant member of the plasma membrane MS salts, 1 Gamborg B5 vitamins (Ducheva), 5% sucrose, 0.044 M benzyl aminopurine (Sigma) pH 5.7 with KOH, and 0.005% SILWET subgroup of the syntaxin family but has no close counterpart L77 (Osi Specialities). T1 seeds were bulk-harvested (each seed among yeast or animal syntaxins (Lukowitz et al., 1996; representing a single transformation event; Bechtold et al., 2000; Sanderfoot et al., 2000). However, syntaxins with analogous Desfeux et al., 2000; Ye et al., 1999), sown on soil and selected for roles in membrane fusion during cellularisation or cytokinesis transformants by spraying BASTA® (183 g/l Glufosinate, AgrEvo™; have been described in animals. The Drosophila syntaxin 1 Düsseldorf, Germany; 1:1000) twice. BASTA®-resistant plants were gene is required for cellularisation of the blastoderm embryo, genotyped for KNOLLE by PCR with the primers X37-2C and X37- as well as for neural development (Burgess et al., 1997). 2D (Lukowitz et al., 1996), which amplify a 0.7 kb fragment from Likewise, the Caenorhabditis syn-4 gene is involved in X37-2 and a 1.7 kb fragment from wild type. Seeds containing knolle mutant embryos are shrunken and darker than wild-type seeds. For embryo cleavage divisions but also plays a role in nuclear × membrane reformation (Jantsch-Plunger and Glotzer, 1999). confirmation of the genotype, mutant seeds were germinated on 0.5 MS salts, 1% Select Agar plates, and seedlings were examined for the In contrast to the other two syntaxins, KNOLLE is required knolle mutant phenotype. only for de novo formation of the partitioning plasma Plants heterozygous for kn-X37-2 were transformed with the membrane during cytokinesis, and its expression is tightly KNRescue construct (see Fig. 1). Selfing of the T0 plants gave three regulated during the cell cycle, suggesting a unique role in distinguishable genotypes of BASTA®-resistant T1 progeny bearing cytokinesis. the KN transgene: (1) KN/KN, (2) kn/KN and (3) kn/kn. PCR with We have addressed the biological significance of the tight KNOLLE-specific primers amplified the kn-X37-2 fragment from the regulation of KNOLLE expression by replacing the genotypes (2) and (3) (Lukowitz et al., 1996). Selfing of T1 plants endogenous 5′ regulatory region with promoters that are active with genotype (2) or (3) produced 6.25% or 25% knolle mutant T2 in both proliferating and non-proliferating cells. The seeds, respectively. Reduction of phenotypically mutant seeds from transgenic plants were phenotypically normal, although 25% to 6.25% for genotype (2) indicated complementation. One-third of the Basta-resistant T2 plants derived from genotype (3) were KNOLLE protein accumulated strongly in non-proliferating homozygous for the transgene and produced only phenotypically cells and was mistargeted to the plasma membrane. normal seeds, although the embryos were homozygous for the kn- Conversely, the KNOLLE transgene did not rescue knolle X37-2 mutant allele (T3 generation). mutant embryos, which correlated with low-level accumulation of mRNA from the KNOLLE transgene in Molecular biology proliferating embryonic tissue, when compared with the Constructs for plant transformation were introduced into pBar vectors. activity of the endogenous KNOLLE gene. Our observations pBarA (AJ251013) was used for AP3::KN misexpression and for the suggest that the tight regulation of KNOLLE expression meets rescue of the knolle mutant phenotype by a KNOLLE SacI-SnaI two opposing requirements. First, the KNOLLE gene must be genomic fragment. pBar-35S (AJ251014) was used for 35S::KN misexpression. pBar vectors were a gift from G. Cardon (MPI, strongly expressed to produce sufficient KNOLLE protein Cologne, Germany). PCR was carried out according to standard during M phase for the efficient execution of cytokinetic procedures using TaqPlus Precision™ (Stratagene, La Jolla, CA), vesicle fusion. Second, degradation of KNOLLE mRNA and Expand High Fidelity™ (Roche Mannheim, Germany) and Taq DNA protein prevents the accumulation of large quantities of Polymerase™ (Roche). The AP3 promoter fragment was amplified by useless molecules. PCR using pD1075:AP3 (−650Æ −1, a gift from T. Jack) with the Regulation of KNOLLE syntaxin 3003 forward primer AP3-98a (5′AATTCTAGACAAGGATCTTTAGTT- AAGGC 3) introducing a XbaI 5′ restriction site and with the reverse primers AP3-46a (5′ATACTGCAGATTTGGTGGAGAGGACAAG 3′) and AP3-47: (5′ ATACTGCAGGAAGAGATTTGGTGGAGAG- GACAAG 3′) introducing a consensus transcription start (Joshi, 1987) and a 3′ Pst1 restriction site. The KNOLLE fragments KN1 (−217 to +1500) and KN2 (−4 to +1500) were PCR amplified with either forward primer KNstart1 (5′TTTCTGCAGCTTTCTCTCATCTCAC- A AATC 3′) or KNstart2 (5′ATACTGCAGAAGATGAACGACTTG- ATGACG 3′), introducing a PstI restriction site at the 5′ end, and reverse primer KNstop (5′ATAGAATTCATGACCTTGTTCCAGAG- ATTG 3′), introducing an EcoRI cloning site at the 3′ end. AP3::KN1 (KNOLLE coding sequence with intron; see Fig. 1) and AP3::KN2 (coding sequence without intron) were used for KNOLLE misexpression, which gave essentially the same results (data not shown). XbaI-AP3::KN-EcoRI was cloned into pBarA. p35S::KN constructs were obtained by restriction digest of Bluescript SK::KN2 with XhoI, introducing 40 nucleotides non-coding sequence 5′ to the KNOLLE translational start. The 3′ end of the KN2 fragment was a XbaI restriction site within the Bluescript SK vector, introducing another 30 nucleotides. This fragment was cloned into pBar-35S SmaI Fig. 1. Constructs for KNOLLE expression in transgenic plants. The + XbaI. A genomic SnaI and SacI 4.75 kb fragment containing constructs were cloned into the multiple cloning site of the pBar KNOLLE was directly subcloned into pBarA SacI, SmaI. The transformation vectors indicated (see Materials and Methods). constructs were confirmed by sequencing and transformed into Numbers indicate distances (in bp) from ATG (+1) of the KNOLLE Agrobacterium tumefaciens strain GV3101 (GV3101 + pM90: gift gene. The arrow above the gene indicates the orientation of from G. Cardon, MPI Cologne, Germany). The p35S::KN transgene transcription. KNRescue (top), genomic DNA fragment containing was re-amplified from transgenic plants by using the forward primer the KNOLLE gene and fragments of the adjacent genes. Thiored., 35SPromoter2 (5′ACGCACAATCCCACTATCCTT 3′) and the gene encoding thioredoxin-1-like protein (AF144387); VI, reverse primer 35STerminator2 (5′AAGAACCCTAATTCCCTTATC- hypothetical gene possibly encoding a protein similar to violaxanthin TGG 3′) close to the multiple cloning site of pBar-35S, and de-epoxidase (CAB59211.1); AP3::KN1 (middle), AP3 promoter sequenced. The 35S::GUS reporter construct from the pBIC20 cosmid (pAP3) fused to a genomic DNA fragment with KNOLLE intron; vector (Meyer et al., 1994) was used to monitor 35S promoter activity p35S::KN (bottom), CaMV 35S promoter (p35S) of the pBAR-35S in embryos and seedlings. Molecular work was carried out according vector fused to a genomic DNA fragment without KNOLLE intron. to standard protocols (Sambrook et al., 1989). Restriction enzymes Restriction sites: E, EcoRI; P, PstI; Sc, SacI; Sn, SnaI; X, XbaI. were purchased from NEB (New England Biolabs, Hitchin, UK); polyA, poly-adenylation site; L, left border of the T-DNA; R, right synthetic oligonucleotides were from ARK Scientific (Sigma, border of the T-DNA. Germany).

In situ hybridisation Protein concentration was measured using a Bradford assay, Sample preparation and in situ hybridisation of KNOLLE antisense and equal amounts of protein were loaded onto the gel. riboprobes transcribed in vitro were carried out according to Mayer Immunolocalisation was carried out with rabbit anti-KNOLLE et al. (Mayer et al., 1998), using a 300 bp KNOLLE fragment from antiserum diluted 1:2,000 or with mouse monoclonal anti-plasma the 5′ end of the coding region (Lukowitz et al., 1996). Paraffin- membrane H+-ATPase antibody diluted 1:500. Root tissue was fixed embedded material was cut into 8 µm sections for embryos and 12.5- with 4% paraformaldehyde (Sigma) in MTSB (pH 7.0) for 0.5 hours. 17.5 µm for seedlings. Digoxigenin-labelled probes were detected Goat anti-rabbit secondary antibody was coupled to Cy3™ (Dianova, with Boehringer anti-Dig FAB (Roche) coupled to alkaline Hamburg, Germany) or Alexa-m488 (Molecular Probes, Eugene OR, phosphatase. Western Blue® alkaline phosphatase (Promega, USA), goat anti-mouse secondary antibody was coupled to Cy3 Madison, USA) colour reaction was carried out for 1-4 days. Samples (Dianova). Instead of MTSB, PBS (pH 7.2) was used in all steps after were mounted in 50% glycerol. Images were taken with a Zeiss fixation of the plant material. The primary antibody was incubated Axiophot (Carl Zeiss Inc., Thornwood, NY), using a Nikon Coolpix for 3 hours at 37°C after blocking for 1 hour with 1% BSA in PBS, 990 digital camera with 3.34 Mio pixels. the secondary antibody was incubated for 3 hours at 37°C. Nuclei were stained with 1 mg/ml DAPI. After mounting in Citifluor (Agar, Immunoblotting and whole-mount immunofluorescence Amersham), specimens were analysed with a Zeiss Axiophot microscopy epifluorescence microscope equipped with a Nikon Digital camera Immunoblotting and immunolocalisation were as previously (3.34 Mio pixels) or with a Leica confocal laser scanning microscope described (Lauber et al., 1997; Steinmann et al., 1999). For (CLSM) with Leica TCS-NT software. The CLSM standard objective separation of proteins, 12 to 15% polyacrylamide SDS (Sigma) gels was 63× (water immersion), scanning was carried out with electronic were used. Protein extraction was achieved by grinding plant material magnification. with sand and boiling in 1× Laemmli buffer, except for cell fractionation experiments. Western analysis with rabbit anti- Histochemical GUS staining KNOLLE antiserum was performed as previously described (Lauber GUS staining was as previously described (Sundaresan et al., 1995). et al., 1997). Protein concentrations were estimated by Coomassie Plant tissue was incubated in the detection solution (500 mM NaPO4 Blue staining. Cell fractionation was done as previously described buffer pH 7.0, 500 mM EDTA pH 8.0, 150 mM potassium (Lauber et al., 1997). S10 was the supernatant of a 10,000 g pre- ferrocyanide (K4Fe(CN)6 3H2O), 5% Triton X100, 40 mM X-Gluc in centrifugation, S100 and P100 were the supernatant and the pellet of dimethylformamide) in the dark at 37°C for several hours until the a 100,000 g centrifugation for 12 hours. Integral membrane proteins blue colour became apparent. After transfer to water, the stained tissue were solubilised with Triton X-100 (Sigma; Lauber et al., 1997). was examined by bright-field light microscopy. 3004 JOURNAL OF CELL SCIENCE 114 (16)

Fig. 2. Protein blots of extracts from KNOLLE transgenic plants. Protein extracts were separated by SDS-PAGE and analysed by immunoblotting with anti-KN antiserum (see Materials and Methods). (A) Misexpression of KNOLLE from the flower-specific AP3 promoter. Lanes 1-4: total protein extracts of wild-type (wt) flowers (lane 1) and AP3::KN1 transgenic petals (lanes 2-4). Lanes 5-7: cell fractionation of pooled extracts (lanes 2-4). (B) Organ distribution of KNOLLE protein misexpressed from the CaMV 35S promoter. 35S, T3 transgenic plants homozygous for 35S::KN; wt, wild-type control. (C) Quantitative analysis of KNOLLE expression in 20 days old seedlings grown on different media. Lanes 1-3: 0.5× MS salts with 0%, 1%, 3% sucrose. Lanes 4-6: 1× MS salts with 0%, 1%, 3% sucrose. Upper panel: similar amounts of total protein were loaded. Lower panel: 35S::KN samples 1 and 3 were diluted 1:100 and 1:500; wild-type (wt) samples 2, 3 and 5 were not diluted. (D) Membrane integration properties of KNOLLE protein from T3 generation plants homozygous for the 35S::KN transgene. Protein extracts from leaves of 35S::KN plants (35S, top) or from flowers and siliques of wild-type control (wt, middle) were differentially centrifuged. The P100 fraction (lane 3) was resuspended with or without Triton X-100 followed by a second 100,000 g centrifugation (lanes 4-7). Plants heterozygous for the knolle mutant allele UU1319 (bottom) express normal KN protein (arrow) and a truncated KN protein without the membrane anchor that was not pelleted by 100,000 g centrifugation (arrowhead). Arrow: position of the 34 kDa KN protein; S10, supernatant of 10,000 g centrifugation; S100, supernatant; P100, pellet of 100,000 g centrifugation.

Immunolocalisation on cryosections and EM analysis Roots of 5-day-old plants grown on 0.5× MS salts, 1% sucrose and 1% Select Agar were fixed with 4% formaldehyde in MTSB (pH 7.0) for 60 minutes and embedded in 1% agarose. After infiltration with 20% (w/v) polyvinylpyrrolidone (MW 10,000, PVP-10; Sigma) in 1.8 M sucrose (Tokuyasu, 1989) and freezing in liquid nitrogen, cells were sectioned at –85°C (400 nm, semithin) or at –100°C (100 nm, ultrathin) with a Leica Ultracut S/FCS. Cryosections were transferred to poly-L- lysine-coated (Sigma) coverslips for immunofluorescence or collected on electron microscopy copper grids. After blocking (1% skim milk/0.5% BSA in PBS, pH 7.2), labelling with rabbit anti-KNOLLE antiserum (1:1000) was performed for 60 minutes followed by incubation with Cy3-conjugated goat anti-rabbit secondary antibody (Dianova) or protein A-15 nm gold for 60 minutes (Slot and Geuze, 1985). For immunofluorescence, sections were stained with DAPI and embedded in Mowiol 4.88 (Hoechst, Frankfurt/Main, Germany; Rodriguez and Deinhardt, 1960) containing DABCO (25 mg/ml; Sigma; Langanger et al., 1983). For electron microscopy, grids were stained with uranyl acetate and embedded in methyl cellulose (Sigma) according to Griffiths (Griffiths, 1993). Cy3-labeled cryosections were Staining of root hair membranes with fluorescent dyes viewed with a Zeiss Axioplan, gold-labelled cryosections with a FM4-64 and FM1-43 Philips 201 electron microscope at 60 kV accelerating voltage. For Wild-type and transgenic seedlings grown on agar plates were stained ultrastructural analysis, 5-day-old roots were cryofixed in liquid with the lipophilic steryl-dyes FM4-64 and FM1-43 (1 mg/mL; propane, freeze-substituted in acetone containing 0.5% glutaraldehyde Molecular Probes, Eugene OR, USA). Both dyes show a Stoke shift and 0.5% osmium tetroxide and embedded in Spurr. Ultrathin sections when integrated into membranes, depending on the membrane were stained with uranyl acetate and lead citrate. composition. Whole seedlings were stained alive or after fixation with 4% paraformaldehyde (in MTSB) for 5 minutes and then destained Processing of digital pictures for 15 minutes in water. For confocal laser scanning microscopy, All images shown were processed with Photoshop 5.0 and Illustrator FM4-64 and FM1-43 were excited by the laser at 568 nm and 488 8.0 (Adobe Mountain View, CA). nm, respectively, and emitted light at >600 nm and 500-530 nm plus 580-630 nm. In live root hair cells, FM4-64 stains predominantly endomembranes and also the plasma membrane, whereas the less RESULTS hydrophobic dye FM1-43 preferentially stains the plasma membrane. FM4-64 endomembrane labelling correlated with the number and size of vacuolar structures in root hairs grown on 0.5× MS medium by Ectopic expression of KNOLLE protein has no varying sucrose content: increasing sucrose concentration (0% to 3%) phenotypic effect resulted in enlarged but fewer vacuoles. Because of the tight regulation of KNOLLE expression in Regulation of KNOLLE syntaxin 3005

Fig. 3. KNOLLE protein immunolocalisation in 35S::KN seedling roots. Seedlings were grown on 1% agar with 0.5× MS salts and 1% sucrose. Dark-field images of wild-type root tip (B) and root differentiation zone (E) are shown as reference for the epifluorescence images of roots homozygous for the 35S::KN transgene (A,D,G) and of wild-type control roots (C,F,H). Broken lines delineate root tips (A,C) and root hairs and lateral surface of root (D,F). (A,C) Root tip with KN-labelled cell plates (arrowheads). KN is also expressed in the non- proliferating cells of the central cylinder in the transgenic root (A; see also D) but not in the wild- type root (C). (D,G,F,H) Differentiation zone of the root. KN is expressed in the central cylinder (cr) and in root hair tips (arrowheads) of transgenic roots (D,G) but not in wild-type control roots (F,H). (D,F) Brightness of images increased; (F) maximum brightness to visualise the wild-type root in the absence of the KN signal. (G,H) Original images; (H) wild-type root is not visible. (I- L) GUS staining (2 hours, blue colour) to visualise 35S promoter activity in 6-day-old 35S::GUS transgenic seedlings. (I) Seedling with high GUS activity in root tip, root hairs and tips of the cotyledons. Boxed areas are magnified in J (young root hair), K (central part of the root) and L (root tip). Scale bar: 50 µm.

KNOLLE, an integral (Lauber et al., 1997; see Fig. 2D). To determine possible effects of excessive amounts of KNOLLE protein in other organs and developmental stages, we generated 35S::KN transgenic plants (Fig. 1; Benfey et al., 1989). The 35S::KN transgenic plants were also morphologically normal, although misexpressed KNOLLE protein was detected at very high levels in protein extracts from mature leaves and stem, both of which consist of differentiated cells that normally contain no KNOLLE protein (Fig. 2B). By semi-quantitative analysis of protein extracts, 35S::KN transgenic seedlings contained at least 100-fold more KNOLLE protein than did wild-type seedlings (Fig. 2C). Furthermore, in leaves of 35S::KN plants, KNOLLE protein was localised to the membrane fraction from which it could be removed by Triton-X100 (Fig. 2D), as reported for endogenous KNOLLE protein (Lauber et al., 1997). This subcellular localisation indicated that the membrane anchor of misexpressed KNOLLE protein was functional. For comparison, truncated KNOLLE protein from the knolle allele UU1319 that shows the complete loss-of-function phenotype lacks the membrane anchor and was detected in the soluble fraction (Fig. 2D; see Materials and Methods). In summary, misexpression from two different promoters resulted in high- level accumulation of membrane-integrated KNOLLE protein in non-proliferating cells, but did not interfere with essential proliferating cells, we considered the possibility that its cellular or developmental processes. Because it had no deregulated expression might be deleterious, as has been deleterious effect, misexpression of KNOLLE offered the shown for other cell cycle-regulated genes (Ito, 2000) and unique possibility of exploring how the cell deals with a syntaxins (Zhou et al., 2000). We therefore generated syntaxin that cannot be trafficked to its proper destination, transgenic plants that expressed KNOLLE protein under the which in this case is the cell plate. control of the flower-specific APETALA3 (AP3) promoter (Fig. 1; AP3::KN1). The AP3 promoter is active in the petal and Mistargeting of KNOLLE protein in 35S::KN stamen primordia of the flower (Jack et al., 1992; Jack et al., transgenic seedling roots 1994). If misexpression of KNOLLE interfered with cellular To determine the fate of misexpressed KNOLLE protein, or developmental processes, this would be recognised by floral 35S::KN transgenic embryos and seedlings were analysed by abnormalities of viable transgenic plants. Surprisingly, the whole-mount immunolocalisation. Whereas no distinct pattern transgenic plants were morphologically normal, although of abnormal localisation was detected in embryos (data not KNOLLE protein accumulated to high levels in petals (Fig. shown), deviation from the wild-type pattern was observed in 2A). In addition, KNOLLE protein was detected in the roots of transgenic seedlings carrying two copies of the microsomal fraction (Fig. 2A), as described for endogenous 35S::KN transgene outside the meristematic region of the root 3006 JOURNAL OF CELL SCIENCE 114 (16)

Fig. 4. KNOLLE protein immunolocalisation in developing root hairs of 35S::KN seedlings. Seedlings were grown on 1% agar with 0.5× MS salts and 1% sucrose. (A) Dark-field, (B) epifluorescence microscopy, (C-Q) optical sections, confocal laser- scanning microscopy; n, nucleus; pm, plasma membrane. (A-M,O-Q) 35S::KN, (N) wild-type control. (B,C,E- G,N) Anti-KN antiserum (Cy3, red) and DAPI (blue) staining. (D,H-M) Double labelling with anti-KN antiserum (Alexa-m488, green) and anti-PM- ATPase monoclonal antibody (Cy3, red) and DAPI (blue) staining. (O-Q) Membrane labelling of live seedling roots with lipophilic dye FM4- 64. The surface of root hairs and the root is delineated by broken lines in (E-G,K-O,Q). (A-D) Overviews. (A) Seedling root with root hairs of increasing age. (B-D) KN accumulation in root hairs: (B) young root hair tip (boxed) of similar age as those in (E,H-J); (C) old root hairs; boxed area is enlarged in G and similar region is shown in Q; (D) old root hairs; boxed area is shown in (K-M). Surface of root hairs is delineated by white lines in (C,D). (E-G) KN accumulation in young (E), mid-age (F) and old (G) transgenic root hairs. (H-M) Co- localisation of KN and PM-ATPase in young (H-J) and old (K-M) root hairs. (H,K) KN staining, (I,L) PM-ATPase staining, (J,M) overlays; pink colour in J is due to additional DAPI nuclear signal. Note co-localisation of the two antigens in old root hair (K-M), which is partly collapsed owing to fixation- induced plasmolysis. (N) Young root hairs of wild-type control showing no KN signal; compare with B,E. (O-Q) Endomembrane staining of young (O), mid-age (P) and old (Q) root hairs. Compare (O,P) with strong KN accumulation in root hair tips (E,H) filled by vesicles, ER and Golgi (see Fig. 5K,L). In old root hairs (Q), FM4-64 staining resembles the KN staining (C,G). tip (Fig. 3). A clear difference between 35S::KN and wild-type protein was analysed in detail in readily accessible root hair roots was noted in the cells of the adjacent elongation zone and cells by confocal laser scanning microscopy (Fig. 4). These in the more mature part of the root, but not in the root tip, where non-proliferating epidermal cells form a local outgrowth, the cell divisions take place. The expanding cells of the central root hair, which undergoes tip growth by targeting membrane cylinder and the tip-growing root hairs strongly accumulated vesicles from the trans-Golgi to the apical plasma membrane KNOLLE protein in transgenic, but not in wild-type seedlings (for a review, see Yang, 1999; see also Fig. 5J,K). In a sense, (compare Fig. 3D,G with Fig. 3F,H). The activity of the 35S this preferential vesicle targeting to the growing tip of the root promoter was independently monitored by the expression of a hair resembles the vesicle targeting to the cell division plane GUS reporter gene (Fig. 3I-L). GUS staining was observed in during cytokinesis. Growing and mature root hairs were root hairs, the central part of the root and the root tip, as immunostained with anti-KNOLLE antiserum and with a described previously for the tobacco seedling root (Benfey et monoclonal antibody directed against the plasma membrane al., 1989). Although the 35S promoter was active in the H+-ATPase (PM-ATPase; Lauber et al., 1997). KNOLLE proliferating cells of the root tip, KNOLLE protein failed to protein accumulated strongly in growing and mature root hairs accumulate to high levels, presumably owing to cell cycle- of transgenic seedlings, in contrast to the wild-type control dependent degradation. In summary, KNOLLE protein (Fig. 4B-D,E-G; compare with Fig. 4N). In young root hairs, expressed from the transgene accumulated in root cells that KNOLLE was concentrated at the tip (Fig. 4B,E,H), whereas were no longer dividing. older root hairs also accumulated KNOLLE away from the tip The subcellular localisation of misexpressed KNOLLE (Fig. 4C-D,F,G,K). For comparison, we used the lipophilic Regulation of KNOLLE syntaxin 3007

Fig. 5. Subcellular immunolocalisation of KNOLLE protein in root cells of 35S::KN seedlings. (A-C) Confocal laser-scanning microscopy of dividing cells in wild-type root tip stained for KN (Cy3, red) and DAPI (false colour, green) to show the temporal and spatial dynamics of KN relocalisation during cytokinesis; cells are delineated by broken lines. (A) Prophase: KN patches, presumably Golgi. (B) Telophase: KN in plane of cell division and nearby patches. (C) Late phase of cytokinesis: KN-positive cell plate extends to the parental cell wall, only few patches remain. (D-L) 35S::KN transgenic root cells. (D-F) Semi-thick cryo-sections through the central cylinder of root. (D,E) Cross section: (D) KN immunofluorescence; (E) phase contrast of D; cells close to the phloem show strong signals (D) at the cell surface and inside the cells (asterisks, D; arrowheads, E) point to the xylem. (F) KN immunofluorescence of longitudinal section: the plasma membrane (arrowhead, pm) and Golgi-like intracellular structures (arrow, g) are labelled. (G-I,L) Ultrathin cryo sections of root (G-I) and root hair (L) labelled with anti-KN immunogold. Cryo sections were required because the anti- KNOLLE antiserum does not detect the integral membrane protein KNOLLE on conventional EM sections after chemical fixation (Lauber et al., 1997). Cryosections do not preserve membrane structures very well. (G) Root cell adjacent to xylem cell (overview). Gold labelling within areas delineated by dashed or dotted lines is highlighted by yellow (internal staining) or red (plasma membrane) asterisks to visualise gold label that is not readily detectable at this low magnification. g, Golgi; m, mitochondrion; pm, plasma membrane. (H) Higher magnification of upper boxed area in G rotated 90° clockwise showing KN immunogold label at the plasma membrane (pm), Golgi (g) and post-Golgi area. m, mitochondrion. (I) KN immunogold labelling of Golgi (g) and trans-Golgi (t) compartments. (J,K) Ultrathin section of chemically fixed and embedded root hair. (K) Higher magnification of boxed area in J: root hair tip filled with rough endoplasmic reticulum (er) and numerous electron-dense vesicles (v). (L) Strongly KN-labelled vesicles (v) and trans-Golgi network (t) from root hair tip. Scale bar: 20 µm in D-F; 250 nm in H,I,K,L.

fluorescent dye, FM4-64, which labels predominantly the KNOLLE-positive aggregates (Fig. 4P,Q; compare with Fig. membrane of the vacuole in yeast (Vida and Emr, 1995). In 4F,G). To determine the fate of misexpressed KNOLLE protein root hairs, FM4-64 labelled predominantly endomembranes more precisely, root hairs of 35S::KN seedlings were and to some extent the plasma membrane (Fig. 4O). Most of simultaneously immunostained for PM-ATPase (Fig. 4H-M). the FM4-64 label was located below the KNOLLE-positive In young root hairs, both proteins co-localised to the apical tip apical region in young root hairs (compare Fig. 4P with 4E). region, not only at the surface but also internally (Fig. 4E-G; In older root hairs, however, the FM4-64 label resembled additional optical sections not shown). Away from the tip 3008 JOURNAL OF CELL SCIENCE 114 (16) independent transformants, all of which produced viable and fertile kn-X37-2/knX37-2 homozygous plants (for details, see Materials and Methods). By contrast, none of the 18 kn-X37- 2/KN heterozygous independent transformants bearing the 35S::KN transgene gave rise to kn-X37-2/knX37-2 homozygous plants. Thus, the 35S::KN transgene did not rescue kn-X37-2 mutant embryos. The transgene also did not promote growth of kn-X37-2 mutant seedlings on callus- inducing medium (data not shown, see Materials and Methods). Sequencing of the transgene after re-isolation from the transgenic plants did not show any deviation from the Fig. 6. KNOLLE protein expression in a kn-X37-2 mutant seedling wild-type sequence (see Materials and Methods). We therefore carrying the 35S::KN transgene. (A) Cartoon of knolle mutant checked kn-X37-2 mutant seedlings carrying the 35S::KN seedling. Boxed area is shown in B. (B) Root hairs stained for transgene for KNOLLE protein accumulation by whole-mount KNOLLE protein (green) and nuclei (DAPI, blue). Arrows indicate immunolocalisation (Fig. 6). KNOLLE protein was detected KNOLLE protein accumulation in root hair tips. in root hair cells of those knolle mutant seedlings. Thus, the 35S::KN transgene produced KNOLLE protein in knolle region, both proteins were strictly localised to the surface, mutants, but failed to rescue cytokinesis-defective kn-X37-2 resembling the plasma membrane staining with the lipophilic mutant embryos and seedlings. fluorescent dye, FM1-43 (data not shown; see Materials and Methods). Older root hairs also displayed almost perfect co- Low-level accumulation of KNOLLE mRNA localisation of KNOLLE and PM-ATPase (Fig. 4K-M). Thus, transcribed from the 35S promoter in the embryo KNOLLE protein was targeted like a protein destined to the To determine why 35S::KN did not rescue the knolle mutant apical plasma membrane of the growing root hair. phenotype, although transgenic plants expressed high levels of To reveal the ultrastructural localisation of KNOLLE protein KNOLLE protein (see Fig. 2B,C), we analysed the transcript in non-proliferating cells of 35S::KN transgenic seedlings, root accumulation by in situ hybridisation of a KNOLLE antisense cryosections were prepared for immunogold-labelling electron riboprobe to sections of 35S::KN transgenic embryos that microscopy (Fig. 5; see Materials and Methods). In the central carried two functional copies of the endogenous KNOLLE gene cylinder of the root, expanding cells displayed cytoplasmic (Fig. 7). Up to the torpedo stage of embryogenesis, KN mRNA patches of KNOLLE immunofluorescence that resembled accumulated in a patchy pattern that was indistinguishable those in dividing cells (Fig. 5D,F; compare with Fig. 5A). from the wild-type control (Fig. 7A,B; Lukowitz et al., 1996). However, KNOLLE was also located at the plasma membrane, At the bent-cotyledon stage, transgenic embryos showed whereas dividing cells accumulated KNOLLE in the forming diffuse expression predominantly in the cotyledonary cell plate (Fig. 5F, arrowhead; compare with Fig. 5B,C). By primordia, whereas only a few cells in the wild-type control immunogold labelling, KNOLLE was detected in Golgi stacks, embryos were labelled (compare Fig. 7D,E,G with Fig. 7C). the trans-Golgi network and at the plasma membrane of The intensity of diffuse staining was stronger in embryos with expanding root cells (Fig. 5G-I). We also analysed tip-growing two copies of the transgene than in those with only one, young root hairs, which accumulate vesicles underneath the indicating that the staining was due to transgene expression apical plasma membrane (Fig. 5J,K; Galway et al., 1997). Anti- (compare Fig. 7E,G with Fig. 7D). Transgene expression was KNOLLE immunogold label was concentrated at the tip of the strongest in presumptive vascular cells of the cotyledons (Fig. root hair, with vesicles giving the strongest signal (Fig. 5L). In 7E,F) and in adjacent internal cell layers (Fig. 7G,H). Within summary, these data support the light microscopy observation these stained tissues, individual cells gave stronger signals that that KNOLLE protein is mistargeted to the plasma membrane resembled in intensity the stained cells in the hypocotyl, and in non-proliferating cells of 35S::KN transgenic plants. In probably reflect the expression of the endogenous KNOLLE addition, this is the first time that KNOLLE protein has been gene (Fig. 7H, arrowheads). In summary, mRNA transcribed localised in cells at the ultrastructural level. from the 35S::KN transgene accumulated detectably only at advanced stages of embryogenesis, at which its level of No rescue of knolle mutant embryos by 35S::KN accumulation was lower than that of the endogenous KNOLLE transgene expression mRNA. This result was consistent with the observation that Expression of the 35S::KN transgene lacked any detectable developing 35S::GUS embryos showed comparable temporal biological effect in the wild-type genetic background. We and spatial distribution and intensity of GUS expression (data therefore examined whether the transgene could functionally not shown), and that KNOLLE protein immunolocalisation did replace the endogenous KNOLLE gene. We transformed plants not reveal any difference between 35S::KN transgenic and heterozygous for the kn-X37-2 mutation (Lukowitz et al., wild-type embryos. 1996) with the 35S::KN transgene and analysed their progeny for the occurrence of mutant seeds. As a control, we 35S::KN transgene expression in the seedling root transformed kn-X37-2/KN heterozygous plants with a genomic To compare KNOLLE mRNA accumulation from 35S::KN DNA fragment that differed from the 35S::KN transgene in transgene expression with the immunolocalisation of the 5′ region, whereas their 3′ regions were nearly identical misexpressed KNOLLE protein, we examined seedling roots (see Fig. 1). The control construct gave 22 kn-X37-2/KN by in situ hybridisation. Unlike the situation in the embryo, heterozygous and 13 kn-X37-2/knX37-2 homozygous most cells of the seedling are mitotically inactive. Exceptions Regulation of KNOLLE syntaxin 3009

Fig. 7. KNOLLE mRNA expression pattern in 35S::KN transgenic embryos and seedlings. In situ analysis using KN antisense probe. Embryos 8 µm, seedlings 17.5 µm sections. Unless stated otherwise, 35S::KN material is homozygous for the transgene. (A-H) embryos, (I-U) seedlings. (A) Wild-type, torpedo-stage embryo. (B) 35S::KN, torpedo-stage embryo. (C) Wild-type, bent-cotyledon stage embryo. (D) 35S::KN, bent- cotyledon stage embryo carrying only one copy of the transgene. Note slight KNOLLE mRNA overexpression in cotyledons. (E-H) 35S::KN, bent-cotyledon stage embryo. KNOLLE mRNA is overexpressed in cotyledons. (E) Overview; (F) higher magnification of area boxed in E – expression in pre-vasculature of cotyledon (arrowheads). (G) Overview; (H) higher magnification of area boxed in G – arrowheads indicate cells with stronger signals resembling the patchy mRNA pattern observed in wild type. (I) Wild-type, root with root hairs. (J) 35S::KN, root with root hair. KN mRNA is accumulated in central part of the root. (K) 35S::KN, lateral root primordium with strongly labelled cells (asterisks) and young root hair (arrowhead) show mRNA accumulation. cot, cotyledon. Scale bars: 50 µm. are the meristems of the shoot and the root as well as the DISCUSSION primordia of leaves and lateral roots. The mature root of wild- type seedlings showed little or no distinct staining (Fig. 7I). By Consistent with its role as a cytokinesis-specific syntaxin, contrast, the root of transgenic seedlings gave strong signals in Arabidopsis KNOLLE protein accumulates during M phase, the expanding central cells (Fig. 7J) and also in tip-growing relocates to the plane of cell division during telophase and root hairs (Fig. 7K). These observations were consistent with disappears at the end of cytokinesis (Lauber et al., 1997). The the immunolocalisation of KNOLLE protein in transgenic transient accumulation of KNOLLE protein closely follows roots and with the 35S::GUS expression pattern (see Fig. 3). that of KNOLLE mRNA (Lukowitz et al., 1996), suggesting In addition, strong expression of the 35S::KN transgene was that synthesis and degradation of both mRNA and protein are observed in lateral root primordia (Fig. 7K), with patches of regulated in a cell cycle-dependent manner. We have addressed dark staining above a lighter background, resembling the the biological significance of this tight regulation. situation in embryogenesis. Thus, the activity of the 35S::KN transgene yielded high levels of KNOLLE mRNA mainly in KNOLLE protein targeting depends on the cell cycle non-proliferating cells that did not express the endogenous By expressing KNOLLE in non-proliferating cells, we created KNOLLE gene. an abnormal situation in which vesicles budding from the 3010 JOURNAL OF CELL SCIENCE 114 (16) trans-Golgi could deliver KNOLLE to several potential target redirection of membrane flow may be involved in plant membranes. However, KNOLLE co-localised with the plasma cytokinesis. Supporting evidence comes from two recent membrane H+-ATPase in tip-growing root hairs, thus behaving observations. The Arabidopsis putative auxin efflux carrier like an integral plasma membrane protein. This result was PIN1, an integral membrane protein, which is normally located confirmed by ultrastructural immunolocalisation of KNOLLE in the basal plasma membrane of non-proliferating vascular protein not only in root hairs but also in expanding cells of the cells, also accumulates at the forming cell plate (Steinmann central cylinder of the root. Thus, the plasma membrane et al., 1999). Furthermore, a secreted enzyme, cell wall- appears to be the destination of KNOLLE protein in non- associated endoxyloglucan transferase, has been reported to proliferating cells. traffic to the plasma membrane during interphase and to the Mistargeting of KNOLLE to the plasma membrane did not cell plate during cytokinesis via the endoplasmic reticulum- interfere with essential cellular processes that are required Golgi pathway in tobacco BY-2 cells (Yokoyama and Nishitani, for normal plant development. Tip-growing root hairs 2001). These observations suggest that the vast majority of were morphologically indistinguishable between 35S::KN vesicles budding from the trans-Golgi during M phase traffic transgenic and wild-type plants, although only the former to the plane of cell division. These vesicles may incorporate accumulated large amounts of KNOLLE protein in the apical any membrane or soluble cargo protein that passes through the growth zone. This observation suggests that KNOLLE did not Golgi stacks at that time and lacks a retention signal. obviously interfere with the interaction of SNARE complex Accordingly, KNOLLE protein would not need a sorting motif. partners involved in apical membrane fusion, which would be Whatever the underlying mechanism, only KNOLLE protein consistent with recent findings of yeast SNARE pairing that is synthesised during M phase can be targeted to the plane specificity (McNew et al., 2000). Although we cannot rule out of cell division. Consequently, the level of KNOLLE the fact that KNOLLE interacts with non-cognate SNARE expression during M phase may be a crucial parameter for partners without obvious deleterious effects, it is also cytokinetic vesicle fusion. conceivable that in non-proliferating cells, KNOLLE might be a biologically inactive passenger protein on vesicles destined Cytokinesis requires strong KNOLLE expression to the plasma membrane. The 35S::KN transgene yielded approximately hundred-fold Why does KNOLLE traffic to the plasma membrane in non- accumulation of KNOLLE protein in seedlings, when proliferating cells? As eukaryotic cells express several compared with the wild-type control. This result is consistent syntaxins, each of which resides in a distinct membrane with the common use of the 35S promoter for transgene compartment, there must be a sorting mechanism to ensure that overexpression in plants (Holtorf et al., 1996; Lermontova and each syntaxin is delivered to its proper destination. For Grimm, 2000; Sentoku et al., 2000). However, the 35S::KN example, the Arabidopsis syntaxin AtPEP12 is targeted to the transgene did not complement knolle mutant embryos. One vacuolar pathway (da Silva Conceicao et al., 1997). Although difference between developing embryos and seedlings is that the mechanism of syntaxin sorting is unknown in plants, recent embryo cells are proliferating, whereas most cells in the observations suggest that sorting occurs during vesicle budding seedling are not. As shown by in situ hybridisation and from the trans-Golgi donor membrane in yeast. An acidic di- immunostaining in seedling roots, 35S::KN transgene leucine motif of the vacuolar t-SNARE Vamp3p appears expression resulted in the stable accumulation of KNOLLE essential for sorting mediated by the adaptor protein complex mRNA and protein in non-proliferating cells. Comparative in AP-3 (Darsow et al., 1998), whereas Golgi-associated coat situ hybridisation and immunostaining of 35S::KN transgenic proteins with homology to gamma adaptin appear to interact and wild-type embryos revealed the relative strength of the 35S with a different sorting motif of Pep12p for its targeting to late promoter in proliferating cells. Expression of the 35S::KN endosomes (Black and Pelham, 2000). By contrast, no AP3- transgene, if at all detectable, supplemented the wild-type dependent sorting motif has been identified in the plasma patchy pattern of KNOLLE mRNA accumulation by low-level membrane syntaxins, Sso1 and Sso2 (Tang and Hong, 1999), accumulation of KNOLLE mRNA in the primordia of the and KNOLLE lacks the consensus acidic di-leucine motif. If cotyledons. The 35S promoter activity was independently post-Golgi trafficking to the plasma membrane were a default assessed in 35S::GUS embryos, which gave a similar pathway in the absence of specific targeting cues, mistargeting developmental expression pattern and level of expression as of KNOLLE protein might reflect the lack of such a sorting 35S::KN (data not shown). Our results are consistent with signal. This does not exclude the possibility that normal previous results demonstrating 35S promoter activity only from targeting of KNOLLE to the plane of cell division may involve the heart stage on in 35S::GUS transgenic tobacco embryos a hypothetical sorting-signal receptor that is not present in non- (Odell et al., 1994). Furthermore, no additional KNOLLE proliferating cells. The existence of an active sorting protein accumulation was detected in 35S::KN transgenic mechanism for proteins destined to the cell plate has been embryos, when compared with wild-type embryos. Taken hypothesised based on the behaviour of GFP-KOR, a together, these observations suggest that the expression level GFP fusion to the Arabidopsis endo-1,4-β-D-glucanase of the 35S::KN transgene in proliferating cells was insufficient KORRIGAN (Nicol et al., 1998), in a heterologous expression to rescue knolle mutant embryos. system (Zuo et al., 2000). KORRIGAN and KNOLLE share a The difference between the 35S::KN and the KNOLLE YVDL sequence that may act an AP-dependent sorting motif, rescue (KNRescue) constructs was confined to the 5′ region, although its physiological significance and specificity remain whereas both constructs contained the same genomic 3′ to be determined (for a review, see Bonifacino and sequence. Thus, any difference in expression pattern and Dell’Angelica, 1999). intensity between the two constructs can be attributed to Independently of specific sorting signals, a general different 5′ sequences, the KNOLLE cis-regulatory region as Regulation of KNOLLE syntaxin 3011 opposed to the 35S promoter. The KNOLLE 5′ sequence appears Assaad, F. F., Huet, Y, Mayer, U. and Jürgens, G. (2000). The cytokinesis to integrate signals that link KNOLLE expression to the cell gene KEULE encodes a Sec1 protein which binds the syntaxin KNOLLE. cycle as indicated by the patchy pattern of mRNA J. Cell Biol. 152, 531-543. Bechtold, N. and Pelletier, G. (1998). In planta Agrobacterium-mediated accumulation. Promoter elements conferring M phase-specific transformation of adult Arabidopsis thaliana plants by vacuum infiltration. transcription have been identified in mitotic cyclin genes (Ito et Methods Mol. Biol. 82, 259-266. al., 1998). The KNOLLE 5′ UTR should also contain a sequence Bechtold, N., Jaudeau, B., Jolivet, S., Maba, B., Vezon, D., Voisin, R. and that enables translation of the mRNA during M phase, when Pelletier, G. (2000). The maternal chromosome set is the target of the T- DNA in the in planta transformation of Arabidopsis thaliana. Genetics 155, most protein expression is shut down (for a review, see Sachs, 1875-1887. 2000). However, KNOLLE expression is not strictly linked to Benfey, P. N., Ren, L. and Chua, N. H. (1989). The CaMV 35S enhancer karyokinesis, as KNOLLE protein accumulates between non- contains at least two domains which can confer different developmental and mitotic nuclei during endosperm cellularisation (Lauber et al., tissue-specific expression patterns. EMBO J. 8, 2195-2202. 1997; Otegui and Staehelin, 2000). Black, M. W. and Pelham, H. R. (2000). A selective transport route from golgi to late endosomes that requires the yeast GGA proteins. J. Cell Biol. In contrast to the KNOLLE promoter, the 35S promoter 151, 587-600. appears to be active in a cell cycle-independent manner. Blatt M. R., Leyman B. and Geelen D. (1999). Molecular events of vesicle KNOLLE mRNA accumulated stably in non-proliferating cells trafficking and control by SNARE proteins in plants. New Phytol. 144, 389- of 35S::KN transgenic seedlings but only transiently in 418. Bonifacino, J. S. and Dell´Angelica, E. C. (1999). Molecular bases for the proliferating cells. Instability of short-lived mRNAs has been recognition of tyrosine-based sorting signals. J. Cell Biol. 145, 923-926. attributed to specific sequences in the 3′ UTR (Gutierrez et al., Burgess, R. W., Deitcher, D. L. and Schwarz, T. L. (1997). The synaptic 1999; Sachs, 2000). In the case of KNOLLE mRNA, an as yet protein syntaxin1 is required for cellularization of Drosophila embryos. J. undefined degradation signal appears to be linked to the M Cell Biol. 138, 861-875. phase and/or cytokinesis. The lack of strong accumulation of Clague, M. J. and Herrmann, A. (2000). Deciphering fusion. Curr. Biol. 10, R750-R752. KNOLLE mRNA in proliferating cells of 35S::KN transgenic Clough, S. J. and Bent, A. F. (1998). Floral dip: a simplified method for embryos and seedlings suggests that the activity of the 35S Agrobacterium-mediated transformation Arabidopsis thaliana. Plant J. 16, promoter is too low in proliferating cells or does not counteract 735-743. efficiently the cell cycle-dependent mRNA degradation. By Darsow, T., Burd, C. G. and Emr, S. D. (1998). Acidic di-leucine motif essential for AP-3-dependent sorting and restriction of the functional contrast, the endogenous KNOLLE promoter is strong enough specificity of the Vam3p vacuolar t-SNARE. J. Cell Biol. 142,:913-922. to yield high levels of mRNA and protein accumulation in da Silva Conceicao, A., Marty-Mazars, D., Bassham, D. C., Sanderfoot, proliferating cells, although it is only active during a brief A. A., Marty, F. and Raikhel, N. V. (1997). The syntaxin homolog period of the cell cycle. Thus, during its period of activity, the AtPEP12p resides on a late post-Golgi compartment in plants. Plant Cell endogenous KNOLLE promoter is clearly stronger than the 35S 9, 571-582. Desfeux, C., Clough, S. J. and Bent, A. F. (2000). Female reproductive tissues promoter, and this difference appears to be crucial for the are the primary target of Agrobacterium-mediated transformation by the execution of cytokinesis. Arabidopsis floral-dip method. Plant Physiol. 123, 895-904. In summary, there is no obvious need for the observed tight Fasshauer, D., Antonin, W., Margittai, M., Pabst, S. and Jahn, R. (1999). regulation of KNOLLE expression, provided enough Mixed and non-cognate SNARE complexes. Characterization of assembly KNOLLE protein is available during M phase to ensure and biophysical properties. J. Biol. Chem. 274, 15440-15446. Fuerst, R. A., Soni, R., Murray, J. A. and Lindsey, K. (1996) Modulation efficient vesicle fusion during cell-plate formation. This of cyclin transcript levels in cultured cells of Arabidopsis thaliana. Plant demand is met by the strong KNOLLE promoter, which is Physiol. 112, 1023-1033. highly active during the period preceding cytokinesis. If there Fukuda, R., McNew, J. A., Weber, T., Parlati, F., Engel, T. Nickel, W., were no cell cycle-dependent degradation of KNOLLE mRNA Rothman, J. E. and Söllner, T. H. (2000). Functional architecture of an intracellular membrane t-SNARE. Nature 407, 198-202. and protein, both would stably accumulate. Although Galway, M. E., Heckman, J. W., Jr. and Schiefelbein, J. W. (1997). Growth mistargeting of KNOLLE protein to the plasma membrane and ultrastructure of Arabidopsis root hairs: the rhd3 mutation alters vacuole appears not to be harmful, it is also not useful. Perhaps enlargement and tip growth. Planta 201, 209-218. proliferating cells gain some selective advantage from linking Griffiths, G. (1993). In Fine Structure Immunocytochemistry (ed. G. Griffiths), the degradation of KNOLLE mRNA and protein, as well as the pp. 174-177. Heidelberg: Springer Verlag. Gutierrez, R. A., MacIntosh, G. C. and Green, P. J. (1999). Current promoter activity, to the cell cycle. perspectives on mRNA stability in plants: multiple levels and mechanisms of control. Trends Plant Sci. 4, 429-438. We thank W. Lukowitz for kindly providing the cosmid 74-4, G. Heese, M., Mayer, U. and Jürgens, G. (1998). Cytokinesis in flowering Cardon for the pBarA and pBar-35S vectors and the Agrobacterium plants: cellular process and developmental integration. Curr. Opin. Plant strain GV3101, T. Jack for the pAP3 vector, U. Mayer for the knolle Biol. 1, 486-491. allele UU1319, and W. Michalke for the monoclonal anti-plasma Holtorf, S., Apel, K. and Bohlmann, H. (1996). Comparison of different membrane H+-ATPase antibody. We also thank S. Mangold and T. constitutive and inducible promoters for the overexpression of transgenes in Pacher for help with the immunoblotting, and N. Geldner, M. Heese, Arabidopsis thaliana. Plant Mol. Biol. 29, 637-646. Ito, M., Iwase, M., Kodama, H., Lavisse, P., Komamine, A., Nishihama, M. Lenhard, Jaideep Mathur, U. Mayer and K. Schrick for critical R., Machida, Y. and Watanabe, A. (1998). A novel cis-acting element in reading of the manuscript. This work was supported by grant SFB promoters of plant B-type cyclin genes activates M phase-specific 446/B-8 from the Deutsche Forschungsgemeinschaft. transcription. Plant Cell 10, 331-341. Ito, M. (2000). Factors controlling cyclin B expression. In The Plant Cell Cycle (ed. D. Inzé). Plant Mol. Biol. 43, 677-690. Jack, T., Brockman, L. L. and Meyerowitz, E. M. (1992). The homeotic REFERENCES gene APETALA3 of Arabidopsis thaliana encodes a MADS box and is expressed in petals and stamens. Cell 68, 683-697. Assaad, F. F., Mayer, U., Wanner, G. and Jürgens, G. (1996). The KEULE Jack, T., Fox, G. L. and Meyerowitz, E. M. (1994). Arabidopsis homeotic gene is involved in cytokinesis in Arabidopsis. Mol. Gen. Genet. 253, 267- gene APETALA3 ectopic expression: transcriptional and posttranscriptional 277. regulation determine floral organ identity. Cell 76, 703-716. 3012 JOURNAL OF CELL SCIENCE 114 (16)

Jahn, R. and Südhof, T. C. (1999). Membrane fusion and . Annu. in tobacco BY-2 and root tip cells: a new model of cell plate formation in Rev. Biochem. 68, 863-911. higher plants. J. Cell Biol. 130, 1345-1357. Jantsch-Plunger, V. and Glotzer, M. (1999). Depletion of syntaxins in the Sanderfoot, A. A., Assaad F. F and Raikhel N. V. (2000). The Arabidopsis early Caenorhabditis elegans embryo reveals a role for membrane fusion Genome. An abundance of soluble N-Ethylmaleimide-sensitive factor events in cytokinesis. Curr. Biol. 9, 738-745. adaptor protein receptors. Plant Physiol. 124, 1558-1569. Joshi, C. P. (1987). An inspection of the domain between putative TATA box Sentoku, N., Sato, Y. and Matsuoka, M. (2000). Overexpression of rice OSH and translation start site in 79 plant genes. Nucleic Acids Res. 15, 6643- genes induces ectopic shoots on leaf sheaths of transgenic rice plants. Dev. 6653. Biol. 220, 358-364. Langanger, G., de Mey, J. and Adam, H. (1983). 1,4-Diazobizyklo-[2.2.2.]- Slot, J. W. and Geuze, H. J. (1985). A new method for preparing gold probes Octan verzögert das Ausbleichen von Immunfluoreszenzpräparaten. for multiple-labeling microscopy. Eur. J. Cell Biol. 38, 87-93. Mikroskopie (Wien) 40, 237-241. Söllner, T., Bennett, M. K., Whiteheart, S. W., Scheller, R. H. and Lauber, M. H., Waizenegger, I., Steinmann, T., Schwarz, H., Mayer, U., Rothman, J. E. (1993a). A protein assembly-disassembly pathway in vitro Hwang, I., Lukowitz, W. and Jürgens, G. (1997). The Arabidopsis that may correspond to sequential steps of docking, KNOLLE protein is a cytokinesis-specific syntaxin. J. Cell Biol. 139, 1485- activation, and fusion. Cell 75, 409-418. 1493. Söllner, T., Whiteheart, S. W., Brunner, M., Erdjument-Bromage, H., Lermontova, I. and Grimm, B. (2000). Overexpression of plastidic Geromanos, S., Tempst, P. and Rothman, J. E. (1993b). SNAP receptors protoporphyrinogen IX oxidase leads to resistance to the diphenyl-ether implicated in vesicle targeting and fusion. Nature 362, 318-324. herbicide acifluorfen. Plant Physiol. 122, 75-84. Soni, R., Carmichael, J. P., Shah, Z. H. and Murray J. A. H. (1995). A Lukowitz, W., Mayer, U. and Jürgens, G. (1996). Cytokinesis in the family of cyclin D homologs from plant differentially controlled by growth Arabidopsis embryo involves the syntaxin-related KNOLLE gene product. regulators and containing the conserved retioblastoma protein interaction Cell 84, 61-71. motiv. Plant Cell 7, 85-103. Mayer U., Torres-Ruiz R. A., Berleth T., Miséra S. and Jürgens G. (1991). Staehelin L. A. and Hepler P. K. (1996). Cytokinesis in higher plants. Cell Mutations affecting body organization in the Arabidosis embryo. Nature 84, 821-824. 353, 402-407 Steinmann, T., Geldner, N., Grebe, M., Mangold, S., Jackson, C. L., Paris, Mayer, K. F. X., Schoof, H., Haecker, A., Lenhard, M., Jürgens, G. and S., Galweiler, L., Palme, K. and Jürgens, G. (1999). Coordinated polar Laux, T. (1998). Role of WUSCHEL in regulating stem cell fate in the localization of auxin efflux carrier PIN1 by GNOM ARF GEF. Science 286, Arabidopsis shoot meristem. Cell 95, 805-815. 316-318. McNew, J. A., Parlati, F., Fukuda, R., Miller, R., Söllner, T. H. and Straight, A. F. and Field, C. M. (2000). Microtubules, membranes and Rothman, J. E. (2000). Compartmental specificity of cellular membrane cytokinesis. Curr. Biol. 10, R760-R770. fusion encoded in SNARE proteins. Nature 407, 153-159. Sundaresan, V., Springer, P., Volpe, T., Haward, S., Jones, J. D., Dean, C., Meyer, K., Leube, M. P. and Grill, E. (1994). A protein phosphatase 2C Ma, H. and Martienssen, R. (1995). Patterns of gene action in plant involved in ABA signal transduction in Arabidopsis thaliana. Science 264, development revealed by enhancer trap and gene trap transposable elements. 1452-1455. Genes Dev. 9, 1797-810. Nacry, P., Mayer, U. and Jürgens, G. (2000). Genetic dissection of Tang, B. L. and Hong, W. (1999). A possible role of di-leucine-based motifs cytokinesis. In The Plant Cell Cycle (ed. D. Inzé). Plant Mol. Biol. 43, 719- in targeting and sorting of the syntaxin family of proteins. FEBS Lett. 446, 733. 211-212. Nicol, F., His, I., Jauneau, A., Vernhettes, S., Canut, H. and Höfte, H. Tokuyasu, K. T. (1989). Use of poly(vinylpyrrolidon) and poly(vinyl alcohol) (1998). A plasma membrane-bound putative endo-1,4-b-D-glucanase is for cryoultramicroscopy. Histochem J. 21, 163-171. required for normal wall assembly and cell elongation in Arabidopsis. Vida, T. A. and Emr, S. D. (1995). A new vital stain for visualizing EMBO J. 17, 5563-5576. vacuolar membrane dynamics and endocytosis in yeast. J. Cell Biol. 128, Odell, J. T., Hoopes, J. L. and Vermerris, W. (1994). Seed-specific gene 779-792. activation mediated by the Cre/lox site-specific recombination system. Plant Waizenegger, I., Lukowitz, W., Assaad, F., Schwarz, H., Jürgens, G. and Physiol. 106, 447-458. Mayer, U. (2000). The Arabidopsis KNOLLE and KEULE genes interact to Otegui, M. and Staehelin, L. A. (2000). Syncytial-type cell plates: a novel promote vesicle fusion during cytokinesis. Curr. Biol. 10, 1371-1374. kind of cell plate involved in endosperm cellularization of Arabidopsis. Yang, Z. (1999). Signaling tip growth in plants. Curr. Opin. Plant Biol. 1, 525- Plant Cell 12, 933-947. 530. Parlati, F., McNew, J. A., Fukuda, R., Miller, R., Söllner, T. H. and Ye, G. N., Stone, D., Pang, S. Z., Creely, W., Gonzalez, K. and Hinchee, Rothman, J. E. (2000). Topological restriction of SNARE-dependent M. (1999). Arabidopsis ovule is the target for Agrobacterium in planta membrane fusion. Nature 407, 194-198. vacuum infiltration transformation. Plant J. 19, 249-257. Robinson, D. N. and Spudich, J. A. (2000). Towards a molecular Yokoyama, R. and Nishitani, K. (2001). Endoxyloglucan transferase is understanding of cytokinesis. Trends Cell Biol. 10, 228-237. localized both in the cell plate and in the secretory pathway destined for the Rodriguez, J. and Deinhardt, F. (1960). Preparation of semipermanent apoplast in tobacco cells. Plant Cell Physiol. 42, 292-300. mounting medium for fluorescent antibody studies. Virology 12, 316-317. Zhou, Q., Xiao, J. and Liu, Y. (2000). Participation of syntaxin 1A in Sachs, A. B. (2000). Cell cycle-dependent translation initiation: IRES membrane trafficking involving neurite elongation and membrane elements prevail. Cell 101, 243-5. expansion. J. Neurosci. Res. 61, 321-328. Sambrook, J., Fritsch, E. F. and Maniatis, T. (1989). Molecular Cloning: A Zuo, J., Niu, Q.-W., Nishizawa, N., Wu, Y., Kost, B. and Chua, N.-H. Laboratory Manual, 2nd ed. Cold Spring Harbor, NY: Cold Spring Habor (2000). KORRIGAN, an Arabidopsis endo-1,4-b-glucanase, localizes to the Laboratory Press. cell plate by polarized targeting and is essential for cytokinesis. Plant Cell Samuels, A. L., Giddings, T. H., Jr and Staehelin, L. A. (1995). Cytokinesis 12, 1137-1152.