Endocytosis of Cell Surface Material Mediates Cell Plate Formation During Plant Cytokinesis

Home , Cell plate

Developmental Cell 10, 137–150, January, 2006 ª2006 Elsevier Inc. DOI 10.1016/j.devcel.2005.11.015 Endocytosis of Cell Surface Material Mediates Cell Plate Formation during Plant Cytokinesis

Pankaj Dhonukshe,1,2,* Frantisˇek Balusˇka,3,4 Introduction Markus Schlicht,3 Andrej Hlavacka,3 Jozef Sˇ amaj,3,5 Jirı´ Friml,2,* and Theodorus W.J. Gadella, Jr.1,* Cytokinesis in immobile plants strikingly differs from 1 Section of Molecular Cytology and Center for that in other higher eukaryotes, as a new plasma mem- Advanced Microscopy brane (PM) and cell wall (CW) separating the daughter Swammerdam Institute for Life Sciences cells form as a new structure (called the cell plate) within University of Amsterdam the cytoplasm rather than by inward growth or constric- Kruislaan 316 tion of the existing PM and CW (Staehelin and Hepler, 1098 SM Amsterdam 1996; Verma, 2001). This process has been studied for The Netherlands more than a century, as the term ‘‘cell plate’’ can be 2 Zentrum fu¨ r Molekularbiologie der Pflanzen tracked back to Strasburger (1875). Cell plate bio- Universita¨ tTu¨ bingen genesis process requires the creation of one-third of Auf der Morgenstelle 3 the original cell surface area during 4% of the cell cycle 72076 Tu¨ bingen duration and hence demands extremely fast and tar- Germany geted vesicular trafficking. The cell plate formation is as- 3 Institute of Cellular & Molecular Botany sisted by a plant-specific polarized microtubular array University of Bonn sandwiching the forming cell plate known as a phragmo- Kirschallee 1 plast (Ledbetter and Porter, 1963). This highly special- 53115 Bonn ized structure also contains actin and assists in the de- Germany livery of the vesicles containing material for cell plate 4 Institute of Botany formation. Slovak Academy of Sciences Genetic studies in the model plant Arabidopsis thali- Du´ bravska´ cesta 14 ana led to identification of mutants with defects in cyto- 842 23 Bratislava kinesis (Assaad et al., 2001; Heese et al., 2001; Kang Slovak Republic et al., 2003a; Lukowitz et al., 1996; Nishihama et al., 5 Institute of Plant Genetics and Biotechnology 2002; Waizenegger et al., 2000). The typical phenotypic Slovak Academy of Sciences features include formation of incomplete CWs and mul- Akademicka´ 2 tinucleate cells along with normal cells and CWs in the P.O. Box 39 A same tissue (Assaad et al., 2001; Heese et al., 2001; SK-950 07 Nitra Kang et al., 2003a; Lukowitz et al., 1996; Nishihama Slovak Republic et al., 2002). These genetic studies highlighted a role of syntaxin KNOLLE and its interactor Sec1 homolog KEULE in cell plate formation (Assaad et al., 2001; Luko- Summary witz et al., 1996; Waizenegger et al., 2000). KNOLLE is expressed only in dividing cells, and it localizes strongly Dividing plant cells perform a remarkable task of build- at the cell plate beside intracellular structures (Lauber ing a new cell wall within the cytoplasm in a few et al., 1997). knolle and keule mutants display incom- minutes. A long-standing paradigm claims that this plete CWs (formed by the fusion of vesicles) as well primordial cell wall, known as the cell plate, is gener- as accumulation of nonfused vesicles, consistent with ated by delivery of newly synthesized material from a known role of syntaxin and Sec1 in vesicle fusion. Golgi apparatus-originated secretory vesicles. Here, This suggests a role for KNOLLE and KEULE in vesicle we show that, in diverse plant species, cell surface ma- fusion at the cell plate, but not in vesicle formation or tar- terial, including plasma membrane proteins, cell wall geting (Assaad et al., 2001; Waizenegger et al., 2000). In components, and exogenously applied endocytic trac- addition, cytokinesis defects occur when vesicle trans- ers, is rapidly delivered to the forming cell plate. Im- port is inhibited, by repression of kinesin-like proteins portantly, this occurs even when de novo protein NACK1 and NACK2 (Nishihama et al., 2002), or by de- synthesis is blocked. In addition, cytokinesis-specific fects in dynamin-like regulators of vesicle budding, syntaxin KNOLLE as well as plasma membrane (PM) adl1A adl1E (Hinshaw, 2000; Kang et al., 2003b). This resident proteins localize to endosomes that fuse implies that vesicle formation, trafficking, and fusion to initiate the cell plate. The rate of endocytosis is events are crucial for proper execution of plant cyto- strongly enhanced during cell plate formation, and kinesis. its genetic or pharmacological inhibition leads to cyto- For roughly 40 years, it has generally been assumed kinesis defects. Our results reveal that endocytic de- that Golgi apparatus-derived secretory vesicles are the livery of cell surface material significantly contributes sole source of material required for cell plate initiation to cell plate formation during plant cytokinesis. and expansion (Alberts et al., 2002; Frey-Wyssling et al., 1964). This textbook view of cell plate biogenesis is mainly based on electron microscopy (EM) observa- *Correspondence: [email protected] tions showing massive accumulation of vesicles with (P.D.); [email protected] (J.F.); gadella@science. a diameter of 60–80 nm (resembling transport and se- uva.nl (T.W.J.G.) cretory vesicles) at the spindle and phragmoplast areas Developmental Cell 138

during mitosis (Otegui et al., 2001). It is further sup- compare the speed of FM4-64 delivery to the cell plate ported by drug treatments showing that Brefeldin A in the case of unattached and attached cell plates, we (BFA), a presumptive inhibitor of secretion, slows down photobleached FM4-64 at the cell plate and followed cell plate expansion (Lippincott-Schwartz et al., 1989; its recovery kinetics. After complete photobleaching of Yasuhara et al., 1995). However, a number of observa- FM4-64 at initiating cell plates (that are not yet attached tions cannot be easily reconciled with the notion that to any side of the parental PM) in both BY-2 (not shown) Golgi apparatus (GA)-derived exocytic vesicles alone and Arabidopsis cells (Figure 1B; Movie S2), a rapid re- drive plant cytokinesis: (1) that BFA retards expansion, appearance of about 40% of the initial FM4-64 fluores- but cannot inhibit cell plate initiation (Yasuhara et al., cence was observed at the photobleached cell plate 1995), (2) that composition of the early cell plate largely area with a half time of w80 s (Figure 1D). FRAP analysis excludes GA-derived matrix polysaccharides (Moore of cell plates attached to their parental PMs displayed a and Staehelin, 1988), and (3) that interfering with di- much more extensive (80%) and faster (half time of w8s) leucine or tyrosine endocytic motifs affects localization reappearance of FM4-64 (Figures 1C and 1D; Movie S3). of key molecules on cell plates (Zuo et al., 2000). There- This result demonstrates that FM4-64 is typically de- fore, to extend our understanding of the process of livered to nonattached cell plates by endocytic ves- cell plate formation, we undertook the analysis of the icles and not by a direct connection with the parental sources of material contributing to cell plate formation. PM. Next, we performed FM4-64 pulse-chase experiments Results in BY-2 cells by complete washing out of FM4-64 after its application. Clearly, FM4-64 labeled first the periph- Endocytic and Pinocytic Markers Are Targeted eries of already initiated cell plates, and, afterward, it into Cytokinetic Cell Plates was distributed all along the cell plate (Figure S1A). The fluorescent lipophilic styryl dye FM4-64 has been This shows polarized delivery of FM4-64 at the edges used in a number of plant studies as an authentic in of the forming cell plate. Furthermore, FM4-64 fluores- vivo endocytic marker. It labels the PM, is taken up cence gradually increased on the expanding cell plate into the cell interior only by endocytosis, and, afterward, at the expense of its fluorescence at the PM (see the it gradually labels the entire endosomal pathway toward successive frames in Figure S1A), implying active trans- the vacuole (Geldner et al., 2003; Grebe et al., 2003; port against a concentration gradient. Together, these Meckel et al., 2004; Murphy et al., 2005; Paciorek et al., results imply active and selective internalization as well 2005; Samaj et al., 2005; Shope et al., 2003; Tse et al., as polarized delivery of the endocytic marker FM4-64 2004; Ueda et al., 2001, 2004; Yamada et al., 2005; from the PM to forming cell plates. Yano et al., 2004). Therefore, we employed FM4-64 To investigate whether, besides membrane material, for studying the dynamics of endocytosis during cyto- extracellular fluid can also be delivered to the cell plate kinesis in real time. FM4-64 was applied to a well- by endocytosis, we used the membrane-impermeable established tobacco bright yellow-2 (BY-2) cell line, pinocytic fluid-phase markers Alexa 633 (Alexa) and Lu- transformed with a GFP fusion protein labeling microtu- cifer Yellow (LY) (Baluska et al., 2004; Emans et al., bules in vivo, facilitating visualization of the cell cycle 2002). Strikingly, we observed that both Alexa (Figures stage (Dhonukshe and Gadella, 2003). FM4-64 internal- 1E and 1F) and LY (Figure 1G) internalized in dividing ized in mitotic BY-2 cells and accumulated inside the cells and accumulated within the volume occupied by spindle region within a few minutes (Figure 1A). In telo- forming cell plates. Again, these cell plates were not at- phase, FM4-64-labeled vesicles occupied the newly avail- tached to the parental PM (see the top view of a cell in able space between separating chromosomes, and, sub- Figure 1F). Also, in maize root cells, LY localized to form- sequently, FM4-64 extensively labeled emerging cell ing cell plates, which were specifically marked with the plates (even stronger than the parental PM), from the very callose marker Aniline Blue (Figure 1G). Together these first signs of their appearance (Figure 1A; Movie S1; see results demonstrate that endocytic lipophilic and (extra- the Supplemental Data available with this article online). cellular) fluid-phase markers are constitutively internal- Also during cell plate expansion, the constitutive incor- ized and rapidly targeted into initiating and expanding poration of many FM4-64-stained endomembrane cell plates. structures at cell plate edges was observed, suggesting continuous delivery of endocytosed material (Movie S1). PM and CW Markers Are Internalized and Targeted Identical results were obtained in dividing root tip cells into Cytokinetic Cell Plates after FM4-64 application to Arabidopsis seedlings (not As the results with fluorescent endocytic markers sug- shown). To test the possibility that rapid delivery of gested continuous delivery of cell surface-derived mate- FM4-64 into cell plates occurs via transient connections rial to the cell plate, we investigated what happens to between parental PM and the forming cell plate (Cutler PM-resident proteins during cytokinesis. For this, we and Ehrhardt, 2002), we performed both 3D optical sec- performed coimmunolocalization studies in Arabidopsis tioning and fluorescence recovery after photobleaching root tips by using the cell plate marker KNOLLE (Lauber (FRAP) analysis. We followed dividing cells from meta- et al., 1997; Lukowitz et al., 1996). All tested PM proteins, phase onward and acquired time-lapse data by scan- including the auxin transport facilitator PIN2 (Friml, ning median longitudinal sections. 3D optical sectioning 2003)(Figure 2A), the PM water channel PIP2 (Cutler comprised of both median longitudinal and top views et al., 2000)(Figure 2C), the integral PM protein LTI6b showed the emergence and expansion of the cell plates (Cutler et al., 2000)(Figure 2E), dynamin (Kang et al., at the cell center without their immediate attachment 2003a)(Figure S2C), and the brassinolide receptor to any side of the parental PM (Figures S1B–S1D). To BRI1 (Wang et al., 2001)(Figure S2E), clearly localized Endocytosis Drives Plant Cytokinesis 139

Figure 1. Endocytic and Pinocytic Markers Localize to the Forming Cell Plates (A–G) (A, E, and F) Tobacco BY-2 suspension cells, (B and C) Arabidopsis root cells, and (G) maize root cells. (A) Endocytic marker FM4-64 labels the cell plate from initiation until completion. Microtubules are shown in green, FM4-64 is shown in red, and time is given in minutes (see also Movie S1). (B and C) FRAP of FM4-64 ([B], forming cell plate; [C], attached cell plate). The photobleached areas are marked by circles, and time is given in seconds (see also Movies S2 and S3). (D) Kinetics of FM4-64 FRAP depicted in (B) and (C). FRAP is 10-fold faster in cell plates attached to parental PM compared to the forming cell plates. (E) The endocytic ﬂuid phase marker Alexa labels cell plate volume. (F) Cross- section of a cell from (E) shows no attachment of the cell plate to the parental CW. Microtubules, green; Alexa, red. (G) An internalized LY-labeled (green) callosic (red) cell plate. The scale bars are 5 mm. Developmental Cell 140

Figure 2. Endocytosed PM Proteins and CW Pectins Localize to the Forming Cell Plates (A–L) (A–F) Arabidopsis root cells, (G–J) tobacco BY-2 suspension cells, and (K and L) maize root cells. DAPI staining is in blue. (A–B) PIN2 (green) and KNOLLE (red) colocalize on cell plates and cytosolic punctae in (A) control cells as well as (B) CHX-treated cells. (C–D) PM-localized water channel PIP2 (green) and KNOLLE (red) relocalize on cell plates and cytosolic punctae in (C) control cells as well as (D) CHX-treated cells. (E–F) PM-localized LT16b (green) and KNOLLE (red) colocalize on cell plates and cytosolic punctae in (E) control cells as well as (F) CHX-treated cells. (G and H) RGII pectin labels (G) early and (H) expanding cell plates. (I) JIM5 pectin, but not (J) JIM7 pectin, labels the cell plate. (K) JIM5 (red) and PIN1 (green) colocalize (yellow) on the cell plate. (L) RGII labels the cell plates formed in CHX-treated maize cells.

to KNOLLE-labeled forming cell plates as well as these results suggest that both parental PM- and CW- KNOLLE-labeled endomembrane structures (Figures derived material is targeted to the cell plate. 2A, 2C, and 2E; Figures S2C and S2E). These results demonstrate targeting of the PM proteins/receptors to Cell Plate Initiation Does Not Require De Novo the forming cell plate. Protein Synthesis Next, we tested the fate of the CW-derived material Our results indicated that already preexisting material at during cytokinesis. Early cell plates are enriched with the cell surface is used for cell plate formation. There- CW pectins. However, unlike the mature CWs, they fore, next, we investigated the contribution of de novo also contain callose, but hardly any cellulose (Matar synthesis. Cycloheximide (CHX) is used generally as and Catesson, 1988; Moore and Staehelin, 1988; Sa- a potent blocker of protein synthesis. It is also highly ef- muels et al., 1995). Recently, it has been reported that fective in the plant systems at a 50 mM concentration for interphase maize root cells internalize pectins from 30 min, as was demonstrated by [35S]-methionine incor- parental CWs (Baluska et al., 2002). We performed a poration analysis in Arabidopsis root cells (Geldner pectin-immunolocalization study in both BY-2 cells et al., 2001; Paciorek et al., 2005). After FM4-64 applica- and maize root cells by using different sets of antibodies tion to BY-2 cells pretreated with 50 mM CHX for 1 hr, we specifically recognizing pectins found either exclusively found that cell plates can still be initiated and labeled by in mature CWs or in Golgi apparatus-derived vesicles. FM4-64 (not shown). Also, Arabidopsis cells pretreated The antibodies used include rhamnogalacturonan II with CHX for 1–2 hr displayed cell plate initiation and lo- (RGII), recognizing de muro-formed RGII dimers cross- calization of all tested PM proteins to KNOLLE-labeled linked by borate diol diester (O’Neill et al., 2001); JIM5, forming cell plates (Figures 2B, 2D and 2F; Figures recognizing partially esterified (up to 40%) CW homoga- S2B, S2D, and S2F). In addition, maize root cells pre- lacturonan pectins; and JIM7, labeling esterified (up to treated with CHX for 1–2 hr displayed forming cell plates 80%) homogalacturonan pectins present mainly within together with cell plate targeting of RGII (Figure 2L). Golgi apparatus-derived vesicles (Baluska et al., 2002). These results demonstrate that cell plate initiation and Remarkably, the forming cell plates in both maize and localization of PM-resident proteins and CW pectins to BY-2 cells were strongly labeled with the RGII (Figures forming cell plates do not strictly depend on de novo 2G and 2H; Figures S2G and S2H) and JIM5 (Figure 2I; synthesis. This further confirms that cell surface material Figures S2I and S2J) antibodies rather than the JIM7 can be used for cell plate formation. antibodies (Figure 2J; Figure S2K). The crosslinking of RGII pectins is well known to occur exclusively within Endosomes, but Not cis-ortrans-Golgi Markers, mature parental CW (Kobayashi et al., 1999; O’Neill Are Targeted toward the Initiating Cell Plate et al., 2001), demonstrating that parental CW-derived Delivery of cell surface material to the cell plate likely in- material gets internalized into the cell plate. In summary, volves endocytosis, whereas newly synthesized material Endocytosis Drives Plant Cytokinesis 141

is supposed to arrive by Golgi apparatus-dependent to the cell plate and surrounding endosomes, whereas secretion. Therefore, we examined the dynamics of en- it revealed 3.9 6 1.3 gold particles to be cytoplasmic. dosomes and Golgi apparatus during cytokinesis by us- In the case of Ara7, immunogold labeling in dividing cells ing the established markers. We used BY-2 cells and (n = 3 in two independent experiments) revealed 6.7 6 1.3 Arabidopsis plants transformed with GFP-Ara7 (a Rab5 gold particles to be targeted to the cell plate and sur- homolog labeling endosomes) (Geldner et al., 2003; rounding endosomes, whereas it revealed 1.6 6 0.5 gold Ueda et al., 2001, 2004) and ST-YFP (sialyl transferase la- particles to be cytoplasmic. Analysis of different cell beling trans-Golgi) (Boevink et al., 1998; Grebe et al., plate stages revealed that both the cell plate and cell 2003). The authenticity of endosomal and Golgi labeling plate vesicles contained these endosomal markers dur- in BY-2 and Arabidopsis cells was thoroughly investi- ing initiation (Figures 4A and 4B) and expansion (Fig- gated and confirmed in interphase cells by using a variety ure 4C) of cell plates. The gold label was mostly associ- of colocalization and drug studies (see Figure S3, Sup- ated with the surface of electron-transparent vesicles plemental Data, and Movies S4 and S5). In dividing and less with electron-dense cell plate vesicles with di- BY-2 and Arabidopsis cells during spindle assembly, ameters between 50 and 65 nm. The ultrastructural anal- the endosomes were always distributed throughout the ysis shows that both cell plate matrix and cell plate spindle area, and their apparent fusion coincided and co- vesicles contain endosomal markers Ara6 and Ara7 localized with cell plate initiations (Figures 3A–3F and (Grebe et al., 2003; Ueda et al., 2001, 2004). Similar re- 3H; Figure S4A; Movies S6–S8). In contrast, trans-Golgi sults were obtained when antibodies, which recognize remained around the spindle apparatus and never local- parental CW pectins, were used for immunogold label- ized into its inside space during cell plate initiation (Fig- ing (Baluska et al., 2005). Together, these results reveal ures 3A and 3E; Movie S6 and S8). The trans-Golgi moved that endosomal vesicles are included in initiating and ex- inward toward assembling cell plates only after cell panding cell plates, whereas fusion of endomembranes plates were initiated (Figure S4A; Movies S6–S8) and labeled with a trans-Golgi marker for cell plate initiation likely supplied exocytic products needed for their expan- is not detected. This again highlights the involvement of sion. During these later stages, endosomal and trans- endocytosis in cell plate formation. Golgi organization around the cell plate appeared more complex and interconnected (Figure 3G; Movie S9). Cytokinesis-Specific Syntaxin KNOLLE Localizes The secretory inhibitor BFA has been used in many to Endosomes studies to visualize endosomal aggregations (Baluska The cytokinesis-specific syntaxin KNOLLE is one of the et al., 2002; Geldner et al., 2001, 2003; Grebe et al., best-characterized mediators of vesicle fusion for cell 2002, 2003; Murphy et al., 2005; Paciorek et al., 2005; plate formation (Lauber et al., 1997; Lukowitz et al., Samaj et al., 2004). Therefore, we used BFA to better vi- 1996). Surprisingly, not only the cell plate-based, but sualize endosomes and investigate whether endosomal also the intracellular KNOLLE signal, colocalized with aggregations can initiate and expand cell plates. When endosomal markers Ara7 (Figure 5B) and Ara6 (Figures BFA was applied to either BY-2 or Arabidopsis cells, 5C and 5D). The endosomal identity of the KNOLLE- FM4-64-labeled endosomal aggregates (but not the containing intracellular structures was further confirmed ST-YFP-labeled trans-Golgi) were observed to partici- by wortmannin treatment. Wortmannin is known to in- pate both in cell plate initiation (Figure 4D; Figure S6A; duce the swelling of specific endosomal structures Movies S10 and S11) and expansion (Figure 4D; Figures (Vieira et al., 2003) (see Figures S3Q–S3S). Significantly, S6B and S6C; Movies S10 and S11). These in vivo real- KNOLLE displayed its localization on these swelled time observations were further confirmed by immuno- vesicles (Figure 5E) together with endosomal marker localization studies in Arabidopsis. In fixed Arabidopsis Ara7 (see inset in Figure 5E). In the case of BFA-treated roots, the cis-Golgi marker g COP (Geldner et al., 2003) cells, in which the cell plates have already been initiated, (Figure S4B) and the trans-Golgi network (TGN) marker KNOLLE mainly localized to the interior of BFA compart- TLG2a (Bassham et al., 2000; Geldner et al., 2003) ments that fused with the edges of the cell plate (Fig- (Figure S4C) did not localize significantly to the KNOLLE- ure 5H). The interior of these compartments also contains positive cell plate. However, all endosomal markers endosomal markers such as GNOM-Myc (Figure S6G) tested, such as GNOM-Myc (Geldner et al., 2003), Ara6 and recycling PM proteins such as PIN2 (Figure S6F), (Grebe et al., 2003; Ueda et al., 2001, 2004), and Ara7 the PM water channel PIP2 (Figure S6H), and PM-local- (Geldner et al., 2003; Ueda et al., 2001, 2004) clearly lo- ized LTI6b (Figure S6I). In contrast, g COP-labeled cis- calized to the forming cell plates (Figure 5). To examine Golgi localized to the periphery (Figure 5H). Also, in the localization of endosomal markers to cell plates at maize, BFA treatment induced formation of large blobs the substructural level, we prepared ultrathin sections and colocalization of CW/endocytic pectin RGII and from high-pressure, fixed-freeze substituted Arabidop- PIN1 at the expanding peripheries of the plates resem- sis roots and subjected them to immunogold labeling bling ‘‘cell plate edge BFA compartments’’ (Figures S6D with Ara6 and Ara7 antibodies that have been exten- and S6E). Together, these results suggest that during sively tested before (Ueda et al., 2001). This analysis cytokinesis KNOLLE localizes to endosomes, and fusion showed Ara6 (Figures 4A–4C) and Ara7 (Figure S5A) la- of these endosomes initiates or expands cell plates. beling at the cell plate matrix and associated vesicles, It has been shown that during preprophase (well whereas no labeling was detected when primary anti- before cell plate initiation) KNOLLE accumulates intra- bodies were omitted (Figure S5B). Quantitative evalua- cellularly, and that during cytokinesis it is included in the tion of Ara6 immunogold labeling in dividing cells (n = 6 cell plate (Lauber et al., 1997). Accordingly, in premitotic in three independent experiments) revealed 14.3 6 2.2 cells, KNOLLE colocalized with endosomal Ara6 (Fig- (mean 6 standard error) gold particles to be targeted ure 5D), and upon BFA treatment KNOLLE colocalized Developmental Cell 142

Figure 3. Fusion of Endosomes Initiates and Expands the Cell Plates (A–H) (A–G) Tobacco BY-2 suspension cells. (H) Arabidopsis root cells. (A) ST-YFP-labeled trans-Golgi (green) and FM4-64-labeled endosomes (red) during the spindle stage. Note that Golgi apparatus remain around the spindle and that FM4-64-labeled endosomes occupy the inside space. (B) FM4-64-labeled endosomes (red) and ST-YFP-labeled trans-Golgi (green) during cell plate initiation. Note that endosomes fuse to initiate the cell plate, whereas trans-Golgi remain away from the initiating cell plate. (C and D) GFP-Ara7- (green) and FM4-64-labeled (red) endosomes during (C) spindle and (D) cell plate initiation. Note that the fusion of FM4-64- and Ara7-stained endosomal structures initiates the cell plate (see also Movie S7). (E and F) GFP-Ara7-labeled endosomal (green), ST-YFP-labeled trans-Golgi (blue), and FM4-64-labeled endosomal (red) localizations during (E) spindle and (F) cell plate initiation (see also Movie S8). (G) Close-up view of one side of an expanding cell plate, with the above triple markers revealing complex interactions between trans-Golgi, endosomes, and the cell plate during its expansion (see also Movie S9). (H) Colocalization of GFP-Ara7- (green) and FM4-64-labeled (red) endosomes on the forming cell plates. The scale bar in (G) is 3 mm, and the scale bars in (A)–(F) and (H) are 5 mm. with internalized PIN1 (Figure 5F), PIN2 (Figure S6F), and marker 58K (Baluska et al., 2002)(Figure 5I). Together, the PM water channel PIP2 (Figure S6H) in BFA com- these results demonstrate that the regulator of cell plate partments that were surrounded by g COP-labeled cis- formation KNOLLE at least partially localizes to endo- Golgi (Figure 5G). We further conﬁrmed endosomal lo- somes before cell plate initiation. calization of KNOLLE by biochemical means by using protein cofractionation analysis. On sucrose gradients, Endocytosis Rate Is Increased during Cytokinesis KNOLLE cofractionated with endosomal markers Ara7 Since our data showed that cell plate material is rap- and FYVE-GFP (Samaj et al., 2005), but not with Golgi idly derived from the cell surface by endocytosis, we Endocytosis Drives Plant Cytokinesis 143

Figure 4. Contribution of Endosomes to Cell Plate Formation (A–C) Electron micrographs showing immunogold (10 nm) localization of Ara6 in cytokinetic Arabidiopsis root cells. Overview of Ara6 immunogold labeling in the (A) early and (C) late cell plate. Note the Ara6 label on the cell plate vesicles of different electron densities within the developing cell plate matrix (marked by yellow outline). (B) A detailed view showing the Ara6 label on the surface of a marginally expanding cell plate and an electron-transparent round-shaped cell plate vesicle with a diameter of 65 nm. (D) 50 mM BFA treatment causes aggregation of FM4-64-labeled endosomes (surrounded by ST-YFP-labeled trans-Golgi) to form BFA compartments that participate in both cell plate initiation (see arrowheads) and expansion (see arrows). Note that all of the BFA compartments fuse to the cell plates. Time is given in minutes (see also Movie S11). The scale bar is 5 mm.

determined the rate of endocytosis in interphase and Inhibition of Endocytosis Interferes with Cell mitotic cells by quantifying FM4-64 internalization in Plate Formation Arabidopsis cells. In interphase cells, internalization Our results strongly suggest the importance of endo- of FM4-64 was at least 2-fold lower compared to that cytosis for cell plate formation. To test this notion, we of cytokinetic cells (Figures 6A and 6B) (i.e., even ex- examined the impact of inhibition of endocytosis on cy- cluding the cell plate-localized FM4-64 signal during tokinesis. quantiﬁcation of FM4-64 intensity). In support, it has The endocytic pathway can be inhibited by cold treat- been shown that expression of endosomal regulator ment, wortmannin [affecting endosome morphology Ara7 is upregulated during cell division (Ueda et al., (Vieira et al., 2003 and this study)], and sodium azide 2001). These observations show that the rate of endo- (Parton et al., 2001). All of these treatments reversibly cytosis is cell cycle-dependent and upregulated during slowed down or blocked FM4-64 internalization and cytokinesis. also cell plate formation (not shown). To interfere with Developmental Cell 144

Figure 5. Localization of KNOLLE to Endo- somes (A–C) Colocalization of the endosomal markers GNOM-Myc (red in [A]), Ara7 (green in [B]), and Ara6 (green in [C]) and KNOLLE (green in [A] and red in [B] and [C]) on cell plates and cytosolic punctae. (D) Colocalization of Ara6 (green) and KNOLLE (red) on cytosolic punctae in a premitotic cell. (E) After wortmannin treatment, KNOLLE (red) labels enlarged vesicles and colocalizes with Ara7 (green) on them (inset). (F) Colocalization of KNOLLE (red) with PIN1 (green) in the BFA compartment formed in premitotic cell. (G and H) KNOLLE (red) labels the BFA compartment/cell plate that is surrounded by cis- Golgi (green) in (G) premitotic and (H) cytokinetic cells. (I) Protein cofractionation Western blot showing KNOLLE cofractionating with endosomal, but not Golgi, markers. DAPI is in blue in (C), (D), and (H).

Figure 6. Endocytosis during Cell Plate For- mation (A and E–K) Arabidopsis root cells and (C and D) tobacco BY-2 suspension cells. (A and B) Quantiﬁcation of FM4-64 endocytosis in interphase and mitotic Arabidopsis cells. A rep- resentative region of interest used for quan- tiﬁcation is depicted by a white circle. The columns show the means, whereas the error bars depict standard deviations (C) After internalization, FM4-64 (red) colocalizes with GFP-Ara7-labeled endosomes. (D) DNAra7 (green) inhibits endocytosis of FM4-64 (red). (C) Note that the GFP-DNAra7 labels mainly the cytosol, which is in contrast to control GFP-Ara7. (E) The phenotype of wild-type and RPS5A>>GFP-DNAra7-expressing plants grown on MS plates for 3 weeks (the bars on the left of the panel indicate millimeters). (F and G) A DIC image of (F) wild-type and an (G) RPS5A>>GFP-DNAra7-expressing plant root tip. (H and I) Enlarged multinucleate cells observed with (H) DIC and (I) DAPI staining. (J) Binucleate cell expressing RPS5A>> GFP-DNAra7. (K) Formation of partial CW attached on its one side to the parental CW in the case of tamoxifen-induced GFP-DNAra7. Note that even after 30 min, FM4-64 hardly goes inside the cell with unnoticeable cell plate expansion. The scale bars are 5 mm. Endocytosis Drives Plant Cytokinesis 145

endocytosis more specifically, we genetically manipu- contribution of other pathways such as endocytosis lated Rab5 (Ara7) function. In mammalian cells, Rab5 was not seriously considered (Alberts et al., 2002; Jur- (ortholog of Ara7) is required for early endosome fusion gens, 2005). However, material already delivered for in vitro (Gorvel et al., 1991; Zerial and McBride, 2001), and cell plate initiation has been proposed to be retrieved it regulates transport between the PM and early endo- from the assembled cell plate area and reused together somes, in vivo (Bucci et al., 1992; van der Bliek, 2005; with Golgi apparatus-derived exocytic vesicles at its ex- Zerial and McBride, 2001). Proper functioning of Rab5 panding edges (Bednarek and Falbel, 2002). Conse- depends on its alternating GTP and GDP bound states. quently, it is believed that the cell plate consists of de By a specific mutation, Rab5 can be locked in its GDP novo-synthesized material. In this study, we show by state, which interferes with its function since it keeps multiple independent approaches that, in both monocot Rab5 cytosolic rather than in its active state on mem- and dicot plants, endocytic delivery of the cell surface branes (Stenmark et al., 1994). It has been shown before material significantly contributes to cell plate initiation that GDP-locked Rab5 interferes with endocytosis (van and expansion. der Bliek, 2005). To test the effects of this genetic inter- ference with endocytosis on cell plate formation, we Endocytosed PM and CW Material Is Used made a dominant-negative (DN) GDP-locked state Ara7 for Cell Plate Formation by replacing S24N (Stenmark et al., 1994) and introduced Our combined results from Arabidopsis, maize, and it into cowpea protoplasts and into BY-2 cells by trans- tobacco cells clearly show that material originating fection. In both cell systems, DN-Ara7 appeared in cyto- from the cell surface is delivered to the forming cell sol (compare Figure 6C with Figure 6D and Figure S7A plate. Live cell studies revealed endocytic delivery of with Figure S7C), and FM4-64 internalization was drasti- the exogenously applied endocytic tracer FM4-64 as cally reduced (compare Figure 6C with Figure 6D and well as extracellular fluid markers. Targeting of FM4-64 Figure S7B with Figure S7D), confirming an inhibitory ef- to cell plates has been shown before in Fucus zygotes, fect of DN-Ara7 on endocytosis. The extent of endocyto- with the interpretation of FM4-64 targeting first to the sis inhibition by DN-Ara7 was always correlated to its ex- Golgi and then being delivered from Golgi to cell plates pression level (not shown). (Belanger and Quatrano, 2000). We and others (Grebe To investigate the effects of DN-Ara7 expression on et al., 2003; Tse et al., 2004) observed no FM4-64 label- cell plate formation in Arabidopsis, we attempted to ing of the Golgi in a time span when FM4-64 does inter- obtain DN-Ara7-transformed plants. We failed to re- nalize from the parental PM and labels forming cell cover Arabidopsis plants transformed with a 35S::DN- plates. Our results of FM4-64 labeling to the cell plates Ara7 (not shown). Therefore, we expressed DN-Ara7 in BFA-treated BY-2 cells (where Golgi cisternae are re- only within specific cell types (by using a two-compo- sorbed into endoplasmic reticulum [ER]) and Arabidop- nent transactivation system) (Haseloff, 1999) or condi- sis root cells (where Golgi aggregates around the BFA tionally (by using a tamoxifen-inducible expression sys- compartments) further support direct delivery of FM4- tem) (Friml et al., 2004). DN-Ara7 expression was driven 64 to the cell plate without its intermediate targeting to by the Arabidopsis ribosomal protein S5A (RPS5A) the Golgi. Experiments with FRAP analysis or 3D recon- promoter, which is active specifically in dividing cells struction suggest that the endocytic markers do not ar- (Weijers et al., 2001). RPS5A-transactivated plants rive to the forming cell plate by a direct transient con- were selected based on their resistance to the selec- nection between the cell plate and the parental PM. In tion marker as well as on the basis of the expression addition, due to the high plant cell turgor pressure, of coregulated GFP, which was mainly found in the such direct connections, if present at all below the con- root and shoot meristems as reported before (Weijers focal microscope detection limit, are energetically un- et al., 2001). The positive plants expressing RPS5A>> favorable and thus extremely unlikely. DN-Ara7 remained small compared to the controls even Also, PM-localized proteins such as PIN1 (Geldner 3 weeks after germination (Figure 6E). In addition, these et al., 2001) and the SNAP25 homolog AtSNAP33 (Heese plants displayed abnormalities in their root tip architec- et al., 2001) localize at cell plates. Our immunolocaliza- ture (compare Figure 6G with Figure 6F). Importantly, tion studies incorporating other PM-localized proteins closer examination revealed the presence of bi- and and CW-originated pectins revealed their strong accu- multinucleate cells (Figures 6H–6J), demonstrating cy- mulation within both initiating and expanding cell plates. tokinesis defects. Also, when DN-Ara7 expression was Biochemical studies revealed that RGII dimers are induced by tamoxifen treatment, endocytosis, as visu- formed via borate-diester crosslinks between apiose alized by FM4-64 uptake (Figure 6K), but not exocyto- residues in the RGII molecule localized predominantly sis (Figure S8), was dramatically reduced. Concomi- within CWs (O’Neill et al., 2004; Ridley et al., 2001). Im- tantly, characteristic cytokinesis defects such as the portantly, in vitro studies revealed that RGII dimer for- formation of partial CWs and the reduction of cell plate mation is a spontaneous self-assembly process driven expansion were induced (Figure 6K). Together, these by acidic pH values (5 and less) (Ishii et al., 1999) and experiments clearly show that inhibition of endocytosis high calcium levels (10 mM and more) (Matoh et al., interferes with cytokinesis and cell plate formation. 1998), conditions that are known to be characteristic only for CWs. RGII synthesis and transport into CWs is Discussion accomplished within compartments (Golgi and Golgi- derived vesicles) that do not satisfy requirements for di- So far, the prominent view of cell plate assembly in plant mer formation. The used RGII antibody specifically rec- cells assumed that Golgi apparatus-derived exocytic ognizes RGII-B-RGII dimers (Matoh et al., 1998), which vesicles alone drive plant cytokinesis, and the possible are formed only in muro (within CWs) (Ishii et al., 1999; Developmental Cell 146

Matoh et al., 1998; Ridley et al., 2001). Hence, our results with the RGII antibody suggest a CW origin of endosomal and cell plate RGII-B-RGII dimers. They further may indicate that the cell plate will bear lower pH and higher calcium levels, a finding that is also supported by previous studies showing the influence of the calcium chelator caffeine on cell plate formation (Becerra and Lopez-Saez, 1978). Importantly, cell plates still initiate, and cell surface material is included also when de novo protein synthesis is blocked. Also, earlier studies showed that mitotic pro- cesses that develop after metaphase are independent of fresh protein synthesis (Garcia-Herdugo et al., 1974). This further corroborates that preexisting material is used for cell plate formation. In order to arrive at the cell plate, cell surface material must first be endocytosed. Previous studies on interphase cells have shown that many PM proteins and CW pectins cycle between the PM and intracellular compartments via endocytosis (Baluska et al., 2002; Geld- ner, 2004; Geldner et al., 2001, 2003; Grebe et al., 2002, 2003; Murphy et al., 2005; Paciorek et al., 2005). This is in agreement with the fact that many of these proteins have functional endocytic sorting motifs (Zheng et al., 2002; Zuo et al., 2000). The constitutive endocytic recycling provides a likely means for continuous translo- cation of material from the cell surface to the forming cell plate (see our model in Figure 7).

Endosomes Fuse to Initiate and Expand the Cell Plate The delivery of endocytosed material to the cell plate implies a connection between endosomes and the cell plate. Our live cell studies in BY-2 cells and Arabidopsis Figure 7. Membrane Trafficking Pathways during Cell Plate Forma- with available specific endosomal and Golgi apparatus tion markers (Boevink et al., 1998; Geldner et al., 2003; Grebe The model depicts the mass flow of endocytic material from the PM to the cell plate. ER, Endoplasmic Reticulum; GA, Golgi apparatus; et al., 2003; Ueda et al., 2001, 2004) show that, preferen- TGN, Trans Golgi Network; PM, plasma membrane; EE, early endo- tially, endosomes, but not Golgi-localized endomem- some; RE, recycling endosome; LE, late endosome. The sizes of the branes, aggregate to mark cell plate initiation. Further- arrows show the extent of membrane transport. Red arrows indicate more, blocking Golgi apparatus-dependent secretion a novel pathway revealed by our results. Red lines indicate inhibi- in BY-2 cells by BFA (Ritzenthaler et al., 2002) does tory effects of BFA. not prevent cell plate initiation, whereas endosomal aggregates induced by BFA treatment initiate or expand cent detailed EM tomography revealing the existence of the cell plate (this study). During cell plate expansion, two kinds of vesicles during cell plate formation (Segui- endosomal membranes continue to be targeted to the Simarro et al., 2004) and the presence of MVB at the con- cell plate (revealed by real-time and FRAP analysis), verging point of exocytic and endocytic pathways would and Golgi apparatus can be observed in close vicinity be consistent with a view that both endosome- and to expanding cell plates. In support, (1) available endo- Golgi apparatus-derived vesicles are targeted to the somal markers localize to the cell plate, as visualized cell plate. at the ultrastructural level, (2) an earlier study docu- The SNARE hypothesis for vesicle fusion predicts mented the fusion of coated vesicles to initiate cell the involvement of three proteins, namely, Syntaxin plates (Frank and Herth, 1974), and (3) another endoso- (Q-SNARE), SNAP25, and R-SNARE in the formation of mal structure, the multivesicular body (MVB), has been the SNARE complex required for proper vesicle fusion observed to fuse with initiating cell plates (Lehmann (Chen and Scheller, 2001). A crystal structure of neuro- and Schulz, 1969), to attach to the expanding cell plate nal SNARE complexes shows a bundle of four helices (Otegui et al., 2001), or to remain in very close proximity for which Q- and R-SNARE each contributes one helix to the cell plate (Segui-Simarro et al., 2004). Our data and SNAP25 provides two helices (Chen and Scheller, concerning the effect of wortmannin (inhibiting MVB 2001). The cytokinesis-specific syntaxin KNOLLE (Q- formation [Fernandez-Borja et al., 1999]) on cell plate SNARE) and its interactor SNAP25 homolog AtSNAP33 formation and newly emerging genetic studies indicat- are supposed to play a role in the fusion of vesicles at ing a requirement of the MVB components ESCRT-I the cell plate (Heese et al., 2001; Lukowitz et al., 1996). (Schellman and Hu¨ lskamp 2005) and ESCRT-III (Winter Furthermore, a plant-specific Q-SNARE NPSN11 local- et al., 2005; Schlager et al., 2005) for proper cytokinesis izes at the cell plate, interacts with KNOLLE, and co- support the role of MVB in cytokinesis. In addition, a re- localizes with KNOLLE-positive cytosolic structures Endocytosis Drives Plant Cytokinesis 147

(Zheng et al., 2002). Findings that (1) KNOLLE at least either bypass the endosomal compartment or result partially localizes to endosomes and cofractionates in Golgi apparatus-mediated direct ‘‘intracellular cell with endosomal, but not with Golgi, markers (this study), plate-targeted exocytosis’’, or it may merge the secre- (2) AtSNAP33 cycles between the PM and endosomes tory and endocytic cargo for fast lateral growth of the ex- (Heese et al., 2001), and (3) NPSN11 (together with panding cell plate (see our model in Figure 7). In accor- KNOLLE) neither localizes nor cofractionates with Golgi dance, recent discoveries in the mammalian field show membranes (Zheng et al., 2002) strongly suggest the im- that endocytosis is required for cytokinesis (Schweitzer portance of endosomes for vesicle fusion contributing et al., 2005), and recycling endosomes (receiving mate- to cell plate formation. In support of this, viral movement rial from both early endosomes and trans-Golgi) are proteins (which have recently been shown to traffic via crucial for proper membrane addition at the cleavage the endocytic recycling pathway [Haupt et al., 2004]) di- furrow during cytokinesis (Albertson et al., 2005). In rectly interact with KNOLLE and also localize to the cell C. elegans embryos, inhibition of recycling endosomal plate (Laporte et al., 2003). Requirement of a specialized Rab11 does not interfere with the initial cytokinesis R-SNARE has been implicated mostly in heterotypic events, but it affects late events of cytokinesis. In con- membrane fusion events in which a vesicular R-SNARE trast, in Drosophila embryos, injection of dominant- recognizes the Q-SNARE present on the target mem- negative Rab11 causes cleavage furrow defects, specif- brane to which the vesicle fuses. The existence of ically during the initial phase (Albertson et al., 2005). In KNOLLE-interacting, plant-specific Q-SNARE NPSN11 addition, many aspects of the cell plate formation in and the localization of KNOLLE (and possibly that of plants display homology to endosomal recycling path- NPSN11) on endosomes may implicate homotypic fu- ways implicated in calcium-regulated PM repair of torn sion of endosomes as a mechanism for cell plate forma- cell periphery (McNeil and Steinhardt, 2003) and those tion and its subsequent attachment to the parental PM. involved in generating secretory lysosomes (Blott and Griffiths, 2002). In this view, the cell plate can be re- Endocytosis Is Upregulated during and Is Required garded as some sort of specialized endosomal com- for Cell Plate Formation partment that receives material from the Golgi appara- During cell plate formation, almost one-third of the cell tus directly as well as from the PM through sorting and surface has to be created in a matter of minutes. This re- recycling endosomes (see our model in Figure 7). Our quires a huge capacity of the material delivery mecha- view of cell plate formation involving multiple membrane nisms. Consistently, our measurements of the rate of en- fusion events and cargo received from multiple path- docytosis reveal a strong upregulation of endocytosis ways would also provide more stability and back up during cell plate formation. Conversely, inhibition of en- for this process. One can imagine that inhibition of one docytosis by using cold and drug treatments or, more pathway may temporarily upregulate the other to rescue specifically, by genetic manipulation of endosomal the defects. This would explain the robustness of the cy- Rab Ara7, slows down or even abolishes cell plate for- tokinetic machinery, which is indicated by multiple ob- mation and leads to phenotypic aberrations indicative servations of CW stubs and normal, complete CWs in of cytokinesis defects. This reveals that endocytosis various reported cytokinetic mutants. bears a rate-limiting effect on cell plate formation. Exocytosis and endocytosis are supposed to be bal- Experimental Procedures anced to maintain the cell surface area constant. None- theless, BFA inhibits secretion in a matter of minutes in Plant Material and Growth Conditions BY-2 cells, as the Golgi cisternae get resorbed into the Maize roots were obtained as described before (Baluska et al., ER, and it affects recycling endosomes carrying pro- 2002). Tobacco BY-2 cells were cultured as reported previously teins such as PIN1 to the plasma membrane in Arabi- (Dhonukshe and Gadella, 2003). Cowpea plants were grown, and protoplasts were isolated as described before (Dhonukshe and Ga- dopsis (Samaj et al., 2005). BFA allows endocytosis to della, 2003). In the case of Arabidopsis, 4-day-old seedlings grown be continued in both systems (with it even being en- in growth chambers at 23ºC on vertically oriented plates containing hanced in BY-2 cells) without plasmolysis (Emans Arabidopsis medium (AM, 1/2 MS agar, 1% sucrose [pH 5.8]) were et al., 2002; Geldner et al., 2001; Paciorek et al., 2005). used. The details of the two-component transactivation system have Apparently, there is some excess material at the PM been described before (Friml et al., 2004). that still can be internalized when exocytosis is blocked Construction of Reporter Genes and Transgenics without compromising the cell surface integrity. Construction of GFP-MAP4 was described before (Dhonukshe and Gadella, 2003). Please see Supplemental Data for cloning details Biological Implications of Endocytic Involvement of GFP-Ara7, constrcution of DN-Ara7, and its transfection, transfor- for Cell Plate Formation mation, and transactivation. Transgenic Arabidopsis lines carrying Considering the fast speed of cell plate formation, the the ADL1A-GFP (Kang et al., 2003a), BRI1-GFP (Friedrichsen et al., use of prefabricated building blocks delivered via endo- 2000), ST-YFP (Grebe et al., 2003), and GNOM-myc (Geldner et al., 2003) gene fusions have been described before. cytosis is highly efficient. Besides speeding up the maturation of the cell plate by delivery of fully matured pro- Fluorescent Dyes and Inhibitor Treatments teins and CW molecules, the presence of endocytosed FM4-64 (Molecular Probes) dissolved in water was applied at a 2 mM PM components would also contribute to defining the final concentration for 5 min to BY-2 cells or Arabidopsis plants, and nature of the cell plate membrane as a future PM. This both were washed with medium to remove excess dye and were may be instrumental for delivery of regular secretory observed immediately. Alexa 633 (Molecular Probes; catalog no. A30634) dissolved in water was applied at a 2 mM final concentration, Golgi apparatus-based vesicles as well, which would al- and cells were observed immediately. BFA (Sigma), wortmannin ready at early stages of cell plate formation recognize (Sigma), LY294002 (Sigma), and CHX (Sigma) were used from this as a target membrane. Such a mechanism would DMSO-dissolved stock solutions and were applied to cells at final Developmental Cell 148

concentrations of 50 mM, 10 mM, 10 mM, and 50 mM, respectively, for COP antibody, Akihiko Nakano for the GFP-Ara7 construct and the indicated periods. LY (Sigma) was used at a 1% concentration. For Ara7 and Ara6 antibodies, W. Michalke for the PM-ATPase antibody, visualizing nuclei in live Arabidopsis cells, the seedlings were J. Carette for the ST-YFP construct, and Gerd Ju¨ rgens for the incubated with 40-6-diamidino-2-phenylindole (DAPI) solution. So- KNOLLE antibody and the GNOM-myc-transformed Arabidopsis dium Azide (Sigma) was diluted in water (100 mM). Aniline blue dis- line. We also thank Dolf Weijers for the RPS5A activator line, Remko solved in Gly/NaOH buffer (pH 9.5) was used at a 0.1% concentration Offringa for the Tamoxifen induction activator line, Markus Grebe for (w/v). Incubation of Arabidopsis seedlings with various chemicals the ST-YFP-transformed Arabidopsis line, and David Ehrhardt for was carried out in 24-well cell culture plates in liquid AM medium. the PIP2- and LTI6b-transformed Arabidopsis lines. We thank Tom- In the case of Arabidopsis, these experiments were performed in asz Paciorek for initial help with Arabidopsis immunolocalizations, triplicate with at least 60 roots evaluated in total for each experimen- Ursula Mettbach for excellent technical assistance with EM, and tal condition, the experiments with BY-2 cells were performed at other members of all of the labs for their timely help. P.D. and least four times with hundreds of cells, and those with maize were T.W.J.G. were supported by the Netherlands Organization for Scien- repeated three times. tific Research (NWO FOM-ALW) 805.47.012, by NWO van der Leeuw 835.25.004, and by the European Union integrated project on molec- Immunofluorescence Microscopy ular imaging (LSHG-CT) 2003-503259; F.B, A.H., and M.S. were sup- Whole-mount immunofluorescence preparations for Arabidopsis ported by the European Commission Human Potential Programme (Friml et al., 2003) and antibody staining of maize tissue sections (HPRN-CT) 2002-00265, J.F. and P.D. were supported by Volkswa- (Baluska et al., 2002) were done as described before. Please see genstiftung to J.F., and F.B. and J.S. were supported by Grant Agen- Supplemental Data for the immunolocalizations in BY-2 cells and an- cies Vega (Project Nr.2/5085/25) and Slovak grant Agency for Sci- tibody concentrations. The cells were then viewed either with a Zeiss ence and Technique ([APVT], Bratislava [APVT-51-002302]). Axiovert 405M microscope equipped with a Zeiss AxioCam HR dig- ital camera or with a Leica TCS-SP2 AOBS CLSM by using a 488 nm Received: June 22, 2005 Ar/Kr laser line exciting FITC, a 561 nm Diode laser line exciting Cy3, Revised: September 15, 2005 and a 405 nm Diode laser line exciting DAPI in a sequential scan Accepted: November 14, 2005 mode. FITC emission was captured with a 500–525 nm slit detector Published: January 9, 2006 setting, Cy3 was captured with a 570–640 nm slit detector setting, and DAPI was captured with a 410–460 nm slit detector setting. Ac- quired images were processed with Adobe Photoshop 7. All of the References experiments were repeated at least three times. Alberts, B., Johnson, A., Lewis, J., Raff, M., Roberts, K., and Walter, Cofractionation Analysis P. (2002). Molecular Biology of the Cell, Fourth Edition, (New York: Please see Supplemental Data for details of cofractionation analy- Garland Science). sis. Albertson, R., Riggs, B., and Sullivan, W. (2005). Membrane traffic: a driving force in cytokinesis. Trends Cell Biol. 15, 92–101. Live Cell Analysis For live cell analysis of BY-2 cells, the Zeiss CLSM510 system imple- Assaad, F.F., Huet, Y., Mayer, U., and Jurgens, G. (2001). The cyto- mented on an inverted (Axiovert 100) microscope was used. The mi- kinesis gene KEULE encodes a Sec1 protein that binds the syntaxin croscopy system, sample preparation, single wavelength scanning, KNOLLE. J. Cell Biol. 152, 531–543. image processing, and movie generation were described before Baluska, F., Hlavacka, A., Samaj, J., Palme, K., Robinson, D.G., (Dhonukshe and Gadella, 2003). Dual and triple color imaging was Matoh, T., McCurdy, D.W., Menzel, D., and Volkmann, D. (2002). performed by using dual or triple excitation/emission scanning, re- F-actin-dependent endocytosis of cell wall pectins in meristematic spectively, in multitracking mode (for further details, please see Sup- root cells. Insights from brefeldin A-induced compartments. Plant plemental Data). For live Arabidopsis root tip cells analysis, the Leica Physiol. 130, 422–431. TCS-SP2 AOBS CLSM was used (for further details, please see Sup- Baluska, F., Samaj, J., Hlavacka, A., Kendrick-Jones, J., and Volk- plemental Data). Fluorescence recovery after photobleaching mann, D. (2004). Actin-dependent fluid-phase endocytosis in inner (FRAP) analysis is described in Supplemental Data. For quantifica- cortex cells of maize root apices. J. Exp. Bot. 55, 463–473. tion of FM4-64 internalization, w25 mm2 circular areas (representing 2000 pixels) were marked at various positions in interphase, and cy- Baluska, F., Liners, A., Hlavacka, A., Schlicht, M., van Custem, P., tokinetic (excluding the cell plates in cytokinetic cells) cells and pixel McCurdy, D.W., and Menzel, D. (2005). Cell wall pectins and xyloglu- intensities were acquired. For quantification of FM4-64 internaliza- cans are internalized into dividing root cells and accumulate within tion, 40 interphase and 40 cytokinetic cells were measured, and cell plates during cytokinesis. Protoplasma 225, 141–155. the values were gathered and processed in Microsoft Excel. Bassham, D.C., Sanderfoot, A.A., Kovaleva, V., Zheng, H., and Rai- khel, N.V. (2000). AtVPS45 complex formation at the trans-Golgi net- Differential Interface Contrast Microscopy work. Mol. Biol. Cell 11, 2251–2265. For differential interface contrast (DIC) microscopy, seedlings were Becerra, J., and Lopez-Saez, J.F. (1978). Effects of caffeine, calcium placed in 2% glycerol and were observed with a Zeiss axioplan II mi- and magnesium on plant cytokinesis. Exp. Cell Res. 111, 301–308. croscope equipped with DIC optics and a CCD camera. Bednarek, S.Y., and Falbel, T.G. (2002). Membrane trafficking during Electron Microscopy plant cytokinesis. Traffic 3, 621–629. Immunogold EM labeling was performed as described before Belanger, K.D., and Quatrano, R.S. (2000). Membrane recycling oc- (Reichelt et al., 1999), with few modifications (see Supplemental curs during asymmetric tip growth and cell plate formation in Fucus Data). distichus zygotes. Protoplasma 212, 24–37. Blott, E.J., and Griffiths, G.M. (2002). Secretory lysosomes. Nat. Rev. Supplemental Data Mol. Cell Biol. 3, 122–131. Supplemental Data including additional text for the Results section, additional information on Experimental Procedures, 8 figures, and Boevink, P., Oparka, K., Santa Cruz, S., Martin, B., Betteridge, A., 11 movies are available at http://www.developmentalcell.com/cgi/ and Hawes, C. (1998). Stacks on tracks: the plant Golgi apparatus content/full/10/1/137/DC1/. traffics on an actin/ER network. Plant J. 15, 441–447. Bucci, C., Parton, R.G., Mather, I.H., Stunnenberg, H., Simons, K., Acknowledgments Hoflack, B., and Zerial, M. (1992). The small GTPase rab5 functions as a regulatory factor in the early endocytic pathway. Cell 70, 715– We thank Toru Matoh for the RGII antibody, Klaus Palme for the 728. maize PIN1 antibody, Sebastian Bednarek for the ADL1 antibody, Chen, Y.A., and Scheller, R.H. (2001). SNARE-mediated membrane Christian Luschnig for the PIN2 antibody, David Robinson for the g fusion. Nat. Rev. Mol. Cell Biol. 2, 98–106. Endocytosis Drives Plant Cytokinesis 149

Cutler, S.R., and Ehrhardt, D.W. (2002). Polarized cytokinesis in vac- Hinshaw, J.E. (2000). Dynamin and its role in membrane fission. uolate cells of Arabidopsis. Proc. Natl. Acad. Sci. USA 99, 2812– Annu. Rev. Cell Dev. Biol. 16, 483–519. 2817. Ishii, T., Matsunaga, T., Pellerin, P., O’Neill, M.A., Darvill, A., and Cutler, S.R., Ehrhardt, D.W., Griffitts, J.S., and Somerville, C.R. Albersheim, P. (1999). The plant cell wall polysaccharide rhamnoga- (2000). Random GFP::cDNA fusions enable visualization of subcellu- lacturonan II self-assembles into a covalently cross-linked dimer. lar structures in cells of Arabidopsis at a high frequency. Proc. Natl. J. Biol. Chem. 274, 13098–13104. Acad. Sci. USA 97, 3718–3723. Jurgens, G. (2005). Cytokinesis in higher plants. Annu. Rev. Plant Dhonukshe, P., and Gadella, T.W., Jr. (2003). Alteration of micro- Biol. 56, 281–299. tubule dynamic instability during preprophase band formation re- Kang, B.H., Busse, J.S., and Bednarek, S.Y. (2003a). Members of the vealed by yellow fluorescent protein-CLIP170 microtubule plus- Arabidopsis dynamin-like gene family, ADL1, are essential for plant end labeling. Plant Cell 15, 597–611. cytokinesis and polarized cell growth. Plant Cell 15, 899–913. Emans, N., Zimmermann, S., and Fischer, R. (2002). Uptake of a fluo- Kang, B.H., Rancour, D.M., and Bednarek, S.Y. (2003b). The dyna- rescent marker in plant cells is sensitive to brefeldin A and wortman- min-like protein ADL1C is essential for plasma membrane mainte- nin. Plant Cell 14, 71–86. nance during pollen maturation. Plant J. 35, 1–15. Fernandez-Borja, M., Wubbolts, R., Calafat, J., Janssen, H., Dive- Kobayashi, M., Nakagawa, H., Asaka, T., and Matoh, T. (1999). cha, N., Dusseljee, S., and Neefjes, J. (1999). Multivesicular body Borate-rhamnogalacturonan II bonding reinforced by Ca2+ retains morphogenesis requires phosphatidyl-inositol 3-kinase activity. pectic polysaccharides in higher-plant cell walls. Plant Physiol. Curr. Biol. 14, 55–58. 119, 199–204. Frank, W.W., and Herth, W. (1974). Morphological evidence for de Laporte, C., Vetter, G., Loudes, A.M., Robinson, D.G., Hillmer, S., novo formation of plasma membrane from coated vesicles in expo- Stussi-Garaud, C., and Ritzenthaler, C. (2003). Involvement of the nentially growing cultured plant cells. Exp. Cell Res. 89, 447–451. secretory pathway and the cytoskeleton in intracellular targeting Frey-Wyssling, A., Lopez-Saez, J.F., and Muhlethaler, K. (1964). For- and tubule assembly of Grapevine fanleaf virus movement protein mation and development of the cell plate. J. Ultrastruct. Res. 10, in tobacco BY-2 cells. Plant Cell 15, 2058–2075. 422–432. Lauber, M.H., Waizenegger, I., Steinmann, T., Schwarz, H., Mayer, Friedrichsen, D.M., Joazeiro, C.A., Li, J., Hunter, T., and Chory, J. U., Hwang, I., Lukowitz, W., and Jurgens, G. (1997). The Arabidopsis (2000). Brassinosteroid-insensitive-1 is a ubiquitously expressed KNOLLE protein is a cytokinesis-specific syntaxin. J. Cell Biol. 139, leucine-rich repeat receptor serine/threonine kinase. Plant Physiol. 1485–1493. 123, 1247–1256. Ledbetter, M.C., and Porter, K.R. (1963). A ‘microtubule’ in plant cell Friml, J. (2003). Auxin transport—shaping the plant. Curr. Opin. Plant fine structure. J. Cell Biol. 19, 239–250. Biol. 6, 7–12. Lehmann, H., and Schulz, D. (1969). Elekronnenmikroskopische Un- Friml, J., Benkova, E., Mayer, U., Palme, K., and Muster, G. (2003). tersuchungen von Differenzierungsvorgangen bei Moosen. II. Die Automated whole mount localisation techniques for plant seedlings. Zellplatten-und Zellwandbildung. Planta 85, 313–325. Plant J. 34, 115–124. Lippincott-Schwartz, J., Yuan, L.C., Bonifacino, J.S., and Klausner, Friml, J., Yang, X., Michniewicz, M., Weijers, D., Quint, A., Tietz, O., R.D. (1989). Rapid redistribution of Golgi proteins into the ER in cells Benjamins, R., Ouwerkerk, P.B., Ljung, K., Sandberg, G., et al. treated with brefeldin A: evidence for membrane cycling from Golgi (2004). A PINOID-dependent binary switch in apical-basal PIN polar to ER. Cell 56, 801–813. targeting directs auxin efflux. Science 306, 862–865. Lukowitz, W., Mayer, U., and Jurgens, G. (1996). Cytokinesis in the Garcia-Herdugo, G., Fernandez-Gomez, M.E., Hidalgo, J., and Lo- Arabidopsis embryo involves the syntaxin-related KNOLLE gene pez-Saez, J.F. (1974). Effects of protein synthesis inhibition during product. Cell 84, 61–71. plant mitosis. Exp. Cell Res. 89, 336–342. Matar, D., and Catesson, A.M. (1988). Cell plate development and Geldner, N. (2004). The plant endosomal system–its structure and delayed formation of the pectic middle lamella in root meristems. role in signal transduction and plant development. Planta 219, Protoplasma 146, 10–17. 547–560. Geldner, N., Friml, J., Stierhof, Y.D., Jurgens, G., and Palme, K. Matoh, T., Takasaki, M., Takabe, K., and Kobayashi, M. (1998). Im- (2001). Auxin transport inhibitors block PIN1 cycling and vesicle traf- munocytochemistry of rhamnogalacturonan II in cell walls of higher ficking. Nature 413, 425–428. plants. Plant Cell Physiol. 39, 483–491. Geldner, N., Anders, N., Wolters, H., Keicher, J., Kornberger, W., McNeil, P.L., and Steinhardt, R.A. (2003). Plasma membrane disrup- Muller, P., Delbarre, A., Ueda, T., Nakano, A., and Jurgens, G. tion: repair, prevention, adaptation. Annu. Rev. Cell Dev. Biol. 19, (2003). The Arabidopsis GNOM ARF-GEF mediates endosomal recy- 697–731. cling, auxin transport, and auxin-dependent plant growth. Cell 112, Meckel, T., Hurst, A.C., Thiel, G., and Homann, U. (2004). Endocyto- 219–230. sis against high turgor: intact guard cells of Vicia faba constitutively Gorvel, J.P., Chavrier, P., Zerial, M., and Gruenberg, J. (1991). rab5 endocytose fluorescently labelled plasma membrane and GFP- controls early endosome fusion in vitro. Cell 64, 915–925. tagged K+-channel KAT1. Plant J. 39, 182–193. Grebe, M., Friml, J., Swarup, R., Ljung, K., Sandberg, G., Terlou, M., Moore, P.J., and Staehelin, L.A. (1988). Immunogold localization of Palme, K., Bennett, M.J., and Scheres, B. (2002). Cell polarity signal- the cell-wall-matrix polysaccharides rhamnogalacturonan I and xy- ing in Arabidopsis involves a BFA-sensitive auxin influx pathway. liglucan during cell expansion and cytokinesis in Trifolium pratense Curr. Biol. 12, 329–334. L.; implications for secretory pathways. Planta 174, 433–445. Grebe, M., Xu, J., Mobius, W., Ueda, T., Nakano, A., Geuze, H.J., Murphy, A.S., Bandyopadhyay, A., Holstein, S.E., and Peer, W.A. Rook, M.B., and Scheres, B. (2003). Arabidopsis sterol endocytosis (2005). Endocytotic cycling of PM proteins. Annu. Rev. Plant Biol. involves actin-mediated trafficking via ARA6-positive early endo- 56, 221–251. somes. Curr. Biol. 13, 1378–1387. Nishihama, R., Soyano, T., Ishikawa, M., Araki, S., Tanaka, H., Haseloff, J. (1999). GFP variants for multispectral imaging of living Asada, T., Irie, K., Ito, M., Terada, M., Banno, H., et al. (2002). Expan- cells. Methods Cell Biol. 58, 139–151. sion of the cell plate in plant cytokinesis requires a kinesin-like pro- Haupt, S., Cowan, G.H., Ziegler, A., Roberts, A.G., Oparka, K.J., and tein/MAPKKK complex. Cell 109, 87–99. Torrance, L. (2004). Two plant-viral movement proteins traffic in the O’Neill, M.A., Eberhard, S., Albersheim, P., and Darvill, A.G. (2001). endocytic recycling pathway. Plant Cell 17, 164–181. Requirement of borate cross-linking of cell wall rhamnogalactur- Heese, M., Gansel, X., Sticher, L., Wick, P., Grebe, M., Granier, F., onan II for Arabidopsis growth. Science 294, 846–849. and Jurgens, G. (2001). Functional characterization of the KNOLLE- O’Neill, M.A., Ishii, T., Albersheim, P., and Darvill, A.G. (2004). Rham- interacting t-SNARE AtSNAP33 and its role in plant cytokinesis. nogalacturonan II: structure and function of a borate cross-linked J. Cell Biol. 155, 239–249. cell wall pectic polysaccharide. Annu. Rev. Plant Biol. 55, 109–139. Developmental Cell 150

Otegui, M.S., Mastronarde, D.N., Kang, B.H., Bednarek, S.Y., and van der Bliek, A.M. (2005). A sixth sense for Rab5. Nat. Cell Biol. 7, Staehelin, L.A. (2001). Three-dimensional analysis of syncytial-type 548–550. cell plates during endosperm cellularization visualized by high reso- Verma, D.P. (2001). Cytokinesis and building of the cell plate in lution electron tomography. Plant Cell 13, 2033–2051. plants. Annu. Rev. Plant Physiol. Plant Mol. Biol. 52, 751–784. Paciorek, T., Zazimalova, E., Ruthardt, N., Petrasek, J., Stierhof, Y.- Vieira, O.V., Bucci, C., Harrison, R.E., Trimble, W.S., Lanzetti, L., D., Kleine-Vehn, J., Morris, D.A., Emans, N., Jurgens, G., Geldner, N., Gruenberg, J., Schreiber, A.D., Stahl, P.D., and Grinstein, S. (2003). and Friml, J. (2005). Auxin inhibits endocytosis and promotes its own Modulation of Rab5 and Rab7 recruitment to phagosomes by phos- efflux from cells. Nature 435, 1251–1256. phatidylinositol 3-kinase. Mol. Cell. Biol. 23, 2501–2514. Parton, R.M., Fischer-Parton, S., Watahiki, M.K., and Trewavas, A.J. Waizenegger, I., Lukowitz, W., Assaad, F., Schwarz, H., Jurgens, G., (2001). Dynamics of the apical vesicle accumulation and the rate of and Mayer, U. (2000). The Arabidopsis KNOLLE and KEULE genes growth are related in individual pollen tubes. J. Cell Sci. 114, 2685– interact to promote vesicle fusion during cytokinesis. Curr. Biol. 2695. 10, 1371–1374. Reichelt, S., Knight, A.E., Hodge, T.P., Baluska, F., Samaj, J., Volk- Wang, Z.Y., Seto, H., Fujioka, S., Yoshida, S., and Chory, J. (2001). mann, D., and Kendrick-Jones, J. (1999). Characterization of the un- BRI1 is a critical component of a plasma-membrane receptor for conventional myosin VIII in plant cells and its localization at the post- plant steroids. Nature 410, 380–383. cytokinetic cell wall. Plant J. 19, 555–567. Weijers, D., Franke-van Dijk, M., Vencken, R.J., Quint, A., Hooykaas, Ridley, B.L., O’Neill, M.A., and Mohnen, D. (2001). Pectins: structure, P., and Offringa, R. (2001). An Arabidopsis Minute-like phenotype biosynthesis, and oligogalacturonide-related signaling. Phytochem- caused by a semi-dominant mutation in a RIBOSOMAL PROTEIN istry 57, 929–967. S5 gene. Development 128, 4289–4299. Ritzenthaler, C., Nebenfuhr, A., Movafeghi, A., Stussi-Garaud, C., Winter, V., Mu¨ ller, S., and Hauser, M.T. (2005). Characterization of Behnia, L., Pimpl, P., Staehelin, L.A., and Robinson, D.G. (2002). HYADE, an Arabidopsis gene involved in cytokinesis. XVII Interna- Reevaluation of the effects of brefeldin A on plant cells using to- tional Botanical Congress, Book of Abstracts, P0068, p. 250. bacco Bright Yellow 2 cells expressing Golgi-targeted green fluores- Yamada, K., Fuji, K., Shimada, T., Nishimura, M., and Hara- cent protein and COPI antisera. Plant Cell 14, 237–261. Nishimura, I. (2005). Endosomal proteases facilitate the fusion of Samaj, J., Baluska, F., Voigt, B., Schlicht, M., Volkmann, D., and endosomes with vacuoles at the final step of the endocytotic path- Menzel, D. (2004). Endocytosis, actin cytoskeleton, and signaling. way. Plant J. 41, 888–898. Plant Physiol. 135, 1150–1161. Yano, K., Matsui, S., Tsuchiya, T., Maeshima, M., Kutsuna, N., Hase- Samaj, J., Read, N.D., Volkmann, D., Menzel, D., and Baluska, F. zawa, S., and Moriyasu, Y. (2004). Contribution of the plasma mem- (2005). The endocytic network in plants. Trends Cell Biol. 15, 425– brane and central vacuole in the formation of autolysosomes in cul- 433. tured tobacco cells. Plant Cell Physiol. 45, 951–957. Samuels, A.L., Giddings, T.H., Jr., and Staehelin, L.A. (1995). Cytoki- Yasuhara, H., Sonobe, S., and Shibaoka, H. (1995). Effects of brefel- nesis in tobacco BY-2 and root tip cells: a new model of cell plate for- din A on the formation of the cell plate in tobacco BY-2 cells. Eur. J. mation in higher plants. J. Cell Biol. 130, 1345–1357. Cell Biol. 66, 274–281. Schellman, S., and Hu¨ lskamp, M. (2005). The Arabidopsis elch Zerial, M., and McBride, H. (2001). Rab proteins as membrane organ- mutant: part of a new ubiquitin-dependent protein degradation izers. Nat. Rev. Mol. Cell Biol. 2, 107–117. pathway in plants? XVII International Botanical Congress, Book of Zheng, H., Bednarek, S.Y., Sanderfoot, A.A., Alonso, J., Ecker, J.R., Abstracts, 5.3.2, p. 79. and Raikhel, N.V. (2002). NPSN11 is a cell plate-associated SNARE Schlager, N., Torres-Acosta, J.A., Winter, V., and Hauser, M.T. protein that interacts with the syntaxin KNOLLE. Plant Physiol. (2005). Characterization of the plant ESCRT-III complex: protein in- 129, 530–539. teraction studies between its putative Arabidopsis thaliana compo- Zuo, J., Niu, Q.W., Nishizawa, N., Wu, Y., Kost, B., and Chua, N.H. nents. XVII International Botanical Congress, Book of Abstracts, (2000). KORRIGAN, an Arabidopsis endo-1,4-b-glucanase, localizes P0021, p. 275. to the cell plate by polarized targeting and is essential for cytokine- Schweitzer, J.K., Burke, E.E., Goodson, H.V., and D’Souza-Schorey, sis. Plant Cell 12, 1137–1152. C. (2005). Endocytosis resumes during late mitosis and is required for cytokinesis. J. Biol. Chem. 280, 41628–41635. Segui-Simarro, J.M., Austin, J.R., 2nd., White, E.A., and Staehelin, L.A. (2004). Electron tomographic analysis of somatic cell plate formation in meristematic cells of Arabidopsis preserved by high- pressure freezing. Plant Cell 16, 836–856. Shope, J.C., DeWald, D.B., and Mott, K.A. (2003). Changes in surface area of intact guard cells are correlated with membrane internalization. Plant Physiol. 133, 1314–1321. Staehelin, L.A., and Hepler, P.K. (1996). Cytokinesis in higher plants. Cell 84, 821–824. Stenmark, H., Parton, R.G., Steele-Mortimer, O., Lutcke, A., Gruen- berg, J., and Zerial, M. (1994). Inhibition of rab5 GTPase activity stimulates membrane fusion in endocytosis. EMBO J. 13, 1287– 1296. Strasburger, E. (1875). Zellbilding und Zelltheilung (Jena: Herman Dabis). Tse, Y.C., Mo, B., Hillmer, S., Zhao, M., Lo, S.W., Robinson, D.G., and Jiang, L. (2004). Identification of multivesicular bodies as prevacuo- lar compartments in Nicotiana tabacum BY-2 cells. Plant Cell 16, 672–693. Ueda, T., Yamaguchi, M., Uchimiya, H., and Nakano, A. (2001). Ara6, a plant-unique novel type Rab GTPase, functions in the endocytic pathway of Arabidopsis thaliana. EMBO J. 20, 4730–4741. Ueda, T., Uemura, T., Sato, M.H., and Nakano, A. (2004). Functional differentiation of endosomes in Arabidopsis cells. Plant J. 40, 783– 789.