The Plant Journal (2002) 29(5), 661±678 Redistribution of membrane proteins between the Golgi apparatus and endoplasmic reticulum in plants is reversible and not dependent on cytoskeletal networks Claude M. Saint-Jore1, Janet Evins1, Henri Batoko2, Federica Brandizzi1, Ian Moore2 and Chris Hawes1,* 1 School of Biological and Molecular Sciences, Oxford Brookes University, Headington, Oxford OX3 0BP, UK 2 Department Plant Sciences, Oxford University, South Parks Road, Oxford OX1 3RB, UK. Received 21 August 2001; revised 5 November 2001; accepted 6 December 2001. *For correspondence (e-mail chawes@brookes.ac.uk). Summary We have fused the signal anchor sequences of a rat sialyl transferase and a human galactosyl transferase along with the Arabidopsis homologue of the yeast HDEL receptor (AtERD2) to the jelly®sh green ¯uorescent protein (GFP) and transiently expressed the chimeric genes in tobacco leaves. All constructs targeted the Golgi apparatus and co-expression with DsRed fusions along with immunolabelling of stably transformed BY2 cells indicated that the fusion proteins located all Golgi stacks. Exposure of tissue to brefeldin A (BFA) resulted in the reversible redistribution of ST-GFP into the endoplasmic reticulum. This effect occurred in the presence of a protein synthesis inhibitor and also in the absence of microtubules or actin ®laments. Likewise, reformation of Golgi stacks on removal of BFA was not dependent on either protein synthesis or the cytoskeleton. These data suggest that ER to Golgi transport in the cell types observed does not require cytoskeletal-based mechanochemical motor systems. However, expression of an inhibitory mutant of Arabidopsis Rab 1b (AtRab1b(N121I) signi®cantly slowed down the recovery of Golgi ¯uorescence in BFA treated cells indicating a role for Rab1 in regulating ER to Golgi anterograde transport. Keywords: Golgi apparatus, endoplasmic reticulum, green ¯uorescent protein, cytoskeleton, brefeldin A. Introduction In higher plants, it is generally accepted that, with a few exceptions (Hara-Nishimura et al., 1998), transport of material from the endoplasmic reticulum (ER) along the secretory pathway, to either the cell surface or the vacuolar system, is via the Golgi apparatus (Hawes et al., 1999a). Likewise, it has been suggested that the vectorial transport of material out of the ER is mediated by membranebounded vesicles (Movafeghi et al., 1999; Pimpl et al., 2000) as is the case in yeast and mammalian cells (Klumperman 2000 and references therein). However, both in vivo and in vitro evidence for the existence of such vesicles in plants is scant. With the exception of various unicellular algae, electron microscopy has failed to produce convincing evidence of either exit sites from the ER, i.e. areas from which vesicle bud or for transition vesicles themselves (Hawes et al., 1996). In contrast, there have been various reports of direct tubular connections between the ER and Golgi in cells producing storage ã 2002 Blackwell Science Ltd protein such as those found in the developing cotyledons of legume seeds (Harris and Oparka, 1983; Hawes et al., 1996). The spatial relationship between the organelles of the secretory pathway is maintained by the cytoskeleton. In mammalian cells, both the organisation of, and communication between the endoplasmic reticulum and the Golgi apparatus, are dependent on the microtubule cytoskeleton and its associated proteins such as dynein and kinesin (Allan and Schroer 1999; Harada et al., 1998; Roghi and Allan, 1999). However, Golgi membranes also appear to contain myosins, which may indicate the ability to undertake actin based motility under certain conditions (Allan and Schroer, 1999; Buss et al., 1998). In plants it appears that the actin network is the major cytoskeletal component maintaining the spatial organisation of the Golgi apparatus (Boevink et al., 1998; NebenfuÈhr et al., 1999; SatiatJeunemaitre et al., 1996), and regulating the organisation 661 662 Claude M. Saint-Jore et al. ã Blackwell Science Ltd, The Plant Journal, (2002), 29, 661±678 Redistribution of membrane proteins and movement of the ER network (Boevink et al., 1998; Liebe and Quader, 1994; Quader, 1990). Application of ¯uorescent protein technology has started to revolutionise plant cell biology and the study of the secretory pathway is no exception. It is now possible to observe in vivo the dynamic events of secretion and endomembrane organisation (Boevink et al., 1996, 1998, 1999; NebenfuÈhr et al., 1999). Delivering GFP to the Golgi apparatus by the addition of targeting regions of Golgi transferases (Boevink et al., 1998; Essl et al., 1999) or complete enzymes (NebenfuÈhr et al., 1999) has demonstrated that in suspension culture cells, leaf epidermal cells, hypocotyls, and roots (Saint-Jore, Moore and Hawes unpublished) individual Golgi stacks to be highly dynamic, moving not only in the general cytoplasmic stream but in more de®ned patterns at the cortex of cells. With a construct comprising the Arabidopsis HDEL receptor homologue (Bar-Peled et al., 1995) and GFP (AtERD2GFP) we have also demonstrated that both the ER and Golgi can be targeted in leaf epidermal cells of Nicotiana clevelandii (Boevink et al., 1998). The surprising result was that Golgi bodies were apparently closely associated with, and tracked over the surface of the polygonal network of cortical ER tubules. This ER network overlaid the cortical actin cytoskeleton, which when disrupted resulted in cessation of Golgi movement without disrupting the ER geometry. Use of the fungal macrocyclic lactone brefeldin A (BFA) induced retrograde transport of these constructs into the ER of tobacco leaf cells, a phenomenon which was reversible on removal of the drug (Boevink et al., 1998). Although it appears that Golgi movement in plants is mediated by the actin cytoskeleton, the role of the cytoskeleton in the vectorial transport of material between putative exit sites on the ER and the cis-Golgi is unknown, as is the mode of retrograde retrieval of Golgi proteins to the ER. By utilising the BFA effect combined with transient expression (Batoko et al., 2000) of Golgi targeted GFP constructs or stably transformed BY-2 cells, cytoskeletal inhibitors and immunocytochemistry, we show here that BFA mediated retrograde transport to the ER and ante- 663 rograde recovery of Golgi membrane from the ER is insensitive to the action of cytoskeletal inhibitors. However, expression of a dominant-inhibitory mutant of the Arabidopsis Rab GTPase AtRab1b (Batoko et al., 2000) had no effect on BFA induced retrograde transport of Golgi targeted GFP but inhibited recovery of Golgi on BFA release. These observations pose important questions on the nature of the physical relationship between the two organelles. Results GFP constructs target the Golgi apparatus in tobacco leaves and BY-2 cells In tobacco leaves, GFP when fused to the signal anchor sequences of a rat sialyl transferase (ST-GFP) and transiently expressed using Agrobacterium in®ltration targeted a system of highly motile ¯uorescent structures (Figure 1a). A human galactosyl transferase signal anchor sequence construct (GT-GFP, Zaal et al., 1999) and the Arabidopsis ERD2-GFP also targeted motile ¯uorescent structures (Figure 1b,c), but also often showed ¯uorescence of the ER. It was assumed that the ¯uorescent bodies were individual Golgi stacks as described previously by Boevink et al. (1998). In transformed BY2 cells STGFP again labelled Golgi-like structures (Figure 1d,e). One of the properties of GFP is that the ¯uorophore retains its capability to ¯uoresce after chemical ®xation. This enabled the staining ST-GFP cells with JIM84, a Golgi marker antibody, revealing that JIM84 epitopes co-localised with the GFP tagged structures (Figure 1f,g,h). To con®rm that our constructs labelled the same population of Golgi we co-expressed, in tobacco leaves, AtERD2-GFP and the red ¯uorescent protein, DsRed, fused to the sialyl transferase signal anchor sequence. Figure 1(i,j,k) shows that in cells expressing low levels of the DsRed construct, the two ¯uorescent proteins co-localise, indicating that it is likely that each construct targets every Golgi body in the cell. Figure 1. Expression of Golgi targetted constructs in tobacco leaves and BY2 cells. (a) ST-GFP locates to Golgi bodies in the cortical cytoplasm of a leaf epidermal cell. Note lack of expression in guard cells (arrows) which are rarely infected by the agrobacteria. Bar = 25 mm. (b) GT-GFP locates to Golgi bodies and also faintly highlights the ER network. Bar = 25 mm. (c) AtERD2-GFP locates to the Golgi and to the ER network. Visualisation of the ER depends on expression levels and often necessitates saturation of the Golgi signal (see also Boevink et al., 1998). Bar = 15 mm. (d,e) Stable expression of ST-GFP in BY2 cells showing distribution of Golgi bodies in the cortical cytoplasm (d) and in the same cell at a different focal plane surrounding the nucleus and in transvacuolar strands (e). Bar = 10 mm (d) and 25 mm (e). (f±h) Colocalisation of ST-GFP with the Golgi marker antibody JIM84 in transformed BY2 cells. ST-GFP distribution in a ®xed cell (f) colocalises (h) with the JIM84 antigen detected with a Texas red conjugated second antibody (g). Bar = 25 mm. (i±k) Co-localisation of AtERD2-GFP (i) with ST-DsRed (j) represented in blue (k) with the Zeiss LSM410 co-localisation software. Bar = 10 mm. (l) Actin labelling with rhodamine phalloidin in BY2 cells expressing ST-GFP. Actin and Golgi are closely associated and Golgi bodies align on actin cables (insert). Bar = 10 mm. (m) Immunolocation of cortical microtubules (red) in BY2 cells expressing ST-GFP. Note, in many cells Golgi ¯uorescence was often located in a different focal plane to the microtubules (data not shown). Bar = 10 mm. ã Blackwell Science Ltd, The Plant Journal, (2002), 29, 661±678 664 Claude M. Saint-Jore et al. ã Blackwell Science Ltd, The Plant Journal, (2002), 29, 661±678 Redistribution of membrane proteins Staining of the actin cytoskeleton in BY2 cells with rhodamine labelled phalloidin revealed a close association between Golgi bodies and actin (Figure 1l), with Golgi bodies aligned along actin cables (Figure 1l insert) as previously reported for tobacco leaves (Boevink et al., 1998). However, there generally appeared to be less coalignment between the cortical microtubules and Golgi bodies (Figure 1m) with the cortical Golgi bodies most often being located in a different confocal (focal) plane to the cortical microtubules (data not shown). 665 cence in the ER indicating successful inhibition of GFP synthesis (Figure 2f). Also, leaves in®ltrated with higher titres of Agrobacterium expressing secGFP showed low level of ER ¯uorescence indicating build up of GFP in the ER lumen prior to secretion. Treatment of these leaves with cycloheximide resulted in gradual loss of ¯uorescence in the ER over 4 h consistent with inhibition of new BFA induces redistribution of Golgi proteins into the ER Treatment of tobacco leaves expressing ST-GFP with 180 mM BFA for 30±60 min, results in cessation of Golgi movement over the ER followed by some clumping of the Golgi and ®nally the disappearance of ¯uorescent Golgi stacks and redistribution of ¯uorescence in the polygonal ER network (Figure 2a). Likewise, a similar treatment of AtERD2-GFP expressing leaves resulted in an increase in ER ¯uorescence and disappearance of Golgi (Figure 2b). This agrees with the data of Boevink et al. (1998) who used a viral expression system to target GFP to the Golgi. BFA is known to inhibit export of some newly synthesised proteins from the ER (Batoko et al., 2000; Boevink et al., 1999). To rule out the possibility that the BFAinduced accumulation of ST-GFP and ERD2-GFP in the ER resulted from the synthesis of new ST-GFP in the ER and loss from the Golgi by secretory activity, we conducted the experiments above in the presence of cycloheximide. To verify that cycloheximide can inhibit the accumulation of new GFP in this system, we investigated the effect of cycloheximide on the accumulation of a secreted GFP, secGFP (Batoko et al., 2000). Leaf cells expressing a secretory form of GFP at levels suf®cient to cause visible accumulation of ¯uorescence in the apoplast (Figure 2d, see also Batoko et al., 2000) were treated with BFA for up to 4 h to block secretion and trap GFP in the ER (Figure 2e). Treatment with 100 mM cycloheximide along with the BFA prevented any signi®cant accumulation of GFP ¯uores- Figure 3. Modi®cation of glycans on an ER-resident protein after BFA treatment. Leaf samples expressing a myc-tagged N-GFP-HDEL were incubated in water or water plus 180 mM BFA for 3 h. Protein extracts were subjected to endo-H digestion prior to SDSA-PAGE. N-GFP-HDEL was detected with a cMyc antibody. BFA treatment confers endoglycosidase H resistance on the glycosylated construct indicating a retrograde transfer of medial to trans-Golgi glycosyl transferases to the ER in response to the drug (a). In contrast, expression of the mutant (N121I) form of At Rab1b did not confer resistance to endoglycosidase H (b). Figure 2. Effect of BFA on Golgi targetted constructs. (a) Treatment of ST-GFP expressing tobacco leaf epidermal cells with 180 mM BFA for 4 h results in complete loss of Golgi ¯uorescence accompanied a dramatic increase in ER ¯uorescence. Note lamellate areas of ER in the cortical network, a common feature of BFA treatment. Bar = 25 mm. (b) BFA at 180 mM for 2.5 h causes the redistribution of Golgi ¯uorescence into the ER in AtERD2-GFP expressing tobacco leaf cells. Bar = 25 mm. (c) Redistribution of ST-GFP into the ER of transformed BY2 cells after treatment at 180 mM for 2.5 h. Bar = 25 mm. (d) Secreted GFP locates to the apoplast in transiently expressing tobacco leaf epidermal cells. Bar = 50 mm. (e) BFA treatment at 360 mM for 3 h results in blockage of secretion and retention of GFP in the ER. Bar = 10 mm. (f) Leaf tissue treated as in e but also in the presence of 100 mm cycloheximide for 3 h. Note there is no new synthesis of GFP in the ER. Bar = 50 mm. (g±l) Time course of BFA (360 mM) treatment on BY2 cells. Initially Golgi show some clumping before ER and nuclear envelope ¯uorescence can be observed. Note that not all Golgi are reabsorbed into the ER. Bar = 50 mm. (m) The cortical network of ER in a ST-GFP transformed BY2 cells after 180 mM BFA for 6 h. Note that the polygonal network of ER tubules are interdispersed with small patches of ER lamellae. Bar = 25 mm. (n) Location of Golgi in ®xed wild type BY2 cells with the monoclonal antibody JIM84. Bar = 25 mm. (o) Immuno-location of the JIM 84 epitope in BY2 cells after 180 mM BFA for 2 h showing labelling of the ER and Golgi clumps. Note that in such preparations the morphology of the ER network is affected by the ®xation protocol. Bar = 10 mm. ã Blackwell Science Ltd, The Plant Journal, (2002), 29, 661±678 666 Claude M. Saint-Jore et al. Figure 4. BFA treatment reveals ER in the mitotic apparatus. (a) Metaphase showing aggregations of ER expressing ST-GFP around the spindle. Bar = 25 mm. (b) Late telophase showing ER accumulating at the phragmoplast and extending between the phragmoplast and the daughter nuclei. Bar = 25 mm. (c) Late telophase as in (b) but after depolymerisation of the actin cytoskeleton with latrunculin B prior to BFA treatment. Bar = 25 mm. (d±f) Location of Golgi (d) and microtubules (e) in a telophase cell showing almost complete exclusion of Golgi from the phragmoplast area (f). Bar = 10 mm. (g±i) Location of ST-GFP (g) and microtubules (h) in a BFA treated cell showing ER interdispersed with the phragmoplast microtubules array (i). Bar = 10 mm. GFP synthesis and secretion of the residual ER located secGFP (data not shown). A similar redistribution of ¯uorescence was obtained on treating ST-GFP BY2 cells with BFA (Figure 2c) where a time series of micrographs revealed an initial clumping of Golgi, followed by redistribution into the ER over 25 min (Figure 2g-l). A full movie sequence of this event can be viewed at http://www.brookes.ac.uk/schools/bms/research/ molcell/hawes/gfpmoviepage.html. Such redistribution into the ER revealed the classical cortical network arrangeã Blackwell Science Ltd, The Plant Journal, (2002), 29, 661±678 Redistribution of membrane proteins 667 Figure 5. BFA induces retrograde transport of Golgi targetted ST-GFP in the presence of cycloheximide and cytoskeletal inhibitors. (a±d) BY2 cells expressing ST-GFP showing BFA induced redistribution of ¯uorescence into the ER in the presence of 100 mM cycloheximide for 2 h (a) 28.9 mM oryzalin for 1.5 h (b), 39.4 mM cytochalasin D for 1.5 h (c) and 265 mM latrunculin B for 1 h (d). Note the almost complete conversion of the polygonal cortical ER network to a large fenestrated sheet in one cells (c). Bars = 25 mm. (e±f) BFA induced retrograde transport of ST-GFP in tobacco leaf epidermal cells in the presence of 39.4 mM cytochalasin D for 2 h (e) and 28.9 mM oryzalin for 1.5 h (f). Bars = 25 mm. ment of ER tubules (Figure 2m). BFA treatment also resulted in distinct morphological changes to the cortical ER network in that with increasing time in the drug the network formed small lamellae at the vertices of the ã Blackwell Science Ltd, The Plant Journal, (2002), 29, 661±678 network tubules (Figure 2m) and in many cells the network converted into large fenestrated sheets of ER membrane (Figure 5c) as has been previously reported for a soluble ER targeted GFP (Boevink et al., 1999). 668 Claude M. Saint-Jore et al. ã Blackwell Science Ltd, The Plant Journal, (2002), 29, 661±678 Redistribution of membrane proteins Immuno¯uorescence of BFA treated cells also revealed a redistribution of the JIM84 epitope into the ER of wild type BY2 cells, although some clumps of Golgi material also remained in the cytoplasm (Figure 2n,o). As this antibody recognises a glycan antigen resident on many Golgi proteins (Fitchette et al., 1999; Horsley et al., 1993) the effect of BFA is not restricted to the redistribution of our chimeric GFP constructs. To con®rm the above data indicating that the BFA effect was a true retrograde redistribution of Golgi membrane proteins into the ER we biochemically tested the endoglycosidase H (endo H) sensitivity of a glycosylated ER targeted GFP construct (N-GFP-HDEL) after BFA treatment. Treatment of protein extracts from tobacco leaves transiently expressing N-GFP-HDEL revealed that the construct in the ER was glycosylated and sensitive to the enzyme. However, after treatment with 180 mM BFA for 3 h all the GFP had acquired endo H resistance indicating the BFA induced presence of mid to late Golgi glycosyl transferases in the ER (Figure 3a). We carried out a similar experiment on leaves transiently expressing N-GFP-HDEL along with a mutant form of the Arabidopsis small GTPase Rab1b (AtRab1b(N121I; Batoko et al., 2000) which has been shown to inhibit ER to Golgi transport and to cause increased accumulation of ST-GFP in the ER. As can be seen in Figure 3b N-GFP-HDEL in leaves expressing AtRab1b(N121I) remained Endo H sensitive indicating that inhibition of forward membrane traf®c by At-Rab1b(N121I) was not suf®cient to cause accumulation of detectable Golgi transferase activity in the ER. These biochemical data are consistent with the complete loss of morphologically distinct Golgi structures in BFA-treated cells in contrast to the persistence of GFP-labelled Golgi stacks in the presence of At-Rab1b(N121I) (Batoko et al., 2000). One unexpected feature of BFA treatment on BY2 cells expressing ST-GFP cells was the highlighting of mitotic pro®les by a characteristic distribution of ER around the spindle apparatus and phragmoplast. Thus, highly characteristic patterns of ER surrounding and running through metaphase spindles were revealed (Figure 4a) and in telophase, the new nuclear envelopes, ER compacted around the phragmoplast and membrane spanning the 669 nuclear envelopes and phragmoplast were revealed (Figure 4b). These pro®les were also observed after BFA treatment following actin depolymerisation in the presence of latrunculin B (Figure 4c). Immunostaining of telophase cells with anti-tubulin revealed a classic phragmoplast array of microtubules, which appeared to exclude the majority of Golgi bodies (Figure 4d,e,f). However, BFA treatment appeared not to interfere with the cytoskeleton but revealed ER tubules interdispersed with the phragmoplast microtubules (Figure 4g,h,i). Brefeldin A induces redistribution of Golgi proteins in the presence of cytoskeletal inhibitors In plant cells the movement, and to a certain extent the distribution, of ER and Golgi appears to be dependent on the actin cytoskeleton (Boevink et al., 1998; Quader, 1990). To test whether cytoskeletal elements are necessary for the redistribution of Golgi membrane proteins into the ER we carried out a series of experiments in the presence of various inhibitors. In BY2 cells the redistribution of ST-GFP with BFA was not affected by pre-treatment with the protein synthesis inhibitor cycloheximide (Figure 5a), by the microtubule inhibitor oryzalin (Figure 5b) or by the actin inhibitors cytochalasin D and latrunculin B (Figure 5c,d). Similar results were obtained for tobacco leaf cells expressing STGFP (Figure 5e,f). The effects of the cytoskeletal inhibitors were also assessed by the use of af®nity probes and immuno¯uorescence. In BY2 cells expressing ST-GFP the cortical actin cytoskeleton was unaffected by BFA treatment whilst the ER appeared in the form of cortical sheets and tubules (compare Figure 6a with Figure 6b). Treatment with cytochalasin D resulted in the disappearance of major actin cables leaving short stubby actin ®laments in the cytoplasm (Figure 6c) whereas latrunculin B resulted in the depolymerisation of the majority of the actin cytoskeleton (Figure 6d). In both cases, depolymerisation of the actin cytoskeleton induced cessation of long-range Golgi movement but BFA treatment still resulted in disappearance of the ¯uorescent Golgi bodies and ¯uorescence of the ER Figure 6. Merged images showing af®nity labelling of actin and microtubules (red channel) in BY2 cells expressing ST-GFP (green channel) before and after treatment with cytoskeletal disrupting agents and BFA. (a,b) The cortical actin cytoskeleton (a) is not disrupted by 180 mM BFA treatment for 5 h and large sheets of green ER can be seen associated with the cortical actin (b). Bars = 10 mm. (c,d) Disruption of the actin cytoskeleton with 39.4 mM cytochalasin D for 1.3 h, which leaves some short stubs of ®laments (c), and 10 mm latrunculin B for 1 h which completely disperses the actin cytoskeleton (d), results in some clumping of Golgi stacks. Bars = 10 mm (c), 25 mm (d). (e,f) Disruption of the actin cytoskeleton with 39.4 mM cytochalasin D for 1.5 h (e) and 25 mm latrunculin B for 1 h (f) does not disturb the effect of BFA which results in the formation of large cortical sheets of ER. Bars = 25 mm. (g,h) The cortical microtubule cytoskeleton (g) is not disrupted by 180 mM BFA treatment for 3 h (h). Bars = 10 mm (g), 25 mm (h). (i,j) Depolymerisation of the microtubule cytoskeleton with 28.9 mM oryzalin for 1 h (i) does not affect Golgi distribution (or movement) and does not disturb the BFA affect (j). Bars = 10 mm. ã Blackwell Science Ltd, The Plant Journal, (2002), 29, 661±678 670 Claude M. Saint-Jore et al. Figure 7. Recovery of Golgi ¯uorescence after BFA removal does not require the cytoskeleton and can be independent of protein synthesis. (a) Reformation of ¯uorescent Golgi in BY2 cells expressing ST-GFP after removal into fresh medium minus BFA for 3.5 h. Note some residual ER ¯uorescence remains. Bar = 25 mm. (b) Reformation of Golgi in the presence of 28.9 mM oryzalin for 7 h. Bar = 10 mm. (c,d) Reformation of Golgi in the presence of 39.4 mM cytochalsin D for 7 h (c) and 25 mM latrunculin B for 15 h (d). Bars = 10 mm (c), 25 mm (d). (e) Reformation of Golgi in the presence of 100 mm cycloheximide for 15 h. Bar = 10 mm. (f,g) ST-GFP in ®xed cells after 7 h recovery from BFA (f) colocates with the JIM84 epitope (g) con®rming anterograde transport of the construct into newly formed Golgi bodies. Bar = 10 mm. (h±k) Reformation of Golgi in tobacco leaf epidermal cells after incubation of BFA treated leaves for 15 h in water (h), cycloheximide (i), cytochalasin D (j) and 9 h in oryzalin (k). Note, some residual ¯uorescence in the ER can be seen in the cells. Bars = 25 mm. ã Blackwell Science Ltd, The Plant Journal, (2002), 29, 661±678 Redistribution of membrane proteins 671 Figure 8. Co-expression of AtRab1b(N121I) with ST-GFP and N-secYFP in tobacco leaf epidermal cells slows the recovery of Golgi ¯uorescence after BFA treatment. Treatment of ST-GFP and N-secYFP expressing tobacco leaf epidermal cells (a) with 180 mM BFA for 3 h results in ¯uorescence of the ER (b) and Golgi ¯uorescence recovers after 7 h incubation in water (c).Co-expression of ST-GFP/N-secYFP with AtRab1b(N121I) results in ¯uorescence of ER and Golgi (d, see also Batoko et al., 2000) and treatment with 180 mM BFA for 4 h results in disappearance of Golgi ¯uorescence along with an increase in ER ¯uorescence (e). After 8 h incubation in water, recovery from the drug can be seen to be impeded by co-expression of AtRab1b(N121I) while the ER remains brightly ¯uorescent (f). A comparison of the two BFA recovery experiments at higher magni®cation: (g) recovery in absence of Rab mutant, (h) recovery in presence of Rab mutant. Bars = 25 mm. often in the form of large cortical sheets (Figure 6e,f). Similarly, BFA had no effect on the arrangement of the cortical microtubule cytoskeleton (Figure 6g,h), whilst depolymerisation of the microtubule cytoskeleton with oryzalin had no effect on the distribution of the Golgi apparatus (Figure 6i) nor on BFA induced redistribution of the Golgi targeted GFP constructs (Figure 6j and compare with Figure 2o). ã Blackwell Science Ltd, The Plant Journal, (2002), 29, 661±678 The effect of cytochalasin D on ST-GFP labelled Golgi has previously been reported in tobacco leaves (Boevink et al., 1998). Using the freeze-shatter technique to immunostain tobacco leaf epidermal cell microtubules we con®rmed that oryzalin induced depolymerisation of leaf microtubules. In the presence of either drug, BFA-induced redistribution of ST-GFP into the ER was observed as in BY2 cells (data not shown). 672 Claude M. Saint-Jore et al. Reformation of Golgi stacks can take place in the absence of the cytoskeleton On removal of BFA from treated BY2 cells, new Golgi bodies are formed over a period of about 5 h along with a loss of ¯uorescence intensity in the ER (Figure 7a) as has previously been demonstrated in tobacco leaves using a viral expression system (Boevink et al., 1998). This recovery can also occur in the presence of the three cytoskeletal inhibitors (Figure 7b,c,d) and remarkably in the presence of cycloheximide (Figure 7e). This reformation of the Golgi can also be monitored not only by the presence of the GFP constructs in the ¯uorescent Golgi bodies but also by the presence of the JIM84 epitope after immuno¯uorescence staining (Figure 7f,g). A similar pattern of recovery after treatment with cycloheximide, cytochalasin D and oryzalin, was also observed in BFA treated leaf segments (Figure 7h-k). At-Rab1b(N121I) inhibits recovery of Golgi stacks on withdrawal of BFA To investigate whether BFA-induced alterations in ER and Golgi organisation are dependent on normal vesicle transport, we investigated the effect of a mutant form of the regulatory GTPase, At-Rab1 which has been shown to inhibit transport of a secretory GFP marker between the ER and Golgi (Batoko et al., 2000). In order to con®rm that mutant protein was expressed in the cells imaged, plants were co-in®ltrated with ST-GFP and a secretory version of YFP (N-secYFP) which, like secGFP (Batoko et al., 2000), accumulates in the ER upon any blockage of ER to Golgi transport (¯uorescence data from GFP and the accumulated N-secYFP were collected from the same confocal detector). Using transient expression in tobacco leaves, BFA induced accumulation and/or redistribution of ¯uorescence in the ER, with recovery of Golgi bodies after removal of the drug (Figure 8a-c,g). Co-expression of the AtRab1b(N121I) mutant protein with ST-GFP and N-secYFP in tobacco leaves resulted in increased accumulation of ¯uorescence in the ER compared with control leaves as reported by Batoko et al. (2000), without signi®cant loss of Golgi ¯uorescence (compare Figure 8a and Figure 8d). This con®rmed some inhibition of ER to Golgi transport by the dominant-inhibitory Rab1 mutant. Treatment of such leaf cells with BFA for 4 h resulted in total loss of Golgi ¯uorescence along with an increase in intensity of ER ¯uorescence (Figure 8e), due to ST-GFP redistribution alongside accumulation of N-secYFP, indicating that redistribution of ST-GFP is not affected by Rab1 inhibition. In control leaves incubated in water for an equivalent period there was no disruption of Golgi. However, in the leaves expressing the At-Rab1b(N121I) mutant there was a greatly reduced recovery of Golgi ¯uorescence 7 h after removal of BFA while the ER remaining brightly ¯uorescent as expected in the presence of N-secYFP and the AtRab1b(N121I) mutant (Figure 8f, and compare Figure 8g with Figure 8h). This retention of ER ¯uorescence was apparent up to 20 h after removal of the BFA and was not observed in leaves recovering from BFA treatment without AtRab1b(N121I) expression where Golgi reformed. The BFA recovery phenotype was also rescued by the coexpression of wild type AtRab1b along with the mutant (data not shown) indicating that the inhibition of recovery from the BFA phenotype was indeed due to loss of AtRab1b function (see Batoko et al., 2000 for discussion). Discussion Targeting GFP to the plant Golgi apparatus It has been shown that the signal anchor sequences of both mammalian and plant glycosyl transferases are suf®cient to target GFP to the Golgi apparatus in tobacco (Boevink et al., 1998; Essl et al., 1999). Here we also show that the C-terminal 60 amino acids of a human b-1,4galactosyl transferase composing the signal anchor sequence is also suf®cient to target GFP to the plant Golgi. Targeting of GFP constructs to the plant Golgi has now been reported using various expression systems: viral (Boevink et al., 1998; Essl et al., 1999); biolistics (Baldwin et al., 2001; Takeuchi et al., 2000); Agrobacterium-mediated (Batoko et al., 2000 and this report); and stable in suspension culture cells (this report). Similar ¯uorescent patterns also result from the expression of complete Golgi proteinGFP constructs in leaves and suspension cultures (Baldwin et al., 2001; NebenfuÈhr et al., 1999; this report). We have also successfully expressed the AtERD2 and ST constructs reported here in transformed Arabidopsis plants (SaintJore, Moore and Hawes unpublished). Co-expression of the AtERD2-GFP with ST-DsRed (Matz et al., 1999) in leaves and immunolabelling of ST-GFP cells with JIM 84 in BY2 cells con®rmed that in all cases, all the Golgi stacks in cells were targeted with the ¯uorescent proteins and we have no evidence of differential targeting to subsets of Golgi. Expression of AtERD2-GFP and GT-GFP also con®rmed the previous report (Boevink et al., 1998) that leaf Golgi are closely associated with the cortical endoplasmic reticulum and track over the polygonal network of ER tubules. Of the three constructs, as ST-GFP showed the lowest level of ER labelling, it was decided to use this as the marker of choice for further experiments. BFA induces retrograde redistribution of GFP constructs to the ER BFA has long been used as a drug by which the mechanisms of protein transport within the secretory system can ã Blackwell Science Ltd, The Plant Journal, (2002), 29, 661±678 Redistribution of membrane proteins be investigated (Klausner et al., 1992; Satiat-Jeunemaitre et al., 1996). The most apparent effect in mammalian cells is the induction of a rapid recycling of Golgi enzymes and membrane into the ER via a tubule mediated fusion process driven by microtubules (Lippincott-Schwartz et al., 1990). At the molecular level BFA prevents the budding of COPI vesicles on membranes by inhibiting the recruitment of the GTP form of ADP ribosylation factor (ARF) onto cisternae, thus preventing the construction of the vesicle coat from the coatomer subunits (Donaldson et al., 1992; Helms and Rothman, 1992). One molecular target of the drug is a 200 amino acid domain identi®ed from the sec7 gene product found in various guanine nucleotide exchange factors (GEF, Mansour et al., 1999; Robineau et al., 2000). Speci®c residues in Sec7p appear to be critical for BFA action (Peyroche et al., 1999; Sata et al., 1998) preventing the GDP/GTP exchange on ARF (Donaldson et al., 1992). Before the advent of GFP technology the structural effects of BFA on plant cells were described, mainly form ultrastructural observations coupled with a few immunocytochemical reports (Satiat-Jeunemaitre and Hawes, 1992; Satiat-Jeunemaitre et al., 1996; Staehelin and Driouich, 1997). In roots it had been shown that in some species the drug induces a trans-face vesiculation of the Golgi and the formation of so called `BFA compartments' (Satiat-Jeunemaitre and Hawes, 1992; Wee et al., 1998). In other tissues authors reported on the redistribution of Golgi membranes to the ER, but with no marker proteins to visualise the Golgi, such evidence was weak (Rutten and Knuiman, 1993; Yasuhara et al., 1995). With the publication of the ®rst Golgi targeted GFP constructs in planta, it was shown that in tobacco leaves BFA appeared to induce a retrograde delivery of Golgi targeted ST signal anchor sequence and AtERD2-GFP chimeras back to the ER (Boevink et al., 1998). The wild type, slow folding GFP, was utilised in this latter study, suggesting that the BFA induced ER ¯uorescence was due to retrograde transport and not the translation and folding of new GFP within the ER. However, no evidence was available to prove that the ER resident GFP originated from the Golgi. In this report we have shown the protein synthesis inhibitor cycloheximide can inhibit the production of a secretory form of GFP, as the BFA induced ER retention of secGFP, previously reported by Batoko et al. (2000), was inhibited by the drug. We therefore carried out BFA experiments in the presence of the inhibitor demonstrating that ER ¯uorescence is indeed a result of redistribution of pre-existing GFP chimeric proteins and not a result of blockage of GFP export from the ER followed by accumulation of newly folded GFP in the ER. This result was also con®rmed by the presence of an endo-H resistant form of a glycosylated GFP construct in BFA treated leaves suggesting the relocation of some Golgi transferase ã Blackwell Science Ltd, The Plant Journal, (2002), 29, 661±678 673 activity to the ER. In contrast, inhibition of anterograde transport by the expression of a dominant inhibitory mutant of At-Rab1 did not cause accumulation of an endo-H resistant form of ER resident GFP. Similar GFP redistribution was also observed in both tobacco leaves and BY2 cells, and immunocytochemical staining with the Golgi marker antibody JIM 84 showed redistribution of the JIM84 Lewis A type glycan epitope into the ER of BY2 cells. Previously, such immunocytochemistry on BFA treated maize roots showed JIM84 staining to be located in the Golgi derived `BFA compartments' and not in the ER (Satiat-Jeunemaitre and Hawes, 1992; Wee et al., 1998). Recently BFA has been shown to induce clumping of Golgi in onion epidermal cells expressing a GFP tagged Arabidopsis GDP-mannose transporter with no apparent redistribution of the protein into the ER (Baldwin et al., 2001), indicating that there can indeed be a variety of responses of plant tissues to BFA. The molecular target and major site of action of BFA in plant cells has yet to be con®rmed. However, at the structural level it has been shown, using both virus and Agrobacterium-based expression systems in tobacco leaves, that BFA treatment results in the inhibition of the export of a secretory form of GFP and its accumulation in the fused ER-Golgi compartment that arises from BFA treatment (Batoko et al., 2000; Boevink et al., 1999). If BFA initially causes inhibition of export from ER prior to its fusion with the Golgi, and, if a continual recycling of Golgi material to and from the ER occurs, inhibition of the anterograde pathway by BFA could result in an imbalance in the usual Golgi membrane cycling, resulting in reabsorption into the ER. However, this data does not preclude a lesser effect of BFA on the trans-Golgi as previously reported in roots. Interestingly, reports in the mammalian literature have suggested that remnants of Golgi can remain after BFA treatments which are free of glycosylation enzymes but contain matrix proteins such as Golgin (Seeman et al., 2000). The other, and more generally accepted possibility is that the principal site of BFA action would be, as suggested for mammals and yeast, in the inhibition of COPI coats on retrograde Golgi derived transport vesicles (Donaldson et al., 1992; Helms and Rothman, 1992). This would result in the breakdown of the physical discontinuity between Golgi and ER and the formation of tubular connections between the organelles resulting in an effective merging of the two structures (Lippincott-Schwartz et al., 1989, 1990). The intracellular accumulation of secreted markers such as secGFP may then result secondarily from the inability of this fused organelle to sustain anterograde transport to post-Golgi compartments. However, tubulation of the GFP expressing Golgi at the confocal level was not seen in any of our experiments, though considering the close physical proximity of the two organelles and the possibility that sites of 674 Claude M. Saint-Jore et al. membrane exchange are localised and transient, any such tubules may well be missed. As previously reported, the BFA phenotype can easily be reversed on removal of the drug (Boevink et al., 1998; Satiat-Jeunemaitre and Hawes, 1992) with the Golgi reforming, over a number of hours, apparently from the ER network. Surprisingly there was also some reformation of Golgi-like structures in the presence of cycloheximide, which suggests that the proteins required for Golgi reformation, including the many that are required to sustain ER to Golgi transport (Andreeva et al., 2000), can retain activity over the time course of such an experiment. A feature of BFA treated cells was the appearance of mitotic pro®les due to extensive ER ¯uorescence, con®rming the extensive accumulation of ER at the mitotic spindle poles, around the spindle and at the phragmoplasts, as previously reported by electron microscopy (Hawes et al., 1981) and immunolabelling (Gunning and Steer, 1996). Such a result indicates that redistribution of membrane from the Golgi to the ER can take place during the division process and that the membrane protein can disperse around the ER into areas that were previously devoid of Golgi (compare Figure 3f and Figure 3i). Alternatively, and in our opinion a more unlikely scenario, is that BFA treated cells could continue the mitotic process in the absence of functional Golgi stacks. It is accepted that in mammalian cells during mitosis, anterograde export from the ER is inhibited and the Golgi is dispersed (Lippincott-Schwartz and Zaal 2000), but experiments with GFP tagged galactosyl transferase have shown continued retrograde transport of the construct into the ER during early stages of mitosis resulting in a complete reabsorption of Golgi enzymes into the ER (Zaal et al., 1999). In plant cells Golgi bodies remain intact during mitosis and cytokinesis and there is no information as to whether the secretory pathway is inactivated at any stage during the division process. Interestingly NebenfuÈhr et al. (2000) reported faint ER ¯uorescence during anaphase in a BY2 line expressing a soybean a-1,2-mannosidase-GFP chimera. One explanation was that this represented an up-regulation of the retrograde pathway during mitosis. However, in our STGFP cells we saw no evidence of ER ¯uorescence during mitosis and cytokinesis (Figure 3d). The distribution of ER at the spindle poles, around the spindle apparatus and at the phragmoplast revealed by BFA treatment, corresponded exactly to that previously reported in non-drug treated plant cells by electron microscopy (Hawes et al., 1981), by immuno¯uorescence (Gunning and Steer, 1996) and with a GFP-HDEL construct (NebenfuÈhr et al., 2000). The cytoskeletal networks do not mediate the BFA effect Transport from the ER to the Golgi in mammalian cells is mediated by the microtubule cytoskeleton and is sensitive to microtubule inhibitors (Thyberg and Moskalewski, 1999). However, as it has previously been shown that actin is responsible for Golgi positioning and movement in plant cells (Boevink et al., 1998; NebenfuÈhr et al., 1999; Satiat-Jeunemaitre et al., 1996), the role of the cytoskeleton in the BFA effect in leaves and BY2 cells was investigated. Both F-actin inhibiting drugs, cytochalasin D and latrunculin B suppressed Golgi movement as expected but surprisingly did not inhibit the retrograde transport of ST-GFP in the presence of BFA. This indicates that an actin network is not required for the BFA induced redistribution of Golgi to ER. Likewise the microtubule depolymerising drug oryzalin did not inhibit either Golgi movement or the BFA effect. Recovery of Golgi bodies was observed on removal of BFA after depolymerisation of the cytoskeleton indicating that formation of new Golgi membrane from the ER is also independent of the microtubule and actin networks. Therefore, it is tempting to speculate that direct transport from the ER to the Golgi, in the systems investigated here, may be free from cytoskeletal regulation even though their motility is actin dependent. This could well be a consequence of the close proximity of the two organelles to each other, moving over actin tracks, indicating a fundamental difference between plant and mammalian early secretory pathways. In the leaf system it has been shown that the Golgi stacks and ER are very closely aligned (Batoko et al., 2000; Boevink et al., 1998). So much so that only very rarely can Golgi stacks be visualised apart from an ER tubule or lamellar region in the slower streaming cortical cytoplasm. As a result of this observation the `vacuum cleaner' model for Golgi ER exchange has been postulated, whereby the Golgi stacks travel over the ER network picking up products from the ER either through direct connections, tubules or vesicles (Boevink et al., 1998; Hawes et al., 1999a). An alternative `recruitment' model, based on the fact that some Golgi stacks undergo periods of stasis on ER tubules, suggests they rest, in response to a `stop' signal, to pick up vesicles produced at exit sites on the ER (NebenfuÈhr and Staehelin 2001; NebenfuÈhr et al., 1999). This is in effect a close range and transient version of the mammalian model of ER to Golgi traf®c. The results presented here indicate that the short distance travelled by any cargo-carrying vector between ER and Golgi is insuf®cient to warrant a mechanochemical motor system based on the cytoskeleton. Therefore, one has to postulate a short-range regulated transport of ER derived vesicles or some form of direct connection between the two organelles either permanent or transient. As we also show BFA induced redistribution and recovery of the Golgi can take place when Golgi bodies are static, they must halt at putative exit sites, if these events depend on normal retrograde traf®cking pathways. Previously it has been shown that on disruption of the actin cytoskeleton with ã Blackwell Science Ltd, The Plant Journal, (2002), 29, 661±678 Redistribution of membrane proteins cytochalasin D, Golgi clustered on small islands of ER lamellae at the vertices of the tubules of the polygonal cortical network (Boevink et al., 1998). How these structures relate to exit sites has yet to be determined. It is therefore pertinent to investigate the location and role of the gamut of accessory proteins such as Rab1, Sar1 and Sec12, that regulate the formation of putative ER/Golgi transport vesicles in leaves and also in a system that is not so spatially limited. One of the advantages of the transient expression system in leaves described in this and previous papers (Andreeva et al., 2000; Batoko et al., 2000) is that not only can a number of different constructs be tested quickly, but also several genes can be co-expressed at the same time, such as the combination of GFP and DsRed tagged proteins. Also the system enables the testing of the function of regulatory proteins within the secretory pathway. In a recent publication Batoko et al. (2000) presented the ®rst evidence for the membrane traf®cking function of a plant Rab GTPase in plant cells. Expression of AtRab1b(N121I) inhibited transport of a secretory form of GFP and the membrane bound glycosylated form of STGFP (N-ST-GFP) out of the ER. Here we have shown that expression of the same mutant Rab1, which is defective in its ability to be converted to the active GTP-bound form, suppresses the recovery of Golgi after BFA treatment but has no effect on the retrograde transport of ST-GFP back to the ER. This indicates that even if the transport distance between ER and Golgi is minimal, regulatory proteins still exert a level of control over the ER to Golgi transport step. 675 Construction of STtmd-GFP, STtmd-DsRed and GTtmdGFP The trans-membrane domain and short cytoplasmic tail (52 amino acids) of a rat a-2,6-sialyltransferase (gift of S. Munro, Cambridge, UK) was ampli®ed by PCR (ST5¢: 5¢-aaatctagaccatgattcataccaacttgaag; ST3¢: 5¢-ccaaagtcgacatggccactttctcctg) and fused to the 5¢ end of GFP5 in place of ERD2 in p pVKH-ERD2-GFP using SalI and XbaI restriction sites, in pVKHEn6Erd2 plasmid. DsRed (BD Clontech UK, Cat. # 6923±1) was ampli®ed by PCR (RFP5¢- ggcggcgtcgactatgaggtcttccaagaatgttatcaaggagttcatgagg; RFP3¢: 5¢-gcgcggggatccctaaaggaacagatggtggcgt-ccctcgg). GFP5 was then replaced by DsRed (BD Clontech UK, Cat. # 6923±1) using SalI and BamHI restriction sites. The trans-membrane domain and short cytoplasmic tail (60 amino acids) of a human b-1,4-galactosyltransferase (gift of J. Lippincott-Schwartz, Bethesda, USA) was ampli®ed by PCR (GT5¢: 5¢aaaatctagaccatgaggcttcgggagccg; GT3¢: 5¢aaaaagtcgactgcagcggtgtggagactccg) and fused to the 5¢ end of GFP5 at the place of ERD2 using SalI and XbaI restriction sites, in the pVKH-Erd2-GFP plasmid. Agrobacterium-mediated BY2 cells transformation pBINPLUS, which carries a kanamycin resistance marker, was used to select transgenic BY2 cells. ST-GFP was subcloned in pBINPLUS using HindIII and BglII/BamH I restriction sites (Figure 1). The construct derived (pBIN-ST-GFP) was transferred into Agrobacterium (strain GV3101 pMP90, Koncz and Schell, 1986) by electroporation. Transgenic Agrobacterium were selected on YEB medium (per litre: beef extract 5 g, yeast extract 1 g, sucrose 5 g, MgSO4±7H2O 0.5 g) containing kanamycin (100 mgml±1) and gentamycin (10 mgml±1) and were used to transform BY-2 cells, as described in Gomord et al. (1998). Transformed tobacco cells were selected in the presence of kanamycin (100 mgml±1) and cefotaxime (250 mgml±1). After screening by ¯uorescence microscopy and immunodetection, calli expressing ST-GFP were used to initiate suspension cultures of transgenic cells. Experimental procedures Construction of AtERD2-GFP The binary vector pVKH18En6-GUS was generated by inserting the expression cassette of pE6113-GUS (gift of M. Ugaki, Ibaraki, Japan; see Mitsuhara et al., 1996) as a PvuII and HindIII into the SacI and HindIII sites of the polylinker of binary vector pVKH18, a derivative of pVK18 (Moore et al., 1998) where the methotrexate resistance marker of pVK18 had been replaced by an hygromycin selectable marker (Zheng and Moore, unpublished). To create pVKH18En6-ERD2-GFP, the coding sequence of GFP5 (Haseloff et al., 1997) was ampli®ed from pBINmGFP5ER (kindly provided by J. Haseloff, Cambridge, UK) by PCR using primers GFP5¢: 5¢tttaagcttcctgcgtcgactttcagtaaaggagaagaacttttca and GFP3¢: 5¢tttggatccttacaaatcctcctcagagataagtttctgctctttgtatagttcatccatgc (cmyc epitope tag sequence is in italics). The cDNA encoding the Arabidopsis H/KDEL receptor ± AtERD2 ± (provided by N. Raikhel, Michigan, USA) was inserted as a HindIII fragment at the HindII site introduced into the 5¢ end of GFP5. The GFP-fusion was subcloned in pVKH18En6 as a BamHI fragment, replacing the GUS coding region of the E6113 cassette to generate plasmid pVKH-ERD2-GFP. ã Blackwell Science Ltd, The Plant Journal, (2002), 29, 661±678 Agrobacterium-mediated transient expression in Nicotiana tabacum pVHK18En6GFP fusions transformed A. tumefaciens was cultured at 28°C, until the stationary phase ( approximately 24 h), washed and resuspended in in®ltration medium (MES 50 mM pH 5.6, glucose 0.5% (w/v), Na3PO4 2 mM, acetosyringone (Aldrich) 100 mM from 10 mM stock in absolute ethanol). The bacterial suspension was pressure injected into the abaxial epidermis of plant leaves using a 1-ml plastic syringe by pressing the nozzle against the lower leaf epidermis (Hawes et al., 1999b). Plants were incubated for 2±3 days at 20±25°C. GFP ¯uorescence was detected by illuminating leaves with a long wavelength UV lamp. Cycloheximide treatments Tobacco leaves were incubated for 3 days after in®ltration with Agrobacterium transformed with pVKH-secGFP (see Batoko et al., 2000) resulting in the expression of a secreted form of GFP. Leaf segments were then incubated in 100 mm cycloheximide, 180 mm BFA (from 36 mM stock in DMSO, Sigma) or a mixture of BFA and cycloheximide for 3-4 h before confocal analysis. 676 Claude M. Saint-Jore et al. Brefeldin A and cytoskeletal inhibitor treatments All the experiments were carried out using 3- to 4-day-old-cells or 2-3 day post-infection leaves. BFA was kept as a 36-mM stock solution in DMSO, at ±20°C. BY2 cells were resuspended in culture medium with brefeldin A (36 or 180 mM). Transfected leaf segments were incubated in a solution of brefeldin A (180 mM). For time-lapse visualisation of BFA on BY2 cells, cells were lain on a thin layer of 2% agar made with culture medium on a slide, BFA added and the cells viewed immediately by confocal microscopy using the Zeiss LSM410 time lapse software. Leaves transiently expressing a myc-tagged, glycosylated NGFP-HDEL construct were treated with BFA (180 mM for 3 h) and proteins extracted to biochemically test the retrograde transport of Golgi enzymes. Extraction, endoglycosidase H (endo H) treatment and immuno-blot analysis were as previously described in Batoko et al. (2000). Cytochalasin D (Sigma, 1.97 mM stock in DMSO, ±20°C) or latrunculin B (Calbiochem, 1 mM stock in DMSO, ±20°C) was used to depolymerise actin. BY2 cells were resuspended in culture medium containing cytochalasin D (39.4 mM) or latrunculin B (25 mM). Transfected leaf discs were incubated in a solution of cytochalasin D (39.4 mM) or latrunculin B (10±25 mM). To disrupt microtubules, BY2 cells were resuspended in culture medium containing 28.9 mM oryzalin (Dow Elanco) from a stock solution (0.15 M in acetone). Transfected leaf discs were incubated in a solution of oryzalin (28.9 mM). Controls contained an equivalent concentration of DMSO or acetone. Immunolabelling For microtubules, BY2 cells were ®xed for 1 h in 4% (w/v) paraformaldehyde (PFA) in microtubule stabilising buffer (MTSB, 0.1 M PIPES, pH 6.9, EGTA 10 mM, MgSO4 10 mM), were digested for 20 min in 1% (w/v) cellulase (Onozuka R10, Yakult Honusha Co. Ltd, Japan), 0.1% (w/v) pectinase (Sigma), 1% (w/v) BSA in buffer, and were permeabilised with 0.5% (v/v) Triton X100 in MTSB, for 15 min. For immunolocalisation, cells were treated with 1% BSA and 1% ®sh gelatin before incubation with anti a-tubulin (YOL1/34, Serotec, UK) (diluted 1 : 20) for 2 h at room temperature followed by incubation with Texas-Red-conjugated goat antirat antibodies (Molecular Probes) for 1 h at room temperature, washing and mounting in Citi¯uor antifade (Agar Scienti®c, UK). For Golgi labelling, cells were incubated in the presence of JIM84 culture supernatant (Horsley et al., 1993) for 2 h at room temperature followed by Texas-Red-conjugated antirat as above. Actin staining BY2 cells were pre-incubated for 2 min in 100 mM 3-maleidobenzoyl N-hydroxysuccinimide ester (MBS, Sigma) and 0.025% Triton X-100 to stabilise the actin and permeabilise the cells, and stained in a solution of rhodamine-phalloidin (SIGMA, 10±5 M in MTSB) buffer for 4 min. Excess stain was removed by washes in MTSB. Expression of AtRab1b(N121I) Plasmids encoding the Arabidopsis Rab GTPase AtRab1b and the dominant interfering mutant N121I were described in Batoko et al. (2000). The YFP variant of N-secGFP was constructed by amplifying the chitinase signal peptide and synthetic N-glycosylated peptide exactly as described for N-secGFP (Batoko et al., 2000) and inserting it as an XbaI ± XhoI fragment into the XbaI and SalI sites upstream of YFP in the plasmid pVKH-N-ST-YFP to generate pVKH-N-secYFP. pVKH-N-ST-YFP is identical to pVKH-N-ST-GFP (Batoko et al., 2000) except that the GFP coding region was replaced with a YFP coding region using the SalI and BamHI sites. Co-expression of these constructs with the Golgi-targeted ST-GFP into tobacco leaf epidemal cells was achieved by Agrobacteriummediated transient expression. ST-GFP was introduced at OD600 0.1, AtRab1b(N121I) at OD600 0.03 and N-secYFP at OD600 0.01. The infected plant was incubated at 20°C for 3 days and samples from infected areas subjected to BFA treatment as described above. Confocal microscopy All specimens were imaged with a Zeiss LSM-410 or LSM 510 confocal laser scanning microscope (CLSM) with a 488-nm argon ion laser and a 510±525 or 505±530 nm bypass ®lter to exclude chlorophyll auto¯uorescence. Texas-Red and rhodamine were excited with a 543-nm argon ion laser line with a 570-nm bypass ®lter. Note added in proof In a recently published paper immuno¯uorescence and protein blot labelling of BY2 cells expressing a GFP-Golgi marker has demonstrated a relatively rapid loss of g-COP and a slower loss of Arf1 from Golgi in BFA treated cells (Ritzenthaler et al., 2002). Acknowledgements We acknowledge the BBSRC for a grant to CH supporting this work. Jennifer Lippincott-Schwartz (NIH Bethesda) for the b1±4 galactosyl transferase, Jim Haseloff (Cambridge) for pBIN mGFP5ER, Huanquan Zheng (Oxford Brookes) for the sec-GFP construct and Ulla Neumann (Oxford Brookes) for taking images in Figure 1i-k. References Allan, V.J. and Schroer, T.A. (1999) Membrane motors. Curr. Op. Cell Biol. 11, 476±482. Andreeva, A.V., Zheng, H., Saint-Jore, C.M., Kutuzov, M.A., Evans, D.E. and Hawes, C. (2000) Organization of transport from the endoplasmic reticulum to the Golgi in higher plants. Biochem. Soc. Trans. 28, 505±512. Baldwin, T.C., Handford, M.G., Yuseff, M.-I., Orellana, A. and Dupreee, P. (2001) Identi®cation and characterization of GONST1, a Golgi-localised GDP-mannose transporter in Arabidopsis. Plant Cell 13, 2283±2295. Bar-Peled, M., da Silva ConceicË|o, A., Frigerio, L. and Raikhel, N.V. (1995) Expression and regulation of aERD2, a gene encoding the KDEL receptor homolog in plants, and other genes encoding protein involved in ER-Golgi vesicular traf®cking. Plant Cell 7, 667±676. Batoko, H., Zheng, H.-Q., Hawes, C. and Moore, I. (2000) A Rab1 GTPase is required for transport between the endoplasmic reticulum and Golgi apparatus and for normal Golgi movements in plants. Plant Cell 12, 2201±2218. ã Blackwell Science Ltd, The Plant Journal, (2002), 29, 661±678 Redistribution of membrane proteins Boevink, P., Martin, B., Oparka, K., Santa Cruz, S. and Hawes, C. (1999) Transport of virally expressed green ¯uorescent protein through the secretory pathway in tobacco leaves is inhibited by cold shock and brefeldin A. Planta 208, 392±400. Boevink, P., Oparka, K., Santa Cruz, S., Martin, B., Betteridge, A. and Hawes, C. (1998) Stacks on tracks: the plant Golgi apparatus traf®cs on an actin/ER network. Plant J. 15, 441±447. Boevink, P., Santa Cruz, S., Hawes, C., Harris, N. and Oparka, K.J. (1996) Virus-mediated delivery of the green ¯uorescent protein to the endoplasmic reticulum of plant cells. Plant J. 10, 935±941. Buss, F., Kendrick-Jones, J., Lionne, C., Knight, A.E., Cote, G.P. and Luzio, J.P. (1998) The localization of myosin VI at the Golgi complex and leading edge of ®broblasts and its phosphorylation and recruitment into membrane ruf¯es of A431 cells after growth factor stimulation. J. Cell Biol. 143, 1535±1545. Donaldson, J.G., Finazzi, D. and Klausner, R.D. (1992) Brefeldin A inhibits Golgi membrane-catalyzed exchange of guanine nucleotide onto ARF protein. Nature 360, 350±352. Essl, D., Dirnberger. D., Gomord, V., Strasser, R., Faye, L., GloÈssl, J. and Steinkellner, H. (1999) The N-terminal 77 amino acids from tobacco N-acetylglucosoaminyltransferase I are suf®cient to retain a reporter protein in the Golgi apparatus of Nicotiana benthamiana cells. FEBS Lett. 453, 169±173. Fitchette, A.-C., Cabanes-Macheteau, M., Marvin, L., Martin, B., Satiat-Jeunemaitre, B., Gomord, V., Crooks, K., Faye, L. and Hawes, C. (1999) Biosynthesis and immunolocalization of Lewis a-containing N-glycans in the plant cell. Plant Physiol. 121, 333± 343. Gomord, V., Fitchette-LaineÂ, A.-C., Michaud, D. and Faye, L. (1998) Production of foreign proteins in tobacco suspension culture. In: Methods in Biotechnology, Vol. 3 (Cunningham C. and Porter, A.J.R. eds). Humana Press, Towota, NJ. pp. 155± 164. Gunning, B.E.S. and Steer, M.W. (1996) Plant Cell Biology: Structure and Function. Boston: Jones and Bartlett Publishers. Harada, A., Takei, Y., Kanai, Y., Tanaka, Y., Nonaka, S. and Hirokawa, N. (1998) Golgi vesiculation and lysosome dispersion in cells lacking cytoplasmic dynein. J. Cell Biol. 141, 51±59. Hara-Nishimura, I., Schimada, T., Hatano, K., Takeuchi, Y. and Nishimura, M. (1998) Transport of storage proteins to protein storage vacuoles is mediated by large precursor accumulating vesicles. Plant Cell 10, 825±836. Harris, N. and Oparka, K. (1983) Connections between dictyosomes, ER and GERL in cotyledons of mung bean (Vigna radiata L.). Protoplasma 114, 93±102. Haseloff, J., Siemering, K.R., Prasher, D.C. and Hodge, S. (1997) Removal of a cryptic intron and subcellular localization of green ¯uorescent protein are required to mark transgenic Arabidopsis plants brightly. Proc. Natl Acad. Sci. USA 94, 2122±2127. Hawes, C., Boevink, P. and Moore, I. (1999b) Green ¯uorescent protein in plants. In: Protein Localization by Fluorescence Microscopy: a Practical Approach. (Allen, V. ed). Oxford: Oxford University Press, pp. 163±177. Hawes, C.R., Brandizzi, F. and Andreeva, A. (1999a) Endomembranes and vesicle traf®cking. Curr. Op. Plant Biol. 2, 454±461. Hawes, C., Faye, L. and Satiat-Jeunemaitre, B. (1996) The Golgi apparatus and pathways of vesicle traf®cking. In: Membranes: Specialized Functions in Plants (Smallwood, M., Knox J.P. and Bowles, D.J., eds). Oxford: Bios Scienti®c Publishers, pp. 337± 365. Hawes, C.R., Juniper, B.E. and Horne, J.C. (1981) Low and high ã Blackwell Science Ltd, The Plant Journal, (2002), 29, 661±678 677 voltage electron microscopy of mitosis and cytokinesis in maize roots. Planta 152, 397±407. Helms, J.B. and Rothman, J.E. (1992) Inhibition by Brefeldin A of a Golgi membrane that catalyses exchange of guanine nucleotide bound ARF. Nature, 360, 352±354. Horsley, D., Coleman, J., Evans, D., Crooks, K., Peart, J., SatiatJeunemaitre, B. and Hawes, C. (1993) A monocolonal antibody, JIM 84, recognizes the Golgi apparatus and plasma membrane in plant cells. J. Exp. Bot. 44, 223±229. Klausner, R.D., Donaldson, G.J. and Lippincott-Schwartz, J. (1992) Brefeldin A: Insights into the control of membrane traf®c and organelle structure. J. Cell Biol. 116, 1071±1080. Klumperman, J. (2000) Transport between ER and Golgi. Curr. Op. Cell Biol. 12, 445±449. Koncz, C. and Schell, J. (1986) The promoter of T-DNA gene 5 controls the tissue-speci®c expression of chimeric genes carried by a novel type of Agrobacterium binary vector. Mol. General Genet. 204, 383±396. Liebe, S. and Quader, H. (1994) Myosin in onion (Allium cepa) bulb scale epidermal cells: involvement in dynamics of organelles and endoplasmic reticulum. Physiol. Plant. 90, 114±124. Lippincott-Schwartz, J., Donaldson, J.G., Schweizer, A., Berger, E.G., Hauri, H.D., Yuan, L.C. and Klausner, R.D. (1990) Microtubule-dependant retrograde transport of proteins into the ER in the presence of Brefeldin A suggests an ER recycling pathway. Cell 80, 821±836. Lippincott-Schwartz, J., Yuan, L., Bonifacino, J. and Klausner, R. (1989) Rapid redistribution of Golgi proteins into the ER in cells treated with Brefeldin A: evidence for membrane recycling from Golgi to ER. Cell 56, 801±813. Lippincott-Swartz, J. and Zaal, K.J.M. (2000) Cell cycle maintenance and the biogensis of the Glgi complex. Histochem. Cell Biol. 114, 93±103. Mansour, S.J., Skaug, J., Zhao, X.-H., Giordana, J., Scherer, S.W. and MelancËon, P. (1999) p200 ARF-GEP1: a Golgi-localized guanine nucleotide exchange protein whose Sec7 domain is targeted by the drug brefeldin A. Proc. Natl Acad. Sci. USA 96, 7968±7973. Matz, M.V., Fradkov, A.F., Labas, Y.A., Savitsky, A.P., Zaraisky, A.G., Markelov, M.L. and Lukyanov, S.A. (1999) Fluorescent proteins from nonbioluminescent Anthozoa species. Nature Biotech. 17, 969±973. Mitsuhara, I., Ugaki, M., Hirochika, H., Ohshima, M., Murakami, T., Gotoh, Y., Katayose, Y., Nakamura, S., Honkura, R., Nishimiya, S., Ueno, K., Mochizuki, A., Tanimot, H., Tsukawa, H., Otsuki, Y. and Ohashi, Y. (1996) Ef®cient promoter cassettes for enhanced expression of foreign genes in dicotyledonous and monocotyledonous plants. Plant Cell Physiol. 37, 49±59. Moore, I., Galweiler, L., Grosskopf, D., Schell, J. and Palme, K. (1998) A transcription activation system for regulated gene expressioon in transgenic plants. Proc. Natl Acad. Sci. USA 95, 376±381. Movafeghi, A., Happel, N., Pimpl, P., Tai, G.H. and Robinson, D.G. (1999) Arabidopsis Sec21p and Sec23p homologs. Probable coat proteins of plant COP-coated vesicles. Plant Physiol. 119, 1437±1445. NebenfuÈhr, A., Frohlick, J.A. and Staehelin, L.A. (2000) Redistribution of Golgi stacks and other organelles during mitosis and cytokinesis in plant cells. Plant Physiol. 124, 135± 151. NebenfuÈhr, A., Gallagher, L.A., Dunahay, T.G., Frohlick, J.A., Mazurkiewicz, A.M., Meehl, J.B. and Staehelin, L.A. (1999) 678 Claude M. Saint-Jore et al. Stop-and-go movements of plant Golgi stacks are mediated by the acto-myosin system. Plant Physiol. 121, 1127±1141. NebenfuÈhr, A. and Staehelin, L.A. (2001) Mobile factories: Golgi dynamics in plant cells. Trends Plant Sci. 6, 160±167. Peyroche, A., Antonny, B., Robineau, S., Acker, J., Cher®ls, J. and Jackson, C.L. (1999) Brefeldin A acts to stabilize an abrortive ARF-GDP-sec7 domain protein complex: Involvement of speci®c residues of the sec7 domain. Mol Cell. 3, 275±285. Pimpl, P., Movafeghi, A., Coughlan, S., Denecke, J., Hillmer, S. and Robinson, D.G. (2000) In situ localization and in vitro induction of plant COPI-coated vesicles. Plant Cell 12, 2219± 2236. Quader, H. (1990) Formation and disintegration of cisternae of the endoplasmic reticulum visualized in live cells by conventional ¯uorescence and confocal laser scanning microscopy: evidence for the involvement of calcium and the cytoskeleton. Protoplasma 155, 166±175. Ritzenthaler, C., NebenfuÈr, A., Movafeghi, A., Stussi-Garaud, C., Behnia, L., Pimpl, P., Staehelin, L.A. and Robinson, D.G. (2002) Reevaluation of the effects of brefeldin A on plant cells using tobacco bright yellow 2 cells expressing Golgi-targeted green ¯uorescent protein and COPI antisera. Plant Cell 14, 237±261. Robineau, S., Chabre, M. and Antonny, B. (2000) Binding site of brefeldin A at the interface between the small G protein ADPribosylation factor 1 (ARF1) and the nucleotide-exchange factor Sec 7 domain. Proc. Natl Acad. Sci. USA 97, 9913±9918. Roghi, C. and Allan, V.J. (1999) Dynamic association of cytoplasmic dynein heavy chain 1a with the Golgi apparatus and intermediate compartment. J. Cell Sci. 112, 4673±4685. Rutten, T.L.M. and Knuiman, B. (1993) Brefeldin A effects on tobacco pollen tubes. Eur. J. Cell Biol. 61, 247±255. Sata, M., Donaldson, J.G., Moss, J. and Vaughan, M. (1998) Brefeldin A-inhibited guanine nucleotide-exchange activity of Sec7 domain from yeast Sec7 with yeast and mammalian ADP ribosylation factors. Proc. Natl Acad. Sci. USA 95, 4204±4208. Satiat-Jeunemaitre, B., Cole, L., Bourett, T., Howard, R. and Hawes, C. (1996) Brefeldin A effects in plant and fungal cells: Something new about vesicle traf®cking? J. Microsc. 181, 162± 117. Satiat-Jeunemaitre, B. and Hawes, C. (1992) redistribution of a Golgi glycoprotein in plant cells treated with Brefeldin A. J. Cell Sci. 103, 1153±1166. Seeman, J., Jokitalo, E., Pypaert, M. and Warren, G. (2000) Matrix proteins can generate the higher order architecture of the Golgi apparatus. Nature 407, 1022±1026. Staehelin, L.A. and Driouich, A. (1997) Brefeldin A effects in plants. Plant Physiol. 114, 410±403. Takeuchi, M., Ueda, T., Sato, K., Abe, H., Nagata, T. and Nakano, A. (2000) A dominant negative mutant of Sar1 GTPase inhibits protein transport from the endoplasmic reticulum to the Golgi apparatus in tobacco and Arabidopsis cultured cells. Plant J. 23, 517±525. Thyberg, J. and Moskalewski, S. (1999) Role of microtubules in the organisation of the Golgi complex. Exp. Cell Res. 246, 263± 279. Wee, E.-G., Sherrier, D.J., Prime, T. and Dupree, P. (1998) Targeting of active sialyltransferase to the plant Golgi apparatus. Plant Cell 10, 1759±1768. Yasuhara, H., Sonobe, S. and Shibaoka, H. (1995) Effects of Brefeldin-a on the formation of the cell plate in tobacco BY-2 cells. Europ, J. Cell Biol. 66, 274±281. Zaal, K.J.M., Smith, C.L., Polishchuk, R.S., Altan, N., Cole, N.B., Ellenberg, J., Hirschberg, K., Preseley, J.F., Roberts, T.H., Siggia, E., Phair, R.D. and Lippincott-Schwartz, J. (1999) Golgi membranes are absorbed into and reemerge from the ER during mitosis. Cell 99, 599±601. ã Blackwell Science Ltd, The Plant Journal, (2002), 29, 661±678