Pausing of Golgi Bodies on Microtubules Regulates Secretionof Cellulose Synthase Complexes in Arabidopsis W

Elizabeth Faris Crowell,a Volker Bischoff,a Thierry Desprez,a Aurelia Rolland,a York-Dieter Stierhof,b KarinSchumacher,c Martine Gonneau,a Herman Hofte,a and Samantha Vernhettesa,1

a Laboratoire de Biologie Cellulaire, Institut National de la Recherche Agronomique, 78026 Versailles cedex, Franceb Center for Plant Molecular Biology, Microscopy, University of Tubingen, 72074 Tubingen, Germanyc University of Heidelberg, Heidelberg Institute for Plant Sciences, Department VI Developmental Biology, 69120 Heidelberg,

Germany

Plant growth and organ formation depend on the oriented deposition of load-bearing cellulose microfibrils in the cell wall.

Cellulose is synthesized by plasma membrane–bound complexes containing cellulose synthase proteins (CESAs). Here, we

establish a role for the cytoskeleton in intracellular trafficking of cellulose synthase complexes (CSCs) through the in vivo

study of the green fluorescent protein (GFP)-CESA3 fusion protein in Arabidopsis thaliana hypocotyls. GFP-CESA3 localizes

to the plasma membrane, Golgi apparatus, a compartment identified by the VHA-a1 marker, and, surprisingly, a novel

microtubule-associated cellulose synthase compartment (MASC) whose formation and movement depend on the dynamic

cortical microtubule array. Osmotic stress or treatment with the cellulose synthesis inhibitor CGA 325’615 induces

internalization of CSCs in MASCs, mimicking the intracellular distribution of CSCs in nongrowing cells. Our results indicate

that cellulose synthesis is coordinated with growth status and regulated in part through CSC internalization. We find that

CSC insertion in the plasma membrane is regulated by pauses of the Golgi apparatus along cortical microtubules. Our data

support a model in which cortical microtubules not only guide the trajectories of CSCs in the plasma membrane, but also

regulate the insertion and internalization of CSCs, thus allowing dynamic remodeling of CSC secretion during cell expansion

and differentiation.

INTRODUCTION

Anisotropic cell expansion drives plant growth and organ forma-

tion and depends primarily on the properties of the cell wall

(Baskin, 2005). The load-bearing component of the plant cell wall

consists of cellulose microfibrils, which play a key role in

generating differential resistance to turgor pressure to allow

anisotropic expansion (Green and Selker, 1991). Shortly after the

discovery that cellulose microfibrils are deposited in precisely

oriented layers during development (Probine and Preston, 1961,

1962), it was proposed that long cytoplasmic elements sensitive

to colchicine could direct the orientation of cellulose microfibrils

in the wall (Green, 1962). The discovery of cortical microtubules

soon afterward, and the observation of their parallel alignment

with cellulosemicrofibrils, gave rise to themicrotubule-microfibril

alignment hypothesis (Ledbetter and Porter, 1963). Subse-

quently, numerous studies reported correlations in the orienta-

tion of cortical microtubules and cellulose microfibrils (reviewed

in Baskin, 2001). However, the alignment hypothesis became

increasingly controversial as evidence for absence of correlation

was found in certain cell types (Emons, 1982;Himmelspach et al.,

2003; Sugimoto et al., 2003).

In vascular plants, cellulose is synthesized by a plasma mem-

brane–localized cellulose synthase complex (CSC), recognizable

as a hexameric rosette structure with a diameter of 25 to 30 nm

(Mueller and Brown, 1980) and a cytoplasmic component;45 to

50 nm in diameter (Bowling and Brown, 2008). The only known

component of the CSC is the cellulose synthase catalytic subunit

(CESA) (Kimura et al., 1999). Identification of the genes encoding

CESAs in higher plants (reviewed in Somerville, 2006) and

advances in live-cell imaging techniques permitted visualization

of fluorescently tagged CSCs and microtubules simultaneously

in living cells and provided powerful evidence that cortical

microtubules define the trajectories of CSCs (Paredez et al.,

2006). Experimental evidence presented by Paredez et al. lends

strong support to the microtubule-microfibril alignment hypoth-

esis. Nevertheless, the mechanism by which microtubules guide

CSC trajectories remains unknown and represents an important

focus of current research.

The importance of the actin cytoskeleton in cell expansion

in non-tip-growing cells is only beginning to be revealed (Smith

and Oppenheimer, 2005). Wightman and Turner (2008) recently

showed that actin participates in secondary cell wall synthesis by

delivering organelles containing fluorescently labeled CESA to

bands of secondary cell wall thickenings in developing xylem

cells. Thus, the role of actin in cellulose synthesis seems to be

related to the intracellular trafficking of cellulose synthases, but

this needs to be further confirmed.

1 Address correspondence to samantha.vernhettes@versailles.inra.fr.The author responsible for the distribution of materials integral to thefindings presented in this article in accordance with the policy describedin the Instructions for Authors (www.plantcell.org) is: Herman Hofte(herman.hofte@versailles.inra.fr).WOnline version contains Web-only data.www.plantcell.org/cgi/doi/10.1105/tpc.108.065334

The Plant Cell, Vol. 21: 1141–1154, April 2009, www.plantcell.org ã 2009 American Society of Plant Biologists

While it is generally accepted that the microtubule cytoskel-

eton influences patterns of deposition of cellulose, the actin and/

or microtubule cytoskeleton may also influence the rate of cell

wall synthesis by regulating the delivery of cell wall–synthesizing

enzymes to the plasma membrane. The regulation of cellulose

synthesis activity through cytoskeleton-dependent intracellular

trafficking has remained largely unexplored. Here, we studied the

intracellular trafficking of CESA3, an essential enzyme for pri-

mary cell wall cellulose synthesis. We demonstrate that Golgi

bodies carrying CSCs are distributed throughout the cell via the

acto-myosin system and that Golgi bodies pause on cortical

microtubules for the targeted delivery of CSCs to the plasma

membrane. Under conditions of growth cessation or treatment

with a cellulose synthesis inhibitor, CSCs are internalized into

compartments that show movement dependent on cortical

microtubule array dynamics. Together, our data establish novel

roles for the actin and microtubule cytoskeleton in regulating

cellulose synthesis in the primary cell wall.

RESULTS

CSC Density and Trajectories Are

Developmentally Regulated

To examine regulation of the localization of CESA proteins, we

used stably transformed Arabidopsis thaliana lines expressing

green fluorescent protein (GFP)-tagged CESA3 fusion protein

under the control of the CESA3 endogenous promoter. This

construct has been shown to complement the cesa3 mutant

phenotype (Desprez et al., 2007). We studiedGFP-CESA3 on the

outer face of epidermal cells in 3-d-old etiolated hypocotyls, in

which epidermal cells show an acropetal developmental gradient

(Gendreau et al., 1997). Cells near the apical hook elongate

slowly and deposit a thick, external cell wall, cells in themiddle of

the hypocotyl elongate rapidly, and cells at the base are fully

elongated (Refregier et al., 2004). As described previously,

discrete GFP-CESA3 particles were found aligned at the cell

surface (Figure 1A; Desprez et al., 2007). These particles ex-

hibited steady, bidirectional movement in the focal plane of the

plasmamembrane (Figures 1B and 1C; see Supplemental Movie

1 online). Immunogold labeling of cryosections confirmed the

plasma membrane localization of the GFP-CESA3 surface par-

ticles (Figure 1D), leading us to conclude that these particles

represent active CSCs in the plasma membrane. CSC density

was highest in cells in the apical hook and decreased toward the

base of the hypocotyl (Figure 2). The orientation of CSC trajec-

tories and cortical microtubules visualized using a GFP fusion to

Arabidopsis a-tubulin6 (GFP-TUA6) varied from longitudinal to

transverse in the apical hook and hypocotyl top (Figures 2A and

2B), becoming strictly longitudinal at the base of the hypocotyl

(Figures 2C and 2D). Notably, the densities of cortical microtu-

bules and CSCs were also correlated (Figure 2).

GFP-CESA3 Accumulates in Golgi Bodies and in the

VHA-a1 Compartment

GFP-CESA3 labels ring-shaped structureswith bright, peripheral

punctae, previously shown to be Golgi bodies (Figure 1H;

Paredez et al., 2006). Golgi bodies circulate through the cyto-

plasm via the actomyosin system (Sparkes et al., 2008) and are

characterized by rapid, nonlinear movement (Figure 1I; see

Supplemental Movie 1 online). Electron microscopy (EM) images

of immunogold-labeled cryosections of hypocotyl epidermal

cells confirmed that GFP-CESA3 was present at the periphery

of the Golgi stack, consistent with their ring shape in confocal

images (Figures 1P and 1Q). In addition to Golgi bodies, we

identified a second, relatively rare population (<1 observed per

cell per 10 min imaging period) of smaller, homogeneously

labeled particles with the same rapid, nonlinear movement as

Golgi bodies (Figures 1J and 1K). These particles were invariably

labeled with the early endosome/trans-Golgi network (TGN)

marker VHA-a1-mRFP (Dettmer et al., 2006) in transgenic lines

expressing both markers (41 time series analyzed; Figures 1N

and 1O). VHA-a1 compartments rapidly associated and disso-

ciated with Golgi bodies (Figures 1L to 1O; see Supplemental

Movie 2 online). Most often a single VHA-a1 compartment was

found centered beneath aGolgi stack (Figure 1L). The occasional

presence of GFP-CESA3 in the TGNwas further confirmed in EM

micrographs (Figures 1Q and 1R). Thus, we will refer to these

particles as VHA-a1/GFP-CESA3–containing compartments.

Regulated Distribution of GFP-CESA3 in a Unique

Population of Intracellular Particles

We identified a third population of small, bright particles just

below the plasma membrane, which were distinguished from

CSCs by their fluctuating velocities and intracellular localization

(Figures 1E to 1G; see Supplemental Figures 2A to 2F online).

These particles were likewise distinguished from Golgi bodies

and VHA-a1/GFP-CESA3–containing compartments by their

smaller size and linear trajectories (Figures 1E to 1G). In basal

hypocotyl cells with low CSC density, GFP-CESA3 signal was

preferentially found in these bright intracellular compartments

(Figure 2D). Interestingly, treatment with 500 mM mannitol to

induce osmotic stress or treatment with 50 mM cycloheximide

both caused a rapid decrease inCSCdensity and a simultaneous

increase in the density of the same bright compartments (Figures

3A and 3B). To determine the response kinetics, we examined

the redistribution of GFP-CESA3 under mannitol treatment in the

root tip, a widely used, well-characterized system (Robinson

et al., 2008) where the absence of a thick cuticle allows rapid

penetration of treatment solutions. In the root tip, CSC redistri-

bution was complete within 6 min (five independent experi-

ments), suggesting that these bright intracellular compartments

result from internalization of CSCs as opposed to a block in

secretion (Figures 3C and 3D).

CGA 325’615 (CGA), a potent inhibitor of cellulose synthesis in

cotton fibers (Peng et al., 2001), causes reduced crystalline

cellulose content and isotropic cell expansion inArabidopsis (see

Supplemental Figure 3 online). CGA treatment did not affect

microtubules labeled with either GFP-TUA6 or a GFP fusion to

theArabidopsismicrotubule end binding protein1A (EB1-GFP) or

actin filaments marked by an N-terminal GFP fusion to the

second actin binding domain of Arabidopsis fimbrin 1 (GFP-

FABD2) or alter the morphology of the Golgi apparatus or

organelles observed by EM (see Supplemental Figure 4 online).

1142 The Plant Cell

Figure 1. Four Distinct Localizations of GFP-CESA3 in Epidermal Cells of Etiolated Arabidopsis Hypocotyls.

(A) to (D) Particles corresponding to CSCs in the plasma membrane.

(A) Maximum projected z-stack of first time point in a series showing CSC arranged in linear tracks in the plasma membrane focal plane.

(B) Average projection of the same time series showing organized CSC trajectories.

(C) Kymograph along the trajectory indicated by a line in (B). Time is represented along the vertical axis, and a 5-min scale bar is shown at left (see

Supplemental Figure 1 online for information on interpreting kymographs). CSC movement is steady and bidirectional (see Supplemental Movie

1 online).

(D) Ultrathin cryosection of epidermal cells in the upper hypocotyl of seedlings expressing GFP-CESA3, labeled with anti-GFP primary antibody and

a Nanogold-coupled secondary antibody. Control labeling of wild-type seedlings was negative. Discrete CSCs are labeled in the plasma membrane.

Bars = 500 nm.

(E) to (G) Small, bright, intracellular particles with fluctuating velocity (referred to as MASCs).

(E)Maximum projected z-stack of first time point in a series acquired in a cell near the apical hook; a small, bright, intracellular particle is indicated by an

arrowhead.

(F) Average projection of the time series.

(G) Kymograph along the trajectory indicated in (E). The movement of the small, bright particle is along a linear track, but different from that of the

surrounding CSCs. Initial unsteady translocation is followed by acceleration in the opposite direction.

(H) and (I) GFP-CESA3–labeled Golgi bodies.

(H) GFP-CESA3 is found exclusively at the periphery of Golgi bodies and is characterized by bright punctae (deconvolved image).

(I) Trajectories of four Golgi bodies indicated by colored lines overlaid on an image from the time series. Golgi body movement is nonlinear and

interrupted by pauses.

(J), (K), (N), and (O) VHA-a1/GFP-CESA3–containing compartments.

(J) Image from a time series showing a GFP-CESA3–labeled compartment (arrowhead) just below the plasma membrane.

(K) The white line illustrates the trajectory of the compartment marked in (I). The movement pattern is similar to that of Golgi. The particle associates with

a Golgi body before disappearing into a lower focal plane.

Cellulose Synthase Trafficking 1143

Time-lapse imaging of GFP-CESA3 revealed that CGA treatment

resulted in clearance of CSC from the plasma membrane,

concomitant with an increase in the density of small, bright

intracellular compartments (Figure 3E) identical in appearance,

behavior, and velocity to those observed in untreated basal

hypocotyl cells and those induced by osmotic stress or cyclo-

heximide treatment (cf. Figures 2D, 3A, 3B, and 3E; see Supple-

mental Figure 2G online). Mock treatment with an identical

dilution of DMSO did not observably alter GFP-CESA3 localiza-

tion (see Supplemental Figure 5A online).

Characterization of CGA-Induced Compartments

CGA-induced compartments were observed to move with fluc-

tuating velocities along linear tracks (Figure 3F; see Supplemen-

tal Movie 3 online). These linear tracks could be visualized in

average projections of time series, and their position was ob-

served to change over time (Figure 3G). In image acquisitions of

10 to 15 min, individual compartments could be found that were

stationary, migrating steadily at 10 to 3000 nm/min, or moving as

fast as 8900 nm/min (Figure 3N; see Supplemental Movie 3

online). The linear trajectories of CGA-induced compartments

contrast with the movement of intracellular compartments

known to be regulated by the actomyosin system, such as Golgi

bodies. To test if an intact actin cytoskeleton is required for the

movement of CGA-induced compartments, we treated GFP-

CESA3–expressing seedlings first with CGA and then with the

actin-disrupting agent Cytochalasin D (CytD). CytD causes frag-

mentation of actin microfilaments and inhibits their dynamics

(see Supplemental Figures 5D and 5E online; Foissner and

Wasteneys, 2007). Treatment with CytD resulted in aggregation

of GFP-CESA3–labeled Golgi bodies within 25 min, but no

difference was found in the density or velocity distribution of

CGA-induced compartments compared with CGA treatment

alone (Figures 3H, 3I, and 3N).

To further study the nature of CGA-induced compartment

movement, seedlings were pretreated with the microtubule-

stabilizing drug taxol and then treated with taxol + CGA. The

resulting CGA-induced compartments showed a density similar

to that in CGA treatment alone (Figure 3J); however, the addition

of taxol greatly reduced themotility of the compartments (Figures

3K and 3N). Out of 989 individual particles measured, only 11

were found with velocities equal to or greater than the mean

velocity of 543 nm/min measured in CGA treatment (Figure 3N).

As a third test, seedlings were pretreated with 7 mM oryzalin,

which results in depolymerization of most cortical microtubules

within 4 h (see Supplemental Figures 5B and 5C online). The

subsequent treatment with oryzalin + CGA resulted in reduced

CSC density; however, we never observed characteristic CGA-

induced compartments with normal density near the cell surface

(41 cells from two independent treatments; Figure 3L). In one cell,

we found a small group of aligned particles resembling CGA-

induced compartments, but these particles were stationary

(Figure 3M), suggesting they were associated with fragments

of stable, resistant microtubules.

Based on these results, we conclude that an intact, dynamic

cortical microtubule array is essential for the movement of CGA-

induced compartments, and the actin cytoskeleton does not

appear to have a role in their movement. Henceforth, wewill refer

to CGA-induced compartments as microtubule-associated cel-

lulose synthase compartments (MASCs). To verify that MASCs

observed in untreated cells were also associated with cortical

microtubules, we transiently transformed cotyledons of theGFP-

CESA3–expressing line with the microtubule binding domain

(MBD) of the microtubule-associated protein 4 fused to mono-

meric red fluorescent protein (mRFP). Although plants stably

overexpressing GFP-MBD and cultured on kanamycin were

observed to show growth defects in one study (Granger and

Cyr, 2001), MBD remains a widely used marker for microtubules

in plants and yields signals highly similar to that obtained with

immunolabeling of tubulin (Sainsbury et al., 2008). In the absence

of any drug treatments, we observed MASCs migrating along

cortical microtubules labeled by mRFP-MBD (Figure 3O; see

Supplemental Movie 4 online). Intriguingly, MASC particles

moving along linear tracksmergewith one another, progressively

increasing in intensity and cotranslocating until a split occurs

(Figure 3P; see Supplemental Movie 3 online). This merging

behavior is reminiscent of that observed for the DAM1 ring

protein, known to move on depolymerizing microtubule ends

(Westermann et al., 2006). Indeed, MASC velocities (0 to 8.9 mm/

min) are within the range of velocities measured for depolymer-

ization at leading or lagging microtubule ends (0 to 30 mm/min)

and polymerization at leading ends (0 to 15 mm/min) (Shaw et al.,

2003).

The Golgi Apparatus Pauses on Cortical Microtubules, and

Golgi Bodies and the VHA-a1 Compartment Physically

Interact with MASCs

High-frame-rate high-resolution imaging of GFP-CESA3 in com-

bination with mRFP-MBD or VHA-a1-mRFP markers yielded

several salient observations. GFP-CESA3–labeled Golgi bodies

Figure 1. (continued).

(L) to (O) Confocal images from plants coexpressing GFP-CESA3 (green) and VHA-a1-mRFP (red). Bar = 5 mm.

(L) The VHA-a1 compartment was often found centered on the Golgi stack, without any detectable colocalization of VHA-a1-mRFP and GFP-CESA3.

(M) and (N) VHA-a1 compartments partially dissociated from Golgi show variable levels of colocalization with GFP-CESA3 (see Supplemental Movie 2

online).

(O) Two VHA-a1/GFP-CESA3–containing compartments physically separated from Golgi and labeled by GFP-CESA3. PM, plasma membrane; CP,

cytoplasm.

(P) to (R) Ultrathin cryosections of epidermal cells in the upper hypocotyl of seedlings expressing GFP-CESA3, labeled with anti-GFP primary antibody

and a Nanogold-coupled secondary antibody. Control labeling of wild-type seedlings was negative. G, Golgi stack; T, TGN. Bar = 500 nm.

(P) and (Q) Gold marker can be detected at the rims of many Golgi stacks and, less often, also surrounding the TGN ([Q] and [R]).

Bars = 10 mm except where noted.

1144 The Plant Cell

were observed to pause at discrete sites on microtubules (Fig-

ures 4A to 4C; see Supplemental Movie 4 online). In one repre-

sentative time series, 63 Golgi bodies were tracked and pause

events were quantified for each individual Golgi body. A pause

event was defined as a period of immobility lasting at least 8 s.

One or more pause events were observed for 48 of the 63 Golgi

bodies (76%), and in each instance, the pause event occurred on

a microtubule. The remaining 24% of the Golgi bodies moved

rapidly in and out of framewithout undergoing a pause event.We

observed multiple Golgi bodies successively pausing at the

same positions (45 independent time series analyzed; presented

in the following section). Detailed analysis in untreated cells in the

hypocotyl revealed that Golgi pauses occurred in proximity to

MASCs (Figures 4D and 4E) and could last as long as 73 s (n = 34,

mean = 22 s). VHA-a1 compartments were also observed to

pause at MASCs, even in the absence of an associated Golgi

body (seven events from five independent experiments; Figure

4D). Visible connections between MASCs and Golgi bodies or

VHA-a1 compartments were prominent in CGA-, mannitol-, and

cycloheximide-treated cells (see Supplemental Movies 3 and 5

online; Figure 4E).

CSC Delivery to the PlasmaMembrane Depends on

Interactions between the Golgi Apparatus and

the Cytoskeleton

The high density of CSCs in cells near the top of the hypocotyl

has thus far prevented visualization of individual CSC particle

insertions (DeBolt et al., 2007b). We circumvented this prob-

lem by focusing our attention on cells closer to the hypocotyl

base, where CSC density is lower (Figure 2C). Through the

rigorous analysis of time series acquired in these untreated

cells, we were able to identify seven CSC insertion events. Each

insertion event was analyzed both manually and through

kymographs to verify the absence of the CSC in the initial

frames, to monitor the movement of all particles in proximity to

the identified insertion site, and to measure the velocity of the

newly inserted CSC. The newly inserted CSCs all began trans-

locating with a velocity approximating 270 nm/min, as expected

Figure 2. CSC Density and Trajectory Change with the Developmental

Gradient along the Length of Etiolated Hypocotyls.

(A) to (D) In the left-most and central column are maximum projected

z-stacks and average projections of time series, respectively, of CSCs

visualized in the outer face of epidermal cells in GFP-CESA3–expressing

plants. The right-most column shows maximum projected z-stacks of

GFP-TUA6–labeled microtubules in the same cell types.

(A) Mixed CSC trajectories and microtubule orientations in slowly

expanding cells of the apical hook. The high density of CSCs in these

cells correlates with the high rate of wall synthesis.

(B) A wide range of orientations from transverse to longitudinal is found in

rapidly expanding cells ;2 mm below the apical hook.

(C) Approximately 6 mm below the apical hook, in cells beginning to

cease elongation, CSC density is reduced and trajectories are consis-

tently longitudinal.

(D) CSCs are nearly absent from cells 10 mm below the hook, and

cortical microtubules are equally sparse. Small, bright, intracellular

compartments, referred to as MASCs (arrowheads), are abundant in

these basal hypocotyl cells. Bar = 10 mm.

Cellulose Synthase Trafficking 1145

Figure 3. Small, Bright, Intracellular Compartments Represent a MASC Whose Movement Is Dependent on Cortical Microtubule Dynamics.

(A) and (B) Maximum projected z-stacks showing GFP-CESA3 intracellular distribution in Arabidopsis epidermal hypocotyl cells. The ring-shaped

labeling of Golgi bodies is present in all treatments. Treatments with 50 mM cycloheximide (CHX) for 4 h (A) or 500 mM mannitol for 1.5 h (B) result in

disappearance of CSCs (cf. to Figure 1A) and accumulation of MASCs.

(C) and (D) Maximum projected z-stacks showing GFP-CESA3 intracellular distribution in the root tip upon treatment with 500 mM mannitol.

(C) After 2 min of treatment, CSCs are still visible in linear tracks in the plasma membrane, as they are in control treatments.

(D) After 6 min of treatment, CSCs have been internalized in MASCs. In five independent experiments, MASCs formed after ;5 to 6 min.

(E), (H), (J), and (L)Maximum projected z-stacks showing the GFP-CESA3 intracellular distribution in Arabidopsis epidermal hypocotyl cells. Labeling of

Golgi bodies is found in all treatments. Representative images are shown from treatments with 5 nM CGA for 5 h (E), 5 nM CGA followed by 5 nM CGA +

50 mM CytD (H), 4 mM taxol followed by 5 nM CGA + 4 mM taxol (J), and 7 mM oryzalin followed by 5 nM CGA + 7 mM oryzalin (L). In (E), (H), and (J),

MASCs formed at high densities near the cell surface, as they did in the presence of 5 nMCGA treatment alone, for an equal time duration. In (L), MASCs

did not form normally and are largely absent.

(F) Representative kymograph of a time series for 5 nM CGA treatment (corresponding to the region indicated by a line in [E]). Time is represented along

the vertical axis, and a 10-min scale bar is shown at left (see Supplemental Figure 1 online for information on interpreting kymographs). With 5 nM CGA

treatment alone, MASCs are highly motile and have inconstant velocities (cf. to Figure 1F; see Supplemental Movie 3 online).

(G)MASCs move along linear tracks that are remodeled over time. From left to right: the first time point in a series, average projection of the first 5 min of

the series, average projection of the last 6 min of the series, overlay of the two average projections. The red lines and arrowheads indicate new

trajectories that were not present at the 5 min time point.

(I), (K), and (M) Representative kymographs of time series for treatments with 5 nMCGA followed by 5 nMCGA + 50 mMCytD (I), 4 mM taxol followed by

1146 The Plant Cell

for GFP-CESA3–labeled CSCs (Desprez et al., 2007). In each

insertion event, appearance of new CSC particles was coinci-

dent with the pause of a Golgi body just beneath the plasma

membrane (Figures 5A to 5C). In two of these events, this pause

of the Golgi body resulted in the insertion of multiple particles in

a row (Figure 5B).

To further study the role of the Golgi apparatus and VHA-a1/

GFP-CESA3–containing compartments in CSC trafficking, we

inhibited the movement of these two compartments using

actin-depolymerizing agents. Incubation with CytD or Latrun-

culin B caused aggregation of Golgi bodies and VHA-a1/

GFP-CESA3–containing compartments and led to localized

reductions in CSC density (Figure 5D). Regions of the plasma

membrane containing high CSC density were found exclusively

above Golgi aggregates (seven independent experiments; Fig-

ures 5D to 5F).

We further explored the insertion of new CSC particles using

fluorescence recovery after photobleaching (FRAP) experiments

in untreated seedlings. We found that CSCs are not inserted

randomly, but in linear tracks resembling those present before

bleaching (Figure 5G). In these experiments, we could also

observe multiple CSCs inserted one after another in rows, and

these insertions were coincident with the passage of Golgi

bodies (13 experiments analyzed). Based on our observations

of the frequent pauses of Golgi bodies on microtubules (Figures

4A to 4C; see Supplemental Movie 4 online), and the insertion of

CSCs in rows along linear tracks, we hypothesized that interac-

tions between Golgi bodies and microtubules were directing the

plasma membrane delivery of CSCs.

To test if cortical microtubules define the insertion sites of

CSCs, we examined the effects of treatment with oryzalin or taxol

on the organization of CSCs in the plasmamembrane. Complete

removal of microtubules by treatment with 20 mM oryzalin

resulted in a uniform distribution of CSCs that lacked row

organization compared with controls (five independent experi-

ments; Figure 5I). By contrast, microtubule stabilization through

taxol treatment resulted in an enhanced alignment of CSCs (five

independent experiments; Figure 5J). In light of our results, we

propose that disorganization of CSC insertion in the presence of

microtubule-disrupting drugs is partly due to deregulation of

interactions between the Golgi/VHA-a1/GFP-CESA3–containing

compartments and cortical microtubules.

DISCUSSION

Synthesis of the plant cell wall requires secretion of cell wall

components and CSCs to specific locations at the cell periphery

bymotile organelles. In this study, we establish a role for the actin

and cortical microtubule cytoskeletons and the Golgi apparatus

in the delivery of CSCs to the plasma membrane and shed light

on the mechanism of endocytosis of CSCs. A schematic model

of the intracellular trafficking of GFP-CESA3 is presented in

Figure 6.

Four Distinct Localizations of GFP-CESA3

It had been previously established using confocal microscopy

that GFP-CESA3 and GFP-CESA6 (another related primary cell

wall CSC isoform) are present as discrete particles at the cell

surface and in punctae at the periphery of Golgi stacks (Paredez

et al., 2006; Desprez et al., 2007). In this study, the plasma

membrane localization of GFP-CESA3 surface particles was

confirmed by EM (Figure 1D), lending further support to the idea

that these motile, discrete particles represent CSCs actively

synthesizing cellulose. We also confirmed the localization of

GFP-CESA3 at the periphery of medial- and trans-Golgi cister-

nae by EM (Figures 1P and 1Q). Early reports show that the CSC

is fully assembled upon arrival in theGolgi apparatus (Haigler and

Brown, 1986; Rudolph, 1987). The bright punctae observed at

the periphery of Golgi stacks may therefore correspond to

assembled CSCs, as represented in Figure 6.

In our experiments, we were able to identify two previously

unknown localizations for the CESA protein. We found GFP-

CESA3 in a subset of VHA-a1 compartments (Figures 1J to 1O)

and a unique compartment we refer to as a MASC (Figures 1E to

1G and 3; see Supplemental Figure 2 online). In our analyses,

pausing of the Golgi apparatus was always observed prior to

insertion of a new CSC in the plasma membrane, whereas VHA-

a1/GFP-CESA3–containing compartments were not detected in

proximity to these insertion sites (Figures 5A to 5C). Although we

cannot exclude that CSCs pass through VHA-a1/GFP-CESA3–

containing compartments on a time scale too rapid to resolve in

our experiments, it is more likely that GFP-CESA3 passes

through the VHA-a1 compartment only on the endocytic route,

as shown for PIN2 (Robert et al., 2008). Since VHA-a1 labels the

Figure 3. (continued).

5 nM CGA + 4 mM taxol (K), or 7 mM oryzalin followed by 5 nM CGA + 7 mM oryzalin (M). Time is represented along the vertical axis, and a 10 min scale

bar is shown at left (see Supplemental Figure 1 online for information on interpreting kymographs). The kymograph in (M) represents the region indicated

by a line in (L).

(N) Frequency histogram comparing MASC velocities calculated from kymographs in replicates of the experiments described in (C), (H), and (I) (CGA

alone, n = 278; CGA + taxol, n = 989; CGA + CytD, n = 342). The points indicate the percentage of particles with velocities in the range indicated on the x

axis. Lines have been added to facilitate visualization of each population distribution.

(O) Cotyledon cells of untreated GFP-CESA3–expressing plant transiently transformed with mRFP-MBD (red). The maximum projected z-stack (left

panel) shows a GFP-CESA3-labeled MASC particle (green) marked by an arrowhead. Right panel, the average projection of the time series showing

translocation of the MASC along the mRFP-MBD–labeled microtubule (see Supplemental Movie 4 online).

(P)Montage of a deconvolved time series illustrating representative merging behavior of four aligned MASCs in cells treated with 50 mM cycloheximide

for 4 h. At 8.8 s, the fourth particle (from the top) merges with the third, and both continue cotranslocating. A second encounter at 19.8 s results in a rapid

split of the particle. The resulting three particles then remain relatively stable in their positions.

Bars = 10 mm.

Cellulose Synthase Trafficking 1147

TGN (Dettmer et al., 2006), these combined data question the

role of the TGN in secretion in plants.

Cellulose Synthesis Is Regulated by Rapid Internalization

into MASCs

Under conditions of osmotic stress, protein synthesis inhibition,

or cellulose synthesis inhibition, we findGFP-CESA3 present in a

novel compartment we refer to as a MASC (Figures 3A, 3B, and

3E). Lower densities of CSCs in fully elongated basal hypocotyl

cells correlated with a higher abundance of MASCs, consistent

with a reduced rate of primary cell wall cellulose synthesis in

these mature cells (Figure 2D). We conclude that MASCs result

from internalization of CSCs, rather than a block in secretion,

based on the following three observations: formation of MASCs

correlates with disappearance of CSCs at the plasmamembrane

(Figure 2D), redistribution of CSCs intoMASCsoccurswithin only

6 min in mannitol-treated roots (Figures 3C and 3D), and MASCs

still form despite prolonged inhibition of protein synthesis using

the drug cycloheximide (Figure 3A). Clathrin-mediated endocy-

tosis has recently been shown to be a predominant endocytic

pathway in plants (Dhonukshe et al., 2007). Given that the 45- to

50-nm diameter cytoplasmic domain of the CSC (Bowling and

Brown, 2008) would be expected to interfere with the formation

of the clathrin coat, it is likely that a differentmechanismoperates

in the endocytosis of CSCs (Figure 6). Indeed, the reduced

formation of MASCs in the presence of oryzalin suggests that

microtubules may have an active role in the internalization of

CSCs under CGA treatment. Oryzalin also prevented the redis-

tribution of YFP-CESA6 into similar small, bright compartments

Figure 4. Golgi Bodies Pause along Microtubules.

Golgi bodies and the VHA-a1 compartment exhibit physical interactions with MASCs. Bars = 10 mm.

(A) to (C) Deconvolved images of a cotyledon cell from an untreated GFP-CESA3–expressing plant transiently transformed with mRFP-MBD (red).

(A) CSCs are observed along the cortical microtubules. The arrowhead shows a future pause site of the Golgi body marked by an arrow.

(B) Average projection of the time series. The yellow color indicates colocalization between motile CSCs and microtubules.

(C) A blue line indicating the trajectory of the Golgi body is overlaid on the average projection. The Golgi body follows this microtubule and pauses

directly on it. See Supplemental Movie 4 online for a second example.

(D) Deconvolved time series from untreated plants coexpressing GFP-CESA3 (green) and RFP-VHA-a1 (red). An associated Golgi/VHA-a1

compartment (arrow) moves rapidly into view and pauses at an immobile MASC (arrowhead) for ;20 s. At the 52 s time point, the Golgi dissociates,

but the VHA-a1 compartment remains colocalized with the MASC. On the right-most image, the blue line indicates the trajectory of the VHA-a1

compartment.

(E) Connections are visible between Golgi and MASCs during interactions. Representative deconvolved time series (same series as in Figure 3L) from a

50-mM cycloheximide-treated seedling showing a GFP-CESA3–labeled Golgi body (blue arrow) changing trajectory and pausing at a MASC (red

arrowhead).

1148 The Plant Cell

in seedlings treated with Morlin, an inhibitor of CESA and

microtubule dynamics (DeBolt et al., 2007a). Together, these

results show that cellulose synthesis is regulated in part through

the internalization of CSCs in intracellular compartments and that

internalization of CSCs intoMASCs is environmentally regulated.

The higher abundance of MASCs in fully elongated basal hypo-

cotyl cells compared with growing apical hypocotyl cells sug-

gests that internalization in MASCs is also developmentally

programmed.

MASCMovement Is Dependent on an Intact, Dynamic

Cortical Microtubule Array

The unusual movement pattern of MASCs (see Supplemental

Movies 3 and 5 online) has not been reported for any other plant

Figure 5. CSC Delivery to the Plasma Membrane Depends on Interac-

tions between the Golgi Apparatus and the Cytoskeleton.

(A) to (C) Images from times series showing three examples of de novo

CSC insertion in the plasma membrane focal plane by GFP-CESA3–

labeled Golgi bodies in untreated seedlings. Kymographs of the time

series are at the far right, with time represented along the vertical axis as

indicated by the scale bars (see Supplemental Figure 1 online for

information on interpreting kymographs).

(A) and (B) In the left-most image, no particle can be detected at or near

the future position of insertion (corresponding to black region above the

trace in the kymograph). The central image shows a Golgi body pausing

beneath the plasma membrane, which can also be detected on the

kymograph at right (white smear marked by arrow). In (A), a single CSC is

inserted and begins moving nearly immediately, as evidenced by the

angled white trace on the kymograph. In (B), three new CSC are inserted

in a row and begin moving steadily (only one is kymographed).

(C) Image 1 shows a first Golgi body pausing at a specific site marked by

an arrowhead. Image 2 shows this pause does not result in an insertion

event. In image 3, a new Golgi body moves into frame and pauses at the

same site (arrowhead), resulting in insertion of a CSC, marked by the

arrowhead in image 4. The white smears on the kymograph (marked by a

gray arrow and black arrow) illustrate the pauses of successive Golgi and

the newly inserted CSC that begins moving with a steady velocity.

(D) to (F) GFP-CESA3 intracellular distribution in epidermal hypocotyl

cells treated with 1 mM Latrunculin B.

(D) Maximum projected z-stack in the plasma membrane focal plane

shows localized reductions in CSC density.

(E) Average projection of the time series illustrates that CSC are migrat-

ing normally along linear trajectories.

(F) Maximum projected z-stack ;600 nm below the plasma membrane

showing Golgi body aggregates (arrows). Positions of Golgi body ag-

gregates correspond to the areas of highest CSC density in (A).

(G) FRAP in untreated epidermal hypocotyl cells. The prebleach image is

an average of 10 frames showing CSCs arranged in obliquely oriented

rows. The central image was acquired ;500 ms after bleaching the

indicated area. In the average projection of the postbleach frames, newly

inserted CSCs are arranged in oblique rows and migrate along linear

tracks similar to those present in the prebleach image.

(H) to (J) Images of a single time point showing the organization of GFP-

CESA3–labeled CSCs in epidermal hypocotyl cells.

(H) Untreated seedling showing CSCs arranged in rows.

(I) After an 8-h treatment with 20 mMoryzalin, row organization is lost and

CSCs are scattered uniformly throughout the plasma membrane.

(J) A 7-h treatment with 4 mM taxol increases the level of CSC row

organization.

Bars = 10 mm.

Cellulose Synthase Trafficking 1149

intracellular compartments or markers thus far, to our knowl-

edge. The merging behavior of MASCs and their movement

along linear tracks clearly distinguish them from the actomyosin-

based movement of Golgi bodies (Sparkes et al., 2008) and the

nonlinear movement of the VHA-a1 compartment (cf. Supple-

mental Movies 2 and 3 online). MASC movement is not signif-

icantly changed by the fragmentation of actin or the inhibition of

cytoplasmic streaming (Figure 3I). Both the velocity distribution

and the qualitative characteristics of MASC movement remain

the same in controls and CytD-treated cells (Figure 3N). By

contrast, the depolymerization of the cortical microtubule array

or the inhibition of array dynamics by treatment with taxol has a

striking effect on the formation and velocity ofMASCs (Figures 3J

to 3N). The fact that stabilization of microtubule array dynamics

also immobilizes MASCs indicates that their movement may not

be kinesin driven, but rather depends on intrinsic properties of

the microtubule array dynamics itself.

Our findings lend support to the hypothesis that CSCs tran-

siently interact with cortical microtubules, rather than merely

passively diffusing between them. For themicrotubule-dependent

Figure 6. Schematic Illustration of CESA Trafficking.

Objects were drawn to scale based on published data and EM micrographs obtained in this study. CSCs may be secreted directly to the plasma

membrane from the Golgi apparatus or may transit rapidly through the TGN. CSC insertion is accompanied by interactions between the Golgi apparatus

and cortical microtubules. We define MASCs as a microtubule-associated compartment containing the CSC, which may encompass both short-lived

secretory vesicles and more stable endocytic vesicles. Although we were not able to directly observe CSC secretory vesicles, we hypothesize that they

may associate with microtubules if the CSC itself is competent for interactions with the microtubule array. It is possible that CSC secretory vesicles do

not associate with microtubules or that secretion occurs in the absence of a vesicular intermediate, though we find these hypotheses less likely. Once

fused with the plasma membrane, CSCs are propelled by the polymerization of glucan chains into cellulose microfibrils, and their trajectories may be

guided by CSC-cortical microtubule interactions of a transient nature. Internalization of CSCs is promoted under certain conditions (e.g., osmotic

stress, cell growth cessation, and cellulose synthesis inhibitor treatment). Consideration of the size of the CSC rosette (;25-nm diameter; Mueller and

Brown, 1980) and the CSC cytoplasmic domain (45 to 50-nm diameter; Bowling and Brown, 2008) with respect to the size of clathrin-coated vesicles

(CCV) (30-nm diameter excluding coat; Dhonukshe et al., 2007) suggests that endocytosis of CSC does not occur via conventional clathrin-coated

vesicle–mediated endocytosis. When internalized, CSCs show surprisingly stable associations with microtubules, which may be due to an interaction

between a microtubule-associated protein and the CSC itself or a coat protein specifically recruited to MASC vesicles. Together, our results support a

model in which interactions with cortical microtubules have a role both in plasma membrane delivery and internalization of CSCs. These interactions

provide a regulatory mechanism for the dynamic remodeling of CSC secretion during cell expansion and differentiation.

1150 The Plant Cell

movement of MASCs to be possible, either an interaction exists

between a microtubule-associated protein and a coat protein

specifically recruited to MASC vesicles or an interaction exists

between a microtubule-associated protein and a component of

the CSC itself. It will be interesting to determine if microtubule-

associated proteins can be copurified with CSCs under CGA

treatment. The discovery of proteins able tomediate interactions

between CSC and microtubules will be essential to reveal

the mechanism underlying microtubule-microfibril alignment in

growing cells.

Trafficking of CSCs Is Dependent on Actin and

Microtubule Cytoskeletons

Our combined results strongly support a role for the microtubule

cytoskeleton in targeting CSC secretion. CSC density and cor-

tical microtubule density are correlated during cellular differen-

tiation in the hypocotyl (Figure 2), suggesting that cortical

microtubules may be involved in regulating CSC delivery to the

plasmamembrane.Wedemonstrate that newCSC insertions are

preceded by pauses of the Golgi apparatus at specific sites

beneath the plasmamembrane (Figures 5A to 5C). The so-called

stop-and-go behavior of Golgi bodies was proposed to be due to

pauses at ER export sites (Nebenfuhr et al., 1999); however,

recent evidence suggests that ER export sites and Golgi stacks

may form mobile secretory units (daSilva et al., 2004). Our data

reveal that Golgi bodies pause at cortical microtubules, and

these pauses often result in insertion of CSCs at the plasma

membrane (Figures 5A to 5C).

Evidence of Golgi bodies pausing at cortical microtubules has

also been observed in cells undergoing high rates of secondary

cell wall synthesis. In developing xylem vessels, secondary cell

wall synthesis is nonuniform and characterized by cell wall

thickenings in a banded pattern. Within these bands, CSCs

are more abundant in the plasma membrane (Herth, 1985),

and microtubules are found in organized bundles (Hepler and

Newcomb, 1964). It was recently shown that organelles labeled

by YFP-CESA7, of which a subset colocalizes with the Golgi

marker mannosidase I, pause at bands marked by microtubules

(Wightman and Turner, 2008). These organelles appear to pause

less frequently following depolymerization of the microtubule

bands by oryzalin, although the statistical significance of this

observation can be questioned. Thus, there is limited evidence

that delivery of secondary cell wall CSCs to the plasma mem-

brane in developing xylem vessels may also be related to the

pausing of Golgi bodies on microtubules.

In plants, Golgi bodies, mitochondria, and peroxisomes are

three organelles known to move along actin filaments via myo-

sins (Sparkes et al., 2008). Although the actomyosin systemmay

drive the movement of these organelles, there is increasing

evidence showing their proper function also depends on inter-

actions with cortical microtubules. The plant peroxisomal mul-

tifunctional protein localizes to peroxisomes and decorates the

cortical microtubule array, thus mediating interactions between

peroxisomes and cortical microtubules (Chuong et al., 2005).

The Arabidopsis internal motor kinesin, Kinesin-13A, and the

homologousGossypium hirsutumKinesin-13A have been shown

to localize to Golgi bodies (Lu et al., 2005). Arabidopsis kinesin-

13amutants show aggregation of Golgi bodies in trichomes and

have defects in trichome morphogenesis, suggesting a role for

Kinesin-13A in regulating secretion of cell wall components by

the Golgi apparatus (Lu et al., 2005). Similarly, mutants in the

kinesin FRAGILE FIBER1 show alterations in cellulose deposition

and reduced fiber strength (Zhong et al., 2002). Finally, cortical

microtubules were found in close association with electron-

dense secretory vesicles targeted to pectic mucilage pockets in

Arabidopsis seed coat cells (McFarlane et al., 2008). Thus, the

actin filament and microtubule components of the cytoskeleton

are highly interdependent, and each plays an important role in

secretion and organelle movement.

Our data support a model in which the actomyosin system

serves to distribute Golgi bodies throughout the cell, and sites of

secretion are defined through interactions with the microtubule

cytoskeleton. This finding explains the diffuse, randomized ap-

pearance of CSCs at the plasma membrane following oryzalin

treatment and isalsosupportedbyobservationsofenhancedCSC

alignment under taxol treatment. In conclusion, wepropose that in

growing cells, cortical microtubules not only direct the trajectories

of CSCs in the plasmamembrane, but also regulate the delivery of

CSCs to the plasma membrane and possibly their internalization.

This allows the dynamic remodeling of CSC secretion patterns

during the expansion and differentiation of plant cells.

METHODS

Plant Material and in Vitro Growth Conditions

Seeds of Arabidopsis thaliana ecotype Columbia (Col) were provided by

K. Feldman (University of Arizona, Tucson, AZ), and GFP-CESA3 and

VHA-a1-mRFP lines were described previously (Dettmer et al., 2006;

Desprez et al., 2007). GFP-TUA6 (Ueda et al., 1999) and EB1-GFP (Chan

et al., 2003) lines were a generous gift of J. Chan (John Innes Centre,

Norwich, UK). GFP-FABD2 was a gift of M. Pastuglia and has been

described previously (Voigt et al., 2005). For imaging, seedlings were

cultured in chambers as described (Chan et al., 2007). For hypocotyl

measurements, seedlings were grown at 208C in Petri dishes on Estelle

and Somerville medium (Estelle and Somerville, 1987) without sucrose.

Seeds were cold-treated at 48C for 48 h, exposed to fluorescent white

light (150 mmol m22 s21, True Light; Philips) for 6 h, and then wrapped in

aluminum foil. The age of the seedlings was defined with respect to the

end of the cold treatment.

Hypocotyl Length Measurements

Etiolated seedlings were fixed in 0.2% paraformaldehyde. The lengths of

at least 50 hypocotyls per treatment were measured using Optimas 5.2

software as described (Refregier et al., 2004).

Generation of the GFP-CESA3/RFP-VHA-a1 Transgenic Line

Pollen from GFP-CESA3–expressing plants in je5cesa3 background

(Desprez et al., 2007) was used to fertilize plants expressing VHA-a1-

mRFP (Dettmer et al., 2006). F1 seeds were collected and amplified. F2

seedlings were screened by PCR for the presence of both markers

and the cesa3 mutation. F3 selected seedlings were used for imaging.

Drug Treatments

For hypocotyl measurements, Col-0 seeds were germinated on media

containing CGA 325’615 (CGA) in the range of 0 (mock treatment DMSO

Cellulose Synthase Trafficking 1151

dilution) to 5 nM and measured after 4 d. For Fourier transform infrared

microspectroscopy, Col-0 seeds were germinated on media plates

containing drugs at the indicated concentrations, and mutant seeds

were germinated on normal media. Measurements were taken after 4 d.

For live-cell imaging, treatment and mock treatment solutions were

injected into chambers using a fine-needle syringe under green light.

Seedlings expressing markers to verify drug treatment efficacy were

cultured in the same chambers as GFP-CESA3 seedlings. Single treat-

ments were performedwith 5 nMCGA for 5 to 6 h, 1 mMLatrunculin B for 6

h, 20mMoryzalin for 8 h, or 500mMmannitol for 3 h. Treatments combining

two different inhibitors were performed as follows: 7 mM oryzalin for 7 h,

then 5 nM CGA + 7 mM oryzalin for 5 h before analysis; 4 mM taxol for 8 h,

then 5 nMCGA + 4 mM taxol for 6 h; and 5 nM CGA for 5.5 h followed by 5

nM CGA + 50 mM CytD for 30 min. All chemicals (Sigma-Aldrich) were

dissolved in DMSO stock solutions and kept at2208C for no longer than 6

months. Stocks were diluted at least 1000-fold in buffered water before

use. CGA 325’615 was kindly provided by Bayer CropScience.

For the experiment represented in Supplemental Figures 2A to 2F

online, 3-d-old light-grown seedlings were first incubated in a solution of

dilute DMSO (mock treatment) or 5 nM CGA for 10 to 15 min and then

mounted in a solution of 50 mM FM4-64 and imaged within 5 min.

Transient Expression in Arabidopsis Seedlings

The mRFP-MBD vector was constructed by M. Pastuglia (Institut Jean-

Pierre Bourgin-Institut National de la Recherche Agronomique). TheMBD

of microtubule-associated protein 4 was amplified from the GFP-MBD

construct (Marc et al., 1998) and cloned into Gateway vector pDONR207

(Invitrogen). To subclone the mRFP-MBD fusion downstream of the 35S

promoter, pDONR207:MBD was used in an LR reaction with destination

vector pH7WGR2 (Karimi et al., 2002). mRFP-MBD and GFP-CESA3

expression vectors were electroporated in Agrobacterium tumefaciens.

For transient expression studies,A. tumefaciens cultures were cultured

overnight at 288C and diluted to OD600 = 1 in 5% sucrose with 200 mM

acetosyringone. Infiltration of 4-d-old Arabidopsis seedlings was per-

formed as described (Marion et al., 2008).

Spinning Disk Microscopy and Image Analysis

During all experiments, cell viability was verified by monitoring cytoplas-

mic streaming. Hypocotyls of 3-d-old etiolated seedlings or root tips of

3-d-old light-grown seedlings were analyzed on an Axiovert 200M

microscope (Zeiss) equipped with a Yokogawa CSU22 spinning disk,

Zeiss 100/1.4 numerical aperture oil objective, and Andor EMCCD iXon

DU 895 camera (Plateforme d’Imagerie Dynamique, Institut Pasteur,

Paris, France). A 488- or 560-nm diode-pumped solid-state laser was

used for excitation, and emission was collected using band-pass 488/25

and 568/25 filters (Semrock) for GFP andmRFP, respectively. Projections

presented in figures were acquired with aminimum of three z-slices using

a 300-nm step interval. For analysis of particle interactions, time series

were acquired at a frame rate of 2 to 4 frames/second.

FRAP experiments were performed using an inverted microscope

(Nikon TE2000) equipped with a cooled interline CCD detector (Roper

CoolsnapHQ2) and a piezoelectric stage driver (LVDT; Physik Instrument)

and operated withMetaMorph 7.1.6 software (Molecular Devices). Series

were acquired at 2 frames/second with a FRAP duration of 20 ms and an

xy resolution of 130 nm/pixel (2 3 2 binning).

When required, fluorescent images were deconvolved using a classic

maximum likelihood estimation deconvolution algorithm (Huygens Pro-

fessional v. 3.1; Scientific Volume Imaging). The integrated morphometry

analysis function in Metamorph Offline (V. 7.0; Molecular Devices) was

used to detect and measure the orientations of microtubules. Particle

velocities were calculated from kymographs created in Image J (W.

Rasband, National Institutes of Health). Kymograph montages were

generatedwith the help of an Image J plug-in developed by F. Cordelieres

(Plateforme d’Imagerie Cellulaire et Tissulaire , Unite Mixte de Recherche

146, Institut Curie). Image size and contrast were adjusted when needed

using Adobe Photoshop CS v. 8.0.1 (Adobe Systems).

Confocal Laser Scanning Microscopy

The images presented in Supplemental Figures 2A to 2F online were

obtained using a spectral Leica SP2 AOBS confocal microscope (Leica

Microsystems). GFP-CESA3 and the sterol dye FM4-64 (Molecular

Probes) were excited at 488 nm, and emission was collected at 492 to

551 nm and 645 to 833 nm, respectively. A 363 water immersion

objective (numerical aperture 1.2) and a zoom factor of eight were used

to collect z-series in a 512-scan format.

Fourier Transform Infrared Microspectroscopy

Four-day-old seedlings were crushed between two barium fluoride

windows and rinsed abundantly with distilled water for 2 min, before

drying at 378C for 20 min. For each mutant, 20 spectra were collected

from hypocotyls from four independent cultures (five seedlings from each

culture) as described (Mouille et al., 2003). The collected spectra were

baselined and normalized as described, and statistical analysis was

performed by a Student’s t test (Robin et al., 2003).

Crystalline Cellulose Measurements

Wild-typeCol-0 plantswere grown in darkness on plates containing 0.625

nM CGA or a DMSO mock treatment solution. Four-day-old seedlings

were harvested and stored in 70%ethanol at 48C. Crude cell wall extracts

were prepared as described (Reiter et al., 1993)withminor changes. Seed

coats were removed and plant material was incubated twice during 1 h at

708C in 70% ethanol. Pellets were washed in methanol for 5 min and

acetone for 2 min and were vacuum-dried. The dry cell wall pellets were

weighed in 2-mL reaction tubes, suspended in 3-mL Updegraff reagent

(acetic acid:nitric acid:water; 8:1:2; v/v), and further incubated in a boiling

water bath for 30 min. Crystalline cellulose was determined as described

(Scott and Melvin, 1953; Updegraff, 1969). Absorbance was determined

at 630 nm (Labsystems iEMS Reader MF). Measurements were per-

formed in duplicate using independent seed batches.

Transmission Electron Microscopy

Degassed seedlings were high-pressure frozen, freeze-substituted, and

rehydrated (Ripper et al., 2008) before samples were infiltrated with the

cryoprotectant polyvinylpyrrolidone and sucrose, mounted on a speci-

men holder, and frozen in liquid nitrogen for ultrathin cryosectioning as

described before (Reichardt et al., 2007). Ultrathin thawed cryosections

were treated with 0.5% BSA, 1% milk powder in PBS (20 min) before

labeling with rabbit anti-GFP antibodies (Torrey Pines Biolabs; 1:500)

and silver-enhanced Nanogold coupled to goat anti-rabbit IgG (1:60)

(HQSilver, 8.5 min; Nanoprobes) for 60 min each.

Accession Numbers

Sequence data from this article can be found in the Arabidopsis Genome

Initiative or GenBank/EMBL databases under the following accession

numbers: CESA3, At5g05170; VHA-a1, At2g28520; EB1a, At3g47690;

TUA6, At4g14960; MBD from MAP4, M72414; and FABD2 from At FIM1,

At4g26700.

Supplemental Data

The following materials are available in the online version of this article.

1152 The Plant Cell

Supplemental Figure 1. Image Analysis Methods.

Supplemental Figure 2. Localization of GFP-CESA3 following CGA

Treatment in FM4-64–Labeled Cells.

Supplemental Figure 3. CGA Causes Specific Reductions in Crys-

talline Cellulose and Inhibits Cellular Elongation.

Supplemental Figure 4. CGA Does Not Have an Observable Effect

on the Cytoskeleton or Golgi.

Supplemental Figure 5. Effects of Drug Treatments Used in This

Study.

Supplemental Movie 1. CSCs Move Bidirectionally with Constant

Velocities.

Supplemental Movie 2. GFP-CESA3–Labeled Golgi Bodies and the

VHA-a1 Compartment Rapidly Associate and Dissociate.

Supplemental Movie 3. CGA-Induced MASCs Have a Wide Range of

Velocities and Merge with One Another along Linear Tracks.

Supplemental Movie 4. GFP-CESA3–Labeled Golgi Bodies Pause on

Microtubules.

Supplemental Movie 5. Physical Interactions Occur between

Mannitol-Induced MASCs and GFP-CESA3–Labeled Golgi Bodies.

ACKNOWLEDGMENTS

We thank Fabrice Cordelieres (Plateforme d’Imagerie Cellulaire et

Tissulaire, Unite Mixte de Recherche 146, Institut Curie, Gif-sur-Yvette,

France) for development of plug-ins to facilitate image analysis; Vincent

Fraisier (Institut Curie, PICT-IBiSA Imaging Facility, Paris, France) for

assistance with FRAP experiments; Nicolas Chenouard (Unite Analyse

d’Images Quantitative, Institut Pasteur, Paris, France), Anne Danckaert,

and Christophe Machu for helpful advice and image analysis tools

(Plateforme d’Imagerie Dynamique, Institut Pasteur, Paris, France);

Olivier Grandjean (Plateforme de Cytologie et d’Imagerie Vegetale,

Institut Jean-Pierre Bourgin, Institut National de la Recherche Agrono-

mique) for technical assistance with the Leica SP2; and Jordi Chan

(John Innes Centre, Norwich, UK) and Martine Pastuglia (Institut Jean-

Pierre Bourgin, Institut National de la Recherche Agronomique) for

contributing fluorescently tagged cytoskeleton markers. E.F.C. was

supported by the National Agency for Research Project ‘‘IMACEL’’

ANR-06-BLAN-0262. V.B. was supported by the European Union

Framework Program 6 (FP6) “CASPIC” NEST-CT-2004-028974 and

“Wallnet” RTN 512265. Part of this work was supported by FP6 program

037704 “AGRON-OMICS.” The Plateforme de Cytologie et d’Imagerie

Vegetale is supported in part by the Region Ile-de-France.

Received December 19, 2008; revised March 10, 2009; accepted March

19, 2009; published April 17, 2009.

REFERENCES

Baskin, T.I. (2001). On the alignment of cellulose microfibrils by cortical

microtubules: A review and a model. Protoplasma 215: 150–171.

Baskin, T.I. (2005). Anisotropic expansion of the plant cell wall. Annu.

Rev. Cell Dev. Biol. 21: 203–222.

Bowling, A.J., and Brown, R.M., Jr. (2008). The cytoplasmic domain of

the cellulose-synthesizing complex in vascular plants. Protoplasma

233: 115–127.

Chan, J., Calder, G.M., Doonan, J.H., and Lloyd, C.W. (2003). EB1

reveals mobile microtubule nucleation sites in Arabidopsis. Nat. Cell

Biol. 5: 967–971.

Chan, J., Calder, G.M., Fox, S., and Lloyd, C.W. (2007). Cortical

microtubule arrays undergo rotary movements in Arabidopsis hypo-

cotyl epidermal cells. Nat. Cell Biol. 9: 171–175.

Chuong, S.D., Park, N.I., Freeman, M.C., Mullen, R.T., and Muench,

D.G. (2005). The peroxisomal multifunctional protein interacts with

cortical microtubules in plant cells. BMC Cell Biol. 6: 40.

daSilva, L.L., Snapp, E.L., Denecke, J., Lippincott-Schwartz, J.,

Hawes, C., and Brandizzi, F. (2004). Endoplasmic reticulum export

sites and Golgi bodies behave as single mobile secretory units in plant

cells. Plant Cell 16: 1753–1771.

DeBolt, S., Gutierrez, R., Ehrhardt, D.W., Melo, C.V., Ross, L., Cutler,

S.R., Somerville, C., and Bonetta, D. (2007a). Morlin, an inhibitor of

cortical microtubule dynamics and cellulose synthase movement.

Proc. Natl. Acad. Sci. USA 104: 5854–5859.

DeBolt, S., Gutierrez, R., Ehrhardt, D.W., and Somerville, C. (2007b).

Nonmotile cellulose synthase subunits repeatedly accumulate within

localized regions at the plasma membrane in arabidopsis hypocotyl

cells following 2,6-dichlorobenzonitrile treatment. Plant Physiol. 145:

334–338.

Desprez, T., Juraniec, M., Crowell, E.F., Jouy, H., Pochylova, Z.,

Parcy, F., Hofte, H., Gonneau, M., and Vernhettes, S. (2007).

Organization of cellulose synthase complexes involved in primary cell

wall synthesis in Arabidopsis thaliana. Proc. Natl. Acad. Sci. USA 104:

15572–15577.

Dettmer, J., Hong-Hermesdorf, A., Stierhof, Y.D., and Schumacher,

K. (2006). Vacuolar H+-ATPase activity is required for endocytic and

secretory trafficking in Arabidopsis. Plant Cell 18: 715–730.

Dhonukshe, P., Aniento, F., Hwang, I., Robinson, D.G., Mravec, J.,

Stierhof, Y.D., and Friml, J. (2007). Clathrin-mediated constitutive

endocytosis of PIN auxin efflux carriers in Arabidopsis. Curr. Biol. 17:

520–527.

Emons, A.M.C. (1982). Microtubules do not control microfibril orienta-

tion in a helicoidal wall. Protoplasma 113: 85–87.

Estelle, M.A., and Somerville, C.R. (1987). Auxin-resistant mutants of

Arabidopsis thaliana with an altered morphology. Mol. Gen. Genet.

206: 200–206.

Foissner, I., and Wasteneys, G.O. (2007). Wide-ranging effects of eight

cytochalasins and latrunculin A and B on intracellular motility and

actin filament reorganization in characean internodal cells. Plant Cell

Physiol. 48: 585–597.

Gendreau, E., Traas, J., Desnos, T., Grandjean, O., Caboche, M.,

and Hofte, H. (1997). Cellular basis of hypocotyl growth in Arabidop-

sis thaliana. Plant Physiol. 114: 295–305.

Granger, C.L., and Cyr, R.J. (2001). Spatiotemporal relationships

between growth and microtubule orientation as revealed in living

root cells of Arabidopsis thaliana transformed with green-fluorescent-

protein gene construct GFP-MBD. Protoplasma 216: 201–214.

Green, P.B. (1962). Mechanism for plant cellular morphogenesis. Sci-

ence 138: 1404–1405.

Green, P.B., and Selker, J.M.L. (1991). Mutual alignments of cell walls,

cellulose and cytoskeletons: Their role in meristems. In The Cytoskel-

etal Basis of Plant Growth and Form, C.W. Lloyd, ed (London:

Academic Press), pp. 303–322.

Haigler, C.H., and Brown, R.M. (1986). Transport of rosettes from the

Golgi apparatus to the plasma membrane in isolated mesophyll cells

of Zinnia elegans during differentiation to tracheary elements in

suspension culture. Protoplasma 134: 111–120.

Hepler, P.K., and Newcomb, E.H. (1964). Microtubules and fibrils in the

cytoplasm of Coleus cells undergoing secondary wall deposition.

J. Cell Biol. 20: 529–532.

Herth, W. (1985). Plasma-membrane rosettes involved in localized wall

thickening during xylem vessel formation of Lepidium sativum L.

Planta 164: 12–21.

Cellulose Synthase Trafficking 1153

Himmelspach, R., Williamson, R.E., and Wasteneys, G.O. (2003).

Cellulose microfibril alignment recovers from DCB-induced disruption

despite microtubule disorganization. Plant J. 36: 565–575.

Karimi, M., Inze, D., and Depicker, A. (2002). GATEWAY vectors for

Agrobacterium-mediated plant transformation. Trends Plant Sci. 7:

193–195.

Kimura, S., Laosinchai, W., Itoh, T., Cui, X., Linder, C.R., and Brown,

R.M., Jr. (1999). Immunogold labeling of rosette terminal cellulose-

synthesizing complexes in the vascular plant vigna angularis. Plant

Cell 11: 2075–2086.

Ledbetter, M.C., and Porter, K.R. (1963). A ‘‘microtubule’’ in plant cell

fine structure. J. Cell Biol. 19: 239–250.

Lu, L., Lee, Y.R., Pan, R., Maloof, J.N., and Liu, B. (2005). An internal

motor kinesin is associated with the Golgi apparatus and plays a role in

trichome morphogenesis in Arabidopsis. Mol. Biol. Cell 16: 811–823.

Marc, J., Granger, C.L., Brincat, J., Fisher, D.D., Kao, Th.,

McCubbin, A.G., and Cyr, R.J. (1998). A GFP-MAP4 reporter gene

for visualizing cortical microtubule rearrangements in living epidermal

cells. Plant Cell 10: 1927–1940.

Marion, J., Bach, L., Bellec, Y., Meyer, C., Gissot, L., and Faure, J.D.

(2008). Systematic analysis of protein subcellular localization and

interaction using high-throughput transient transformation of Arabi-

dopsis seedlings. Plant J. 56: 169–179.

McFarlane, H.E., Young, R.E., Wasteneys, G.O., and Samuels, A.L.

(2008). Cortical microtubules mark the mucilage secretion domain of

the plasma membrane in Arabidopsis seed coat cells. Planta 227:

1363–1375.

Mouille, G., Robin, S., Lecomte, M., Pagant, S., and Hofte, H. (2003).

Classification and identification of Arabidopsis cell wall mutants using

Fourier-transform infrared (FT-IR) microspectroscopy. Plant J. 35:

393–404.

Mueller, S.C., and Brown, R.M., Jr. (1980). Evidence for an intra-

membrane component associated with a cellulose microfibril-

synthesizing complex in higher plants. J. Cell Biol. 84: 315–326.

Nebenfuhr, A., Gallagher, L.A., Dunahay, T.G., Frohlick, J.A.,

Mazurkiewicz, A.M., Meehl, J.B., and Staehelin, L.A. (1999). Stop-

and-go movements of plant Golgi stacks are mediated by the acto-

myosin system. Plant Physiol. 121: 1127–1142.

Paredez, A.R., Somerville, C.R., and Ehrhardt, D.W. (2006). Visuali-

zation of cellulose synthase demonstrates functional association with

microtubules. Science 312: 1491–1495.

Peng, L., Xiang, F., Roberts, E., Kawagoe, Y., Greve, L.C., Kreuz, K.,

and Delmer, D.P. (2001). The experimental herbicide CGA 325’615

inhibits synthesis of crystalline cellulose and causes accumulation of

non-crystalline beta-1,4-glucan associated with CesA protein. Plant

Physiol. 126: 981–992.

Probine, M.C., and Preston, R.D. (1961). Cell growth and the structure

and mechanical properties of the wall in internodal cells of Nitella

opaca. I. Wall structure and growth. J. Exp. Bot. 12: 261–282.

Probine, M.C., and Preston, R.D. (1962). Cell growth and the structure

andmechanical properties of thewall in internodal cells ofNitella opaca.

II. Mechanical properties of the walls. J. Exp. Bot. 13: 111–127.

Refregier, G., Pelletier, S., Jaillard, D., and Hofte, H. (2004). Interac-

tion between wall deposition and cell elongation in dark-grown

hypocotyl cells in Arabidopsis. Plant Physiol. 135: 959–968.

Reichardt, I., Stierhof, Y.D., Mayer, U., Richter, S., Schwarz, H.,

Schumacher, K., and Jurgens, G. (2007). Plant cytokinesis requires de

novo secretory trafficking but not endocytosis. Curr. Biol. 17: 2047–2053.

Reiter, W.D., Chapple, C.C., and Somerville, C.R. (1993). Altered

growth and cell walls in a fucose-deficient mutant of arabidopsis.

Science 261: 1032–1035.

Ripper, D., Schwarz, H., and Stierhof, Y.D. (2008). Cryo-section

immunolabelling of difficult to preserve specimens: advantages

of cryofixation, freeze-substitution and rehydration. Biol. Cell 100:

109–123.

Robert, S., Chary, S.N., Drakakaki, G., Li, S., Yang, Z., Raikhel, N.V.,

and Hicks, G.R. (2008). Endosidin1 defines a compartment involved

in endocytosis of the brassinosteroid receptor BRI1 and the auxin

transporters PIN2 and AUX1. Proc. Natl. Acad. Sci. USA 105: 8464–

8469.

Robin, S., Lecomte, M., Hofte, H., and Mouille, G. (2003). A procedure

for the clustering of cell wall mutants in the model plant Arabidopsis

based on fourier-transform infrared (FT-IR) spectroscopy. J. Appl.

Stat. 30: 669–681.

Robinson, D.G., Jiang, L., and Schumacher, K. (2008). The endo-

somal system of plants: Charting new and familiar territories. Plant

Physiol. 147: 1482–1492.

Rudolph, U. (1987). Occurrence of rosettes in the ER membrane of

young Funaria hygrometrica protonemata. Naturwissenschaften 74:

439.

Sainsbury, F., Collings, D.A., Mackun, K., Gardiner, J., Harper, J.D.I.,

and Marc, J. (2008). Developmental reorientation of transverse cor-

tical microtubules to longitudinal directions: A role for actomyosin-

based streaming and partial microtubule-membrane detachment.

Plant J. 56: 116–131.

Scott, T.A., and Melvin, E.H. (1953). Determination of dextran with

anthrone. Anal. Biochem. 25: 1656–1661.

Shaw, S.L., Kamyar, R., and Ehrhardt, D.W. (2003). Sustained micro-

tubule treadmilling in Arabidopsis cortical arrays. Science 300: 1715–

1718.

Smith, L.G., and Oppenheimer, D.G. (2005). Spatial control of cell

expansion by the plant cytoskeleton. Annu. Rev. Cell Dev. Biol. 21:

271–295.

Somerville, C. (2006). Cellulose synthesis in higher plants. Annu. Rev.

Cell Dev. Biol. 22: 53–78.

Sparkes, I.A., Teanby, N.A., and Hawes, C. (2008). Truncated myosin

XI tail fusions inhibit peroxisome, Golgi, and mitochondrial movement

in tobacco leaf epidermal cells: A genetic tool for the next generation.

J. Exp. Bot. 59: 2499–2512.

Sugimoto, K., Himmelspach, R., Williamson, R.E., and Wasteneys,

G.O. (2003). Mutation or drug-dependent microtubule disruption

causes radial swelling without altering parallel cellulose microfibril

deposition in Arabidopsis root cells. Plant Cell 15: 1414–1429.

Ueda, K., Matsuyama, T., and Hashimoto, T. (1999). Visualization of

microtubules in living cells of transgenic Arabidopsis thaliana. Proto-

plasma 206: 201–206.

Updegraff, D.M. (1969). Semi-micro determination of cellulose in bio-

logical materials. Anal. Biochem. 32: 420–424.

Voigt, B., Timmers, A.C., Samaj, J., Muller, J., Baluska, F., and

Menzel, D. (2005). GFP-FABD2 fusion construct allows in vivo visu-

alization of the dynamic actin cytoskeleton in all cells of Arabidopsis

seedlings. Eur. J. Cell Biol. 84: 595–608.

Westermann, S., Wang, H.W., Avila-Sakar, A., Drubin, D.G., Nogales,

E., and Barnes, G. (2006). The Dam1 kinetochore ring complex

moves processively on depolymerizing microtubule ends. Nature 440:

565–569.

Wightman, R., and Turner, S.R. (2008). The roles of the cytoskeleton

during cellulose deposition at the secondary cell wall. Plant J. 54:

794–805.

Zhong, R., Burk, D.H., Morrison III, W.H., and Ye, Z.H. (2002). A

kinesin-like protein is essential for oriented deposition of cellulose

microfibrils and cell wall strength. Plant Cell 14: 3101–3117.

1154 The Plant Cell

 This information is current as of January 11, 2010

 DOI: 10.1105/tpc.108.065334

2009;21;1141-1154; originally published online Apr 17, 2009; PLANT CELLSchumacher, Martine Gonneau, Herman Höfte and Samantha Vernhettes

Elizabeth Faris Crowell, Volker Bischoff, Thierry Desprez, Aurélia Rolland, York-Dieter Stierhof, Karin Arabidopsis

Pausing of Golgi Bodies on Microtubules Regulates Secretion of Cellulose Synthase Complexes in

Supplemental Data http://www.plantcell.org/cgi/content/full/tpc.108.065334/DC1

References http://www.plantcell.org/cgi/content/full/21/4/1141#BIBL

This article cites 57 articles, 27 of which you can access for free at:

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

eTOCs http://www.plantcell.org/subscriptions/etoc.shtml

at: THE PLANT CELLSign up for eTOCs for

CiteTrack Alerts http://www.plantcell.org/cgi/alerts/ctmain at: Plant CellSign up for CiteTrack Alerts for

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

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

ADVANCING THE SCIENCE OF PLANT BIOLOGY  

© American Society of Plant Biologists