Patterning and Lifetime of Plasma Membrane-Localized Cellulose Synthase Is Dependent on Actin Organization in Arabidopsis Interphase Cells^,

The rate of insertion and lifetime of cellulose-synthesizing complexes at the plasma membrane is dependent on the organization of the actin cytoskeleton.

Abstract

The actin and microtubule cytoskeletons regulate cell shape across phyla, from bacteria to metazoans. In organisms with cell walls, the wall acts as a primary constraint of shape, and generation of specific cell shape depends on cytoskeletal organization for wall deposition and/or cell expansion. In higher plants, cortical microtubules help to organize cell wall construction by positioning the delivery of cellulose synthase (CesA) complexes and guiding their trajectories to orient newly synthesized cellulose microfibrils. The actin cytoskeleton is required for normal distribution of CesAs to the plasma membrane, but more specific roles for actin in cell wall assembly and organization remain largely elusive. We show that the actin cytoskeleton functions to regulate the CesA delivery rate to, and lifetime of CesAs at, the plasma membrane, which affects cellulose production. Furthermore, quantitative image analyses revealed that actin organization affects CesA tracking behavior at the plasma membrane and that small CesA compartments were associated with the actin cytoskeleton. By contrast, localized insertion of CesAs adjacent to cortical microtubules was not affected by the actin organization. Hence, both actin and microtubule cytoskeletons play important roles in regulating CesA trafficking, cellulose deposition, and organization of cell wall biogenesis.

Plant cells are surrounded by a flexible yet durable extracellular matrix that makes up the cell wall. This structure offers mechanical strength that counters osmotically driven turgor pressure, is an important factor for water movement in plants, acts as a physical barrier against pathogens (Somerville et al., 2004), and is a determining factor for plant cell morphogenesis. Hence, the cell wall plays a central role in plant biology.

Two main types of cell walls can typically be distinguished: the primary and the secondary cell wall. The major load-bearing component in both of these cell walls is the β-1,4-linked glucan polymer cellulose (Somerville et al., 2004). Cellulose polymers are synthesized by plasma membrane (PM)-localized cellulose synthase (CesA) complexes (Mueller and Brown, 1980), which contain several CesA subunits with similar amino acid sequences (Mutwil et al., 2008a). The primary wall CesA complexes are believed to be assembled in the Golgi and are subsequently delivered to the PM via vesicular trafficking (Gutierrez et al., 2009), sometimes associated with Golgi pausing (Crowell et al., 2009). Furthermore, the primary wall CesA complexes are preferentially inserted into the PM at sites that coincide with cortical microtubules (MTs), which subsequently guide cellulose microfibril deposition (Gutierrez et al., 2009). Hence, the cortical MT array is a determinant for multiple aspects of primary wall cellulose production.

The actin cytoskeleton plays a crucial role in organized deposition of cell wall polymers in many cell types, including cellulose-related polymers and pectins in tip-growing cells, such as pollen tubes and root hairs (Hu et al., 2003; Chen et al., 2007). Thus, actin-depolymerizing drugs and genetic manipulation of ACTIN genes impair directed expansion of tip-growing cells and long-distance transport of Golgi bodies with vesicles to growing regions (Ketelaar et al., 2003; Szymanski, 2005). In diffusely growing cells in roots and hypocotyls, loss of anisotropic growth has also been observed in response to mutations to vegetative ACTIN genes and to actin-depolymerizing and -stabilizing drugs (Baluska et al., 2001; Kandasamy et al., 2009). While actin is clearly important for cell wall assembly, it is less clear what precise roles it plays.

One well-known function of actin in higher plants is to support intracellular movement of cytoplasmic organelles via actomyosin-based motility (Geisler et al., 2008; Szymanski, 2009). During primary wall synthesis in interphase cells, treatment with the actin assembly inhibitor latrunculin B (LatB) led to inhibition of Golgi motility and pronounced inhomogenities in CesA density at the PM (Crowell et al., 2009; Gutierrez et al., 2009) that coincided with the density of underlying and immobile Golgi bodies (Gutierrez et al., 2009). These results suggested that Golgi motility is important for CesA distribution (Gutierrez et al., 2009). The actin cytoskeleton also appears to be important for secondary wall cellulose microfibril deposition. For example, longitudinal actin filaments (AFs) define the movement of secondary wall CesA-containing Golgi bodies in developing xylem vessels (Wightman and Turner, 2008). In addition, it has been proposed that the AFs also can regulate the delivery of the secondary wall CesA complex to the PM via pausing of the Golgi (Wightman and Turner, 2008). It is therefore clear that actin organization is important for CesA distribution and for the pattern of cellulose microfibril deposition.

Despite the above findings, very few reports have undertaken detailed studies to elucidate the role of the actin cytoskeleton in the distribution and trafficking of specific proteins in plant cells. Here, we have investigated the intracellular trafficking of CesA-containing vesicles and delivery of CesAs to the PM, in the context of the actin cytoskeleton. We quantitatively demonstrate that the organization of the actin cytoskeleton regulates CesA-containing Golgi distribution and the exocytic and endocytic rate of the CesAs. However, actin organization has no effect on the localized insertion of CesAs at sites of MTs at the PM.

RESULTS

Motility of CesA-Containing Golgi Bodies Is Facilitated by the Actin Cytoskeleton

CesA-containing Golgi motility depends on the organization of the actin cytoskeleton (Crowell et al., 2009; Gutierrez et al., 2009). We confirmed this using a yellow fluorescent protein (YFP):CesA6-expressing line (Paredez et al., 2006) in which we monitored interphase epidermis cells in 3-d-old etiolated hypocotyls exposed to LatB (1 µm). In agreement with previous reports, we observed aggregation of the Golgi in response to LatB (Fig. 1; Supplemental Fig. S1, A and B). While these effects are most likely caused by impairment of the actin cytoskeleton, our aim was to further characterize the role of the actin cytoskeleton in cellular CesA distribution. In addition to actin-disrupting drugs, we therefore also took a genetic approach using mutant analyses. The Arabidopsis (Arabidopsis thaliana) genome contains 10 ACTIN genes, of which three, ACT2, ACT7, and ACT8, are expressed in vegetative tissues, two are potential pseudogenes, and the rest, ACT1, ACT3, ACT4, ACT11, and ACT12, are expressed in reproductive tissues (Supplemental Fig. S2, A and B; Gilliland et al., 2002). ACT2, ACT7, and ACT8 are assumed to be functionally redundant, and single mutants only result in weak phenotypes (Kandasamy et al., 2009). In agreement with these assumptions, act2act7 displayed additive phenotypes, including severe dwarfism and organ twisting (Gilliland et al., 2002; Kandasamy et al., 2009). Using live-cell imaging of GFP fused to the F-actin-binding domain (GFP:FABD) lines, we confirmed that the act2act7 mutant held disorganized AFs, with fewer AF bundles and fragmented cortical AFs with more free or broken ends (Fig. 1, A and B; Supplemental Movie S1).

Figure 1.

Defects in actin organization cause reduced Golgi motility and accumulation of post-Golgi vesicles. A and B, Actin organization in elongating hypocotyl cells of 3-d-old etiolated control (A) and act2 act7 mutant (B) seedlings expressing GFP:FABD. Red arrowheads indicate thin transverse AFs that typically appear shorter in the mutant compared with control; yellow arrowheads indicate free GFP. Bars = 5 μm. C to E, Transmission electron micrographs of control seedlings (C), seedlings grown on 100 nm LatB (D), and act2 act7 mutant seedlings (E). Red arrowheads indicate Golgi; green arrowheads indicate regions with dense populations of post-Golgi aggregates in seedlings grown on 100 nm LatB and in the act2 act7 mutant. Bars = 500 nm.

To monitor CesA-containing Golgi in the act2act7 mutant, we made reciprocal crosses between the mutant and YFP:CesA6-expressing plants. Similar to the LatB-treated seedlings, we observed some cells showing aggregation of Golgi (Supplemental Fig. S1, A–C). Transmission electron microscopy (TEM) supported these observations, showing regions with high Golgi density as well as high densities of vesicles adjacent to the Golgi (Fig. 1, C–E), presumably representing post-Golgi vesicles, in the act2act7 mutant and LatB-treated seedlings. Furthermore, to quantify Golgi movement in the different conditions, we classified static Golgi as those that remained paused for periods up to 30 s. In the wild-type control, only 3% of the Golgi (n = 216 Golgi, three cells) were classified as static, whereas 98% (n = 81 Golgi, three cells) and 21% (n = 109 Golgi, three cells) were classified as static in the LatB-treated and act2act7 mutant seedlings, respectively (Supplemental Fig. S1, A–G). These results indicate that the organization of the actin cytoskeleton is important for CesA-containing Golgi distribution in Arabidopsis interphase cells and that the Golgi motility is only partially affected in the act2act7 mutant.

Defects in the Organization of the Actin Cytoskeleton Result in Cellulose Deficiencies

The reduced motility of the CesA-containing Golgi in the actin-deficient cells suggested the possibility that the deposition of cellulose might be affected. Several cellulose-deficient mutants, including procuste1 (prc1-1, affecting CesA6) and cobra (required for cellulose deposition), display isotropic root cell growth and ectopic lignin deposits (Desprez et al., 2002). Close examination of the act2act7 mutants revealed clear root cell swelling in light-grown seedlings, and both mutant and LatB-treated seedling roots and hypocotyls accumulated ectopic lignin, which frequently accompanies diminished cellulose deposition (Caño-Delgado et al., 2003), as assessed by phloroglucinol staining (Jensen, 1962; Fig. 2, A–F; Supplemental Fig. S1, I and J). To examine further the impact on cellulose production by actin impairment, we measured the cellulose content in 5-d-old etiolated act2act7 mutant seedlings using the Updegraff method (Updegraff, 1969) and in seedlings grown on low levels of LatB (100 nm). Both the mutant and the LatB-treated seedlings contained significantly less cellulose per unit dry mass compared with control seedlings (Fig. 2G). In contrast with the LatB-treated seedlings, seedlings grown on media containing the MT-destabilizing drug oryzalin (500 nm) did not show reduced levels of cellulose (Fig. 2G). One possible reason for the reduction in cellulose content in the actin-deficient seedlings could arise from slower movement of CesA complexes at the PM. To investigate this, we measured the velocity of CesA particles in the actin double mutant and LatB-treated seedling; however, there was no significant difference observed between control and the treatment or mutant (P > 0.05, Student’s t test; n = 123, 126, and 108 CesA particles for wild-type, act2act7, and LatB three seedlings, respectively; Supplemental Fig. S1K).

Figure 2.

Actin deficiency mimics cellulose-related mutants and leads to aberrant cell wall deposition. A to C, Root phenotypes of 5-d-old light-grown control (A), prc1-1 (affecting CesA6; B), and act2 act7 mutant (C) seedlings. Inserts display magnified regions of the seedling roots. Bars = 100 μm. D to F, Phloroglucinol-HCl staining of control (D), prc1-1 (E), and act2 act7 mutant (F) seedlings grown as described in A to C. Bars = 100 μm. G, Cellulose levels in etiolated 5-d-old seedlings (genotypes and drugs as indicated in figure). Error bars indicate sd, and asterisk shows significance (P < 0.05, Student’s t test). Calcofluor-stained root sections of 5-d-old wild-type (H) and act2 act7 mutant (I) seedlings. Arrowheads indicate regions of uneven cell wall deposition. Bar = 50 μm. J to L, Transmission electron micrographs of control seedlings (J), act2 act7 mutant seedlings (K), and seedlings treated with 100 nm LatB (L). Uneven cell walls are indicated with red arrowheads. CW, Cell wall. Bars = 500 nm. M, Variability in cell wall thickness is indicated by the mean sd of the wall thickness arbitrary unit; error bars represent se. Asterisk shows significance (P < 0.0001, Student’ t test; n = 70–80 cells per genotype).

Several cellulose-deficient mutants, including prc1-1, also develop incomplete cell walls (Fagard et al., 2000). To assess whether this also is the case in the actin-impaired seedlings, we took two different approaches. First, we analyzed time series projections of YFP:CesA6 at the PM in etiolated hypocotyl cells in the wild type (Supplemental Fig. S3A) and act2act7 mutants (Supplemental Fig. S3B). Similar to what was reported in LatB-treated cells, we observed uneven distribution of CesA complexes at the PM (Supplemental Fig. S3B; Supplemental Movie S2), a phenotype correlated with and explained by uneven distribution of Golgi (Gutierrez et al., 2009). Second, we stained transverse sections of seedling roots of the act2act7 double mutant with calcofluor, which stains cellulose, callose, and other simple β-linked glucans (Maeda and Ishida, 1967). Consistent with the uneven CesA distribution at the PM and the reduced cellulose levels in the act2act7 mutants, the cell wall material appeared unevenly distributed in the epidermal, cortical, and endodermal cells of the mutant seedling roots when compared with the wild type (Fig. 2, H and I). To corroborate this observation, we analyzed TEM sections of act2act7 mutant and LatB-treated seedling roots. These analyses confirmed that both the mutant and LatB-treated seedlings had cell walls with varying thickness compared with wild-type sections (Fig. 2, J to M). These results show that the uneven distribution of Golgi and CesA complexes in the PM observed during impairment of actin are accompanied by defects in cellulose deposition and cell wall thickness distribution. While the uneven cell wall distribution could be due to an uneven distribution of the CesAs at the PM, it is also plausible that these irregularities are caused by defects in the distribution of other cell wall components.

Motility of Small CesA-Containing Compartments Is Facilitated Both by the Actin and MT Cytoskeleton

In addition to Golgi, CesA is also observed in small motile CesA compartments (SmaCCs; Supplemental Fig. S4A; Crowell et al., 2009; Gutierrez et al., 2009). A heterogeneous collection of compartments (Gutierrez et al., 2009), SmaCCs associate with cortical MTs under conditions of osmotic stress or inhibition of cellulose synthesis and can be translocated by tracking depolymerizing MT ends in a polarity-independent fashion (Gutierrez et al., 2009). When associated with MTs, these compartments are also termed MT-associated CesA compartments (MASCs; Crowell et al., 2009). Considering that some SmaCC compartments are observed associated with Golgi (Gutierrez et al., 2009) and that Golgi have been observed to be translocated along the actin cytoskeleton (Akkerman et al., 2011), we investigated whether SmaCCs also may be associated with and perhaps be translocated along the actin cytoskeleton. To investigate the relationship between SmaCCs and the actin cytoskeleton, we generated a dual-labeled GFP:FABD- and tdTomato red fluorescent protein (tdTomatoRFP):CesA6-expressing Arabidopsis line. Importantly, the phenotypes of the majority of the progeny from these crosses were indistinguishable from wild-type plants (Supplemental Fig. S3D), indicating that the fluorescently labeled proteins did not interfere with plant growth to any major extent. We then monitored epidermal interphase cells in 3-d-old etiolated seedlings at focal planes ranging from the cortex to approximately 0.6 to 1 µm below the PM. At the latter subcortical focal planes (at 0.8 µm below the PM focal plane), we observed evidence for migration along AFs by some SmaCCs for periods lasting from 2 s to more than 30 s at a density of 2.4 ± 0.55 particles per 100 μm² under control conditions (Fig. 3A; Supplemental Fig. S4, F–H; Supplemental Movies S3 and S4). Unlike the MT-associated SmaCCs/MASCs that display alternating saltatory and smooth and steady movement with an average velocity of 4.06 µm min^–1 (n = 122 SmaCCs; Supplemental Fig. S4, B–D), the subcortical SmaCCs/MASCs displayed highly erratic movement with an average velocity of 59.27 µm min^–1 (n = 150 smaCCs; Fig. 3, B and D). The measurement of the subcortical SmaCCs/MASCs was done in conditions where seedlings were treated with isoxaben (100 nm) and oryzalin (20 µm), which results in removal of cortical SmaCCs/MASCs, thereby ensuring that the analyzed compartments are subcortical. These velocities are in close agreement with previously published velocities of particles that track along AFs (Boevink et al., 1998). Furthermore, treatment with LatB abolished the movement of the subcortical CesA vesicles (Fig. 3C). Removal of cortical SmaCCs resulted in a significant increase in the number of subcortical CesA compartments (Fig. 3E), suggesting either that the MTs and actin cytoskeleton may compete for SmaCC interaction or that SmaCCs might be handed off from the actin to the MT cytoskeleton. These results support the existence of CesA-containing vesicle compartments that interact with both MTs and the actin cytoskeleton.

Figure 3.

Small subcortical CesA compartment movement is dependent on actin cytoskeleton A, Focal planes below the cell cortex displayed erratic and rapid movement of SmaCCs along the actin cytoskeleton in 3-d-old etiolated dual-labeled GFP:FABD and tdTomatoRFP:CesA6 hypocotyl cells. Green enclosures show rapid movement of the subcorticular CesA compartment along an AF (red arrowhead). B, Kymographs of the subcorticular CesA compartment moving along the actin cytoskeleton in A. C, LatB treatment abolishes movement of the subcortical CesA compartments. Bars = 5 μm. D, Histogram showing the velocity of subcorticular CesA compartments (n = five seedlings, 134 particles). E, Differences in density of CesA compartments at focal plane below the cortex after different drug treatments (n = five seedlings for each treatment). Error bars represent se. Asterisks show significant difference (P < 0.05) between the wild type and the treatments.

Stabilization of Cortical AFs Influences the Linear Movement of CesA Particles in the PM

Because cortical MTs guide PM-localized CesA particles (Paredez et al., 2006) and AFs coalign with cortical MTs for short periods of time (Sampathkumar et al., 2011), it is plausible that the PM-tracking CesAs and cortical AFs can come into contact and that the AFs could exert an influence on CesA tracking behavior at the PM.

To first investigate whether cortical AFs can cooccur with tracks of PM-located CesA particles, we imaged cells in etiolated dual-labeled GFP:FABD and tdTomatoRFP:CesA6 seedlings. Interestingly, in several instances per cell, we observed coalignment between cortical AFs and linear tracks of CesA particles at the PM in time average images (Supplemental Fig. S3E; Supplemental Movie S5). The cooccurrence of the AFs and the CesA trajectories typically lasted from several seconds up to more than 40 s (Supplemental Fig. S3E; Supplemental Movie S5). Time series overlays and kymograph analyses confirmed that the PM-localized CesAs that were analyzed followed linear trajectories and moved at constant velocity, being unaffected during instances of actin colocalization (Supplemental Fig. S3, F and G). However, the alignment of cortical AFs and CesA tracks only took place at few regions at any given point of time in the cell.

To investigate the potential functional relationship of the observed alignment of cortical AFs and CesA tracks, we exposed GFP:CesA3-expressing seedlings to jasplakinolide (5 µm for 3 h), an actin-stabilizing drug, and then observed the behavior of the CesAs. The CesA particles were distributed throughout the PM, similar to mock-treated cells, but when we assembled time average images to trace CesA migration, we observed disorganized trajectories and impaired Golgi movement (Supplemental Fig. S3C; Supplemental Movie S6). Considering that actin stabilization also influences the MT organization (Sampathkumar et al., 2011), we hypothesize that the observed changes in CesA trajectories are a consequence of the changes in MT organization facilitated by the stabilized actin cytoskeleton.

The Bulk of CesA Delivery to the PM in Hypocotyl Cells Are Not Associated with Paused Golgi Bodies

Given that the actin cytoskeleton facilitates the distribution of CesA-containing Golgi bodies and that it is regarded as an important structure for exocytosis and endocytosis in many organisms, including tip-growing cells in higher plants (Moscatelli et al., 2012), we aimed to explore how defects in the actin cytoskeleton affect CesA delivery rates to the PM. The delivery of CesA particles to the PM has been described to derive from pausing Golgi bodies at the cortex (Crowell et al., 2009), a conclusion derived from seven observations in cells at the base of the hypocotyl. This technique is, however, tedious due to the identification of new CesA insertion events in an already existing pool of CesA particles. In observations of CesA delivery in hypocotyl cells (Gutierrez et al., 2009; Bringmann et al., 2012), Golgi bodies were occasionally observed at the location and time of CesA delivery, but frequently no Golgi body was apparent. To put these observations on a quantitative basis, we analyzed CesA delivery events in hypocotyl cells using a photobleaching assay and asked how frequently organelles consistent with Golgi bodies were present when newly arrived tdTomato:CesA6 was first detected at a stable position in the plane of the PM (Fig. 4A; Supplemental Movie S7). We counted 60 delivery events following photobleaching (n = three cells from three seedlings) and observed associated Golgi in 38 (63%) cases. Of these events, 18 (30% of all observed events) occurred following sustained Golgi pausing (>8 s of immobility). No Golgi bodies were observed in the vicinity at the time of CesA delivery in 22 (36%) events. Thus, it is clear that while pausing Golgi bodies precedes some insertion events, the bulk of delivery events in hypocotyl cells either involve rapidly moving Golgi or vesicles no longer associated with Golgi (Supplemental Fig. S5A; Supplemental Movie S7).

Figure 4.

Localized insertion of CesA particles in the PM is independent of the actin cytoskeleton. A and B, Different types of CesA insertion mechanisms. A, CesA insertion can occur without appearance of Golgi body in the frames prior to insertion. Right section shows kymograph along the cyan line. B, CesA insertion can occur from Golgi body that moves rapidly. Right section shows kymograph along the cyan line. Red arrowheads indicate Golgi body; yellow arrowheads indicate CesA particle at the PM. C, Single image frame showing CesA insertions (white arrowheads) on MTs in a GFP:CesA3 TUA5:mCherryRFP dual-labeled hypocotyl cell from LatB-treated 3-d-old etiolated seedlings. D, Map of CesA delivery events in LatB-treated seedlings (C). Sites were mapped onto a 6-min average projection of MT signal. Red plus symbols indicate CesA insertion sites. E, TEM sections of 5-d-old act2 act7 double mutant in which vesicular compartments (yellow arrowheads) can be seen in close proximity to cortical MT sites (red arrows). CW, Cell wall. Bars = 500 nm. F, Histogram showing quantification of average distance between vesicles and the nearest MT in the wild type and act2 act7 double mutant. Error bars represent se (P > 0.05, Student’s t test; n = 161 vesicles for the wild type and n = 273 for act2 act7).

MT-Localized Insertion of CesA Particles into the PM Is Independent of the Actin Cytoskeleton

In addition to its role in guiding CesA particles at the PM, the cortical MTs also define preferred insertion sites of CesAs into the PM (Gutierrez et al., 2009). The brief coordinated actions between MTs and the actin cytoskeleton (Sampathkumar et al., 2011), and the observation that AFs can coalign with CesA trajectories, prompted us to ask whether the positional information for CesA insertion is affected when the actin cytoskeleton is impaired. To test this possibility, we treated seedlings expressing GFP:CesA3 and mCherry fused to a tubulin marker (mCherry:TUA5; Gutierrez et al., 2009) with 1 μm LatB for a period of 4 h and photobleached the GFP channel at the PM. We subsequently analyzed the positions of CesA delivery to the PM with respect to cortical MTs as previously described (Gutierrez et al., 2009; Bringmann et al., 2012). Fifty delivery events were observed in the LatB-treated seedlings, of which 38 (76%) occurred in locations defined by MT label (44% of the image area, P < 0.001, binomial test; Fig. 4, C and D; Supplemental Movie S8). Surprisingly, these results are very similar to those reported for untreated seedlings (Gutierrez et al., 2009; Bringmann et al., 2012). Furthermore, observation at higher resolution using TEM showed electron-dense vesicles at the cortex adjacent to MTs in the act2act7 mutant seedlings (Fig. 4, E and F). We did not observe any significant differences in the average distance between cortical vesicles to the closest MT between the Columbia ecotype and act2act7 (Fig. 4F; cortical vesicles 600 nm or less below the PM were included in analysis). In this analysis, we distinguished between clathrin-coated vesicle, dense vesicle, and other secretory vesicle (Supplemental Fig. S4A); however, in no cases did we observe any differences in MT associations. We conclude that the organization of the actin cytoskeleton does not influence the positional insertion of CesAs next to MTs.

The Organization of the Actin Cytoskeleton Influences Exocytic Rates of CesAs

Next, we explored if and how the actin cytoskeleton influences the rate of CesA delivery to the PM. This rate can be determined in cells containing GFP-labeled CesA by photobleaching a region of the PM containing active CesA complexes and measuring the number of newly delivered CesA particles per unit area per unit time (Gutierrez et al., 2009). We initially performed this procedure in cells with and without disrupted AFs to determine if normal actin organization is necessary for efficient delivery of complexes to the PM. However, this technique has a notable complication. Cytoplasmic GFP:CesA protein, mostly residing in Golgi bodies near the PM, are susceptible to bleaching; therefore, some CesA cargo destined for the PM would be invisible and could potentially cause the rate of CesA delivery to be underestimated. This source of error is likely greater in cells with a compromised actin cytoskeleton because cytoplasmic streaming is severely inhibited. With robust streaming in normal cells, the local cytoplasmic pool in the bleach region is rapidly replaced with unbleached CesA protein. To address this issue in actin-compromised cells (act2act7 double mutants and LatB-treated seedlings), we performed two types of measurements. First, we photobleached relatively small regions of 36 µm² to minimize bleaching of the cytoplasmic GFP:CesA pool (Supplemental Fig. S5B) and measured recovery of optically resolved GFP:CesA punctae that met the criteria of active CesA complexes (linear motility at slow and steady velocity). Only punctae within a central 28-µm² subregion were measured to avoid false positives due to active CesA complexes migrating into the region of observation (Supplemental Fig. S5B). We observed an initial lag phase in CesA recovery, but after 2 min, we observed a gradual and linear increase in CesA recovery in all genotypes and treatments (Supplemental Fig. S5, C and D). If bleached CesAs contribute significantly to the inserted pool of CesAs in the actin deficient cells and the insertion rate is constant, a nonlinear pattern of recovery would be expected. This is because bleached CesAs would initially contribute more to the insertions, whereas newly synthesized CesAs would subsequently become more prevalent. We therefore concentrated on the linear portion of the recovery curves for further analysis. It should be noted that there was not a complete recovery of the CesA fluorescence during the observed time frame, including for the wild type (Supplemental Fig. S5C), indicating bleaching of the sample by imaging during recovery or the existence of a pool of signal that recovers on a longer time frame. The rate of recovery was reduced by nearly 50% in the act2act7 mutants (4.24 ± 1.59 punctae μm^–2 h^–1; n = six cells) and LatB-treated seedlings (5.05 ± 1.29 punctae μm^–2 h^–1; n = six cells) compared with control seedlings (11.2 ± 2.57 punctae μm^–2 h^–1; n = six cells; Fig. 5A; Supplemental S5C). These results might be obtained either by a decrease in the delivery rate of new label to the PM or by an increase in the rate of their removal. To isolate the delivery rate, individual CesA deliveries were measured at three time intervals during the liner phase of recovery using the criteria described by Gutierrez et al. (2009; Fig. 5B). During recovery, label density at the PM was low enough to permit this more precise analysis, whereas density at the end of recovery was too high to assay these events with confidence. We observed that new deliveries were, in fact, significantly higher in the wild type than in either act2act7 mutants or in LatB-treated cells (Fig. 5C). Furthermore, the consistent delivery rates measured at each of the three time intervals provided further evidence that the measurements were not significantly influenced by delivery of bleached complexes at this stage of recovery.

Figure 5.

The actin cytoskeleton is important, but not essential, for CesA vesicle trafficking to the PM. A, Box plot showing rates of recovery of CesA punctae in wild-type, act2 act7 double mutant, and LatB-treated seedlings. Asterisks show significant difference (P < 0.0003, Student’s t test) between wild-type and actin-compromised conditions (n = six cells for each condition; total area of 170 μm for each condition). B, Frames representing delivery of CesA particle. Arrowheads mark appearance of CesA particle and steady movement and lower section shows kymograph of typical CesA insertion event following FRAP. Bars = 1 μm. C, Histogram showing the number of CesA delivery events after photobleaching in wild-type, act2 act7, and LatB-treated seedlings during the indicated time frames. CesA insertions were defined as in Gutierrez et al. (2009). Asterisks show significant difference (P < 0.047, Student’s t test). D, Average lifetime of a CesA complex in wild-type, act2 act7, and LatB-treated seedlings (n = six cells for each condition). Error bars represent sd, and asterisks show significant difference (P < 0.005, Student’s t test) between the wild type and the mutant/treatment.

Because the reduced delivery rate could be a consequence of fewer underlying Golgi in the actin-compromised cells, we measured the density of the labeled CesAs at the cortex in frames that just preceded photobleaching in the observed regions. If, in fact, the observed regions in the actin-compromised cells were devoid of Golgi within the range that vesicles can be efficiently delivered to the PM, then we might expect to see that the starting density of CesAs was lower in the actin-compromised regions. No significant difference (P = 0.1895, Student’s t test; P = 0.25, Mann-Whitney U test) in mean CesA particle density was observed between the mutants (1.09 ± 0.09 particles μm^–2; n = five cells), LatB-treated seedlings (1.005 ± 0.19 particles μm^–2; n = five cells), and control seedlings (1.34 ± 0.08 particles μm^–2; n = five cells; Supplemental Fig. S5C) in regions that were used for fluorescence recovery after photobleaching (FRAP) analyses. Thus, there was no evidence that the regions chosen for analysis were depleted of CesAs in the actin-compromised cells prior to photobleaching.

The observation that the densities of CesAs at the PM were similar, together with our observation that CesA delivery was significantly lower in actin-compromised cells, has interesting implications for both the internalization of CesA complexes and the lifetime of CesA complexes at the PM. If we assume that the observed cells were at a steady state, i.e. delivery rates are equal to internalization rates, than maintenance of the same density at the PM implies that the internalization rate declines along with the delivery rate in actin-compromised cells. Likewise, the lifetimes of the CesAs at the membrane must also be longer. We calculated the lifetime of a typical CesA particle in act2act7 double mutants to be approximately 15.42 min (1.09 particles μm^–2/4.24 ± 1.59 events μm^–2 h^–1) and in the LatB-treated seedlings to be approximately 10.5 min (1.005 particles μm^–2/5.05 ± 1.29 events μm^–2 h^–1), whereas the lifetime of a typical CesA in the control cells was approximately 7.15 min (1.34 particles μm^–2/11.2 ± 2.57 events μm^–2 h^–1; Fig. 5D).

If CesA internalization is reduced in actin-compromised cells, then other measures of membrane and protein internalization might also be affected. To test if this might be true, we used the endocytic marker FM4-64 and stained light-grown hypocotyl cells of wild-type, act2act7 double mutant, and LatB-treated seedlings. After a period of 10 min, we quantified the amount of FM4-64-localized vesicles in the cells and found that there was a significant reduction in the number of FM4-64-marked endosomes in the act2act7 double mutant (4.48 ± 1.217 vesicles per 100 µm²; P = 0.04907, Student’s t test) and LatB-treated (2.99 ± 0.83 vesicles per 100 µm²; P = 0.00021, Student’s t test) cells compared with the wild-type cells (5.94 ± 1.47 vesicles per 100 µm²; Supplemental Fig. S5, E–H). Thus, the actin organization is important for efficient exocytic and endocytic exchange at the PM and therefore has an influence on the lifetime of a typical CesA particle in the PM.

DISCUSSION

Actin is a major component of the cytoskeleton and is involved in vesicle trafficking and cargo distribution in eukaryotic cells (Mooren et al., 2012). We show that while the actin cytoskeleton is not essential for exocytosis and endocytosis in interphase plant cells, an intact actin organization influences the rate of CesA delivery and the time a typical CesA spends in the PM. In addition, the actin cytoskeleton does not seem to impact specific positional insertion of the CesAs into the PM.

In plant cells, the bulk of research regarding the role of the actin cytoskeleton in membrane exchange at the PM comes from studies on tip-growing cells. For example, treatment of pollen tubes and root hairs with actin-depolymerizing drugs inhibit growth and cytoplasmic streaming (Miller et al., 1999; Vidali et al., 2001; Ketelaar et al., 2003). The importance of the actin organization for tip-growing cells is therefore well established. However, with the notable exceptions of early endocytic sterol trafficking (Grebe et al., 2003) and cellular CesA distribution (Crowell et al., 2009; Gutierrez et al., 2009), the role of the actin cytoskeleton in interphase cells is less well explored. In addition, only very few studies have substantiated the observations with robust quantitative analyses for specific PM components.

The AFs can roughly be divided in two populations. One is characterized by highly dynamic fine cortical AFs, and the other displays less dynamic and thicker AF bundles at subcortical focal planes (Staiger et al., 2009; Sampathkumar et al., 2011). While the precise functions of these populations are not clarified, a recent study demonstrated differential Golgi motility along the two populations (Akkerman et al., 2011). Golgi bodies associated with fine AFs at the cell cortex displayed slow nondirectional movement, perhaps engaged in exocytic activities, and subcortical Golgi bodies associated with thicker actin bundles displayed unidirectional movement that may signify long-distance transport. These data explain the subcortical CesA-containing Golgi aggregates that accumulated in response to actin impairment. Moreover, closer observation of Golgi in actin-compromised cells using TEM revealed hyperaccumulation of vesicles adjacent to Golgi, a pattern consistent with continued vesicular biogenesis from the Golgi but a lack of transport away from Golgi due to defects in actin-mediated mobility.

Association of post-Golgi CesA-containing compartments, i.e. SmaCCs/MASCs, was previously observed at cortical MTs in stressed cells (Crowell et al., 2009; Gutierrez et al., 2009), where they undergo tip-tracking (Gutierrez et al., 2009) motility driven by MT depolymerization (Crowell et al., 2009; Gutierrez et al., 2009). MT association and tip-tracking events were also observed in unstressed cells for brief durations (Gutierrez et al., 2009), and delivery of CesAs was found to be strongly associated with the positions of cortical MTs. In this study, subcortical imaging revealed that a population of these vesicles also migrates along AFs and cables. While we do not conclude that this migration occurs through a direct association between the smaCCs and the AFs, it is important to note that the movement of the subcortical smaCCs was substantially reduced when the actin cytoskeleton was impaired. In addition, when actin was disassembled with inhibitors, SmaCCs showed increased association with MTs, an outcome that might indicate interruption of an exchange of SmaCCs between the actin and MT cytoskeletons or perhaps stimulation of SmaCC interaction with MTs, similarly as observed by imposition of osmotic stress (Crowell et al., 2009; Gutierrez et al., 2009).

At least a subset of SmaCCs/MASCs appears to be involved in exocytosis of the CesA complexes. Gutierrez et al. (2009) showed that the compartments could deliver CesA complexes to the PM both in growing cells and after washout of mannitol in osmotically challenged cells. Furthermore, only a small population of the compartments colocalized with the lipophilic tracer FM4-64 after short-term treatment (Gutierrez et al., 2009), a result confirmed by Sánchez-Rodríguez et al. (2012), who also observed only a small overlap in FM4-64 distribution and a vesicle population containing the chitinase-like protein, which largely colocalized with the SmaCCs/MASCs. Taken together, these observations suggested that MTs position CesA insertion at a local scale, perhaps by interacting directly with secretory compartments.

If SmaCCs/MASCs are involved in exocytosis, and considering that disruption of the actin cytoskeleton leads to a larger population of MT-associated compartments, it is not difficult to envision that the AFs may not influence the localized insertion of the CesAs into the PM. Hence, the actin cytoskeleton would promote cellular distribution of the compartments, whereas the cortical MT-associated compartments would signify preparation for CesA delivery to the PM. Impairment of the actin cytoskeleton would therefore not disturb the preferential insertion of CesAs adjacent to the MTs. Recent reports also show that actin depolymerization did not affect targeted accumulation of several polar markers in the endodermis (Alassimone et al., 2010). By scoring the sources of CesA delivery, we found that most CesAs were delivered in regions where no Golgi bodies were observed prior to insertion. We hypothesize that the source of these deliveries are through Golgi-derived vesicles distributed by streaming in both the cortical and subcortical planes.

By quantifying the CesA density and delivery rate at the PM, we estimated the average lifetime of a CesA particle in the PM. Assuming that the density of the CesA particles is at steady state in both control and LatB-treated seedlings, we found that the lifetime of a typical CesA is around 7 to 8 min and that the lifetime is longer in actin-defective cells. These data are in agreement with earlier estimates using amino acid pulse-chase experiments, coupled with protein synthesis inhibitors, in cotton (Gossypium hirsutum; Jacob-Wilk et al., 2006). If our assumption about steady state is correct, then the endocytic rate in the actin-deficient cells would also be reduced. In mammals and yeast (Saccharomyces cerevisiae), it is well established that the actin cytoskeleton is important for endocytic activities (Engqvist-Goldstein and Drubin, 2003). However, this relationship is less clear in plant cells. Studies on endocytosis of the auxin efflux carrier PIN1 showed that the removal of PIN1 from the membrane was unaffected in actin-compromised root cells (Boutté et al., 2006). By contrast, recycling of sterols and certain cell wall components is perturbed upon actin depolymerization in roots (Baluska et al., 2002; Grebe et al., 2003). Furthermore, live-cell imaging studies of GFP-tagged endocytic markers in Arabidopsis roots showed that endocytic vesicles could form in actin-impaired cells (Konopka et al., 2008). However, the movement of these vesicles away from the site of internalization was most clearly affected, and the majority of the GFP-tagged endocytic markers remained immobile for long periods of time upon treatment with either LatB or with 2,3-butanedione monoxime, which inhibits myosin activity (Konopka et al., 2008). Our observation using the lipophilic tracer FM4-64 in interphase cells of hypocotyls suggests that endocytosis is affected in actin-perturbed cells. While the nature of this perturbance remains to be investigated, this could contribute to the increased lifetime of the CesA particles at the PM that we estimated in the actin-compromised cells.

The actin-deficient seedlings contained reduced levels of cellulose and displayed uneven wall thickness in transverse sections. While these two measures are not directly connected, i.e. many other cell wall components contribute to cell wall thickness, it is likely that the impaired actin cytoskeleton delivers material unevenly to the apoplast. We also report that CesAs in these seedlings stay at the PM for a longer time compared with the wild type. Arguably, if the CesAs stay at the PM for a longer time and the velocities are similar (and thus the presumed rates of synthesis), then each complex should produce more cellulose, which in the end would result in similar levels of cellulose as in the wild type. An obvious question is then, why do we see lower levels of cellulose? One likely reason could be that the actin-deficient cells on average contain lower levels of PM-based CesA molecules. While we ensured that the regions, and even cells, we used for our FRAP analyses contained similar CesA densities, it is important to note that this may not be the case for all the cells in the seedling. Hence, a lower overall number of active CesAs in many of the actin-deficient cells would also lead to reduced cellulose levels. Another reason could be that the CesA complexes perhaps are active for a certain amount of time. In such a scenario, the CesAs would move at a constant speed for a certain distance and then perhaps become deactivated and wait for internalization. The delayed endocytosis observed in the actin-deficient cells could cause CesAs to be retained in the PM after deactivation. Perhaps such a deactivation could involve posttranslational modifications, such as phosphorylation. The CesA complexes can be phosphorylated, and changes in phosphorylation can cause aberrant behavior of the CesAs (Chen et al., 2010; Bischoff et al., 2011). Although the distribution of CesA lifetimes at the PM remain to be determined, it seems reasonable to propose that CesA internalization is regulated, both to avoid internalization before cellulose synthesis has been accomplished and as a means to control microfibril length. However, we did not observe any significant increase in the number of stalled CesAs in the actin-deficient cells when compared with wild-type cells. While this does not completely rule out that a certain fraction of CesAs are moving slower, we favor the first scenario in which the actin-impaired cells contain lower levels of active CesAs over the whole of the cell.

The actin cytoskeleton plays a crucial role in regulating cell shape, from bacteria to animals to higher plants (Huang et al., 2012; White and Gober, 2012). In bacteria, the actin-like protein MreB appears to a play role that is somewhat related to that of the MT cytoskeleton in high plants; both form cortical arrays that play a role in positioning cell wall biosynthetic enzymes and patterning the spatial organization of the cell wall (Huang et al., 2012; White and Gober, 2012). Here, we found that the actin cytoskeleton in Arabidopsis also regulates aspects of cell wall synthesis by regulating the spatial distribution of the CesAs at a cell-wide level and by governing CesA lifetime in the PM. Hence, whereas the actin cytoskeleton directly controls cell shape in animal cells, the bacterial and plant actin-related cytoskeleton appears to indirectly manage cell shape by controlling cell wall deposition.

MATERIALS AND METHODS

Plant Materials and Growth Conditions

Arabidopsis (Arabidopsis thaliana) Columbia ecotype seedlings were grown as described in Sampathkumar et al. (2011). Homozygous act2-1 act 7-1 (Gilliland et al., 2002) double mutant seeds were kindly provided by Richard Meagher (University of Georgia), and prc1-1 (Fagard et al., 2000) was a kind gift from Herman Höfte. The mCherry:TUA5 (Gutierrez et al., 2009), FABD:GFP (Ketelaar et al., 2004), CesA6:YFP (Paredez et al., 2006), and CesA3:GFP (Desprez et al., 2007) lines have been described as indicated. All plants used for the imaging are F2 segregating lines and seedlings were selected based on phenotype and presence of either one or both fluorescent markers.

Constructs

To generate the tdTomato::CesA6 construct, the Gateway cassette was amplified from pMDC32 (Curtis and Grossniklaus, 2003) with primers 5′-ACCGGTTATCAAACAAGTTTGTACAAAAAAGC-3′ and 5′-ACCGGTAACCACTTTGTACAAGAAAGCTGA-3′, containing flanking AgeI sites, and cloned into pGEM-T Easy (Promega). The resulting plasmid was cut with AgeI and ligated into the compatible SgrAI site of CesA6-pBluescript SK+ (Paredez et al., 2006), which is located at the N terminus of CesA6. The tdTomato gene was amplified from pRSET-B tdTomato (Shaner et al., 2004) with primers 5′-GGGGACAAGTTTGTACAAAAAAGCAGGCTTCATGGTGAGCAAGGGCGAGGA-3′ and 5′-GGGGACCACTTTGTACAAGAAAGCTGGGTCCTTGTACAGCTCGTCCATCCGT-3′ and cloned into CesA6-pBluescript SK+ via BP and LR reactions. The resulting plasmid was cut with SacI and XmaI and ligated into pCAMBIA2300. The tdTomato::CesA6 construct was then introduced into prc1-1 plants by Agrobacterium tumefaciens-mediated transformation as described previously (Paredez et al., 2006).

Drug Treatments

For biochemical analysis, TEM, and lignin staining, seedlings were germinated and grown on Murashige and Skoog media plates supplemented with inhibitors, i.e. LatB (100 nm), oryzalin (500 nm), and isoxaben (2 nm). For short-term treatment, seedlings were incubated in 2 mL of water containing inhibitors (1 µm LatB, 100 nm isoxaben, or 20 µm oryzalin) in 12-well cell culture plates in darkness. The seedlings were subsequently mounted for confocal imaging as described in Sampathkumar et al. (2011). Stock solutions of LatB, isoxaben, and oryzalin were dissolved in methanol or in dimethyl sulfoxide, and working stocks were made freshly by further dilution in water. FM4-64 was dissolved in water and used at a concentration of 20 µm.

TEM

Five-day-old seedlings were high-pressure frozen, freeze substituted, embedded, sectioned, and viewed according to McFarlane et al. (2008). Briefly, samples were cryofixed using a Leica HPM 100 in B-type sample holders (Ted Pella) using 1-hexadecene (Sigma) as a cryoprotectant. Samples were freeze substituted in 2% (w/v) osmium tetroxide (Electron Microscopy Sciences) and 8% (v/v) 2,2-dimethoxypropane (Sigma) in acetone for 5 d at –78°C, then slowly warmed to room temperature over 2 d and infiltrated with Spurr’s resin (Reynolds, 1963) over 4 d. Samples were sectioned to approximately 70 nm using a Leica UCT microtome, suspended on copper grids (Gilder) coated with 0.3% (w/v) formvar (Electron Microscopy Sciences), stained with 2% (w/v) uranyl acetate in 70% (v/v) methanol and Reynolds’ lead citrate (Spurr, 1969), and then viewed with a Hitachi H7600 equipped with an Advanced Microscopy Techniques Advantage CCD camera (Hamamatsu ORCA) at an accelerating voltage of 80 kV.

Cellulose Measurement

Cellulose was measured from 5-d-old etiolated seedlings. Fresh plant material was harvested in 2-mL Eppendorf tubes and homogenized with a Retsch mill. The powdered plant material was washed two times with 70% (v/v) aqueous ethanol and centrifuged. The supernatant was discarded, and after air drying the pellet, 1 mL of methanol:chloroform (1:1, v/v) was added and sonicated for 10 min and spun down. The supernatant was discarded, and the remaining pellet was washed twice with acetone and air dried for 1 h to obtain cell wall material. The crystalline cellulose content was determined using a modified protocol of Updegraff (1969) described in Neumetzler et al. (2012). Data were collected from four technical replicates for each treatment and mutant lines. Statistical significance between the wild type and other treatments was determined using Student’s t test.

Lignin Staining

To analyze lignification pattern, Arabidopsis wild-type, drug-treated, and mutant 5-d-old seedlings were harvested under sterile conditions. Seedlings were placed onto a microscope slide and incubated with two drops of freshly prepared phloroglucinol-HCl solution (25 mg phloroglucinol, 25 mL 100% [v/v] methanol, and 25 mL 37% [v/v] HCL) for 5 min before observation under a light microscope.

Calcofluor Staining

Sectioning and staining of roots with calcofluor was performed according to Fagard et al. (2000).

Microscopy

Light microscopy was performed using a Leica Stereomicroscope (Leica MZ12.5, Leica DFC420 digital camera). Seedlings expressing GFP:CesA3, YFP:CesA6, GFP:FABD, and all the dual-labeled lines of FABD:GFP, tdTomato:CesA6 and GFP:CesA, mCherry:TUA5 were imaged on confocal microscope equipped with a Yokogawa spinning disc head fitted to a Nikon Ti-E inverted microscope as described in Sampathkumar et al. (2011). Photobleaching was achieved using a FRAP/photoactivation system (Roper Scientific) as described in Gutierrez et al. (2009). Images of seedlings stained with FM4-64 was acquired on a Zeiss LSM780, fluorescence was excited using a 514-nm laser, and emission recoded through a 603- to 651-nm filter.

Image Analysis and Processing

All images were processed and analyzed using the ImageJ software. Background correction was performed using the “Subtract Background” tool (rolling ball radius, 30–40 pixels), and StackReg was used to correct focal drift. Kymograph analysis for measurement of CesA and SmaCC velocity was also done using ImageJ as described in Gutierrez et al. (2009).

Measurement of Cell Wall Thickness

Length of three random points on the wall was measured using ImageJ, and average sd was calculated within these three values as a measure of variation in cell wall thickness. Seventy-five to 80 cell walls each for the different conditions and genotypes were measured. Arbitrary units were used as a measurement for the relative cell wall thickness within one image to account for differences in sectioning planes. Significance was estimated using Student’s t test.

Density Measurement of CesAs at the PM

First frame of a time series data of the PM focal plane was used for quantifying number of CesAs. Regions of cell that contained an even distribution of CesAs at the PM were chosen to avoid complexities arising due to differences in underlying Golgi numbers in the mutant and LatB-treated cells when compared with the wild type. A total area of 432 µm² was analyzed for the wild-type, act2act7 mutant, and LatB-treated seedlings. The total density was calculated as the number of particles per 100 µm² based on 190 CesA particles in control, 186 in act2act7 mutant, and 173 in LatB-treated seedlings (n = six cells in six seedlings for each condition). Significance was estimated using Student’s t test.

Density Measurement of smaCCs

SmaCCs were identified as described in Gutierrez et al. (2009) at different focal planes. Cortical smaCCs were scored based on motility of the particles described previously. Subcortical smaCCs were identified based on focal plane (0.6 to 1 µm below the PM) and on the erratic movement of the particles assessed in kymographs. Particle density was calculated as number of particles per 100 µm² (total area of 500 µm²; n = five cells in five seedlings for all conditions). Significance was estimated using Student’s t test.

FM4-64 Staining and Density Measurements

Five-day-old light-grown wild type and act2act7 mutants were incubated for 10 min in 20 µm FM4-64. Epidermal hypocotyl cells were imaged under a Zeiss LSM780 confocal microscope as indicated above. The density of internalized FM4-64 was determined by counting the number of FM4-64-labeled vesicles in control (total area of 14,011 µm²; n = 25 cells), LatB-treated (total area of 19,770 µm²; n = 34 cells), and act2act7 (total area of 16,072 µm²; n = 23 cells) seedlings. In total, 2,089 FM4-64-labeled particles were counted. Total density was calculated as number of particles per 100 µm². Significance was estimated using Student’s t test.

Delivery Rate Analysis of CesAs

The PM focal plane of wild-type, act2act7 mutant, and LatB-treated seedlings expressing CesA6:GFP were photobleached using the 491-nm laser. After bleach, the new CesA insertions events in the PM plane were identified in control (total area, 171 µm²; 151 insertion events), LatB (total area, 171 µm²; 97 insertion events), and act2act7 (total area, 171 µm²; 101 insertion events). We scored the insertion events based on the following criteria. First, prior to CesA delivery, the immediate region surrounding the delivery should be devoid of any fluorescent particles, and second, subsequent frames were examined using kymographs for characteristic CesA tracking behavior. The images were analyzed for a new insertion for periods of 10 to 13 min after bleaching. Five cells in five different seedlings for control and six cells from six different seedlings for act2act7 and LatB-treated seedlings were analyzed. Significance was estimated using Student’s t test.

Lifetime of CesA Complex

The lifetime of CesA complexes was estimated by dividing CesA particle density by the rate of CesA insertion.

Supplemental Data

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

• Supplemental Figure S1. Actin defects affect Golgi motility and cellulose deposition.

• Supplemental Figure S2. Expression patterns of ACTIN genes in Arabidopsis.

• Supplemental Figure S3. AFs and bundles can align with CesA trajectories at the PM and affect the distribution of CesA molecules.

• Supplemental Figure S4. Movement of MT- and AF-associated SmaCCs.

• Supplemental Figure S5. Global view of CesA insertion patterns.

• Supplemental Movie S1. AF dynamics in the wild type and act2act7 mutant.

• Supplemental Movie S2. CesA distribution is affected in act2act7 mutant.

• Supplemental Movie S3. Tracking of SmaCCs/MASCs along AFs.

• Supplemental Movie S4. Tracking of SmaCCs/MASCs along AFs.

• Supplemental Movie S5. AFs dynamically coalign with CesA trajectories.

• Supplemental Movie S6. Stabilization of AF with jasplakinolide affects CesA trajectories.

• Supplemental Movie S7. Photobleaching reveals bulk of the CesA delivery does not precede sustained pausing of Golgi bodies.

• Supplemental Movie S8. MT-localized delivery of CesA occurs independent of the actin cytoskeleton.

Glossary

CesA

cellulose synthase

PM

plasma membrane

MT

microtubule

AF

actin filament

LatB

latrunculin B

TEM

transmission electron microscopy

SmaCC

small motile cellulose synthase compartment

MASC

microtubule-associated cellulose synthase compartment

FRAP

fluorescence recovery after photobleaching

YFP

yellow fluorescent protein

tdTomatoRFP

tdTomato red fluorescent protein

FRAP

fluorescence recovery after photobleaching