Cellulose microfibril alignment recovers from DCB-induced disruption despite microtubule disorganization

Authors


For correspondence (fax +61 2 6125 4331; e-mail geoffw@rsbs.anu.edu.au).

Summary

Cellulose microfibril deposition patterns define the direction of plant cell expansion. To better understand how microfibril alignment is controlled, we examined microfibril orientation during cortical microtubule disruption using the temperature-sensitive mutant of Arabidopsis thaliana, mor1-1. In a previous study, it was shown that at restrictive temperature for mor1-1, cortical microtubules lose transverse orientation and cells lose growth anisotropy without any change in the parallel arrangement of cellulose microfibrils. In this study, we investigated whether a pre-existing template of well-ordered microfibrils or the presence of well-organized cortical microtubules was essential for the cell to resume deposition of parallel microfibrils. We first transiently disrupted the parallel order of microfibrils in mor1-1 using a brief treatment with the cellulose synthesis inhibitor 2,6-dichlorobenzonitrile (DCB). We then analysed the alignment of recently deposited cellulose microfibrils (by field emission scanning electron microscopy) as cellulose synthesis recovered and microtubules remained disrupted at the mor1-1 mutant's non-permissive culture temperature. Despite the disordered cortical microtubules and an initially randomized wall texture, new cellulose microfibrils were deposited with parallel, transverse orientation. These results show that transverse cellulose microfibril deposition requires neither accurately transverse cortical microtubules nor a pre-existing template of well-ordered microfibrils. We also demonstrated that DCB treatments reduced the ability of cortical microtubules to form transverse arrays, supporting a role for cellulose microfibrils in influencing cortical microtubule organization.

Introduction

One of the main questions challenging plant cell and developmental biologists is how plant cell shape and cell growth are controlled. Most plant cells grow anisotropically along one preferred axis. Anisotropic growth is an important feature of tissue differentiation, organ expansion and cell fate determination. Plant cell growth can be highly polarized in the case of tip-growing pollen tubes and root hairs, but most plant cell expansion is axial and diffuse, with new cell wall material added fairly evenly over the entire length of the cell (Wasteneys and Galway, 2003). It is clear that the cytoskeleton, especially the cortical microtubule component, plays an important role in regulating the direction of diffuse growth. De-polymerizing microtubules with drugs such as oryzalin reduces or eliminates anisotropic growth (Baskin et al., 1994). Comparable defects occur in Arabidopsis mutants that specifically disrupt cortical microtubule arrays, such as fass/ton (McClinton and Sung, 1997; Traas et al., 1995), mor1-1 (Sugimoto et al., 2003; Wasteneys, 2002; Whittington et al., 2001) and bot1/fra2/erh3/lue1 (Bichet et al., 2001; Bouquin et al., 2003; Burk et al., 2001; Webb et al., 2002).

A similar but generally more pronounced loss of growth anisotropy can be generated by inhibiting cellulose biosynthesis, either chemically or genetically. Herbicides including 2,6-dichlorobenzonitrile (DCB) and isoxaben reduce cellulose synthesis and generate rapid loss of growth anisotropy (Heim et al., 1991; Hogetsu et al., 1974; Sabba and Vaughn, 1999). The effects of DCB and isoxaben phenocopy cellulose-deficient mutants like rsw1-1 (Arioli et al., 1998; Baskin et al., 1992; Williamson et al., 2001), korrigan (Lane et al., 2001; Nicol et al., 1998; Sato et al., 2001) and kobito (Pagant et al., 2002).

To support anisotropic growth, it is essential that microtubules and cellulose microfibrils be oriented at right angles to the main axis of cell expansion, especially in rapidly elongating cells (Sugimoto et al., 2000; for reviews, see Baskin, 2001; Cyr, 1994; Wasteneys, 2000; Williamson, 1991). The cellulose synthase constraint hypothesis proposes that cortical microtubules control the direction of cellulose deposition by guiding the cellulose synthase enzyme complexes in the plasma membrane. This hypothesis (Giddings and Staehelin, 1991; Heath, 1974; Robinson and Quader, 1982) is mainly based on the following observations: (i) cortical microtubule and cellulose microfibril alignment are, in many cases, identical or similar (Bergfeld et al., 1988; Falconer and Seagull, 1985; Lang et al., 1982; Seagull, 1986; Vesk et al., 1996); (ii) microtubule-depolymerizing drugs have been shown to alter the orientation of cellulose microfibrils (Robinson and Quader, 1982); and (iii) during changes in growth, cortical microtubule re-orientation seems to precede re-alignment of cellulose microfibril deposition (Abe et al., 1995; Lloyd, 1984; Robinson and Quader, 1982).

This model, however, is incompatible with some experimental evidence (Emons et al., 1992; Preston, 1988). The alignment of cortical microtubules and cellulose microfibrils is not always similar, even in diffusely expanding cells (Sugimoto et al., 2000). To test the cellulose synthase constraint model, Sugimoto et al. (2003) analysed cellulose microfibril patterns in cells that had developed either with microtubules completely de-polymerized with the drug oryzalin, or disorganized in the mor1-1 mutant at its restrictive temperature. Cellulose microfibrils continued to be deposited in parallel order under these conditions (Sugimoto et al., 2003). This result suggested that organized cortical microtubules are not required for oriented deposition of microfibrils. Nevertheless, this study did not rule out the possibility that the deposition of microfibrils was simply maintained because of pre-existing well-ordered microfibrils, which once established, might serve as a template for the continued deposition of transverse, parallel microfibrils (Baskin, 2001).

In this study, we tested this possibility by specifically examining microfibril deposition during microtubule disruption in the mor1-1 mutant after first abolishing the transverse, well-aligned pattern of cellulose microfibrils with the drug DCB, which generates disordered microfibrils as it reduces the rate of cellulose synthesis (Satiat-Jeunemaitre, 1987; Sugimoto et al., 2001). Our results demonstrate that cellulose microfibrils can recover in well-ordered, transverse patterns without a pre-existing transverse microfibril template, lending further support to the idea that cellulose microfibril orientation is largely generated by self-assembly mechanisms that have little reliance on cortical microtubule organization or pre-existing microfibrils. We further observed that DCB treatment partially disorders cortical microtubules, supporting the possibility that microtubule orientation is influenced by cellulose microfibril organization.

Results

We devised an experimental strategy to test the ability of cells to generate parallel cellulose microfibril patterns in the absence of organized cortical microtubules and without a pre-existing microfibril template. This strategy combined the use of a cellulose synthesis-inhibiting drug DCB and the conditional microtubule disruption phenotype of the mor1-1 mutant. DCB is known to be a strong inhibitor of cellulose biosynthesis (Delmer, 1999; Sabba and Vaughn, 1999; and references therein). It has also been shown to randomize cell wall microfibril texture (Satiat-Jeunemaitre, 1987; Sugimoto et al., 2001), a phenomenon that we exploited in designing our experimental strategy. By allowing cellulose synthesis to recover from the DCB treatment while maintaining microtubule disruption, we were able to rigorously assess the requirement for organized microtubules to orient cellulose microfibril deposition and also determine if existing microfibrils are required to serve as a template for maintaining microfibril alignment. It has previously been determined that cellulose synthesis takes place at normal rates in mor1-1 at its restrictive temperature (Sugimoto et al., 2003), so recovery of cellulose synthesis after DCB removal should be the same as would occur in the wild type.

Figure 1 depicts, for a representative seedling, the full experimental procedure used in this study. Seedlings were first germinated and grown in Petri plates on agar-based nutrient medium for 5 days at the permissive temperature of 21°C for mor1-1, then they were transferred to otherwise identical Petri plates supplemented with 1 µm DCB (Figure 1a). After 4 h of DCB treatment (Figure 1b), the general shape of the root tip appeared unchanged, but root elongation rate and the length of the elongation zone were reduced, as indicated by the emergence of root hairs closer to the root tip apex (Figure 1b). Control treatments in which seedlings were transferred to fresh DCB-free plates showed, by comparison, only a slight reduction in root elongation rate (Figure 1f, filled squares).

Figure 1.

Root growth and morphology during the experimental procedure.

Sequential images of a representative mor1-1 root (a–e) are shown at the time points at which microtubule and microfibril orientation patterns were analysed (see Figures 2 and 3). The accompanying line diagram indicates the temperature, presence or absence of DCB and the time points at which the root was photographed. Stars and asterisks indicate fixed points at non-growing regions of the root. Dotted lines indicate the region of the root tip from which microtubule and cellulose microfibril images (Figure 2) were taken.

(a) After 5 days at 21°C.

(b) Following transfer and incubation for 4 h on a plate supplemented with 1 µm DCB at 21°C.

(c) After 2 h at 31°C on the 1 µm DCB-supplemented plate.

(d) After transfer to a DCB-free plate, but continuing at 31°C for 18 h.

(e) After a further 4 h without DCB but at 21°C.

DCB inhibits root elongation, induces root swelling and causes root hairs to emerge closer to the root tip (b,c). The restrictive temperature, 31°C, leads to the typical mor1-1 phenotype with root swelling and distorted root hairs (d,e). Scale bar = 500 µm for (a)–(e).

(f) mor1-1 root elongation rates were measured from 3 days (t = 0) after germination. All seedlings were transferred to fresh plates at day 5 (t = 48 h). Elongation rate on DCB-free plates was only slightly reduced (filled squares), whereas on 1 µm DCB-supplemented plates at 48 h, root elongation decreased drastically (filled triangles). Shifting the temperature to 31°C did not inhibit root elongation any further (open triangles). When DCB was again removed 6 h later, but plants were still kept at 31°C, elongation growth transiently recovered (open squares).

To stimulate the mor1-1 microtubule disruption phenotype, plates were transferred to a growth cabinet set at 31°C but with otherwise identical settings. After 2 h, the root tip showed noticeable radial swelling (Figure 1c) and greatly reduced elongation (Figure 1f, open triangles), an effect attributable mainly to the 6-h DCB treatment. At this point, seedlings were transferred to DCB-free plates, but were kept at 31°C for a period of 18 h. Elongation growth rates recovered significantly (Figure 1d,f, open squares), but were reduced relative to wild-type controls (see Figure S1). Root tips developed typical mor1-1 mutant features including distorted and short root hairs, reduced elongation rates and moderate radial swelling (Figure 1d). Unlike previous studies (Sugimoto et al., 2003; Whittington et al., 2001), left-handed root twisting and consequent skewing of the roots to the left side of the plate were not apparent. This was likely to be the result of both pre-treatment with DCB, which generates radial swelling without twisting, as well as the use of a slightly higher restrictive temperature of 31°C, which stimulates a more rapid onset of radial swelling than does 29°C. Finally, after 18 h, seedlings were transferred to the original growth cabinet set at 21°C for 4 h. This allowed cortical microtubules to recover, but it was too short a treatment to show any obvious recovery of root morphology (Figure 1e).

Figure S1.

At 31°C, wild-type roots elongate faster than mor1-1 roots after removing DCB.

Representative images of wild-type (a–c) and mor1-1 (d–e) roots from seedlings that were grown for 5 days at 21°C, followed by 4 h on 1 µm DCB at 21°C, followed by 2 h on 1 µm DCB at 31°C, followed by 18 h without DCB at 31°C (identical to Figure 1d). The diagram shows growth conditions and indicates at which stage the roots were photographed. Because of the 6-h DCB treatment, roots in both the wild type and mor1-1 have bulged and temporarily stopped elongating (marked by arrows). After removing DCB, restrictive temperature for mor1-1 (31°C) generates the typical mor1-1 phenotype with root swelling, inhibition of root/cell elongation and crooked root hairs (d–f). Wild-type roots, by comparison, have elongated more extensively after removing DCB (a–c). Scale bar = 500 µm for (a)–(f).

We also monitored root growth by marking the position of the root tip every 12 h (Figure 1f). The act of transferring mor1-1 seedlings from one plate to another caused only a slight reduction in elongation rate (filled squares), whereas root elongation was strongly reduced when seedlings where transferred to 1 µm DCB-containing plates (filled triangles). After 12 h on 1 µm DCB, the average elongation rate was reduced to 0.08 mm h−1 compared to 0.3 mm h−1 in the control. After 24 h, root elongation was almost completely abolished (Figure 1f, filled triangles). A similar effect was observed in the wild type (data not shown). These data are in close agreement with previous measurements. The I50 value for root growth inhibition by DCB in wild-type Arabidopsis roots is 0.4 µm (Heim et al., 1998). Shifting the temperature to 31°C does not reduce root elongation any further (Figure 1f, open triangles), suggesting that the 1-µm DCB treatment masks any radial swelling attributable to the mor1-1 phenotype. Removing DCB after 6 h, as carried out in the actual experiment, while keeping the plants at 31°C, restores elongation growth to some degree (Figure 1f, open squares), although, later on, the growth inhibition caused by the mor1-1 mutation at 31°C takes effect and growth rate drops again.

We used a method previously developed to correlate cortical microtubule and cellulose microfibril orientation in identically treated cells (Sugimoto et al., 2000). At all stages of treatment, several seedlings were fixed (while still at culture temperature) and were processed either for examining microtubules by immunofluorescence microscopy or for examining microfibrils by field emission scanning electron microscopy. Microtubule and microfibril patterns were observed (Figure 2) at developmentally equivalent positions in all cases in the upper half of the root elongation zone, where relative elemental growth rates were starting to decline (Figure 1a–e, dotted lines).

Figure 2.

Cellulose microfibrils can be deposited in a parallel order, even though cortical microtubules are disorganized.

mor1-1 seedlings were treated, as indicated in the accompanying time scale, fixed and prepared either for microtubule immunostaining (a,c,e,g,i), or for viewing cellulose microfibrils by field emission scanning electron microscopy (b,d,f,h,j). Representative images of epidermal cells in the upper half of the root elongation zone (700 µm from the root tip for (a) and (b) and 450 µm from the root tip for (c–j)) are shown (a–j). At the start of the experiment (5 days at 21°C), both microtubules (a) and cellulose microfibrils (b) are well aligned and transverse to the cell long axis. After 4 h on 1 µm DCB at 21°C, microtubules are still well organized (c), but cellulose microfibrils were very disordered (d). Additional transfer to 31°C for 2 h disorganized the microtubules (e), and microfibrils remain disordered (f). Eighteen hours after removing the DCB, but keeping the seedlings at 31°C, cellulose microfibrils regained their parallel alignment (h), although microtubules remained disorganized (g). Transferring back to 21°C restored the parallel microtubule order (i) and cellulose microfibrils were also well ordered (j). Double-headed arrows indicate the long axis of the cell. Scale bars are 50 µm for (a), (c), (e), (g) and (i) and 200 nm for (b), (d), (f), (h) and (j).

DCB rapidly disrupts cellulose microfibril alignment

Prior to DCB treatment, cortical microtubules (Figure 2a) and cellulose microfibrils (Figure 2b) both showed predominantly transverse orientation relative to the long axis of the root. After 4 h of treatment with 1 µm DCB, microtubule patterns were essentially unchanged (Figure 2c), while the wall texture changed dramatically (Figure 2d). Loss of parallel microfibril order and changes in the appearance of microfibrils were similar to what has previously been shown for wild-type seedlings (Sugimoto et al., 2001). Microfibrils either developed a lumpy and short texture (Figure 2d) or remained fairly long but lost parallel order (not shown). The disruptive effect of DCB on microfibril alignment was especially evident in cells that were very close (<200 µm) to the root tip (not shown). The same cells were in the elongation zone at the time of examination after 18 h at 31°C (not shown).

Cellulose microfibrils recover parallel order during cortical microtubule disruption

After 2 h at 31°C, cortical microtubules became disorganized (Figure 2e), while cellulose microfibrils continued to show the DCB-dependent disordered texture (Figure 2f). Transferring seedlings to DCB-free plates while keeping the seedlings at 31°C maintained microtubule disorganization (Figure 2g). The ensuing loss-of-growth anisotropy in these cells was attributable to the combined effects of the 6-h DCB pulse treatment and the 20-h treatment at the mor1-1 restrictive temperature. Despite these perturbations to microtubule organization and anisotropic expansion, cellulose microfibril patterns recovered parallel order, predominantly transverse to the root's long axis (Figure 2h). Wall texture was remarkably similar to that found in cells prior to treatment, with only slight deviation from a common orientation.

The reversible nature of the microtubule disruption in mor1-1 was demonstrated by returning seedlings to 21°C. After 4 h, cortical microtubules in the elongation zone were more abundant, well ordered and predominantly transverse (Figure 2i). Cellulose microfibrils continued to be deposited in transverse, parallel patterns (Figure 2j), although qualitatively, these were very similar to the microfibril patterns observed at 31°C. Cells in the second half of the elongation zone, 4 h after restoring the permissive temperature, had clearly defined long axes, demonstrating that elongation growth had been restored even to cells that had undergone some irreversible radial swelling in the early stages of expansion.

We repeated this experiment several times and consistently found that microfibrils were able to recover parallel alignment and predominantly transverse orientation after removing DCB, regardless of the microtubule disorganization. To quantify the orientation of microfibrils and microtubules, we determined their orientation relative to the cell's long axis from images taken 18 h after removing seedlings from the DCB-containing plates (e.g. Figure 2g,h). Images were from cells in the upper half of the elongation zone, at points 400–500 µm from the root tip apex. The angular distribution data show that, at this stage in the treatment, microtubule orientation is greatly dispersed about the cell's transverse axis, with only 31% of microtubules oriented within ±20 degrees of the transverse axis (Figure 3; 90°). In contrast, the majority (67%) of cellulose microfibrils at this stage in recovery from the DCB treatment were oriented within ±20 degrees of the transverse axis (Figure 3). These angular distribution patterns of microtubule and cellulose microfibrils are significantly different from one another according to an F-test (see Experimental procedures). The cellulose microfibril pattern is, however, somewhat less organized than in untreated mor1-1 at the permissive temperature; at 21°C, 86.5% of microfibrils were oriented within ±20 degrees of the transverse axis (data not shown), suggesting that the recovery of transverse microfibril orientation, while substantial, is not complete. From this experimental analysis, we conclude that cellulose deposition can recover parallel order and net transverse orientation when microtubules are disrupted. Furthermore, we conclude that oriented microfibril deposition does not depend on cues from the existing template of previously deposited microfibrils.

Figure 3.

After recovery from DCB, cellulose microfibrils are preferentially oriented transversely, although cortical microtubule orientation is greatly dispersed.

Orientation of microtubules (grey columns, n = 336) and microfibrils (black columns, n = 760) was measured relative to the cell long axis in epidermal cells of mor1-1 roots that were allowed to recover from DCB treatment for 18 h, but were still kept at 31°C (identical conditions as in Figure 2g,h). Note that microtubule orientation deviates much more than microfibril orientation. These distributions seem significantly different according to an F-test (see Experimental procedures).

Cellulose synthesis inhibition impairs microtubule alignment

Upon transfer of seedlings to DCB-free plates, we observed that microtubules in mor1-1 appeared to be less disrupted (compare Figure 2e with Figure 2g). To investigate this in closer detail, we compared microtubule patterns in mor1-1 seedlings with and without DCB at permissive and restrictive temperatures. As shown in Figure 4, microtubules appear slightly disorganized in mor1-1 at 21°C after 4 h of treatment with 1 µm DCB (Figure 4b,f). At 31°C, DCB treatment also clearly generates greater misalignment of microtubules than that generated by the mor1-1 phenotype alone (compare Figure 4c,g with Figure 4d,h). DCB treatment had a similar effect on cortical microtubule alignment in the wild type. Microtubule orientation in epidermal cells, 250–350 µm from the root tip, was measured to generate frequency distribution histograms, as shown in Figure 5. Compared to the microtubules in the DCB-free control treatment (black columns), microtubules in the presence of DCB deviated at a wider variety of angles from the transverse axis (grey columns). This difference in frequency distribution is statistically significant (see Experimental procedures). Taken together, these data suggest that the changes in the organization of the cell wall have some influence on the mechanisms that orient cortical microtubules.

Figure 4.

Microtubule organization in mor1-1 is even more disturbed in the presence of DCB.

Seedlings grown for 5 days at 21°C were treated as follows for 4 h: (a) 21°C, (b) 21°C in the presence of 1 µm DCB, (c) 31°C and (d) 31°C in the presence of 1 µm DCB. Frames indicate areas where higher magnification images are shown in (e)–(h). Immunolabelling of microtubules in mor1-1 roots shows that microtubule arrays are well ordered and transverse to the cell long axis at 21°C (a,e), slightly less well ordered after 4 h on 1 µm DCB (21°C; b,f), disorganized after 4 h at 31°C (c,g) and strongly disorganized after 4 h at 31°C in the presence of 1 µm DCB (d,h). Scale bar = 50 µm for (a)–(d) and 20 µm for (e)–(h).

Figure 5.

Cortical microtubules in wild-type roots are less transverse in the presence of DCB.

Microtubule orientation relative to the cell long axis in root epidermal cells (250–350 µm from the root tip) was measured in wild-type seedlings treated with 1 µm DCB for 4 h (grey columns) and in untreated seedlings (black columns). Microtubules in DCB-treated seedlings have a more dispersed orientation and are less transverse (grey columns) compared to the control (black columns). n = 830 for DCB treatment and 950 for control. An F-test determined that these distribution patterns are significantly different (see Experimental procedures).

Discussion

In this study, we devised a strategy that enabled us to rigorously test the idea that cellulose microfibril deposition in the absence of organized cortical microtubules can be maintained by co-alignment with previously deposited microfibrils. We made two general observations. First, despite disrupting the cellulose microfibril deposition pattern with DCB treatment, the cellulose synthesizing machinery can re-establish well ordered and predominantly transverse microfibril deposition during microtubule disruption. Second, abolishing transverse cellulose microfibrils in expanding cells by DCB treatment leads to increased deviation of cortical microtubules about the cell's transverse axis. These results support the concept that the alignment of cellulose microfibrils is under the control of a mechanism that depends largely on the rate of cellulose synthesis and not on the accurate transverse orientation of cortical microtubules. Moreover, transverse cortical microtubule orientation relies, to some extent, on the existence of well-ordered cellulose microfibrils.

Templated incorporation is not an essential part of microfibril alignment

Cortical microtubules clearly have a major influence on the mechanical properties of the cell wall. Our current results and those of a previous study (Sugimoto et al., 2003), however, cast considerable doubt on a specific role for cortical microtubules in the guidance of cellulose synthase complexes, and hence, in the oriented deposition of cellulose microfibrils.

Tobias Baskin recently summarized the literature on microtubule and cellulose microfibril co-alignment (Baskin, 2001). In this extensive review, he suggested that although there were many exceptions to the rule, the evidence was still good that microtubules influenced cellulose microfibril deposition. Baskin has put forward a model, called templated incorporation, to account for past experimental evidence that appeared to refute the cellulose synthase constraint model. According to Baskin's model, the information to align cellulose microfibrils comes from a scaffold that is either oriented by already incorporated microfibrils or by cortical microtubules. Consequently, cellulose microfibril orientation can be maintained even after microtubule de-polymerization, if the synthesizing complexes are able to use the previously deposited microfibrils as a template. Our experimental procedure specifically tested this mechanism by disrupting the template of well-ordered microfibrils and found no evidence for templated incorporation. Our data do, however, support the concept that cues from the cell wall can influence cortical microtubule stability and/or orientation at the plasma membrane.

Our experimental system provided a rigorous test of microtubule-dependent microfibril deposition. Nevertheless, the mor1-1 phenotype does not completely remove all microtubules, so it can be argued that the residual, heavily disrupted microtubules have at least some indirect influence on the cellulose-orienting mechanism. Complete microtubule disassembly can be achieved with drugs like oryzalin. Unfortunately, these drugs also de-polymerize microtubules in dividing cells, blocking mitosis and cytokinesis, leading to arrest of root elongation. The mor1-1 mutation conveniently disrupts cortical microtubules, while the pre-prophase band, spindle and phragmoplast arrays are much more resistant to de-polymerization and disorganization (Whittington et al., 2001). Cell production and root elongation thus continued over the course of time used in our current study.

Are microtubules necessary for cellulose alignment during very rapid phases of growth?

It may also be argued that microtubules are only essential for maintaining or fine-tuning cellulose microfibril orientation during the most rapid elongation phase and that our experimental procedure reduced elongation rates below this critical level. Sugimoto et al. (2000) determined that in the A. thaliana primary root, microtubule–microfibril co-alignment only occurs during the very brief period of early elongation, when the relative elemental elongation rate is accelerating. Such rapid elongation rates are not achieved at restrictive temperature for mor1-1, although radial swelling compensates for some of this loss. At these moderate elongation rates, cellulose deposition may be able to proceed without a well-organized cortical microtubule array. While this issue identifies a potential shortcoming of our experimental design, it remains clear that expanding cells can establish and maintain a microfibril alignment that is more accurately transverse than the microtubule array.

Self-ordering mechanisms – what evidence is there?

The establishment of microfibrils that are more accurately transverse than the cortical microtubules, as observed in our study, supports the existence of a self-ordering mechanism. Emons (1994) first proposed a model based on geometrical constraints to account for microfibril deposition in secondary walls in the non-growing regions of root hairs. In this model, the density of active synthases in the plasma membrane, the distance between individual microfibrils and the geometry of the cell are considered as factors influencing cellulose microfibril deposition (Emons and Kieft, 1994; Emons et al., 2002). The observation that reducing cellulose biosynthesis, either chemically (DCB) or genetically (Pagant et al., 2002; Sugimoto et al., 2001), causes microfibrils to lose parallel orientation and indicates the importance of a certain rate of cellulose production possibly related to synthase complex density. The possibility that microfibril alignment may be linked to the rate of cellulose synthesis has been discussed in detail by Sugimoto et al. (2001).

Does microtubule organization depend on cues from the cell wall?

We observed that in the presence of DCB, the degree of cortical microtubule transverse orientation was reduced, both in the wild type and in mor1-1. This suggests that cortical microtubule orientation is influenced, at least to some extent, by the organization of the cell wall. This idea of signalling from the cell wall to cortical microtubules is not new. Several earlier studies have suggested that cortical microtubule organization depends on signals from the cell wall, and a role of anisotropic wall stresses on microtubule orientation has been considered (Williamson, 1990). Enzymatic digestion of the cell wall during protoplast production disrupts cortical microtubule arrays (Hasezawa et al., 1988), and cell wall regeneration in protoplasts increases the resistance of cortical microtubules to chilling-induced de-polymerization (Akashi et al., 1990). Other experiments have shown that the inhibition of cellulose synthesis with the drug isoxaben prevents recovery of transverse microtubule orientation in tobacco culture cells (Fisher and Cyr, 1998) and leads to disorganization of microtubules in elongating pollen tubes (Lazzaro et al., 2003).

Experimental procedures

Plant material and growth conditions

The A. thaliana (ecotype Columbia) mor1-1 temperature-sensitive mutant (Whittington et al., 2001) was used throughout this study. Seeds were surface-sterilized in a mixture of 3% (v/v) hydrogen peroxide and 50% (v/v) ethanol for 2 min. After rinsing in sterilized water, seeds were planted on nutrient-solidified 1.2% (w/v) agar (Bacto Agar, Difco Laboratories, Detroit, MI, USA) plates (2 mm KNO3, 5 mm Ca(NO3)2, 2 mm MgSO4, 1 mm KH2PO4, 90 µm EDTA, 46 µm H3BO3, 9.2 µm MnCl2, 0.77 µm ZnSO4, 0.32 µm CuSO4, 0.11 µm MoO3, and 3% (w/v) sucrose). Plates were sealed with surgical tape (Micropore, 3M, USA) and held vertically in a growth cabinet under constant light (80 µmol m−2 sec−1) at 21°C. For temperature shifts, plates were transferred to a cabinet with similar light conditions but with a constant temperature of 31°C. For treatment with DCB, whole seedlings were transferred onto nutrient-solidified agar plates containing 1 µm DCB. DCB plates were always made up fresh from a 10 mm DCB stock solution in dimethylsulfoxide (DMSO). Control experiments demonstrated that plates containing 0.01% DMSO alone had no effect on growth or cortical microtubule organization.

For Figure 1, roots on agar plates were photographed with a digital camera (Leica DC 200) coupled with a dissecting microscope (Leica MZ FL III).

Root elongation measurements

Every 12 h, beginning 3 days after sowing, the position of the root tip apex was marked on the base of the plates with a razor blade under a stereomicroscope. After 7 days, images of the bottom surface of the plates were recorded in adobe photoshop 6.0 with a resolution of 600 pixels per inch using a flatbed scanner. Distances along the root between two successive marks were measured using the scion image software (available at http://www.scioncorp.com; Scion Corporation, MD, USA), and elongation rates (mm h−1) were plotted against time. For every treatment, at least 20 individual seedlings were measured.

Immunofluorescence microscopy

Immunofluorescent labelling of microtubules in whole roots was carried out as described by Sugimoto et al. (2000). Anti-α-tubulin antibody (clone B-5-1-2; Sigma-Aldrich, St Louis, MO, USA) was used at a dilution of 1 : 2000. Fluorescein–isothiocyanate (FITC)-conjugated antimouse IgG (Silenus/Chemicon, Boronia, Victoria, Australia), diluted to 1 : 100, was used as a secondary antibody.

Fluorescent images were collected with an MRC-600 confocal laser scanning microscope (Bio-Rad, Microscience Division, Hemel, Hempstead, UK) coupled to a Zeiss Axiovert IM-10 inverted microscope. The 488-nm line of an argon ion laser was used for FITC excitation. Images were collected (Kalman averaging six scans) using a 63× oil immersion objective lens. Each image shown represents a projection of approximately 20 individual images taken as a Z-series in 0.8-µm increments. Images were processed with image processing software, including comos 7.0 (Bio-Rad), confocal assistant (written by Todd Clark Brelje) and adobe photoshop 4.0.

Cell wall preparation

Seedling preparation for cell wall visualization was only slightly modified from the method described by Sugimoto et al. (2000). Whole seedlings were fixed in 4% (v/v) formaldehyde made up in Pipes-Mg-EGTA (PME) buffer (25 mm Pipes, 0.5 mm MgSO4, and 2.5 mm ethylene-glycol-bis (β-aminoethyl ether)N,N,N′,N′-tetraacetic acid (EGTA) (pH 7.2)) for 10 min, rinsed three times for 10 min in PME buffer and cryoprotected in 25% followed by 50% (v/v) DMSO in PME buffer (10 min each). Root tips were excised with a small pair of scissors, placed on a nail head and frozen in liquid nitrogen. To cut the epidermal cells open, the root surface was sliced off with a glass knife on a cryo-ultra-microtome (Ultracut E with FC 4 cryo-microtomy attachment; Reichert-Jung, Vienna, Austria) at −100°C. The remaining part of the root was thawed in 50% (v/v) DMSO in PME buffer and was then transferred into PME buffer. To remove cytoplasmic material and membranes, sectioned roots were treated with 0.1% sodium hypochlorite for 15 min. After thoroughly rinsing in distilled water, samples were osmicated in 0.5% (w/v) cold OsO4 for 15 min, followed by rinsing in distilled water. Specimens were then dehydrated in a graded ethanol series (30, 50, 70, 95 and three times 100%, each step 15 min) and CO2 critical point-dried (Union FL-9496, Balzers, BAL-TEC AG, Liechtenstein).

Field emission scanning electron microscopy

Specimens were mounted onto microscope stubs with double-sided sticky carbon tape, cut surface facing upward, and coated with platinum for 195 sec at 2.5 mA (Bio-Rad, Polaron Division, E 7400 Sputtering Module). Cellulose microfibril images were taken on a Hitachi 4500 field emission scanning electron microscope (Hitachi, Tokyo) at 3 kV, using the upper secondary electron detector. The aperture was set to 3, and the condenser lens was set to 10, with a working distance between 5 and 6. Scanned images were recorded with image slave 2.11 (Meeco, North Parramatta, Sydney, Australia).

Measurement and analysis of microtubule and cellulose alignment

Microtubule and cellulose microfibril orientation angles relative to the cell long axis were measured from digital images using the scion image software (available at http://www.scioncorp.com; Scion Corporation, MD, USA). For each treatment, orientation angles from at least three different roots were measured. For cellulose microfibrils, at least two different areas in each sampled cell were measured. If not mentioned otherwise, all images analysed were taken from root epidermal cells from the area between 400 and 500 µm away from the root tip. To determine whether the recorded distributions of microtubule or cellulose microfibril angles were significantly different for the different treatments, we calculated the ratio of the variances and applied an F-test as follows.

  • • Figure 5: The variance of microtubule angles is 1170.1 with and 408.6 without DCB treatment. The ratio of these variances is 2.86. Using an F-test, these data sets are significantly different from 1 at the 0.1% level. This means that the chance of getting a similar variance ratio (just by chance) is lower than 0.1%.
  • • Figure 3: The variances are 1630.5 for the microtubule orientation and 638.1 for the cellulose microfibril orientation. The ratio of these variances is 2.55. According to an F-test, this is significantly different from 1 at a similar level as for Figure 5. However, we feel that as the methods for measuring microfibrils and microtubules differed, these data sets are not strictly comparable in a statistical sense.

Conclusion

Our study demonstrates that parallel cellulose microfibril deposition can be re-established in the presence of disorganized cortical microtubules and without a template of well-ordered microfibrils. These findings are in conflict with both the cellulose synthase constraint (Giddings and Staehelin, 1991) and the templated incorporation (Baskin, 2001) hypotheses. On the other hand, our data strongly support the concept that cellulose microfibril orientation depends on the rate of cellulose synthesis and that, to some extent, microtubule orientation mechanisms depend on cues from the cell wall. We stress that microtubules clearly have a major function in regulating plant cell wall mechanical properties and morphogenesis (Sugimoto et al., 2003), even if a role in constraining the movement of cellulose synthase complexes now seems in doubt. Future efforts to understand the function of cortical microtubules in cell morphogenesis will benefit from this new information.

Acknowledgements

We thank Keiko Sugimoto for helpful advice. Research was, in part, supported by ARC Discovery Project Grant no. DP0208872 to G.O.W. We thank Cheng Huang and Sally Stowe from the ANU Electron Microscopy Unit for their assistance with field emission scanning electron microscopy, and Jeff Wood from the Statistical Consulting Unit at the ANU for help with statistical analysis.

Supplementary Material

The following material is available from http://www.blackwellpublishing.com/products/journals/suppmat/TPJ/TPJ1906/TPJ1906sm.htm

Representative images of wild-type (a–c) and mor1-1 (d–e) roots from seedlings that were grown for 5 days at 21°C, followed by 4 h on 1 µm DCB at 21°C, followed by 2 h on 1 µm DCB at 31°C, followed by 18 h without DCB at 31°C (identical to Figure 1d). The diagram shows growth conditions and indicates at which stage the roots were photographed. Because of the 6-h DCB treatment, roots in both the wild type and mor1-1 have bulged and temporarily stopped elongating (marked by arrows). After removing DCB, restrictive temperature for mor1-1 (31°C) generates the typical mor1-1 phenotype with root swelling, inhibition of root/cell elongation and crooked root hairs (d–f). Wild-type roots, by comparison, have elongated more extensively after removing DCB (a–c). Scale bar = 500 µm for (a)–(f).

Ancillary