Hypersensitivity to cytoskeletal antagonists demonstrates microtubule–microfilament cross-talk in the control of root elongation in Arabidopsis thaliana


  • David A. Collings,

    1. Plant Cell Biology Group, Research School of Biological Sciences, Australian National University, GPO Box 475, Canberra, ACT 2601, Australia;
    Search for more papers by this author
  • Adrian W. Lill,

    1. Plant Cell Biology Group, Research School of Biological Sciences, Australian National University, GPO Box 475, Canberra, ACT 2601, Australia;
    Search for more papers by this author
  • Regina Himmelspach,

    1. Plant Cell Biology Group, Research School of Biological Sciences, Australian National University, GPO Box 475, Canberra, ACT 2601, Australia;
    Search for more papers by this author
  • Geoffrey O. Wasteneys

    1. Plant Cell Biology Group, Research School of Biological Sciences, Australian National University, GPO Box 475, Canberra, ACT 2601, Australia;
    2. Department of Botany, University of British Columbia, Vancouver, BC, Canada V6T 1Z4
    Search for more papers by this author

Author for correspondence:David A. Collings Tel: +61 2 6125 3952 Fax: +61 2 6125 4331 Email: collings@rsbs.anu.edu.au


  • • Elongation of diffusely expanding plant cells is thought to be mainly under the control of cortical microtubules. Drug treatments that disrupt actin microfilaments, however, can reduce elongation and induce radial swelling. To understand how microfilaments assist growth anisotropy, we explored their functional interactions with microtubules by measuring how microtubule disruption affects the sensitivity of cells to microfilament-targeted drugs.
  • • We assessed the sensitivity to actin-targeted drugs by measuring the lengths and diameters of expanding roots and by analysing microtubule and microfilament patterns in the temperature-sensitive Arabidopsis thaliana mutant microtubule organization 1 (mor1-1), along with other mutants that constitutively alter microtubule arrays.
  • • At the restrictive temperature of mor1-1, root expansion was hypersensitive to the microfilament-disrupting drugs latrunculin B and cytochalasin D, while immunofluorescence microscopy showed that low doses of latrunculin B exacerbated microtubule disruption. Root expansion studies also showed that the botero and spiral1 mutants were hypersensitive to latrunculin B.
  • • Hypersensitivity to actin-targeted drugs is a direct consequence of altered microtubule polymer status, demonstrating that cross-talk between microfilaments and microtubules is critical for regulating anisotropic cell expansion.


The plant root is ideal for investigating growth and morphogenesis. Transverse orientation of cortical microtubules is essential for controlling the anisotropic properties of the expanding cell wall (Wasteneys, 2002; 2004; Wasteneys & Collings, 2004). In contrast, the function of microfilaments in root elongation is less clear. Microfilament-disrupting drugs reduce cell and overall elongation in onion (Thomas et al., 1973), Arabidopsis (Baskin & Bivens, 1995) and maize roots (Baluška et al., 2000; Blancaflor, 2000). Microfilament disruption also causes the elongating root tip to swell, a process that involves an increase in lateral growth (Thomas et al., 1973; Baskin & Bivens, 1995; Blancaflor, 2000). This suggests that, like microtubules, microfilaments play a role in regulating anisotropic expansion.

How do microfilaments contribute to anisotropic cell expansion? Microfilaments are essential for basic processes associated with overall cell growth, including cytoplasmic streaming and the delivery of Golgi to sites of active exocytosis (Boevink et al., 1998; Nebenführ et al., 1999; Wasteneys & Galway, 2003). In cells with rapid exocytosis, such as root cap peripheral cells, the microfilament-disrupting drugs cytochalasin B and D cause a marked build up of secretory vesicles (Vaughan & Vaughn, 1987). This explains how microfilament disruption might reduce cell elongation (Baskin & Bivens, 1995), but not why it should cause root expansion or swelling. Another possibility is that the loss of microfilament function interferes with the normal function of microtubules, a process that would require interactions between microfilaments and microtubules. Various reports indicate that elongating plant cells contain transverse cortical microfilaments which lie parallel to transverse cortical microtubules (Collings & Allen, 2000), although observation of these microfilaments in Arabidopsis roots is difficult and requires pretreatment with the cross-linking drug, MBS, which may induce artefactual microfilament polymerization (Collings & Wasteneys, 2005). Drug studies demonstrate that microtubules and microfilaments can interact with each other under specific conditions, although not whether the interactions are direct or indirect (Collings & Allen, 2000).

Recent genetic manipulations in A. thaliana support a role for microfilaments in controlled cell expansion in roots. The act2-2D mutant has shorter microfilament bundles in root epidermal cells and shows reduced cell elongation and defects in the radial expansion of trichoblasts (Nishimura et al., 2003). Mutations affecting the ACT7 gene cause reduced root growth, increased root twisting and aberrant cell divisions (Gilliland et al., 2003). Over- and under-expression of actin-binding proteins, such as profilin, actin-depolymerization factor and cyclase-associated protein (CAP), affect cell elongation in roots (Ramachandran et al., 2000; Dong et al., 2001; Barrero et al., 2002).

Regulation of the microfilament assembly in the diffusely expanding cells of the root remains largely unexplored. By comparison, in some cell types that show variations from stereotypical anisotropic growth, such as leaf trichomes, leaf epidermal pavement cells and root hairs, there is evidence that the actin related protein 2/3 (ARP2/3) complex functions in the nucleation and turnover of microfilaments. In these cell types, distorted class mutants that lack the ARP2/3 complex show aberrant cellular morphogenesis (Hülskamp et al., 1994; Mathur et al., 1999; Szymanski et al., 1999; Le et al., 2003; Mathur et al., 2003). However, while ARP2/3 complex proteins are expressed throughout the plant, including the root (Li et al., 2003), distorted class mutants, and more recently identified knockout mutants of ARP2/3 complex proteins, have no reported effects on root growth or morphogenesis (Li et al., 2003; Mathur et al., 2003). Other pathways must exist to regulate microfilament polymerization in the Arabidopsis root.

Recent investigations have tested the sensitivity of Arabidopsis mutants with impaired microtubule arrays to drugs that target the microtubule cytoskeleton. The lefty1 and lefty2 mutants have identical point mutations in the α-tubulin 6 (TUA6) and α-tubulin 4 (TUA4) genes, respectively (Thitamadee et al., 2002; Abe et al., 2004), and are hypersensitive to microtubule disruption, with the inhibition of root elongation and the induction of root swelling occurring at lower concentrations of microtubule-disrupting drugs (Thitamadee et al., 2002). Comparable studies with the spiral1 mutant, in which a microtubule-associated protein is absent, also show hypersensitivity to microtubule-disrupting drugs (Furutani et al., 2000).

In this study, we sought to investigate how interactions between actin microfilaments and microtubules modulate cell expansion in the Arabidopsis root by measuring the effects of several cytoskeleton-targeted drugs on the temperature-sensitive microtubule organization 1 (mor1-1) mutant (Whittington et al., 2001). Our analysis of root growth and cytoskeletal organization demonstrate that the mor1-1 mutant shows hypersensitivity not only to microtubule-disrupting drugs, such as oryzalin and taxol, but also to the microfilament-disrupting drugs, latrunculin B and cytochalasin D. Our findings are discussed in terms of potential cross-talk between microtubules and microfilaments in the control of anisotropic cell expansion.

Materials and Methods

Plant growth

Wild-type A. thaliana (L.) Heynh ecotype Columbia and numerous mutant lines were used in this study. Seed was surface sterilized [90 s, 50% (v/v) ethanol : 3% (v/v) hydrogen peroxide], washed extensively with distilled water and plated onto 1.2% (w/v) agar plates in Hoagland's solution that was modified with 3% (w/v) sucrose, 0.53 mm inositol and 50 µm thiamine (Baskin & Wilson, 1997). Plates were sealed with Nescofilm and stored at 4°C for several days to synchronize germination. Plants were grown vertically at 21°C for 5 d under continuous light (80–100 µE m−2 s−1).

Inhibitor studies

We prepared stock solutions of drugs in dimethylsulfoxide (DMSO), and stored them at −20°C. Stocks were 20 mm oryzalin (Lilley Research Laboratories, Greenfield, IN, USA), 10 mm taxol (Molecular Probes, Eugene, OR, USA), 2 mm latrunculin B (MP Biomedicals, Seven Hills, NSW, Australia), 33 mm propyzamide (a gift from Dr Kevin Vaughn, USDA Stoneville, MS, USA) and 20 mm cytochalasin D (Sigma, St Louis, MO, USA). For experiments, varying concentrations of different drugs were added to 50 ml of liquid agar (described under ‘Plant growth’) and the mixture adjusted to 0.2% (v/v) DMSO. Two replicate 90 mm-diameter plates were poured for treatments at 21° and 29°C, 5-d-old wild-type and mutant seedlings were transferred to these plates, and the plates were sealed with Nescofilm. Plates were then incubated at either 21° or 29°C for 48 h. Root diameters following treatments were determined using the ‘measure’ tool in Adobe Photoshop (version 6.0) (Adobe, San Jose, CA, USA) from photographs taken with a digital camera (model DC200; Leica, Wetzlar, Germany) mounted on a dissecting microscope (model MZ FLIII; Leica). Root elongation was determined from digital scans of plates, measuring the distances between the root tip and marks made on the rear of the plates corresponding to the locations of root tips at the beginning of experiments.

Immunofluorescence microscopy

Plants were processed for immunofluorescence and confocal microscopy, as described previously (Collings & Wasteneys, 2005). Five-day-old seedlings were fixed in 50 mm Pipes (pH 7.2), 2 mm EGTA, 2 mm MgSO4 and 0.1% (v/v) Triton X-100 (PME) containing 2 mm phenylmethylsulfonyl fluoride (PMSF), 400 µm maleimidobenzoyl-N-hydroxysuccinimide ester (MBS; Pierce, Rockford, IL, USA), 4% (v/v) formaldehyde (ProScitech, Thuringowa, Qld, Australia) and 1% (v/v) glutaraldehyde (Sigma) (40 min), washed in PME, and extracted with PME solution containing 1% (v/v) Triton X-100 (1 h). After washing in PME, cell walls were digested for 20 min with 1% (w/v) cellulysin Y6 and 0.1% (w/v) pectolyase Y23 (MP Biomedicals) dissolved in PME buffer containing 1% (w/v) bovine serum albumin (BSA) and 0.4 m mannitol, washed in PME, permeabilized in methanol (−20°C, 10 min) and rehydrated in PBS (131 mm NaCl, 5.1 mm Na2HPO4, 1.56 mm KH2PO4, pH 7.2). Free aldehyde groups were reduced with fresh sodium borohydride (5 mg ml−1) in phosphate-buffered saline (PBS) (20–40 min), and after several rinses in PBS, seedlings were laid on parafilm and blocked in incubation buffer [PBS containing 1% (w/v) BSA] (30 min). Primary antibody incubations (overnight, 4°C) were polyclonal anti-maize actin (Dr Chris Staiger, Purdue University, IN, USA) and monoclonal anti-α-tubulin (clone B512; Sigma) concurrently diluted 1 : 200 and 1 : 1000 in incubation buffer. Roots were washed in PBS (four washes, 15 min each) and incubated overnight at 4°C in secondary antibodies [sheep anti-rabbit immunoglobulin G (IgG)-conjugated fluorescein isothiocyanate (FITC), diluted 1 : 100 (Silenus-Amrad, Boronia, Vic., Australia) and goat anti-mouse IgG-conjugated Cy-5, diluted 1 : 200 (Jackson, West Grove, PA, USA)]. Root tips were washed in PBS (three washes, 15 min each) and mounted in Citifluor antifade solution AF1 (Citifluor, London, UK). Samples were viewed using a confocal microscope (model SP2; Leica) with a ×40 NA 1.25 oil-immersion lens. FITC was excited with a 488-nm laser and fluorescence collected from 500 to 600 nm, while Cy-5 was excited at 633 nm and fluorescence collected from 650 to 760 nm. Optical sections were collected at 0.5-µm intervals, and computer-generated cross-sections through the roots were used to measure the widths of epidermal cells. All images were processed using Adobe Photoshop, version 6.0.


Concentration-dependent responses of microtubules and microfilaments to oryzalin and latrunculin B

We used an immunofluorescence procedure developed specifically to examine co-immunolabelled microtubules and microfilaments in Arabidopsis root tips (Collings & Wasteneys, 2005) to compare changes in the arrangement of microtubules and actin microfilaments in individual wild-type (Columbia ecotype) root epidermal cells following 48 h of treatment with oryzalin and latrunculin B at 21°C (Fig. 1). Control roots contained transverse cortical microtubules and extensive arrays of microfilaments, although few transverse cortical microfilaments were observed in the absence of MBS pretreatment (Fig. 1e). Compared with controls, 100 nm oryzalin had little effect on microtubule organization (Fig. 1a), while microtubules were clearly disrupted at 330 nm oryzalin (Fig. 1b) and at higher concentrations that generated progressively shorter, sparser and more disordered cortical microtubules (Fig. 1c,d). Higher concentrations of oryzalin generated a higher degree of isotropic expansion (Fig. 1c,d). Cell shapes became highly irregular, especially at intermediate oryzalin concentrations that disrupted phragmoplasts in dividing cells. Irregular divisions were not observed at higher oryzalin concentrations, probably because cell division was completely arrested (data not shown). Oryzalin treatments did not significantly modify microfilament distribution patterns (Fig. 1a–d).

Figure 1.

Oryzalin and latrunculin B cause concentration-dependent changes in the root cytoskeleton. Confocal immunofluorescence images are maximum projections covering between 3 and 5 µm, and show microtubules and microfilaments in Arabidopsis thaliana root epidermal cells following 48 h of treatment with oryzalin (Ory) at 100, 330, 1000 and 3300 nm (a–d) and latrunculin B (LatB) at 100, 330, 1000 and 3300 nm (e–i). A control root (control), grown in the presence of 0.2% dimethylsulfoxide (DMSO) (e). Bar, 20 µm for all images.

In wild-type roots at 21°C, microfilament bundles first appeared shorter when exposed to 100 nm latrunculin B (Fig. 1f). This effect became more pronounced at 330 nm latrunculin B (Fig. 1g), and few microfilaments were visible at or above 1 µm latrunculin B (Fig. 1h). Irregularly shaped cells were also observed, especially at higher concentrations of latrunculin B (Fig. 1i) but rarely at low or intermediate concentrations (Fig. 1f–h). Exposure to latrunculin B had no noticeable effect on microtubule organization (Fig. 1f–i).

mor1-1 plants are hypersensitive to microtubule-targeted drugs

We compared the sensitivities of root elongation and swelling to microtubule- and actin-targeted drugs in wild-type plants with those of the temperature-sensitive mor1-1 mutant (Whittington et al., 2001) (Fig. 2; Table 1). As a control, we also compared the effects of these drugs on the CesA1 mutant, root swelling 1 (rsw1-1) (Arioli et al., 1998). Both mor1-1 and rsw1-1 generate temperature-sensitive reductions in root elongation and induced root swelling. In mor1-1, these effects are caused by disruption to microtubule arrays with no effect on cellulose synthesis (Whittington et al., 2001; Himmelspach et al., 2003; Sugimoto et al., 2003), but in rsw1-1 these effects are primarily caused by reduced cellulose synthesis (Arioli et al., 1998; Sugimoto et al., 2001). From the dose–response curves shown in Fig. 2, we estimated the drug concentrations at which root diameters showed half the maximal increase (inline image) and the drug concentrations at which roots showed a 50% reduction in elongation (inline image) (Table 1).

Figure 2.

Measurements of drug-induced root swelling and drug-dependent inhibition of root elongation reveal that mutants show hypersensitivity to specific drugs. We transferred 5-d-old Arabidopsis thaliana seedlings of wild-type plants (left-hand column), mor1-1 plants (centre column) or rsw1-1 plants (right-hand column) to drug-containing plates and measured dose–response curves for root diameters (a,c) and root elongation (b,d) after 48 h at either 21°C or 29°C. The drug regimes used generated microtubule disruption with oryzalin (a,b) and microfilament disruption with latrunculin B (c,d). The position of arrows indicate the approximate direction and magnitude of hypersensitivity. Data points are the means ± standard error (SE) for ≥ 30 roots, from four or more replicate experiments.Wild type (21° and 29°C), mor1-1 (21°C) and rsw1-1 (21°C) plants show similar root swelling (a) and root elongation (b) responses to oryzalin treatments, but at the restrictive temperature (29°C), mor1-1 plants are sensitized to microtubule disruption with swelling occurring at significantly lower oryzalin concentrations (a, vertical arrow), and with root elongation being inhibited at lower concentrations (b, arrow). Wild-type (21° and 29°C), mor1-1 (21°C) and rsw1-1 (21°C) plants show similar root swelling (c) and root elongation (d) responses to treatment with latrunculin B, with wild-type and mor1-1 roots showing slight but significant increases in root elongation at low concentrations of latrunculin B (d, asterisks). At the restrictive temperature, however, root swelling in mor1-1 plants begins at significantly lower concentrations of latrunculin B than it does at 21°C (c, arrow), although significant differences were not visible in root elongation (d). Furthermore, whereas latrunculin B induces root swelling at 21°C in rsw1-1 plants, at 29°C it inhibits the phenotype-related root swelling over the same concentration range (c, hatch-mark). Swelling and elongation data are also shown for the mor1-1 rsw1-1 double mutant in the right-hand panels of (c) and (d), with the inhibition of root swelling occurring at similar concentrations of latrunculin B when compared with the rsw1-1 mutant (c, double hatch-mark).

Table 1.  Induction of root swelling and inhibition of root elongation in Arabidopsis mutants reveals hypersensitivity to different drugs
 Induction of root swellingInhibition of root elongation
Wild-type inline imageamor1-1 inline imagersw1-1 inline imageWild-type inline imagebmor1-1 inline imagersw1-1 inline image
  • a

    Half-maximal swelling concentrations (inline image) estimated as the drug concentration that causes root swelling half-way between the zero-drug response and the maximal swelling response. Estimates derive from adjacent points on the dose–response curves. Statistical modelling to determine values and errors was impractical owing to a lack of convergence with some curves.

  • b

    Half-minimal elongation concentrations (inline image) estimated as the drug concentration that causes root elongation half-way between the zero-drug response and the maximum effect on elongation. Estimates derive from adjacent points on the dose–response curves. Statistical modelling to determine values and errors was impractical owing to a lack of convergence with some curves.

  • c

    Dose–response curves show a hypersensitive drug response at low drug concentrations with only a minimal change in inline image.

  • d

    Hypersensitive response at 21°C compared with wild-type plants.

  • e

    Hypersensitive response evident by a temperature shift from 21°C to 29°C, and in comparison to wild-type plants.

  • f

    Estimated half-minimal concentration for swelling inhibition by latrunculin B in rsw1-1 plants at 29°C. ind, root diameter is independent of drug application, indicating that the mutant phenotype masks effects of the drug being observed.

Microtubule-related drugs
 Oryzalin (nm)460.6536.6455.3302.4479.3ind234.1295.3180.3d124.7e263.7ind
 Propyzamide (µm) 13.0  9.7  9.7 10.1c  9.2ind  6.1  6.8  4.8d  2.9e  6.2ind
 Taxol (µm)  2.7  2.8  1.7  0.2e  3.1ind  1.6  1.9  1.0  0.4e  1.0ind
Microfilament-related drugs
 Latrunculin B (nm)212.7196.3211.1 83.0e181.5226.4f 92.4152.8104.5 94.3133.5ind
 Cytochalasin D (µm)  6.4  5.2  5.9  2.3e  6.2ind  2.4  2.9  3.0  2.5  3.1ind

Seven-day-old wild-type plants maintained an average root diameter of between 130 and 150 µm at both 21°C and 29°C, but swelled in response to microtubule disruption when exposed for 48 h to the herbicide, oryzalin (Fig. 2a). The dose–response curves demonstrate that the lowest concentration of oryzalin to cause significant swelling at either 21°C or 29°C was 180 nm, and the inline image values were estimated to be approx. 500 nm (Table 1). Oryzalin reduced root elongation in a concentration-dependent manner, and although elongation was greater at 29°C, the similar response curves at 21°C and 29°C suggest that the effects of oryzalin were comparable at either temperature (Fig. 1b). inline image values were estimated to be 230 nm and 300 nm at 21°C and 29°C, respectively (Table 1). Compared with root swelling, root elongation was first affected at slightly lower concentrations.

The roots of mor1-1 and rsw1-1 mutants grew normally at the permissive temperature of 21°C, and had elongation rates and diameters similar to those of wild-type plants (Fig. 2a,b). However, when moved to 29°C, root elongation was significantly reduced and, after 48 h, mor1-1 and rsw1-1 root tips had swelled to approx. 220 and 400 µm in diameter, respectively (Fig. 2a). At 21°C, both mor1-1 and rsw1-1 plants responded to oryzalin in the same way as wild-type plants, with similar dose–response curves (Fig. 2a,b) and inline image and inline image values (Table 1). The dramatically reduced elongation and increased diameter of rsw1-1 roots at 29°C masked the effects of microtubule depolymerization. However, in mor1-1 at 29°C, the synergistic effects of oryzalin on the root swelling response were obvious. At 29°C, root swelling was stimulated at 33 nm oryzalin instead of the 180 nm required at 21°C. Much of the difference detected between 21°C and 29°C was apparent in the lower range of oryzalin concentrations tested (Fig. 2a; arrow); the inline image value was only reduced from 460 nm to 300 nm (Table 1). Root elongation in mor1-1 at 29°C was also more sensitive to oryzalin (Fig. 2b; arrow), with the inline image value reduced from 180 nm to 130 nm. We detected similar hypersensitive responses to propyzamide and taxol in mor1-1 plants, but not in rsw1-1 plants (Table 1). We also measured a similar oryzalin hypersensitivity response in a second temperature-sensitive allele, mor1-2 (Table 2). Thus, two conditional mor1-1 mutant alleles are hypersensitive to several distinct microtubule-targeted drugs under restrictive conditions.

Table 2. Arabidopsis root diameter changes correlate with cell-width changes
  ControlOryzalin (100 nm)Latrunculin B (100 nm)
Root diameter (µm)Cell width (µm)Root diameter (µm)Cell width (µm)Root diameter (µm)Cell width (µm)
  • a

    Calculated as the percentage change from 21° to 29°C.

  • **

    Values significantly different between 21°C and 29°C (P < 0.05, Student's t-test).

  • n > 20 replicate roots for all measurements of root diameter; n > 10 roots for measurements of cell widths (or > 5 roots for oryzalin treatments).

WT21°C149.3 ± 1.2 17.7 ± 0.6152.3 ± 2.5 16.3 ± 0.4156.0 ± 4.1 17.8 ± 0.5
29°C140.0 ± 1.6 15.7 ± 0.4154.0 ± 2.3 17.4 ± 0.8144.3 ± 2.8 16.7 ± 0.6
% Changea 93.8 88.4101.1107.1 92.5 93.9
mor1-121°C148.7 ± 1.6 17.3 ± 0.5176.8 ± 5.9 16.1 ± 0.8161.1 ± 5.0 17.0 ± 0.7
29°C194.7 ± 3.5** 19.4 ± 0.6302.9 ± 8.8** 23.1 ± 2.1260.1 ± 10.8** 22.0 ± 1.3
% Change130.9112.0171.3143.4161.5129.5

mor1-1 plants are hypersensitive to microfilament-targeted drugs

Exposure for 48 h to the microfilament-disrupting drug, latrunculin B, induced swelling in wild-type Arabidopsis roots, although this was not as pronounced as the microtubule disruption caused by oryzalin (Fig. 2c). Significant swelling was first detected between 100 nm and 330 nm at both 21°C and 29°C (inline image = 200 nm). Latrunculin B reduced root elongation at concentrations of > 100 nm (Fig. 2d) (inline image = 100 nm) (Table 1). Interestingly, low concentrations of latrunculin B often caused slight increases in root elongation (Fig. 2d; asterisks). The occurrence of these latrunculin B-induced growth increases, and similar growth increases measured with low concentrations of cytochalasin D, was, however, variable, as indicated by the relatively large size of the standard errors when compared with those found with drug-free controls and low oryzalin concentrations. Thus, our growth analysis shows that the effects of actin-targeted drugs on root growth are complex.

At 21°C, the swelling responses of mor1-1 and rsw1-1 mutants to latrunculin B were similar to those of wild-type roots (Fig. 2c; inline image = 210 nm and 180 nm for mor1-1 and rsw1-1, respectively, Table 1). At 29°C, mutation-dependent root swelling in rsw1-1 was reduced over a similar latrunculin B concentration range that promoted swelling at 21°C (Fig. 2c, hatch-mark; inline image = 230 nm). In contrast, mor1-1 roots at 29°C were hypersensitive to latrunculin B, and swelled at considerably lower concentrations than required for swelling at 21°C (Fig. 2c, arrow; inline image = 80 nm, Table 1). This response was also found for cytochalasin D, with the inline image value reduced from 5.9 µm to 2.3 µm when the temperature was raised from 21°C to 29°C (Table 1). The mor1-2 allele was also hypersensitive to latrunculin B, not only at 29°C, but also apparently at 21°C (Table 2). This finding is particularly interesting given that mor1-2 also has reduced seed set at 21°C, suggesting that mor1-2 has a partially constitutive phenotype. Compared with measurements of root swelling, analysis of root elongation in mor1-1 showed only a limited hypersensitivity response to latrunculin B (Fig. 2d; Table 1) and, as with latrunculin B, root elongation showed little difference in sensitivity to cytochalasin D at 21°C and 29°C (Table 1).

The opposite swelling responses of mor1-1 and rsw1-1 roots at higher latrunculin B concentrations prompted us to measure the effects of latrunculin B on root growth in the mor1-1 rsw1-1 double mutant. At 21°C, the double mutant responded to latrunculin B in a similar manner to the mor1-1 and rsw1-1 single mutants, though all latrunculin B concentrations generated slightly greater root diameters in the double mutant. At 29°C, the double mutant had consistently greater root diameters than either the mor1-1 or rsw1-1 single mutants, but the response to latrunculin B was essentially the same as in rsw1-1, as root diameters declined at latrunculin B concentrations of > 100 nm, and there was no sign of hypersensitivity to latrunculin B (Fig. 2c; double hatch-mark). This analysis suggests that the more severe effects of the rsw1-1 mutation on root growth mask any increased sensitivity to these drugs by mor1-1.

The hypersensitivity of the mor1-1 mutant to oryzalin and latrunculin B is clearly illustrated in the micrographs of root tips shown in Fig. 3. Wild-type root tips showed no swelling response when exposed to 100 nm oryzalin or 100 nm latrunculin B for 48 h at either 21°C or 29°C (Fig. 3a–f). Similarly, at 21°C, the mor1-1 (Fig. 3g–i) and rsw1-1 (Fig. 3m–o) mutants did not show any response to the drug treatments. At 29°C, however, the increased root diameter of mor1-1 (Fig. 3j) was greatly augmented by the low concentrations of both oryzalin (Fig. 3k) and latrunculin B (Fig. 3l). The impressive radial swelling of rsw1-1 at 29°C was not affected by either 100 nm oryzalin (Fig. 3q) or 100 nm latrunculin B (Fig. 3r).

Figure 3.

Arabidopsis thaliana roots show increased sensitivity to microfilament disruption when the microtubule cytoskeleton is disturbed. Wild-type roots (a–f) at either 21° (a–c) or 29°C (d–f) show no swelling response following 48 h of exposure to 0.2% dimethylsulfoxide (DMSO) controls (a,d), 100 nm oryzalin (Ory) (b,e) or to 100 nm latrunculin B (LatB) (c,f). mor1-1 roots (g–l) also fail to respond to these drugs by swelling at 21°C (g–i). However, while mor1-1 controls at 29°C show only a limited swelling response (j), mor1-1 plants respond strongly to the same low concentrations of oryzalin and latrunculin B (k,l). rsw1-1 plants (m,r), which grow normally at 21°C (m) but swell significantly in DMSO controls at 29°C (p), do not respond to either oryzalin (n) or latrunculin B (o) at 21°C. At 29°C, oryzalin does not modify root swelling in rsw1-1 (q). Although latrunculin B causes a significant reduction in swelling at higher concentrations, at 100 nm it has no inhibitory effects on rsw1-1 root swelling (r). Bar, 200 µm for all images.

Hypersensitivity at the cellular level

Immunofluorescence microscopy of the microtubule and microfilament cytoskeletons demonstrated hypersensitivity at the cellular level (Figs 4 and 5). In epidermal cells in the distal elongation zone of wild-type Arabidopsis roots, cortical microtubules were transverse at either 21°C or 29°C, consistent with these cells starting rapid elongation (Fig. 4a,b). Treatment of wild-type plants with 100 nm oryzalin for 48 h caused minimal root swelling and only subtle changes in microtubule organization (Fig. 4e,f). Unswollen mor1-1 roots grown at 21°C had microtubules organized in transverse arrays (Fig. 4c), and treatment of mor1-1 seedlings with 100 nm oryzalin at at 21°C had little effect on microtubules (Fig. 4g). As previously published (Whittington et al., 2001; Sugimoto et al., 2003; Kawamura et al., 2005), mor1-1 seedlings grown at 29°C showed some root swelling, had partially disorganized cortical microtubule arrays, and showed aberrant cell-division planes (Fig. 4d). The treatment of mor1-1 roots with 100 nm oryzalin at 29°C eliminated all but a few short and disordered microtubules (Fig. 4h). Under these conditions, microfilaments in mor1-1 plants remained abundant and of normal appearance (Fig. 4h).

Figure 4.

Hypersensitivity of mor1-1 to latrunculin B is demonstrated by immunofluorescence. Four- or 5-d-old wild-type and mor1-1 Arabidopsis thaliana plants were transferred to and grown on fresh plates at either 21°C (a,c,e,g,i,k) or 29°C (b,d,f,h,j,l) in the presence of 0.2% dimethylsulfoxide (DMSO) (control, a–d), 100 nm oryzalin (e–h) or 100 nm latrunculin B (i–l). Images show epidermal cells in the distal elongation zone of the root, and are projections of between five and 10 confocal optical sections covering a depth of between 3 and 5 µm. Bar, 20 µm for all images.In wild-type plants, little difference occurs between the microtubules and microfilaments at either 21°C (a) or 29°C (b), and mor1-1 plants at 21°C (c) appear similar. Raising the temperature to 29°C has a disorganizing effect on microtubules, but not microfilaments, in mor1-1 plants (d). Oryzalin (100 nm) has little effect on either microtubules or microfilaments at 21°C in wild-type plants (e) or mor1-1 plants (g), and similarly has little effect on microtubules and microfilaments in wild-type plants at 29°C (f). However, although either 100 nm oryzalin or 29°C alone have relatively few effects on microtubule organization in mor1-1 plants, when combined, microtubule depolymerization is almost complete although microfilaments appear normal (h). Latrunculin B (100 nm) has little effect on either microtubules or microfilaments at 21°C in either wild-type (i) or mor1-1 plants (k), and similarly has little effect on microtubules or microfilaments in wild-type plants at 29°C (j). However, although neither 100 nm latrunculin B nor 29°C alone have effects on cytoskeletal organization in mor1-1 plants, when combined, microtubules show significant disruption, even though microfilaments appear normal (l).

Figure 5.

Hypersensitivity is demonstrated by immunofluorescence of epidermal cells in the distal elongation zone of the root. Wild-type (a,b) and mor1-1 (c,d) Arabidopsis thaliana plants were transferred to fresh plates at 29°C containing either 0.2% dimethylsulfoxide (DMSO) (control; a,c) or 100 nm latrunculin B (LatB; b,d). Latrunculin treatments caused similar minor levels of disruption to the microfilaments in both wild-type (b) and mor1-1 plants (d). However, while microtubule organization in wild-type plants (b) was relatively similar to the control (a), in mor1-1 plants, treatment with 100 nm latrunculin resulted in considerably more microtubule disruption (d) than in the drug-free treatment (c). Bar, 20 µm for all images.

Latrunculin B (100 nm) caused disruption to the microfilament cytoskeleton, with microfilament bundles becoming shorter and more scattered. In wild-type plants, this response was independent of temperature, being similar at both 21°C and 29°C, and there was little change in microtubule organization (Figs 4i,j and 5b). In mor1-1 plants, 100 nm latrunculin also caused similar levels of microfilament disruption at both 21° (Fig. 4k) and 29°C, with the response of the microfilament cytoskeleton being unaffected by the microtubule disruption that occurs at 29°C (Figs 4l and 5d). Unexpectedly, the hypersensitivity of mor1-1 to latrunculin B (100 nm) at 29°C was manifested by increased disruption of microtubules, with only scattered and sparse microtubules remaining (Figs 4l and 5d). This microtubule disruption was considerably greater than observed in drug-free mor1-1 plants (Figs 4d and 5c).

Using measurements of cell diameter calculated from computer-generated cross-sections through immunolabelled roots, we determined that hypersensitivity to oryzalin in mor1-1, visible as increased root diameters, was also visible as an increase in epidermal cell widths (Table 2). In wild-type roots, cell widths did not vary significantly at either 21°C or 29°C in response to either 100 nm oryzalin or 100 nm latrunculin. Similarly, mor1-1 roots, exposed to oryzalin or latrunculin at 21°C, showed little change in cell width. These measurements were consistent with the root diameter changing little with these treatments. In mor1-1 roots at 29°C, there was a slight, but nonsignificant, increase in cell width of 12%, even though there was a significant increase in root diameter of 31%. This difference was probably a result of the aberrant cell divisions found in mor1-1 at the restrictive temperature (Kawamura et al., 2005), confounding the measurements of cell widths. However, the slight increase in cell widths caused by the mutant phenotype was vastly increased by the application of either 100 nm oryzalin or 100 nm latrunculin, although these increases of 43% and 30%, respectively, were of only borderline statistical significance. Nevertheless, our observations of cell widths were consistent with both the observed disruption in microtubule labelling (Fig. 4h) and increased root diameters (Figs 2a and 3k).

Oryzalin-induced disruption of microtubules in wild-type plants also enhanced sensitivity to latrunculin B

If microtubule destabilization at 29°C is the basis for the increased sensitivity to actin-targeted drugs in mor1-1 plants, then drug-induced disruption of microtubules in wild-type plants at 21°C should generate a similar response. To test this, we measured latrunculin B-induced root swelling in the presence of different concentrations of oryzalin (Fig. 6). At oryzalin concentrations of < 180 nm, latrunculin B stimulated additional radial swelling at concentrations that, in the absence of oryzalin, had no effect (Fig. 6a; arrow). The inline image value for latrunculin B was estimated to be 150 nm in the absence of oryzalin, and decreased to 90 nm in the presence of 180 nm oryzalin. At oryzalin concentrations of > 330 nm, root swelling was sufficiently great to mask latrunculin B-specific swelling. However, at the highest oryzalin concentration tested (1 µm), latrunculin B inhibited the full root-swelling response at similar concentrations to those causing swelling in the absence of oryzalin. Calculations also showed that the latrunculin B dose–response effects on root elongation were not altered by oryzalin treatments, with inline image values remaining constant over the full range of oryzalin (data not shown).

Figure 6.

Microtubule disruption with oryzalin (Ory) sensitizes wild-type Arabidopsis thaliana plants to microfilament disruption with latrunculin B (LatB). (a) Wild-type plants were treated for 48 h at 21°C with varying concentrations of latrunculin B in a background of differing amounts of oryzalin. Between 100 and 180 nm oryzalin, concentrations that do not themselves significantly induce swelling, root swelling is induced at latrunculin B concentrations that would otherwise show no response (arrow). (b) The same data set replotted against oryzalin. Low latrunculin B concentrations (100 nm) sensitize plants to microtubule disruption at oryzalin concentrations that would otherwise not cause root swelling (arrow). At higher latrunculin B concentrations, the amount of swelling is essentially independent of the state of the microtubule cytoskeleton (asterisk). Data points represent the mean values ± standard error (SE) for > 15 roots, from three replicate experiments.

Replotting the root swelling data to follow the effects of changing oryzalin concentrations, while keeping latrunculin B levels constant, revealed interesting relationships (Fig. 6b). At the highest latrunculin B concentrations (330 and 1000 nm), root swelling was only moderately but steadily increased by the addition of oryzalin, without the sharp increase in root diameter at concentrations of > 180 nm oryzalin observed in the latrunculin B-free treatments (Fig. 6b; asterisk). Latrunculin B, at concentrations of < 100 nm, had no effect on the oryzalin responses, but at 100 nm latrunculin B, the oryzalin response curve was altered, suggesting that the sensitivity to oryzalin was increased (Fig. 6b, arrow).

Cross-sensitivity in cytoskeletal mutants

We tested a selection of Arabidopsis mutants with modified cytoskeletal properties and a range of transgenic lines expressing reporter fusion proteins that are suspected of altering the dynamic properties of the cytoskeleton for root swelling sensitivity to latrunculin B and oryzalin (Fig. 7) (Table 3). As these mutant and transgenic lines originated in different genetic backgrounds, we conducted controls for both Columbia and Landsberg erecta ecotypes, which revealed no significant differences in root-swelling sensitivities to latrunculin B or oryzalin (Fig. 7a,b; Table 3).

Figure 7.

Some known cytoskeleton-related mutants show hypersensitivity. Arabidopsis thaliana plants were treated for 48 h with varying concentrations of either oryzalin (a,c) or latrunculin (b,d) at 21°C. Higher temperatures were not used, as no temperature-sensitive mutations were studied. Compared with Columbia and Landsberg erecta controls, botero and spiral1 are hypersensitive to latrunculin B (b, arrow and double arrow, respectively), whereas only botero is hypersensitive to oryzalin (a, arrow). lefty1 is not hypersensitive to either latrunculin B or oryzalin. Furthermore, compared with Columbia wild-type controls, neither rhd3 nor the distorted-class mutant gnarled are hypersensitive to either latrunculin B (d) or oryzalin (c). Data points represent the mean values ± standard error (SE) for ≥ 10 roots, from two or more replicate experiments.

Table 3. Arabidopsis mutants and GFP-cytoskeletal lines demonstrate hypersensitivity to different drugs
Mutant/GFP lineProteinPhenotypeBackgroundRef.Latrunculin B inline imagea (nm)Oryzalin inline image (nm)
 Columbia  (C) 210200460540
 Landsberg erecta  (L) 220290320380
 mor1-1217 kDa MAPTS MT disruption causes root twisting & swellingC121080c460300c
 mor1-2217 kDa MAPTS MT disruption causes root twisting & swellingC1100d30c410410
 botero-157 kDa small subunit of kataninConstitutively swollen and short rootsC2, 350e 120e 
 spr1-112 kDa plus-end MAPRight-handed organ twistingL4, 5, 620e 420 
 lefty1α-tubulin 6, TUA6Left-handed root twistingL7300 350 
 GFP-MBDFusion of GFP and the MT-binding domain of mouse MAP4Can generate right-handed root twistingC8200 560 
 GFP-tubulinFusion protein of GFP and α-tubulinCan generate right-handed root twistingC9200 440 
Microfilament-relatedFusion of GFP and the actin-binding domain of Arabidopsis fimbrinNo apparent phenotypeC10200 520 
 GFP-talinFusion of GFP and the actin-binding domain of human talinPlants can show a dwarfed phenotypeC11190 400 
 distorted1:: GFP-talindistorted class mutant; encodes an ARP2/3 complex protein, in a background of a fusion between GFP and the actin-binding domain of mouse talinTrichome disruptions, but no root phenotypeL12160 440 
 gnarleddistorted class mutant; an ARP2/3 complex-regulating proteinTrichome disruptions, but no root phenotypeC13190 410 
 rhd3-1Putative GTP-binding protein aberrant root microfilamentsDefective root hairs andC14240 290 
Cellulose-relatedcesA1, a cellulose synthaseTS block in celluloseC15, 16180230f480ind
 rsw1-1 synthesis causes swelling      
 mor1-1 rsw1-1C17100240fndnd

We investigated three Arabidopsis mutants in which constitutive microtubule-related defects are manifested in roots. The botero1 (bot1) mutant is a null allele of the p60 subunit of the microtubule-severing protein, katanin (Bichet et al., 2001; Burk et al., 2001) and fails to organize transverse cortical arrays in the elongation zone, resulting in constitutive radial swelling and reduced elongation. Compared with wild-type controls, bot1 was hypersensitive to oryzalin (Fig. 7a; arrow), with the inline image value decreasing from approx. 400 nm in wild type to 120 nm in the presence of oryzalin (Table 3). However, bot1 was also hypersensitive to latrunculin B with a swelling response occurring in as little as 10 nm latrunculin B (Fig. 7b; arrow; inline image value reduced from 200 nm to 50 nm). spiral1 (spr1), a null allele of a 12-kDa plus-end microtubule-binding protein (Nakajima et al., 2004; Sedbrook et al., 2004), characterized by constitutive right-handed twisting of organs (Furutani et al., 2000), was hypersensitive to latrunculin B (Fig. 7b; double headed arrow; inline image = 20 nm) but, interestingly, not to oryzalin (Fig. 7a). The lefty1 mutant, which has left-handed organ twisting because of a mutation in TUA6, and which has reported hypersensitivity to microtubule disruption with taxol and propyzamide (Thitamadee et al., 2002), displayed normal sensitivity to both oryzalin (Fig. 7a) and latrunculin B (Fig. 7b), with inline image values similar to those of wild-type plants (Table 3).

We also tested the drug responses of two lines expressing green fluorescent protein (GFP) fusions that label microtubules [GFP–TUA6 (Ueda et al., 1999) and GFP–MBD (Granger & Cyr, 2001)] which are noted for displaying right-handed root twisting (Ueda et al., 1999; Hashimoto, 2002). Neither of these lines had altered sensitivity to either latrunculin B or oryzalin (Table 3).

Three Arabidopsis mutants that have microfilament-related defects were also tested for their sensitivities to oryzalin and latrunculin B by the root swelling assay (Fig. 7c,d). These were root hair defective3 (rhd3) (Hu et al., 2003), which has aberrant microfilament arrays within root epidermal cells, distorted1 (dis1), a member of the distorted class of trichome mutants (Hülskamp et al., 1994; Mathur et al., 2003) (data not shown), and an allele of GNARLED, a NAP125 homologue that regulates actin organization (Brembu et al., 2004; El-Assal et al., 2004; Zimmermann et al., 2004). None of these mutants showed heightened sensitivity to oryzalin (Fig. 7c) or latrunculin B (Fig. 7d). The dis1 line also carried a transgene encoding a GFP–talin fusion protein. We found that constitutive expression of a related GFP–talin construct (Takemoto et al., 2003, based on Kost et al., 1998) in a wild-type background did not alter sensitivity to either latrunculin B or oryzalin. Similarly, expression of GFP::fABD2 (Arabidopsis fimbrin second actin-binding domain) (Sheahan et al., 2004) had no effect (Table 3).


Using measurements of root swelling to compare drug sensitivities, we have shown that microtubule disruption stimulates hypersensitivity to the microfilament-disrupting drug, latrunculin B, in temperature-sensitive mor1 mutants at 29°C, in the constitutive mutants, botero or spiral1, and in wild-type plants treated with the microtubule-targeted herbicide oryzalin. mor1-1 plants at 29°C are also hypersensitive to another actin-targeted drug, cytochalasin D. Because these experiments used mor1-1 and other microtubule-related mutants, we initially interpreted our hypersensitivity results as a response to microfilament disruption. However, hypersensitivity to actin-targeted drugs by disruption of microtubules also works in the opposite direction. Experiments using a combination of low concentrations of latrunculin B and oryzalin show that low concentrations of latrunculin B make plants hypersensitive to oryzalin. Moreover, while low concentrations of latrunculin B cause slight disruption to microfilaments in a manner independent of microtubule status, in mor1-1 plants at the restrictive temperature, these concentrations of latrunculin also cause further disruption to the microtubules. Our data suggest that cross-talk between microfilament and microtubule organization is important for regulating anisotropic cell expansion.

Microtubule-dependent hypersensitivity to actin-targeted drugs as a general phenomenon in plants

Two other experimental systems have revealed changes in the properties of the actin microfilament cytoskeleton as a direct consequence of microtubule disruption. In characean algal internodal cells, we demonstrated that although microtubule depolymerization does not inhibit the rapid, microfilament-dependent cytoplasmic streaming, it does hypersensitize cells to cytochalasins B and D (Wasteneys & Williamson, 1991; Collings et al., 1996; Foissner & Wasteneys, 2000). The spatial separation of the cortical microtubules, and the prominent subcortical microfilament bundles that generate the streaming, suggest that the mechanism underlying this hypersensitivity response does not depend on direct microtubule–microfilament interactions. Instead, we proposed that some factor, normally bound to microtubules, is released upon their depolymerization, and that this factor interacts with and renders the microfilament cytoskeleton hypersensitive to the destabilizing effects of cytochalasins (Collings et al., 1996). Recently, a similar mechanism has been proposed to account for interactions in the development of the interdigitated pavement cells of the Arabidopsis leaf, where spatially separated microtubule- and microfilament-dependent growth processes are regulated by rho of plants (ROP) GTPases (Fu et al., 2002; 2005). Using measurements of fluorescence resonance energy transfer to directly monitor interactions between ROP2/4 and one of their effectors, RIC4, Fu et al. (2005) demonstrated that interactions increase in the mor1-1 mutant at restrictive temperature and in wild-type plants during drug-induced microtubule depolymerization (Fu et al., 2005). As with the characean internodal system, microtubule disruption alters the properties of microfilaments, identifying a feedback loop that regulates localized cortical microfilament polymerization according to microtubule polymerization elsewhere in the cell. The identity of these microtubule-associated signalling elements remains unknown (Fu et al., 2005).

Our demonstration of changes in the properties of microfilaments as a result of microtubule disruption, along with our previous observations (Collings et al., 1996) and those of Fu et al. (2005), suggest that regulatory cross-talk between microtubules and microfilaments may be a general phenomenon in plant cells, but one that is not easily detected. In all cases so far, observing cross-talk required the use of specific experimental conditions. In our current study, we relied on the highly reproducible, anisotropic growth of the Arabidopsis primary root. Measurements of root diameter clearly revealed hypersensitivity of mor1-1 to latrunculin B and cytochalasin D, whereas concurrent recordings of root elongation showed a much less clear effect. One explanation for this could be that the increased variability in root elongation following microfilament disruption with low drug concentrations, visible as the large standard errors in our measurements, obscures any hypersensitivity response. Alternatively, the weakly bundled cortical microfilaments involved in anisotropic expansion might be more prone to modification than the other microfilament arrays, such as the thick cytoplasmic bundles of microfilaments responsible for organelle motility. Differential microfilament sensitivity has been demonstrated in Nitella internodal cells (Collings et al., 1995), pollen tubes (Gibbon et al., 1999) and root hairs (Ketelaar et al., 2003). Verifying differential sensitivity in expanding Arabidopsis root cells has not been possible by conventional approaches (Collings & Wasteneys, 2005) but GFP–fimbrin-expressing lines show much promise (Sheahan et al., 2004).

Hypersensitivity and microfilament–microtubule cross-talk

Identifying the mechanisms that might underlie microtubule–microfilament cross-talk, identified through the use of drugs, is challenging. In part, this is because the fundamental and diverse roles of microfilaments in cells make the effects of actin-disrupting drugs, such as latrunculin, pleiotropic. For example, as microfilaments function in the regulation and movement of the endomembrane system (Boevink et al., 1998; Nebenführ et al., 1999), might hypersensitivity be a manifestation of general defects there?

Our analysis of hypersensitivity as a general response suggests that this is unlikely. In this study, we have already shown that hypersensitivity to drugs which target the cytoskeleton occurs only in mutants in which the cytoskeleton is disrupted, and that hypersensitivity in these mutants (if it occurs at all) is limited to drugs that target the cytoskeleton. This is consistent with a more generalized effect. We used the root swelling assay, and a wide range of root growth mutants and drugs, to demonstrate that a mutant will only show hypersensitivity to a drug if the drug and the mutation target the same developmental process. Thus, mor1-1 is hypersensitive to microtubule-disrupting drugs (this study), but lacks hypersensitivity to cellulose biosynthesis inhibitors or to drugs that target the endomembrane system (D. A. Collings et al., unpubl. res.). Similarly, rsw1-1 is not hypersensitive to microtubule-disrupting drugs (this study) or to drugs that target the endomembrane system, but is hypersensitive to both dichlobenil (DCB) and isoxaben (D. A. Collings et al., unpubl. res.), and in mutants where vesicle trafficking is compromised, hypersensitivity does not occur with either cytoskeletal or cellulose biosynthesis antagonists, but can occur with drugs that target the endomembrane system (D. A. Collings et al., unpubl. res.).

We therefore conclude that latrunculin hypersensitivity in mutants in which microtubules are compromised suggests that there is indeed some form of microtubule–microfilament cross-talk. In identifying the mechanisms that underlie microtubule–microfilament cross-talk, it is necessary to distinguish mechanisms that rely on close physical interactions between the two cytoskeletal elements and those involving indirect regulatory networks (Collings & Allen, 2000). Indirect mechanisms, such as those identified in the Nitella internodal cell system (Collings et al., 1996) and in the ROP GTPase signalling system of the Arabidopsis leaf pavement cells (Fu et al., 2005), discussed above (under ‘Microtubule-dependent hypersensitivity to actin-targeted drugs as a general phenomenon in plants’), remain plausible. Alternatively, interactions could be more direct, mediated perhaps by a kinesin with microfilament-binding capability. Arabidopsis has kinesins that are closely related to a central motor domain, minus end-directed kinesin from Gossypium hirsutum (cotton) whose microfilament-binding capabilities have been demonstrated in vitro (Preuss et al., 2004). We have demonstrated, in a separate study, that microfilaments lie parallel to the prominent transverse microtubules of elongating Arabidopsis root cells, and that disruption of microtubules prevents the retention of transverse cortical microfilaments (Collings & Wasteneys, 2005). Similarly, microtubule stabilization can promote the formation of parallel microfilaments in tobacco protoplasts (Collings et al., 1998) and maize roots (Blancaflor, 2000). Interactions between microtubules and microfilaments may also stabilize cortical microtubules. In cotton fibres, microfilament disruption eventually leads to microtubule reorientation (Seagull, 1990), while in azuki bean epidermis, microfilament disruption modifies the cyclic reorientation of the cortical microtubules (Takesue & Shibaoka, 1998; Collings & Allen, 2000).

Microfilament disruption and the growth of Arabidopsis roots

Microfilament disruption with latrunculin B, and its effects on root growth, are complex. In addition to hypersensitivity to actin-targeted drugs that occurs when microtubules are disrupted, we identified a further five general responses following treatment with either latrunculin B or cytochalasin D. These concentration-dependent responses include: the presence of aberrant cell divisions; inhibition of root elongation; promotion of radial root swelling that begins at similar drug concentrations to those that inhibit elongation; inhibition of root swelling, again over similar concentrations, but specifically in roots that would otherwise undergo massive root swelling because of microtubule disruption or cellulose biosynthesis inhibition; and a slight and variable increase in root elongation at very low latrunculin B and cytochalasin D concentrations. This final effect was not always observed, and Thomas et al. (1973) also reported an inconsistent stimulation of onion root elongation at low cytochalasin B concentrations. A mechanism that might account for this growth-promoting effect is not known, but understanding the basis for the first four of the microfilament-disruption responses listed above would provide insight into the microtubule–microfilament cross-talk phenomenon.

In dividing cells, the microfilaments present in and around the phragmoplast function during cytokinesis. Disrupting the organization of these microfilaments with drugs (Palevitz & Hepler, 1974; Palevitz, 1980) or the actin-binding protein, profilin (Valster et al., 1997), or disrupting their function through drugs that block myosin action (Molchan et al., 2002), results in the inhibition of cell plate expansion and the presence of aberrant cell divisions. Could these aberrant cell divisions explain root swelling? We think not. (And even if they did, this could not explain the development of hypersensitivity.) In Arabidopsis roots, aberrant cell divisions resulting in skewed cross-walls occurred in response to high (> 1 µm) concentrations of latrunculin (Fig. 1i), but were rarely observed in response to lower latrunculin concentrations (100 nm) that did not promote swelling (Figs 1f and 4i,j). Even at intermediate latrunculin concentrations that did cause significant swelling (330 nm), aberrant cell divisions were not commonly observed (Fig. 1g,h). Furthermore, root swelling is not caused by an increased number of cells within the root. This has previously been demonstrated in mor1-1 (Whittington et al., 2001), and is consistent with our measurement of increased cell width in all swollen roots, regardless of whether or not they became hypersensitive (Table 2).

Microfilament disruption modifies root growth because microfilaments are necessary for cytoplasmic streaming, exocytosis and endocytosis (Wasteneys & Galway, 2003). It has been demonstrated that microfilament disruption causes a build-up of secretory vesicles in root tip cells (Vaughan & Vaughn, 1987), and prevents the cycling to and from the plasma membrane (via exocytosis and endocytosis) of auxin efflux facilitator proteins (Geldner et al., 2001) and other molecules. This role in cycling probably includes both myosin-dependent movement of Golgi stacks along microfilament bundles and possible further roles in exocytosis and endocytosis. Thus, microfilament disruption reduces root elongation because streaming, exocytosis and endocytosis are impaired. However, by blocking the cycling of auxin efflux facilitators, microfilament disruption will also modify their polar distribution and alter the flow of auxin within the root. Thus, the effects of microfilament disruption on root development will be compounded (Wasteneys & Collings, 2004).

Reduced cytoplasmic streaming, exocytosis and vesicle cycling during treatment with latrunculin B explain the inhibition of root elongation and also the inhibition of swelling in roots that undergo massive swelling responses, such as those treated with high concentrations of oryzalin or in the rsw1-1 mutant at 29°C. Both of these responses occur over the same range of latrunculin B concentrations and neither response is enhanced. For example, inhibition of root swelling in the mor1-1 rsw1-1 double mutant occurs over the same latrunculin B concentration range as in the rsw1-1 single mutant. Thus, perturbing microfilaments has a general inhibitory effect on growth, probably via inhibition of vesicle trafficking events, and does not appear to depend on microtubule function. In contrast, microtubule and microfilament systems are not only implicated in the stimulation of radial swelling, but appear to work synergistically in regulating growth anisotropy.

Hypersensitivity responses in other cytoskeletal mutants

Chemical genetics is an emerging tool for dissecting the function of many biological activities. Comparative sensitivity to microtubule-targeted drugs has recently been demonstrated. For example, the two spiral mutants, isolated on the basis of their right-handed organ twisting, show hypersensitivity to microtubule disruption (Furutani et al., 2000). spiral1, which encodes a plus-end microtubule-interacting protein (Nakajima et al., 2004; Sedbrook et al., 2004), shows growth changes at much lower propyzamide and taxol concentrations than in wild-type plants (Furutani et al., 2000; Sedbrook et al., 2004). The two lefty mutants, with similar mutations in two different α-tubulin isoforms, are also hypersensitive to microtubule disruption (Thitamadee et al., 2002), and propyzamide hypersensitive1, isolated in a screen for hypersensitivity to microtubule disruption with propyzamide, encodes a mitogen-activated protein kinase phosphatase-like gene (Naoi & Hashimoto, 2004).

Our measurements of hypersensitivity in some of these mutants stand in contrast to previously published data. We measured no hypersensitivity to oryzalin in lefty1, while Thitamadee et al. (2002) observed hypersensitivity to propyzamide. Similarly, while we observed hypersensitivity to latrunculin B in spr1, Furutani et al. (2000) detected no hypersensitivity to cytochalasin D. Apart from differences in the drugs tested, such differences may also be explained by differences in measurement protocol. Our observations of sensitivity (in mutants other than mor1-1) were restricted to measuring the highly reproducible dose–responses for root diameter, whereas other studies have used root elongation, root skewing, or direct observation of microtubules by immunofluorescence microscopy (Furutani et al., 2000; Thitamadee et al., 2002).

We also measured dose–responses in Arabidopsis mutants and transgenic lines in which the microfilaments are altered or labelled, expecting to demonstrate hypersensitivity responses to microfilament-disrupting drugs. The distorted class of Arabidopsis mutants are not hypersensitive to latrunculin B. The distorted mutants were initially isolated on the basis of aberrant trichomes (Hülskamp et al., 1994) whose phenotypes can be mimicked by microfilament disruption (Mathur et al., 1999; Szymanski et al., 1999). These mutants all encode components of the ARP2/3 complex, which functions in the nucleation and turnover of actin microfilaments, or proteins involved in signalling the ARP2/3 complex (Le et al., 2003; Mathur et al., 2003; Zimmermann et al., 2004). However, neither dis1, which encodes a component of the ARP2/3 complex, nor gnarled, a positive regulator of the ARP2/3 complex (Brembu et al., 2004; El-Assal et al., 2004; Zimmermann et al., 2004), showed any form of latrunculin B hypersensitivity. These observations are consistent with ARP2/3 complex component protein mutants and T-DNA knockouts lacking phenotypes in the root elongation zone (Li et al., 2003; Mathur et al., 2003) even though the complex is expressed there (Li et al., 2003). Taken together, the lack of a distorted phenotype in elongating roots, and our observations showing no hypersensitivity response, indicate that pathways distinct from the ARP2/3 complex might exist to regulate microfilament polymerization in the Arabidopsis root.

Summary and directions

In summary, we have coupled a conventional drug assay approach with the use of known and characterized Arabidopsis mutants to investigate the function of actin microfilaments in the control of cell shape. We show that microfilament–microtubule interactions are important in the regulation of anisotropic cell expansion, and our data confirm that the measurement of sensitivity to drug-based disruption in mutants is a powerful tool for using to investigate the functions affected by the mutations. We are now investigating the mechanism underlying microtubule–microfilament cross-talk, in part using mor1-1 plants expressing GFP::fABD2.


DAC is the recipient of an Australian Research Council (ARC) Research Fellowship and Discovery Grant DP0208806, while GOW receives funding from ARC Discovery Grant DP0208872. AWL acknowledges funding from an RSBS summer scholarship. Seed for various mutant and transgenic Arabidopsis lines were generously provided by different laboratories, and we thank Herman Höfte for botero, Takashi Hashimoto for spiral1, lefty1 and GFP-TUA6, John Schiefelbein for rhd3, Jaideep Mathur and Martin Hülskamp for dis1, Richard Williamson for rsw1, Richard Cyr for GFP-MBD, Daigo Takemoto and David Jones for GFP-talin, and Michael Sheahan and David McCurdy for GFP::fABD2. We also acknowledge numerous discussions with researchers in the field, including Zhenbiao Yang and others, whose comments and thoughts have proved invaluable. We thank John Harper for his valuable comments on the manuscript.