| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
|
First published online May 12, 2006; 10.1105/tpc.105.039644 The Plant Cell 18:1412-1425 (2006) © 2006 American Society of Plant Biologists
|
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMETHODSREFERENCES |
|---|
-Tubulin is known as one of the key molecular players for microtubule nucleation in animal and fungal cells. Here, we provide genetic evidence that in Arabidopsis thaliana, -tubulin is required for the formation of spindle, phragmoplast, and cortical microtubule arrays. We used a reverse genetics approach to investigate the role of the two Arabidopsis -tubulin genes in plant development and in the formation of microtubule arrays. Isolation of mutants in each gene and analysis of two combinations of -tubulin double mutants showed that the two genes have redundant functions. The first combination is lethal at the gametophytic stage. Disruption of both -tubulin genes causes aberrant spindle and phragmoplast structures and alters nuclear division in gametophytes. The second combination of -tubulin alleles affects late seedling development, ultimately leading to lethality 3 weeks after germination. This partially viable mutant combination enabled us to follow dynamically the effects of -tubulin depletion on microtubule arrays in dividing cells using a green fluorescent protein marker. These results establish the central role of -tubulin in the formation and organization of microtubule arrays in Arabidopsis. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMETHODSREFERENCES |
|---|
ß-tubulin heterodimers that are major structural components of the cytoskeleton in eukaryotic cells. In animal and fungal cells, microtubule nucleation takes place at conspicuous microtubule-organizing centers (MTOCs), such as the centrosome or the spindle pole body, whose activity determines the spatial and temporal organization of the microtubule cytoskeleton. Higher plant cells lack discrete MTOCs but assemble highly ordered arrays of microtubules that coordinate cell division and expansion (Wasteneys, 2002
). The preprophase band forms during late G2 at the cell cortex and delineates the future site of division. It is replaced during late prophase by an acentriolar mitotic spindle. The phragmoplast, assembled during late anaphase, promotes the synthesis of the new cell plate separating daughter cells. As they enter G1, a cortical microtubule array is formed at the cell cortex and participates in the control of cell elongation and cell wall deposition.
Different microtubule nucleation sites accounting for the assembly of plant microtubule arrays have been established in plant cells. The nuclear envelope has been shown to retain nucleating activity in vitro (Stoppin et al., 1994). Using fluorescent-tagged microtubule markers, in vivo nucleation has been revealed at the cell cortex, either at sites linked to preexisting microtubules (Shaw et al., 2003; Van Bruaene et al., 2004; Murata et al., 2005) or de novo at sites with no other detectable microtubules (Shaw et al., 2003). Fluorescent-tagged protein markers, such as ATEB1A–green fluorescent protein (GFP), have also provided support for the existence of diffuse, mobile nucleation sites at spindle poles (Chan et al., 2003). However, our knowledge about the molecular composition of plant microtubule nucleating sites is still in its infancy. In animals and fungi, a large body of evidence strongly implicates the highly conserved -tubulin protein, the third member of the tubulin protein family, as a key element for microtubule nucleation at MTOCs (reviewed in Jeng and Stearns, 1999; Moritz and Agard, 2001; Job et al., 2003; Jaspersen and Winey, 2004). Apart from the -tubulin pool present at the centrosome, most soluble cytoplasmic -tubulin is part of large complexes named -tubulin ring complexes (-TuRCs) containing 10 to 13 -tubulin molecules per complex and at least eight proteins in addition to -tubulin. -Tubulin present at the centrosome likely comes from association of -TuRCs with the pericentriolar matrix. Animal -TuRC is seen as an open ring of the approximate diameter of a microtubule that caps the minus end of a microtubule, but the precise mechanism of microtubule nucleation by -TuRCs is still being debated. Smaller complexes called -tubulin small complexes have also been found in many cell types and are clearly a component of -TuRCs. They consist of two -tubulin molecules and one molecule each of two related proteins known as Spc97 and Spc98 proteins in Saccharomyces cerevisiae.
In plants, the role of -tubulin is still a debated question. Plant -tubulin is present in protein complexes of various sizes in maize (Zea mays), Arabidopsis thaliana, and fava bean (Vicia faba) (Stoppin-Mellet et al., 2000; Drykova et al., 2003), and large -tubulin–containing complexes have nucleation activity in Arabidopsis (Drykova et al., 2003). The ability of plant -tubulin to nucleate microtubules was demonstrated by heterologous expression of Arabidopsis -tubulin in fission yeast lacking endogenous -tubulin. Arabidopsis -tubulin was able to bind MTOCs and nucleate microtubule assembly in Schizosaccharomyces pombe (Horio and Oakley, 2003). Another piece of evidence favoring a role of -tubulin in microtubule nucleation comes from a study on the subcellular localization of the plant Spc98 homologue. This protein, which was recently shown to be required for microtubule nucleation on isolated plant nuclei, colocalized with -tubulin on the nuclear surface (Erhardt et al., 2002). However, the unusual subcellular localization of -tubulin in plant cells, which seemed incompatible with a mere role in nucleation, puzzled cell biologists for long. Indeed, in addition to its presence at established nucleation sites, such as the nuclear surface in Arabidopsis, wheat (Triticum aestivum), soybean (Glycine max), and BY2 cells (Liu et al., 1993, 1994; Joshi and Palevitz, 1996; Erhardt et al., 2002), or the acentriolar polar organizer of basal land plants (Brown et al., 2004; Shimamura et al., 2004), plant -tubulin associates to all microtubule arrays along microtubules in a punctuate manner and is not restricted to microtubule minus ends, which would be expected for a protein supposedly involved in nucleation (Liu et al., 1993, 1994; Joshi and Palevitz, 1996; Panteris et al., 2000). Some aspects of this unusual localization have been clarified recently: Using a cell-free system, Murata et al. (2005) demonstrated that this punctuate labeling on cortical microtubules represents bona fide sites of nucleation on the sides of extant microtubules, resulting in branched structures (Murata et al., 2005). They also showed that -tubulin is required for this process. Although in such a cell-free system many parameters (e.g., local tubulin concentrations) may differ markedly from in vivo conditions, these results provide direct evidence of the involvement of -tubulin in nucleation of cortical microtubules. Whether, quantitatively speaking, this process represents a major mechanism in vivo still needs to be evaluated in addition to its involvement in the formation of mitotic arrays. In particular, the fact that the Arabidopsis Spc98 homologue, although present at the cortex, is not codistributed with -tubulin on the whole length of microtubules needs to be clarified (Erhardt et al., 2002).
Apart from its key role in microtubule nucleation, additional functions for -tubulin have been suggested in several eukaryotic systems. Recent genetic and molecular studies in fission and budding yeast, in Aspergillus nidulans, and in Drosophila melanogaster have revealed that -tubulin could be involved in microtubule dynamics or organization (Paluh et al., 2000; Jung et al., 2001; Vogel et al., 2001) and in the control of mitotic checkpoint and coordination of late mitotic events (Hendrickson et al., 2001; Sampaio et al., 2001; Prigozhina et al., 2004). Evidence for such roles is needed for plant -tubulin.
To gain further insights into the function of -tubulin in plant cells, we have obtained T-DNA insertion mutant lines for the two genes encoding -tubulin in Arabidopsis (Liu et al., 1994). Here, we show that the two Arabidopsis genes are functionally redundant. We studied two double mutant combinations of TUBG1 and TUBG2 insertion alleles and investigated the effect of -tubulin depletion on plant development and organization of mitotic and interphase microtubules arrays. Our results demonstrate the in vivo role of -tubulin in the organization of all microtubule structures in plant cells, both during interphase and cell division.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMETHODSREFERENCES |
|---|
-Tubulin Mutants-tubulin isoforms, TUBG1 (At3g61650) and TUBG2 (At5g05620) (Liu et al., 1994; Arabidopsis Genome Initiative, 2000), which share 98% protein sequence identity. RT-PCR analysis using gene-specific primers indicated that both -tubulin genes are constitutively expressed at high levels in all organs tested (Figure 1C
). Affymetrix ATH1 data, corresponding to the combined expression of both -tubulin genes (probe set 251331_s_at), confirms that global transcript level is high and rather constant during development, although especially high in the shoot apical meristem and in cultured cells, and very low in pollen (Zimmermann et al., 2004). To obtain mutants for both genes, we used either PCR screening of a T-DNA–mutagenized Arabidopsis population (Bechtold et al., 1993) or in silico searching in public T-DNA databases (Alonso et al., 2003). We identified one TUBG1 and two TUBG2 insertion lines: tubg1-1 harbors a T-DNA insert in the first exon of TUBG1 associated with a 55-bp deletion of the coding region, and tubg2-1 carries an insertion and a large deletion of most of the TUBG2 gene, whereas tubg2-2 has a T- DNA insert in the 5' untranslated region of TUBG2 (Figures 1A and 1B). DNA sequencing of the T-DNA flanking regions and DNA gel blot analysis enabled us to characterize the three insertion loci at the molecular level (Figures 1A and 1B).
|
All three homozygous single mutant lines had a wild-type phenotype in terms of growth and development and were fully fertile. In order to study the effects of a simultaneous deficiency for both -tubulin isoforms, we crossed homozygous tubg1-1 plants with either tubg2-1 or tubg2-2 homozygous plants. In both cases, neither double heterozygote F1 progenies nor homozygote/heterozygote F2 plants showed any vegetative defects, showing that one functional copy of -tubulin (out of four in the wild type) is enough to sustain growth and development of plants in standard conditions. However, the tubg1-1 tubg2-1 combination induced severe gametophytic defects, and double mutant seedlings were never recovered. For the other allelic combination, tubg1-1 tubg2-2 double mutant seedlings displayed morphological defects and developmental arrests a few days after germination. Detailed analysis of the phenotypes is given below.
The severity of the phenotypes observed (gametophytic lethality in one case and seedling lethality in the other) suggested a drastic depletion of -tubulin levels in both double mutant combinations. Transcript levels are generally poor indicators of molecular defects in insertional mutants. Therefore, we used an anti--tubulin antibody to determine protein levels in double mutant plantlets by protein gel blot analysis. To get sufficient mutant material, we used tubg1-1 tubg2-2 double mutant seedlings of 5 to 7 d, the earliest stage where morphological defects are clear enough to unambiguously distinguish mutants from wild-type plantlets. This experiment shows that, in comparison with the strong signal observed in wild types of the same age, -tubulin is undetectable in such double mutant plantlets (Figure 1D).
Taken together, these results establish the following: (1) Genetically speaking, the observed phenotypes are strictly linked to simultaneous disruption of both -tubulin genes since they appear when insertion alleles are combined and only in this case. In addition, a TUBG2 cDNA expressed under the control of the promoter of the 35S gene of Cauliflower mosaic virus fully complements the phenotype of the tubg1-1 tubg2-2 mutant (Figure 5C), which further demonstrates the link between the mutations and the observed phenotypes. (2) The absence of defect in single mutants indicate high, if not complete, functional redundancy between TUBG1 and TUBG2. (3) -Tubulin is drastically reduced and below detection level in tubg1-1 tubg2-2 seedlings. Given its even more severe defects, the tubg1-1 tubg2-1 combination presumably represents a null mutant, consistent with the molecular nature of the tubg1-1 (coding sequence insertion plus a 55-bp deletion) and tubg2-1 mutations (large deletion, as opposed to a 5' untranslated region insertion in tubg2-2).
|
-tubulin depletion during both male and female gametogenesis. A GFP–microtubule binding domain (MBD) marker was introduced into the tubg1-1 tubg2-2 background to have a more dynamic view of microtubule arrays in -tubulin–depleted cells during postembryonic development.
-Tubulin Depletion Affects Gamete Transmission
tubg1-1 tubg2-1 double heterozygote F1 plants (hereafter referred to as F1 plants) had a reduced seed set. Mature siliques of these F1 plants contained 
24.0% desiccated ovules (484/2015), presumably harboring a defective female gametophyte, a figure highly significantly different from the <1% observed in the wild-type control (18/1847). This suggested abortion of most double mutant female gametes since this frequency is very close to the 25% expected in case of complete female gametophytic lethality.
To determine the transmission efficiency of female and male tubg1-1 tubg2-1 gametes, we used F1 plants in reciprocal crosses with wild-type plants. The progeny of these crosses were genotyped by PCR to score the transmission rate of double mutant gametes. When F1 plants were used as female, only 2.1% (5/244) of the progeny carried both tubg1-1 and tubg2-1 mutant alleles, instead of 25% expected for full transmission. When F1 plants were used as a male donor, the double mutant tubg1-1 tubg2-1 gametes represented 9.4% of the gametes transmitted to the progeny (26/276 plants). Therefore, transmission of double mutant gametes was reduced by >90% on the female side and by >60% on the male side, showing that the tubg1-1 tubg2-1 double mutation drastically alters formation and functioning of both male and female haploid gametophytes.
In order to detail such defects, we then followed gametophytic development in tubg1-1 tubg2-1 plants. As tubg1-1 and tubg2-1 mutations are in different genetic backgrounds (in ecotypes Wassilewskija [Ws] and Columbia [Col], respectively), all the following experiments were conducted on F1 hybrid plants [TUBG1/tubg1-1; TUBG2/tubg2-1] to ensure a constant genetic background and compared with Ws/Col F1 hybrid plants as wild-type controls. In such plants, 25% of produced gametes are expected to carry mutations in both genes.
Gametophytic Defects Induced by -Tubulin Depletion
We used a procedure enabling visualization of the fine structure of the embryo sac by confocal microscopy (Christensen et al., 1997, 1998). Using this method, ovules exhibit autofluorescence with nucleoli appearing extremely bright, whereas cytoplasm and nucleoplasm fluoresce moderately, and vacuoles do not autofluoresce (Figure 2
).
|
5% each). Fewer mutant gametophytes contained six (Figure 2G), two, or one nuclei. Nucleoli in these abnormal gametophytes were often larger than in the wild type (compare Figure 2F with Figures 2G to 2J), suggesting a difference in ploidy level. Presence of large nuclei in tubg1-1 tubg2-1 gametophytes was confirmed after clearing of tissues and observation by differential interference contrast microscopy (data not shown).
To determine the stage where gametophytic defects first appear, we compared several stages of ovule development between wild-type and F1 plants, starting from the one-nucleate stage to the mature gametophyte. Gametophyte development is synchronous within one pistil (Christensen et al., 1997), which enabled us to use wild-type gametophytes as internal calibration for developmental stages. At early stages, all ovules of F1 plants were of wild-type appearance. At the eight-nucleate stage, before nuclear migration, 36 abnormal gametophytes were observed from 228 ovules (15.8%) in F1 plants. As above for mature gametophytes, defects included abnormal number, position, and appearance of nuclei.
Therefore, among mutant tubg1-1 tubg2-1 gametophytes (25% of total), 3/5 (15.8%) present detectable morphological defects; the remaining ones, although undistinguishable from the wild type at this stage, are mostly nonfunctional, since genetic analyses had previously showed that >90% of tubg1-1 tubg2-1 female gametes were defective.
The observation of male gametogenesis by differential interference contrast microscopy in anthers of F1 plants revealed that meiosis and tetrad formation was not notably impaired by the tubg1-1 tubg2-1 double mutation. During later stages, however, the fraction of male gametophytes showing abnormalities was significantly higher in F1 plants compared with the wild type (Table 1 ), as revealed by 4',6-diamidino-2-phenylindole (DAPI) staining (Figure 3 ). The most frequent aberrant class with respect to the wild type (Figure 3A) was binucleate pollen, with one nucleus resembling a vegetative nucleus and the other nucleus more condensed and similar to a sperm nucleus (Figures 3B and 3C). Some pollen had two diffusely stained nuclei (Figures 3D and 3E), and very few abnormal pollen had only one diffusely stained nucleus (Figure 3F). As judged by Alexander staining, there is no sign of pollen lethality in anthers of F1 plants (see Supplemental Figure 1 online), and double mutant pollen grains are viable. Moreover, in vitro germination (see Supplemental Figure 1 online) did not reveal any notable difference between F1 and wild-type pollen. Given the transmission rate obtained in genetic analyses, this indicates that, as was the case for the female side, a high proportion of double mutant pollen is morphologically normal and able to produce a pollen tube but not competent for efficient fertilization.
|
|
-tubulin stock. The timing of appearance of defects is variable from one embryo sac to another, which suggests differences in initial amounts and/or in turnover of -tubulin. The same kind of mitotic defects are noted on the male side, although less pronounced, which could be related to the smaller number of divisions during pollen development and/or to a smaller dilution factor of sporophytic proteins due to a smaller cellular volume.
Microtubule Defects during Mutant Pollen Development
Cells of -tubulin mutants isolated in animal, fungus, and yeast exhibit aberrant microtubule organization (Oakley et al., 1990; Horio et al., 1991; Sobel and Snyder, 1995; Sunkel et al., 1995). Defects in nuclear division observed during tubg1-1 tubg2-1 gametogenesis are consistent with abnormalities in microtubule nucleation and organization. We consequently used -tubulin immunolabeling to compare microtubule organization during mutant and wild-type (Figure 4
; see Supplemental Figure 2 online) pollen development, from the first division of meiosis to trinucleate pollen. Observation of pollen in anthers of double heterozygous F1 plants revealed that microtubule organization is undistinguishable from the wild type up to the uninucleate pollen stage. Thereafter, abnormal microtubular mitotic structures (24 out of 78 figures observed) were detected (Figures 4C to 4H), which were never seen in wild-type pollen (Figures 4A, 4B, and 4I; see Supplemental Table 1 online). A common defect was abnormal spindles, which appeared bent or collapsed and associated with unaligned chromosomes along the equatorial plane (compare Figures 4A and 4B with Figures 4C to 4F; see Supplemental Table 1 online). Another defect consisted of dense accumulations of microtubules at the cell's periphery in two-nucleate pollen just exiting division, suggesting that these structures correspond to collapsed phragmoplast (Figures 4G and 4H; see Supplemental Table 1 online). Therefore, the tubg1-1 tubg2-1 mutation clearly affects both spindle and phragmoplast function during pollen development.
|
; Twell et al., 2002).
Sporophytic Lethality Induced by tubg1-1 tubg2-2 Mutations
In contrast with gametophytic defects of the tubg1-1 tubg2-1 combination, the tubg1-1 tubg2-2 mutant combination allows normal gametophytic development and fertilization. Double homozygous mutant seeds develop and germinate normally, producing tubg1-1 tubg2-2 plantlets, which are similar to the wild type up to 3 d after germination. At this stage, double mutant plantlets start to display several morphological defects, eventually leading to lethality after 3 weeks of in vitro culture. The first observable anomaly is a reduction of the cotyledons' expansion, which remain small and distorted (Figure 5B
) with respect to the wild type (Figure 5A). After 7 d of in vitro growth, it becomes apparent that meristem activity is also impaired, as no more than two leaf primordia are initiated at the shoot apex (compare Figures 5D and 5F with 5E, 5G, and 5H). These primordia undergo little growth and form small abnormally shaped leaves (Figure 5H). Consistent with defective leaf formation, the shoot apical meristem region is highly perturbed in the double mutant: the characteristic dome-shaped organization of the wild type (Figure 5D) is absent in tubg1-1 tubg2-2 double mutants (Figure 5E). Rather, the apex consists of a reduced number of large cells, reminiscent of what occurs in mutants affected in the maintenance of stem cell population (Barton and Poethig, 1993; Laux et al., 1996; Moussian et al., 1998). This is an indication for cell division arrest in the entire shoot apex and initiation of cell differentiation.
Similarly, root growth is inhibited after 3 to 4 d of in vitro culture (Figure 6
). In the wild type, the root apex can be subdivided into three zones: the distal division zone, where cells divide and undergo radial expansion; the elongation zone, where considerable longitudinal cell expansion occurs; and the differentiation zone, marked by the emergence of root hairs (Dolan et al., 1993; Sugimoto et al., 2000). In 3- to 4-d-old tubg1-1 tubg2-2 mutants, the overall root organization is conserved, but the sequence of events is perturbed: Compared with cells in the wild-type root tip (Figure 6A), cells start to elongate in the division zone of double mutant seedlings (Figure 6B), resulting in a twofold to threefold increase in cell length. Together with inhibition of root growth, this indicates a premature arrest of cell division in the mutant root meristem. Concomitantly, differentiating cells in the root apex undergo considerable radial expansion, resulting in swelling of the root apex (compare Figures 6C and 6E with 6A and 6D). No more growth occurs following maximal swelling, and root morphology does not evolve past this point. Reflecting the growth arrest in the root apex, the root hair differentiation zone eventually reaches the distal part of the root (Figure 6E).
|
-tubulin depletion on microtubule organization, we studied microtubule organization in tubg1-1 tubg2-2 mutants using a reporter protein (GFP-MBD) (Marc et al., 1998). tubg1-1 tubg2-2 mutants expressing the GFP-MBD reporter were recovered in the progeny of crosses between [tubg1-1/tubg1-1; tubg2-2/TUBG2] or [tubg1-1/TUBG1; tubg2-2/tubg2-2] plants and a GFP-MBD–expressing line (Camilleri et al., 2002). Double mutant progenies carrying the reporter gene were selected and observed by confocal microscopy. During the first 3 d after germination, mitotic arrays, such as preprophase bands, spindles, and phragmoplasts, were properly assembled during mitosis as in the wild type (data not shown). Similarly, cortical microtubule arrays in differentiating cells of roots or hypocotyls were indistinguishable in appearance from the wild type. Therefore, at this stage of development, where no signs of morphological alteration were visible in double mutant plants, the tubg1-1 tubg2-2 combination of alleles did not prevent formation of typical microtubule structures.
Starting at 4 d after germination, microtubule defects become visible in dividing and differentiating cells (Figure 7
). The mitotic activity is reduced by 60% in double mutant root tips at 8 d after germination (see Supplemental Table 2 online), and premature differentiation of cells in the division zone is noted (Figure 6), indicating a severe block of cell division in the root meristem. Time-lapse observations of several dividing cells indicate that duration of the cell cycle in mutant cells is longer and more variable than in the wild type (Figure 8; see Supplemental Figures 3 and 4 online). In these cells, spindles (Figures 7E, 7F, and 8B to 8G) and to a lesser extent phragmoplasts (Figures 7G and 8H to 8J) are strongly perturbed compared with the wild type (Figures 7A to 7D). Condensed microtubule stacks (Figure 7H) are also observable either at the center of the cell or at the cell periphery, likely deriving from collapsed spindles or phragmoplasts. Cortical arrays appear to be less affected (Figure 8; see Supplemental Figures 3 and 4 online).
|
|
-tubulin is required for proper organization of the cortical network during postembryonic development. |
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMETHODSREFERENCES |
|---|
-Tubulin-Tubulin has been shown to be essential for microtubule nucleation at MTOC in all species where it has been analyzed. In this article, we have investigated the consequences of -tubulin deficiency on plant development and microtubule organization in acentrosomal Arabidopsis cells.
The three insertion lines used in this study all display normal phenotypes in terms of gametogenesis, embryogenesis, and further seedling and plant development. However, combining mutations in both -tubulin genes leads to severe developmental defects and, depending on the allelic combination assayed, to gametophytic or sporophytic lethality. This demonstrates that the two Arabidopsis -tubulin genes have overlapping, if not fully redundant functions, as was predictable from their high sequence identity and overlapping expression patterns.
The molecular defects of the tubg1-1 and tubg2-1 alleles (i.e., deletion of almost the entire coding sequence for tubg2-1 and an insertion plus a 55-bp deletion for tubg1-1) likely induce complete loss of gene function in double mutant cells, consistent with the gametophytic lethality observed. Indeed, combination of these two alleles blocks nuclear division at various stages during gametophyte development and results in mainly abnormal, nonfunctional gametophytes. Genetic transmission of double mutant gametes is very poor, especially for the female ones. Nevertheless, a small proportion of double mutant gametophytes is functional and can proceed through fertilization. These variations suggest that a pool of -tubulin from parental sporocytes is carried over into the gametophytes; depending on the amount and/or stability of this parental stock, gametophytes are able to sustain cell division and proceed into their development up to a point where -tubulin reaches a critical level. This critical concentration must be significantly lower than the normal physiological concentration since the -tubulin pool of the meiocyte is sufficient to sustain a few rounds of cell division. The fact that development of the female gametophyte involves three successive mitoses and a larger cellular volume could explain why embryo sac development is more severely affected than the pollen.
The tubg1-1 tubg2-2 combination of alleles induces a weaker phenotype since gametogenesis, embryogenesis, and early seedling development are normal in a double mutant background. However, thereafter, perturbation of cell division and elongation strongly affects seedling development and leads to seedling lethality after 3 weeks. Comparison of phenotypes induced by the two different combinations suggests that tubg2-2 is a leaky allele and likely allows synthesis of some residual -tubulin, undetectable by protein gel blot analysis but sufficient to sustain embryogenesis and early seedling development. Synthesis of residual -tubulin in tubg1-1 tubg2-2 cells likely comes from fusion transcripts generated by the T-DNA insertion.
-Tubulin Deficiency Induces Mitotic Defects
Our results reveal a range of abnormal mitotic microtubular arrays during tubg1-1 tubg2-1 male gametophyte development, including abnormally shaped spindles and collapsed phragmoplasts. This is confirmed by the study of microtubule organization in tubg1-1 tubg2-2 seedlings, where the same types of mitotic defects are seen in meristematic cells. Altogether, these results indicate that -tubulin is necessary during mitosis for proper formation and/or function of the spindle and phragmoplast. Defective spindles are a common feature of -tubulin mutants of several species: S. pombe, S. cerevisiae, and Drosophila. In all cases, spindles do assemble but are generally abnormal in shape and lead to chromosome segregation defects (Horio et al., 1991; Sobel and Snyder, 1995; Sunkel et al., 1995; Marschall et al., 1996; Spang et al., 1996). In A. nidulans, -tubulin depletion completely abolishes nucleation of spindle microtubules (Martin et al., 1997). Whether Arabidopsis -tubulin is required for proper functioning of mitotic arrays, is only involved in microtubule nucleation, or both remains to be determined.
Defective spindles likely result in unequal segregation of chromosomes and formation of abnormally sized nuclei and presumably aneuploid cells, as observed in root tip cells (see Supplemental Figure 3H online). Similarly, spindle misfunction during mutant pollen development, as exemplified by the observation of distorted spindles (Figure 4), likely results in spermatic nuclei with unbalanced chromosome stocks. Such gametic defects are not expected to impair pollen tube growth, which relies on the activity of the vegetative nucleus. Indeed, our observations show that a high proportion of double mutant pollen is morphologically normal and able to produce a pollen tube. However, aneuploid gametes, even if correctly delivered to the embryo sac, are not likely to produce viable fertilization products, a hypothesis consistent with the poor transmission of double mutant male gametes observed in our genetic analyses.
-Tubulin is clearly required for initiation and proper functioning of mitotic spindles, and its depletion can result in complete inhibition of nuclear division and polyploid cells. The abnormally large nuclei observed in female tubg1-1 tubg2-1 gametophytes (Figure 8
) could correspond to such polyploid nuclei that went through multiple cycles of replication without division. Inhibition of nuclear division and increased ploidy have already been reported for -tubulin mutants in A. nidulans (Oakley et al., 1990; Martin et al., 1997; Jung et al., 2001), S. pombe (Paluh et al., 2000), and Drosophila (Sunkel et al., 1995) as well as in S. pombe cells expressing mutant forms of human -tubulin (Hendrickson et al., 2001).
-Tubulin has also been shown to play an essential role in mitotic checkpoint (Hendrickson et al., 2001; Prigozhina et al., 2004), and mutant forms of -tubulin inhibit anaphase A and induce a delay in mitosis, with cells reentering interphase without dividing (Prigozhina et al., 2004). The variation in cell cycle length observed in Arabidopsis tubg1-1 tubg2-2 root tip dividing cells may reflect a similar function for -tubulin in plants (Figure 8; see Supplemental Figures 2 and 3 online).
-Tubulin Deficiency Affects the Organization of Interphase Microtubules
Analysis of the tubg1-1 tubg2-2 combination of alleles indicates that in Arabidopsis, -tubulin is also required for proper organization of cortical interphase microtubules in elongating cells. This finding is in agreement with a role of -tubulin in cortical microtubule nucleation on extant microtubules, as evidenced by Murata et al. (2005), but do not exclude additional effects on polymerization, organization, or dynamics of microtubules. Indeed, in S. pombe and A. nidulans, -tubulin is involved in the spatial arrangement of cytoplasmic microtubules in addition to its documented role in nucleation (Paluh et al., 2000; Jung et al., 2001). Nucleation, dynamics, and spatial organization of microtubules are tightly coupled processes in plant cells and seem all affected in tubg1-1 tubg2-2 mutants. In double mutant cells, we observed fragmentation of cortical microtubules and ultimately their complete disappearance, consistent with defective nucleation and alteration of dynamic properties of microtubules.
The phenotypic syndrome displayed by tubg1-1 tubg2-2 plantlets seems to be a common feature of plants with perturbed microtubule arrays. For example, similar cytoskeletal defects and root swelling are induced both by oryzalin and taxol, despite their opposed effects on microtubule stability (Baskin et al., 1994), and root swelling is also induced by the mor1 mutation (Whittington et al., 2001) as well as in plants partially depleted in -tubulin (Bao et al., 2001). The precise timing of appearance of defects in the tubg1-1 tubg2-2 weak combination at 3 to 4 d after germination is intriguing. At this stage, residual -tubulin synthesis either stops or, in any case, becomes no longer sufficient to sustain growth and development. A similar timing in the appearance of a root swelling phenotype has been reported in Arabidopsis plants engineered for reduced expression of -tubulin genes (Bao et al., 2001). Bao et al. (2001) proposed that the phenotype appears as roots attain a maximal growth rate. The similarity with the mor1 phenotype (Whittington et al., 2001) is in accordance with a role of -tubulin in the control of dynamic properties of cortical microtubules. The mor1 temperature-sensitive mutation affects a plant homologue of the TOGp/XMAP215 family of microtubule-associated proteins. When shifted at restrictive temperature, cortical microtubules in the mor1 mutant first lose their parallel alignment and eventually depolymerize. This correlates with radial swelling of the root (Whittington et al., 2001).
As in other eukaryotes, depletion of -tubulin leads to severe cellular defects and developmental arrests. These defects concerned all microtubule arrays, either mitotic, as exemplified by defects observed during male and female gametogenesis, or interphase cortical arrays, as observed in a less severe, partially viable mutant background. Minute amounts of -tubulin seem sufficient for mitotic divisions since parental sporophytic stocks are sometimes sufficient to sustain formation of fully functional gametophytes.
Surprisingly enough, all Arabidopsis mutant cells observed contained substantial amounts of microtubules, and defects in such cells were rather related to organization and/or functioning of microtubule arrays. In such mutant cells, where most probably no de novo synthesis occurs, most residual -tubulin is presumably titrated out of the cytoplasm and involved in -TuRC complexes. Such nucleation complexes could allow de novo assembly of microtubules to a certain extent, but the dynamic organization of structured arrays may require constant spatial redistribution and turnover of -tubulin between the cytoplasmic pool and the complexed form.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMETHODSREFERENCES |
|---|
) using the Agrobacterium tumefaciens strain MP5-1 carrying the transformation vector pGKB5 (Bouchez et al., 1993). The SALK_004612 (tubg2-1) line was obtained from the ABRC and derives from a T-DNA–mutagenized population of the Col ecotype (Alonso et al., 2003). GFP-MBD–expressing lines have been described previously (Camilleri et al., 2002). Growth conditions are as described by Nacry et al. (1998).
Molecular Techniques
Cloning, sequencing, and DNA gel blot analysis were performed essentially as previously described (Sambrook et al., 1989; Nacry et al., 1998). Plant DNA extraction was performed as described by Nacry et al. (1998). For PCR screening of the Ws insertion lines, primers corresponding to the T-DNA left border (TAG5: 5'-CTACAAATTGCCTTTTCTTATCGA-3') or right border (TAG3: 5'-CTGATACCAGACGTTGCCCGCATAA-3') were used in pair combinations with the following gene-specific primers: LeaderTubG1 (5'-TCCTCACAGTCTCGAAACCC-3') and TubG2F0 (5'-TACAAGTATTGTTAGAGAAG-3'). Genotypes of the Ws mutant lines were determined by PCR using the following primer pairs: LeaderTubG1/TubGR01 (5'-TATAGTGTTGGTCATCCG-3') for TUBG1; LeaderTubG1/TAG3 for insertion in the tubg1-1 line; TubG2F0/TubGR01 for TUBG2; and TubG2F0/TAG5 for insertion in the tubg2-2 line. Primer pairs used for screening of the tubg2-1 line are as follows: TubG2F (5'-CCTCTTCAGGCGTAGTAGTCTCGAAAC-3') and TubGR0 (5'-TGTAGGGCTGGACAACAACGTCACT-3') for TUBG2 and TubG2F/LBb1 (5'-GCGTGGACCGCTTGCTGCAACT-3') for the tubg2-1 line.
The probes used in DNA gel blot analysis of the tubg1-1 and tubg2-2 lines were a 3088-bp PstI-SstI fragment for the left border probe and a 2072-bp SstI-SstI fragment for the right border probe of the pGKB5 T-DNA (Bouchez et al., 1993). The probes used in DNA gel blot analysis of the tubg2-1 insertion locus were a 457-bp PvuII-PvuII fragment (left border probe) and a 1185-bp ApaI-NheI fragment (right border probe) of the pROK2 T-DNA (http://signal.salk.edu/tdna_FAQs.html) and genomic probes amplified from Arabidopsis thaliana genomic DNA with the TUBG2F/TUBG2-7794 (5'-GTCAATCATAACATTCAGAAGTCA-3') and TUBG2-9280 (5'-GAATGTGTTTTTTTTGGG-3')/TUBG2-17954 (5'-TCTATAACGCCACTTAGC-3') primer pairs.
For RT-PCR analysis, single-stranded cDNAs were obtained from total RNA as described by Camilleri et al. (2002). PCR was performed on diluted cDNA samples using the primers LeaderTubG1/TubGRint1 (5'-ACATCTTTTCTATCACCTCCCTGA-3') and TubG2F0/TubGRint1 for amplification of TUBG1 and TUBG2 cDNAs, respectively. APT1 cDNA was used as an internal control (Moffatt et al., 1994) and was amplified using primers APT-RT1 (5'-TCCCAGAATCGCTAAGATTGCC-3')/APT-RT2 (5'-CCTTTCCCTTAAGCTCTG-3').
Protein Extraction and Protein Gel Blot Analysis
To obtain antibodies directed against -tubulin, a tobacco (Nicotiana tabacum) cDNA was subcloned into pQE60 Escherichia coli His-tagged expression vector and protein purified as recommended by the supplier (Qiagen). Rabbit immunization was done according to Evrard et al. (2002) using 200 µg of recombinant protein for each injection. IgGs directed against -tubulin were purified using thiophilic uniflow resin from Clontech. Purified IgGs were concentrated and stored at –20°C in PBS containing 50% glycerol (v/v) and 1% BSA (w/v). Total protein extracts were prepared as described by Liu et al. (1994). The protein concentration of the extracts was determined using Bradford reagent (Sigma-Aldrich). Proteins were separated on a 10% acrylamide gel and transferred to Immobilon-P membranes according to the specification of the manufacturer (Millipore). The membranes were treated with rabbit anti--tubulin antibody in the presence or absence of 20 nM recombinant -tubulin. The secondary antibody used was anti-rabbit IgG antibody linked to horseradish peroxidase (Sigma-Aldrich). Signals were revealed using the ECL system (Amersham).
Anatomical and Indirect Immunofluorescence Microscopy of Arabidopsis Gametophytes
Female gametophyte development was analyzed using a method developed by Christensen et al. (1997) (1998). The 488-nm laser line of an argon laser of a Leica SP2 AOBS confocal laser scanning microscope was used to illuminate and observe female gametophytes. Optical sections of 1 µm were collected with a Leica HCX PL APO x63/1.20 NA water objective and the Leica LCS software.
For the phenotypic analysis of pollen, four to five flowers were brushed on a microscope slide. Pollen grains were observed under a Leica DMRB microscope after Alexander staining (Alexander, 1969). DAPI staining solution (100 µL) (0.1% Nonidet P-40, 10% DMSO, 50 mM PIPES, pH 6.9, 5 mM EGTA, pH 7.5, and 0.4 µg/mL of DAPI) was added to the slide, and the pollen was covered with cover slips. The pollen was viewed by UV epi-illumination using a Leica DMRB microscope. Pollen germination tests were performed as described (Derksen et al., 2002).
Floral apical meristems were excised and were fixed, embedded, and processed for immunofluorescence as described by Baluska et al. (2002). Sections were incubated for 1 h with B-5-1-2 monoclonal anti--tubulin (T5168; Sigma-Aldrich) diluted 1:400 (w/v) in PBS supplemented with 1% BSA (w/v). After rinsing in PBS, sections were incubated for 1 h with the secondary antibody Alexa Fluor 488 goat anti-mouse IgG (A-11017; Molecular Probes) diluted 1:200 in PBS supplemented with 1% (w/v) BSA. DAPI (1 µg/mL) was then applied for 10 min to label DNA. After rinsing in PBS for 10 min, sections were mounted under cover slips and were examined using a Leica SP2 AOBS confocal laser scanning microscope.
Anatomical Analysis of Arabidopsis Seedlings
Histological sections were prepared as follows. After overnight fixation in 4% paraformaldehyde in PBS, pH 7, samples were rinsed for 2 h in PBS and dehydrated in a graded series of 10, 30, 50, 70, 90, and 100% ethanol. Technovit 7100 (Kulzer, Heraeus) infiltration was performed according to the manufacturer's instructions. Sections (5 or 10 µm) were made on a Leica microtome and stained for a few seconds with 0.1% toluidine blue. Sections were mounted in water and photographed on a Leiz microscope. Seedling structure was studied using low-temperature scanning electron microscopy as described by Traas et al. (1995). For FM1-43 staining, living seedlings were stained with 2 µg·mL–1 FM1-43 (Molecular Probes) in water for 10 min and washed twice after staining, and excised roots were mounted in water for observation. Imaging was performed using a Leica TCS-NT confocal laser scanning microscope. GFP-MBD fluorescence was observed on living seedlings mounted in low-melting-point agarose (0.4% in water). Imaging was performed using a Leica TCS-NT confocal laser scanning microscope.
Accession Numbers
Arabidopsis Genome Initiative locus identifiers for the genes mentioned in this article are as follows: At3g61650 (TUBG1), At5g05620 (TUBG2), and At1g27450 (APT1).
Supplemental Data
The following materials are available in the online version of this article.
Supplemental Table 1. Mitotic Defects in Mutant Pollen.
Supplemental Table 2. Mitotic Defects in tubg1-1 tubg2-2 Roots.
Supplemental Figure 1. Pollen Viability from F1 Plants.
Supplemental Figure 2. Microtubule Organization during Male Gametophyte Development.
Supplemental Figure 3. Time-Lapse Observations of Dividing Root Tip Cells of the tubg1-1 tubg2-2 Double Mutant.
Supplemental Figure 4. Time-Lapse Observations of Dividing Cells in the tubg1-1 tubg2-2 Double Mutant Root Tip.
|
Acknowledgments |
|---|
-Tubulin Is Essential for Microtubule Organization and Development in Arabidopsis
