Two Arabidopsis Phragmoplast-Associated Kinesins Play a Critical Role in Cytokinesis during Male Gametogenesis

Loss-of-Function Mutations at the Kinesin-12A/B Loci

In order to test whether Kinesin-12A/B are indeed critical for the organization of phragmoplast microtubules, T-DNA insertional mutations were recovered (Figure 1A ). Mutations at either the Kinesin-12A or Kinesin-12B locus do not cause a noticeable defect in the growth and reproduction of the plant (Pan et al., 2004). We reasoned that these genes might function redundantly during cytokinesis. Homozygous double mutants were generated using different alleles of T-DNA insertional mutations (Figure 1A). The kinesin-12a-1 and kinesin-12a-2 mutations had the T-DNA sequence inserted in the 2nd and 12th exons, respectively, and kinesin-12b-1 and kinesin-12b-2 had T-DNA inserted in the 1st and 15th exons, respectively. In contrast with single mutants, the kinesin-12a-1 kinesin-12b-2 and kinesin-12a-2 kinesin-12b-2 double mutants consistently produced significantly fewer seeds in their fruits. The result shown in Figure 1B, as well as other results shown here, were from the kinesin-12a-1 kinesin-12b-2 mutant, unless noted otherwise. Mature siliques of wild-type plants had ∼80% mature seeds. Mature siliques of the double mutants, however, had <40%. The mutant siliques often contained small white structures, which could be either unfertilized ovules or fertilized ovules containing early aborted embryos. While the double mutants produced ∼50% fewer seeds than the wild type, the single mutants produced seeds comparable to those of wild-type plants (Figure 1C). Because the F1 plants resembled their homozygous mutant parent (Figure 1C), we concluded that the incomplete penetrance of the low-fertility phenotype was inheritable.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Mutations at the Kinesin-12A and Kinesin-12B Loci.

(A) Diagrammatical representation of the Kinesin-12A and Kinesin-12B gene structures. Introns are shown as lines, and exons are shown as boxes. Positions of the T-DNA mutational insertions of kinesin-12a-1, kinesin-12a-2, kinesin-12b-1, and kinesin-12b-2 are shown at top of the diagrams.

(B) The homozygous double mutant of kinesin-12a-1 and kinesin-12b-2 produced significantly fewer seeds in siliques. Red arrows point to ovule positions where no seeds were found.

(C) Quantification of seed production in wild-type and mutant siliques. The y axis represents the percentage of ovule positions with seeds. Plants bearing homozygous single mutations at either locus produced similar numbers of seeds as their wild-type counterparts. The homozygous double mutant and its F1 progeny produced ∼50% fewer seeds. Genotypes of the plants are as follows: 12A 12B for plants with wild-type alleles at both Kinesin-12A and Kinesin-12B loci; 12a 12B for plants with a homozygous mutation at the Kinesin-12A locus; 12A 12b for plants with a homozygous mutation at the Kinesin-12B locus; 12a 12b for the homozygous double mutant; and 12a 12b (F1) for the F1 progeny of 12a 12b.

Mutations in the Kinesin-12A/B Genes Resulted in Defective Pollen Grains

Because Arabidopsis is a self-pollinating plant, reduction of seed formation could be due to defects in either the male or female gametophyte, or both, or to defects in embryogenesis. At first, prefertilization and postfertilization ovules of the mutant were cleared for microscopic examination using Normaski optics. No observable defect was detected in either the developing embryo sacs or developing embryos (data not shown). In order to reveal whether gametophytes were responsible for the phenotype, we then performed reciprocal pollination experiments by applying mutant pollen grains onto wild-type stigma, and vice versa. Our results indicated that pollen grains of double mutant plants caused reductions of seed formation in wild-type siliques (average seed number/silique = 7; n = 7) and wild-type pollen grains restored the seed production level in mutant siliques (average seed number/silique = 43; n = 7).

Furthermore, reciprocal crosses were performed to analyze genetic transmission via gametophytes. When pollen grains of the Kinesin-12A/kinesin-12a-1; kinesin-12b-2/kinesin-12b-2 or kinesin-12a-1/kinesin-12a-1; Kinesin-12B/kinesin-12b-2 mutant were used to pollinate the stigma of a wild-type female parent, transmission of the double mutant allele (kinesin-12a-1; kinesin-12b-2) was severely reduced (male transmission efficiency = 12.9 and 20.7%, respectively) (Table 1 ). When these mutants were used as female receivers of the wild-type pollen grains, transmission of the double mutant allele through the female gamete was moderately reduced (female transmission efficiency = 64.3 and 61.7%, respectively) (Table 1). Thus, in the double mutants, the phenotype of reduction of seed formation was mainly caused by defective pollen grains. Because the transmission through the female gamete was not perfect, it was also not ruled out that these kinesins might play a role in multiple phragmoplasts associated with the cellularization of the female gametophyte.

We further analyzed male gametophytes of the double mutant by fluorescence and electron microscopy. The pollen grain is a young male gametophyte in angiosperms. The mature male gametophyte of flowering plants is composed of three haploid cells, of which two sperm cells are suspended in the cytoplasm of the vegetative (pollen tube) cell (McCormick, 2004). Pollen grains collected from open flowers of both wild-type and double mutant plants were then examined by staining with the DNA-specific dye 4′,6-diamidino-2-phenylindole (DAPI). When wild-type flowers were open, pollen grains were already mature, and they had two brightly stained sperm nuclei and a faintly stained vegetative nucleus (Figure 2A , a and b). Pollen grains isolated from mutant flowers often contained two loosely packed DNA masses (Figure 2B, a and b), and they resembled the DNA mass of the vegetative nucleus in wild-type pollen grains. To test whether the abnormal mutant pollen grains produced a generative cell or sperm, we examined them by transmission electron microscopy. In wild-type pollen grains, the vegetative nucleus and sperm (only one revealed in this section) were found in the pollen cytoplasm (Figure 2A, c). The sperm nucleus and cytoplasm were completely isolated from the vegetative cytoplasm by the sperm cell wall (Figure 2A, d). Defective mutant pollen grains, however, lacked sperm. Instead, two identical nuclei were found in single sections (Figure 2B, c). The two nuclei were suspended in the vegetative cytoplasm, and no cell wall–like structure was detected between them (Figure 2B, d).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

The Double Mutant Failed to Produce Male Gametes.

(A) In the wild type (a), a mature pollen grain contains two sperm and one vegetative nucleus. The sperm nuclei (arrows) and the vegetative nucleus (arrowhead) were revealed by DAPI staining. In (b), a differential interference contrast image shows the pollen appearance. Transmission electron microscopy images ([c] and [d]) show the vegetative nucleus (VN) and one sperm cell (arrow) in the pollen cytoplasm. Note that the sperm cytoplasm was physically separated from the pollen cytoplasm by a barrier (arrows in [d]). The other sperm cell was not included in this section. Bars = 10 μm in (b) for (a) and (b), 4 μm in (c), and 1 μm in (d).

(B) In the double mutant, a defective pollen grain failed to produce sperm. DAPI staining (a) showed two loosely packed DNA masses (arrowheads) resembling the vegetative nucleus. A differential interference contrast image of this pollen is shown in (b). Transmission electron microscopy images ([c] and [d]) show two similar nuclei (N) suspended in the pollen cytoplasm. Note that between the nuclear envelopes of the two nuclei (arrowheads), there was no barrier as seen in ([A], [d]). Bars = 10 μm in (b) for (a) and (b), 4 μm in (c), and 1 μm in (d).

(C) Quantification of pollen grains in the categories of two sperm nuclei plus one vegetative nucleus (2+1), two identical nuclei (1+1) or one DNA mass (1), and shrunken appearance (s). The y axis represents the proportion of pollen grains in each category. Pollen grains in the three categories were quantified in the wild type (12A/12A; 12B/12B), single mutants (12a/12a; 12B/12B and 12A/12A; 12b/12b), various heterozygous double mutants (12A/12a; 12B/12b, 12a/12a; 12B/12b, and 12A/12a; 12b/12b), and the homozygous double mutant (12a/12a; 12b/12b).

While many defective pollen grains were consistently detected in the double mutants, very few such pollen grains were observed in the wild type and single mutants. The difference among pollen grains of these different genetic backgrounds was striking when they were classified into three categories: those with two sperm nuclei and one vegetative nucleus (2+1); those with two similar nuclei (1+1) or those with one large DNA mass (1); and those aborted ones with a shrunken appearance (s) (Figure 2C). In the pollen grains with one large DNA mass, the DNA mass likely resulted from the overlap of two indiscernible nuclei by fluorescence microscopy, as sister chromatids segregated normally in the mutants (data not shown). Alternatively, it was also not ruled out that two nuclei might fuse to become one. Nevertheless, both scenarios reflected the failure of spermatogenesis. Ninety-seven percent of wild-type pollen grains contained three nuclei, 3% contained one or two nuclei, and 0% were aborted (n = 103). In the kinesin-12a-1 single mutant, the distribution of pollen grains in these three categories was 96, 3, and 1%, respectively (n = 529). The homozygous kinesin-12b-1 mutant had very similar distribution of different pollen grains, as did the heterozygous Kinesin-12A/kinesin-12a-1; Kinesin-12B/kinesin-12b-2 mutant. In mutants bearing one copy of either Kinesin-12A or Kinesin-12B, defective pollen grains were detected more frequently than in the aforementioned mutants (Figure 2C). In the kinesin-12a-1 kinesin-12b-2 homozygous double mutant, however, only 42% of pollen grains contained three nuclei; 41% of pollen grains were either binucleate or uninucleate, and 17% of pollen grains were shrunken (n = 1825). These data, therefore, suggested that the increased loss of functional Kinesin-12A/B genes was accompanied by the increase of defective pollen grains. It was noticed that mutants bearing one functional copy of either Kinesin-12A or Kinesin-12B produced significantly more than half of normal pollen grains, which could result from possible inheritance of the wild-type product through meiosis. Thus, our data suggested that in the double mutant, the failure of gametogenesis frequently took place due to the failure in cytokinesis, resulting in the absence of the male gamete sperm.

Pollen grains isolated from open flowers of the homozygous double mutant were subject to germination in vitro. It was found that defective pollen grains with two similar nuclei were able to produce pollen tubes as trinucleate pollen grains (data not shown). Thus, a defective pollen grain and resulting tube would fail to reach the ovule, which would reduce the fertility in the double mutant.

Kinesin-12A/B Play a Role in the Organization of Phragmoplast Microtubules

It was hoped that the T-DNA insertions would inactivate Kinesin-12 gene expression in homozygous mutants. The Kinesin-12A and/or Kinesin-12B transcripts were detected in the wild type (12A/12A; 12B/12B) by RT-PCR (Figure 3A ). Moreover, a corresponding transcript was detected in mutants bearing either one or two copies of the wild-type Kinesin-12A or Kinesin-12B gene (12a/12a; 12B/12B, 12A/12A; 12b/12b, 12a/12a; 12B/12b, and 12A/12a; 12b/12b). While the At1g13320 transcript encoding protein phosphatase 2A, as a positive control, was detected in the wild type and all mutants, corresponding transcripts were not detected in homozygous mutants (Figure 3A). Thus, the insertions inactivated the expression of the Kinesin-12A/B genes.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Alteration of At Kinesin-12 Expression by T-DNA Insertional Mutations.

(A) Absence of the Kinesin-12A and/or Kinesin-12B transcripts in single and double homozygous mutants by RT-PCR. The transcripts were detected in the wild type (12A/12A; 12B/12B), and either transcript was detected in mutants bearing either one or two copies of the wild-type Kinesin-12A or Kinesin-12B gene (12a/12a; 12B/12B, 12A/12A; 12b/12b, 12a/12a; 12B/12b, and 12A/12a; 12b/12b). The At1g13320 transcript encoding protein phosphatase 2A (PP2A) was used as a positive control.

(B) Localization of At Kinesin-12A in dividing microspores of the wild type (12A 12B) and the double mutant (12a 12b). The At Kinesin-12A signal is pseudocolored green, and DNA is pseudocolored red. While in the wild type, microspore-specific signals (white arrowheads) were detected between the vegetative nucleus (blue arrows) and the generative nucleus (purple arrows), no such signal was detected in the microspore of the double mutant. The peripheral signal was due to the autofluorescence of the pollen coat. Bars = 5 μm.

To determine the activity of Kinesin-12A/B in the dividing microspores, antibodies raised against the C-terminal domain of Kinesin-12A were used. The antibodies recognize both Kinesin-12A and Kinesin-12B (Pan et al., 2004). In the wild-type microspores, Kinesin-12A/B appeared as discrete signals between the forming vegetative nucleus and the generative nucleus (Figure 3B). No detectable signal was revealed in double mutant microspores at identical stages, indicating that Kinesin-12A/B were absent from the phragmoplast in these mutant cells (Figure 3B).

To further elucidate what had caused the frequent failure of cell division, antitubulin immunofluorescence was performed in microspores of both the wild type and the double mutant. In wild-type microspores, upon the completion of mitosis, microtubules were gradually organized into the typical phragmoplast array between two reforming daughter nuclei located toward one end of the microspore (Figure 4A , a to c). At first, microtubules were organized into an antiparallel array, and a clear dark line was revealed in the middle (Figure 4A, a). Such a dark line marks the plus end of phragmoplast microtubules. The progression of cytokinesis was accompanied by the expansion of the phragmoplast microtubule array toward the periphery and the shortening of microtubules at their ends (Figure 4A, b). Once the cell plate was laid down centrifugally, the phragmoplast microtubule array appeared in a hollow cylinder-like configuration (Figure 4A, c). The cell plate eventually met the plasma membrane of the microspore to render a convex lens–shaped generative cell and a larger vegetative cell (McCormick, 2004).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

Comparisons of Microtubule Organization and Cell Plate Development during the First Mitotic Cell Division in the Microspore.

Microtubules are pseudocolored green, and DNA is pseudocolored blue.

(A) In wild-type microspores (a), upon the completion of mitosis, microtubules were organized into an antiparallel array between two DNA masses. A typical phragmoplast microtubule array (b) had a dark line in the middle. The phragmoplast microtubule array (c) appeared in a barrel-like shape at this late stage of cytokinesis.

(B) In defective microspores of the homozygous double mutant, microtubules failed to be organized into an antiparallel phragmoplast array. Microtubules (a) polymerized into bundles between two DNA masses after mitosis. More microtubules were formed between two reforming nuclei (b), and they did not appear to have a dark line by tubulin immunofluorescence in the middle (c). Aggregates/bundles of microtubules ([d] and [e]) remained to be associated with one nucleus toward the periphery. Bar = 5 μm.

In dividing microspores of the double mutant, however, microtubules were frequently disorganized after mitosis was complete (Figure 4B, a to e). In all microspores upon completion of anaphase, coalesced microtubule bundles were found between the two sister chromatid masses (Figure 4B, a). At later stages, microtubule bundles remained unshortened, and their distal ends toward the reforming daughter (son) nuclei extended at or near the nuclear envelope (Figure 4B, b and c). The antitubulin immunofluorescence often did not reveal a dark line in the center of the microtubule mass (Figure 4B, c). This result suggested that these microtubules had not had their plus ends organized in the middle of the phragmoplast, as was seen in the wild-type cells. In pollen grains collected from open flowers of the homozygous double mutant, microtubule bundles were still detected, and they were associated with the nucleus toward the periphery of the cytoplasm (Figure 4B, d). Microtubule bundles often became randomly organized, with no reminiscence of the phragmoplast array (Figure 4B, e).

Among wild-type microspores, 95% demonstrated normal phragmoplast arrays with mirrored microtubule sets during their first cytokinesis (n = 56). Only 26% of mutant microspores, however, showed similar phragmoplast arrays at comparable cytokinesis stages as wild-type microspores (n = 118).

Mutant Microspores Failed to Form the Cell Plate

Earlier reports on cytokinesis mutants revealed that fragments of cell plate–like structures were still formed even though cytokinesis failed (Lauber et al., 1997). In those mutants, cells undergoing cytokinesis have microtubules organized into the mirrored phragmoplast array. Here, we report that in kinesin-12a kinesin-12b homozygous double mutants, microtubules no longer bear the typical phragmoplast array. Thus, we wanted to examine whether the defective mutant microspores assembled the cell plate when microtubules were disorganized. Immunolocalization of KNOLLE, a syntaxin-like protein localized at the cell plate (Lauber et al., 1997), was performed in both wild-type and mutant microspores. In the wild-type microspore bearing a mature phragmoplast, KNOLLE was densely accumulated at the division site (Figure 5 ). In the mutant microspore bearing aberrantly organized microtubules, KNOLLE had a diffuse localization pattern among microtubule bundles (Figure 5). This result suggests that the accumulation of KNOLLE-bearing vesicles at the division site is dependent on the antiparallel microtubule array of the phragmoplast, in which the plus ends of microtubules face each other in the middle. Hence, in the defective mutant microspores, disorganized microtubules, when their plus ends were not placed in the middle of the phragmoplast, prevented KNOLLE from accumulating at the cell division site.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

The Cell Plate Failed To Be Formed in Defective Microspores.

(A) Localization of the cell plate marker KNOLLE in developing pollen grains of the wild type and the double mutant. In a wild type (12A 12B) microspore, the developing cell plate marked by the syntaxin-like protein KNOLLE (arrowheads) was formed in the middle region of the phragmoplast. A dark midline was clearly seen by tubulin immunofluorescence (arrowheads). In defective microspores of a homozygous double mutant (12a 12b), microtubules failed to be organized into a phragmoplast array with a dark midline. KNOLLE accumulated around microtubules in a diffuse fashion. The peripheral signal was due to the autofluorescence of the pollen coat. Bars = 10 μm.

(B) Callose accumulation in the wild type and the double mutant. In the wild type, callose (small arrowheads), labeled by aniline blue, appeared between the reforming vegetative nucleus (large arrowhead) and the generative nucleus (arrow) stained by DAPI. The completion of cytokinesis left a callose-rich cell plate (arrowheads) separating the cytoplasms of the generative cell and the vegetative cell (top right). In defective pollen grains in the double mutant, callose accumulated as a large aggregate at the cell cortex (small arrowhead), while two identical nuclei (large arrowheads) were positioned away from the aggregate. Such a callose-rich aggregate (arrowhead) was not organized in a cell plate–like configuration at the cell cortex (bottom right). Bar = 10 μm.

Callose deposition in developing pollen grains of the double mutant was also compared with that in wild-type pollen. After mitosis was completed in wild-type microspores, callose was deposited between two reforming nuclei (Figure 5B). Eventually, a complete callose-rich cell plate was formed (Figure 5A). In the double mutant, defective pollen grains bearing two identical nuclei lacked callose deposition between the nuclei (Figure 5B). Instead, callose was detected at the cell cortex as a large aggregate that did not resemble the cell plate (Figure 5B).

The Organization of Phragmoplast Microtubules

Upon the completion of sister chromatid segregation, newly polymerized microtubules coalesce in the spindle midzone and later form the phragmoplast microtubule array (Zhang et al., 1993). Microtubule-associated proteins, known as MAPs, play a regulatory role in microtubule organization (Lloyd and Hussey, 2001; Jürgens, 2005). A group of evolutionarily conserved MAPs with molecular masses of ∼65 kD, known as MAP65/Ase1p/PRC1, have been shown to decorate phragmoplast microtubules (Jiang and Sonobe, 1993; Hussey et al., 2002). Particular isoforms of MAP65s show discrete patterns of localization in the phragmoplast (e.g., preferentially toward the plus or the minus end of phragmoplast microtubules) (Smertenko et al., 2000; Müller et al., 2004; Van Damme et al., 2004). By exclusively decorating the spindle midzone at late anaphase and the phragmoplast midzone, At MAP65-3/PLE contributes to limiting the dimension of the phragmoplast microtubule array, as loss-of-function mutations lead to expansion of the phragmoplast microtubule midzone (Müller et al., 2004). Separately, tobacco (Nicotiana tabacum) MAP65-1 was shown to be a substrate of a phragmoplast-specific mitogen-activated protein kinase cascade, and its phosphorylation is required for the timely depolymerization of phragmoplast microtubules (Sasabe et al., 2006).

The other likely organizing factor of phragmoplast microtubules is the MOR1/GEM1 protein in the XMAP215/Dis1/TOG family (Hussey et al., 2002). The gem1-2 and mor1 mutations at a locus encoding an evolutionarily conserved XMAP125-like MAP also leads to a failure of cytokinesis in the microspore and abnormal cell plate formation in somatic cells (Whittington et al., 2001; Twell et al., 2002; Eleftheriou et al., 2005; Kawamura et al., 2006). But the antiparallel pattern of phragmoplast microtubules is not altered in vegetative cells of the mor1 mutant (Eleftheriou et al., 2005; Kawamura et al., 2006).

While the aforementioned MAPs likely contribute to the general operation of phragmoplast microtubules, their functions may be limited to regulation of the stability and/or dynamics of the microtubules. In other words, unlike Kinesin-12A/B, they are unlikely to contribute to establishing the fundamental pattern of the phragmoplast microtubule array.

Specialized Roles of Distinct Kinesins in Plant Cytokinesis

Besides Kinesin-12A/B, other kinesins have also been found to actively participate in cytokinesis in Arabidopsis and other plants (Lee and Liu, 2004). Those in the Kinesin-5/BIMC subfamily are among the most well characterized. Kinesin-5 in both tobacco and carrot (Daucus carota) cells decorates phragmoplast microtubules with an emphasis toward the plus end and plays a role in microtubule–microtubule sliding (Asada et al., 1997; Barroso et al., 2000). Functions of similar kinesins have yet to be tested in Arabidopsis by genetic means.

Conversely, members in the Kinesin-7 subfamily exhibit spatially and temporally specific localization in the middle region of the phragmoplast. Among them, the NACK1/HIK kinesin interacts physically with a cytokinesis-important mitogen-activated protein kinase cascade to allow a mitogen-activated protein kinase kinase kinase to be targeted to the division site and is required for cytokinesis (Nishihama et al., 2002; Strompen et al., 2002). A similar kinesin, NACK2/TES, is required exclusively for male meiotic cytokinesis (Yang et al., 2003). An identified function of this kinesin–mitogen-activated protein kinase alliance is to downregulate Nt MAP65-1's microtubule-bundling activity by phosphorylation, which is required for the timely execution of cytokinesis in tobacco cells (Sasabe et al., 2006). But the inactivation of the NACK1/HIK kinesin does not alter the pattern of the phragmoplast microtubule array (Nishihama et al., 2002).

Other kinesins have been implicated in the spatial regulation of cytokinesis in Arabidopsis. A novel kinesin, as a cyclin-dependent kinase substrate, localizes to the division site and cortex except for the site once occupied by the preprophase band (Vanstraelen et al., 2006). How this intriguing localization pattern might be linked to phragmoplast operation awaits further examination. In addition, two novel extra-large kinesins, POK1 and POK2, play a redundant role in guiding phragmoplast and cell plate orientation during cytokinesis (Müller et al., 2006). However, unlike At Kinesin-12A/B, these kinesins again do not affect the overall antiparallel organization of phragmoplast microtubules.

Kinesin-12 as a Microtubule Plus End–Associated Motor in the Cell Plate Assembly Matrix

As demonstrated by electron microscopic tomography, antiparallel phragmoplast microtubules do not interdigitate, and plus ends of one group of these microtubules are inserted in an amorphous structure termed the cell plate assembly matrix at the cell division site in dividing somatic cells (Austin et al., 2005). Because a dark line was revealed by antitubulin immunofluorescence in the phragmoplast of dividing microspores (Figure 4A, b and c), the microtubules were probably organized as in somatic cells. It has been postulated that microtubule-stabilizing factors like EB1 could have contributed to capturing the blunt, metastable plus ends of these microtubules within the matrix (Austin et al., 2005). Unfortunately, the composition of the proposed microtubule plus end–capturing complex has not been determined.

Our results have provided evidence that by acting directly at the microtubule plus end, Kinesin-12A/B play a critical role in allowing phragmoplast microtubules to be organized in two mirrored sets with a gap between them (Figure 6A ). We suggest that these motors likely are part of the microtubule plus end–capturing complex in the phragmoplast. However, the mechanism that regulates the temporally specific association of Kinesin-12 with the plus end of phragmoplast microtubules is not clear. The kinesin may be targeted there by interacting with certain anchoring factor(s) residing in the cell plate assembly matrix. Such a targeting mechanism has been reported for the XKLP2 kinesin in Xenopus (Wittmann et al., 2000). Together with such a putative anchoring factor, Kinesin-12A/B can stabilize the plus ends of the phragmoplast microtubules while still allowing tubulin dimers to be added to the plus ends. Kinesin-12 would then drive newly assembled microtubule segments to be translocated in an outward manner by acting as a plus end–directed motor.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6.

Models of the Function of Kinesin-12.

(A) Kinesin-12A/B and their putative anchoring factor(s) form a protein complex that interacts with the plus ends of phragmoplast microtubules located in the middle region. They function in translocating newly polymerized microtubule segments and allow the plus ends to be stably located in the middle region.

(B) The presence of Kinesin-12A/B allows the formation and maintenance of the antiparallel phragmoplast microtubule array. Consequently, successful cytokinesis brings about the cell plate (red), which separates the generative cell from the vegetative cytoplasm. The generative cell undergoes mitosis to produce two sperm cells. The absence of Kinesin-12A/B causes microtubules to be bundled together with mixed polarities. Consequently, materials for cell plate formation do not accumulate in the middle region. Ultimately, two nuclei are suspended in the vegetative cytoplasm. Microtubules are shown in green, and nuclei are shown in blue.

During male gametogenesis, the cytoplasm of the generative cell is physically separated from that of the vegetative cell, as a result of cell plate formation via the phragmoplast (Figure 6B). To date, mutations in genes including TIO, encoding a member of the FUSED kinase family, which acts at the phragmoplast midzone, lead to the failure of cytokinesis after microspore mitosis (Oh et al., 2005). However, it has not been observed that the reported mutations cause a loss of the bipolarity of the phragmoplast microtubule array. Because mutants such as those bearing null tio mutations produce a partial cell plate, one would predict that at least an initial phragmoplast array is established, which might fail to expand. In the absence of Kinesin-12A/B, microtubules frequently fail to be organized into a mirrored phragmoplast array. Consequently, defective microspores failed to produce the generative cell because of the failure of cell plate formation (Figure 6B).

The fact that some microspores still divide normally in the double mutant suggests that there are other factor(s) that play a redundant role like Kinesin-12A/B. In addition, although both kinesins also decorate the plus end of phragmoplast microtubules in somatic cells (Pan et al., 2004), we did not observe any defect in cytokinesis during vegetative growth in the homozygous double mutant. Thus, it further suggests that one or more of the remaining 59 kinesins encoded by the Arabidopsis genome (Reddy and Day, 2001) may play a redundant role like Kinesin-12A/B. We suggest that each of these functionally related kinesins contributes to establishing the phragmoplast microtubule array quantitatively. A qualitative effect is generated when quantitative functions of individual kinesins are combined. In the homozygous double mutant reported here, a significant number of pollen grains were defective despite the fact that no noticeable defect in diploid cells was detected.

In summary, our results demonstrate that the failure of cytokinesis caused by the inactivation of Kinesin-12 resulted directly from the disorganization of phragmoplast microtubules. Such a direct connection between kinesin motors and cell plate formation establishes the significance of proper microtubule organization in plant cytokinesis.

Plant Materials and Growth Conditions

Arabidopsis thaliana plants bearing T-DNA insertion mutations were either the Wassilewskija ecotype (kinesin-12a-1 and kinesin-12a-2) or Columbia (kinesin-12b-1 and kinesin-12b-2). The kinesin-12a-1 and kinesin-12b-1 lines were reported previously (Pan et al., 2004). The kinesin-12a-2 line was recovered from the collection at the Arabidopsis Knockout Facility of the University of Wisconsin Biotechnology Center. The kinesin-12b-2 (SALK_027020) line was recovered from the Sequence-Indexed Library of Insertion Mutations in the Arabidopsis Genome at the Salk Institute Genome Analysis Laboratory. Seedlings were grown under 24 h of light at 22°C and 70% RH. Standard genetic crosses were performed between mutant lines. Progeny from crosses between wild-type Wassilewskija and Columbia plants were used as controls.

PCR-Based Screening

Positive T-DNA insertions were confirmed by a PCR-based method (Krysan et al., 1996). Primers for kinesin-12a-2 were the gene-specific primer YT3 (5′-TACATGTCAGTAAAAGGGTAATGCAATCA-3′) and the T-DNA border–specific primer JL202 (5′-CATTTTATAATAACGCTGCGGACATCTAC-3′) for testing T-DNA insertion and the gene-specific primers 3726F2 (5′-GATGTTTACCACAAGATGAAATTATCAAC-3′) and 3726R (5′-GCTTCTGTAACTAAATTTTCTCCTTCAC-3′) for testing homozygosity. Primers for kinesin-12b-2 were YBTA1 (5′-CTATGGGATTTTGTGGCTCTGC-3′) and the T-DNA border–specific primer LBa1 (5′-ATGGTTCACGTAGTGGGCCATC-3′) for detecting the T-DNA insertion and the gene-specific primers YBTA1 and JPF (5′-TTAGAAGTTTATTGAATCAATGCAGATATG-3′) for testing homozygosity.

RNA Extraction and RT-PCR

Total RNA was isolated from flower buds using PureLink Plant RNA reagent (Invitrogen) as described by the manufacturer. The expression of RNA was detected using PCR amplification of reverse transcription products. Potential genomic DNA contaminants of RNA samples were eliminated by digesting with DNase I before the reverse transcription step. The primers used for RT-PCR were PA1-5 (5′-GCTGGAGAGTTACTTGTTCGG-3′) and PA1-3 (5′-TCCATTGCTGCTCACTACTTG-3′) for Kinesin-12A and PA1L-5 (5′-TGTTCAAGCAGCAGGAGAGTTAC-3′) and PA1L-3 (5′-GCCATAGCATCGTCATTACAAGAAG-3′) for Kinesin-12B. RT-PCR of At1g13320, which encodes a subunit of Ser/Thr protein phosphatase 2A, served as a positive control (Czechowski et al., 2005). After 40 amplification cycles, PCR products were analyzed by gel electrophoresis.

Fluorescence Microscopy

Nuclei in microspores were stained with the dye DAPI according to a published protocol (Park et al., 1998). Immunolocalization experiments were performed according to a published study (Terasaka and Niitsu, 1990). Briefly, developing pollen grains were mechanically released from anthers and fixed with 4% formaldehyde for 1 h at room temperature. Fixed pollen grains were collected by centrifugation at 3000 rpm for 5 min and then digested with 1% Cellulase RS and 1% Pectolyase Y-23 (both from Yakult Honsha) in 50 mM PIPES buffer, pH 5.3, for 2 h with gentle rocking. The pollen grains were then immobilized on poly-l-Lys–coated slides prior to incubation with antibodies. Microtubules were labeled by the DM1A anti-α-tubulin antibody diluted at 1:400 (Sigma-Aldrich); At Kinesin-12A/B were stained by anti-PAKRP1-C at 1:400 (Lee and Liu, 2000); and KNOLLE was stained with the anti-KNOLLE antibodies at 1:100 (Rose Biotechnology). Wide-field fluorescence images were acquired with a CCD camera (Hamamatsu Photonics) using the ImageProPlus 4.0 software package (Media Cybernetics) on an Eclipse E600 microscope equipped with epifluorescence optics (Nikon). Confocal images were collected with a TCS-SP laser scanning confocal microscope (Leica) using argon and krypton lasers. Images were assembled in the Adobe Photoshop 7.0 software package.

Developing pollen grains from wild-type and mutant flowers were stained for callose with aniline blue solution (0.05% [w/v] in 100 mM potassium phosphate buffer, pH 8.5) for 5 min and observed by fluorescence microscopy as described elsewhere (Park and Twell, 2001). Conventional transmission electron microscopy was performed according to a published protocol (Park and Twell, 2001) using the low-viscosity Spurr mini kit (Ted Pella). Samples were observed with a JEM-100S transmission electron microscope (JEOL).

Accession Numbers

The Arabidopsis Genome Initiative locus identifiers for the major genes mentioned in this study are as follows: Kinesin-12A (At4g14150), Kinesin-12B (At3g23670), and protein phosphatase 2A (At1g13320).