Two Class XI Myosins Function in Organelle Trafficking and Root Hair Development in Arabidopsis thaliana 1

Multigene families encoding class XI myosins are conserved in higher plants, however, little information is available on specific functions of these ubiquitous molecular motors. We isolated gene knockout mutants for all thirteen class XI myosins present in Arabidopsis thaliana genome. Inactivation of 11 myosin genes resulted in no discernible phenotypes under the normal growth conditions. In contrast, the knockouts of the remaining two myosin genes, XI-2 (formerly MYA2) and XI-K, exhibited similar defects in root hair elongation suggesting that the myosin-driven motility plays a significant role in a polar tip growth. Strikingly, inactivation of each of these myosins also reduced trafficking of Golgi stacks, peroxisomes and mitochondria in root hairs and in leaf epidermal cells. These results indicate that myosins XI-K and XI-2 play major and overlapping roles in the cell dynamics in Arabidopsis and highlight the redundant nature of myosin function in plants.

Myosins are signature molecular motors of eukaryotes that are involved in a broad spectrum of actin cytoskeleton-associated types of cellular dynamics (Vale, 2003). Comparative genomics revealed that myosins are conserved throughout the eukaryotic domain of life (Richards and Cavalier-Smith, 2005;Foth et al., 2006). Land plants possess two myosin classes, XI and VIII, each of which is evolutionary related to animal and fungal class V myosins (Desnos et al., 2007) suggesting their origin from a common ancestor which antedated the divergence of Plantae and Opisthokonts (Foth et al., 2006). Subsequent evolution of myosins was dominated by gene duplication and diversification that resulted in the presence of more than 10 myosin genes in all plant genomes sequenced so far (see the accompanying paper). In particular, Arabidopsis thaliana encodes 13 class XI and 4 class VIII myosins (Reddy and Day, 2001). The studies that mostly involved cytoskeletal inhibitors have demonstrated the principal role of actomyosin motilty in plant cell dynamics including organelle trafficking, remodeling, and inheritance (Boevink et al., 1998;Nebenfuhr et al., 1999;Sheahan et al., 2004;Kim et al., 2005;Runions et al., 2006). Some of the class XI myosins were found in association with organelles suggesting their involvement in organelle transport (Wang and Pesacreta, 2004;Hashimoto et al., 2005;Li and Nebenfuhr, 2007;Reisen and Hanson, 2007). However, information on functional profiles of individual myosin motors is very limited. Two recent publications have demonstrated the role of A. thaliana myosin XI-K in root hair growth (Ojangu et al., 2007), and implicated rice myosin XI-B in pollen development (Jiang, 2007).
Here we screen the gene knockouts of all 13 class XI myosins of A. thaliana to show that, in addition to myosin XI-K, myosin XI-2 (MYA2) is also required for root hair development.
Furthermore, we demonstrate that each of these two highly-expressed myosins functions in the rapid movement of the Golgi stacks, peroxisomes and mitochondria in roots and leaves. Interestingly, inactivation of the genes encoding the most closely related paralogs of myosins XI-K and XI-2, myosins XI-1 (MYA1) and XI-B, respectively, did not impair root hair growth or organelle trafficking. These results indicate that evolution of myosins in plants combines opposing tendencies of functional specialization and functional redundancy.

Isolation of the Homozygous Knockout Lines
Sixteen homozygous lines in which each of the 13 class XI myosin genes of A. thaliana was inactivated by T-DNA insertion were obtained and the exact localization of the insert was determined by sequencing ( Fig. 1A and data not shown). Furthermore, inactivation of the target genes was demonstrated using semi-quantitative RT-PCR (e.g., Fig. 1B). The corresponding mRNAs were undetectable by RT-PCR analysis therefore confirming complete abolishment of myosin expression for each of the 16 knockout lines. To ensure that the observed phenotype is attributed solely to the inactivation of the myosin XI-K locus, line SALK_067972 was further backcrossed twice to the wild type plant and the homozygous progeny was selected and used for the experiments described below. In addition, two independent lines for myosin XI-K and three for myosin XI-2 genes were selected for further experiments to ensure that the observed phenotypes were due to the T-DNA insertion in the corresponding locus rather than to secondary site mutations. For both of the obtained independent myosin XI-K gene knockout lines, the lack of an expressed protein was confirmed using a polyclonal antibody specific to this myosin (Fig.   1C).

Inactivation of the Myosins XI-2 and XI-K Induces Similar Defects in Root Hair Growth
Inspection of the insertional lines in which each of the thirteen class XI myosin genes present in A. thaliana genome was inactivated ( Fig. 1 and data not shown) revealed no detectable developmental defects in the aerial organs of plants grown under normal conditions. Therefore, the seeds of each line were planted to vertical plates to screen for root morphology. Although the overall root sizes and shapes appeared normal in all knockout lines, four lines exhibited obvious defects in the elongation of the root hairs ( Fig. 2A). These lines were knockouts of two class XI myosin genes, namely, XI-2 and XI-K. Quantification of the root hair length for each of these lines revealed a dramatic reduction in the mean root hair length that varied from 28% to 40% of that in the parental Columbia line (Fig. 2B). Comparative analysis showed that the differences in the root hair length between each of these four knockout lines and the wild type line were statistically significant with p<0.001 for all lines.
Interestingly, analysis of the light-regulated chloroplast relocalization performed as described in an accompanying work by Avisar et al. revealed that none of the myosin gene knockouts exhibited observable defects in chloroplast movement (data not shown). This result suggested that several myosin motors may have overlapping functions in chloroplast movement.
Alternatively, it is possible that although chloroplast movement requires intact actin cytoskeleton , it does not involve myosins.
To confirm the results of the gene knockout experiments by an independent approach and to examine the usefulness of the dominant negative inhibition of the myosin function that we used for Nicotiana benthamiana in an accompanying paper (Avisar et al., 2007) for Arabidopsis, we generated several transgenic lines that expressed the cargo-binding, globular tale domain (GTD) of myosin XI-K. The HA-epitope tagging of this domain was used to select transgenic lines that exhibited the highest levels of GTD expression (Fig. 2C). Because the mean root hair length in these selected lines was only 38% of that in the control ( Fig. 2A, right-most panel), very similar to that found for myosin XI-K knockout mutants, these data validated the use of the dominant negative inhibition approach in Arabidopsis.
Given that the principal function of myosin motors is physical translocation of various cargoes, our findings suggest that such translocation powered by myosins XI-2 and XI-K is required for the rapid polarized growth of the root hairs.

Myosins XI-2 and XI-K Are Required for the Rapid Organelle Transport in Root Hairs and Leaves
What are the cargoes transported by the myosins XI-2 and XI-K within the growing root hairs? To address this question, we studied trafficking of Golgi stacks and peroxisomes in the gene knockout lines that were transformed to express the fluorophore-tagged, organelle-specific reporters ( Fig. 3 A-C). The mitochondrial transport was examined using a vital fluorescent die Rhodamine 123 (Fig. 3E). Organelle trafficking in the live root hairs was visualized by confocal microscopy and the resulting movies were used for computer-assisted organelle tracking and velocity measurements.
As expected, predominant pattern of Golgi movement was along the root hair longitudinal axis (Fig. 3A). It should be noted, however, that not infrequently the neighboring individual Golgi stacks were moving in the opposite directions or with the drastically distinct velocities (data not shown). This pattern and a mean Golgi velocity of ~1 µm/sec were very similar to those recently described for Arabidopsis root hairs (Campanoni et al., 2007). It seems, therefore, that Golgi stacks move independently of each other rather than in a uniform stream-like pattern.
Strikingly, we found that inactivation of either myosin XI-2 or XI-K resulted in an approximately twofold reduction in the mean velocity of the Golgi stacks and peroxisomes ( Fig.   3D). In the case of mitochondria, myosin XI-K knockout line exhibited more than 3-fold lower velocity than that in the wild type, while less dramatic, ~1.5-fold velocity reduction was observed in the myosin XI-2 knockout line (Fig. 3F). In all cases, differences between the organelle velocities in knockout lines versus parental Columbia line were statistically significant with p<0.001. Taken together, these observations demonstrate that myosins XI-K and XI-2 each make a significant contribution into the rapid transport of Golgi stacks, peroxisomes, and mitochondria in the root hairs. This, however, does not necessarily imply that the defects in root hair growth seen in the corresponding knockout lines can be directly attributed to the slower organelle movement.
To determine whether or not myosins XI-2 and XI-K are required for rapid organelle trafficking in organs other than root hairs, we examined motility of Golgi stacks and peroxisomes in the elongated epidermal cells found along the central vein on the leaf underside (central area in Fig. 4B). These elongated cells are most amenable to observing and measuring organelle movement in leaves. Interestingly, the mean velocity of these organelles in leaf cells was at least 30% greater than that in the root hairs (cf. Figs. 3D and 4D). Inactivation of either myosin XI-2 or XI-K resulted in ~3 to ~5-fold reduction of the mean velocity of Golgi stacks and peroxisomes ( Fig. 4 A, C, D), also a stronger effect compared to that in the root hairs. The dominant negative inhibition of myosin XI-K closely mimicked the effect of the gene knockout, once again confirming the utility of this approach for the study of myosin function in plants ( Collectively, these results demonstrated that two class XI myosins, XI-2 and XI-K, make comparable contributions to rapid trafficking of Golgi stacks and peroxisomes in the Arabidopsis roots and leaves. It seems, however, that myosin XI-K plays more significant role in the translocation of mitochondria than myosin XI-2 and that this difference is more pronounced in leaves than in root hairs.

Hair Growth
Phylogenetic analysis of plant myosins presented in Fig. 1A of an accompanying paper (Avisar et al., 2007) shows that Arabidopsis myosins XI-2 and XI-K each possess closely related paralogs, myosins XI-B and XI-1, respectively, suggesting that the myosins in each paralogous pair may perform similar functions. To test this possibility, we examined root hair growth and trafficking of the mitochondria and peroxisomes in the knockout lines with inactivated myosins XI-B or XI-1. Our analyses revealed no defects in root hair growth in either of the mutant lines ( Fig. 5A). Examination of mitochondria in the root hairs ( Fig. 5 B and C) or leaf epidermal cells (Fig. 5 D,E) showed no significant changes in the mean organelle velocities due to mutational inactivation of either of two myosin paralogs. Only a modest reduction of the peroxisome velocity in leaf epidermal cells was observed in the myosin XI-1 knockout line compared to that in the parental Columbia line (Fig. 5 F and G).
It was also found that transgenic expression of the myosin XI-1 GTD (Fig. 5H) had only moderate negative effects on the root hair growth (Fig. 5A) or peroxisome trafficking (Fig. 5 F and G). Taken together, these observations indicate that myosins XI-B and XI-1 might play only a relatively minor roles in the root hair development and organelle translocation in the presence of intact paralogous myosins XI-2 and XI-K.

DISCUSSION
It is well established that actin cytoskeleton dynamics plays a paramount role in the polarized growth of the root hair cells (Hepler et al., 2001;Carol and Dolan, 2002;Smith and Oppenheimer, 2005;Staiger and Blanchoin, 2006). However, there is only a limited insight into the significance of the myosin-powered motility of vesicles and organelles for the elongation of the root hairs. In fact, only one class XI Arabidopsis myosin, XI-K, was recently implicated into this process (Ojangu et al., 2007), and even in this study, the potential role of this myosin in organelle movement has not been addressed.
Here we describe systematic screening of all 13 class XI myosins of Arabidopsis for their potential functions in the root hair development. Using gene knockout and dominant negative inhibition approaches, we reveal that two class XI myosins, XI-2 and XI-K are essential for the normal elongation of the root hairs. Inactivation of each of these myosins in four distinct insertion lines (Fig. 1) results in very similar phenotypes with a mean root hair length of ca. one third that in the parental Columbia line (Fig. 2). In addition, we investigated the roles of myosins XI-2, XI-K and their most closely related paralogs, myosins XI-B and XI-1 in the trafficking of Golgi stacks, peroxisomes and mitochondria in the root hairs. Conspicuously, we found that myosins XI-2 and XI-K are each required for the rapid movement of all three organelles (Fig. 3).
Inactivation of the corresponding genes reduced the mean velocity of organelles from 1.5 to 3fold depending on the organelle or myosin identity. In contrast, inactivation of the closely related, paralogous myosins XI-B and XI-1, had no detectable effects on either root hair growth or organelle motility (Fig. 5); additional work is needed to identify functions of these myosins.
The observed correlation in the roles of myosins XI-2 and XI-K in root hair growth and organelle trafficking suggests that the latter might be functionally required for the former.
Indeed, the rapid movement of the Golgi stacks, peroxisomes and mitochondria in a 'reverse fountain' manner (Hepler et al., 2001) may be needed to enhance protein secretion and turnover of metabolites and energy in the growing root hair tip. It also seems possible that the vesicular trafficking that delivers building materials for the plasma membrane and cell wall expansion in the tip is powered by myosins XI-K and XI-2. Finally, the possibility that these myosins are also involved in microfilament remodeling or even more general processes required for maintenance of cell metabolism cannot be ruled out. However, because interference with cargo binding by myosin XI-K using transgenic expression of the cognate GTD mimics the phenotype of XI-K gene knockout in defective root hair growth and slow organelle movement (Figs. 2 and 3), both of these effects are likely due to the lack of the proper cargo translocation in both knockout and dominant negative inhibition approaches used to determine myosin XI-K function.
http://www.weigelworld.org/resources) make these myosins unlikely candidates for the major role in this process. In contrast, myosins XI-A, XI-B, XI-C, XI-D, and XI-J are almost exclusively expressed in pollen and therefore are much better candidates for direct involvement in pollen tube growth. Therefore, it seems likely that myosin's XI-2 closest paralog, myosin XI-B, has evolved to mediate organelle and vesicle transport in the pollen tubes.
We found that myosins XI-2 and XI-K are required for the rapid organelle trafficking not only in root hairs, but also in the leaves (Fig. 4). This conclusion resonates with the relatively high levels of these myosins in leaves (http://www.weigelworld.org/resources). In fact, myosins XI-2, XI-K, and XI-1 are the most abundant myosins in the entire Arabidopsis plants.
Surprisingly, this abundance and major roles played by myosins XI-2 and XI-K in organelle movement do not translate into substantial defects in the development of leaves, stems, or flowers in the corresponding gene knockout lines. It appears that the reduction in the leaf organelle velocity in these lines is not critical for leaf development under normal growth conditions. Therefore, perhaps due to their rapid elongation, root hairs are a more sensitive indicator for the defects in organelle trafficking than the leaf cells.
The lack of major developmental defects in two myosin XI-2 knockout lines characterized here and in an additional line described elsewhere (Hashimoto et al., 2005) is in a stark contrast to the earlier paper that reported dwarf growth and flower sterility in a single knockout line (Holweg and Nick, 2004). The reason for this severe phenotype could be the secondary site mutation(s).
Rapid organelle trafficking is a hallmark of plant cell physiology that is traditionally referred to as 'cytoplasmic streaming ' (Hepler et al., 2001;Smith and Oppenheimer, 2005;Taiz and Ziegler, 2006;Shimmen, 2007 1998;Nebenfuhr et al., 1999) demonstrated that individual stacks are moved independently of each other by the actomyosin motility system rather than being passively carried along by indiscriminate cytosol flow (Nebenfuhr and Staehelin, 2001). Our analyses of the trafficking of Golgi, peroxisomes, and mitochondria in N. benthamiana (Avisar et al., 2007) stressed the need to revise the concept of cytoplasmic streaming. The patterns of organelle movement in Arabidopsis (e.g., Fig. 4) are also incompatible with passive 'floating with the flow' and further support introduction of a mechanistically more relevant concept of active organelle translocation defined by the actomyosin transport network.
On a broader scale, this work is relevant to the problem of the multigene families evolution in eukaryotes: to what extent the lineage-specific expansion of these families is due to adaptation as opposed to stochastic gene birth and death processes (Lynch, 2007)? The myosin gene families provide an excellent model to address this problem because they are nearly ubiquitous in eukaryotes. Our phylogenetic analysis (Avisar et al., 2007) suggested that the common ancestor of higher plants possessed five class XI myosins. This number has more than doubled in Arabidopsis, the plant that encodes more class XI myosins than other three plants with completely sequenced genomes. The functional analysis of the Arabidopsis class XI myosins described here revealed a rather complex picture. The lack of major developmental defects in 11 myosin gene knockouts attests to a redundant nature of myosin functions in plants. The functional profiles of myosins XI-2 and XI-K that belong to distinct paralogous families largely overlap: both myosins are involved in the organelle transport and in root hair growth. The only functional specialization apparent between these Arabidopsis myosins is somewhat larger contribution of myosin XI-K in the movement of mitochondria, an effect that is even more pronounced in N. benthamiana (Avisar et al., 2007). More extensive characterization of the myosin functions is required to determine how many and which myosins are essential for plant development.

MATERIALS AND METHODS
Upon request, all novel materials described in this publication will be made available in a timely manner for non-commercial research purposes, subject to the requisite permission from any third-party owners of all or parts of the material. Obtaining any permissions will be the responsibility of the requestor.

Root Hair Morphology
Seeds were surface sterilized and grown on vertical plates containing 0.5 x Murashige and Skoog (MS) medium, 5mM MES pH 5.8, 1% sucrose, 0.6% Phytogel under a 16 hr light/8hr dark cycle. Root hair phenotypes of the 5-day-old seedlings were photographed using Leica MZ6 stereozoom microscope equipped with CCD camera and measured using the Image-Pro (Media Cybernetics, Inc., Bethesda, MD) software. At least two hundred root hairs from four or more individual plants were measured to determine the mean length for each line shown in Figs. 2 and 5.

Arabidopsis Transformation
A Golgi-specific reporter was obtained by fusing rat A-2,6-sialyltransferase (Saint-Jore et al., 2002) with YFP, while fusion between mCherry and a signal peptide of the pumpkin hydroxypyruvate reductase (Mano et al., 2002) was used to tag the peroxisomes. Each of these cDNAs was cloned into binary vector pMDC32 (Curtis and Grossniklaus, 2003). The resulting plasmids were mobilized into A. tumefaciens strain GV3101 and used to transform homozygous myosin gene knockout plant lines SAIL_632_D12 and SALK_067972 by floral dipping. Seeds harvested from treated flowers were surface sterilized and grown under 16 hrs light/8 hrs dark cycles at 22°C. Plant lines exhibiting Hygromycin resistance were selected and grown in soil.
The gene knockout/transgenic lines expressing organelle markers were selected using epifluorescent microscopy. At least two, but normally, 4-6 independent transgenic lines were analyzed for each experimentl design. Analogous protocol was used to generate double transgenic lines expressing either myosin XI-K GTD and Golgi marker or myosin XI-1 GTD and peroxisome marker. Each of the GTD domains was tagged with the triple hemagglutinin epitope at the N-terminus. The borders of GTD domains were from amino acid residue 1,100 to the Cterminus for myosin XI-K (predicted mol. wt ~54 kDa) and from residue 1,082 to the C-terminus for myosin XI-1 (predicted mol. wt ~53 kDa). The expression of GTD in transgenic lines was assayed by immunoblotting using anti-HA monoclonal antibody (Roche) in 1:3000 dilution. At least three independent transgenic lines we used for analyses (Figs. 2C and 5H).

Organelle Trafficking
For visualization of live mitochondria in root hairs seeds were germinated and grown for 5 days on the same medium supplemented with 50nM of Rhodamine 123 (Invitrogen, Eugene OR). A series of 8 consecutive images was acquired using a Zeiss 510 Meta (Zeiss, Thronwood, NY) confocal microscope and used to measure the organelle velocities. Samples were excited using Argon laser at 488nm, emission signal was collected through bandpass 505-530nm filter.
The Golgi stacks and peroxisomes were observed using the following configurations of excitation and emission filters respectively: 488 nm and 508 nm for GFP, 513 nm and 527 nm for YFP, and 587 nm and 610 nm for mCherry. For time-lapse experiments, the consecutive images were taken at 1 sec intervals for mitochondria or 2 sec for Golgi and peroxisomes. For the root hair and leaf epidermis observations, more than 150 and 300 individual organelles, respectively, were traced. Tracking and measurements of velocities of individual organelles was performed using the Volocity3.7.0 Classification software (Improvision, Image Processing and Vision Company, Ltd, Lexington, MA). Statistical analysis of the data was done using t-test and Excel software. Additional details of organelle trafficking analyses are provided in an accompanying work by Avisar et al.

ACKNOWLEDGMENTS
We are grateful to Eugene Koonin for critical reading of the manuscript, to Amit Gal-On for kindly providing lab space to D.A., and to Maria Ivanchenko and Rex Cole for their technical help. The publication was made possible in part by grant 1S10RR107903-01 from the National Institutes of Health. The authors wish to acknowledge the Confocal Microscopy Facility of the OSU Center for Genome Research and Biocomputing. Structure Display Server (http://gsds.cbi.pku.edu.cn). The diagram corresponding to the gene At5g20490 (myosin XI-K) was modified to accommodate corrections in the exon structure (Ojangu et al., 2007). B, semi-quantitative RT-PCR analysis of the insertion line SALK_055785 in which myosin XI-2 gene was inactivated. Primers complementary to sites flanking the insertion site in the myosin XI-2 mRNA (lines under XI-2) or complementary to two regions within the myosin XI-K mRNA (lines under XI-K) were used. Control, the RNA used for analysis was isolated from the parental Columbia line. The expected position of the DNA amplification products for each mRNA are shown by arrows. C, immunoblot analysis of the protein extracts from the control (Col) and myosin XI-K knockout plants using specific polyclonal antibody. Samples from two plants for each line are shown.