|Home | About | Journals | Submit | Contact Us | Français|
While previous studies have shown that microtubules (MTs) are essential for maintaining the highly biased axial growth of the Drosophila bristle, the mechanism for this process has remained vague. We report that the MT minus-end marker, Nod-KHC, accumulates at the bristle tip, suggesting that the MT network in the bristle is organized minus end out. Potential markers for studying the importance of properly polarized MTs to bristle axial growth are Ik2 and Spindle-F (Spn-F), since mutations in spn-F and ik2 affect bristle development. We demonstrate that Spn-F and Ik2 are localized to the bristle tip and that mutations in ik2 and spn-F affect bristle MT and actin organization. Specifically, mutation in ik2 affects polarized bristle MT function. It was previously found that the hook mutant exhibited defects in bristle polarity and that hook is involved in endocytic trafficking. We found that Hook is localized at the bristle tip and that this localization is affected in ik2 mutants, suggesting that the contribution of MTs within the bristle shaft is important for correct endocytic trafficking. Thus, our results show that MTs are organized in a polarized manner within the highly elongated bristle and that this organization is essential for biased bristle axial growth.
Polarized cell growth, manifested as cellular growth biased toward one pole of a cell, is the result of dynamic developmental processes that require an extensive reorganization of the cytoplasm in response to both intracellular and extracellular signals. Essentially, all cells can polarize in response to internal and/or external cues, such as matrix components, cell-cell contacts, or chemical gradients. Eukaryotic cells generally interpret these cues by assembling a polarized actin cytoskeleton at the cortex, which in turn coordinates with microtubules to guide internal membranes. This network ultimately polarizes events that occur internally and at the cell surface (10). A critical issue in this respect concerns how the cytoskeleton responds to those cues that lead to polarized growth.
During development, Drosophila epidermal cells form a variety of polarized structures. These include the epidermal hairs that decorate much of the adult cuticular surface, the shafts of the bristle sense organs, the lateral extensions of the arista, and the larval denticles. These cuticular structures are produced by cytoskeleton-mediated outgrowths of the epiderma (13, 16). Since alterations in bristle morphology are easy to detect in living flies and since small changes in the actin cytoskeleton, as induced by drugs or mutations, often result in an easily detectable phenotype, the growth of the bristle cell is used to define the role of the cytoskeleton in polarized cell growth.
Bristle cells sprout during metamorphosis and elongate over the course of ~18 h. Growth is driven by actin filament polymerization (41). The actin bundles in bristle sprouts begin as microvilli (45) and are cross-bridged into modular bundles 1 to 5 μm in length by at least two cross-linking proteins, forked and fascin (43, 45, 46). These modules are then grafted together by end-to-end joining into stiff bundles (15) which run longitudinally along the bristle shaft, attached to the plasma membrane (40), to support the cell extension as well-spaced ribs. Bundles are tapered, with the largest cross-sectional area of individual bundles found at the base, containing >500 filaments (40). In Drosophila pupae, developing bristles contain 7 to 11 (microchaeta) or 12 to 18 (macrochaeta) bundles of cross-linked actin filaments and a large population of microtubules (MTs) that run longitudinally along the bristle shaft. It was suggested that bristle MTs are highly stable, forming at the start of elongation and then moving out along the shaft as the cell elongates (44). Inhibitor studies suggest that MTs are essential for maintaining bristle axial growth, since injection of microtubule antagonists, such as vinblastine, into pupae resulted in short and fat bristles (13).
It was previously demonstrated that mutations in the Drosophila ikk homologue, ik2, and in the novel gene spindle-F (spn-F), which is not conserved outside insect species, affect both egg chamber polarity and bristle development (1, 37). During oogenesis, both ik2 and spn-F affect mRNA localization due to their effects on actin and MT minus-end organization. Moreover, we were able to show that Ik2 and Spn-F form a complex that regulates cytoskeleton organization during Drosophila oogenesis, with Spn-F serving as the direct regulatory target for Ik2 kinase activity (11). Further evidence for the role of ik2 in cytoskeleton-related processes comes from its interaction with the Drosophila inhibitor of apoptosis 1 (DIAP1). It was suggested that ik2 acts as a negative regulator of F-actin assembly and maintains the fidelity of polarized elongation during cell morphogenesis by modulating DIAP1 levels (22, 29). Recently it was shown that ik2 regulates the dendrite pruning involved in MT disassembly (23).
Since ik2 and spn-F affect bristle polarity organization, we investigated the role of these genes in shaping bristle morphology. We report that MTs within the bristle are organized in a polarized manner, minus-end out. We also demonstrate that both the Spn-F and Ik2 proteins are localized to the bristle tip. Close examination during the bristle elongation period revealed that mutations in either gene affect cytoskeleton organization. Specifically, upon mutation of ik2, the MT minus-end marker is no longer accumulated at the bristle tip. Moreover, we found that the Hook protein is localized at the bristle tip and that such localization is affected in spn-F and ik2 mutants, suggesting that MT functionality within the bristle is essential for recruitment of components of the endocytic trafficking to the tip of the bristle. Thus, we suggest that ik2 and spn-F affect MT functions which are required for the biased axial shape of the bristle. This, in turn, affects the localization of the endocytic trafficking machinery to the bristle tip.
The Oregon-R strain of Drosophila melanogaster was used as the wild type in these studies. The following mutant and transgenic flies were used: spn-F234 (1), pUAS ik2 RNAi (v12485) and pUAS spn-F RNAi (v17015), pUAS ik2K41A (29), pUAS GFP-IK2 (11), Nod-KHC-β-Gal and pUAS KHC-β-Gal (7), pUAS GFP-Nod-KHC (2), and hook11 (21) flies. Bristle expression was induced under the control of the neur-Gal4 driver. Stages of all animals were determined from the point of puparium formation (5). White prepupae were collected and placed on double-sided Scotch tape in a petri dish that was placed in a 25°C incubator, as previously described (44). At the appropriate time of incubation (44 to 46 h, unless indicated otherwise), the petri dish was removed and the pupae were dissected for live imaging or fixation.
To create green fluorescent protein (GFP)-tagged Spn-F as a genomic construct in the CaSpeR4 vector, the GFP-coding sequence was fused by PCR in-frame to the 5′-end region of the spn-F sequence using the internal primers SpnF F (5′-GCTGTACAAGATGGAGGCATCTGCTGCCAAAATCACGCC-3′), EGFP R (5′-GCCTCCATCTTGTACAGCTCGTCCATGCCGAGAGTG-3′), and the external modified primers to create a SacII restriction site at the 5′ end (EGFP-SacII F (5′-CCGCGGATGGTGAGCAAGGGCGAGGAGC-3′)), and a NotI site at the 3′ end (SpnF-NotI Rev [5′-GCGGCCGCCTGGGTCAGAAGTCACC-3′]) of the construct. Two-kb stretches upstream to the spn-F gene were amplified using modified primers to create EcoRI restriction sites at the 5′ end (5′-GAATTCTCCGCTCCTGTCTGCAATGTGG-3′) and 3′ end (5′-GAATTCTCTGCAGTATTGGAGTTCCTTG-3′). Two-kb regions downstream to the spn-F gene were amplified using modified primers to create NotI restriction sites at the 5′ end (5′-GCGGCCGCCCCAGTAGGCGCATTTATTTCGC-3′) and 3′ end (5′-CTCGAGGGTAATGGGTGTGCGCAAGA CC-3′). The inserts were digested with the appropriate restriction enzymes and cloned into the CaSpeR4 vector. P-element-mediated germ line transformation of this construct was carried out according to standard protocols. Ten independent lines from each construct were established.
After removal of the pupal case, pupae were filleted as outlined in detail by Tilney et al. (43). The fillet, which consists of the dorsal surface of the thorax, was placed on its back and cleaned from the large tracheoles and fat bodies, as described by Tilney et al. (44). The isolated thoracic fillets were then taken for live imaging or antibody staining.
Isolated thoracic fillets were placed on a drop of Halocarbon-700 oil (Halocarbon Products, River Edge, NJ) on a 24-by-66-mm cover glass. Confocal images were taken immediately after dissection on an Olympus FV1000 laser-scanning confocal microscope.
Procedures for fixation and staining were described previously (17). Briefly, cut and cleaned thorax was transferred to 4% paraformaldehyde in phosphate-buffered saline (PBS) for 20 min. The tissue was then incubated in 4% paraformaldehyde containing 0.1% Triton X-100 in PBS for an additional 20 min, washed three times in 0.1% Triton X-100 in PBS, and then blocked in 0.1% Triton X-100 containing 4% bovine serum albumin and 0.1% NaN3 for 1 h. The samples were then incubated with a primary antibody in blocking solution overnight at 4°C, washed three times in 0.1% Triton X-100 in PBS, and incubated with secondary antibodies in blocking solution for 2 h at room temperature. The samples were washed three times in 0.1% Triton X-100 in PBS and incubated with phalloidin in PBS overnight at 4°C, washed three times in 0.1% Triton X-100 in PBS, and then placed on a slide and mounted in Citifluor glycerol (Ted Pella, Redding, CA). A coverslip was applied, and the preparation was sealed with nail polish. Confocal images were taken on an Olympus FV1000 laser-scanning confocal microscope and are presented as Z-projections of several optical sections that collectively cover the bristle diameter.
Primary antibodies used here were monoclonal mouse anti-Spn-F (1:10; clone 8C10) (1), monoclonal mouse anti-α-tubulin (1:250; Sigma), monoclonal rabbit anti-β-galactosidase (β-Gal) (1:250; Promega), monoclonal mouse anti-γ-tubulin (1:250; Sigma), and rabbit antihook (20) and monoclonal mouse anticentrosomin (1:250) antibodies, the latter kindly provided by T. Kaufman (18). The secondary antibodies used were Cy3-conjungated goat anti-mouse (1:100; Jackson Immunoresearch) and Alexa Fluor 488-conjugated goat anti-rabbit (1:100; Molecular Probes) antibodies. For actin staining, we used Oregon Green 488- or Alexa Fluor 568-conjugated phalloidin (1:250; Molecular Probes).
The three-dimensional (3D) animated confocal projections (shown in Movies S1 and S2 in the supplemental material) were generated using the Bitplane Imaris (version 5.5.3) software program.
To analyze the fluorescent signal intensity of the various markers used in this study, we used the Olympus FV10-ASW (version 1.7a) software program. We generated an intensity projection over the z axis. We then drew a free area region of interest (ROI) that surrounds the bristle borders as they appear using all channels available and generated a fluorescent signal intensity chart of this area while presenting only the relevant channel in the chart.
To quantitatively analyze Spn-F and Ik2 protein localization within the bristle shaft, we used the Olympus FV10-ASW (version 1.7a) software program. We generated an intensity projection over the z axis and then drew two free area ROIs. The first surrounded the bristle borders, while the second surrounded the bristle tip. We defined the tip region as the area that ranged from the most distal part of the shaft toward the area in which the Spn-F or Ik2 intensity measurement decreased 5-fold. We then used the software colocalization option on each ROI to calculate Pearson's correlation coefficient. We repeated this measurement on 12 different bristles from 4 different pupae and found significantly high correlation of the intensity distribution between Spn-F and Ik2 signals. For each ROI, we then measured the total value of intensity in the region specified and the area of each ROI. To assess Spn-F and Ik2 proportional tip localization, we calculated the ratio of intensity between the tip area and the shaft area and the proportion between the total value of intensity at the tip and the total intensity in the entire shaft.
Adult Drosophila flies were fixed and dehydrated by immersion in increasing concentrations of alcohol (25%, 50%, 75%, and 2× 100%, each for 10 min). The samples were then completely dehydrated using increasing concentrations of hexamethyldisilazane (HMDS) in alcohol (50%, 75%, and 2× 100%, each for 2 h), air-dried overnight, placed on stubs, coated with gold, and examined with a scanning electron microscope (Jeol model JSM-5610LV). Length measurements of adult bristles were made using the Image J (version 1.40j) software program. To test for differences in bristle length between the wild type and the different mutants, we used one-way analysis of variance (ANOVA) followed by the Bonferroni post hoc test. To test if the bristle traits are expressed differently between wild-type flies and each of the mutants, we used the G test of independence.
Since MT polarization in bristles had yet to be defined, we examined the localization of the MT minus ends within Drosophila bristles using MT minus-end reporter, the Nod-KHC-β-Gal fusion protein (7), a chimeric protein comprised of the Nod motor domain fused to the kinesin 1 (Kin1) coiled-coiled domain and β-galactosidase (7). Although full-length Nod preferentially binds MT plus ends (8), the Nod-KHC-β-Gal chimera localizes to the MT minus end in multiple cell types and hence is commonly used as a marker of this end of the filament (7). Immunolocalization analysis using antibodies to β-galactosidase assigned this reporter to the bristle tip (Fig. (Fig.1A).1A). Since Nod-KHC-β-Gal stock is an enhancer trap line, we decided to verify our findings by overexpressing GFP-Nod-KHC, also shown to be a minus-end reporter in Drosophila neurons (2), in the bristle using the Gal4/UAS expression system. We found, as with Nod-KHC-β-Gal, that the GFP-fused marker is also enriched at the bristle tip (Fig. (Fig.1B),1B), again suggesting MTs to be polarized within the bristle. To further analyze MT plus-end polarity within the bristle, we used kinesin-β-Gal as a marker to report the location of MT plus ends (7) and found that this reporter protein was localized at the bristle base in the cell nucleus (Fig. (Fig.1C).1C). A similar localization pattern was obtained using another plus-end marker, GFP-Nod (data not shown) (8). Our results thus demonstrate, for the first time, that bristle MTs are uniquely polarized, with minus-end-out polarity.
Since we found that MTs are polarized within the bristle shaft, with their minus ends directed toward the bristle tip, we sought to localize the microtubule organizing center (MTOC) using two known centrosome markers, namely, γ-tubulin, thought to be the primary source of microtubule-nucleating activity (47), and centrosomin (Cnn) (25). Using antibodies directed against both proteins, we were unable to detect any MTOCs in the cell body or in the bristle shaft during the elongation period (data not shown). These results are not surprising, since it was recently reported that interphase Drosophila cells rely on a distinctive “canonical” centrosome functional cycle, where interphase centrosomes are inactivated and functionally replaced by an alternative mechanism for interphase MT array organization (33).
To further understand the importance of MT polarity in bristle development, we focused on investigating the roles of the spn-F and ik2 genes, previously found to affect MT polarity during oogenesis (1, 11). Mutations in these genes also resulted in bristle development defects (1, 37). Accordingly, we carefully analyzed defects in bristle morphology associated with mutations in these genes. Drosophila bristles are normally highly polarized and elongated structures that sprout from the cell body and grow in a single direction (40). Bristles exhibit a large, almost circular diameter at the base and taper along the bristle shaft into the bristle tip. Detailed studies have demonstrated how these massive structures initiate (9, 41) and elongate (13, 15, 40, 44) while being self-supported by massive, polarized, and well-cross-linked actin filament bundles.
Since ik2 is an essential gene, severe loss-of-function alleles are lethal (37). With this in mind, we reduced ik2 activity in a tissue-specific manner by either overexpressing a dominant-negative form of ik2 (29) or overexpressing ik2 RNA interference (RNAi) constructs (29) in bristle cells. Both approaches resulted in phenotypes that were similar to ik2 loss-of-function alleles. We and others (1, 37) have previously reported that spn-F and ik2 mutant bristles are shorter and thicker than the wild-type structures (Fig. (Fig.2).2). We measured the lengths of adult anterior and posterior scutellar macrochaetae and found that spnF and ik2 bristles are significantly shorter than wild-type bristles (Table (Table1),1), as tested by one-way ANOVA followed by a Bonferroni post hoc test. No significance differences were found between the different mutants.
The external cuticular structure of ridges and valleys viewed by scanning electron microscopy (SEM) reflects the internal organization of actin bundles during bristle elongation (30, 40). The valleys which decorate the adult bristle represent actin bundles that were bound to the cell membrane during bristle elongation, while the ridges represent places where the cytoplasm projected between these actin bundles prior to chitin deposition (30, 40). Due to this correlation between actin bundle organization and the final shape of the adult bristle, the characteristics of the underlying actin bundles during development (early to late) can be deduced by examining external bristle structure (base to tip). We analyzed the arrangement of spn-F and ik2 scutellar bristles based on the following three categories. The first is focused on the overall shape of the shaft, as manifested in the change of its diameter along the bristle length. We found that only 20% to 29% of the mutant bristles had wild-type-like morphology (Fig. (Fig.3A;3A; Table Table2),2), in which the shaft diameter is the widest at the bristle base and tapered toward the tip. In most of the mutant fly bristles (Table (Table2),2), the widest shaft diameter is not found at the base of the bristle (Fig. (Fig.3B3B and Table Table2).2). Importantly, using SEM analysis, we could not distinguish whether the bristle widening we observed was due to increased shaft diameter or due to a change in the circular shaft shape that could result in a flattened wide shape.
In the second category, we examined alterations in bristle growth direction. We observed that 59% to 75% of the mutant bristles exhibited an altered growth direction that was not axially biased (Fig. 3C and D and Table Table2).2). In the third category, we examined the bristle tip shape and found that 24% to 60% of the mutant bristles terminate in several minitips (Fig. (Fig.3F)3F) instead of a single tip (Fig. (Fig.3E3E and Table Table2).2). The expression of all the above bristle traits varied significantly between each of the mutants and the wild type (Table (Table2),2), as analyzed by a G test of independence. However, no significance difference in bristle traits between the different mutants could be detected. These results suggest that in both spn-F and ik2 mutants, the polarized bristle organization is severally affected.
Closer examination of the basal third of spn-F and ik2 mutant macrochaetae revealed a conventional morphology, suggesting that initial bristle formation and growth proceed normally in the mutants. The middle third of the mutant bristles showed moderately disorganized ridges that are shallower and more numerous (Fig. 2E and H) than the wild-type bristle surface, which is characterized by straight ridges and valleys (Fig. (Fig.2B),2B), suggesting that alterations in actin organization had occurred in the mutants during the middle phase of bristle elongation. Overall, adult spn-F and ik2 mutant bristles show problems in cytoskeleton organization during the growth stage.
We have shown that Spn-F and Ik2 localize within the Drosophila oocyte in a polarized manner, with both proteins accumulating at the anterior end in stage 8 to 9 egg chambers (1, 11). To address whether Spn-F in bristles also localizes in a polarized fashion, we dissected pupae at various times during the 16-h elongation period (44) and assayed for protein localization by immunolocalization after fixation. We found that Spn-F is always localized to the tip throughout the elongation period (Fig. 4A and B, representing two time points during this period). To verify these results, we examined developing bristles using a GFP-Spn-F fusion protein expressed under the control of the endogenous spn-F promoter. This protein was able to rescue all mutant-induced spn-F ovarian and bristle defects (data not shown). Similar to the antibody staining pattern observed, confocal imaging of live specimens showed that most of the GFP-Spn-F fusion protein accumulated at the bristle tip (Fig. (Fig.4C),4C), as shown in fixed samples. Moreover, live imaging allowed us to detect a subcompartment of the tip that could not be detected in fixed specimens, in the shape of a single centered extension initiating from the bristle tip (Fig. (Fig.4D),4D), decorated with the GFP-Spn-F protein. This extension is probably sensitive to the fixation protocol used in this study and could represent the elongating part of the tip.
Given our previous demonstration (11) that Spn-F and Ik2 colocalized at the anterior end of the Drosophila stage 8 to 9 egg chamber, we addressed the localization of Ik2 within the bristle. However, since the appropriate antibodies (i.e., anti-Ik2 antibodies) are not available, we expressed a GFP-Ik2 protein fusion using the Gal4/UAS system (12), employing the neur-GAL4 driver. Like Spn-F, the GFP-tagged Ik2 protein was localized to the bristle tip (Fig. (Fig.4E4E).
Next, we examined whether Spn-F and Ik2 colocalized. Here we employed immunolocalization to detect endogenous Spn-F in bristles expressing the GFP-Ik2 chimera and found that Spn-F is completely colocalized with Ik2 in the bristle tip during bristle elongation (Fig. (Fig.5).5). We quantitatively analyzed the colocalization pattern of Spn-F and Ik2 in bristles from 37-h-old pupae, representing an early elongation stage, and a 44-h-old pupa, representing a late elongation stage, by measuring Pearson's correlation coefficient (26). At the early stage of bristle elongation (Fig. 5A to C), we found a highly significant correlation of intensity distribution (Pearson's correlation coefficient = 0.968). The same significant correlation was also observed in later stages of bristle elongation (Fig. 5D to F) (Pearson's correlation coefficient = 0.961). Then, we calculated the proportional fraction of each of the proteins at the bristle tip. We found that in both stages of development, the two proteins accumulated at the bristle tip in similar proportions. At the early elongation stage (Fig. 5A to C), 20.52% of Spn-F and 21.73% of Ik2 were found at the bristle tip in an area that covers only 5.35% of the bristle shaft. At the later stage of elongation (Fig. 5D to F), 32.22% of Spn-F and 38.92% of Ik2 were found at the bristle tip in an area that covers only 5.66% of the bristle shaft. The same pattern of colocalization, as described above, was found in all other analyzed bristles.
Finally, since our previous results (11) showed that Spn-F localization to the anterior end of the oocyte is affected in ik2 mutants, we analyzed whether Spn-F localization at the bristle tip is also dependent on ik2. Immunolocalization of Spn-F in bristles expressing ik2-RNAi constructs revealed that Spn-F is no longer localized to the bristle tip but is instead aberrantly localized in clumps along the bristle shaft (Fig. 6A and B; see also Movie S1 in the supplemental material). Similar results were obtained with bristles expressing a dominant-negative form of ik2 (data not shown). Thus, the polarized localization of Spn-F in the elongating bristle is dependent on Ik2.
Since spn-F and ik2 mutant adult bristles exhibit a range of phenotypes that appear to be cytoskeleton related, we analyzed actin cytoskeleton organization during the development of these mutants. Phalloidin staining of actin filaments during the elongation phase revealed that in both spn-F and ik2 mutants, actin bundles were less organized than were those in the wild type (Fig. 7A and D). A 3D animated confocal projection of an ik2 mutant bristle (see Movie S2 in the supplemental material) demonstrates that although the actin bundles within the shaft appear intact and continuous, they were not restricted to the shaft periphery. The actin bundles are loosely oriented within the shaft such that a single actin bundle can be traced as it goes in various directions. This can also be visualized by examining digital cross sections through the bristle shaft in which the actin bundles are not restricted to the shaft membrane as in the wild type (Fig. (Fig.7C,7C, inset) but instead are distributed throughout the bristle cytoplasm (Fig. (Fig.7F,7F, inset). Mutant tips appeared misshapen and did not taper in a straight direction as in the wild type. In some cases, it appeared as if the actin bundles had even aligned opposite to the normal growth direction (Fig. 7D and G).
MTs are a major component of the bristle cytoplasm and run longitudinally along the core of the bristle shaft (44). Since previous findings (1, 37) implicated both spn-F and ik2 in MT organization during oogenesis, we tested whether the same mutations affect MT organization in bristles. Immunostaining of developing bristles with anti-α-tubulin antibodies revealed that whereas in wild-type bristles the MTs appear as thick stripes that probably represent the superposition of many MTs (Fig. 7B and J), bristles with reduced levels of ik2 contained extremely disorganized MTs (Fig. 7E, H, and K). These often appeared as aggregates, found at various locations along the bristle shaft (Fig. 7H and K).
In spn-F and ik2 mutants, there are defects in the organization of the MT minus end within stage 8 to 9 egg chambers (1, 11). Thus, we decided to analyze whether ik2 function is also required for bristle MT polarization. For this purpose, we analyzed the polarization of MTs in the ik2 mutant using the Nod-KHC-β-Gal and GFP-Nod-KHC reporters (2, 7). We found that in bristles expressing ik2 in a dominant-negative manner, both minus-end reporters failed to accumulate at the bristle tip (Fig. 8B and C). These results demonstrate that in similarity to its effect on MT organization in the oocyte, ik2 also plays a role in MT polarization in bristle cells.
It was previously reported that hook mutant bristles exhibit sharp bends and aberrant tip morphology (20). Such defects suggest that bristle polarized organization is affected in hook mutants. Also, it was shown that hook is required for endocytic trafficking in both the Drosophila eye and nervous system (20, 39, 36, 28). Several studies have shown that MTs are required for endocytic trafficking in polarized cell growth (reviewed in reference 3). Considering both the bristle defects in hook mutants and the known role of the protein in endocytic trafficking, hook appears to be a good candidate reporter for better understanding the contribution of MT polarity to bristle shape. Accordingly, we first studied the localization of the Hook protein in developing bristles using antibodies specific for Hook (20). We found that the Hook protein is localized at the bristle tip (Fig. (Fig.9A).9A). In ik2 mutants, Hook is no longer found at the bristle tip but rather is aberrantly localized in the middle of the shaft (Fig. (Fig.9B).9B). Similar defects in Hook localization were obtained in spn-F mutant bristles (data not shown). Thus, our results demonstrate that spn-F and ik2 are required for Hook localization at the bristle tip.
MTs are well known for their roles in polarized cell growth, maintaining the steady-state positions of cellular organelles, including the nucleus, Golgi apparatus, endoplasmic reticulum, mitochondria, and lysosomes and mediating the polarized transport of vesicles, organelles, protein assemblies, mRNA, and other cytoskeletal elements (6). The Drosophila mechanosensory bristles serve as a model tissue for understanding cytoskeleton-related processes during polarized cell growth. Previously it was shown that in Drosophila bristles, the MT network is essential for biased axial growth (13). In this study, we focused on understanding the constitution of the MT network in the elongating bristle. To study the polarized organization of the MT within the bristle, we employed several MT polarity markers. We found that the minus-end marker, Nod-KHC, accumulates at the tip of the bristle. We also showed that kinesin-β-Gal, an MT plus-end reporter, is not found at the bristle shaft but instead localized to the bristle cell nucleus. We believe that this nuclear localization does not reflect that the microtubule plus ends reach into the nucleus but rather refers to the function of this motor protein within the nucleus, as suggested for other kinesins of this family (8). While this plus-end reporter construct has been reported to accumulate in regions of high density of plus ends of microtubules, such as in midstage oocytes of Drosophila (52), it does not necessarily mark all plus ends of microtubules, which may be found in different arrangements inside various cells. The localization of the minus-end marker to the bristle tip, as well as the absence of the plus-end marker from the bristle shaft, suggests a minus-end-out orientation as was described for Drosophila dendrites (2, 7, 24, 34, 35, 38, 48, 49). It was shown that MTs have opposite orientations in axons and dendrites of Drosophila neurons, with all axonal microtubules having a plus-end-out orientation and dendritic microtubules being oriented with their minus ends distal to the cell body. Thus, our interpretation is that the overall organization of MT polarity in the developing bristle is minus end out.
It was suggested that mutations in spn-F and ik2 affect MT minus-end organization in the oocyte (1, 11, 37). Since these genes also affect bristle morphology (1, 37), we decided to analyze their roles in MT organization within developing bristles. Examination of cytoskeleton organization during bristle elongation in ik2 and spn-F mutants showed that both the actin and MT networks are disorganized. SEM analysis of adult bristles and actin staining of developing bristles showed that in spn-F and ik2 mutants, there are actin bundles that are not restricted to the shaft periphery and the actin modules are poorly oriented, resulting in adult actin ridges that run in different directions. These phenotypes could be the outcome of defects in maintaining actin stability, as was suggested for other actin-regulating proteins (14). It was suggested that the actin bundles, which drive bristle elongation, are formed as filaments being cross-linked and hexagonally packed (43, 45, 46). Stability of actin filaments increases as a result of filament cross-bridging into bundles and/or attachment of the filaments to the plasma membrane. These stabilized actin filaments turn over at a lower rate, such that eventually only bundles that are packed and attach to the shaft membrane prevail (42). Since in ik2 and spn-F mutant bristles we observed actin bundles that are not restricted to the shaft membrane, we can reason that such bundles are stabilized regardless of the fact that they are not attached to the shaft membrane. It was previously reported that ik2 acts as a negative regulator of F-actin assembly and maintains the fidelity of polarized elongation during cell morphogenesis by regulating the function of DIAP1 (22, 29). According to that study, with low levels of ik2 one would expect an increase in actin filament stability that may lead to extra actin bundles that are not attached to the bristle membrane, thus enabling them to run in different direction within the shaft.
As mentioned above, the MT networks in ik2 and spn-F mutants are also disorganized. When staining for α-tubulin was carried out, we found that ik2 and spn-F bristle mutants contained disorganized MTs, which often appeared as aggregates in various locations along the bristle shaft. We also found that Spn-F and Ik2 are located at the bristle tip, where the MT minus-end marker accumulates. In the ik2 mutant, Spn-F is mislocalized and the MT minus-end marker failed to accumulate at the bristle tip. In a previous study, we reported similar results for Spn-F localization in wild-type and ik2 mutant oocytes (11). These results collectively suggest that defects in MT functionality could lead to the morphology phenotypes of the mutant adult bristle.
What could be the possible pathways by which ik2 and spn-F affect MT function in the developing bristle? It was suggested that the majority of bristle MTs are formed early during bristle development and are than dispersed along the shaft as the bristle elongates (44). Any disturbance of this process has the potential to affect MT-dependent processes that transpire during bristle development, especially processes that rely on the polar architecture of the MTs. One possible explanation for the way ik2 and spn-F affect MT function could be that these phenotypes are due to defects in the regulation of bristle microtubule orientation. Our results suggest that the MTs within the bristle are oriented minus ends out. Thus, it is possible that in the wild-type bristle, MT filaments slide into the bristle with their minus end distal to the bristle tip. Therefore, spn-F and ik2 may assume roles ensuring that the correct MT orientation occurs during the sliding process. A second possible explanation is that the defects in MT function could arise from effects on MT stability. The axonal sliding-filaments model suggests that for MTs to be mobilized, they should be severed into shorter mobile polymers (4) in a process called cut and run. Recently it was reported that ik2 is required for dendrite pruning of sensory neurons during metamorphosis, probably by regulating MT stability at the dendrite severing sites (23). Thus, it could be that in ik2 mutant bristles, the MT filaments fail to slide properly due to defects in MT stability in turn affecting MT functionality.
What could be the role of spn-F and ik2 in shaping the Drosophila bristle? Our results point to three possible explanations for the defects in organization of both actin and MT networks in ik2 and spn-F mutants. In the first, ik2 and spn-F play direct roles in coordinating interactions between MT minus ends and actin. Indeed, mediation of this interaction was previously suggested as a role for ik2 during oogenesis (37). Another possible explanation is that ik2 and spn-F affect MT function. As a secondary effect, the actin bundles that were supposed to run parallel along the bristle shaft are misoriented in the mutants. A third explanation is that spn-F and ik2 affect actin in a manner that leads to defects in the functionality of the MT network. Fei et al. (13) demonstrated that MTs are important for the biased axial growth of bristles through the injection of vinblastine before bristle initiation. The introduction of vinblastine, which binds to tubulin and inhibits microtubule formation, resulted in short and fat ocellar bristles, suggesting that MTs play an important role in biased axial growth of bristles (13). On the other hand, the introduction of actin depolymerization drugs resulted in shorter and thinner bristles, suggesting that actin is required for specific axial growth (13). In their study, Fei et al. (13) noted the lack of available known mutations that produce bristles with morphological defects equivalent to those induced by the injection of vinblastine or colchicine (13). Our results provide the missing genetic data to support the role of MTs in biased axial growth according to this hypothesis. Both spn-F and ik2 mutant bristles are much shorter and thicker than wild-type bristles and exhibit defects in the bristle biased axial growth similar to the defects induced by drugs affecting MT. Although we could not exclude the role of spn-F and ik2 in actin and MT coordination, our data suggest that Spn-F and Ik2 act as part of a complex that regulates MT function in the bristles.
What could be the contribution of MT polarity to bristle axial growth? Previous studies have shown that MTs are important for endocytic trafficking in polarized cell growth (reviewed in references 3 and 27). It was suggested that in polarized cells, endocytosis underlies polarized delivery of lipids and membrane proteins (32). Several lines of evidences suggest that endocytic trafficking plays an important role in bristle development. Mutations in several members of the Rab protein family known to mediate endocytic intracellular vesicle trafficking lead to defects in bristle development (19, 31, 50, 51). Specifically, mutations in rab6 and rab11 lead to considerably shortened bristles (19, 31, 50), whereas a rab35 mutant expressed bristles presenting sharp bends, kinks, and forks (51), suggesting several different roles of endocytic trafficking members during bristle development. Previously it was shown that Drosophila hook is required for endocytic trafficking both in the eye and in the nervous system (20, 28, 34, 39). Early on, it was shown that hook is required for the accumulation of internalized transmembrane ligands in multivesicular bodies (MVBs) (20). Later, it was demonstrated that hook is required for normal maturation of multivesicular endosomes (39). It was also suggested that hook function in later stages of endocytosis is essential for regulating synaptic plasma membrane composition but not synaptic vesicle recycling. hook mutant bristles have defects in polarity and shape but are not substantially shorter than wild-type bristles (20), suggesting that hook is not required for bristle growth but probably plays a role in bristle polarized morphology. Since the most prominent features in hook mutants are their defects in polarized bristle shape (20), we decided to use the Hook protein as a proxy for testing the importance of MT polarity on the localization of members of the endocytic trafficking pathway. Our results show that the Hook protein is asymmetrically localized in the bristle, accumulating at the bristle tip. We found that in ik2 and spn-F mutants, Hook is mislocalized to the middle of the shaft. Thus, we suggest that mutations in ik2 and spn-F lead to defects in protein trafficking within the bristle shaft. To better characterize the mislocalization defects of the Hook protein, we will use a GFP-tagged version of this protein for in vivo imaging analysis with ik2 and spn-F mutants. What could be the reason for the defects in Hook protein trafficking in ik2 and spn-F mutants? One possible option is that the defects in Hook localization could be parallel consequences of absent ik2 and spn-F function. Another possible explanation is that the mutations in ik2 and spn-F affect MT function, which in turn affects Hook localization. Thus, the polarized organization of MTs in the bristle could serve as a platform for the correct localization of part of the endocytic compartments at the bristle tip, which in turn leads to polarized organization of the bristle shaft.
We thank Trudi Schüpbach, Damian Brunner, Shigeo Hayashi, Masayuki Miura, VDRC, and the Bloomington stock center for generously providing fly strains and reagents. We also thank Ofer Ovadia and Hagai Guterman for their help in statistical analysis.
This research was supported by the United States-Israel BiNational Science Foundation (BSF-2005137) and in part by Israel Science Foundation grant 166/06 (to U.A.) and by a grant from the National Science Foundation to G.M.G. (MCB-0344136).
Published ahead of print on 16 November 2009.
†Supplemental material for this article may be found at http://mcb.asm.org/.