Research Article 3531 Microvilli appear to represent the first step in bundle formation in Drosophila bristles

Lewis G. Tilney, Patricia S. Connelly and Gregory M. Guild* Department of Biology, University of Pennsylvania, Philadelphia, PA 19104-6018, USA *Author for correspondence (e-mail: [email protected])

Accepted 10 March 2004 Journal of Cell Science 117, 3531-3538 Published by The Company of Biologists 2004 doi:10.1242/jcs.01215

Summary During bristle development the emerging bristle shaft, Evidence is presented showing that socket cells do not socket cell, and the apical surface of thoracic epithelial cells contain forked protein crossbridges, a fact that may explain form tiny protuberances or pimples that contain electron- why cortical bundles only appear in bristle shaft cells. dense material located on the cytoplasmic surface of the Furthermore, as pimples and microvilli form in the absence pimple tip. In a few cases short actin filaments extend from of both forked and fascin crossbridges, we also conclude this material into the cortical cytoplasm. When cultured that neither of these crossbridges account for core bundle in the presence of jasplakinolide, an agent that prevents formation in microvilli, but there must exist a third, filament disassembly, pimples elongate to form microvilli as yet unidentified crossbridge in this system. containing a core of crosslinked filaments. Emerging- Immunocytochemisty suggested that this new crossbridge bristle mutants delay cortical bundle formation and are is not Drosophila villin. Finally, ultrastructural aggregated by forked protein crossbridges. Using these comparisons suggest that microspikes and microvilli form mutants and enhancing core bundle formation with very differently. jasplakinolide we found that microvillar formation represents the first stage in the morphogenesis of much Key words: Microvilli, Bristles, Drosophila, Jasplakinolide, Actin, larger actin bundles in Drosophila bristle shaft cells. Microvillar formation, Filament assembly

Introduction morphogenesis, we can now interpret results presented earlier Microvilli are common extensions of the apical surface of for which we then had no explanation. We show that the epithelial cells. Examples include intestinal epithelial cells, formation of microvilli in situ occurs by filament elongation the convoluted tubules of the kidney, as well as cells of the from small densities located on the plasma membrane, not by reproductive system such as oocytes, newly fertilized eggs, bundling of filaments in the cell cortex as recently suggested developing embryos and nurse cells (reviewed by DeRosier for microspike formation in motile cells (Svitkina et al., 2003). and Tilney, 2000). Microvillar extensions are uniform in Although descriptive, a format viewed by some as passé, this diameter and have a core of crosslinked actin filaments that study is paramount because it answers in-depth questions run from the plasma membrane, where they are often furthering our understanding of the formation of bristles and embedded into some electron-dense material, down into the pattern development in this model system. We now have to cortex of the cell. Furthermore microvilli can be modified search for a third, as yet unidentified crossbridge as well as try secondarily to form the of hair cells of the cochlea to determine exactly how and by which proteins the precursor or the actin bundles in Drosophila nurse cells (Guild et al., stage in the formation of microvilli in this as well as other 1997). model systems are orchestrated. In this paper we present evidence that during Drosophila bristle formation, microvilli are also secondarily modified to provide the basis of large cortical bundles. In a recent review Materials and Methods (DeRosier and Tilney, 2000) we speculated that such a Drosophila stocks, developmental staging, dissection and modification might exist. We also show that the crossbridge culture used in the generation of the core bundles in microvilli is The Oregon-R strain of Drosophila melanogaster was used as the wild 3 36a neither the forked protein nor fascin, two known crossbridges type in these studies. The singed stock (sn ), forked stock (f ) and singed-forked (sn3 f 36a) double-mutant stock were maintained as required for the generation of large cortical bundles in bristles. X2 Additionally, we show that it cannot be villin, a crossbridging viable homozygotes. The In(1)dl-49, singed chromosome was maintained over the In(1)Mud, Mud1 chromosome. Flies were protein found in nurse cells of developing Drosophila egg maintained on standard cornmeal-molasses-yeast food at 25°C, 60- chambers (Mahajan-Miklos and Cooley, 1994). Armed with 70% relative humidity, with a 12 hour/12 hour day/night cycle. the fact that microvilli are found at bristle tips and that the Complete descriptions of genes and symbols can be found elsewhere lateral aggregation of the core bundles in these microvilli (Lindsley and Zimm, 1992) and on FlyBase (Flybase Consortium, account for the intermediate steps in mature bundle 2003). Developmental staging, thoracic dissections and culturing in 3532 Journal of Cell Science 117 (16) the presence of jasplakinolide (3 µM) were as described (Tilney et al., cytoplasmic surface of the plasma membrane (Figs 2 and 3). 2003). In some pimples short actin filaments (approximately 0.1 µm in length) extend from the dense material into the cortex. From this morphology and from what we know about other systems Confocal and electron microscopy (see Discussion) we suspected that the pimples might represent The procedures for fixation, antibody and phalloidin staining, and a precursor stage in microvillar formation. To test this idea, we confocal microscopy were described previously (Guild et al., 2002). The rabbit polyclonal antibody directed against the forked proteins incubated cultured pupal thoraces with the sponge toxin was also described previously (Guild et al., 2003). The mouse jasplakinolide before fixation (Fig. 2b-e). Jasplakinolide is a monoclonal antibody (6B9) directed against the Drosophila quail membrane-permeant phalloidin-like compound (Bubb et al., protein (Drosophila villin) was developed by Mahajan-Miklos and 1994; Bubb et al., 2000) that binds to actin filaments and Cooley (Mahajan-Miklos and Cooley, 1994) and obtained from the prevents their disassembly. If these pimples are indeed Developmental Studies Hybridoma Bank developed under the microvillar precursors then in the presence of jasplakinolide auspices of the NICHD and maintained by the University of Iowa. the filaments in the pimples might elongate as any filament that The procedures used for thin-section transmission EM have been forms would fail to disassemble. Accordingly, the pimples seen described previously (Tilney et al., 1998). in the tip of newly emerging bristles should elongate. In fact this is exactly what we found when we examined thin sections cut through newly emerging bristle shafts and adjacent Results epithelial cells from preparations that had been treated with Microvilli are present on emerging bristle tips jasplakinolide prior to fixation (Fig. 2b-c). We found large Bristles begin to emerge in 32-hour pupae and elongate over numbers of microvilli at the bristle tips. Furthermore these the next 16 hours after which they reach their mature length microvillar-like extensions from the newly emerging bristle (Tilney et al., 2000b). Examination of thin sections cut through tips all possess an internal core of actin filaments that extend the apical surface of the shaft of a bristle cell just prior to its from dense material attached to the cytoplasmic surface of the emergence from the surface of the thorax reveals the presence microvillar tips (Fig. 2b-e). This is demonstrated particularly of numerous microvilli on this tiny surface (Fig. 1). These clearly when two serial sections of a microvillar-like extension microvilli are at least 1 µm in length. It is as if the apical seen on a bristle tip (Fig. 2c,d) were observed at a greater surface of the shaft anticipates its elongation. magnification (Fig. 2e). Here the core microfilaments extend We next examined thin sections of newly emerging bristles. from the dense material at the tip all the way into Numerous short protrusions or pimples are found at the tips of the cortical cytoplasm of the bristle proper. There are these bristle shafts (Fig. 2a). Similar pimples are present on the transverse stripes on this core bundle which indicate the apical surface of the socket cell (Fig. 2a) as well as along the presence of a crossbridge. apical surface of the epithelial cells that separate adjacent bristles (Fig. 3a,b). All of these pimples are characterized by having some electron-dense material attached to the Core bundles in microvilli aggregate into cortical bundles, a process that requires actin crossbridging proteins Forked protein In an earlier publication (Tilney et al., 1998) we used mutants to show that the forked crossbridges were instrumental in aggregating tiny core bundles of actin filaments into larger bundles. These aggregations differentiate into the cortical bundles present in fully elongated bristles. Subsequently the actin filaments are crossbridged into a paracrystallic array by the fascin crossbridge. Much to our initial surprise when we examined transverse sections through the newly emerging or emerged tips of forked mutants we found not fewer pimples or microvilli, but many more than the wild type (Tilney et al., 1998). This was a surprise for two reasons. First, we assumed that the actin filaments making up the core filament bundles in microvilli would be crosslinked by the same crossbridge used to aggregate the bundle laterally (the forked crossbridge) reasoning that if the forked proteins were missing (in the forked mutant), microvilli would not appear. As microvilli do form it follows that the actin filaments in the core bundles are not crosslinked together by the forked proteins. Second, why the increased number of microvilli? We assume the answer to this question is that lateral aggregation of the core bundles is less Fig. 1. Thin section through the apical end of a bristle shaft cell (B) likely in the absence of the forked protein and this role must from a 32-hour pupa just prior to the emergence of the bristle. The now be taken over by another crossbridge, perhaps fascin. Thus, bristle shaft cell anticipates this event by extending long microvilli. the emerging tips would elongate more slowly and at the same Surrounding the bristle shaft cell is its socket cell (SC). time have an increased number of microvilli. This hypothesis is Microvilli and Drosophila bristles 3533 born out. The bristles of this mutant are 40% shorter than the expressing normal amounts of the forked protein, now becomes wild type and contain, in mature bristles, threefold smaller very revealing. First of all, pimples and microvilli in this bundles than the wild type (Tilney et al., 2000a). mutant appear not only along the epithelial cells between the bristle shafts but are also prominent on the new emerging bristle. This fact means that fascin, like the forked proteins, Fascin protein cannot account for the bundling of actin filaments together in A mutant lacking the other major crossbridge, fascin, but the microvilli and/or pimple precursors. Second, and what is

Fig. 2. Jasplakinolide treatment accentuates microvilli growth on elongating bristle tips. (a) Thin section through the apical portion of an untreated newly emerging bristle shaft cell (B) from a 33-hour pupa. A portion of the neuron (N) that innervates the bristle is present in this section. Note that pimples are present extending from the surface of this emerging bristle shaft cell (arrowheads) as well as from the surface of the adjacent socket cell (SC). (b-d) 33-hour pupal thoraces were cultured in jasplakinolide for 1 hour before fixation and thin sections of emerging tips are shown here. In b-d microvilli are seen extending from the bristle tips. In panels c and d (which are adjacent sections) microvilli have extended from the pimples from adjacent epithelial cells (E). (e) Higher magnification of the microvillus indicated by the squares in panels c and d. As these are adjacent sections we have pasted the squares together to show the entire length of the microvillus. Within this microvillus one can easily see the core of actin filaments. There are transverse stripes on this core bundle (lines) which indicate the presence of crossbridges between the filaments. 3534 Journal of Cell Science 117 (16)

Fig. 3. Jasplakinolide treatment results in actin filament elongation from the surface projections on epithelial cells. The parallel dense lines above the plasma membrane represent an early stage in cuticle deposition (C). (a) Untreated 32-hour pupa. At this stage bristles begin to erupt from the surface of the thorax (not depicted). Extending from the apical surface of this epithelial cell are tiny projections we call pimples. Attached to the cytoplasm surface of the plasma membrane in each pimple is some electron-dense material (vertical arrowheads). (b) Untreated 41-hour pupa. Pimples are present on the apical surface. By this time actin filaments extend from the dense material of some of the pimples into the cortical cytoplasm. Microtubules (MT) are found running parallel to the cell surface. (c,d) Portions of epithelial cells from 41-hour pupal thoraces cultured in jasplakinolide for 4 hours. Note that actin filaments are now commonly seen extending from the electron-dense material at the pimple tips into cortical cytoplasm as a rootlet (horizontal arrowheads). These filaments are three or four times longer than those seen in untreated pupae and are clustered together into discrete core bundles. The pimples in panel c have not elongated outward although they possess long rootlets. The pimples shown in panel d have elongated into short microvilli; the actin core bundles extend from the dense material at the pimple tips into the cortical cytoplasm as rootlets (arrowheads) where actin filaments often fray apart. (e) Epithelial cells from a 41-hour pupa lacking both the forked and fascin proteins cultured in jasplakinolide for 4 hours. Pimples are present and form short microvilli on the surface of the epithelial cells. Extending from the tips of the pimples are actin filaments that are clustered together into bundles (horizontal arrowheads) and then fray apart in the cortical cytoplasm. (f) Transverse section through microvilli lacking both forked and fascin crossbridges. Dots in the center of each are the actin filaments of these core bundles cut in transverse section. Panels a-e are printed at the same magnification. Microvilli and Drosophila bristles 3535

Fig. 4. Bristles lacking fascin crossbridges still form microvilli. (a,b) Transverse sections of newly emerged bristle tip from 34-hour-old singedX2 mutant pupa. A smaller version appeared in Tilney et al., (1998) and is included here at sufficient magnification in panel b so that the aggregated core bundles (arrows) of the microvilli can be clearly visualized. Panel b, published with permission by Rockefeller University Press. The size of these bundles and the number of filaments per bundle are similar to those in microvilli at the bristle tip (e.g. Fig. 2e) or to those in longitudinal section (panel d). (c,d) Grazing longitudinal section through a singedX2 bristle tip at the same magnification as in panels a and b, respectively. The bundles are indicated by arrows. Note that the diameters of the bundles here are the same as in panel b and in Fig. 2e. particularly interesting, is that in the fascin-less mutant To test this possibility we stained socket cells and bristle shafts (singed) we find linear arrays of tiny actin bundles attached to with an antibody prepared against the forked proteins. The the plasma membrane (Fig. 4a) (Tilney et al., 1998). Each of socket cells are not recognized by this antibody, in contrast to these contains approximately the same number of filaments the adjacent shaft that stains heavily (Fig. 5), possibly (Fig. 4b) as those in the microvilli on the bristle tip (Fig. 2e). explaining why the socket cells do not form large actin bundles. We presume that the reason why we see these linear aggregates is that in the absence of fascin these aggregates mature slowly or not at all into normal round cortical bundles. By removing Pimples are microvillar precursors fascin we have frozen a stage in the differentiation of the In the presence of jasplakinolide not only are true microvilli round cortical bundles. If this interpretation is correct then present on the tips of newly emerging bristle shafts (Fig. 2c,d) longitudinal sections through a bristle tip should reveal parallel but also microvilli extend from the epithelial cells adjacent to core bundles. This is indeed the case (Fig. 4c,d). the bristle shaft (right side of these two panels). To prove that If pimples, as hypothesized earlier, are microvillar the pimples are a first step in microvillar formation requires precursors, then the core actin bundles in those microvilli real-time observations of individual pimples growing into would be aggregated by the forked proteins to form the large microvilli; this is just not possible with current technology. We actin bundles present in the bristle shaft. Assuming this to be do have, however, data from examination of the pimples from true, then one wonders why large actin bundles do not form in epithelial cells consistent with this idea. socket cells which as shown in Fig. 2a display pimples on their Tiny pimples are present on the apical surface of epithelial apical surfaces. Perhaps the reason why socket cells do not cells prior to emergence of a bristle, e.g. 32-hour pupae. In thin form large actin bundles is that they lack the forked proteins. section these pimples (Fig. 3a) appear as minute protuberances 3536 Journal of Cell Science 117 (16) other or both cross-linkers, we conclude that the core filaments in the microvilli must be crosslinked by a third yet to be identified crossbridge. Additional evidence for the existence of a third crossbridge comes from recent results (Tilney et al., 2003) on sections of bristles of the singed-forked double mutant cultured in the presence of jasplakinolide. In these bristles because filament disassembly/turnover has been inhibited, we found that the bristle cytoplasm filled with actin filaments. Some of these are free filaments but there were also clusters of filaments. As in this mutant both the forked and fascin proteins are not present, a third crossbridge must exist to bundle the filaments into clusters. Matova et al. (1999) demonstrated that Drosophila villin is essential for actin bundle formation in the nurse cells during oogenesis. As villin crossbridges actin filaments in vertebrate microvilli, we immediately suspected that this crossbridge might be the unknown crossbridge needed for bristle microvillar formation. Accordingly, we stained developing bristles with an antibody prepared against Drosophila villin. Fig. 5. Forked proteins are present in newly emerging bristles but absent from surrounding socket cells. Confocal image of a sprouting Although the antibody stains the nerve that innervates the thoracic macrochaete from a 33-hour pupa stained with anti-forked bristle and the hairs present on the Drosophila wings, the newly antibody (green) and phalloidin (red). Signal overlap is seen as emerging bristles are completely negative (data not shown). yellow. The perimeter of the socket cell (SC) is outlined with a Therefore, the unknown crossbridge cannot be villin. dashed line. Bar, 5 µm. Discussion within which is a small accumulation of electron-dense Epithelial cells lining the thorax as well as emerging bristle material. Actin filaments infrequently extend from these shaft cells and socket cells express on their apical surface tiny densities into the cell. In older pupae actin filaments are often protrusions or pimples, that resemble the tips of mature found extending from these pimple densities into the cortical microvilli. When the thorax is cultured in the presence of cytoplasm. If isolated thoraces of pupae are incubated in jasplakinolide the pimples elongate and now appear as true jasplakinolide then fixed and examined instead of pimples with microvilli. Each microvillus contains a core of actin filaments only a few actin filaments attached to the densities, the majority that extend from the electron-dense material at the tip into the have a core bundle of actin filaments that extend into the cell cortex as a rootlet. Subsequently, these core bundles cortical cytoplasm as a rootlet (Fig. 3c,d) and in many cases become laterally bound together into larger linear aggregates the pimples elongate and appear as short microvilli (Fig. 3d). by the forked crossbridges. These observations allow us to Within the latter is a core bundle of actin filaments that extends make sense of earlier published observations and at the same into the cortical cytoplasm as a rootlet. These core bundles then time group the morphogenesis of the bristle with other model appear to be units or the tiny bundles present in the bristle (Fig. systems such as the elaboration of stereocilia in the ear, 4d) that aggregate together to form the cortical bundles. Thus the formation of the of nurse cells and the jasplakinolide treatment has allowed us to concentrate on the morphogenesis of brush borders (reviewed by DeRosier and steps of bundle formation, namely how the core bundles with Tilney, 2000). Furthermore, knowledge of the intermediate their rootlets are formed from the pimples and how aggregating steps in bristle formation allows us to understand more about these with the forked proteins and their subsequent the morphogenesis of pattern generally and, in particular, to see differentiation by fascin could be the first step in the formation what role individual proteins or their homologs play. of the large cortical bundles. Further studies are necessary to Two examples illustrate this last point. First, from our see what influence other actin binding proteins, e.g. profilin, studies it is clear that three separate actin filament crossbridges capping protein or the Arp 2/3 complex, have on pimple are essential in sequentially building the actin bundles in formation, elongation and microvillar aggregation. mature bristles: fascin, the forked proteins and the so far unidentified crossbridge shown to be present here. A similar philosophy occurs in the formation of stereocilia and microvilli A third crossbridge is required for connecting adjacent in intestinal epithelial cells and in the convoluted tubule cells actin filaments in microvilli of the kidney (DeRosier and Tilney, 2000) although in each In the double mutant lacking both forked and fascin case different crossbridging proteins are used. Second, the crossbridges, pimples are also present on the epithelial cells pimples appear to be precursors to microvilli. Microvilli are as well as on newly emerging bristles. If cultured in the defined as stereotyped linear extensions of the cell surface presence of jasplakinolide the pimples elongate into microvilli containing a central core of crosslinked actin filaments that (Fig. 3e). Transverse sections through the tips of these insert into some electron-dense material attached to the plasma microvilli show actin filaments cut in cross section (Fig. 3f). membrane. Formation of microvilli therefore occurs by a As pimples and microvilli form in the absence of one or the different method to the assembly of microspikes on the Microvilli and Drosophila bristles 3537 pseudopodia of fibroblasts and keratocytes (Svitkina et al., In this last system the rootlets form first and subsequently 2003). zipper the plasma membrane down around them to form microvilli. As the total length of the core actin bundle (including the microvillus and the rootlet) is approximately the Microvilli are key intermediates in bundle assembly same as the length of the rootlet in the crypt cell, it was We now can understand why during bristle elongation in the proposed that the microvillus elongates by membrane addition absence of the forked crossbridge extra tiny actin bundles at the microvillus base (Mamajiwalla et al., 1992; Fath et al., appear at the elongating bristle tip (Tilney et al., 2000b) and 1990). The reason we dwell on this is that it is the core bundles why bristle elongation in the absence of the fascin cross-linker in the rootlets of the microvilli that aggregate together under produces flattened linear aggregates attached to the plasma the influence of the forked proteins to produce the cortical membrane (Tilney et al., 2000b) instead of round bundles. bundles in bristles (Fig. 5) (Tilney et al., 2000b). Likewise it In both cases the lack of one crossbridge slows bundle is the rootlets of the microvilli that associate like extension development sufficiently so we can more easily recognize ladders to form the actin cage in nurse cells (Guild et al., 1997). intermediate steps in bundle formation. Thus without the Our experiments on bristles treated with jasplakinolide (Fig. 2) forked proteins to aggregate the actin filament bundles that emphasize that microvilli can elongate from the bristle tip, but comprise the cores of microvilli or microvillar stubs, many what is key to cortical bundle formation is the aggregation of more cores are seen in thin section than in the wild type. the rootlets by the forked proteins and their subsequent Likewise, without fascin, the aggregated core bundles slowly crosslinking by fascin that accounts for the final shape of the change or fail to change from a linear aggregate of microvillar cortical bundles in the bristles. core bundles to a compact often spherical mass of filaments We should also mention that microvilli are dynamic and can (Tilney et al., 1995; Tilney et al., 1998). shorten or lengthen. Perhaps the most-studied examples are the microvilli of intestinal epithelial cells that shorten during starvation (Misch et al., 1980), after treatment with Microvillar dynamics cyclohexamide (Lecount and Grey, 1972) or lectins (Weinman Jasplakinolide stabilizes actin filaments against et al., 1989) and thereafter re-elongate provided the stimulus depolymerization (Bubb et al., 2000; Tilney et al., 2003). The is removed. Furthermore these microvilli elongate when G- fact that pimples on newly emerged bristles (and epithelial actin is added to the membrane intact ‘brush borders’ derived cells) elongate in the presence of jasplakinolide suggests that from intestinal epithelial cells (Mooseker et al., 1982). these pimples are likely to be microvillar precursors capable of Interestingly the elongation of the core bundle in each extension induced by filament elongation, provided that microvillus occurs at the tips of the microvilli, insertion filament disassembly is inhibited. This behavior is in keeping occurring between the electron-dense material at the tips and with recently published results using jasplakinolide on the ends of the filaments in core bundles. wild-type and mutant bristles (Tilney et al., 2003). This interpretation further explains the elaboration of microvilli on the apical surface of the shaft cell (Fig. 2a) prior to its Formation of microspikes and microvilli seem to occur elongation. However, it is also true that the pimples often by different mechanisms extend their core bundles as rootlets into the cortical Several investigators questioned whether microvilli are the cytoplasm. Rootlets exposed to the forked crossbridges induce same as microspikes. We think they are different for the aggregation into linear arrays, a key step in bundle formation following reasons. First, in contrast to microvilli, microspikes in bristles. found on neuron growth cones and pseudopodia like those From the literature we know that precursors of microvilli present on the surface of fibroblasts and keratocytes are resemble the pimples described here. Examples include (1) the generally not of uniform diameter but often are volcano microvilli on the apical surface of developing intestinal shaped. Within them is a core or cores of actin filaments and epithelial cells that elongate from tiny precursors (Stidwill sometimes microtubules, but unlike microvilli the actin and Burgess, 1986); (2) the ‘short papillae’ or microvillar filaments in microspikes are less regularly crossbridged and the precursors (Schroeder, 1978) that elongate following sea filaments vary in length. Furthermore these two types of urchin egg fertilization (Eddy and Shapiro, 1976); (3) de novo extensions tend to be found on different categories of cells formation of microvilli following the removal of existing although the distinction is far from perfect. Microspikes tend microvilli such as occurs following treatment with hydrostatic to be located on motile cells rather than on cells that are pressure in intestinal epithelial cells (Tilney and Cardell, 1970) attached to others by junctions and are thus less motile. and fertilized sea urchin eggs (Begg et al., 1983; Henson and Second, microvilli seem to be more stable and less likely to Begg, 1988). turnover than microspikes, as for example, the microvillus Particularly relevant to our studies on cortical bundle brush borders can be isolated intact, although they can shorten formation in elongating bristles is the fact that the rootlets of if traumatized. In contrast, growth cones in neurons and microvilli elongate from their precursor pimples or papillae, keratocytes form, retract completely, and regrow within short e.g. after treatment with jasplakinolide (this study) or in sea time courses. urchin eggs following fertilization (Wong et al., 1997). Another This study has pointed out a third difference, namely in their example is during microvillar formation in intestinal epithelial formation, which to our thinking is paramount. As already cells which begins in the crypts of Lieberkuln and continues mentioned, pimples seem to be precursors to microvilli, similar on the villus before its cells are sloughed away from the villus to the papillae noted on the surface of unfertilized sea urchin tip 2-3 days later (Fath et al., 1990; Mamajiwalla et al., 1992). eggs (Henson and Begg, 1988; Schroeder, 1978) and in 3538 Journal of Cell Science 117 (16) intestinal epithelial cells that are reforming microvilli after G. (2002). Actin filament turnover removes bundles from Drosophila bristle treatment with hydrostatic pressure (Tilney and Cardell, 1970). cells. J. Cell Sci. 115, 641-653. In both cases the pimples seem to elongate after fertilization Guild, G. M., Connelly, P. S., Ruggiero, L., Vranich, K. A. and Tilney, L. G. (2003). Long continuous actin bundles in Drosophila bristles are (Begg et al., 1982) or cessation of pressure treatment. Thus in constructed by overlapping short filaments. J. Cell Biol. 162, 1069-1077. systems where microvilli elongate, the filaments within them Henson, J. H. and Begg, D. A. (1988). Filamentous actin organization in the elongate downward from some electron-dense material at their unfertilized sea urchin egg cortex. Dev. Biol. 127, 338-348. tips, which in turn can lead to an elongation of the microvillus Lecount, T. S. and Grey, R. D. (1972). Transient shortening of microvilli induced by cycloheximide in the duodenal of the chicken. J. Cell or the formation of long rootlets that extend into the cell cortex. Biol. 53, 601-605. In contrast, in recent work on microspike formation in Lindsley, D. L. and Zimm, G. G. (1992). The Genome of Drosophila keratocytes it has been proposed that microspikes form by the melanogaster, p. 1133. San Diego, CA: Academic Press. zippering together of cortical actin filaments that act to form a Mahajan-Miklos, S. and Cooley, L. (1994). The villin-like protein encoded microvillus by pushing the membrane from the cortex upwards by the Drosophila quail gene is required for actin bundle assembly during oogenesis. Cell 78, 291-301. (Svitkina et al., 2003). This suggests that the control of these Mamajiwalla, S. N., Fath, K. R. and Burgess, D. R. (1992). Development processes is different, although we must be open to the of the chicken intestinal epithelium. Curr. Top Dev. Biol. 26, 123-143. suggestion that process extension may result from a Matova, N., Mahajan-Miklos, S., Mooseker, M. S. and Cooley, L. (1999). combination of both of these methods. Drosophila quail, a villin-related protein, bundles actin filaments in apoptotic nurse cells. Development 126, 5645-5657. Misch, D. W., Giebel, P. E. and Faust, R. G. (1980). Intestinal microvilli: We would like to express our thanks to the Drosophila Stock Center responses to feeding and fasting. Eur. J. Cell Biol. 21, 269-279. (Indiana University) and the Developmental Studies Hybridoma Bank Mooseker, M. S., Pollard, T. D. and Wharton, K. A. (1982). Nucleated (University of Iowa) for generously making available fly stocks and polymerization of actin from the membrane-associated ends of microvillar monoclonal antibody reagents, respectively. This work was supported filaments in the intestinal brush border. J. Cell Biol. 95, 223-233. by grants from the National Institute of Health (GM-52857) to LGT Schroeder, T. E. (1978). Microvilli on sea urchin eggs: a second burst of and the National Science Foundation (MCB-0077839) to GMG. We elongation. Dev. Biol. 64, 342-346. wish to thank Linda Ruggiero and Kelly Vranich for expert technical Stidwill, R. P. and Burgess, D. R. (1986). Regulation of intestinal brush assistance. We also wish to thank the referees who greatly improved border microvillus length during development by the G- to F-actin ratio. the presentation of our results with their suggestions. Dev. Biol. 114, 381-388. Svitkina, T. M., Bulanova, E. A., Chaga, O. Y., Vignjevic, D. M., Kojima, S., Vasiliev, J. M. and Borisy, G. G. (2003). Mechanism of filopodia initiation by reorganization of a dendritic network. J. Cell Biol. 160, 409- References 421. Begg, D. A., Rebhun, L. I. and Hyatt, H. (1982). Structural organization of Tilney, L. G. and Cardell, R. R., Jr (1970). Factors controlling the actin in the sea urchin egg cortex: microvillar elongation in the absence of reassembly of the microvillous border of the small intestine of the actin filament bundle formation. J. Cell Biol. 93, 24-32. salamander. J. Cell Biol. 47, 408-422. Begg, D. A., Salmon, E. D. and Hyatt, H. A. (1983). The changes in structural Tilney, L. G., Tilney, M. S. and Guild, G. M. (1995). F actin bundles in organization of actin in the sea urchin egg cortex in response to hydrostatic Drosophila bristles. I. Two filament cross-links are involved in bundling. J. pressure. J. Cell Biol. 97, 1795-1805. Cell Biol. 130, 629-638. Bubb, M. R., Senderowicz, A. M., Sausville, E. A., Duncan, K. L. and Tilney, L. G., Connelly, P. S., Vranich, K. A., Shaw, M. K. and Guild, G. Korn, E. D. (1994). Jasplakinolide, a cytotoxic natural product, induces M. (1998). Why are two different cross-linkers necessary for actin bundle actin polymerization and competitively inhibits the binding of phalloidin to formation in vivo and what does each cross-link contribute? J. Cell Biol. F-actin. J. Biol. Chem. 269, 14869-14871. 143, 121-133. Bubb, M. R., Spector, I., Beyer, B. B. and Fosen, K. M. (2000). Effects of Tilney, L. G., Connelly, P. S., Vranich, K. A., Shaw, M. K. and Guild, G. jasplakinolide on the kinetics of actin polymerization. An explanation for M. (2000a). Actin filaments and microtubules play different roles during certain in vivo observations. J. Biol. Chem. 275, 5163-5170. bristle elongation in Drosophila. J. Cell Sci. 113, 1255-1265. DeRosier, D. J. and Tilney, L. G. (2000). F-actin bundles are derivatives of Tilney, L. G., Connelly, P. S., Vranich, K. A., Shaw, M. K. and Guild, G. microvilli: what does this tell us about how bundles might form? J. Cell M. (2000b). Regulation of actin filament cross-linking and bundle shape in Biol. 148, 1-6. Drosophila bristles. J. Cell Biol. 148, 87-100. Eddy, E. M. and Shapiro, B. M. (1976). Changes in the topography of the Tilney, L. G., Connelly, P. S., Ruggiero, L., Vranich, K. A. and Guild, G. sea urchin egg after fertilization. J. Cell Biol. 71, 35-48. M. (2003). Actin filament turnover regulated by cross-linking accounts for Fath, K. R., Obenauf, S. D. and Burgess, D. R. (1990). Cytoskeletal protein the size, shape, location, and number of actin bundles in Drosophila bristles. and mRNA accumulation during brush border formation in adult chicken Mol. Biol. Cell 14, 3953-3966. enterocytes. Development 109, 449-459. Weinman, M. D., Allan, C. H., Trier, J. S. and Hagen, S. J. (1989). Repair FlyBase Consortium (2003). The FlyBase database of the Drosophila genome of microvilli in the rat small intestine after damage with lectins contained projects and community literature. Nucleic Acids Res. 31, 172-175. in the red kidney bean. Gastroenterology 97, 1193-1204. Guild, G. M., Connelly, P. S., Shaw, M. K. and Tilney, L. G. (1997). Actin Wong, G. K., Allen, P. G. and Begg, D. A. (1997). Dynamics of filamentous filament cables in Drosophila nurse cells are composed of modules that slide actin organization in the sea urchin egg cortex during early cleavage passively past one another during dumping. J. Cell Biol. 138, 783-797. divisions: implications for the mechanism of cytokinesis. Cell Motil. Guild, G. M., Connelly, P. S., Vranich, K. A., Shaw, M. K. and Tilney, L. Cytoskeleton 36, 30-42.