Summary
Myosin VI has been implicated in membrane dynamics in several organisms. The mechanism of its participation in membrane events is not clear. We have used spermatogenesis in Drosophila to investigate myosin VI's in vivo role. We demonstrate that myosin VI colocalizes with and is required for the accumulation of the actin polymerization regulatory proteins, cortactin and arp2/3 complex, on actin structures that mediate membrane remodeling during spermatogenesis. In addition, we show that dynamin localizes to these actin structures and that when dynamin and myosin VI function are both impaired, major defects in actin structures are observed. We conclude that during spermatogenesis myosin VI and dynamin function in parallel pathways that regulate actin dynamics and that cortactin and arp2/3 complex may be important for these functions. Regions of myosin VI accumulation are proposed as sites where actin assembly is coupled to membrane dynamics.
Introduction
Myosin VI is a conserved unconventional myosin found in several organisms including worms, flies, mice and humans ( Avraham et al., 1997; Avraham et al., 1995; Hasson et al., 1996; Kelleher et al., 2000; Kellerman and Miller, 1992). Like all members of the myosin superfamily, myosin VI contains an N-terminal motor domain with conserved actin- and ATP-binding motifs. However, unlike most other myosins that have been tested for direction of motility, myosin VI moves towards the slow growing, pointed end of an actin filament ( Homma et al., 2001; Wells et al., 1999). Only one other myosin so far studied, myosin IX, moves towards the pointed end ( Inoue et al., 2002). Recently, a heritable form of human deafness has been linked to mutations in myosin VI ( Melchionda et al., 2001). Given the unusual direction of motility of myosin VI and the clinical relevance of myosin VI, there is much interest in understanding its cellular and molecular functions.
Some clues as to the cellular functions of myosin VI have come from phenotypic analysis of mutations in Mus musculus, Caenorhabditis elegans and Drosophila melanogaster. The defects seen in the deaf Snell's Waltzer mutant mouse are caused by a loss-of-function mutation in the myosin VI gene. These mutants display disruptions in the organization and morphogenesis of the stereocilia in the inner ear ( Avraham et al., 1995; Self et al., 1999). The hair cell stereocilia in these mutants are disorganized, and the membrane between neighboring stereocilia is uprooted, suggesting a defect in the anchoring of stereocilia and the apical membrane ( Self et al., 1999). In C. elegans, loss-of-function mutations in one of two isoforms of myosin VI cause male sterility ( Kelleher et al., 2000). These mutants display defects in the asymmetric sorting of mitochondria and ER/Golgi-derived organelles during spermatogenesis, suggesting defects in organelle trafficking or localization. Consistent with a role in trafficking, myosin VI in vertebrate tissue culture cells localizes to trafficking membranes such as the trans-Golgi network ( Buss et al., 1998) and clathrin-coated pits and vesicles ( Buss et al., 2001a), and overexpression of the tail fragment of myosin VI reduces transferrin uptake ( Buss et al., 2001a). From these studies and those in Drosophila (described below) it has been proposed that myosin VI has a role in organizing and/or trafficking membrane. However, the details of how myosin VI participates in these functions remain unclear.
In Drosophila there is one myosin VI gene. The proposed function for myosin IV is in membrane organization and trafficking. In Drosophila syncytial blastoderm embryos, myosin VI protein localizes to cortical actin in transient mitotic membrane invaginations (pseudocleavage furrows). Inhibition of myosin VI function in the Drosophila embryo causes defects in formation of these transient mitotic membrane invaginations ( Mermall and Miller, 1995). Thus, myosin VI is required for membrane remodeling during embryogenesis.
Myosin VI function is also required for membrane remodeling during the individualization step of spermatogenesis ( Hicks et al., 1999). During individualization, a syncytial membrane encasing a bundle of 64 spermatids is remodeled so that an individual membrane encases each of the 64 mature sperm. Myosin VI localizes to an actin complex, the individualization complex, which assembles at the spermatid heads at the start of individualization ( Hicks et al., 1999). This complex progresses from the spermatid heads to the tips of tails, remodeling membrane as it moves. Loss of myosin VI in the testis leads to the disruption of these actin complexes as they progress and, consequently, individualization is not completed ( Hicks et al., 1999).
The precise function of myosin VI in individualization remains unclear. Does myosin VI catalyze transport of membrane trafficking components at the actin individualization complex? Is it involved in anchoring membrane to actin and/or organizing actin polymerization sites at the membrane? To obtain a clearer understanding of myosin VI's role in the individualization complex, we compared the localization of myosin VI in spermatids with that of proteins known to have roles in regulating actin dynamics or membrane trafficking. We also tested for genetic interactions between mutations in myosin VI and mutations in the membrane trafficking gene, dynamin. We report that myosin VI localizes with regulators of actin dynamics and is required for the localization of these proteins. We also show that myosin VI mutations interact genetically with dynamin mutations to affect actin dynamics. We propose that myosin VI is important for actin assembly at sites of membrane remodeling.
Materials and Methods
Fly culture and fly stocks
All fly stocks were raised and maintained on standard Drosophila cornmeal, agar, sucrose medium at 22°C (except as indicated; see below). The jaguar (jar1) mutant allele ( Castrillon et al., 1993) and shibire (shi1) mutant allele were obtained from the Bloomington Drosophila Stock Center (Bloomington, IN).
Temperature shifts to inactivate shibire function
Newly eclosed adult flies (0-12 hrs after eclosure) were shifted to the non-permissive temperature of 30°C in a water bath for the times indicated in each experiment. At the end of the temperature shift, flies were immediately dissected in prewarmed (30°C) dissection buffer on prewarmed dissection slides. A temperature-controlled rubber mat (controller and heating mat from Thermolyne) set at low temperature was placed under the dissection slide to maintain the non-permissive temperature during dissection. Testes dissected in this manner were immediately fixed as described below.
Testis dissection and fixation
Newly eclosed males (3-12 hrs after eclosure) were collected and incubated on ice until paralyzed (except with shi1 flies, in which case they were already paralyzed by the non-permissive temperature). Testes were dissected in testis isolation buffer (47 mM NaCl, 183 mM KCl, 10 mM Tris pH 6.8) and prepared by a method previously described ( Hime et al., 1996) with minor modifications. Briefly, testes were squashed, incubated in cold 100% ethanol for 10 minutes, incubated in 4% paraformaldehyde for 7 minutes, incubated once in PBT (PBS + 0.3% Triton X-100 + 0.3% sodium deoxycholate) for 15 minutes and washed twice in PBST (PBS + 0.1% Triton) for 10 minutes.
Antibodies and fluorescent markers
Monoclonal anti-Drosophila myosin VI antibodies, 3c7 (1:20), and polyclonal anti-Drosophila myosin VI antibodies (1:100) were previously described ( Kellerman and Miller, 1992). The mouse anti-rat dynamin-1 monoclonal antibody (1:1000) was obtained from Transduction Laboratories (Lexington, KY, catalog number D25520). The rabbit anti-Drosophila dynamin-1 (shibire) polyclonal antibody (1:500) was a gift from Mani Ramaswami ( Estes et al., 1996). Both antibodies to dynamin-1 show the identical localization in the testis. The rat anti-Drosophila cortactin antibody (1:250) was a gift from Manabu Takahisa ( Katsube et al., 1998). The rabbit anti-Drosophila Arp3 (1:50) was a gift from Bill Theurkauf and the rabbit anti-human ARPC2/p34 (1:100) was a gift from Matt Welch ( Welch et al., 1997). The rabbit anti-Drosophila amphiphysin (1:100) and anti-Drosophila alpha adaptin (1:25) were gifts from Andrew Zelhof ( Zelhof et al., 2001) and Nick Gay ( Dornan et al., 1997; Zelhof et al., 2001). The rabbit and rat anti-Drosophila capping protein β (CP-β) antibodies (1:10) were described previously ( Hopmann et al., 1996).
Immunofluorescence staining and imaging
Fixed testes were blocked in PBST + 3% BSA for at least 30 minutes at room temperature or overnight at 4°C. All antibodies and dyes were diluted in PBST + 3% BSA. Antibodies were incubated with tissue samples for 2 hours at room temperature or overnight at 4°C. After incubation with primary antibodies in a humid chamber, fixed testes were washed four times in PBST + 3% BSA for 10 min at room temperature. Then, secondary antibodies were incubated for 2 hours in a humid chamber. Goat anti-mouse Alexa Fluor 488 and goat anti-rabbit Alexa Fluor 488 secondary antibodies were diluted 1:500. Goat anti-mouse Alexa Fluor 568 and goat anti-rabbit Alexa Fluor 568 secondaries were diluted 1:1000. Alexa Fluor 488 phalloidin (1:125) and the DNA dyes TOTO-3 (1:3000) and DAPI (1 μg/ml) were included in incubations with secondary antibodies for 2 hours at room temperature. All Alexa-conjugated secondary antibodies, phalloidin and TOTO-3 dyes were from Molecular Probes. After staining, testes were washed four times in containers filled with PBST for 10 minutes at room temperature and finally mounted in mounting media [500 mg/ml glycerol, 17 mM Tris pH 8.5, 200 mg/ml Mowiole (Calbiochem)]
All imaging of fluorescent stains was performed on a Leica laser scanning spectral confocal microscope (model TCS SP2). Images shown are either single planes or, when noted, multiple planes collapsed into a single images (a projection). Projections were obtained by collecting consecutive planes at 0.5μ m intervals in volume samples that include most of the actin individualization complex.
Results
Myosin VI colocalizes with cortactin and other actin-binding proteins that regulate actin polymerization
During individualization, myosin VI localizes to a complex of filamentous actin, the individualization complex, responsible for remodeling the syncytial membrane around a bundle of 64 spermatids ( Hicks et al., 1999). The actin complex initially assembles around spermatid nuclei and then progresses away from the nuclei and down the tails of the spermatids, extruding cytoplasm and remodeling membrane as it moves ( Fabrizio et al., 1998). Initially, myosin VI accumulates in a particulate fashion ( Fig. 1C) along actin filaments ( Fig. 1B) as they assemble around the nuclei, but just as the actin individualization complex initiates movement away from the nuclei, myosin VI concentrates at the front of each actin cone ( Fig. 1G). As the complex progresses down the length of the spermatid tails, myosin VI further concentrates into a tight band at the front of the actin cones ( Fig. 1K). Myosin VI is also diffusely localized in front of the actin cones in association with the membrane and cytoplasm of the cystic bulge ( Fig. 1K).
(A) Schematic representation of three spermatids undergoing individualization. Spermatogenesis produces 64 elongated spermatids interconnected by cytoplasmic bridges. During individualization, syncytial membrane (black) of the bundled spermatids (far left) is remodeled to individual membranes encasing each of the spermatids separately (far right). The individualization complex is responsible for membrane remodeling, and its central constituent is actin (green; one actin cone is associated with each spermatid in the complex) ( Fabrizio et al., 1998). At the beginning of individualization (far left) the actin cones assemble around the spermatid nuclei (blue), and myosin VI (red dots) coats the actin cones in a particulate fashion. Later, the actin cones progress away from the nuclei and down the length of the spermatid axonemes (orange), extruding the cytoplasm between the spermatid tails, resolving the cytoplasmic bridges (area between ovals of membrane between the spermatid axonemes) and remodeling the syncytial membrane into individual membranes that encase each spermatid (middle). As soon as the actin cones move away from the nuclei, myosin VI localizes to an intense band at the front of the actin cones (red band). A bulge of plasma membrane, syncytial cytoplasm, vesicles and organelles (the cystic bulge) develops. When the actin cones reach the apical end of the testis, membrane and cytoplasm collected are pinched off, leaving the spermatids completely encased in their own membranes. The length of the spermatids is approximately 2 mM, and the cones take an estimated 18 hours to travel this distance. (B-M) Laser scanning confocal images of myosin VI staining at different stages of individualization complex progression. Early (B-E) in the process of individualization, actin cones (B) have assembled around the nuclei (D) of spermatids near the basal end of the testis, and myosin VI (C) coated the surface of actin cones in a particulate fashion. A little later (F-I), the actin cones (F) have moved away from the nuclei (H), myosin VI (G) was concentrated at the front of the actin cones. Much later (J-M), myosin VI accumulated and concentrated in a tight band (K) at the front of the actin cones as the individualization complex moves down the testis. Each image shows a single plane through one group of 64 bundled spermatids with 64 associated actin cones.
The site at which myosin VI concentrates is the junction between a moving actin structure and a zone of active membrane remodeling. This location places myosin VI in an ideal position to link sites of remodeling to actin dynamics. Therefore, we examined the localization of proteins that have been implicated in membrane/actin coordination. One such protein is the actin-binding protein cortactin. Cortactin is thought to link membrane signaling proteins to actin dynamics by virtue of is ability to associate with both actin polymerization components and membrane-associated kinases ( Olazabal and Machesky, 2001; Weed et al., 2000).
We examined the distribution of cortactin in individualizing spermatids with anti-Drosophila cortactin antibodies. Like myosin VI, cortactin concentrated at the front of actin cones ( Fig. 2B), and double labeling of spermatids with cortactin and myosin VI antibodies showed that they colocalized at the front of each actin cone ( Fig. 2M,N). Cortactin was also present on the cyst membrane ( Fig. 3B). The distribution of cortactin in individualization complexes indicates that the fronts of the actin cones are sites where actin polymerization might be coupled with membrane dynamics. Myosin VI colocalization with cortactin at these sites suggests myosin VI may also be involved in these dynamics.
Myosin VI colocalizes with cortactin, arp2/3 complex and capping protein at the front of individualization complexes. (A-L) Confocal images of spermatids double labeled with Alexa Fluor 488 phalloidin (A,D,G,J) and antibodies to cortactin (B), arp3 (E), ARPC2/p34 (H) or capping protein (K). (M-U) Projections of images of spermatids double labeled with anti-myosin VI antibodies (M,P,S) and antibodies to cortactin (N), arp3 (Q) and capping protein (T).
Cortactin and arp2/3 complex accumulation were abnormal on individualization complexes in myosin VI mutants. (A-L) Low magnification of a confocal image of wild-type (A-C, G-I) or myosin VI mutant (jar/jar; D-F, J-L) individualization complexes stained with Alexa Fluor 488 phalloidin (A,D,G,J) and anti-cortactin antibodies (B,E) or anti-arp3 antibodies (H,K). (M-X) Zoomed confocal images of a few actin cones from the individualization complexes in A-L. Cortactin localized to the fronts of the actin cones (blue arrow; B,N) and on the cyst membrane (yellow arrow; B). Cortactin localized to cyst membranes (yellow arrow, E) and actin cones in mutants but was not concentrated at the front of the actin cones (white arrow; E,Q). Cortactin localized normally to the cyst membrane in the myosin VI mutant (yellow arrow, E). Arp3 localized to the front of actin cones (blue arrowhead; H,T) and the cytoplasm of the cystic bulge (H). Arp3 was present on actin cones in mutants but was not well concentrated at the front of actin cones (white arrowhead; K,W). Images of wild-type and myosin VI mutant testes were collected at the same gain.
To further demonstrate that the fronts of the actin cones are sites of regulated actin assembly, we examined the distribution of the arp2/3 complex and capping protein in individualizing spermatids. The arp2/3 complex is a complex of seven proteins that binds actin filaments and nucleates new actin filament assembly ( Cooper et al., 2001; May, 2001). Capping protein is a barbed-end actin-binding protein with a known role in regulating actin polymerization at sites where the arp2/3 complex promotes assembly ( Cooper and Schafer, 2000; Schafer and Cooper, 1995). It is also concentrated in regions of dynamic actin assembly in many cell types ( Schafer et al., 1998; Waddle et al., 1996). In individualizing spermatids, the arp2/3 complex, as demonstrated by arp3 ( Fig. 2E) and ARPC2/p34 ( Fig. 2H) staining, and capping protein, as demonstrated by CP-β staining ( Fig. 2K), concentrated at the front of actin cones. In both cases, staining was also visible generally through the cytoplasm of the cyst and along the actin cones ( Fig. 2H,K). Double labeling experiments showed that myosin VI colocalized with concentrated arp3 ( Fig. 2P,Q) and capping protein ( Fig. 2S,T) at the front of actin cones. The accumulation of proteins involved in actin polymerization at the front of the actin cones supports the idea that the zone where myosin VI concentrates is a zone of active actin assembly.
Myosin VI is required for the proper distribution of cortactin and the arp2/3 complex on individualization complexes
The colocalization of cortactin, arp2/3 complex and myosin VI on individualization complexes prompted us to examine cortactin and arp2/3 complex distribution on individualization complexes in myosin VI mutants. Cortactin could be detected on actin individualization complexes in myosin VI mutants (jar1). However, its distribution was not normal. Cortactin was not concentrated at the front of the actin cones ( Fig. 3E,Q). Instead, it was weakly present uniformly along the cones. The complexes shown have a disrupted morphology and reduced actin staining, as is typically observed for progressed actin cones in myosin VI mutants. When early individualization complexes were examined in myosin VI mutants, no early complexes showed any concentration of cortactin at the front of cones (data not shown). By contrast, in wild-type spermatids, some early individualization complexes had cortactin concentrated at the front and others did not. When doubly stained for myosin VI, those complexes with concentrated myosin VI also showed concentrated cortactin. We conclude that myosin VI is required for the proper asymmetrical distribution of cortactin on actin cones.
In contrast to the localization of cortactin on actin cones, its localization to cyst membrane was unaffected in myosin VI mutants, indicating that myosin VI is not required for its proper localization to cyst membrane ( Fig. 3B).
We also observed defects in arp2/3 complex localization in myosin VI mutants. Arp3 did not concentrate at the front of actin cones, either early (data not shown) or on progressed complexes ( Fig. 3K), in myosin VI mutants. In addition, there appeared to be a higher level of arp3 staining in the cytoplasm of the cysts in myosin VI mutants in comparison to wild-type cysts ( Fig. 3K,H). This may be because arp3 cannot concentrate on the actin cones and, instead, accumulates in the cytoplasm. Like arp3, ARPC2/p34 concentration at the front of actin cones was abolished in myosin VI mutants (data not shown). Therefore, like cortactin, asymmetric distribution of the arp2/3 complex on the actin cones is dependent on myosin VI function. These findings support a role for myosin VI in regulating actin dynamics by participating in the localization of cortactin and arp2/3 complex at the front of the individualization complex.
Dynamin is localized on actin complexes of individualizing spermatids
The close coupling between actin assembly and membrane remodeling during individualization prompted us to examine proteins involved in membrane dynamics to determine if myosin VI works with these proteins. Dynamin, encoded by the Drosophila shibire gene, is a large GTPase with known roles in promoting the fission of clathrin-coated pits into clathrin-coated vesicles during endocytosis ( Hinshaw, 2000; McNiven, 1998). Dynamin binds cortactin and regulates actin dynamics at membrane sites of actin assembly ( Ochoa et al., 2000). Thus, dynamin is a likely molecule to function in individualization complexes.
We examined dynamin distribution on individualizing spermatids and found that it localized along the length of the actin cones ( Fig. 4A-C) in a distribution most similar to actin. As expected from the distribution of dynamin on actin cones, myosin VI concentrated at the front of dynamin-stained cones ( Fig. 4D-F). Dynamin localization along the actin cones suggests that this region might either be an area of high endocytic membrane trafficking or a region where actin dynamics are regulated by dynamin. Furthermore, the close juxtaposition of myosin VI and dynamin suggests that they might participate in the same process during individualization.
Localization of proteins involved in membrane events on actin individualization complexes. Confocal images of individualizing spermatids double-labeled with Alexa Fluor 488 phalloidin (A,G,J) and anti-dynamin antibody (B), anti-α adaptin antibody (H) or anti-amphiphysin antibody (K). In double labeled samples, myosin VI (E) was concentrated at the front of dynamin-stained complexes (D). Dynamin and amphiphysin colocalized with actin in the individualization complexes (C,L); however, unlike dynamin, amphiphysin also was concentrated at the front of the actin complexes.
To gain a clearer understanding of dynamin's function in the actin cones, we examined the distribution of two proteins known to interact with dynamin in vertebrates, but which appear to function in different pathways in Drosophila. α-Adaptin is the α subunit of the AP-2 adaptor complex, which is known to bind clathrin and function in early endocytosis ( Hirst and Robinson, 1998). α-Adaptin is also required for endocytosis in Drosophila ( Gonzalez-Gaitan and Jackle, 1997). Amphiphysin, on the other hand, is not required for endocytosis in Drosophila ( Leventis et al., 2001; Razzaq et al., 2001; Zelhof et al., 2001). However, amphiphysin can influence filamentous actin localization and has been implicated in membrane morphogenesis and organization in Drosophila ( Razzaq et al., 2001; Zelhof et al., 2001).
α-Adaptin was neither concentrated on actin cones nor at the front of the cones, as indicated by anti-Drosophila α-adaptin antibodies ( Fig. 4G-I). Instead, it localized in a particulate fashion throughout the cystic bulge ahead of the actin cones. By contrast, Drosophila amphiphysin antibodies localized to actin cones in a manner similar to dynamin ( Fig. 4K,B). However, unlike dynamin, amphiphysin also concentrated at the front of cones in a manner similar to myosin VI, cortactin and the arp2/3 complex. Thus, its distribution is intermediate between dynamin and cortactin/myosin VI. As amphiphysin localized to the actin cones whereas α-adaptin did not, we conclude that dynamin on the actin cones participates in a non-endocytic function. We hypothesize that its function is related to actin dynamics or organization on the basis of the amphiphysin localization and further experiments (see below).
Myosin VI and dynamin mutations genetically interact
The temperature-sensitive mutation in Drosophila dynamin, shibire (shi1), was used to determine if dynamin function was required for individualization. This mutation is in the GTPase domain of dynamin ( Grant et al., 1998) and is thought to block the GTPase cycle of dynamin ( Hinshaw, 2000), thereby acting as a functional null at the non-permissive temperature. Exposure of the dynamin mutant flies to the non-permissive temperature for up to 6 hours did not have any obvious effect on actin in actin cones or organization of individualization complexes ( Fig. 5A). After dynamin inactivation, myosin VI and cortactin localization and accumulation at the front of actin cones was not significantly affected either, although myosin VI accumulated abnormally elsewhere in individualizing cysts (A.D.R., unpublished). Individualization complexes moved normally when dynamin mutants were shifted to the non-permissive temperature as judged by observations in real time using cysts cultured in vitro (T. Noguchi, personal communication). Thus, dynamin inactivation alone did not have a strong effect on individualization complexes. However, striking defects were observed when the temperature-sensitive dynamin mutation was placed into the background of a hypomorphic myosin VI mutation, jaguar (jar1). When these dynamin myosin VI double mutants were exposed to the non-permissive temperature to inactivate dynamin function, the total number of individualization complexes per testis was dramatically reduced. In addition, most of the individualization complexes that could be observed stained very weakly with phalloidin in comparison with those in flies bearing only the dynamin or myosin VI mutations ( Fig. 5). We conclude that actin assembly or stability is dramatically reduced in the dynamin myosin VI double mutant. Therefore, we suggest that myosin VI and dynamin function in parallel pathways and that each pathway contributes to the regulation of actin structures in individualizing spermatids.
Actin stability defects in dynamin myosin VI double mutants. Confocal projections of mutant testes stained with phalloidin. Volume samples were taken near the basal end of testis where all the actin individualization complexes are associated with nuclei. Within each testis, several groups of 64 bundled spermatids are shown. (A) All actin individualization complexes in dynamin (shi1/Y) single mutants exposed to the non-permissive temperature for 6 hours stained strongly with phalloidin (white arrowhead). (B) Most actin complexes in myosin VI single mutants (jar/jar) exposed to the non-permissive temperature for 6 hours stained brightly with phalloidin, although a few disrupted complexes stained weakly with phalloidin (asterisk). (C) Few actin complexes in dynamin myosin VI double mutants (shi1/Y; jar/jar) exposed to the non-permissive temperature for 6 hours were visible, and those that were visible stained weakly with phalloidin (white arrow). All images are projections of multiple planes. Images of single and double mutants were taken at the same gain and scale.
We quantified the defects in myosin VI dynamin double mutants to demonstrate the severity of the effect on actin structures. In these experiments we exposed dynamin myosin VI double mutants, dynamin single mutants, myosin VI single mutants and wild-type flies (OreR) to the non-permissive temperature for different times then fixed and stained the testes with phalloidin to visualize actin. We counted the total number of early individualization complexes (complexes that are just assembling or that have just begun to move; Fig. 6A) and determined the proportion of individualization complexes that stained strongly or weakly with phalloidin ( Fig. 6B).
Quantification of number of actin complexes in dynamin myosin VI double mutants. Wild-type (OreR), dynamin mutants (shi1/Y), myosin VI mutants (jar/jar) or dynamin myosin VI double mutants (shi1/Y; jar/jar) were exposed to the non-permissive temperature for the times indicated. (A) The average number of early actin individualization complexes per testis is shown for each genotype and condition. (B) The average proportion of early actin complexes that stained weakly with phalloidin are shown.
In wild-type flies there were typically 13 early individualization complexes per testis, of which most stained strongly with phalloidin ( Fig. 6). These numbers were not significantly changed even after 6 hours at the non-permissive temperature. Similar numbers were observed for the dynamin mutants even after 6 hours at the non-permissive temperature. Therefore, exposing flies to as much as 6 hours of dynamin inactivation does not, obviously, affect actin structures.
In the myosin VI single mutants there was a slight reduction in the total number of early actin complexes relative to wild-type and dynamin mutant flies at all time points tested ( Fig. 6A). Approximately 40% of the early complexes stained weakly with phalloidin ( Fig. 6B). These findings are consistent with our previous results ( Hicks et al., 1999). In myosin VI mutants, actin complexes form, but as complexes begin to progress they fall apart and apparently lose actin filaments. This results in fewer total actin complexes in myosin VI mutant testes and accounts for the larger proportion of weakly staining complexes. Thus, absence of myosin VI alone has a slight effect on actin assembly and/or stability in early individualization complexes.
In comparison with wild-type controls and the single mutants, the dynamin myosin VI double mutant at the non-permissive temperature displayed a large reduction in the total number of early actin complexes per testis ( Fig. 6A; 3 hours and 6 hours). Moreover, in the double mutant a large proportion of the actin complexes that remained stained weakly with phalloidin (e.g. 66% after 6 hours at the non-permissive temperature; Fig. 6B), indicating that those few complexes remaining had greatly reduced F-actin levels. By contrast, when the dynamin myosin VI double mutant was exposed to the non-permissive temperature for only 2 minutes, the total number of actin complexes per testis and the number of weakly staining complexes were similar to the myosin VI single mutant ( Fig. 6). This demonstrates that dynamin myosin VI double mutants at the permissive temperature are similar to myosin VI single mutants, in that actin cones can assemble. Only after inactivation of dynamin are filamentous actin structures disrupted.
Since continuous actin assembly is often required to maintain stable actin structures, the loss of individualization complexes in the double mutant could be due to a disruption of actin filament assembly. Alternatively or in addition, the loss of actin filaments could be due to a direct disruption of actin filament stability. Therefore, we conclude that myosin VI and dynamin function in pathways that regulate actin assembly or stability in individualizing spermatids. We also conclude that, although dynamin activity is not required for actin cone formation, it plays a redundant role with myosin VI in regulating actin dynamics.
Discussion
A role for myosin VI in regulating actin dynamics
Our results suggest that myosin VI regulates the dynamics of actin structures during individualization. This idea is supported by these observations: (1) myosin VI colocalizes with the actin polymerization regulatory proteins cortactin, arp2/3 complex and capping protein; (2) cortactin and arp2/3 complex distribution are disrupted when myosin VI function is reduced; and (3) actin individualization complexes disassemble when both myosin VI and dynamin function are compromised. This is the first demonstration of a genetic interaction between myosin VI and dynamin. In addition, the observation that myosin VI can influence the localization of actin regulatory proteins is novel.
We propose a model in which myosin VI acts in a structural capacity in the individualization complex to regulate actin dynamics at sites of active membrane remodeling. Myosin VI might participate in actin dynamics solely by its influence on cortactin and arp2/3 complex localization. Cortactin can bind both actin and the arp2/3 complex ( Weed et al., 2000). Cortactin also enhances actin polymerization and stabilizes actin filaments during arp2/3-complex-dependent actin polymerization through its actin-binding activity ( Olazabal and Machesky, 2001; Weaver et al., 2001). Myosin VI could help localize or maintain the localization of arp2/3 complex and cortactin at the fronts of the actin cones. The localization of these components at the front of the cones would then facilitate actin assembly. This localized actin assembly would drive actin cone movement in a manner similar to the leading-edge protrusion or Listeria motility. In the absence of myosin VI, the fronts of the cones would lose assembly sites and thus actin cones would depolymerize. It is interesting to note that another unconventional myosin, myosin 1 from yeast and Dictyosteliyum, can interact with the arp2/3 complex and may have a direct role in actin assembly ( Evangelista et al., 2000; Geli et al., 2000; Jung et al., 2001).
Alternatively, myosin VI might directly stabilize actin at the front of the actin cones. Since myosin VI has a coiled-coil domain that is thought to mediate its dimerization, myosin VI itself could crosslink actin filaments and provide a stabilizing force to newly generated actin filaments. Loss of myosin VI function in this scenario would result in destabilization of actin filaments and the actin-binding proteins associated with those filaments such as arp2/3 complex and cortactin. Thus, in this case, loss of arp2/3 complex and cortactin concentration would be secondary effects owing to loss of actin filaments at the front of actin cones. At this point, our results do not distinguish between a direct or indirect effect of myosin VI on actin assembly/stability.
A role for dynamin in actin dynamics
To explain the actin defects we observe in dynamin myosin VI double mutants we also suggest that dynamin participates in a structural capacity by helping to organize or stabilize the actin filaments of the cones. The genetic analysis of the double mutants suggests that these genes function in parallel pathways. Moreover, the localization of the two proteins on actin cones is consistent with parallel separate pathways since their localization is distinct from one another. Since dynamin is located along the cone, but is not concentrated at the front, its role would be to help regulate actin dynamics or organize actin away from the front where myosin VI and cortactin are located. In this structural model, loss of both dynamin and myosin VI function would be predicted to have severe effects on the assembly and stability of actin filaments in the individualization complex, because two regulators of actin structure in the individualization complex would be lost. This is consistent with the dynamin myosin VI double mutant phenotype.
The temperature-sensitive mutation in dynamin that we used for these experiments was initially characterized in neurons where neuronal phenotypes can be detected within minutes of dynamin inactivation ( Koenig and Ikeda, 1989; Koenig et al., 1989). We observed that short periods of dynamin inactivation (2 minutes) in the myosin VI background have no effect on actin. The difference in time course of effect of dynamin inactivation might be explained by the inherently slower dynamics of individualization, a process that takes place over the course of many hours, as compared to the very rapid process of synaptic vesicle release and re-uptake in neurons that requires dynamin.
There is the formal possibility that dynamin could participate in clathrin-based endocytosis and not actin dynamics during actin cone movement, but several lines of evidence argue against this possibility. First, a protein normally associated with dynamin during endocytosis, α-adaptin, does not concentrate on actin complexes nor does it concentrate at the front of the complexes like myosin VI does. Instead α-adaptin localizes to the cystic bulge ahead of the cones. Second, electron micrographs of individualizing spermatids do not show evidence of high endocytic activity at the individualization complex ( Tokuyasu et al., 1972). Third, the robust disruption of the actin individualization complexes observed in the dynamin myosin VI double mutant is not easily explained by a defect in endocytosis. Fourth, amphiphysin localizes to actin cones in a similar fashion to dynamin. However, amphiphysin does not appear to be required for endocytosis in Drosophila ( Leventis et al., 2001; Razzaq et al., 2001; Zelhof et al., 2001). Although it is unclear if Drosophila amphiphysin binds directly to dynamin ( Razzaq et al., 2001), it is clear that amphiphysin can function in non-endocytic processes that involve actin structures and membrane morphogenesis ( Razzaq et al., 2001; Zelhof et al., 2001).
Moreover, there is precedence for dynamin involvement in actin dynamics in other organisms. For example, Dynamin-2 concentrates in actin-based podosomes and regulates actin assembly and turnover in these structures ( Ochoa et al., 2000). Dynamin-2 also localizes to actin comet tails of macropinocytes, Listeria and type I PIP-kinase-induced motile vesicles and regulates the assembly and structure of these actin tails ( Lee and De Camilli, 2002; Orth et al., 2002). The implication of dynamin in an actin assembly process is a novel finding in Drosophila, but bolsters the recent observations that dynamin functions to regulate actin structures in mammalian systems. This suggests that actin assembly regulation is another conserved function for dynamin.
A role for myosin VI in linking actin assembly to membrane dynamics
In addition to a role for myosin VI in localizing actin assembly proteins, we speculate that myosin VI might be important for coupling actin assembly sites to regions where the membrane is remodeled at the front of individualization complexes. On the basis of electron micrographs of individualizing spermatids ( Tokuyasu et al., 1972), membrane lies in close proximity to actin along the length of the cone, including the front of the cone, and vesicles, organelles and ribosomes are excluded from this region. The close proximity of the actin cones to the plasma membrane suggests they are tightly linked. The front of the actin cones is also the region where we see concentrated myosin VI and concentrated sites of actin-assembly-regulating proteins. Immediately in front of the cones there is an abrupt transition to the syncytial cytoplasm in the cystic bulge, a region of large vesicles, organelles and ribosomes. Membrane remodeling takes place at the junction between the front of the actin cones and the syncytial cytoplasm. During individualization complex progression, the connection between membrane and actin must be maintained, otherwise membrane remodeling would not progress synchronously with movements of the actin complex. Myosin VI, perhaps through its effects on cortactin localization, might be part of the structure that links actin and membrane as the individualization complex moves down the spermatids.
Moreover, the detection of cortactin at the front of actin cones where myosin VI accumulates indicates that this region is not only a site of active actin assembly but also a region of coupling between membrane and the actin cytoskeleton. Cortactin is thought to act as a link between membrane-associated kinases and actin assembly since it associates with proteins such as Src, Syk and the arp2/3 complex and concentrates in zones of membrane/actin linkage such as lamellipodia (Wu et al., 1993; Okamura and Resh, 1995; Maruyama et al., 1996; Weed et al., 1998).
Myosin VI may be carrying out similar functions in other cellular contexts
Our suggestion that myosin VI facilitates actin assembly through effects on the arp2/3 complex and cortactin and couples actin dynamics to membrane remodeling through either a direct or cortactin-dependent mechanism may be applicable to myosin VI function in other systems. For example, vertebrate myosin VI has been proposed to function in endocytosis ( Buss et al., 2001a; Buss et al., 2001b). During endocytosis, plasma membranes undergo dynamic morphological change in order to invaginate and form a budding vesicle. It has been proposed that F-actin networks help deform membrane during coated-vesicle formation and that F-actin polymerization helps propel endocytic vesicles away from the plasma membrane ( Qualmann et al., 2000). In these scenarios actin assembly is required and membranetopology changes must be coordinated with actin assembly. Given such requirements it is not surprising that cortactin has been suggested to function in endocytosis ( Jeng and Welch, 2001; Kaksonen et al., 2000). We speculate that myosin VI's role in this process may be to link actin assembly sites containing arp2/3 complex and cortactin to sites of membrane dynamics involving dynamin.
It is also tempting to consider that myosin VI might facilitate actin assembly and its coupling with membrane reorganization in other cellular contexts. Examples include membrane invagination in syncytial blastoderm embryos ( Mermall and Miller, 1995), membrane ruffling in EGF stimulated cells ( Buss et al., 1998) and stereocilia morphogenesis in developing hair cells ( Self et al., 1999).
Open questions
It is unclear if the ability of myosin VI to translocate along actin filaments is required for actin cone movement or function. In particular, it is not known whether the minus-end movement of myosin VI is important. The orientation of actin filaments in individualization complexes has not been visualized directly, although the capping protein and arp2/3 complex localization are consistent with the barbed ends of actin filaments at the front of actin cones oriented away from the spermatid nuclei. If this is true, then pointed-end motility would lead to myosin VI walking away from the advancing edge. This seems inconsistent with maintaining a high concentration of myosin VI at the front of the cones. Until the orientation of actin filaments is known, the mobility of membrane and myosin VI observed, and the dynamics of actin in the actin cones determined, it is premature to formulate specific models for the involvement of myosin VI's actin-based motility in this process. However, our studies of spermatid individualization are providing insight into the molecules required for actin cone assembly and movement, as well as myosin VI's specific role in this process. Our continuing studies will allow us to develop and test models regarding myosin VI's in vivo mechanism of action and ultimately determine the importance of its motile properties in vivo.
Acknowledgements
We are grateful to Debbie Frank, Robbie Hopmann and Tatsuhiko Noguchi, Magdalena Bezanilla, Phillip Harries and John Cooper for useful discussions about the project and critical reading of the manuscript. We thank Nick Gay, Mani Ramaswami, Manabu Takahisa, Bill Theurkauf, Andrew Zelhof and Matt Welch for providing antibodies. We also thank the Bloomington Stock center for providing stocks. This work was supported by National Institutes of Health grant GM-60494 (to K.G.M).
- Accepted September 4, 2002.
- © The Company of Biologists Limited 2002
References
Article navigation
Related articles
Cited by...
More in this TOC section
Similar articles
Other journals from The Company of Biologists
2020 at The Company of Biologists
Despite the challenges of 2020, we were able to bring a number of long-term projects and new ventures to fruition. While we look forward to a new year, join us as we reflect on the triumphs of the last 12 months.
Mole – The Corona Files
"This is not going to go away, 'like a miracle.' We have to do magic. And I know we can."
Mole continues to offer his wise words to researchers on how to manage during the COVID-19 pandemic.
Cell scientist to watch – Christine Faulkner
In an interview, Christine Faulkner talks about where her interest in plant science began, how she found the transition between Australia and the UK, and shares her thoughts on virtual conferences.
Read & Publish participation extends worldwide
“The clear advantages are rapid and efficient exposure and easy access to my article around the world. I believe it is great to have this publishing option in fast-growing fields in biomedical research.”
Dr Jaceques Behmoaras (Imperial College London) shares his experience of publishing Open Access as part of our growing Read & Publish initiative. We now have over 60 institutions in 12 countries taking part – find out more and view our full list of participating institutions.
JCS and COVID-19
For more information on measures Journal of Cell Science is taking to support the community during the COVID-19 pandemic, please see here.
If you have any questions or concerns, please do not hestiate to contact the Editorial Office.