Distinct Effects of Abelson Kinase Mutations on Myocytes and Neurons in Dissociated Drosophila Embryonic Cultures: Mimicking of High Temperature

Abelson tyrosine kinase (Abl) is known to regulate axon guidance, muscle development, and cell-cell interaction in vivo. The Drosophila primary culture system offers advantages in exploring the cellular mechanisms mediated by Abl with utilizing various experimental manipulations. Here we demonstrate that single-embryo cultures exhibit stage-dependent characteristics of cellular differentiation and developmental progression in neurons and myocytes, as well as nerve-muscle contacts. In particular, muscle development critically depends on the stage of dissociated embryos. In wild-type (WT) cultures derived from embryos before stage 12, muscle cells remained within cell clusters and were rarely detected. Interestingly, abundant myocytes were spotted in Abl mutant cultures, exhibiting enhanced myocyte movement and fusion, as well as neuron-muscle contacts even in cultures dissociated from younger, stage 10 embryos. Notably, Abl myocytes frequently displayed well-expanded lamellipodia. Conversely, Abl neurons were characterized with fewer large veil-like lamellipodia, but instead had increased numbers of filopodia and darker nodes along neurites. These distinct phenotypes were equally evident in both homo- and hetero-zygous cultures (Abl/Abl vs. Abl/+) of different alleles (Abl1 and Abl4) indicating dominant mutational effects. Strikingly, in WT cultures derived from stage 10 embryos, high temperature (HT) incubation promoted muscle migration and fusion, partially mimicking the advanced muscle development typical of Abl cultures. However, HT enhanced neuronal growth with increased numbers of enlarged lamellipodia, distinct from the characteristic Abl neuronal morphology. Intriguingly, HT incubation also promoted Abl lamellipodia expansion, with a much greater effect on nerve cells than muscle. Our results suggest that Abl is an essential regulator for myocyte and neuron development and that high-temperature incubation partially mimics the faster muscle development typical of Abl cultures. Despite the extensive alterations by Abl mutations, we observed myocyte fusion events and nerve-muscle contact formation between WT and Abl cells in mixed WT and Abl cultures derived from labeled embryos.


Introduction
The mammalian cytoplasmic Abelson tyrosine kinase gene (cabl) is an oncogene first revealed by a specific chromosomal translocation, the Philadelphia chromosome [1], found in patients with chronic myeloid leukemia [2,3]. Studies in vertebrate species have established that Abelson kinase is also important for other cell types, primarily through cytoskeleton organization through actin dynamics regulation [4,5].
In Drosophila, the homologous gene Abl (initially described as abl) [6] is known to exert wide-ranging effects in development and growth. For example, Abl mutant embryos display arrested motor axon outgrowth when targeting peripheral muscles [7]. Furthermore, interactions with fasciclin 1 (fas1), which encodes an adhesion molecule, further disrupts growth cone guidance and nerve track formation in the embryonic central nervous system (CNS) in Abl fas I double mutants [8]. Abl is shown to regulate growth cone motility mediated by actin cytoskeletal organization that is tightly regulated by its phosphorylation substrate Ena (homolog of VASP (Vasodilator-Stimulated Phosphoprotein) in mammals) [9,10]. While less extensively studied in myocytes, Abl has been shown to interact with the gene disabled (dab), causing musculature derangement in double-mutant embryos [11].
In the present study, we furthered the above in vivo findings by utilizing embryonic cell culture system. Our previously work has utilized neuronal cultures derived from dissociated larval CNS [12,13,14,15,16] or embryonic ''giant'' neurons cultures from cell division-arrested neuroblasts, in which Cytochalasin B treatment eliminates muscle cells [17,18,19,20]. To extend our observations to other cell types, in addition to neurons, we carried out experiments using the single-embryo culture system to study muscle cell development and nerve-muscle interaction.
The dissociated cultures were initiated at defined embryonic stages, which enabled us to study the developmental progression of distinct cell types and the interactions among them, as well as to distinguish between mechanisms mediated by cell-cell interactions or cell autonomous processes.
Here we report several findings that have not been characterized previously. First, muscle development in our cultures critically depended upon the stage at which embryos were dissociated. Second, Abl mutations differentially affect various aspects of myocyte and neuronal development. In particular, abundant muscle cells were present in Abl cultures dissociated at embryonic stage 10, while muscle cells were not seen in WT cultures until stage 12. Third, high temperature (HT, 30uC) incubation greatly enhanced neuronal and muscle growth and partially mimicked Abl myocyte phenotypes. Fourth, Abl nerve and muscle cells responded differentially to HT incubation, supporting the notion of distinct interacting partners of Abl in nerve and muscle development.

Drosophila Stocks
The primary wild-type (WT) strain was Canton S (CS), which was used for all statistics and images, except for Rhodamine 123 staining, where a second WT strain, Oregon-R (OR), was used. Two Abl alleles, Abl 1 and Abl 4 , were kept in marked and balanced stocks: Abl 1 kar red e/TM6B, Tb (from Bloomington Stock Center, Bloomington, IN) and Abl 4 kar red e/TM6B, Tb (from Dr.FM Hoffmann), and were used for making the appropriate crosses as specified in our culture studies. The balancer line w*; Sb 1 /TM3, P {ActGFP} JMR2, Ser 1 (Bloomington Stock Center, Bloomington, IN) was used to replace unlabeled balancer TM6B to make labeled heterozygous Abl alleles to facilitate genetic background checks in culture. All the Abl alleles used for cultures are Abl 4 /TM3-ser-GFP and Abl 1 //TM3-Ser-GFP. CS was crossed with both Abl 4 /TM3-ser-GFP and Abl 1 //TM3-ser-GFP to yield heterozygous Abl alleles under wild-type background for cultures. The ubiquitously expressed driver, daughterless (Da-Gal4) (Gift from Toshi Kitamoto) and the pan-neuronal driver, Elav-Gal4, were used to drive UAS-GFP, UAS-RFP (Bloomington Stock Center, Bloomington, IN).

Embryonic Primary Culture
The methodology for preparing single-embryonic cultures is a modified version from two previously described protocols for the Drosophila larval central neuron system (CNS) culture [12], Drosophila embryonic ''giant'' neuron culture system [17] and primary neuron cultures of Drosophila [21]. The culture system in this paper essentially is based on the ''giant'' Drosophila embryonic culture system without adding cytochalasin B into medium. Briefly, the stage 7-8 embryos were selected by the clear gastrulation marker, after egg laying 3:00 h-3:40 h; the stage 7-8 embryos were incubated in medium for an additional 2 hours, where the embryos progressed to late stage 10; 4 hrs after stage 7-8, embryos progressed to early stage 12; 7 hrs after stage 7-8, embryos progressed to stage 13. Tissues of the embryos were sucked by micropipette, and immediately plated in culture medium on a coverslip. The coverslips were put into moisture chambers, which were in turn placed in either room temperature (22-25uC) or a 30uC incubator for heat stress treatment. Throughout the study, observations on long-term HT effects were achieved by placing the moisture chambers in an incubator set to 30uC. The cultures were exposed to HT throughout the growth period, up until the time of microscope observation, as specified. The culture method for Figure S2 was based on previous research [22,23]. We homogenized 4-6 embryos from stage 10 by eppendorf micropestle in 80 ml culture medium, pelleted in a microcentrifuge at 5,000 rpm for 5 minutes at room temperature, resuspended by pipetman, and plated directly.
The culture medium contains 80% Drosophila Schneider medium and 20% fetal bovine serum (GIBCO, Invitrogen, Carlsbad, CA) with the addition of 200 ng/ml insulin, 50 g/ml streptomycin, and 50 U/ml penicillin (all from Sigma, St. Louis, MO). We carried out single-embryo and two-embryos mixed cultures. For Abl 4 /TM3-ser-GFP and Abl 1 //TM3-ser-GFP cultures, we checked maternal fluorescent signal after plating 2 days. Cultures of Abl alleles used for analysis were from both hetero and homozygous stocks, both of which showed similar phenotypes.

Fluorescent Staining
Immunohistochemical staining. The staining protocol was based on our previous protocol for larval CNS culture [24]. Cultures were briefly washed in phosphate-buffered saline (PBS, pH 7.2), fixed for 10 minutes with 3.7% formaldehyde, washed in PBS 5 times for 5 minutes each, and permeabilized with 0.01% Triton X-100 for 2 minutes. After blocking with 3% bovine serum albumin (BSA) for 10 min, cells were washed 5 times for 5 minutes each, incubated at room temperature for 25 minutes with TRITCconjugated phalloidin (0.05 mg/ml, Sigma), followed by 5 washes in PBS. After staining, they were mounted in VECTASHIELDH Mounting Medium with DAPI to stain nuclei or without DAPI (49, 6-diamidino-2-phenylindole) (Vector Lab). The cultures were observed and photographed on a Leica inverted microscope.
Rhodamine 123 staining. Cultures were incubated with 1 mg/ml Rhodamine 123 (molecular probes) in Drosophila culture medium for 2 minutes, washed with fresh medium 3 times, and then mounted in culture medium immediately for observation.

Image Processing and Quantification
We used 20X objective lens (N.A. = 0.4, Nikon) on an inverted microscope (Nikon, Japan), 40X objective lens (N.A. = 0.55,Leica) and 100X oil immersion objective lens (N.A. = 1.3,Leica) on an inverted microscope (DM IRBE, Leica, Wetzlar, Germany) or an upright microscope (DM RBE, Leica, Wetzlar, Germany) for collecting images of cultured cells. Images were processed and analyzed using ImageJ, Adobe Photoshop CS, Gimp and Inkscape. ImageJ was used for quantifications of filopodia length, and lamellipodia area. The cells clusters counted were less than 30 cells with an area of cell bodies occupying less than 2500 mm 2 . To be counted, filopodia along neurites had to be at least 5 mm away from cell bodies. Muscle lamellipodia were defined as muscle terminals with a flattened, well-expanded area of at least 5 mm 2 .

Neurons and Muscle Cells Grown in Dissociated Cultures: Embryonic Stage-dependent Myocyte Development
In this study, we performed single-embryo cultures grown on uncoated glass coverslips to observe the growth and differentiation potentials of dissociated nerve and muscle cells, as well as their interactions. Individual embryos collected at different times after egg-laying were staged and prepared for plating under a dissecting microscope. We strove for the optimal extraction of entire contents from individual embryos with a micropipette. The tissue was subsequently extruded and dispersed in a drop of medium on the coverslip surface, without mechanical agitation or enzyme treatment. Consequently, both isolated cells and cell clusters were present in the culture. Presumably, cell groups with stronger adhesion were more likely to remain in associated cell aggregates. In such cultures, both nerve and muscle cells could be successfully grown, but the appearance of myocytes critically depended upon the developmental stage at which the embryonic tissue was extracted and plated. As Figure 1 shows, there were abundant muscle cells in cultures dissociated from older embryos (stage 13 and beyond) that displayed differentiated features, e.g. striations and occasional contractions, when grown at room temperature for more than 1 day ( Fig. 1C; data not shown). Generally, in cultures from stage 12 or older embryos, muscle cells could be readily found in open fields of coverslip a few hours after plating, whereas stage 10 cultures rarely produced any isolated muscle cells, even after several days of incubation, while neuronal development proceeded vigorously with extended neurites deco-rated with filopodia and lamellipodia (Fig. 2). Thus, there was an apparent critical transition between embryonic stage 10 and stage 12 for myocyte development in single-embryo cultures.
It should be noted that cultures prepared from stage 10 embryos occasionally gave rise to a few myocytes in open fields (data not shown). However, further incubation over several days indicated poor survival of these myocytes and no additional myocytes appeared in culture ( Fig. 2A). Conversely, in stage 12 cultures, myocyte aggregates of various group sizes continued to grow and differentiate into multinucleated fused cells beyond 4 days and up to 18 days (Fig. 2B). Additionally, neuromuscular contacts were frequently encountered in these cultures (Fig. 2C). Notably, in stage 12 cultures, there were two distinct muscle locations; either isolated from neuronal clusters (Fig. 2B1, arrow) or still in the process of emerging from nerve tissue (Fig. 2B1, arrowhead). In both cases, myotube differentiation continued as indicated by the fused, multinucleated muscle cells (Fig. 2C) that often exhibited contraction behavior with or without innervation. Time-lapse observations confirm myocyte movements and changes in shape during this translocation process (data not shown).
This result suggests that in our single-embryo cultures, the tendency of myocyte migration from tissue aggregates and the abundance of individual or groups of myocytes in open fields increases with embryo age. This apparently reflects the in vivo observations that stage 10 embryos contain undifferentiated myoblasts and muscle cell proliferation does not occur until stage 12 [25].

Abelson Kinase Mutations Exert Dominant Effects on Myocyte Development and Neuronal Growth in Culture
In our previous study, we found altered neuronal morphology and growth in dissociated larval CNS cultures of Abl mutants, in which defective cytosolic tyrosine kinase Abl leads to increased plating efficiency, enhanced growth and motility of filopodia, and altered neurites with swollen nodes and growth cone morphology [24]. These cellular phenotypes are likely derived from disrupted actin cytoskeleton regulation.
In the present single-embryo culture system, we discovered striking effects of Abl on embryonic muscle development in addition to abnormal neuronal growth. We observed abundant isolated muscle cells in open fields of coverslips as well as muscle cells emerging from, but still associated with, neuronal clusters in cultures prepared from embryos as early as stage 10 (Fig. 3).
While WT cultures derived from embryonic stage 10 were devoid of exposed, readily detectable muscle cells (Fig. 3A), Abl cultures were populated with single or multinucleated muscle cells, both isolated and still connected with cell clusters (Fig. 3B1). These Abl muscle cells, including multinucleated, fused myocytes, displayed well-expanded lamellipodia with filopodia projections (Fig. 3B2-B3 & see HT effects below). Consistent with the altered neuronal growth in the larval CNS culture [24], neurons in Abl cultures displayed increased numbers of filopodia and more swollen, dark nodes along neurites (Fig. 4B1). However, compared to WT control, Abl neurons exhibited tendency to show fewer veillike structures and well developed, large lamellipodia (.4 mm 2 ) occurred only at a lower frequency (Fig. 5).
As stated in Methods, the majority of homozygous Abl mutant alleles are lethal at pupal or earlier stages, so we performed most of the experiments on Abl 4 /+ (about 70% of total cultures) and some on Abl 1 /+ (about 10%) heterozygous embryos, both of which demonstrated striking phenotypes that were readily detected. We have also collected data from Abl 1 /Abl 1 embryos (about 10%) as well as some Abl 4 /Abl 4 embryos (less than 10%). Surprisingly, the neuronal and myocyte phenotypes observed were similar in all four types of culture ( Figure S1). The representative photomicrographs presented in this paper were collected from Abl 4 allele cultures, but the statistics for phenotypic characterization shown in the bar charts throughout this report are based on pooled data from cultures of all four genotypes. Furthermore, we observed a few cultures derived from Abl 1 /Abl 4 embryos and found similarly extreme phenotypes of neurons and myocytes (data not shown). There is also an increase in dark nodules along the neurite and at the terminal (B1, arrowheads). However, while phase-light growth cones with expanded lamellipodia are rare in Abl, they can readily be seen in WT (A1, arrow). A1 vs. A2) HT incubation of WT neurons produces extremely large growth cones (A2, arrows). B1 vs. B2) HT incubation of Abl neurons enhances development of phase-light lamellipodia at growth cones and along neurites (B2, arrows). All cultures were incubated 3-6 days. Scale bars, 10 mm. doi:10.1371/journal.pone.0086438.g004 The above results indicate that both Abl 4 and Abl 1 mutations exhibit dominant effects on both nerve and muscle cell development.

Enhanced Neuronal Growth Promoted by Hightemperature Incubation in WT and Abl Cultures
The above observations suggest that cytosolic Abl tyrosine kinase may play differential roles in regulating veil-like structures, such as growth cone lamellipodia and muscle lamellipodia, in nerve and muscle development. To accentuate these potential differences, we subjected these two cell types to temperature stress since our previous research has shown striking high temperature (HT)-induced increases in growth cone size and terminal arborization in embryonic ''Giant'' neuron cultures [26].
In stage 10 WT cultures, we found that incubation at HT (30uC) tended to increase the number of neuronal lamellipodia over room temperature (RT) WT cultures (Fig. 5A). Strikingly, HT incubation of stage 10 Abl produced a drastic increase over RT in the number of neuronal lamellipodia (.300%), raising the average to levels comparable to those found in WT (Fig. 5A). Furthermore, HT-incubated neurons in both WT and Abl displayed enhanced neuronal growth via larger lamellipodia with greatly enlarged sizes (.10 mm 2 ), which are normally less than 10 mm 2 at RT (Fig. 5B). Therefore, HT incubation clearly enhanced neuronal growth through the increase in lamellipodia size. Additionally, the length of filopodia projecting along neurites tended to be increased in HT-incubated WT (Fig. 5D). There were more lengthened (.10 mm long) filopodia at HT, which were rare in cultures grown at RT.
However, the number of filopodial projections along neurites decreased in Abl cultures (Fig. 5C-D), further demonstrating that Abl regulates lamellipodia and filopodia in different ways.

High-temperature Incubation of WT Cultures Mimics Abl Muscle Phenotypes but Exerts Little Effect on Cultured Abl Myocytes
We discovered an unexpected, striking effect of HT incubation on myocyte migration and cell fusion in cultures derived from stage 10 embryos (Figs. 6 & 7). This is reminiscent of the normal developmental progression of WT embryos later at stage 12 or ; Abl HT, 151, respectively. Note that over 60% of WT lamellipodia at RT fall between 5-10 mm2, whereas at HT over 70% of lamellipodia are 10 mm2 or larger, even reaching areas over 25 mm2. In contrast, 80% of lamellipodia at RT are less than 10 mm2, whereas at HT over 60% are 10 mm2 or larger in Abl cultures. C) Numbers of filopodia per cell cluster. Neurons in Abl cultures show a greater number of filopodia at RT than WT but show a significant decrease in filopodial numbers following HT incubation (cell clusters numbers: WT RT, 74; WT HT, 66; Abl RT, 72; Abl HT, 102, respectively, **p,0.01, unpaired T-test). D) Distribution of filopodial lengths (mm). HT incubation does not produce a significant effect on filopodial length. There is a slight tendency of increased abundance of longer filopodia in WT cultures following HT incubation but x2 test shows no significant differences. For each culture conditions at least 3 coverslips were used and sequential Bonferroni adjustments for multiple comparisons were performed. All cultures were derived from stage 10 embryos and incubated 2-5 days. Error bars, SEM. The statistics here and in Figure 7 are based on pooled data from Abl4/+, Abl1/+, Abl4/Abl4, and Abl1/Abl1. See text. doi:10.1371/journal.pone.0086438.g005 older, which initiates myocyte migration and myotube formation (Figs. 1, 2&6).
Within 4 days in WT stage 10 cultures grown at HT, a higher percentage of muscle cells were fully separated from nerve clusters (Figs. 6 & 7) and many fused myotubes formed with nuclear numbers beyond 7 (Fig. 7A, C). HT muscle cells did display some well-expanded lamellipodia occasionally (Fig. 7B) and many putative innervations or interactions with neurites or growth cones from nerve cells (Fig. 6B2-B3). We observed muscle cells in the process of elongation and fusion at HT, as shown in the timelapse photography (Fig. 8), mirroring the in vivo muscle developmental sequence [27].
It should be noted that in our study these stage 10 embryos followed a normal developmental course at RT up to the time of plating. Therefore, the intrinsic capability of myocyte migration has developed by stage 10, even though the myocytes do not emerge from cell clusters when grown at RT. In support of this notion, stage 8 cultures also displayed some isolated muscles cells or aggregates when incubated at 30uC, although these muscle cells were short-lived and could only survive approximately 2 days after plating (data not shown). In addition, we carried out stage 12 cultures of Abl at RT and WT at HT, both of which displayed well-developed muscle cells, with enhanced muscle lamellipodia growth and cell fusion (Fig. S2). Together, these results suggest similar enhancements in muscle development by Abl mutation and HT treatment.
One interesting question was whether HT incubation and Abl mutations have synergistic effects on muscle development. Following the same protocol of HT incubation for WT, we found that muscle cells in Abl single-embryo cultures did not appear to be further promoted in muscle phenotypes. There was no significant increase in the numbers of muscle cells and muscle lamellipodia (Fig. 7A-B). However, the myocyte fusion rate appeared to be slightly increased at HT, as indicated by the increase in multinucleated muscle cells (Fig. 7C). The results from stage 10 Abl embryonic cultures demonstrate that HT exerted only limited effects on muscle phenotypes despite its extensive enhancement of Abl neuronal phenotypes. The apparent distinctions of the HT Apparently, high-temperature incubation of WT cultures mimics some, but not all, aspects of Abl muscle phenotypes. Muscle lamellipodia are a characteristic organelle of cultured embryonic muscle cells, seen often in cultures from more advanced embryos (stage 12), which are thought to play important roles in cell movement and cell-cell interactions. We observed that muscle cells in Abl cultures incubated at RT and HT presented far more numerous well-expanded lamellipodia than WT muscle cells grown at HT (Figs. 7B & 9). Furthermore, interacting protrusions from both neuron and muscle in the process of establishing connections were observed using time-lapse photography. This phenomenon was more abundant in Abl cultures at both RT and HT (Fig. 9E).
Despite the apparently altered Abl muscle development, the basic properties that define the muscle cell identity seem to be preserved in cultured Abl myocytes. To examine whether normal WT nerve and muscle cells could interact with their Abl counterparts, we took advantage of RFP-labeled embryos to enable vital marking of developing cells of identified genotypes. The expression on RFP in the tissue of WT embryos was controlled using the ubiquitously expressed Da-GAL4 driver, daughterless. We mixed cultures derived from WT and Abl embryos and found fusions among WT and Abl myocytes to form multinucleated muscle cells (Fig. 10A). Furthermore, neurite projections from WT neurons made contacts with Abl myocytes in culture (Fig. 10B). The converse was also true (data not shown). This observation demonstrates that Abl mutations do not prevent    cell-cell recognition, interaction and fusion with the WT counterparts.

Recapitulation of Embryonic Muscle and Nerve Cell Development and Alterations by Abl Mutations in Dissociated Single-embryo Cell Cultures
We discovered that Abl mutations profoundly alter the developmental progression and properties of myocytes, including myocyte movement, lamellipodia formation, and fusion process in the single-embryonic culture. Multinucleated muscle cells and neuron-muscle contacts were present in open fields of Abl cultures derived from stage 10, while WT embryos only at a later stage (12), but not at stage 10, could produce isolated single or grouped muscle cells in dissociated cultures. Also, similar to our earlier observations on dissociated larval CNS cultures [24], embryonic Abl neurons displayed altered morphology with dark nodules along neurites and more abundant filopodia.
During in vivo development, neuroblasts and myoblasts are determined before stage10 [28]. The nervous system starts to develop after gastrulation, while muscle differentiation occurs at stage 12 with signs of cell fusions [21]. There are two temporal phases of in vivo fusion events. During the first fusion phase (stage 12-13), founder cells (FCs) fuse with fusion competent myoblasts (FCMs) and a 2-3 nucleated precursor cell forms. Subsequently (stage 14-15), more massive fusion events between precursor cells and FCMs occur until the final size of the muscle is reached [29,30].
Factors affecting the above developmental progressions in the nerve and muscle systems still remain to be fully elucidated. The Drosophila primary embryonic culture systems have been contributing to the studies of cellular development under controlled conditions since the 1970s [31,32,33,34]. Cells from different developmental stages of embryos can differentiate into a variety of cell types in culture [35,36].
It has been reported that such primary culture systems maintain the formation of in situ-like neuron lineages [37,38,39]. Our previous studies using the Drosophila primary culture system of dissociated neurons from the larval CNS and the embryonic ''giant'' neuron culture derived from division-arrested neuroblasts have also demonstrated lineage-dependent neuronal development, as well as basic neuronal function via electrophysiology and optical imaging [40,41,42,43].
More recently, muscle cell development has been successfully studied in dissociated cultures to observe morphogenesis and to define effectors in the development process using gastrulae stages. Typically, these studies utilize cultures established by mixing and dispersing the contents from large numbers of embryos that have undergone extensive mechanical dissociation or enzyme treatment [22,23]. However, it should be noted that in our current culture system, muscle cells do not migrate from cell clusters extracted from single embryos prior to stage 12. This enabled us to detect the striking Abl phenotype of muscle migration and maturation.
The crucial technical differences were that we extracted contents from individual embryos and extruded them directly into the culture medium onto untreated glass coverslips. This procedure partially preserved the embryonic environments, especially the aspect of cellular aggregation at the specific stages of embryonic development. This was reflected by the distinct appearance of cell clusters and isolated single cells that enabled us to study, at a subcellular resolution, processes of myocyte movement, fusion and interactions with neurons (Figs. 2 & 3).
In a separate set of experiments, we treated stage 10 embryos with mechanical dissociation and dispersed the mixed contents to establish dissociated cell cultures. Indeed, consistent with published results [22,23], we observed individual muscle cells undergoing migration and fusion, indicating that embryonic myocytes at stage 10 are capable of migration and fusion when isolated in culture (Fig. S3).
Our results also suggest the possibility that developmental progression of muscle cells is preserved in our primary culture (Fig. 2). The serendipitous finding of the striking effect of Abl mutations in this culture system provides a stark contrast of its differential effects on nerve and muscle cells (Fig. 3 & 4), raising the probability that Abl may function in neuron and muscle through distinct interacting partners.

Abl and its Diverse Interaction Partners in Different Cell Types, Cellular Organelles and Subcellular Compartments
The major phenotypic alterations of Abl nerve and muscle cells in our cultures include the enhanced growth of muscle lamellipodia and filopodia along neurites (Figs. 3, 4, 5 & 7). There were also indications that Abl neuronal filopodia became more numerous at the expense of growth cone lamellipodia expansion, a defect that could be overcome by HT incubation (Figs. 4 & 5). However, HT incubation did not reduce the dark nodules along neurites, indicating that the production of dark nodules and excessive filopodia are derived from different mechanisms (Fig. S4). Apparently, Abl normally functions to exert inhibitory growth control of neuronal filopodia, and possibly muscle lamellipodia as well (Figs. 4 & 9).
In Drosophila and other species, Abelson tyrosine kinase has been implicated in cell migration, adhesion, cell-cell interaction, and morphogenesis in a variety of tissues during development [8,11,44,45,46]. Among these functional and developmental regulations, many of the Abl-regulated neuronal and muscular processes are related to cytoskeleton dynamics [5,47]. It is known that Abl family kinases convey signals from cell surface receptors for growth factors, such as platelet derived growth factor (PDGF), epithelial growth factor (EGF), netrin, and adhesion receptors, such as cadherin, to regulate cytoskeletal dynamics by phosphorylating several regulatory proteins, such as b-catenin [45,48,49].
For example, during axonogenesis, in Drosophila, Abl regulates cell adhesion and actin polymerization through suppressive/ antagonistic interactions with its substrates, the Ena/VASP family of actin regulatory proteins [10,50] and Abl interacting protein, Abi [51]. Moreover, Ena functions together with Diaphanous (Dia)-related formins, to regulate the balance between filopodia and lamellipodia. It is known that the presence of either Dia or Ena, is sufficient to induce filopodia growth, but only together will they induce lamellipodia protrusion. Interestingly, constitutively active Dia increased filopodia growth, mimicking our Abl neuronal phenotype [52]. It will be interesting to investigate in future studies whether our observations of filopodia overgrowth in Abl neurons also involves dysregulation of Dia-related formins. In vertebrates, c-Abl and Ena/VASP also regulate the leading edge mediated by a specific intermediary protein, Lamellipodin, which is required for lamellipodia formation during axonal morphogenesis [53,54]. It will be important to identify the Drosophila homologous gene(s) encoding Lamellipodin-related protein(s) in order to elucidate the mechanism of enhancing muscle lamellipodia formation and its relation to muscle cell movement and fusion, and to explain the apparently opposite effects on neuronal growth.
In addition to the identification of Abl interacting partners, another important issue is the temperature dependence of these protein activities. Such information may provide insights into our findings, specifically that HT incubation promotes myocyte migration out of cell clusters (Figs. 6 & 7) mimicking the Abl muscle phenotype, enhances neuronal lamellipodial growth, and reverts Abl neuronal filopodium-lamellipodium phenotypes.
In the current study, we have focused on the alterations of the cellular organelles lamellipodia and filopodia in Abl cultures. However, in Abl neurons, another striking phenotype is the characteristic dark nodules along the neurites and darkened central domains of the growth cone (Fig. 4) [24]. Our initial efforts indicated that these structures may be enriched in membranous organelles, such as mitochondria, as evidenced by Rhodamine 123 staining (Fig. 11). In phalloidin staining experiments, these regions are also enriched with actin, suggesting abnormal actin filament organization (data not shown). It should be noted that dark nodules in Abl neurons grown at HT were still abundant and enriched with organelles such as mitochondria (Fig. S4), a further contrast with the HT mitigation of filopodial defects. Future experiments should determine whether additional cellular organelles accumulate in these structures. This may provide indications for modifications in trafficking of these organelles, which is another important cellular mechanism dependent upon functional cytoskeletal cables.
In future studies, it will be important to determine the differential partners of Abl in the regulation of lamellipodia and filopodia in neuron and muscle cells and their temperaturedependent properties. Our culture system offers the potential of growing nerve and muscle cells with identified lineages and with vital cell markers, thus enabling the study of interactions among nerve and muscle cells of different genotypes which can then be monitored at high temporal and spatial resolutions.