Distinctive Effects of Cytochalasin B in Chick Primary Myoblasts and Fibroblasts

Actin-based structures play fundamental roles in cellular functions. However it remains controversial how cells cope with the absence of F-actin structures. This report focuses on short- and long-term effects of cytochalasin B (CB) on actin-complexes in fibroblasts and myoblasts. Thirty min of CB treatment dispersed subplasma actin cortices, lamellipodia, ruffled membranes, stress fibers and adhesion plaques into actin patches in fibroblasts and muscle cells. In contrast, 72 hrs CB treatment showed distinct morphological effects. Fibroblasts became giant multinucleated-finger shaped with 5 to 10 protrusions, 3–8 μm in width, and >200 μm in length. They lacked cortical actin, stress fibers, adhesion plaques and ruffled membranes but contained immense lamelliopodia with abnormal adhesion plaque protein complexes. Muscle cells transformed into multinucleated globular-shaped but contained normal I-Z-I and A-bands, indicating that CB did not interfere with the assembly of myofibrils. Within 30 min after CB removal, finger-shaped fibroblasts returned to their original shape and actin-containing structures rapidly reappeared, whereas muscle cells respond slowly to form elongated myotubes following CB washout. The capacity to grow, complete several nuclear cycles, assemble intermediate filaments and microtubules without a morphologically recognizable actin cytoskeleton raises interesting issues related to the role of the actin compartments in eukaryotic cells.


Introduction
In many cell types the conserved, actin-based cytoskeleton consists of 5 kinetically labile substructures, each subject to rapid changes. They are the (1) lamellipodia, (2) ruffled membranes, (3) stress fibers, (4) adhesion plaques and (5) subplasma actin cortices. Over the past decades numerous studies have focused on the regulatory molecules controlling the assembly/disassembly of the first 4 sub-structures and how they mediate directional locomotion. The outcome of this work has been the delineation of complex interactions between many signaling and structural molecules. Early studies emphasized the upstream activity of Rho, Rac, and Cdc42 in regulating the assembly, respectively, of stress fibers and adhesion complexes, lamellipodia and/or ruffled membranes and filopodia [1]. These in turn activate formins, WASP, Scar/ Wave, and still more downstream molecules like ERM, Arp2/3 and associated proteins [2][3][4][5][6][7]. These findings have converged with elegant immuno-electron microscopy observations demonstrating that the Arp2/3 complex nucleates and caps actin monomers at the sides of preexisting filaments which subsequently elongate by their barbed ends [8][9][10][11]. These complementary findings have provided a compelling molecular/structural overview suggesting how branched F-actin networks might drive translocating lamellipodia, as well as leading to the assembly of adhesion complexes and their associated stress fibers.
The challenge of how coordinated interactions between these 4 actin sub-structures regulate cell migration has tended to eclipse efforts to understand the assembly and function of the remaining major cytoskeleton component, namely the subplasma cortical actin lattice. In erythrocytes, where it has been studied most intensively, the plasma lipid bilayer is supported by a cortex consisted by a network of short actin/tropomyosin filaments capped at their pointed ends and barbed ends by tropomodulin and adducins, respectively [12]. These in turn are cross-linked into a planar hexagonal lattice by flexible isoforms of spectrin [12][13][14]. The subplasma actin/spectrin network, in conjunction with a variety of transmembrane molecules, conforms in outline with the juxtaposed plasma membrane proper [15].
Primary embryonic chick myogenic cultures consist of approximately equal numbers of precursor replicating fibrogenic and myogenic cells. It is an important tool to address the role of different actin compartments because these two cell types will assemble quite different actin structures upon cell differentiation. Both cell types display actin-based cytoskeletal structures including prominent subplasma actin/spectrin cortices. Previous studies [16][17][18] reported that while some of the actin based structures in these fibrogenic and myogenic cells responded similarly to the inhibitor of actin polymerization cytochalasin B (CB), others responded in a cellspecific manner. For example, while blocking cytokinesis, but not nuclear duplication, CB did not block the terminal differentiation of myogenic cells and accumulation of either fibrogenic or myogenic cells. One consequence of these effects was that after growing for 72 hrs in CB these cultures displayed 2 distinct multinucleated phenotypes. One was a giant, flattened vimentin-positive arborized cell termed a "finger-shaped fibroblast" (see below). The other, a globular, desmin-positive cell that was clearly myogenic, for it assembled I-Z-I bands of polarized 1.0 μm thin filaments that interdigitated with A-bands. These globular cells contracted spontaneously and occasionally. The limitations to these early studies, however, were both methodological and conceptual. These experiments were performed before the advent of fluorescent phalloidin [19] and confocal microscopy, as well as before actin isoforms were discovered [20]. As a consequence, many of the effects of CB on the multiple actin containing structures in both fibrogenic and myogenic cells were missed.
The current availability of fluorescent phalloidin plus many new antibodies to cytoskeletal and myofibrillar proteins, together with confocal microscopy techniques, prompted us to reasses the short-and long-term, as well as the strikingly reversible, effects of CB on the actin based structures in cells in these primary chick myogenic cultures. Here we show that after short-term treatment with CB, both fibroblasts and myogenic cells lose their filamentous cytoplasmic and membrane-actin structures in a very similar way. By contrast, long-term CB treatments lead to distinct morphological impact on both cells, i.e. finger-shaped fibroblasts and mushroom-like myogenic cells. CB-treated fibrogenic cells, despite lacking for 3 consecutive days a morphologically detectable filamentous actin-cytoskeleton, perform a variety of basic functions. They repeatedly cycle through G1, S, G2, prophase and metaphase forming giant, multinucleated, finger-shaped fibroblasts that display enlarged but normal microtubule-and vimentin intermediate filaments-networks. Following CB-washout, these multinucleated fibroblasts assemble a normal complex actin cytoskeleton. In contrast to fibroblasts, the mushroomshaped globular myogenic cells are myofibrillar protein positive and contain mono-, bi-, or 4 + -nuclei. These myogenic cells contract spontaneously and occasionally and show normal I-Z-I-bands interdigitating with normal A-bands, indicating that CB do not interfere with the assembly of myofibrils but did interfere with their elongation mechanisms. We also detect atypical distribution of microtubule and intermediate filament structures after long-term CB treatment, which may be caused by the mechanical modifications induced by the broken down of actin networks.

Effects of 30 min CB treatment on fibroblasts and muscle cells
Since these day-2 (the next day of plating) muscle cell cultures are composed of young myoblasts (~50%) and mesenchymal stem cells (~50%), normally the post-mitotic myoblasts are going through myofibrillogenesis including cell fusion to develop multinucleated myotubes in the coming 2-3 days while the presumptive myoblasts need yet to complete their last 1-3 mitosis before start myofibrillogenesis. On the other hand, mesenchymal stem cells are actively mitotic to produce most of the fibroblasts. We used CB to disrupt F-actin and to study actincontaining structures of these day-2 muscle cell cultures. Firstly, we focused on the short-term treatment on fibroblastic cells. A 30 min treatment with CB was sufficient to round up fibroblasts and muscle cells (compare Fig 1A and 1B with Fig 1C and 1D). Morphologically both cell types were indistinguishable without probing with antibodies. Preliminary observations suggested that the response to CB depended on whether the fibroblastic cells were exposed for 15 or 30 min, or grown continuously in CB for 72 hrs. While a 15 min exposure to CB totally dispersed the stress fibers, adhesion plaques, lamellipodia and ruffled membranes, short fragments of the cortical actin rim often remained (not shown). However, total elimination of the cortical actin rim, which may constitute 60-90% of their perimeter in control spread cells ( Fig  1A and 1B), required 30 min CB exposure (Fig 1C and 1D). Concomitant with the fragmentation of all actin-containing structures, numerous scattered actin aggregates or patches emerged throughout the cell (Fig 1C), whereas microtubules seemed to form tightly fasciculate bundles in CB treated fibroblasts ( Fig 1D). These actin-containing structures varied in their sizes (~3.0 μm) and in their numbers (20-50/cell, n = 100 cells from 3 different experiments, significantly different from untreated cultures, p<0.05). Some were irregular in shape, others more crystalloid. Each patch was precisely stained with Rho-labeled phalloidin. Importantly, despite the displacement of actin from its normal sub-plasma membrane location, under the light microscope the physical integrity of the sub-plasma membrane seems to be maintained ( Fig  2C). Furthermore, the displacement of F-actin complexes from the plasma membrane did not obviously impair the selective permeability or the functions of the treated cells, e.g. roughly only 3% of both control-untreated and treated cells were stained with trypan blue (n = 100 cells from 3 different experiments). Thus, we can conclude from the trypan blue exclusion assay experiments that the disorganization of actin filaments by CB did not impair the selective membrane permeability of the cells. The most convincing evidence for the apparently normal metabolic activity of cells lacking detectable actin-based cytoskeletal structures is illustrated in Fig 1C, 1D, 1E and 1F. Cells that lacked visible actin-based cytoskeleton structures just prior to CB-washout assembled one that was indistinguishable from control-untreated cells (compare Fig 1A and 1B with Fig 1E and 1F).
Given that actin-based cytoskeletal structures play roles in multiple functions and have both a complex molecular composition and global distribution, the implosion of the actin-based cytoskeleton by CB raises interesting issues: (1) the fate of molecules, other than actin, in such treated cells, (2) the molecular composition of the actin patches, which can be particularly relevant to their functions, and (3) the metabolic options of fibroblasts lacking morphological actin cytoskeletal structures for short periods (e.g. 30 min of CB treatment) versus long periods (e.g. 72 hrs of CB treatment).
In a detailed immuno-electron microscopy analysis, coupled with decorating actin filaments with myosin subfragment 1, Verkhovsky and colleagues [21] have reported that in rat fibroblasts CB-induced patches consisted of clustered actin filaments whose barbed ends oriented towards and converged on a myosin II core. Their electron microscopy images encouraged us to further investigate what type of proteins appeared in actin patches under fluorescent light microscopy. Specifically, we asked whether actin patches were heterogeneous or homogeneous in terms of composition, and if the former which proteins might be packed into an actin patch. To this end, CB-treated cultures were double-stained with Rho-phalloidin and one of the following antibodies to adhesion plaque proteins: paxillin, vinculin, or β1-integrin (Fig 2A-2L). Seventy-eighty percent of the F-actin patches in each of the 4 sets of cultures (n = 100 cells from 3 different experiments) were positive for the respective adhesion plaque proteins.  Obviously, the CB-induced actin patches were not promiscuous traps for non-specifically binding of other major cytoplasmic proteins such as tubulin (Figs 1D and 2B).
Interestingly, after a 30 min treatment with CB many β1-integrin positive-intracellular vesicles were observed at the ventral surface (where the cells are attached to the substrate) of the fibroblastic cells (Fig 2J-2L). These vesicles also contained F-actin, since they were positive for phalloidin (Fig 2J-2L). β1-integrin was found in these intracellular vesicles as well as in the Factin patches, together with other adhesion plaque proteins, such as talin, vinculin and paxillin.

Reassembly of the actin-containing structures in fibroblasts after recovery from 30 min CB
The kinetics and sequence of transitory macromolecular structures during recovery was followed by growing CB treated cells in normal medium for 5 and 15 min after CB-washout. While many cells in the 5 min recovery series were shapeless, unexpectedly 55% (n = 100 cells from 3 different experiments) displayed the overall triangular/parallelogram morphology common to many well-spread immotile fibroblasts (Fig 2M-2O). At all stages of recovery F-actin structures were always detected with other associated proteins, such as vinculin (Fig 2M-2O). The earliest recognizable reconstituted F-actin structures at this early stage were the subcortical actin lattices. In 30 min recovered cells, as in control-untreated cells, adhesion plaques rarely associated with the actin cortex but were largely confined to the distal tips of stress fibers ( Fig  2M-2O).
Reliable criteria distinguishing lamellipodia from ruffled membranes have yet to be established. Our findings show that there were significant behavioral and structural differences between the two. Movement in real time of the free edge of the ruffled membrane was readily followed by eye under the phase microscope as a lighting high contrast sharp line. After staining with phalloidin the free edge appears as an intensely fluorescent curlicue-like structure. It did not contact the substrate, did not display adhesion complexes, often extended more than 12 μm into the medium, and during recovery appears earlier than lamellipodia ( Table 1). The majority of the actin containing structures that assembled 30 min after CB-washout was indistinguishable from control-untreated cells (Fig 2M-2O and Table 1). The two earliest actin substructures to reform concurrent with the disappearance of the ectopic actin patches were the actin cortices and ruffled membranes (Table 1). Despite the fact that ruffled membranes were one of the earliest structures to reappear after CB-washout, their presence decreased with time and only 13% of them were found after 30 min CB-washout (Table 1). Ruffled membranes are important in the initial events that occur during the cytoplasmic remodeling and the establishment of polarity in motile cells, such as fibroblasts. The decrease in number of ruffled membranes during the recovery from CB can be explained by the increase in stress fibers and adhesion plaques as the cells became more adherent and less motile.
In summary, (1) the rapid sequential reassembly of the actin cytoskeleton after CB-washout involved transitory combinations of actin-associated proteins that were lost with further completion; (2) the two earliest actin sub-structures to reform concurrent with the disappearance of the ectopic actin patches were the actin cortices and ruffled membranes ( Table 1).

Effects of 30 min CB treatment on muscle cells and recovery
In order to compare fibroblasts with muscle cells, we studied the short-term treatment of CB in myoblasts. A 30 min treatment with CB was sufficient to round up muscle cells (S1 Fig). Myoblasts treated for 30 min with CB lost their cortical actin and their stress fiber-like structures while the cells round-up. Many α-actin/s-α-actinin patches were observed at the surface of these rounded myoblasts. The α-actin/s-α-actinin patches were also positive for adhesion plaque proteins, such as vinculin and paxillin. When recovered in normal medium, these CB treated-myoblasts reassembled their actin-containing filament structures in a way that cells were indistinguishable from control-untreated cultures (data not shown). Interestingly, these results showed that the short-term treatment of CB induced similar changes and recovery in the actin containing structures in both fibroblasts and myoblasts.

Distinct morphological effects of 72 hrs CB treatment on fibroblasts and muscle cells
Given the multiple functions attributed to the actin cytoskeleton, it was important to determine how cells lacking such a system for long time, due to the disruptive effects of CB, would function. Seventy-eighty percent of the cells in day 4 cultures grown in the presence of CB for the last 72 hrs (n = 100 cells from 3 different experiments) were multinucleated (Fig 3). They were differentiated into two distinct phenotypes in approximately even numbers. They were the giant spread finger-shaped fibroblasts and giant globular myogenic cells. The first class of giant multinucleated cells was the finger-shaped fibroblasts (Figs 3 and 4, S2-S5 Figs). The characteristics of these bizarre cells was summarized as follows: (1) over 90% harbor 2 to 5 nuclei at their cell center, which under the confocal microscope measured > 15 μm in the Z-axis (S5 Fig); (2) they had several finger-like protrusions, which varied from 3 to 10 (Fig 4 and S2 Fig). The individual fingers varied from 2 to 7 μm in the x-y axis and from 2-5 μm in the z-axis (S5 Fig); (3) recognizable subplasma F-actin cortices, stress fibers, adhesion plaques or ruffled membranes were absent; (4) the scattered, F-actin patches vary in size and often were > 100 in number (S5 Fig). Interestingly, they were not crystalloid (e.g. Fig 3); (5) as the antibodies to the 4 adhesion plaque proteins (paxillin, β1-integrin, vinculin and talin) diffusely stained in the entire cell, many actin patches were obscured. But from those that were visualized~90% double stain with Table 1. Emergence of actin cytoskeletal subunits after CB washout. Temporal reconstitution of cytoskeletal subunits in cultures exposed to CB for 30 min and then after washout permitted to recover for varying times in normal medium. In each of the 6 series, 3 dishes were double-stained at the indicated times. The first randomly selected 100 cells/dish at each time point were categorized into the different substructures. paxillin, whereas we estimated~60% double stained with β1-integrin, vinculin and talin. Furthermore, the four adhesion plaque proteins assembled a thin, widely distributed, filigree network not observed in control-untreated fibroblasts (Fig 4B and 4C); (6) over 70% of the fingers terminated in lamellipodia of varying sizes. Many of the larger ones exceeded 20 μm in the x-y axis (Figs 3 and 4). Visually we estimated that~90% of the total F-actin per cell was sequestered in their giant lamellipodia (Fig 4D-4H). Intriguingly, (a) unlike lamellipodia in diploid control fibroblasts, these lamellipodia were positive for all 5 of the tested proteins (α-actinin, paxillin, talin, β1-integrin and vinculin), and (b) these lamellipodia did actively translocate; (7) filopodia and ruffled membranes were conspicuously absent; (8) every finger-fibroblast was vimentin positive, but negative for all myofibrillar proteins including desmin (S4 Fig).
In contrast to finger-shaped fibroblasts, mushroom-shaped globular myogenic cells (Figs 5-7, S4 and S6 Figs) which contained mono-, bi-, or 4 + -nuclei were myofibrillar protein positive ( Fig 5). 70% of these globular-shaped myoblasts contained 2 or more nuclei. These myogenic cells also occasionally contracted spontaneously (S7 Fig). Long term CB treatment induced the loss of stress fiber like structures in these cells. However, polarized 1.0 μm α-actin containing thin filaments, which were assembled into normal I-Z-I-bands interdigitating with normal Abands, were observed ( Fig 5 and S6 Fig), indicating that CB did not interfere with the assembly of myofibrils but did interfere with their elongation mechanisms. Furthermore, CB treatment blocked cell fusion since we observed~65% of the round-shaped muscle cells closely contacted with each other but unfused (Figs 5D-5I and 7D). Most of the cell surface of the rounded myogenic cells did not connect to the substrate (S3, S4 and S6 Figs), and there were aggregates of adhesion proteins around the surface of the cell and in the cytoplasm. We found both α-actin and s-α-actinin positive patches at the surface of peripheral sarcolemma of the rounded myogenic cells. These actin-containing patches were positive for focal adhesion proteins ( Fig 4B  and 4C), suggesting that focal adhesion proteins were closely associated with α-actin.
Since we detected actin structures (F-actin patches) with phalloidin in both fibroblasts and muscle cells after long term CB treatment (72 hrs), we decided to analyze the distribution of Arp2/3 proteins, which are known to be involved in the nucleation of actin monomers from existing filaments [8][9][10][11]. Arp2/3 was detected in many cytoplasmic puncta and in the perinuclear region of both myoblasts and fibroblasts that were not treated with CB (arrow in Fig 6B).   In multinucleated myotubes from control-untreated cultures (72 hrs) Arp2/3 was found in many bright cytoplasmic puncta (arrow in Fig 6A). After 72 hrs of CB treatment Arp2/3 was found in puncta in the cytoplasm and at the perinuclear region of round-shaped myoblasts and finger-shaped fibroblasts (Fig 6C and 6D). Interestingly, almost no Arp2/3 was found in the membrane extensions of CB-treated finger-shaped fibroblasts (arrow in Fig 6C).
Finally, to investigate microtubules and intermediate filaments after long-term CB treatment, finger-shaped fibroblasts and in round myogenic cells were visualized with anti-tubulin and anti-desmin antibodies (S4 Fig). Both cell types were positive for tubulin, but in different distributions (S3 and S4 Figs). In fibroblasts, numerous microtubules were found and they were long, softly bended, spreading within the whole cytoplasm and radiating from the cell center at the nuclear vicinity towards the distal tip of each finger-shaped cytoplasm projection the cell whereas adhesion plaque protein clusters localized to the periphery of the cell. Arrowheads in G, H and I indicate a cytoplasmic extension that can be seen using DIC. Note the clusters of vinculin located outside the myofibrils (arrows in J, K and L). Scale bars, 10 μm.
doi:10.1371/journal.pone.0154109.g005   Effects of Cytochalasin B on Fibroblasts and Muscle Cells (Fig 4E and 4G, S3 and S5 Figs). MTOCs were clearly seen in these fibroblastic cells (Fig 4G  and 4H, and S3 Fig). In round-shaped myoblasts, microtubules were shorter, spreading the whole cytoplasm and concentrated around the nuclear membrane and forming a ring line along the border of the rounded cells (Fig 7E and 7F, and S3 Fig). No visible MTOC or microtubules radiating centers were seen. Condensed microtubules accumulated at the peripheral region of the cells were observed in addition to the normal network of microtubules (Fig 7A-7C and 7F). Quantification of the number of myoblasts that present these structures showed that 15% of long-term CB treated myoblasts displayed condensed microtubules in their cytoplasm (n = 100 cells from 3 different experiments, significantly different from untreated cultures, p<0.05). Furthermore, intermediate filament desmin was detected spreading in the whole cytoplasm of round-shaped myoblasts and concentrated in ring structures which located between the myofibrils and condensed microtubule bundles (Fig 7E, and S4 Fig).  Fig 7G) and spreading in the whole cytoplasm of finger-shaped fibroblasts ( Fig 7H). Vimentin filaments were clearly observed at the cytoplasmic extensions of fingershaped fibroblast cells (arrow in Fig 7H). Vimentin labeling in CB-treated myoblasts resembles the desmin positive-ring structures found at the periphery of round shaped myoblasts (compare Fig 7E and 7G

Recovery of actin cytoskeleton in 72 hrs-CB treated fibroblasts
As soon as washed off CB and shifted to normal medium for 20 min, finger-shaped fibroblasts gradually extended their membrane tension, resulting in eliminating finger protrusions, i.e. the morphology of fibroblasts changed from finger-shaped fibroblasts to flatten multinucleated spread cells with immense cytoplasm (Fig 8F and S7 Fig). During the initial 20 min of recovery after 72 hrs of CB treatment, it is possible to observe under phase-contrast microscopy the disappearance of membrane protusions at the periphery of the fibroblastic cells (arrows in Fig 8F  and S7 Fig). Vinculin emerged as dots at the edge of membrane and then gradually formed adhesion plaques where stress fibers start/end. During recovery, not only the spatiotemporal sequence of actin containing structures in the finger-shaped fibroblasts was different from recovery in the 30 min CB-treated diploid cells, but also the transitory structures displayed in the recovery of 30 min CB were not observed in the recovery of 72 hrs CB cultures. The actin cortex did not reconstitute precociously. Rather, even after a 60 min CB-washout, most cells tended to be circular, circumscribed by a~2.0 μm fringe (S7 Fig). Structurally the fringe displays features more reminiscent of lamellipodia than ruffled membranes. During this period virtually all actin patches disappeared. Modest numbers of incipient adhesion plaques and poorly defined stress fibers appeared to be in the process of being nucleated and assembled at the periphery of the cell. Reconstitution and integration of all the subunits of the former fingerfibroblasts required 4-6 hrs (Fig 9A, 9B and 9J-9L). To examine whether de novo synthesized proteins were required for fibroblast recovery from long-term CB treatment, cycloheximide (a protein synthesis inhibitor) was added to recovery medium (Fig 9A). Fibroblasts treated on day 2 with CB for 3 days and recovered in the presence of cycloheximide showed sub-plasma membrane actin cortices and multiple bundles of actin containing stress fibers that were undistinguished from control-untreated cultures and from long-term CB treated cultures allowed to grown in normal media without cycloheximide (Fig 9B). These results suggest that newly synthesized proteins are not necessary for the formation of stress fibers during the recovery of fibroblasts from CB treatment.

Slow recovery of muscle cells from 72 hrs CB exposure
In contrast to relatively quick response of fibroblasts to the removal of CB, myogenic cells recovered very slowly. During the initial 20 min of recovery after 72 hrs of CB treatment, myogenic   (Fig 8F and S7 Fig). It took about 24 hrs to form elongated myotubes from globular myogenic cells. We observed that the α-actin patches disappeared and then were replaced with a veil-like sarcolemma surrounding the shaft of myotubes, which was also observed with phalloidin labeling (Fig 9J). Myofibril structures were still intact during recovery process (Fig 9D-9I). In time-lapse videomicroscopy it was possible to observe that the mushroom-shaped myogenic cells contract spontaneously during treatment and recovery from CB treatment (S7 Fig), which is in accordance with our results showing that these globular and multinucleated cells contain intact and well organized myofibrils (Fig 5 and S6 Fig).
To examine whether de novo synthesized proteins were required for myotube recovery from long-term CB treatment, cycloheximide was added to recovery medium (Fig 9A and 9C). Muscle cells grown in the presence of cycloheximide elongate but fail to fuse, and show a minimal number of sarcomeres ( Fig 9C). As cell adhesion molecules mediate the recognition and adhesion to promote cell fusion [22], we examined whether or not N-cadherin was localized correctly during recovery from CB in the presence of cycloheximide. Anti-N-cadherin antibody specifically captured their signals at cell-cell contact regions of myogenic cells which were recovered from 72 hrs CB treatment in the presence of cycloheximide (Fig 10A-10I). These results confirm that CB treated muscle cells grown in the presence of cycloheximide elongate but fail to fuse. To further investigate whether the myogenic program proceeds or not during the recovery of CB in the presence of cycloheximide, cells were probed with an antibody against Mrf4, one of the myogenic regulatory factors [23]. The Mrf4 signal was detected in myonuclei during recovery from CB in the presence of cycloheximide (Fig 10J-10Q). These results suggest that newly synthesized proteins are necessary for the formation of a mature multinucleated myofiber during the recovery from CB treatment.
One of the most characteristic features during recovery was the appearance of cortical actin bodies (CABs) in the shaft of some myotubes at 1-6 hrs after washing out CB. We found that these CAB structures consisted of a core that was positive for both α-actin/s-α-actinin and a rim containing focal adhesion plaque proteins (Fig 9J-9O and S9 Fig).

Distinct phenotypical impact of long-term CB treatment on fibroblasts versus muscle cells
After short-term CB treatment, the cell shape of both fibroblasts and myogenic cells become globular with very thin cytoplasmic extensions, resulting in indistinguishable cell shape. By contrast, long-term CB treatment led to distinct morphological impact on fibroblasts and muscle cells, i.e. finger-shaped fibroblasts and mushroom-like myogenic cells (Fig 11). Interestingly, most myoblasts and fibroblasts became multinucleated after long-term exposure to CB, with an even number of nuclei. This observation leads to the conclusion that CB blocks cytokinesis and myoblast fusion.
The morphology of the CB-treated fibroblasts is very dramatic: they are immense cells, almost 100 μm in diameter, with long processes and lamellipodia. This finger-shaped fibroblastic cell size and shape has never been described, particularly as an effect of CB treatment. The long-term effects of CB on myoblasts are quite different from the effects on fibroblasts. In fibroblasts CB disassembles stress fibers and adhesion structures preserving lamellipodia,  Effects of Cytochalasin B on Fibroblasts and Muscle Cells whereas in muscle cells the cortical actin cytoskeleton is affected and the myofibrillar cytoskeleton is resistant to CB. After 72 hr CB treatment, myoblasts displayed striated myofibrillar structures that were positive for all the components of the sarcomere: thin filaments and Z lines (actin, tropomyosin, troponin, α-actinin), thick filaments and M-lines (C-protein, myosin heavy chain), and elastic filaments (titin and nebulin). These results differ from previous experiments, which suggested that CB induces the appearance of round myoblast cells containing only a few intact sarcomeres in chick myogenic cell cultures [24]. Our data shows that, even after long term CB treatment, myoblasts contain normal I-Z-I-bands interdigitating with normal A-bands. These data suggest two possibilities: (i) that CB did not interfere with the assembly of myofibrils, and/or (ii) that the myofibrils were preexisting and stable throughout the time course of CB treatment.
Even though most of the cell surface in the rounded CB-myoblasts does not connect to the substrate, there are aggregates of adhesion proteins around the surface of the cell and in the cytoplasm. The morphological differences between muscle and fibroblasts after long-term CB treatment may be caused by the presence of active lamellipodia in finger-shaped fibroblasts but not in mushroom-like myogenic cells. In normal fibroblasts, lamellipodia is a mobile component that drives the cell body [25] in which actin filaments are closely associated with the final ends of microtubules. We observed that almost all cytoplasmic extensions of finger-shaped fibroblasts are terminated with lamellipodia, which are very locomotive even in the presence of CB. We also show that a bundle of microtubules run from the cell body to the final tip of lamellipodia in finger-shaped fibroblasts. These results suggest that mobile lamellipodia followed by microtubules drives the cytoplasm in different directions, resulting in finger-shaped fibroblasts. Since lamellipodia are absent in mushroom-like muscle cells, they maintain a round shape after long term CB treatment.

Different effects of CB on actin-containing structures
Actin-containing filament structures, such as stress fibers, adhesion plaques and cortical actin networks, are disorganized after CB treatment in muscle cells and fibroblasts. CB binds to the barbed end of actin filaments and inhibits the polymerization of actin filaments, resulting in breakdown of actin filament structures [19]. Phalloidin still, however, detects actin structures in both CB treated fibroblasts and muscle cells, indicating that actin molecules are still polymerized into actin oligomers in lamellipodia in fibroblasts and in actin patches in both cells. It is possible that actin oligomers, which consist of less than 10 actin molecules [26], are left behind in the CB treatment and are still polymerized at lamellipodia. As CB primarily caps and fragments long actin filaments into short oligomers, both aggregates in myoblasts and giant lamellipodia in finger-shaped fibroblasts should consist of densely packed oligomers.
To test our hypothesis that actin oligomers are left behind after long-term CB treatment and are still polymerized at lamellipodia, we used latrunculin B in CB treated myogenic cultures to prevent the polymerization of actin oligomers. Latrunculin B binds to actin monomers and prevents their incorporation into actin polymer. Latrunculin B is preferred for short-term studies because it is gradually inactivated by serum [27]. It has been shown that calf serum substantially increases the stability of actin filaments [27]. Therefore, latrunculin B could not work during long-term CB treatment (72 hrs). To test the possible use of latrunculin B in long-term CB treatment, we treated chick myogenic cultures that were in the presence of CB for 72 hrs with 250 mM of latrunculin B using two different culture media composition: normal 8-1-0.5 medium (MEM containing horse serum and chick embryo extract) and MEM without horse serum and chick embryo extract. These experiments showed that myogenic cells grown with MEM without serum suffered and eventually die after 24 hrs, whereas cultures grown in the presence of 8-1-0.5 medium grown well and showed no effects of latrunculin B (CB treated cultures were identical to cultures treated with CB and latrunculin B). The problem here is that we cannot discard the possibility that latrunculin B had no effects in the inhibition of actin oligomers because of its inactivation by the presence of serum in the medium.
Arp2/3 complex was found in both untreated and CB-treated chick myogenic cultures. Changes in Arp2/3 distribution was found after long term CB treatment, i.e., the absence of Arp2/3 labeling at the membrane extensions in finger-shaped fibroblasts. Since Arp2/3 complex is a nucleator of actin polymerization [8][9][10][11], the presence of a high quantity of Arp2/3 puncta in the cytoplasm of both fibroblasts and muscle cells after long term-CB treatment could be caused by an increase in Arp2/3 synthesis as a response to the absence of F-actin structures in these cells. New experiments are needed to confirm this hypothesis. It is important to point out that, to our knowledge, our data represents the first description of the cellular distribution of Arp2/3 proteins in skeletal muscle cells and during skeletal muscle differentiation.
The failure of adhesion plaque proteins to assemble into discrete focal adhesions but instead assemble into long multimolecular filaments, which is termed filigree, is of interest. In normal fibroblasts, focal adhesion plaques are formed behind lamellipodia and capped with stress fibers to regulate cell motility [28]. Our results show that CB treatment to fibroblasts diminished focal adhesion plaques but their components were partially localized at their lamellipodia, indicating that proper focal adhesion plaque complex is not formed in finger-shaped fibroblasts. It is clear that the breakdown of F-actin cytoskeleton induced by CB does not block the continued synthesis of the individual adhesion plaque proteins and that they assemble into ectopic long filament filigree, which is not observed in control-untreated cells. This filigree structure which runs through cytoplasmic extensions in finger-shaped fibroblasts is probably synthesized as a result of the absence of stress fibers, since F-actin stress fibers are essential to initiate the organization of proper focal adhesion plaques in fibroblasts [28].
The distribution of phalloidin-positive structures in chick fibroblasts briefly exposed to CB was identical to those reported by Ballestrem and colleagues [29] on living mouse melanoma cells (B16F1), Chinese hamster ovary cells (CHO), thymic endothelial cells, and 3T3 fibroblasts expressing EGFP/β-actin cDNA and exposed to CB, with one intriguing exception. Chick fibroblasts in CB do not form the conspicuous perinuclear actin rings or "actin clouds" they observed in mouse melanoma cells. Finding that actin containing structures in different cell types or in different structures within the same cell, respond differently to CB is of considerable interest.
Interestingly, we observed changes in microtubules and intermediate filament structures after CB treatment. The fact that changes in one cytoskeletal component can affect, more or less directly, all the cytoskeletal components suggest that there is a much deeper and subtle interaction between each part of the cytoskeleton than just direct molecular interaction as it has been repeatedly shown by sequencing of genes and identification of putative binding sites for several molecules. The changes observed in microtubules and intermediate filaments after CB could be caused directly by the modifications induced in the actin network or indirectly by other molecules. Further studies are necessary to understand the molecular machinery involved in the alterations of microtubules and intermediate filaments in myogenic cells in response to CB.
Given the evidence that the actin cytoskeleton serves as substrates for myosin motors, as well as in the distribution of mRNAs and relaying signals from the extracellular matrix to the nucleus, the finding that cells lacking this system nevertheless increase in mass, duplicate their chromosomes, and replicate their nuclei was unexpected. Even more challenging is how such cells when removed from CB retained mechanisms that permitted the re-expression of a relatively normal cell phenotype.

Recovery from CB treatment
Long-term treatment with CB does not kill the myogenic cells, as demonstrated by their recovery capability. Instead, the cells respond to the stress by creating new morphologies using their cytoskeletons, depending on their genetic and differentiation state, e.g. CB-washout permits fusion into typical elongated myotubes with normal longitudinal striated myofibrils despite blocking fusion into myotubes in the presence of CB.
Our data showing the CB-induced actin disorganization and the CB recovery suggest that this system could be used to understand the roles of actin compartments in different cell functions, such as cell movement, cell fusion and cell adhesion. Association of actin, α-actinin, and the adhesion proteins in actin patches was observed when myogenic cells were exposed to CB for 72 hrs. This close correlation implies that the complexes play as the kernel to initiate the assembly of stress fibers or myofibrils during the recovery process. We could speculate that after CB washout the rapid reassemble of stress fibers in fibroblasts works in pre-existing short actin filaments rather than inducing de novo assembly of filaments, and that this process might be dependent on Rho GTPases [30].
Drastic rearrangements and inter-convertibility of the different actin cytoskeletal subunits are induced by a variety of extra-and intracellular signals. Rous sarcoma virus-transformed fibroblasts lose their global structures, which are replaced by numerous 1-2 μm multi-protein dot-like complexes [31]. These protein complexes, termed podosomes, accumulate at the membrane/substrate interface [32]. These podosomes closely resemble cortical α-actin containing bodies (CABs), which are formed in 12-O-tetradecanoylphorbol-13-acetate (TPA) treated myotubes [33]. Both podosomes and CABs consist of a core of actin and α-actinin surrounded by integrin, vinculin, talin and paxillin positive rims. As formation of CABs is thought to be an indicator of enhancing myofibrillar turnover [34], appearance of CABs reflects that myotubes intensively remodel their myofibrillar proteins during the recovery from CB. Furthermore, our result shows that newly synthesized myofibrillar proteins, including actin, are dispensable for the recovery from CB treatment. These data indicate the presence of a sufficient pool of preexisting myofibrillar proteins to assemble a minimal number of sarcomeres but further protein synthesis is necessary to produce a mature myofiber.
It is important to point out that although CB has been shown to bind to the glucose transporter and interferes with cellular glucose uptake [35], as pointed out by the authors [35], it seems unlikely that the various other biological effects of CB, such as the inhibition of microfilament polymerization, cytokinesis, cell movement, or phagocytosis, can be attributed to the observed interference with the glucose transport system.

Primary skeletal muscle cell culture
The work was performed in accordance with the requirements of the Ethics Committee for Animal Care and Use in Scientific Research of the University of Pennsylvania, which specifically approved this study. Primary myogenic cells were prepared as described previously [36,37]. Briefly, pectoral muscle tissue was removed from 11-day old chicken embryos (CBT Farms, Chestertown, MD). After separation of the connective tissue under a dissecting microscope, muscle tissue was minced with surgery knives and digested with trypsin (0.025%) at 37°C in a 5% CO 2 incubator for 25 min. To stop digestion, a small volume of growth medium (10% horse serum, 10% chick embryo extract, 2 mM L-Glutamine, 50 U/ml penicillin, 50 μg/ml streptomycin, Fungizone in Eagle's Minimum Essential Medium [Gibco]) was added to the sample and then centrifuged for 5 min at 800 rpm. After discard the supernatant, the pellet was suspended in the growth medium and filtered through lens paper to produce mononucleated cell suspension. Isolated mononucleated cells were plated onto collagen-coated Aclar coverslips (Pro-Plastics, Linden, NJ) at initial densities of 4.5 x 10 5 cells per 35 mm culture dish. They were grown continuously in 1.5 ml of the growth medium. Cells were grown under humidified 5% CO 2 atmosphere at 37°C.

Drug treatments
To determine the short-term effect of CB on the actin disorganization, day 2 cultured cells were incubated with 5 μM CB in DMSO (both from Sigma-Aldrich) for 30 min. Some cells were treated with DMSO (2 μL) alone as a control. For recovery experiments, cells were washed with fresh growth medium 3 times, fed with fresh growth medium, kept in CO 2 incubator for 30 min, and fixed. To determine the long-term effect of CB on the actin disorganization, day 2 cultured cells were incubated with 5 μM CB for 72 hrs. Some CB treated cells were recovered in the presence of the protein synthesis inhibitor cycloheximide (Sigma, dissolved in ethanol, working concentration 0.1 mM). Long-term CB dishes were pretreated with 0.1 mM cycloheximide for 1h before washout and then continue to recover in cycloheximide containing medium for designed time points. Some CB treated cultures were treated also with 250 mM of latrunculin B (Sigma-Aldrich). To analyze cell viability, trypan blue exclusion assays were performed in CB treated and non-treated cells. Cells were stained with 0.4% trypan blue and then examined microscopically to determine the total number of cells and trypan blue positive cells. The ratio of trypan blue positive cells was calculated as following: [(the number of trypan blue positive cells) / (the number of total cells)] × 100.

Antibodies and fluorescent probes
The findings in this report rely heavily on the double-staining of cultured cells with Rhodamine

Immunofluorescence microscopy, image acquisition and processing
Indirect immunofluorescence double staining has been described by Lin and collaborators [36]. In brief, cultured cells, with or without drug treatment, were fixed with 2% formaldehyde in PBS for 3 min after washing with PBS at 37°C. Cells were permeabilized with 0.5% Triton X-100 in PBS for 10 min three times. For anti-tyrosylated-α-tubulin antibody staining, cells were fixed with 2% formaldehyde and 0.1% glutaraldehyde in PHEM (60 mM Pipes, 25 mM Hepes, 1 mM EGTA, 2 mM MgCl 2 , pH 6.95) for 10 min after washing with PHEM at 37°C. Fixed cells were permeabilized with 0.5% Triton X-100 in PHEM for 2 min. Subsequently, specimens were rinsed with 1 mg/mL sodium borohydride in PBS three times for 5 min and rinsed with PBS again to quench auto-fluorescence of glutaraldehyde. Specimens were incubated with primary antibody for 1 hour at 37°C. After washing with 0.5% Triton X-100 in PBS, specimens were incubated with secondary antibody for 1 hour at 37°C. Affinity-purified secondary antibodies conjugated with Rho or FITC were utilized (Jackson Immuno Research Labs.). To localize F-actin, some specimens were stained with Rho-phalloidin (3.3 μM, Molecular Probes Inc.). Cells were exposed to phalloidin for 20 min at 37°C. As a control for bleed-through artifact, all double-stained preparations were run in duplicate. In one series, each of the two primary antibodies was stained with either a Rho-or a FITC-conjugated secondary antibody. In the second series, the same primary antibodies were stained in the opposite fashion, thus each antigen was localized by both FITC-or Rho-tagged secondary antibodies. Nuclei were detected with 1 μg/ mL DAPI (4, 6-diamidino-2-phenylindole dihidrochloride; Polyscience). Specimens were mounted in 90% glycerol in PBS containing 2.5% DABCO (1, 4-diazabicyclo (2, 2, 2) octane; Sigma). Cells were observed using a laser confocal microscope (LSM 510; Carl Zeiss), which employed a Zeiss Axiovert 100 inverted microscope. Images were viewed with a Zeiss C-Apo 63 x water immersion objective lens (NA 1.2). Rho, FITC and UV signals were visualized using appropriate lasers and filters. Image processing and stack projections were performed using ImageJ (http://imagej.nih.gov/ij/). Figure plates were mounted with Adobe Photoshop software (Adobe Systems Inc.). Some images were taken with a Leica DM model IRB fluorescence inverted optical microscope using 63 x 1.6 (NA 1.4) oil immersion objective lens.

Videomicroscopy
Chick myogenic cells were grown for 24 hrs, treated with CB for 30 min or for 72 hrs. Live cultures were imaged under phase contrast microscopy in an Axiovert 100 microscope (Carl Zeiss, Germany). Images were recorded with an Olympus DP72 digital camera (Olympus, Japan) and analyzed using Fiji software.

Statistical analysis
All data are represented as mean ± standard error. Statistical analysis was performed with oneway ANOVA on Ranks with Newman-Keuls Post Hoc test and statistical significance was defined as p<0.05. nuclei between finger-shaped fibroblasts and mushroom-like muscle cells after 72 hrs of cytochalasin B treatment. The 3-μm apart images were taken from a stack of 1-μm interval slices, acquired using a laser scanning confocal microscope. Cells were labeled with antibodies against α-tubulin (green) and desmin (red), and with DAPI (blue). Compare the spread and flat fibroblasts with the round and thick desmin-positive myoblasts. Note that the fibroblast nuclei are located in the 6 and 3-μm slices whereas the myoblast nuclei are present in all the slices. Scale bar, 10 μm. (TIF) S4 Fig. 3D reconstruction of chick myogenic cells showing distinct morphological differences between a finger-shaped fibroblast and a mushroom-like muscle cell after 72 hrs of cytochalasin B treatment. Sequential projections in the X and Y-axis from a stack of 1-μm interval images, acquired using a laser scanning confocal microscope. Cells were labeled with antibodies against α-tubulin (green) and desmin (red), and with DAPI (blue). Note the desmin signal was detected at the periphery of the myoblast but not observed in the fibroblast. Microtubules are well spread in the fibroblast and concentrated in the center of the round myoblast. Both cells are binucleated. Scale bar, 10 μm. The sections were reconstructed from a stack of 22 images with 1-μm interval, acquired using a laser scanning confocal microscope. Cells were labeled with Rho-phalloidin (red), an antibody against α-tubulin (green), and DAPI (blue). Note that actin accumulates at the tip of the fibroblast fingers (close to the substrate), whereas microtubules (that label most of the cell cortex) do not localize to the very extremity of the cell. Scale bar, 5 μm. (TIF) S6 Fig. 3D reconstruction of a muscle cell after 72 hrs of cytochalasin B treatment. Sequential projections in the X and Y-axis from a stack of 1.4-μm interval images, acquired using a laser scanning confocal microscope. Cells were labeled with an antibody against sarcomeric-αactinin (red) and DAPI (blue). Note that the round myoblast contains perfectly striated myofibrils, as seen in the Z-line staining of sarcomeric-α-actinin. The round myoblast has 4 nuclei. Scale bar, 5 μm. The images were taken from a stack of 1-μm interval slices, acquired using a laser scanning confocal microscope. Cells were labeled with Rho-phalloidin (red), an antibody against vinculin (green), and DAPI (blue). Actin is found in striated myofibrils at the dorsal part of the multinucleated myotube, whereas at the ventral surface of the cell actin (surrounded by vinculin aggregates) emerges as small and round cortical actin containing bodies (CABs). Scale bar, 5 μm. (TIF)