Myosin II dynamics and cortical flow during contractile ring formation in Dictyostelium cells.

Myosin II is a major component of a contractile ring. To examine if myosin II turns over in contractile rings, fluorescence of GFP-myosin II expressed in Dictyostelium cells was bleached locally by laser illumination, and the recovery was monitored. The fluorescence recovered with a half time of 7.01 +/- 2.62 s. This recovery was not caused by lateral movement of myosin II from the nonbleached area, but by an exchange with endoplasmic myosin II. Similar experiments were performed in cells expressing GFP-3ALA myosin II, of which three phosphorylatable threonine residues were replaced with alanine residues. In this case, recovery was not detected within a comparable time range. These results indicate that myosin II in the contractile ring performs dynamic turnover via its heavy chain phosphorylation. Because GFP-3ALA myosin II did not show the recovery, it served as a useful marker of myosin II movement, which enabled us to demonstrate cortical flow of myosin II toward the equator for the first time. Thus, cortical flow accompanies the dynamic exchange of myosin II during the formation of contractile rings.

Recruitment of a myosin heavy chain kinase to actin-rich protrusions in Dictyostelium.

Nonmuscle myosin II plays fundamental roles in cell body translocation during migration and is typically depleted or absent from actin-based cell protrusions such as lamellipodia, but the mechanisms preventing myosin II assembly in such structures have not been identified [1-3]. In Dictyostelium discoideum, myosin II filament assembly is controlled primarily through myosin heavy chain (MHC) phosphorylation. The phosphorylation of sites in the myosin tail domain by myosin heavy chain kinase A (MHCK A) drives the disassembly of myosin II filaments in vitro and in vivo [4]. To better understand the cellular regulation of MHCK A activity, and thus the regulation of myosin II filament assembly, we studied the in vivo localization of native and green fluorescent protein (GFP)-tagged MHCK A. MHCK A redistributes from the cytosol to the cell cortex in response to stimulation of Dictyostelium cells with chemoattractant in an F-actin-dependent manner. During chemotaxis, random migration, and phagocytic/endocytic events, MHCK A is recruited preferentially to actin-rich leading-edge extensions. Given the ability of MHCK A to disassemble myosin II filaments, this localization may represent a fundamental mechanism for disassembling myosin II filaments and preventing localized filament assembly at sites of actin-based protrusion.

Myosin II-independent cytokinesis in Dictyostelium: its mechanism and implications.

Similar to higher animal cells, ameba cells of the cellular slime mold Dictyostelium discoideum form contractile rings containing filaments of myosin II during mitosis, and it is generally believed that contraction of these rings bisects the cells both on substrates and in suspension. In suspension, mutant cells lacking the single myosin II heavy chain gene cannot carry out cytokinesis, become large and multinucleate, and eventually lyze, supporting the idea that myosin II plays critical roles in cytokinesis. These mutant cells are however viable on substrates. Detailed analyses of these mutant cells on substrates revealed that, in addition to "classic" cytokinesis which depends on myosin II ("cytokinesis A"), Dictyostelium has two distinct, novel methods of cytokinesis, 1) attachment-assisted mitotic cleavage employed by myosin II null cells on substrates ("cytokinesis B"), and 2) cytofission, a cell cycle-independent division of adherent cells ("cytokinesis C"). Cytokinesis A, B, and C lose their function and demand fewer protein factors in this order. Cytokinesis B is of particular importance for future studies. Similar to cytokinesis A, cytokinesis B involves formation of a cleavage furrow in the equatorial region, and it may be a primitive but basic mechanism of efficiently bisecting a cell in a cell cycle-coupled manner. Analysis of large, multinucleate myosin II null cells suggested that interactions between astral microtubules and cortices positively induce polar protrusive activities in telophase. A model is proposed to explain how such polar activities drive cytokinesis B, and how cytokinesis B is coordinated with cytokinesis A in wild type cells.

Molecular biological approaches to study myosin functions in cytokinesis of Dictyostelium.

The cellular slime mold Dictyostelium discoideum is amenable to biochemical, cell biological, and molecular genetic analyses, and offers a unique opportunity for multifaceted approaches to dissect the mechanism of cytokinesis. One of the important questions that are currently under investigation using Dictyostelium is to understand how cleavage furrows or contractile rings are assembled in the equatorial region. Contractile rings consist of a number of components including parallel filaments of actin and myosin II. Phenotypic analyses and in vivo localization studies of cells expressing mutant myosin IIs have demonstrated that myosin II's transport to and localization at the equatorial region does not require regulation by phosphorylation of myosin II, specific amino acid sequences of myosin II, or the motor activity of myosin II. Rather, the transport appears to depend on a myosin II-independent flow of cortical cytoskeleton. What drives the flow of cortical cytoskeleton is still elusive. However, a growing number of mutants that affect assembly of contractile rings have been accumulated. Analyses of these mutations, identification of more cytokinesis-specific genes, and information deriving from other experimental systems, should allow us to understand the mechanism of contractile ring formation and other aspects of cytokinesis.

Novel cellular tracks of migrating Dictyostelium cells.

After Dictyostelium cells were settled on a coverslip and allowed to migrate freely on the surface, they were stained with fluorescently labeled Concanavalin A. Tracks with distinct patterns that consist of dots and short fibers were observed behind the cells. In this study, we refer to these tracks as "cellular tracks", CTs for short. We characterized the biological effect of CTs on cell behavior and development. CTs decreased the strength of cell-substratum adhesion, increased the velocity of cell migration, but did not affect growth of cells. CTs also promoted cell aggregation. When pre-aggregation cells touched the CTs of other cells, they avoided or orthogonally crossed them, but did not migrate along them. These observations suggest that the CTs of pre-aggregation cells prompts cells to disperse uniformly on substratum and may enable cells to sense cell density. On the other hand, when aggregation-competent cells touched the CTs of other aggregation-competent cells, a half of them migrated along the CTs. Pre-aggregation cells did not migrate along the CTs of aggregation-competent cells. The CTs of aggregation-competent cells may help the cells to aggregate toward the aggregation center.

Architectural dynamics of F-actin in eupodia suggests their role in invasive locomotion in Dictyostelium.

Eupodia are F-actin-containing cortical structures similar to vertebrate podosomes or invadopodia found in metastatic cells. Eupodia are rich in alpha-actinin and myosin IB/D, but not a Dictyostelium homologue of talin. In the present study, we localized other actin-binding proteins, ABP120, cofilin, coronin, and fimbrin, in the eupodia and examined the three-dimensional organization of their F-actin system by confocal microscopy and transmission electron microscopy. To examine their function, we analyzed the assembly and disassembly dynamics of the F-actin system in eupodia and its relation to lamellipodial protrusion. Actin dynamics was examined by monitoring S65T-GFP-coronin and rhodamine-actin using a real-time confocal unit and a digital microscope system. Fluorescence morphometric analysis demonstrates the presence of a precise spatiotemporal coupling between F-actin assembly in eupodia and lamellipodial protrusion. When a lamellipodium advances to invade a tight space, additional rows of eupodia are sequentially formed at the base of that lamellipodium. These results indicate that mechanical stress at the leading edge modulates the structural integrity of actin and its binding proteins, such that eupodia are formed when anchorage is needed to boost for invasive protrusion of the leading edge.

Myosin II-independent F-actin flow contributes to cell locomotion in dictyostelium.

While the treadmilling and retrograde flow of F-actin are believed to be responsible for the protrusion of leading edges, little is known about the mechanism that brings the posterior cell body forward. To elucidate the mechanism for global cell locomotion, we examined the organizational changes of filamentous (F-) actin in live Dictyostelium discoideum. We labeled F-actin with a trace amount of fluorescent phalloidin and analyzed its dynamics in nearly two-dimensional cells by using a sensitive, high-resolution charge-coupled device. We optically resolved a cyclic mode of tightening and loosening of fibrous cortical F-actin and quantitated its flow by measuring temporal and spatial intensity changes. The rate of F-actin flow was evaluated with respect to migration velocity and morphometric changes. In migrating monopodial cells, the cortical F-actin encircling the posterior cell body gradually accumulated into the tail end at a speed of 0.35 microm/minute. We show qualitatively and quantitatively that the F-actin flow is closely associated with cell migration. Similarly, in dividing cells, the cortical F-actin accumulated into the cleavage furrow. Although five times slower than the wild type, the F-actin also flows rearward in migrating mhcA- cells demonstrating that myosin II ('conventional' myosin) is not absolutely required for the observed dynamics of F-actin. Yet consistent with the reported transportation of ConA-beads, the direction of observed F-actin flow in Dictyostelium is conceptually opposite from a barbed-end binding to the plasma membrane. This study suggests that the posterior end of the cell has a unique motif that tugs the cortical actin layer rearward by means of a mechanism independent from myosin II; this mechanism may be also involved in cleavage furrow formation.

Spatiotemporal dynamics of actin concentration during cytokinesis and locomotion in Dictyostelium.

To study the spatial and temporal regulation of the actin cytoskeleton, we have analyzed the actin concentration dynamics in live Dictyostelium. The relative actin concentration was analyzed with respect to cell behavior by fluorescence morphometry. We electroporated rhodamine-actin into Dictyostelium cells and acquired images with 200-300 millisecond temporal and approximately 250 nm spatial resolutions. To convert fluorescence intensity into actin concentration, the observation was made on nearly two-dimensional cells, and the actin signal was ratioed over a volume marker (FITC-BSA or GFP). Since the emission of FITC and GFP is pH-dependent, we first measured the cytoplasmic pH in live cells and determined that the pHi in pseudopods is same as that of general cytoplasm. During cytokinesis, the relative concentration of actin in the cleavage furrow was significantly higher than in the general cytoplasm. In migrating cells, actin was recruited surprisingly rapidly, particularly in the pseudopod. We found that the region of high actin concentration moves relative to the leading edge when a pseudopod projects or retracts. When the pseudopod retracts, the actin density dissipates within 5 seconds. We have also found that actin accumulates in developing pseudopods in an oscillatory manner, and this timing coordinates with advancement of the centroid. This is the first study to reveal the dynamic changes in relative concentration of actin in live cells and to quantitatively correlate these changes with the locomotive behavior of the amoeba.

Transport of myosin II to the equatorial region without its own motor activity in mitotic Dictyostelium cells.

Fluorescently labeled myosin moved and accumulated circumferentially in the equatorial region of dividing Dictyostelium cells within a time course of 4 min, followed by contraction of the contractile ring. To investigate the mechanism of this transport process, we have expressed three mutant myosins that cannot hydrolyze ATP in myosin null cells. Immunofluorescence staining showed that these mutant myosins were also correctly transported to the equatorial region, although no contraction followed. The rates of transport, measured using green fluorescent protein-fused myosins, were indistinguishable between wild-type and mutant myosins. These observations demonstrate that myosin is passively transported toward the equatorial region and incorporated into the forming contractile ring without its own motor activity.

Myosin II can be localized to the cleavage furrow and to the posterior region of Dictyostelium amoebae without control by phosphorylation of myosin heavy and light chains.

To elucidate the role of phosphorylation in regulation of intracellular distribution of myosin II, we have characterized mutant Dictyostelium cells expressing myosin II that could not be regulated by the phosphorylation on the mapped heavy chain sites, the light chain site, or both sites. Immunofluorescence microscopy demonstrated that all three mutant myosin IIs were localized in the furrow region of dividing cells and in the tail region of migrating cells, similar to wild-type cells. Thus, regulation by phosphorylation is not required to direct myosin II toward the furrow region and the tail region in Dictyostelium. However, myosins that were deficient in heavy chain phosphorylation were distributed only in the cortical region of interphase cells, whereas some myosin IIs were present throughout the endoplasm in wild-type cells. Video microscopy showed that the rate of cell migration was significantly lower in cells that were deficient in heavy chain phosphorylation- than in light chain phosphorylation-deficient cells, myosin null cells and wild-type cells. Chemotactic behavior of cells that were deficient in heavy chain phosphorylation was also retarded. These results suggest that loss of regulation by heavy chain phosphorylation results in excessive myosin in the cortex, which leads to retarded motility.

Intracellular free calcium responses during chemotaxis of Dictyostelium cells.

A calcium ion indicator, fura-2 bovine serum albumin, was introduced into Dictyostelium discoideum cells by electroporation. The concentration of intracellular calcium ions ([Ca2+]i) increased transiently in vegetative cells upon stimulation with submicromolar concentrations of folic acid, a chemoattractant for this organism at the vegetative stage. Similar [Ca2+]i responses were also observed in aggregation-competent cells upon stimulation with subnanomolar concentrations of cAMP, a chemoattractant at the aggregation stage. The [Ca2+]i response caused by cAMP was 2.1 times higher than that caused by folic acid. The magnitude of these responses depended on the concentration of Ca2+ in the external buffer. The presence of magnesium ions inhibited the [Ca2+]i responses in a dose-dependent manner. [Ca2+]i was higher in the rear region than in the anterior region of cells freely migrating on the surface, although such a gradient was not always maintained. When aggregation competent cells were locally stimulated by the application of a microcapillary containing cAMP, the cells extended pseudopods toward the microcapillary. In these cases, an increase in [Ca2+]i was transiently observed in the region opposite to the tip of the capillary. At the slug stage, [Ca2+]i was higher in prestalk cells than in prespore cells of slugs. The possibility that the [Ca2+]i is spatially regulated within a cell was discussed.

Spatial distribution of fluorescently labeled actin in living Dictyostelium amoebae.

Actin from Dictyostelium was labeled with iodoacetamide tetramethylrhodamine (IAR). The labeled actin retained the ability to polymerize into filaments. The labeled actin was introduced into Dictyostelium cells by electroporation. The introduced IAR-labeled actin was diffusely distributed in the cytoplasm but some of it was concentrated in small and large projections at the cell periphery. IAR-labeled actin was concentrated in pseudopods and in the tail cortical region in actively migrating cells. Intense fluorescence due to labeled actin appeared rapidly and disappeared during the extension and retention of pseudopods. During cytokinesis, IAR-labeled actin was concentrated in both polar regions and slightly concentrated in the furrow region. However, the ratio of intensities due to IAR-labeled actin and fluorescently labeled bovine serum albumin that had been introduced simultaneously into cells showed the absence of any concentration of IAR-labeled actin in the furrow region. Staining of fixed cells with fluorescently labeled phalloidin revealed that filamentous actin was rich in the furrow region. These observations indicate that actin is concentrated at this region in higher ratio of filamentous actin to monomeric actin than in other regions of the cytoplasm. Image analysis revealed that the concentration of IAR-labeled actin was high in a pseudopod of an actively migrating cell. Comparisons with staining by fluorescently labeled phalloidin of fixed cells revealed that there was a decreasing gradient in the concentration of filamentous actin from the tip to the base of a pseudopod. These results reveal the high rate at which actin is coordinately organized during mitosis and locomotion in highly motile Dictyostelium cells.

Introduction of macromolecules into living Dictyostelium cells by electroporation.

An attempt was made to optimize conditions for introduction of macromolecules into Dictyostelium cells by electroporation. The amount of fluorescein-labeled bovine serum albumin (FITC-BSA) introduced into cells was measured by fluorometry after extraction of FITC-BSA from cells with detergent. The amount increased as the applied voltage and capacitance of the discharger were increased. However, the survival of cells decreased at higher voltages and elevated capacitance. FITC-BSA was introduced into 80-90% of treated cells. FITC-BSA at 0.25 mg/ml was introduced into cells under optimum conditions when the concentration of the extracellular protein was 2.5 mg/ml. Several discharges in sequence improved the extent of introduction of FITC-BSA although viability decreased. There was a linear correlation between final intracellular concentration and the initial extracellular concentration of FITC-BSA, suggesting the possible quantitative introduction of the protein into cells. The membrane pores that opened during electroporation closed within 2.5 sec after the discharge. FITC-labeled dextran with molecular weights of less than 5 x 10(5) were able to pass through these pores. Our results show that electroporation provides a quantitative and reproducible method for introduction of macromolecules into living Dictyostelium cells.

Differential association of three actin-bundling proteins with microfilaments in Dictyostelium amoebae.

Three actin-bundling proteins have been isolated from Dictyostelium discoideum cells. To investigate whether these proteins might play different roles in Dictyostelium cells, the distribution of these proteins was examined in cells of the slime mold at various developmental stages by agar-overlay immunofluorescence. ABP-30a was localized in pseudopods and projections, and regions of intercellular contacts, where filaments of actin were present at high levels. ABP-30a was not concentrated in the cleavage furrow of dividing cells or in the tail cortical regions of actively moving cells even though actin was present at high levels in both regions. Some ABP-50 was concentrated at the leading edges of pseudopods and at the distal ends of projections, but most ABP-50 was distributed diffusely in the cytoplasm. ABP-30b was not concentrated in any restricted regions and was diffusely distributed in the cytoplasm. When developed cells were stimulated with cAMP, fluorescence due to ABP-30a and ABP-50 increased in the cortical region 5 to 10 sec after stimulation. Quantitative analysis of levels of both proteins in Triton-insoluble cytoskeletons after stimulation by cAMP also showed that levels of both proteins increased within 5 to 10 sec. These observations indicate that levels of ABP-30a and AMP-50 are regulated by signal transduction during chemotaxis and that the three actin-bundling proteins play different roles in Dictyostelium cells.

Rapid translocation of myosin II in vegetative Dictyostelium amoebae during chemotactic stimulation by folic acid.

Immunofluorescence studies showed that cytoskeletons that were composed of actin and myosin II rapidly reorganized in vegetative Dictyostelium cells upon chemotactic stimulation by folic acid. The amount of F-actin increased biphasically with peaks at 5-10 sec (first peak) and 25-45 sec (second peak) after the addition of folic acid. Filaments of myosin II became associated with cell membrane with increases in the meshwork of actin filaments on the cell membrane at the time of the second peak. The number of actin foci in the actin meshwork on the cell membrane decreased transiently at the time of the second peak. The number of filaments of myosin II on the cell membrane decreased and the number of actin foci recovered concomitantly towards the end of the second peak. This reorganization of cytoskeletons during the chemotactic stimulation was also observed after the application of the calcium ionophore A23187 or cGMP to cells. These observations suggest that increases in intracellular levels of both Ca2+ ions and cGMP may play a crucial role in the rapid translocation of myosin II during stimulation by folic acid.

Reorganization of actin and myosin II in Dictyostelium amoeba during stimulation by cAMP.

The distribution of actin and myosin II was studied by immunofluorescence microscopy in Dictyostelium discoideum cells that had been stimulated with a chemoattractant, cAMP. The amounts of actin but not of myosin II increased in the cortical region of cells after 5-10 seconds of stimulation by cAMP at 21 degrees C (the first peak). After a transient decrease in the amount of actin in the cortical region, the amounts of both actin and myosin II increased in the cortical region after 25-35 seconds of stimulation (the second peak). The amounts of myosin II decreased in the cortical region thereafter but actin became localized in the pseudopods. The stimulated cells extended many short projections that contained actin at the time of the first peak and the cell bodies contracted at the time of the second peak. Foci in the cortical actin network decreased in number at 20-30 sec and recovered thereafter. A quantitative examination using a fluorocytometer revealed that the amounts of actin filaments in the cells increased at both peaks. The calcium ionophore A23187 and Ca2+ ion induced the reorganization of actin and myosin II to a certain extent. Furthermore, Ca2+ ions inhibited the return of myosin II to the endoplasm in cAMP-stimulated cells after preincubation with A23187. Thus, Ca2+ ions may play a dual role, being involved in the regulation of both the initiation and the termination of the reorganization of cytoskeletons. The cAMP-stimulated responses of cytoskeletons were not inhibited by EGTA, La3+ ions, or ruthenium red but were inhibited by preincubation with BAPTA-AM.(ABSTRACT TRUNCATED AT 250 WORDS)

Ciliate protozoa in the rumen of Kafue lechwe, Kobus leche kafuensis, in Zambia, with the description of four new species.

The composition of the rumen ciliate fauna in 76 Kafue lechwe inhabiting a limited area in Zambia was surveyed and five genera containing 24 species with 16 formae belonging to the family Ophryoscolecidae were identified. Four new species belonging to Diplodiniinae were recognized and described as Diplodinium lochinvarense n. sp., Diplodinium leche n. sp., Diplodinium zambiense n. sp., and Metadinium ossiculi n. sp. In addition, Ostracodinium gracile form fissilaminatum Dogiel, 1932 was found for the second time and described as Metadinium fissilaminatum n. comb. The species composition was fairly unusual. Seven of the species have been found only in African wild antelopes and these species were found more frequently than cosmopolitan species. There was no evidence of isotrichid species. The average density of ciliates per 1 ml of the rumen fluid was 25.7 x 10(4), and the number of ciliate species per head of host was 10.8.

Release of myosin II from the membrane-cytoskeleton of Dictyostelium discoideum mediated by heavy-chain phosphorylation at the foci within the cortical actin network.

Membrane-cytoskeletons were prepared from Dictyostelium amebas, and networks of actin and myosin II filaments were visualized on the exposed cytoplasmic surfaces of the cell membranes by fluorescence staining (Yumura, S., and T. Kitanishi-Yumura. 1990. Cell Struct. Funct. 15:355-364). Addition of ATP caused contraction of the cytoskeleton with aggregation of part of actin into several foci within the network, but most of myosin II was released via the foci. However, in the presence of 10 mM MgCl2, which stabilized myosin II filaments, myosin II remained at the foci. Ultrastructural examination revealed that, after contraction, only traces of monomeric myosin II remained at the foci. By contrast, myosin II filaments remained in the foci in the presence of 10 mM MgCl2. These observations suggest that myosin II was released not in a filamentous form but in a monomeric form. Using [gamma 32P]ATP, we found that the heavy chains of myosin II released from membrane-cytoskeletons were phosphorylated, and this phosphorylation resulted in disassembly of myosin filaments. Using ITP (a substrate for myosin II ATPase) and/or ATP gamma S (a substrate for myosin II heavy-chain kinase [MHCK]), we demonstrated that phosphorylation of myosin heavy chains occurred at the foci within the actin network, a result that suggests that MHCK was localized at the foci. These results together indicate that, during contraction, the heavy chains of myosin II that have moved toward the foci within the actin network are phosphorylated by a specific MHCK, with the resultant disassembly of filaments which are finally released from membrane-cytoskeletons. This series of reactions could represent the mechanism for the relocation of myosin II from the cortical region to the endoplasm.

Contraction of Dictyostelium ghosts reconstituted with myosin II.

Cytoskeletons, or 'Triton ghosts,' that contained mainly actin and myosin II were prepared from Dictyostelium discoideum amoebae by extraction with Triton X-100. The Triton ghosts contracted immediately upon addition of ATP. However, under high-salt conditions in the presence of ATP, they did not contract but released myosin II. Almost all of the applied myosin II became associated with ghosts when myosin-free Triton ghosts, prepared in this way, were incubated with purified actin and then with myosin II from Dictyostelium. Immunofluorescence microscopy demonstrated that the associated myosin was localized in the cortical actin layer of the ghosts. Furthermore, the ghosts reconstituted with purified myosin resumed ATP-dependent contraction. Skeletal muscle myosin could also restore contractility to ghosts from which myosin had been extracted. The amount of myosin II necessary for the contraction of the ghosts was calculated by two methods. Less than 10% of the myosin II in intact cells was necessary for the contraction. These results show that myosin II is responsible for the contraction of the Dictyostelium cytoskeleton.

Immunoelectron microscopic studies of the ultrastructure of myosin filaments in Dictyostelium discoideum.

We have examined the characteristics of myosin in situ in Dictyostelium amoebae. By an improved immunofluorescence method, we previously found rod-like structures that contain myosin, which we call "myosin rods", in amoebae (Yumura. S., and Fukui, Y. (1985) Nature, 314: 194-196). Although we prepared samples for electron microscopy using conventional chemical fixation to clarify the ultrastructure of the myosin rods, we could not find any filamentous structures similar to myosin thick filaments. Therefore, we examined the effects of chemical fixatives on the myosin rods in situ by immunofluorescence staining. When cells were fixed in more than 0.05% glutaraldehyde or more than 1% osmium tetroxide at 4 degrees C, the myosin rods disappeared. These effects did not result from loss of the antigenicity, because a monoclonal myosin-specific antibody was able to react with synthetic myosin filaments treated with 0.5% glutaraldehyde or 2% osmium tetroxide. Cells fixed by the procedure used for immunofluorescence staining were post-fixed with permissible concentrations of chemical fixatives and prepared for examination by transmission electron microscopy. We found discrete filaments of about 12 nm thickness between the microfilaments. These filaments were shown to contain myosin by immunoelectron microscopy with an immunogold probe. These filaments were thinner than synthetic myosin thick filaments formed in vitro in the presence of 10 mM MgCl2, but they were similar to those formed in the presence of 2 mM MgCl2, or under nearly physiological ionic conditions. The images after immunofluorescence and immunogold labeling both suggested that these 12-nm-thick filaments in Dictyostelium amoebae were myosin filaments in situ.