ABSTRACT
Native thick filaments from the clam, Mercinaria mercinaria translocate actin filaments both toward and away from the center of the thick filament in an in vitro motility assay. The thick filaments from the adductor muscle are about 10 gm long whereas those from the catch muscle are 30-50 gm long. These thick filaments should prove useful in understanding the mechanism of myosin-dependent movement of actin filaments.
INTRODUCTION
The interaction of actin and myosin has been studied for many years using a wide variety of methods. These could, in general, be grouped into studies of isolated proteins using various biochemical assays or of studies using intact or permeabilized muscle fibers. A great deal has been learnt about actomyosin interactions using these combined approaches, but in some cases there have been difficulties in relating the in vitro biochemical studies with the in vivo physiological studies. The biochemical studies often provide simple interpretations of the results and are free from many of the complications of muscle fibers. Although one could easily measure the actin-activation of MgATPase activity of myosin in vitro, fundamental processes of muscle such as force production and active sliding of the two filament systems were not measured. Similarly in fiber systems, one could measure the latter two parameters but these measurements were complicated by a complex structural system and diffusion barriers.
Recently, two different types of in vitro motility assays have been developed that measure the relative movement of actin and myosin (Sheetz et al. 1984; Kohama and Shimmen, 1985; Kron and Spudich, 1986; Harada et al. 1987). These systems provide a bridge between the schools of biochemistry and physiology. The first system involves visualization of the movement of myosin-coated beads over organized arrays of actin filaments derived from microdissection of giant alga cells such as Nitella axillaris (Sheetz et al. 1984; Kohama and Shimmen, 1985). It has been used to characterize the movement of myosin filaments from a number of sources including vertebrate striated and smooth muscle (Sheetz et al. 1984; Sellers et al. 1985; Umemoto et al. 1989; Umemoto and Sellers, 1990), invertebrate muscle (Yamada et al. 1989; Vale et al. 1984) and nonmuscle myosins from both vertebrate (Umemoto et al. 1989; Umemoto and Sellers, 1990) and lower eukaryotic sources (Kohama and Shimmen, 1985; Albanesi et al. 1985; Flicker et al. 1985). The assay has been shown to be reproducible and quantitative, despite the fact that the actin substratum is derived from microdissection of a plant.
The second system uses only purified proteins. It involves observation of the movement of fluorescently labelled actin filaments over a myosin-coated surface (Kron and Spudich, 1986; Harada et al. 1987). Individual actin filaments, which are too small to be seen in the light microscope, are labelled with rhodamine phalloidin which binds stoichiometrically to the actin molecules within a filament. Upon excitation with light, this labelled actin filament emits a bright image which can easily be detected using fluorescence microscopy. These filaments can be seen to move in a serpentine manner over the myosin-coated surface in the presence of MgATP. Both polymerized myosin filaments and myosin monomers can be used to coat the surface (Umemoto and Sellers, 1990; Toyoshima et al. 1987). In addition, the proteolytic fragments of myosin, heavy meromyosin and subfragment-1 can also be used in the sliding actin assay (Toyoshima et al. 1987; Kishimo and Yanagida, 1988; Takiguchi et al. 1990).
Despite the dissimilarity in the geometry of the two assay systems there is surprisingly good agreement between the relative rates of movement of actin and myosin when the same myosin is used (Umemoto and Sellers, 1990). It is noted in both systems that the source of myosin is the basic determinant of the speed of movement, whereas the polarity of actin is thought to determine the direction of movement. This is particularly important in the sliding actin filament assay system where the myosin molecules or filaments are generally bound to the glass surface in a random manner.
The range of rates produced by different myosin molecules is astounding given the degree of sequence conservation that is seen in the head region of myosin molecules from different sources. In our laboratory we have measured rates of movement of purified myosins that extend from a pedestrian 0.04 gms-1 for human platelet myosin to about 6 gm s-1 for clam adductor myosin under the same ionic conditions. In preliminary experiments, rates of 30–60 gms-1 have been measured for sliding of purified actin filaments over coverslips coated with high speed supernatants of soluble extracts of Nitella axillaris protoplasm, giving a total range of movements of about 1500 fold for the rate of sliding of purified actin filaments under the same ionic conditions (Rivolta et al. 1990).
In the sliding actin assay, the myosin bound to the surface is not usually visualized because of the small size of monomeric myosin and synthetic thick filaments. Recently, we (Sellers and Kachar, 1990) have been observing the movement of fluorescently labelled actin filaments over large native thick filaments isolated from molluscan muscle, using a gentle technique developed by Yamada et al. (1989). These thick filaments consist of myosin molecules packed in a bipolar manner around a core of paramyosin (Fig. IB). It is the paramyosin core which allows these thick filaments to obtain lengths up to 50 μm depending on the source of the muscle. Using videoenhanced differential interference contrast (DIC) microscopy, individual thick filaments can be directly imaged in the light microscope (Fig. IC). The quality of the image is markedly improved by the use of computer image processing such as background subtraction, frame averaging and contrast enhancement.
The ability to image both actin and myosin filaments has allowed us to correlate directly the movement of actin filaments imaged by fluorescence with their position on the myosin filament imaged by DIC. Direct comparison between the images revealed that actin filaments could bind at any position on the myosin filament and commence directed movement (Fig. 2). Actin filaments moved both towards and away from the center of the myosin filaments. The movement of actin filaments away from the center of the myosin filament is opposite that which normally occurs in contracting muscle, where it is thought that myosin pulls actin filaments only toward the center of the thick filament. Analysis of the movement demonstrated that those actin filaments traveling toward the center of the thick filament moved at a rate of 8.8±1.4μms-1, whereas those traveling away from the center of the thick filament moved at a much slower rate of 1.0±0.3qms-1 (Fig. 3). Some actin filaments traveled the entire length of the thick filament, first at the fast rate and then changing abruptly to the slower rate as it encountered myosin heads of the opposite polarity. It was observed that the direction of travel was determined by the polarity of the actin filament. A single actin filament could reverse direction of travel by detaching and rapidly reattaching after undergoing a 180° flip (Fig. 4A). This occurred most frequently with short actin filaments which undergo rapid Brownian movement upon detachment. In some cases the leading end of a longer actin filament sliding off the end of a thick filament will reattach and commence moving back toward the center. Occasionally very small actin filaments can be observed to move back and forth along a thick filament reversing its direction up to 7 times with no apparent dissociation. We believe that in these cases the actin filament may be held instantaneously by only a single myosin head allowing for Brownian movement in two dimensions which can result in a 180° rotation and reversal of direction of travel.
The native thick filaments have a large diameter and can support the movement of more than one actin filament at the same time. Actin filaments can pass one another unimpeded on the same thick filament (Fig. 4B). When this occurred one filament would be moving at the fast rate and the other at the slow rate. Actin filaments can also interact simultaneously with two adjacent myosin filaments after undergoing large changes in angle when switching from one filament to another.
Several reports have shown that the heads of isolated myosin molecules are capable of large rotational movements about the subfragment 1-subfragment 2 junction (Winklemann and Lowey, 1986; Craig et al. 1980). Similar observations have been made on myosin heads projecting from myosin filaments (Walker and Trinick, 1986, 1988). Recently it has been demonstrated that in mutated Drosophila flight muscle, in which peripheral thick and thin filaments are misregistered, the heads of myosin can bind with opposite rigor cross-bridge angles to flanking actin filaments that are of the opposite polarity (Reedy et al. 1989). Toyoshima et al. (1989) described bidirectional movement of actin filaments along tracks of myosin heads created by binding actin filaments that had been decorated with skeletal muscle heavy meromyosin to the surface of the coverslip. It was also shown that the myofibrillar Mg2+ATPase activity does not decrease at short sarcomere length, suggesting that actin of the opposite polarity can activate the Mg2+ATPase activity of myosin (Stephenson et al. 1989). We now provide definitive data that the myosin within a native thick filament can also generate the force to translocate actin both toward and away from the central bare zone.
In summary, we propose a model whereby the heads of myosin have considerable rotational flexibility. The direction of movement is determined by the polarity of actin and myosin heads can rotate perhaps 180° in order to interact with actin (Fig. 5). The reason for the slower rate of motility when actin is sliding over myosin heads of the opposite polarity is not known, but may be related to altered cross-bridge kinetics of strained heads.
Most of our experiments have been performed with native thick filaments isolated from the adductor (pink) or catch (white) muscle of the clam, Mercinaria mercinaria. The thick filaments from the adductor muscle average about 9,"m in length whereas those of the catch muscle are considerably longer (30-50,um). We concentrated most of our effort on the smaller adductor muscle thick filaments in order to avoid complications that may arise from the potentially more complex regulation of the catch muscle thick filaments. We have also prepared thick filaments from the mussel, Mytillus edulis. All of these thick filaments, while exhibiting their own characteristic rates of movement, nonetheless show a 8—10 fold slower rate when the actin filaments are moving away from the center of the thick filaments.
The structures of several invertebrate native thick filaments have been examined by high resolution electron microscopy coupled with optical diffraction techniques (Kensler and Levine, 1982; Castellani et al. 1983; Vibert and Craig, 1983; Crowther et al. 1985). Clam thick filaments have not been studied in such detail at present. For this reason, thick filaments from other species may prove to be more interesting. In particular, Limulus muscle has thick filaments of about 10 μm in length which can be isolated using gentle means. The structure of this thick filament has been extensively studied (Kensler and Levine, 1982). To date, we have not been able to use successfully Limulus thick filaments in the motility assay, probably due to the presence of excess actin in the preparation.
The ability to image both the moving actin filament and the thick filament which is generating the motile force represents a significant advance in the utility of in vitro motility assays. It should allow more direct observation of actin-myosin interactions and may be of use in systems designed to measure force in vitro using such techniques as actin-coated microneedles (Kishino and Yanagida, 1988) and optical tweezers (Block, 1990). In addition, these molluscan myosin filaments are regulated by calcium binding to myosin (Kendrick-Jones et al. 1970). This affords the possibility of studying the regulation of movement in vitro. Preliminary experiments show that movement of actin filaments along clam thick filaments is dependent upon micromolar levels of calcium.
In order to aid others who wish to use the system of native thick filaments in in vitro motility assays, we will conclude by giving the details of the preparation and the assay. The thick filament preparation was modified from that described by Yamada et al. (1989). Muscle from either the pink adductor muscle or the white catch muscle of the clam, Mercinaria mercinaria was removed from the shell, diced into small pieces and placed in buffer A (10 mM ATP, 10mM MgCl2, ImM EGTA, 20mM MOPS (pH7.0), 3mM NaN3, ImM dithiothreitol, 0.1 mM PMSF). The muscle was rinsed twice in buffer A and then homogenized twice in 5 ml of buffer A in a Sorvall Omnimixer (small cup) at a setting of 7.5, for 7 s on ice. The homogenized muscle was mixed with an equal vol of buffer A containing 0.1% Triton X-100. After sitting for 5 min on ice, the homogenate was sedimented at 500 g for 5 min. The supernatant was carefully removed and sedimented for 30 min (adductor muscle) or 20 min (catch muscle) at 5000 g. The pellet was gently resuspended in 30 ml of buffer A and sedimented at 500 g for 5 min. The supernatant was removed and sedimented at 5000 g for 20 or 30 min (see above). The two centrifugation steps were repeated once more and the resulting pellet was gently resuspended in 1–2 ml of buffer A, incubated for 20 min on ice and sedimented at 500 g for 5 min. The supernatant was recovered and stored on ice until used. An SDS polyacrylamide gel of the sample is shown in Fig. IB. A band migrating near the position of actin is observed, but lack of staining of the preparation with rhodamine phalloidin suggests that it is not actin.
We have modified the published procedures for sliding actin motility assays for use with the native thick filaments. Glass coverslips can be used instead of nitrocellulose-coated coverslips. A 40,ul aliquot of thick filaments (30—200 fig ml-1 diluted with buffer A) was applied to the slide and a flow cell consisting of two slivers of coverslip flanked by a ribbon of Apiezon grease topped by a coverslip was constructed around the droplet. This was done in order to avoid orientation of the thick filaments under flow. After 30-60 s the flow cell was washed with 20 mM KC1, 20 mM MOPS (pH 7.2), 5mM MgCl2, O.lmM EGTA, 2mM DTT, O.ômgmU1 bovine serum albumin. After 1 min incubation, the flow cell was washed with motility buffer (20 mM KC1, 10 mM MOPS (pH7.2), 5mM MgCl2, ImM ATP, 0.2mM CaCl2, O.lmM EGTA, 2 mM DTT, 2.5mgml-1 glucose, O.lmgmU1 glucose oxidase, 0.02mgml”1 catalase) containing 2nM rhodamine phalloidin actin. In some cases, care was taken to minimize the contribution of rigor-like heads in the preparation which bind to actin in an ATP-independent manner and tether actin filaments. To accomplish this, after the bovine serum albumin blocking step, the flow cell was washed with motility buffer containing 1,UM phalloidin-labelled actin (note: not rhodamine-labelled phalloidin). After 1—5 min incubation this solution was washed out with motility buffer and replaced with the motility buffer containing 2nM rhodamine phalloidin-labelled actin. Images were recorded on either VHS or U-matic videotapes or in some cases on optical memory disks. In order to obtain nonoverlap of the thick filaments, concentrations between 30-50 .iigml-1 total protein must be applied.
Bright rhodamine-phalloidin labelled actin is prepared by drying 60 ill of a 3.3,UM rhodamine phalloidin stock (Molecular Probes, Oregon, USA) under vacuum. The sample is resuspended in 90 fil of 4mM imidazole (pH 7.0), 2HIM MgCl2, ImM DTT, O.lmM EGTA, 3mM NaN3 and sonicated in an ultrasonic cleaner for 10 min. F-actin (10,id) is added to a final concentration of 2,UM from a freshly diluted 20 μM solution of F-actin in the same buffer as above. This is allowed to sit on ice for at least 1–2 h before use. The rhodamine-phalloidin labelled actin solution is usually good for about 1 week.
ACKNOWLEDGEMENTS
To obtain video-enhanced DIC images of the thick filaments a Zeiss axiomat microscope was used in the critical illumination mode (Kachar et al. 1987) equipped with a ×100, 1.4 NA objective. The video image was background subtracted, averaged and contrast enhanced using the Image 1 image processor (Universal Imaging, Pennsylvania, USA). Higher resolution images were obtained using a Newvicon video camera (Dage-MTI), but for direct comparison between the fluorescent image and the DIC image a SIT camera (Dage-MTI) was used. In some cases the fluorescent image was captured with a intensified Newvicon (KS 1380, Videoscope International). The rate of movement was quantified using the computer assisted tracking program of Steven Block (1990).