Dimerization of the Head–Rod Junction of Scallop Myosin

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS ARTICLE NO.

252, 595– 601 (1998)

RC989603

Dimerization of the Head–Rod Junction of Scallop Myosin Andra´s Ma´lna´si-Csizmadia,* Ethan Shimony,* Gyo¨rgy Hegyi,* Andrew G. Szent-Gyo¨rgyi,† and La´szlo´ Nyitray*,1 *Department of Biochemistry, Eo¨tvo¨s Lora´nd University, Budapest, Hungary, H-1088; and †Department of Biology, Brandeis University, Waltham, Massachusetts 02254-9110

Received September 30, 1998

We have compared the dimerization properties and coiled-coil stability of various recombinant fragments of scallop myosin around the head–rod junction. The heavy-chain peptide of the regulatory domain and its various extensions toward the a-helical rod region were expressed in Escherichia coli, purified, and reconstituted with the light chains. Rod fragments of the same length but without the light-chain binding domain were also expressed. Electron micrographs show that the regulatory domain complex containing 340 residues of the rod forms dimers with two knobs (two regulatory domains) at one end attached to an ;50-nm coiled coil. These parallel dimers are in equilibrium with monomers (Kd 5 10.6 mM). By contrast, complexes with shorter rod extensions remain predominantly monomeric. Dimers are present, accounting for ca. 5% of the molecules containing a rod fragment of 87 residues and ca. 30% of those with a 180-residue peptide. These dimers appear to be antiparallel coiled coils, as judged by their length and the knobs observed at the two ends. The rod fragments alone do not dimerize and form a coiled-coil structure unless covalently linked by disulfide bridges. Our results suggest that the N-terminal end of the coiled-coil rod is stabilized by interactions with the regulatory domain, most likely with residues of the regulatory light chain. This labile nature of the coiled coil at the head–rod junction might be a structural prerequisite for regulation of scallop myosin by Ca21-ions. © 1998 Academic Press

Scallop and other molluscan muscles are triggered by direct Ca21-binding to myosin, representing one of the simplest form of regulation of any motile systems. Like other conventional myosins, the scallop protein 1 To whom correspondence should be addressed at Department of Biochemistry, Eo¨tvo¨s University, Puskin u. 3., H-1088, Budapest, Hungary. Fax: (36-1)-266-7830. E-mail: [email protected]. Abbreviations used: ELC, essential light chain; RLC, regulatory light chain; HC, heavy chain; RD, regulatory domain; RD-R, regulatory domain-rod construct.

has two globular heads attached to an elongated rod. The heads consist of two domains: the motor domain which binds actin and has ATPase activity and the regulatory domain which turns on the motor activity in response to Ca21-binding. The regulatory domain (RD) consists of an approximately 10-kDa heavy chain fragment (HC) and one each of the regulatory (RLC) and essential light chains (ELC). The RD is produced by limited proteolysis of isolated scallop myosin heads and shown to bind Ca21 with the same affinity and specificity as native myosin [reviewed in Ref. 1]. The three-dimensional structure of scallop RD has recently been determined with bound Ca21-ions [2, 3]. Considerably less is known about the structure and stability of the myosin rod. Certainly it is important to regulation, as the isolated heads of both scallop and vertebrate smooth muscle myosins (where regulation is achieved by reversible phosphorylation of RLC) are unregulated. There is also little doubt that the rod is a two-stranded a-helical coiled coil, as inferred from sequence analysis [4, 5] and visualized by electron microscopy of metal shadowed [6 – 8] or negatively stained myosin molecules [9, 10]. However, it is not known where exactly the coiled coil begins and how stable it is. The general view is that the HC sequence immediately after the “last Pro” (residue 835 in the scallop sequence) in the head region forms an a-helix and dimerizes into a coiled coil. Recently it has been demonstrated that smooth muscle myosin fragments are not fully regulated unless they include a portion of the coiled-coil rod [11–13]. In addition, regulation is accomplished by interactions between head groups. Smooth muscle myosin with only one head is constitutively active [14], whereas the analogous fragment from scallop is poorly regulated [15]. Computer modeling of the head/tail junction of scallop myosin indicates some of the possible headhead and head–rod interactions which might be involved in the regulatory switch [16].

595

0006-291X/98 $25.00 Copyright © 1998 by Academic Press All rights of reproduction in any form reserved.

Vol. 252, No. 3, 1998

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

In this paper, we studied dimerization properties of recombinant RD constructs containing various HC subunits extended into the rod domain (RD-R), and rod fragments without the light chain binding domain. We show that the RD-R, with at least 300 residues of rod, forms parallel dimers like myosin molecules with truncated heads. However, shorter extensions result in unexpected antiparallel dimers with the same Ca21binding properties as the parallel dimer or monomer RD-Rs. In contrast to RD-Rs, rod fragments shorter than 340 residues do not dimerize unless forced to do so by disulfide bonding, indicating that the regulatory domain influences the stability of the head–rod junction in scallop myosin. MATERIALS AND METHODS DNA constructs. Scallop myosin cDNA fragments, coding for the light chain binding region of the HC and various length of the rod region were cloned into expression vectors to produce the HC components of the following recombinant RD and RD-Rs: sRD (Asp749Pro835), m1RD-R (Asp749-Arg922), m2RD-R (Asp749-Leu947), m3RD-R (Asp749-Asn1015) and hRD-R (Asp749-Leu1175). The recombinant sRD corresponds to the proteolytic RD [17], while the mRD-R and hRD-R constructs contain additional residues from the rod domain. A parallel set of construct contained cDNA inserts coding for various HC fragments representing only the N-terminal region of the rod (subfragment-2 peptides). These constructs began with the codon for Ala840, a residue near the presumed N-terminus of the coiled coil. S-947 encoded the residues Ala840 to Leu947, S-1015 extended to Asn1015, and S-1175 reached on to Leu1175. Thus, the two longer constructs encoded the cysteine residue naturally found at position 1014. Additional short constructs were engineered so as to encode a cysteine at the N-terminus or one at each termini of the recombinant peptide. Thus S-919N had codons for Cys.Gly preceding the wildtype sequence for Leu836 . . . Leu919. S-905NC encoded the mutated sequence Ser838.Cys839 . . . Cys905 (Residue 839 is normally isoleucine, whereas 905 is leucine). The N- and C-terminal cysteines in the resulting peptide would map to heptad residues a and d in the putative coiled coil. Cloning sites, the 39 stop signals and the mutations were introduced by polymerase chain reaction using synthetic oligonucleotides and a scallop MHC clone as template [18]. All of the inserts were ligated into the pET15b (Novagen) vector between NdeI and BamHI sites. The expression vector introduced a His-tag fusion peptide at the N-terminus of the constructs. Protein expression, purification and RD reconstitution. Recombinant DNA constructs were transformed into E. coli BL21(DE3) pLysS cells. A single colony was inoculated into 1 l 2YT medium containing 100 mg/l ampicillin. Cultures were incubated with shaking at 37°C until OD600 reached 0.5– 0.7 and induced with 1 mM IPTG. After 4 h incubation, the cells were harvested, resuspended in 8 M GuHCl, 0.5 M NaCl, 20 mg/ml leupeptin, 20 mg/ml pepstatin and 0.5 mM PMSF, sonicated to disrupt DNA, and cleared by centrifugation. The pH of the supernatant was adjusted to 7.5 with 20 mM Tris–HCl, and loaded onto His-Bind Resin (Novagen) or ProBond (Invitrogen) nickel-chelating column. After washing the column with 8 M urea, 0.5 M NaCl, 20 mM Tris–HCl, pH 5.3, the target protein was eluted by decreasing the pH to 4.0. Alternatively, the HC of sRD was loaded onto Sep-Pak C18 Cartridge (Waters) in 8 M GuHCL, 0.1% TFA, 10% acetonitrile, washed with 30% acetonitrile and 0.1% TFA, and eluted with 40% acetonitrile, 0.1% TFA. All HC samples were purified by HPLC on C18 MN Nucleosil HPLC column eluting with 0.1% TFA/acetonitrile gradients. The N-terminal His-tag from

the recombinant subfragment-2 peptides was removed by thrombin cleavage as suggested by the manufacturer of the ProBond resin (Invitrogen). Purified recombinant HC fragments were reconstituted into RDs by mixing stoichiometric amount of the appropriate HC and both light chains in 6 M GuHCl, 10 mM DTT, 20 mM Tris–HCl, pH 7.5, dialyzing overnight against renaturation buffer (40 mM NaCl, 2 mM MgCl2, 1 mM CaCl2, 0.1 mM DTT, 20 mM Tris–HCl, pH 7.5) in 6 M urea, then continuously decreasing the urea concentration to zero over 24 h. Protein precipitate was removed by centrifugation and the reconstituted RD-Rs and sRD were purified either on a MonoQ HR 5/5 (Pharmacia) or a Superdex 200 HR 10/30 (Pharmacia) column equilibrated in renaturation buffer. Analytical gel filtration. Superdex 200 HR 10/30 and Superdex 75 HR 10/30 (Pharmacia) FPLC gel exclusion columns were equilibrated with 40 mM NaCl, 2 mM MgCl2, 0.1 mM CaCl2, 20 mM Tris–HCl, pH 7.5. 50 –100 ml of 0.5–1 mg/ml protein was loaded at a flow rate of 0.5– 0.7 ml/min. 0.1– 0.5% isopropanol was also added to reduce protein aggregation. Flow rate reproducibility of the LKB HPLC pumps (Model 2249) was better than 0.3%. Electron microscopy. Samples of different RD-Rs were diluted to about 30 mg/ml in a spraying buffer containing 30% (by volume) of 1 M ammonium acetate, 1 mM MgCl2 and 70% glycerol. They were sprayed onto mica and rotary shadowed with platinum as described in [7]. Electron micrographs were recorded at a 50,0003 magnification with a Philips EM301 electron microscope. Gel electrophoresis. SDS–PAGE was performed on slab gels according to the method of Laemmli [19]. Gels containing 15% acrylamide were prepared from a stock solution of 30:1 acrylamide to N,N-methylene-bis-acrylamide by weight. Samples were mixed with SDS–PAGE sample buffer with or without b-mercaptoethanol. Native gels (containing 10% acrylamide) were prepared as were denaturing gels but without SDS. The gels were fixed and stained in 0.1% Coomassie brilliant blue R250, 10% acetic acid, 50% methanol. Ca21 binding assay. Ca-binding of sRD and RD-Rs were measured at 4°C by equilibrium dialysis in a medium containing 20 mM NaCl, 20 mM Hepes pH 7.6, 2 mM MgCl2, 0.1 mM DTT, 0.11 mM CaCl2 (0.2 mCi/ml) and appropriate amount of EGTA to obtain the desired free Ca21 concentration as described previously [17]. Carboxymethylation and oxidation of cysteines. Recombinant rod fragments containing one or two cysteines per chain were oxidized at 3 mg/ml in 100 mM NaCl, 20 mM Tris–HCl, pH 7.5 by either incubation with about 5-fold molar excess of CuCl2 for 30 min or by air oxidation. Excess reagent was removed by extensive dialysis. To prevent formation of disulfide bonds cysteines were reduced with 5 mM b-mercaptoethanol or modified with iodoacetic acid. The proteins were treated with 1 mM iodoacetic acid in a solution containing 20 mM Tris–HCl pH 8.0, 100 mM NaCl, 1 mM EDTA for 30 min at room temperature. The reaction was terminated by adding 2 mM DTT, and the sample were dialyzed extensively to remove the reagents. CD spectroscopy. CD measurements were carried out on a Jasco J-720 spectropolarometer (Tokyo, Japan). CD spectra were recorded between 190 nm and 250 nm. The optical path of the CD cell was 0.1 mm. The samples were in a buffer of 10 mM Tris–HCl, pH 7.5, 100 mM NaCl, 1 mM EDTA. The protein concentration was 0.8 –1.0 mg/ml. Helix content was calculated using the computer program k2d [20].

RESULTS Recombinant RDs. Five different recombinant HC peptide were expressed. The two light chains plus the shortest of these reconstitutes the sRD, which resembles the proteolytic RD [19]. The HC components of

596

Vol. 252, No. 3, 1998

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

FIG. 1. Analytical gel-filtration chromatography of m1RD-R (solid line), m2RD-R (dashed line) and m3RD-R (dotted line). 50 –100 ml of 0.5–1 mg/ml protein samples were loaded on Superdex 75 HR 10/30 (Pharmacia) FPLC gel exclusion column at 0.6 ml/min flow rate in 40 mM NaCl, 2 mM MgCl2, 0.1 mM CaCl2, 20 mM Tris–HCl, pH 7.5. The first and the second peaks represent dimer and monomer species, respectively.

Electron microscopy of RD-Rs. Topology of the RD-R dimers were examined by rotary shadowed electron microscopy (Fig. 2). The hRD-R molecules form parallel dimers: they look like truncated two-headed myosins. The two regulatory domains are evident as lobes at one end of a 51 nm rod (Fig. 3). This is the expected length of a two-stranded coiled coil of 335 residues (Pro840 to Leu1175) and a rise of 0.15 nm per residue. Surprisingly, in mRD-R molecules the two HC seemingly dimerized into an antiparallel coiled coil. The two regulatory domains are at the opposite ends of the molecules connected by a rod region, most likely an antiparallel coiled coil. The larger the HC component of mRD-R the longer the rod: 10 nm for m1RD-R, 12 nm for m2RD-R, and 18 nm for m3RD-r (see Fig. 3). When visualized on EM grids, sRD molecules appear globular and ill-defined. The largest diameter of the RD is 11-12 nm regardless of the length of the attached rod. This value agrees with the longest dimension of RD in the crystal structure (12 nm) [2,3]. Including Ca21-ions

m1RD-R, m2RD-R, m3RD-R and hRD-R extend towards the rod region (at their C-terminal end) by 87, 112, 180 and 340 residues, respectively. All five sorts of reconstituted RDs (“RD-R” if it contains a portion of the rod) bind Ca21 with the same stoichiometry (0.69 – 0.75 mol Ca21/mol protein at pCa 5) as the proteolytic RD or the native myosin [21]. Dimerization of RD-Rs. The rod portion of myosin or the proteolytic heavy meromyosin form a-helical coiled-coil dimers. To determine whether the recombinant RD-Rs dimerize, they were studied by analytical gel filtration and visualized by EM. As expected sRD formed only monomeric species. Figure 1 shows a chromatographic profile of mRD-R samples. The two peaks correspond to monomers and dimers, respectively, as judged from their apparent sizes and the fact that both peaks have the same subunit composition on SDS– PAGE (not shown). The dimer to monomer ratio increases with the C-terminal extension of the rod. For m1RD-R the dimer peak contains 6.2% of the total protein, but this fraction increases to 26.6% with m2RD-R and to 31.6% with m3RD-R. Apparently, these dimers are not in a rapid equilibrium with monomers. After separation of the monomers and dimers by gelexclusion chromatography they remained stable for a long period of time. On the other hand, dimers of hRD-R are in a rapid equilibrium with monomers. The dissociation constant (Kd) of hRD-R dimer is 10.6 mM, a calculation based on the ratio of dimer and monomer fractions of analytical gel filtration runs (data not shown).

FIG. 2. Rotary shadowed electron microscopic pictures of RD-Rs and myosin. The “nubs” at the end of RD-R molecules correspond to the regulatory domain. hRD-R and mRD-Rs form parallel and antiparallel dimers, respectively. Magnification is 50,0003. The solid bar represents 50 nm.

597

Vol. 252, No. 3, 1998

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

bands are visible on a native gel (Fig. 4B, lane A), indicative of an unstable tertiary structure downstream of the covalent linkage. The native gel in Fig. 4B demonstrates that without introducing a covalent cross-link, S-919N (Fig. 4B lanes a and a9), S-947 (lanes b and b9) and S-1015 (lanes c and c9) do not dimerize spontaneously. The longest S2 fragment (S-1175, lanes d and d9) forms some dimer in the presence of reducing agent, but the bulk remains monomeric. A very small amount of S-1175 dimer is detected by analytical gel filtration at 50 mM concentration and in the presence of DTT (Fig. 5), indicating a Kd for dimerization of larger than 1 mM.

FIG. 3. Histogram of coiled-coil lengths of RD-Rs. Frequency distribution of the rod region in m1RD-R, m2RD-R, m3RD-R, and hRD-R is shown. 25–50 molecules of each sort were measured.

during sample preparation did not have any detectable effect on the shape and topology of RD-Rs. Dimerization of recombinant rod fragments. Myosin rod fragments with different length were produced to examine their dimerization properties. Recombinant S-919N, S-1015, and S-1175 each contain a single cysteine, whereas S-905NC has cysteines at each end. By contrast, S-947 lacks thiol-containing residues. Disulfide bonding by oxidation of these cysteins forces the peptides to associate. Such disulfide bridges can be prevented by carboxymethylating the reduced thiols. Electrophoresis on SDS- (Fig. 4A) and native gels (Fig. 4B) of the reduced versus oxidized samples reveals the contribution of noncovalent forces to dimerization of these model peptides. In the presence or absence of reducing agent, S-947 (Fig. 4A, lanes b and b9; primed letters denote samples containing reducing agent) is monomeric. Oxidation/reduction reversibly interconverts S-919N (lanes a and a9) and S-1015 (lanes c and c9) between 100% monomers and 100% dimers. S-1175 (lanes d and d9) also undergoes reversible dimerization upon formation/hydrolysis of the disulfide bond. However, never more than 70% of this peptide exists as dimer. Flanking sequences must inhibit the cysteine from pairing, despite its occupying an a position of the putative coiled-coil heptad repeat. Thiols in S-919N and S-1015, being located at the N- or C-terminus of their respective peptides, encounter no such steric hindrance. Also noteworthy is the complex pattern seen in the lane containing oxidized S-919N. At least three

CD spectroscopy. CD spectra of the recombinant rod fragments with or without disulfide-bridges are shown in Fig. 6. The spectra of carboxymethylated S-905NC and S-919N peptides are similar. Their low molar ellipticities indicate that their structure is mostly random coil; the calculated a-helix content is indeed low (9%). Oxidized S-905NC, a dimer with a disulfide bond on each end, has a dramatically different CD spectrum: two minima are observed near 220 nm and 208 nm, the signature of a-helices. Furthermore, the ratio of [U]220/[U]208 is 1.00. This value is close to 1.03, the corresponding ratio measured with documented two-stranded coiled coils [22–25]. The transition observed upon oxidation of S-919N is less dramatic: the single disulfide bond formed at its N-terminus drives the peptide into a partially alphahelical conformation (ca. 30%). Even though this cova-

FIG. 4. SDS–PAGE (A), native PAGE (B) of oxidized and reduced forms of recombinant rod fragments. Lanes a and a9, S-919N; lanes b and b9, S-947; lanes c and c9 S-1015; lanes d and d9, S-1175. Primed letters denote samples containing reducing agent.

598

Vol. 252, No. 3, 1998

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

FIG. 5. Analytical gel filtration chromatography of oxidized (solid line) and reduced (dotted line) S-1175. 50 –100 ml of 0.5-1 mg/ml protein was loaded on Superdex 200 column at a flow rate of 0.5 ml/min in 100 mM NaCl, 20 mM Tris–HCl, pH 7.5. The inferred Kd for dimerization is greater than 1 mM.

lent linkage is followed by 25 heptads, the nascent coiled coil remains unstable. DISCUSSION Myosin regulatory domains, containing various HC subunits, can be produced by reconstituting recombinant HC with RLC and ELC. Dimerization and the effect of the head–rod junction can be investigated by using HC peptides extended downstream from Pro-835 into the rod region. Such recombinant RD-Rs display the same affinity and specificity for Ca21 as do native myosin and a regulatory domain obtained by proteolysis. We expected that RD-R molecules containing a sufficient stretch of the myosin rod would self-assemble into a-helical coiled-coil dimers. Recombinant smooth muscle HMM molecules in which the heavy chains are progressively truncated in the rod region require 25–37 heptad repeats (“short HMM”) to form dimers [11–13]. If the HC component of the RD-R contains 340 residues (48 heptad repeats in hRD-R) it dimerizes into a parallel coiled coil which is in equilibrium with monomeric complexes (Kd 5 10.6 mM). The hRD-R dimers has two regulatory domains at the N-terminal end of the coiled coil; they are similar to the two “nubs” previously seen by Winkelmann et al. in a scallop myosin preparation cleaved by a bacterial protease [26]. The most unusual of the recombinant complexes are the mRD-R constructs which are extended by 87–180 amino acid residues (12–26 heptad repeats). They form dimers with one regulatory domain at each end connected by an 11- to 18-nm coiled coil. These structures appear to be antiparallel coiled coils judging from their topology and dimensions as seen by EM and their high

a-helical content determined by CD spectroscopy (data not shown). Monera et al. studied 35-residue synthetic peptides which are composed of five heptad repeats and showed that there is no physical restriction to the packing of the same hydrophobic residues into an antiparallel coiled coil [25]. In this work, we found the first instance of a fragment from a parallel coiled-coil polypeptide forming an antiparallel structure. Previously, Holtzer et al. surmised, but did not demonstrate, that antiparallel coiledcoil dimers of tropomyosin were formed in renatured samples at low pH [27]. They suggested that under native conditions interhelix salt bridges force adjacent peptides to be parallel. By contrast, molecules reassembled from chaotropic solvents in acid exist as a mixture of parallel and antiparallel dimers. In our system, it was essential to remove urea gradually and completely in order to successfully reconstitute mRD-Rs. We suspect that at an intermediate concentration of the chaotropic agent the peptide chains favor antiparallel interactions. We did not observe parallel dimers among the mRD-R molecules. Apparently, the antiparallel dimers are not in rapid equilibrium with monomers and parallel dimers; they may represent a “dead end” conformation which form only during reconstitution/refolding of the RD-R subunits from a chaotropic solvent. Presumably, the monomer to dimer ratio is also established during the reconstitution process. Since the orientation of the a-helical chains in two-stranded synthetic coiled coils was shown to be determined primarily by interchain electrostatic interactions [25, 28], it is possible that electrostatic interactions of RLC with residues of the rod region contribute to the stability of the antiparallel dimers.

FIG. 6. CD spectra of recombinant rod fragments. Carboxymethylated S-905NC (solid line), S-919N (dotted line) and oxidized S-905NC (dashed line), S-919N (crossed line) were recorded at 0.8 – 1.0 mg/ml protein concentration in 10 mM Tris–HCl pH 7.5, 100 mM NaCl, 1 mM EDTA, at 21°C.

599

Vol. 252, No. 3, 1998

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

In RD-Rs, where the number of heptad repeats of the rod region is sufficiently large (48 in hRD-R), the stabilizing effect of the hydrophobic “seam” outweighs presumed electrostatic repulsion between aligned residues. Such pairing occurs for short peptides in which all residues in the a and d positions of the heptad are hydrophobic, as in the synthetic peptide mentioned above. By contrast, the relative abundance of polar and charge groups in the scallop rod sequence precludes such an interaction unless a large stretch of rod is included. Interestingly, recombinant myosin rod fragments lacking the light chain binding domain do not dimerize. (A peptide with 48 heptad repeats shows evidence of weak self association with a Kd . 1 mM.) Interactions between the regulatory domains or of the RD with the rod somehow stabilizes the coiled coil. Coiled-coil dimers of short recombinant rod fragments from the subfragment-2 region (70 –180 residues) can be produced by introducing disulfide-bridges at their ends. An intermolecular crosslink at one end is not sufficient to make a stable coiled coil (only 30% a-helix content was measured in S-919N by CD spectroscopy). However, if both ends of the subunits are coupled by disulfides, a highly stable coiled coil is formed which has a ;100% a-helix content and a CD spectrum characteristic of two-stranded a-helical coiled coils [22–25]. What is the implication of our finding that the rod region close to the head–rod junction of scallop myosin has a labile coiled coil which is stabilized by interactions with the RD? In current views of the regulatory mechanism of molluscan and vertebrate smooth muscle myosins critical head-head and head–rod interactions play an essential role [1, 11, 12, 14, 21]. In the unusual inhibitory state (the “off” state) of regulated myosins, the two heads are thought to interact at their base through the RLCs. Alternatively (or in addition) the heads may bend back toward to rod [32, 33]. Upon Ca21-binding (or phosphorylation) these interactions are weakened, allowing the heads to move freely and independently. The unstable nature of the coiled-coil rod near the fork of the two heads could facilitate this switching mechanism. It may also contribute to the flexibility of the head–rod junction seen by electron microscopy [6 –10]. Whether the conformational change associated with the regulatory switch require unwinding and/or melting of the coiled coil at the fork of the two heads remains to be seen. Interestingly, it has been suggested that separation of the two chain of the coiled coil near the head–rod junction could also occur in vertebrate skeletal myosin [34]. If a dynamic coiled coil at the beginning of the rod is a general feature of all conventional myosins, then it may even have a more fundamental role in the function of these motor proteins besides its role in regulation.

In conclusion, we have demonstrated that the regulatory domain greatly influences the stability of the head–rod junction in scallop myosin. It is likely that the labile nature of the N-terminal end of the coiledcoil rod is a structural prerequisite for regulation of scallop myosin by Ca21-ions. ACKNOWLEDGMENTS We thank Mary Lee York and Peter Vibert for performing EM experiments. We also thank Jo´zsef Kardos for help in carrying out CD measurements, Krisztina Recska for technical assistance, and Professor La´szlo´ Gra´f for helpful suggestions. This study was supported by grants from the U.S.–Hungarian Joint Fund (317) and OTKA (T023618 to L.N. and F023615 to A.M.-C.).

REFERENCES 1. Szent-Gyo¨rgyi, A. G., and Chantler, P. D. (1994) in Myology (Engel, A. G., and Franzini-Armstrong, C., Eds.), pp. 506 –528, McGraw-Hill, New York. 2. Xie, X., Harrison, D. H., Schlichting, I., Sweet, R. M., Kalabokis, V. N., Szent-Gyo¨rgyi, A. G., and Cohen, C. (1994) Nature 368, 306 –312. 3. Houdusse, A., and Cohen, C. (1996) Structure 4, 21–32. 4. McLachlan, A. D., and Karn, J. (1982) Nature 299, 226 –231. 5. McLachlan, A. D., and Karn, J. (1983) J. Mol. Biol. 164, 605– 626. 6. Elliott, A., and Offer, G. (1978) J. Mol. Biol. 123, 505–519. 7. Flicker, P. F., Wallimann, T., and Vibert, P. (1983) J. Mol. Biol. 169, 723–741. 8. Winkelmann, D. A., and Lowey, S. (1986) J. Mol. Biol. 188, 595– 612 9. Walker, M., Knight, P., and Trinick, J. (1985) J. Mol. Biol. 184, 535–542. 10. Jontes, J. D., and Milligan, R. A. (1997) J. Mol. Biol. 266, 331– 342. 11. Trybus, K. M. (1994) J. Biol. Chem. 269, 20819 –20882. 12. Trybus, K. M., Freyzon, Y., Faust, L. Z., and Sweeney, H. L. (1997) Proc. Natl. Acad. Sci. USA 94, 48 –52. 13. Sata, M., Matsuura, M., and Ikebe, M. (1996) Biochemistry 35, 11112–11118. 14. Cremo, C. R., Sellers, J. R., and Facemayer, K. C. (1995) J. Biol. Chem. 270, 2171–2175. 15. Kalabokis, V. N.,Vibert, P., York, M. L., and Szent-Gyo¨rgyi, A. G. (1996) J. Biol. Chem. 271, 26779 –26782. 16. Offer, G., and Knight, P. (1996) J. Mol. Biol. 256, 407– 416 17. Kalabokis, V. N., O’Neall-Hennessey, E., and Szent-Gyo¨rgyi, A. G. (1994) J. Muscle Res. Cell Motil. 15, 547–553. 18. Nyitray, L., Goodwin, E. B., and Szent-Gyo¨rgyi, A. G. (1991) J. Biol. Chem. 266, 18469 –18476. 19. Laemmli, U. K. (1970) Nature 227, 680 – 685. 20. Merelo, J. J, Andrade, M. A., Prieto, A., and Moran, F. (1994) Neurocomp. 6, 1–12. 21. Kalabokis, V. N., and Szent-Gyo¨rgyi, A. G. (1997) Biochemistry 36, 15834 –15840. 22. Lau, S. Y. M., Taneja, A. K., and Hodges, R. S. (1984) J. Biol. Chem. 259, 13253–13261. 23. Hodges, R. S., Zhou, N. E., Kay, C. M., and Semchuk, P. D. (1990) Pept. Res. 3, 123–137.

600

Vol. 252, No. 3, 1998

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

24. Zhou, N. E., Kay, C. M., and Hodges, R. S. (1992) J. Biol. Chem. 267, 2664 –2670. 25. Monera, O. D., Zhou, N. E., Kay, C. M., and Hodges, R. S. (1993) J. Biol. Chem. 268, 19218 –19227. 26. Winkelmann, D. A., Almeda, S., Vibert, P., and Cohen, C. (1984) Nature 307, 758 –760. 27. Holtzer, M. E., Downing, L. J., and Holtzer, A. (1995) Biopolymers 35, 239 –244. 28. Monera, O. D., Kay, C. M., and Hodges, R. S. (1994) Biochemistry 33, 3862–3871.

29. Lupas, A., Van Dyke, M., and Stock, J. (1991) Science 252, 1162–1164. 30. Nozias, M., and Be´chet, J. J. (1993) Eur. J. Biochem. 218, 1049 – 1055. 31. Hardwicke, P. M. D., and Szent-Gyo¨rgyi, A. G. (1985) J. Mol. Biol. 183, 203–211. 32. Onishi, H., and Wakabayashi, T (1982) J. Biochem. 92, 871– 879. 33. Frado, L-L. Y., and Craig, R. (1992) J. Muscle Res. Cell Motil. 13, 436 – 446. 34. Knight, P. J. (1996) J. Mol. Biol. 225, 269 –274.

601