Abstract The interaction of the heavy chain (HC) and the light chain (cdLC1) of cardiac S1 (cdS1) with F-actin was studied by cross-linking, Western blotting, and fluorescence polarization methods. Incorporation of cdLC1 in cross-linked products was examined by Western blots using the primary antibody against 71-74 residues of cdLC1. Cross-linking with zero-length, water-soluble reagent yielded three products with apparent molecular masses of 150, 160, and 210 kD. Like in the case of cross-linking of skeletal S1 with actin, these complexes included only HC of S1 and actin. The composition of the products were as follows: 150 kD, one HC of S1 cross-linked through a primary site (on the C-terminal of the 20-kD fragment) to the N-terminus of actin; 160 kD, one HC of S1 cross-linked through a secondary site (on the 50 kD fragment) to the N-terminus of actin; and 210 kD, one HC of S1 cross-linked through primary and secondary sites to two actins. Four additional products with apparent molecular masses of 66, 120, 185, and 235 kD contained cdLC1 and were identified as cdLC1+actin, cdLC1+HCS1, cdLC1+actin+HCS1, and cdLC1+two actins+HCS1, respectively. The same products were observed when cross-linking was performed in cardiac myofibrils incubated with cdS1. The production of cross-linked complexes of the heavy and light chain with actin decreased with an increase in the molar ratio of cdS1:actin. To test whether the orientation of myosin heads depended on a degree of occupation of thin filaments, myofibrils were irrigated with varying concentrations of cdS1. Fluorescence polarization measurements of cdS1 bound to individual I-bands revealed that the orientation depended on the concentration.
Muscle contraction is caused by cyclic interaction of myosin heads (S1) with actin. Despite the recent solution of the atomic structures of actin and skeletal S1,1 2 3 the molecular mechanism of this interaction remains unknown. Myosin head and F-actin form a stable complex in the absence of ATP. Recently, using skeletal S1, we have shown that depending on the molar ratio of S1 and actin, this complex can consist of one S1 and one actin or one S1 and two actins.4 This has been confirmed by others.5 Further, we have shown that the alkali light chain 1 of skeletal S1 can bind F-actin only when the molar ratio is low.6 In the present work we addressed the question of whether the interaction of cardiac S1 with F-actin has the same features. This is not self-evident at all, despite the fact that there is substantial homology between sequences of the heavy chain of cdS1 and skS1,7 that cdS1 and skS1 have similar binding constants,8 9 and that ATPase activity of both is enhanced significantly by cross-linking to F-actin.10 11 This is because the overall homology is only 73%. Moreover, there are significant differences between the light chains: The cardiac myosin, unlike skeletal, does not contain a shorter isoform of the essential light chain. Both LC1s have an extended positively charged N-terminal extra peptide of 41 residues,12 13 14 15 16 17 but in cardiac and skeletal myosins, this peptide has only 30% homology.12 13 Sliding velocities of F-actin over skeletal and cardiac myosins in in vitro motility assay are very different.18
In the present work we show that depending on the molar ratio of S1:actin, the heavy chain of cdS1—like that of skS1—can be cross-linked to one or to two actin protomers in actin filament by EDC. When F-actin was saturated with heads, cdS1 could only bind to a single monomer: HC of cdS1 was cross-linked to actin only through a primary site located on 20-kD proteolytic fragment. On the other hand, when F-actin was nonsaturated with heads, cdS1 bound to two actins: HC of cdS1 was cross-linked both through a primary and a secondary site located on 50-kD proteolytic fragment. cdLC1 formed a complex with actin only when F-actin was nonsaturated with cdS1. cdS1 produced identical complexes with nonsaturated skeletal F-actin (where it binds to two actins) and with thin filaments of cardiac myofibrils. The polarization of fluorescence of cdS1 saturating and nonsaturating individual I-bands of myofibrils was significantly different.
Materials and Methods
5′-IATR, a 5′ isomer, was from Molecular Probes. EDC and nucleotides were from Sigma. Monoclonal antibody against cardiac cdLC1 was purchased from Alexis Co.
Cardiac S1 was prepared from porcine hearts by the method of Taylor and Weeds (1976).8 Some experiments were done using S1 kindly donated by Dr S. Margossian (Albany Medical College). Skeletal S1, S1 isoforms, actin, and myofibrils were prepared as described elsewhere.19 20 The concentrations of proteins were measured using the following values of the extinction coefficients21 : S1, A1% (280)=7.5, using molecular mass of 120 kD for S1 (A1) and 111 kD for S1 (A2); G-actin, A1% (290)=6.3; F-actin, A1% (290)=6.7. The concentrations of myofibrils was measured by dissolving them in 2% SDS and using A1% (280)=7.0. The quality of proteins was checked by SDS-PAGE.
Labeling of S1
Labeling of S1 was carried out as in Reference 2222 . The concentration of labeled S1 was calculated by absorbance at 280 nm after subtracting absorbance of 5′ IATR at this wavelength.
S1 and F-actin were mixed at different molar ratios and incubated for 1 hour at room temperature; appropriate amounts of EDC were then added. Reactions were stopped by adding an equal volume of electrophoresis sample solution (4% SDS, 24% glycerol, 100 mmol/L Tris, 4% mercaptoethanol, 0.02% bromphenol blue). Unless otherwise indicated, all cross-linking experiments were done in solutions containing 0.2 mmol/L MgCl2, 50 mmol/L KCl, and 10 mmol/L Tris-HCl, pH 7.5. The low concentrations of MgCl2 and KCl were used to prevent actin filament bundle formation.3 23 Light scattering measurements did not detect any bundle formation in this buffer solution.
Gel electrophoresis was carried out according to Reference 2424 , using 8% polyacrylamide gels. After electrophoresis and staining, the slab was dried using a Novex Gel Dryer Kit. The relative intensity of various bands was measured by scanning the dried slab gel (ScanJet, Hewlett Packard) and quantifying the images using Image Pro Plus program (Media Cybernetics). A calibration of the scanner was done using Edmund Scientific Stepped Density Filters.
Measuring Polarization of Fluorescence
Polarization was measured from images of isolated myofibrils as described before.25 Briefly, 0.5 mg/mL myofibrils were incubated for 1 hour (with 2 μmol/L labeled S1) or overnight (with 0.1 μmol/L labeled S1). Free S1 was removed by centrifugation at 3000 rpm at room temperature in a desktop centrifuge. Birefringent crystal (Melles Griot) in the emission path split the fluorescent light into two orthogonally polarized components so two images of each myofibril were obtained. A CCD camera (Photometrics) recorded both images, which were later analyzed to determine polarization using Image Pro Plus program (Media Cybernetics). If ∥I∥ is the intensity of the I-band illuminated and viewed with parallel polarized light and ∥I⊥ is the intensity of the I-band illuminated with parallel polarized light and viewed with perpendicularly polarized light, then P∥=[(∥I∥)/C∥−(∥I⊥)]/[(∥I∥)/C∥+(∥I⊥)]. Similarly, P⊥=[(⊥I⊥)/C⊥−(⊥I∥)]/[(⊥I⊥)/C⊥+(⊥I∥)]. The correction factors C∥=∥I∥/∥I⊥ and C⊥=⊥I⊥/⊥I∥ are necessary because the dichroic mirror transmits the parallel component of emitted light more efficiently than the perpendicular component and are for Zeiss FT 580 mirrors 0.91 and 1.60, respectively. Intensities were corrected for background immediately adjacent to a given I-band.
Cross-linking of the HC of cdS1 to One or Two Actins
Fig 1⇓ shows that cross-linking of cdS1 to F-actin yields the following adducts: 66 kD, 150 to 160 kD doublet, 185, 210, and 235 kD. We will show below that 66, 185, and 235 kD adducts all contain cdLC1. However, we will first discuss the adducts that do not contain cdLC1: 150 to 160 kD doublet and 210 kD band. When actin is in molar excess over cdS1 (nonsaturation), cross-linking with cdS1 yields 150 to 160 kD doublet and 210 kD band (the corresponding bands in the presence of glycine buffer migrate with an apparent molecular mass [Mapp] of 175 to 185 and 265 kD, respectively). Cross-linking of skeletal S1 yields a similar pattern (lane 1). We measured the ratio of S1 to actin in each of the bands. Cys-707 of the HC of cdS1 was labeled with 5′-IATR; Cys-374 of skeletal actin was labeled with 5′-IAF. cdS1 (2 μmol/L) was cross-linked to 8 μmol/L of F-actin for 40 minutes at room temperature using 50 mmol/L of EDC. The adducts (150 to 160 kD doublet and 210 kD band) were separated in 7.5% SDS-PAGE in a tube apparatus (model 150A, BioRad). After the electrophoresis, the fluorescent bands were marked under the UV illumination, the glass tube was mounted at 50° relative to the incident beam in SLM 500C spectrofluorometer, and fluorescence spectra were recorded from the marked bands. Front-face illumination was used to avoid a concentration quenching of fluorescence. The fluorescence of 5-IAF on actin and 5′-IATR on cdS1 was excited at 480 nm and 525 nm, and the emission was recorded at 525 and 580 nm, respectively. Fluorescence contributions of actin and S1 are separated by this choice of wavelengths.3 Contribution of fluorescein and rhodamine to the background was measured separately. Contribution of the background fluorescence to the 150 to 160 kD doublet was recorded from a region of the gel between 120 kD and the doublet bands. Contribution of the background to the 210 kD band was recorded from a region of the gel between 185 kD and the doublet bands. The appropriate background was subtracted from the fluorescence of the bands. In the doublet the ratio of the fluorescence of actin to the fluorescence of cdS1 was 3.42. In the 210 kD band, the ratio was 6.43. Therefore there was 6.43/3.42=1.88 more actin per cdS1 in 210 kD band than in the doublet. Assuming that in the doublet there is one actin per S1,14 we conclude that in the 210 kD band, one cdS1 binds to two actins.
Inhibition of the Production of 160 and 210 Adducts at High cdS1:Actin
An important consequence of the suggestion that the secondary site is physiologically important is the inhibition of cross-linking through the secondary site (on 50 kD fragment of HC) at high S1:actin (saturation). This inhibition is expected to result from the fact that an increase in the S1:actin ratio (ie, progressive filling up of F-actin) leaves no vacant space on F-actin necessary for binding through both sites. Thus all adducts involving binding through the secondary site must disappear at high ratios. Fig 2A⇓ shows that the production of 160 and 210 bands is progressively inhibited as S1:actin increases from 0.17 (lane 1) to 4 (lane 6). The data are summarized in Fig 2B⇓. Earlier, we have demonstrated such inhibition for skS1.6
Cross-linking of cdLC1 to Actin
Fig 1⇑ shows that in addition to the 150 to 160 doublet and 210 kD band, cross-linking also produced 66 kD, 185 kD, and 235 kD adducts. To identify the protein components of these adducts, the peptides were electroblotted into nitrocellulose membrane after the electrophoresis. The membrane was incubated for 1 hour with blocking solution, then with the primary monoclonal antibodies to muscle light chain 1 for 1 hour and finally with the horseradish peroxidase–conjugated secondary antibodies for 1 hour. Luminescence was detected by x-ray film. The result is shown in Fig 3A⇓. The major products were similar to the ones identified for skS1-actin6 and were 235 kD, A+A+cdS1+cdLC1; 185 kD, A+cdS1+cdLC1; 120 kD, cdS1+cdLC1; and 66 kD, A+cdLC1.
Inhibition of the Production of 66, 185, and 235 Adducts at High cdS1:Actin
Fig 3A⇑ shows that the production of 66, 185, and 235 kD adducts is progressively inhibited as S1:actin increases from 0.17 (lane 1) to 4 (lane 6). The amount of actoS1 complexes were almost the same in each sample because the concentration of S1 or actin was kept constant at 1 μmol/L. The intensities of the bands were measured as described in “Materials and Methods” and normalized to that at the molar ratio S1:actin=0.17. The normalized intensities are plotted against molar ratio S1:actin in Fig 3B⇑. The decrease in the intensity is particularly dramatic considering that the amount of cdS1 is 4 times higher in lane 6 than in lane 1.
Cross-linking of cdS1 to Cardiac Thin Filaments
It is essential to demonstrate that, like in isolated acto-S1, the covalent complexes of cdS1 with one and two cardiac actin monomers exist in cardiac myofibrils. This is shown in Fig 4⇓: cardiac myofibrils (4 mg/mL) were incubated with 0.5 μmol/L rhodamine-labeled cdS1 (ie, actin was in excess) and cross-linked with 50 mmol/L EDC for 1 hour at room temperature. The doublet and 210 kD band were produced regardless of whether cdS1 was cross-linked to skeletal F-actin or to cardiac thin filaments.
Polarization of Fluorescence
The foregoing results suggest that cdS1-saturating thin filaments in myofibrils bind to one actin and that cdS1-nonsaturating thin filaments bind to two actins. The following experiments were carried out to test whether the orientation of cdS1 is different under these conditions. For experiments in which F-actin was saturated with S1, myofibrils (0.5 mg/mL) were incubated for 1 hour with 2 μmol/L labeled S1. For experiments in which F-actin was nonsaturated with S1, myofibrils (0.5 mg/mL) were incubated overnight with 0.1 μmol/L labeled S1. After extensive washing to eliminate free S1, myofibrils were placed on a rotating stage of a polarization microscope, which was rotated until a single myofibril was oriented parallel to the (rectangular) excitation slit. A slit in the back focal plane of the objective (Zeiss Planapo ×63, NA=1.4) was narrowed to exclude as much background fluorescence as possible. Polarizations of several I-bands were measured. This has an advantage that the A-band is excluded from the analysis (see “Discussion”). Fig 5⇓ shows two typical pairs of polarized images from a myofibril irrigated with rhodamine-S1. In A, the excitation was parallel to myofibrillar axis; the top image was formed with the light polarized perpendicularly and the bottom one with the light polarized parallel to the myofibrillar axis. In B, the excitation was perpendicular and the top image was formed with the light polarized perpendicularly and the bottom one with the light polarized parallel. Fig 5C⇓ illustrates the procedure used to calculate polarization of an I-band pointed to by an arrow: The images of the I-band were enclosed in an area-of-interest (AOI) and their average intensities calculated. The average intensity #2 was subtracted from the corrected intensity #1 and divided by their sum, giving parallel polarization of fluorescence. The average intensity #4 was subtracted from the corrected intensity #3 and divided by their sum, giving perpendicular polarization of fluorescence. The same experiment was done using a myofibril irrigated with nonsaturating concentration of rhodamine-S1. The results are summarized in the Table⇓. Parallel polarizations of S1 added at saturating and nonsaturating concentrations were not different (P>.05, paired t test), but perpendicular polarizations were statistically significantly different (P=2.71×10−2, t=3.4).
Cross-linking of HC of cdS1
Apparent molecular mass of 150 kD adduct, the fact that it does not contain cdLC1 (Fig 3A⇑), and the fact that it comigrates with skeletal 150 kD band suggest that it is a complex of HCS1 bound to one actin through the primary site. Similarly, 160 kD band is most likely a complex of HCS1 bound to one actin through the secondary site,14 and 210 kD band is a complex of HC+actin+actin.3 As F-actin became progressively saturated with S1, the formation of 160 and 210 kD complexes was inhibited (Fig 2⇑, A and B). This inhibition suggests that there is a difference in conformation between the complexes of cdS1 with one and two actins. Polarization experiments show that this conformational difference is associated with the orientational difference. The same behavior was observed in skeletal S1.6 The primary site on S1 includes lysines 636, 637, 640, 641, and 64226 ; the secondary site is located between Trp-510 and Trp-59514 and may include the residues of the flexible loop 567-578. Three-dimensional reconstruction of myosin head2 supports this assignment.
Cross-linking of cdLC1
In addition to the doublet and 210 kD band, cross-linking of cdS1 yielded products with Mapp of 66, 120, 185, and 235 kD. Mapp of 66 kD and the fact that it contains cdLC1 and actin (Fig 3A⇑) suggest that it is a complex of one actin and one cdLC1. The inhibition of production of 66 kD indicates that cdLC1, like the secondary site on the heavy chain, can only bind to actin at low degrees of saturation of a filament. It has been shown that the N-terminus of skeletal A1 could interact with the C-terminus of actin.6 14 27 28 The present results showed that cardiac LC1, like skeletal, could also interact with actin despite the fact that there is only 30% homology between the 41 N-terminal residues of cardiac and skeletal LC1.12 13 The 120 kD complex contained cdS1, cdLC1, and no actin. Consistent with this identification is the fact that the formation of 120 kD complex is not inhibited at higher ratios of S1:actin. The formation of the 66, 185, and 235 kD complexes was strongly inhibited in fully saturated actin filaments. It follows that the cross-linking of cdLC1 to actin correlates with cross-linking of HC of cdS1 through the secondary site (Fig 3B⇑). Thus 185 kD band contains most likely cdLC1 and one HCS1 cross-linked through the secondary site to one actin, and it is generated from 160 kD peptide by cross-linking of cdLC1 to actin. The 235 kD band contains cdLC1 and one HC cross-linked through the primary site to one actin and through the secondary site to another actin and it is generated from 210 kD adduct by cross-linking of cdLC1 to the second actin. Fig 6⇓, A and B, shows a schematic diagram of cross-linking of cdLC1 to actin at different degrees of saturation.
The present data clearly show that as in skeletal muscle, there is no interaction between cdLC1 and actin under conditions of full saturation. The fact that the 66 kD adduct could not be formed in saturated filaments supports a suggestion that the conformation of S1 is different at different molar ratios. The N-terminus of LC1 can only reach the C-terminus of actin when S1 binds to two actin protomers. Since in muscle fibers actin is in excess, it is likely that interaction of cdLC1 with thin filaments is physiologically important.16
The fact that the pattern of cross-linking is different in saturated and nonsaturated filaments (ie, that 160, 210, and 160 kD adducts could not be formed in saturated filaments) suggests that the conformation of S1 is different at different molar ratios. To quantify this difference, we carried out polarization experiments in myofibrils in which F-actin is naturally aligned along the myofibrillar axis. cdS1 forms the same complex with isolated F-actin and with thin filaments of myofibrils because the pattern of cross-linking was the same in both cases (Fig 4⇑). Polarizations were measured in the individual I-bands. Contribution from the A-bands complicates measurements because the amount of fluorescent light originating from the A-bands is small, giving rise to undefined polarization (the light is a sum of fluorescence emanating from the overlap zone29 and of little or no fluorescence emanating from the rest of the A-band). The results show that as in skeletal muscle, perpendicular polarization is significantly greater for myofibril irrigated with saturating concentration of rhodamine-S1.
Selected Abbreviations and Acronyms
|A1/A2||=||alkali (light chain 1/2)|
|cdS1||=||cardiac subfraction 1|
|skS1||=||skeletal subfraction 1|
This study was supported by grant AR40095 from the National Institution of Arthritis and Musculoskeletal and Skin Diseases.
- Received May 22, 1997.
- Accepted August 28, 1997.
- © 1997 American Heart Association, Inc.
Rayment I, Rypniewski W, Schmidt-Base K, Smith R, Tomchik DR, Benning MM, Winkelman DA, Wesenberg G, Holden HM. Three-dimensional structure of myosin subfragment-1: a molecular motor. Science. 1993;261:50-58.
Fisher AJ, Smith CA, Thoden J, Smith R, Sutoh K, Holden HM, Rayment I. Structural studies of myosin:nucleotide complexes: a revised model for the molecular basis of muscle contraction. Biophys J.. 1995;68:19s-28s.
Bonafe N, Chaussepied P. A single myosin head can be cross-linked to the N termini of two adjacent actin monomers. Biophys J.. 1995;68:35s-43s.
Nakayama S, Tanaka H, Yajima E, Maita T. Primary structure of chicken cardiac myosin S-1 heavy chain. J Biochem.. 1994;115:909-926.
Taylor RS, Weeds AG. The magnesium-ion-dependent adenosine triphosphatase of bovine cardiac myosin and its subfragment-1. Biochem J.. 1976;159:301-315.
Siemankowski RF, White HD. Kinetics of the interaction between actin, ADP, and cardiac myosin-S1. J Biol. Chem.. 1984;259:5045-5053.
Lauer B, Van Thiem N, Swynghedauw B. ATPase activity of the cross-linked complex between cardiac myosin subfragment 1 and actin in several models of chronic overloading: a new approach to the biochemistry contractility. Circ Res.. 1989;64:1106-1115.
Lowey S, Waller GS, Trybus KM. Function of skeletal muscle myosin heavy and light chain isoforms by an in vitro motility assay. J Biol Chem.. 1993;268:20414-20418.
Sweeney HL. Function of the N terminus of the myosin essential light chain of vertebrate striated muscle. Biophys J.. 1995;68:112s-119s.
Yamashita H, Sata M, Sugiura S, Monomura S, Serizawa T, Iizuka M. ADP inhibits the sliding velocity of fluorescent actin filaments on cardiac and skeletal myosins. Circ Res.. 1994;74:1027-1033.
Spudich J, Watt S. Regulation of rabbit muscle contraction. J Biol Chem.. 1971;246:4866-4871.
Ando T, Scales D. Skeletal muscle myosin subfragment 1 induces bundle formation by actin filaments. J Biol Chem.. 1985;260:2321-2327.
Yamamoto K, Sekine T. Interaction of alkali light chain 1 with actin: effect of ionic strength on the cross-linking of alkali light chain 1 with actin. J Biochem.. 1983;94:2075-2078.