Heterogeneity of Atlantic salmon troponin-I.

Three non-identical, full length troponin-I (Tn-I) clones were isolated from an Atlantic salmon myotomal (trunk) muscle cDNA library. The primary structures, which are predicted to range from 172 to 180 amino acids in length, exhibit similar percent identity scores when compared with fast, slow and cardiac specific Tn-Is from higher vertebrates. When the sequence data are considered along with the results of Western blotting it is evident that Tn-I is more heterogeneous in Atlantic salmon than has been previously shown in higher vertebrates.

The effect of cross-innervation on the tropomyosin composition of rabbit skeletal muscle.

Soleus, semitendinosus and crureus muscles of the rabbit were found to contain alpha- and beta-tropomyosin subunits and additional forms that have been provisionally designated gamma and delta. Extensor digitorum longus and psoas muscles contained only alpha and beta subunits, the relative proportions of which varied between single fibres of psoas muscle. On cross-innervation of rabbit soleus and extensor digitorum longus muscles, the fraction of the total tropomyosin present as the beta subunit remained constant. The relative proportions of alpha, gamma and delta subunits changed as would be expected from the change in speed that occurred.

Phosphorylation of tropomyosin during development in mammalian striated muscle.

Alpha-Tropomyosin from rat cardiac muscle was shown by two-dimensional gel electrophoresis to become phosphorylated when tissue slices were incubated in Eagle's medium supplemented with 32Pi. In the adult rat and mouse heart the level of phosphorylation was approximately 30% but the level was much higher in the foetal heart (60-70%). A similar developmental trend was observed in skeletal muscle from the rat and mouse, where phosphorylated forms of both alpha- and beta-tropomyosins were observed. When rat cardiac cells were grown in tissue culture in the presence of 32Pi, radioactivity was incorporated into the region of the gel containing tropomyosin.

Investigation of the effects of phosphorylation of rabbit striated muscle alpha alpha-tropomyosin and rabbit skeletal muscle troponin-T.

FPLC has been employed to prepare the phosphorylated and unphosphorylated forms of rabbit striated muscle alpha alpha-tropomyosin (TM), and the major isoform of rabbit fast-skeletal-muscle troponin-T (Tn-T2f) and corresponding chymotryptic fragment T1 (residues 1-158), in order to investigate the effects which these in vivo modifications have on thin filament function. In all instances, no significance could be attributed to the presence of a phosphate moiety on acetyl serine 1 of Tn-T (or fragment T1). As expected, fragment T1 increased the relative viscosities of solutions of unphosphorylated alpha alpha-TM, but this induction was noticeably lower for phosphorylated alpha alpha-TM. In affinity chromatography experiments, fragment T1 bound equally well to either form of alpha alpha-TM, but the interaction between fragment T2 (residues 159-259) and phosphorylated alpha alpha-TM was strengthened relative to the control. In the presence of alpha alpha-TM (unphosphorylated), fragment T1 was found to down regulate the actin-activated myosin-S1 MgATPase activity, indicating that this portion of Tn-T possesses modulatory properties. Under the same conditions, less inhibition was observed with phosphorylated alpha alpha-TM. When the two different forms of alpha alpha-TM were reconstituted into a complete regulatory system, the activation of myosin-S1 was double for those thin filaments containing the phosphorylated molecule. Dephosphorylation of the phospho alpha alpha-TM reduced the rates to control values. In ATPase Ca2+ titrations, these systems exhibited no difference in the co-operativity of activation and little or no difference in the pCa2+ 1/2 value. Developmentally linked changes in the steady-state phosphorylation of alpha alpha-TM could be a mechanism to increase the activating propensity of thin filaments, by modifying the functional properties of the T1 section of Tn-T.

Interaction of rabbit skeletal muscle troponin T and F-actin at physiological ionic strength.

Troponin T has been shown to interact significantly with F-actin at 150 mM KC1 by using an F-actin pelleting assay and 125I-labeled proteins. While troponin T fragment T1 (residues 1-158) fails to pellet with F-actin, fragment T2 (residues 159-259) mimics the binding properties of the intact molecule. The weak competition of T2 binding to F-actin, shown by subfragments of T2, indicates that the interaction site(s) encompass(es) an extensive segment of troponin T. The extent of pelleting of troponin T (or T2) with F-actin is only marginally altered in the binary complex troponin IT (or T2), indicating that the direct interactions either of troponin T (or T2) or of troponin I, or both, with F-actin are weakened when these components are incorporated into a binary complex. The binding of troponin T (or T2) is moderately (-Ca2+) or more extensively reduced (+Ca2+) in the presence of troponin C. The pelleting of Tn-T seen in the presence of Tn-C (-Ca2+) and Tn-I was further reduced when either Tn-I or Tn-C (-Ca2+) was added, respectively, to form a fully reconstituted Tn complex. As noted by others, whole troponin shows little sensitivity to Ca2+ in its binding to F-actin (-tropomyosin). These and other observations, taken together with the restoration of troponin IC (+/- Ca2+) binding to F-actin by troponin T, implicate a role for the interaction of troponin T and F-actin in the thin filament assembly.

Isolation and characterization of tropomyosin from fish muscle.

A comprehensive survey of tropomyosin from various fish myotomal muscles is reported. The fish tropomyosins were blocked at the N-terminus and, as expected, were found to be of similar amino acid composition, alpha-helical content (> 90% at 10 degrees C) and molecular weight to other vertebrate striated muscle forms. The tropomyosins of salmonids and herring muscle were noticeably heterogeneous when assessed by 2D-PAGE. The distribution of isoforms was tissue-specific: slow muscle contained alpha-type tropomyosin while fast muscle contained beta-type tropomyosin. In other species (cod, haddock, wolf-fish and sharks) alpha-type tropomyosins were present in both kinds of muscle but beta-tropomyosin was absent.

The effects of troponin T fragments T1 and T2 on the binding of nonpolymerizable tropomyosin to F-actin in the presence and absence of troponin I and troponin C.

Using a nonpolymerizable form of tropomyosin (NPTM) we have investigated the interactions between the T1 (residues 1-158) and T2 (residues 159-259) regions of troponin T and the other components of the thin filament at 50 mM KCl +/- Ca2+. Under these conditions the binding of NPTM to F-actin is fully restored by whole troponin (+/- Ca2+), and in each case, retains a residual degree of cooperativity as demonstrated by Scatchard and Hill plots. Fragment T2 alone had a small inductive effect on the interaction of NPTM with F-actin. In the presence of troponin I, this interaction is increased to a level which exceeds that observed with either component alone. The effects of T2 and troponin I are moderately (-Ca2+) and markedly (+Ca2+) reduced by troponin C. While fragment T1 alone did not promote induction, it accentuated the effects of T2 and troponin I. Since T1 does not interact with T2 or troponin I but does interact weakly with the NH2 terminus of tropomyosin and can be expected to bind weakly at the residual interaction site(s) at the COOH terminus of NPTM, the observed effects of T1 have been ascribed to the linking of neighboring NPTM molecules at their ends.

Factors determining the subunit composition of tropomyosin in mammalian skeletal muscle.

Adult rat fast-twitch skeletal muscle such as extensor digitorum longus contains alpha- and beta-tropomyosin subunits, as is the case in the corresponding muscles of rabbit. Adult rat soleus muscle contains beta-, gamma- and delta-tropomyosins, but no significant amounts of alpha-tropomyosin. Evidence for the presence of phosphorylated forms of at least three of the four tropomyosin subunit isoforms was obtained, particularly in developing muscle. Immediately after birth alpha- and beta-tropomyosins were the major components of skeletal muscle, in both fast-twitch and slow-twitch muscles. Differentiation into slow-twitch skeletal muscles was accompanied by a fall in the amount of alpha-tropomyosin subunit and its replacement with gamma- and delta-subunits. After denervation and during regeneration after injury, the tropomyosin composition of slow-twitch skeletal muscle changed to that associated with fast-twitch muscle. Thyroidectomy slowed down the changes in tropomyosin composition resulting from the denervation of soleus muscle. The results suggest that the 'ground state' of tropomyosin-gene expression in the skeletal muscle gives rise to alpha- and beta-tropomyosin subunits. Innervation by a 'slow-twitch' nerve is essential for the expression of the genes controlling gamma- and delta-subunits. There appears to be reciprocal relationship between expression of the gene controlling the synthesis of alpha-tropomyosin and those controlling the synthesis of gamma- and delta-tropomyosin subunits.

Isolation and characterization of an abundant and novel 22-kDa protein (SM22) from chicken gizzard smooth muscle.

Chicken gizzard smooth muscle contains a highly abundant protein (SM22) with an apparent Mr on sodium dodecyl sulfate-polyacrylamide electrophoretic gels of 23,000. The ratio of actin:SM22:tropomyosin in this tissue is estimated to be 6.5(+/- 0.8):2.0(+/- 0.2):1.0. At least three isoelectric isoforms are present in ratios of alpha:beta:gamma of 14:5:1 with alpha the most basic and gamma the most acidic. A method for the purification of SM22 and partial separation of its isoforms is described. Amino acid analyses of purified alpha and beta demonstrate the presence of 1 and 2 half-cystines, respectively, and a lower content of basic amino acids in beta. A value of 22,000 for the Mr of alpha estimated by sedimentation equilibrium indicated its presence as a monomer at physiological ionic strengths. Estimates of the translational frictional coefficient (f/fmin) of alpha calculated from its Stokes radius (25.5 A) and Mr were consistent with its existence as a moderately asymmetric globular protein. Calculations based on its far-ultraviolet CD spectrum provided values of 37% alpha-helix, 31% beta-sheet, 5% beta-turn, and 27% random coil. SM22 was shown not to share functional properties with several proteins of similar Mr and isoelectric point such as myokinase, brain 23-kDa protein, and troponin I. We conclude that it is a novel protein not previously isolated or characterized from any tissue.

Effect of phosphorylation on the interaction and functional properties of rabbit striated muscle alpha alpha-tropomyosin.

Phosphorylated rabbit cardiac alpha alpha-tropomyosin has been prepared either enzymatically (Montgomery, K., and Mak, A.S. (1984) J. Biol. Chem. 259, 5555-5560) or by fractionation of the phosphorylated and nonphosphorylated forms on a Mono Q column in 9 M urea, 50 mM Tris, pH 8.0. Although the phosphorylated and nonphosphorylated forms showed no difference in their F-actin binding properties, the phosphorylated protein had substantially higher viscosities at low ionic strengths, indicating a greater propensity for head-to-tail interaction. Similar measurements showed the strengthening of this interaction by whole troponin to be substantially reduced by phosphorylation even though the binding of whole troponin and troponin T to tropomyosin was demonstrated by affinity chromatography to be, if anything, strengthened by phosphorylation. In a reconstituted actin (4 microM) plus myosin subfragment 1 ATPase assay (50 mM ionic strength), significantly higher activities over a range (1 to 8 microM) of subfragment 1 concentrations were observed with phosphorylated tropomyosin compared with the nonphosphorylated protein. In the fully reconstituted system with troponin, there was no significant difference in the inhibition of ATPase in the absence of Ca2+. However, in its presence, the activities were appreciably increased with the phosphorylated tropomyosin compared to those with the nonphosphorylated form. These differences were eliminated by treatment of the phosphorylated tropomyosin with alkaline phosphatase. This is the first demonstration of an effect of phosphorylation on the functional properties of tropomyosin.

An abundant and novel protein of 22 kDa (SM22) is widely distributed in smooth muscles. Purification from bovine aorta.

Using a rabbit polyclonal-antibody preparation directed against the chicken gizzard protein, we demonstrated by immunoblotting the presence of the 22 kDa protein (SM22) in a variety of chicken smooth-muscle-containing organs, including uterus, intestine, gizzard, oesophagus and aorta. Protein SM22 was present in only trace amounts in brain, liver and heart, and could not be detected in chicken breast muscle. The antibody preparation did not cross-react with extracts of bovine aorta. However, the presence of SM22 as a major component in bovine aorta and pig carotid was demonstrated by its co-migration with the purified chicken gizzard protein on one- and two-dimensional polyacrylamide electrophoretic gels. Its molar abundance relative to actin was estimated to be 0.9:6.0 and 1.4:6.0 for bovine aorta and pig carotid respectively. Like the chicken gizzard protein, it separates on pH-gradient electrophoresis into at least three variants, alpha, beta and gamma, with similar apparent Mr. Purification of the aorta SM22 showed it to have a similar amino acid composition to the chicken gizzard protein. We conclude that SM22 is widely distributed and an abundant and unique protein component of smooth-muscle tissues of birds and mammals.

Effects of deletion of tropomyosin overlap on regulated actomyosin subfragment 1 ATPase.

The role of the overlap region at the ends of tropomyosin molecules in the properties of regulated thin filaments has been investigated by substituting nonpolymerizable tropomyosin for tropomyosin in a reconstituted troponin-tropomyosin-actomyosin subfragment 1 ATPase assay system. A previous study [Heeley, Golosinka & Smillie (1987) J. Biol. Chem. 262, 9971-9978] has shown that at an ionic strength of 70 mM, troponin will induce full binding of nonpolymerizable tropomyosin to F-actin both in the presence and absence of calcium. At a myosin subfragment 1-to-actin ratio of 2:1 ([actin] = 4 microM) and an ionic strength of 50 mM, comparable levels of ATPase inhibition were observed with increasing levels of tropomyosin or the truncated derivative in the presence of troponin (-Ca2+). Large differences were noted, however, in the activation by Ca2+. Significantly lower ATPase activities were observed with nonpolymerizable tropomyosin and troponin (+Ca2+) over a range of subfragment 1-to-actin ratios from 0.25 to 2.5. The concentration of subfragment 1 required to generate ATPase activities exceeding those seen with actomyosin subfragment 1 alone under these conditions was 3-4-fold greater when nonpolymerizable tropomyosin was used. Similar effects were seen at the much lower ionic strength of 13 mM and are consistent with the reduced ATPase activity with nonpolymerizable tropomyosin observed previously [Walsh, Trueblood, Evans & Weber (1985) J. Mol. Biol. 182, 265-269] at low ionic strength and a subfragment 1-to-actin ratio of 1:100. Little cooperativity in activity as a function of subfragment 1 concentration with either intact tropomyosin or its truncated derivative was observed under the present conditions. Further studies are directed towards an understanding of these effects in terms of the two-state binding model for the attachment of myosin heads to regulated thin filaments.

Further characterisation of fast, slow and cardiac muscle tropomyosins from salmonid fish.

Separate cDNA libraries were constructed from cardiac muscle and slow myotomal muscle of mature brown trout (Salmo trutta). The complete sequence of tropomyosin (TM) that is specific to these muscles was determined from full-length transcripts isolated from the corresponding library. The identity of the sequences was supported by protein data. When compared to the sequence of Atlantic salmon fast myotomal TM [Heeley, D. H., Bieger, T., Waddleton, D. M., Hong, C., Jackman, D. M., McGowan, C., Davidson, W. S. & Beavis, R. C. (1995) Characterisation of fast, slow and cardiac muscle tropomyosins from salmonid fish, Eur. J. Biochem. 232, 226-234], the main difference in the N- and C-terminal sequences comprising the site of end-to-end overlap occurs at residue 276 where an asparagine in fast TM is replaced by a histidine in both cardiac and slow TM. Trout cardiac TM exhibited greatest similarity to chicken cardiac TM while trout slow TM exhibited greatest similarity to skeletal alpha-TMs. Thus, none of the three salmonid TM sequences corresponds to a beta-type TM. In calorimetry experiments (0.1 M salt, pH 7.00, t = 10-60 degrees C), in the presence of dithiothreitol, differences were observed in the thermal unfolding profiles of the purified isoforms. A single endotherm (tm = 39.5 degrees C) was noted for cardiac TM. Two endotherms were observed for fast TM [tm = 26.5 degrees C and 39.8 degrees C (main)] and slow TM [tm = 37.4 degrees C and 46.9 degrees C (main)]. Fast TM was cloned and over expressed in the bacterial cell lines JM105 and BL21. Upon cell lysis, recombinant TM (rc TM) made in JM105 was rapidly and quantitatively cleaved between residues 6 and 7. Intact rc TM was produced by using BL21, as shown by Edman-based sequencing, carboxypeptidase digestion and mass analysis. In viscometry assays, performed at low ionic strength (pH 7.00, t = 5 degrees C) the full-length rc TM exhibited markedly lower relative viscosity values than the corresponding wild type.

Characterisation of troponin-T from salmonid fish.

Five major troponin-T isoforms were isolated from the myotomal muscles of Atlantic salmon: three from fast muscle (Tn-T1F, Tn-T2F and Tn-T3F) and two from slow muscle (Tn-T1S and Tn-T2S). In addition to their presence in troponin preparations, these proteins were also recognised to be Tn-T on the basis of immunoreaction with anti-troponin-T antibodies and partial amino acid sequence. The electrophoretic mobility in the presence of SDS of the various Tn-Ts increases in the order: 1S < 1F < 2S < 2F < or = 3F. Compositional analysis shows that the higher M(r) forms (1F and 1S) contain considerably more proline, glutamic acid and alanine than the lower-M(r) forms (2F, 3F and 2S). Every isoform lacks cysteine and phosphoserine is present only in isoforms 2F and 3F. All of the Tn-Ts, with the exception of isoform 1F, are N-terminally blocked. CNBr fragments from same cell type Tn-Ts yield identical sequences over at least fifteen Edman cycles. Two full-length cDNA sequences, presumed to represent 1S and 3F, or isoforms that are highly similar, are reported. As documented for higher vertebrate Tn-Ts, the predicted primary structures display a non-uniform distribution of charged amino acids and greater divergence at each end than in the central section. The most striking difference between the two salmonid proteins is the presence of a N-terminal (proline-, glutamic acid- and alanine-rich) extension of about fifty amino acids in Tn-T1s (278 amino acids) that is missing from the fast muscle Tn-T (223 amino acids). The sequences also differ in that 1S lacks the known phosphorylation site while the fast-type isoform contains serine next to the initiating methionine. Of the two, the slow isoform has accumulated the greater number of substitutions.

Characterisation of fast, slow and cardiac muscle tropomyosins from salmonid fish.

Tropomyosin (TM) has been isolated from the cardiac muscle, and fast and slow trunk (myotomal) muscles of the mature salmonid fish Atlantic salmon (Salmo salar) and rainbow trout (Salmo gairdneri). When examined electrophoretically, isoforms of TM were detected which were specific, and exclusive, to each type of muscle. Cardiac and fast muscles contained single and distinct isoforms, while slow muscle contained two distinct isoforms, closely related in terms of apparent M(r), and pI. There was no detectable difference between the same TM type from either salmon or trout. On a variety of gel systems, the cardiac and slow isoforms migrated in close proximity to each other and to rabbit alpha-TM. The fast isoform comigrated with rabbit beta-TM. In developing salmon fry, a more acidic (unphosphorylated) variant of TM was present in addition to, and of similar M(r) to, the fast adult isoform. This TM declined in steady-state level during maturation and was virtually undetected in adult muscle. All of the isolated TMs contained little or no covalently bound phosphate and were blocked at the N-terminus. The amino acids released by carboxypeptidase A, when ordered to give maximal similarity to other muscle TMs, were consistent with the following sequences: fast (LDNALNDMTSI) and cardiac (LDHALNDMTSL). The C-terminal region of the slow TM contained His but was heterogeneous. In viscosity measurements, performed as a function of increasing protein concentration, at low ionic strength (t = 5 degrees C, pH 7.00), fast TM exhibited the highest relative viscosity values. Lower and equivalent levels of polymerisation occurred with the cardiac and slow TMs. Polymerisation of all three isoforms was temperature-dependent, with cardiac TM being least sensitive and fast TM being most sensitive. Determination of the complete coding sequence of adult fast TM confirmed the findings of the carboxypeptidase analysis, but the remainder of the sequence more closely resembled alpha-type TMs than beta-type TMs. Overall, salmon fast TM contains 20 (mostly conservative) substitutions compared to rabbit striated muscle alpha-TM and 40 (mostly conservative) substitutions compared to rabbit striated muscle beta-TM. This demonstrates that electrophoretic mobility is not, in all instances, a suitable method to assess the isomorphic nature of striated muscle TMs.