Skeletal Muscle Thin Filaments Research Articles

Proposal of alignment-independent classification of electron microscopic images with helical symmetry and its application to reconstituted thin filaments of skeletal muscle

A new method to classify noisy images of filaments with helical symmetry was developed and applied to cryo-electron microscopic images of skeletal muscle thin filaments. In this method, images are transformed to minimum data sets which are independent of the relative view parameters, such as rotation of a three-dimensional object around the axis of helical symmetry, translation along an axis and operation to make a filament upside down, and classified by a principal component analysis. By this method, atypical data can be identified without aligning procedures, and small structural changes and polymorphic variation are expected to be detected objectively.

Tropomodulin caps the pointed ends of actin filaments.

Many proteins have been shown to cap the fast growing (barbed) ends of actin filaments, but none have been shown to block elongation and depolymerization at the slow growing (pointed) filament ends. Tropomodulin is a tropomyosin-binding protein originally isolated from red blood cells that has been localized by immunofluorescence staining to a site at or near the pointed ends of skeletal muscle thin filaments (Fowler, V. M., M. A., Sussman, P. G. Miller, B. E. Flucher, and M. P. Daniels. 1993. J. Cell Biol. 120: 411-420). Our experiments demonstrate that tropomodulin in conjunction with tropomyosin is a pointed end capping protein: it completely blocks both elongation and depolymerization at the pointed ends of tropomyosin-containing actin filaments in concentrations stoichiometric to the concentration of filament ends (Kd < or = 1 nM). In the absence of tropomyosin, tropomodulin acts as a "leaky" cap, partially inhibiting elongation and depolymerization at the pointed filament ends (Kd for inhibition of elongation = 0.1-0.4 microM). Thus, tropomodulin can bind directly to actin at the pointed filament end. Tropomodulin also doubles the critical concentration at the pointed ends of pure actin filaments without affecting either the rate of extent of polymerization at the barbed filament ends, indicating that tropomodulin does not sequester actin monomers. Our experiments provide direct biochemical evidence that tropomodulin binds to both the terminal tropomyosin and actin molecules at the pointed filament end, and is the long sought-after pointed end capping protein. We propose that tropomodulin plays a role in maintaining the narrow length distributions of the stable, tropomyosin-containing actin filaments in striated muscle and in red blood cells.

Isoform-specific interaction of tropomodulin with skeletal muscle and erythrocyte tropomyosins.

Tropomodulin is a tropomyosin-binding protein that localizes to the pointed end of striated muscle thin filaments and caps the pointed end of tropomyosin-actin filaments in vitro. Results from previous studies have suggested that tropomyosin-tropomodulin interactions are isoform-specific. To investigate the molecular basis for the isoform specific interactions of tropomyosin and tropomodulin, we isolated a cDNA for chicken skeletal muscle tropomodulin. The derived amino acid sequence of muscle tropomodulin is 86% identical with that of human erythrocyte tropomodulin, indicating that tropomodulins are highly conserved proteins. Multiple mRNAs seen on Northern blots of chicken skeletal muscle mRNA can be accounted for by multiple polyadenylation signals in the 3'-untranslated region of the cDNA. 125I-Labeled skeletal muscle and erythrocyte tropomyosins were assayed for binding bacterially expressed muscle tropomodulin using a solid phase binding assay. Unexpectedly, skeletal muscle and erythrocyte tropomyosins bound with similar affinities to muscle tropomodulin (Kd values approximately 0.2 microM). However, cross-competition studies using erythrocyte and skeletal muscle tropomyosins indicated that they bound different sites on tropomodulin. Competition studies using recombinant fragments of tropomodulin to map the binding domains for the two tropomyosins demonstrated that residues 6-94 contained the skeletal muscle tropomyosin binding site and residues 90-184 contained the erythrocyte tropomyosin binding site. Binding of different regions of tropomodulin to muscle and non-muscle tropomyosins may permit interaction of tropomodulin with tissue-specific components or may influence the pointed end capping activity of tropomodulin.

Three-dimensional reconstruction of caldesmon-containing smooth muscle thin filaments.

Caldesmon is known to inhibit actomyosin ATPase and filament sliding in vitro, and may play a role in modulating smooth muscle contraction as well as in diverse cellular processes including cytokinesis and exocytosis. However, the structural basis of caldesmon action has not previously been apparent. We have recorded electron microscope images of negatively stained thin filaments containing caldesmon and tropomyosin which were isolated from chicken gizzard smooth muscle in EGTA. Three-dimensional helical reconstructions of these filaments show actin monomers whose bilobed shape and connectivity are very similar to those previously seen in reconstructions of frozen-hydrated skeletal muscle thin filaments. In addition, a continuous thin strand of density follows the long-pitch actin helices, in contact with the inner domain of each actin monomer. Gizzard thin filaments treated with Ca2+/calmodulin, which dissociated caldesmon but not tropomyosin, have also been reconstructed. Under these conditions, reconstructions also reveal a bilobed actin monomer, as well as a continuous surface strand that appears to have moved to a position closer to the outer domain of actin. The strands seen in both EGTA- and Ca2+/calmodulin-treated filaments thus presumably represent tropomyosin. It appears that caldesmon can fix tropomyosin in a particular position on actin in the absence of calcium. An influence of caldesmon on tropomyosin position might, in principle, account for caldesmon's ability to modulate actomyosin interaction in both smooth muscles and non-muscle cells.

Kinetics of structural changes of reconstituted skeletal muscle thin filaments observed by fluorescence resonance energy transfer.

Fluorescence resonance energy transfer was measured between probes attached to Troponin-I and actin in the reconstituted skeletal muscle thin filament. Cys-133 on TnI was labeled with 4-acetamido-4'-maleimidylstilbene-2,2'-disulfonic acid, and Lys-61 on actin was labeled with fluorescein 5-isothiocyanate. A large difference in the efficiency of resonance energy transfer was observed between the presence (0.55) and absence (0.76) of Ca2+, which corresponds to the change in the distance of 0.53 nm and is in good accordance with a previous report (Miki, M. (1990) Eur. J. Biochem. 187, 155-162). The fluorescence titration curve of this reconstituted thin filament was measured versus free Ca2+ concentration under physiological conditions in the presence of 8 mM MgCl2. The transfer efficiency changed sharply between pCa 6.9 and 5.75, with a midpoint at pCa 6.25, suggesting that a conformational change of thin filaments occurs with a highly cooperative mode. The time course of this conformational change was followed by measuring the fluorescence intensity change (energy transfer efficiency change) of the labeled reconstituted thin filament after rapid decrease or increase of free Ca2+ concentration with a stopped-flow fluorometer. When the free Ca2+ concentration was increased in the presence of 5 mM MgCl2, the fluorescence intensity increased (the transfer efficiency decreased) with a rate constant of 530 +/- 170 s-1 at 20 degrees C. On the other hand, when the free Ca2+ concentration was decreased, the fluorescence intensity decreased with a rate constant of 43 +/- 5 s-1. The results indicate that the conformational change observed by fluorescence resonance energy transfer corresponds well to the regulation mechanism by troponin-tropomyosin on the reconstituted skeletal muscle thin filament.

A method for determining the periodicity of a troponin component in isolated insect flight muscle thin filaments by gold/Fab labelling.

Insect flight muscle has a large component (Tn-H) in the tropomyosin-troponin complex that is not present in vertebrate striated muscle thin filaments. Tn-H is shown by gold/Fab labelling to be present at regular intervals in insect flight muscle thin filaments. The Fab fragment of a monoclonal antibody to Tn-H was conjugated directly with colloidal gold and this probe used to label isolated thin filaments from the flight muscle of Lethocerus indicus (water bug). The distribution of gold particles seen in electron microscope images of negatively stained thin filaments was analysed to show that the probe bound to sites having a periodicity of approximately 40 nm, which is the expected value for the tropomyosin-troponin repeat. Conjugates of Fab with colloidal gold particles of 3 nm diameter labelled almost all sites. Conjugates with gold particles of 5 nm and 10 nm diameter labelled less efficiently (70% and 30%, respectively) but analysis of the distribution of inter-particle intervals among a number of filaments again gave the same fundamental spacing of 40 nm. The error in the measurements (standard deviation approximately +/- 4.2 for 5 nm gold/Fab) is less than earlier estimates for the size of the gold/Fab complex. Measurements on gold/Fab in negative stain suggest that the bound Fab contributes a shell about 2 nm in thickness around the gold particle. The radius of the probe (about 4.5 nm for 5 nm gold/Fab) would then be consistent with the value of error found. The size of the probe suggests that the gold particle binds to the side of the Fab molecule, rather close to the antibody combining site. The potential resolution of the technique may thus be better than originally expected.

Muscle abnormalities in Drosophila melanogaster heldup mutants are caused by missing or aberrant troponin-I isoforms.

We have investigated the molecular bases of muscle abnormalities in four Drosophila melanogaster heldup mutants. We find that the heldup gene encodes troponin-I, one of the principal regulatory proteins associated with skeletal muscle thin filaments. heldup3, heldup4, and heldup5 mutants, all of which have grossly abnormal flight muscle myofibrils, lack mRNAs encoding one or more troponin-I isoforms. In contrast, heldup2, an especially interesting mutant wherein flight muscles are atrophic, synthesizes the complete mRNA complement. By sequencing mutant troponin-I cDNAs we demonstrate that the molecular basis for muscle degeneration in heldup2 is conversion of an invariant alanine residue to valine. We finally show that degeneration of heldup2 thin filament/Z-disc networks can be prevented by eliminating thick filaments from flight muscles using a null allele of the sarcomeric myosin heavy chain gene. This latter observation suggests that actomyosin interactions exacerbate the structural or functional defect resulting from the troponin-I mutation.

Calponin and the composition of smooth muscle thin filaments

Recent work suggests that caldesmon and calponin, two actin-binding proteins found in smooth muscle, compete with each other for interaction with F-actin, implying that they may not be subunits of the same complex in vivo. Rather than associating with the same thin filaments in vivo, the two proteins actually may be localized on separate filaments possibly performing distinct functions. It is well-established that actomyosin ATPase and consequently tension in vertebrate smooth muscle are dependent on Ca2+-calmodulin-activated phosphorylation of myosin light chain subunits (reviewed in Kamm & Stull, 1985). The possibility of additional thin filament-linked modulation, raised by the pioneering work of Driska and Hartshome (1975) and Ebashi and colleagues (1977) and detailed extensively by Marston and colleagues (Marston et al., 1980; Marston & Smith, 1984), also has received much attention. This possibility became particularly attractive after Sobue and colleagues (1981) first isolated and characterized the protein caldesmon, which now is known to be localized exclusively on thin filaments (Marston & Lehman, 1985) and capable of modulating actomyosin ATPase in vitro (Sobue eta]., 1981; Ngai & Walsh, I984; Dabrowska et al., 1985; Marston & Lehman, I985; Chacko et al., I987). More recently, a second putative thin filament-linked regulatory protein, calponin, has been characterized by Takahashi and colleagues (1986, 1988, 1990), and their work on this 34-35 kDa molecule has stimulated further interest in the role played by thin filament-binding proteins in the contractility of smooth muscle. In these studies, antibody crossreactivity and molecular cloning demonstrated that regions of calponin have an amino acid sequence homologous to one of the tropomyosin binding domains of skeletal muscle troponinT, suggesting strongly that calponin is a true thin filament binding protein (Takahashi et al., 1988, 1990). Calponin also shares extensive sequence homology with the very abundant and modestly-named smooth muscle protein, SM 22, which as yet has no known function or binding affinities (Vancompernolle et al., 1990; Takahashi et al., 1990; Lees-Miller et al., 1987; Pearlstone et al., 1987). Calponin, like caldesmon, binds to actin with high affinity and also binds calmodulin (Lehman & Kaminer, 1984; Takahashi et al., 1986; Lehman, 1989; Winder & Walsh, 1990; Sutherland eta]., I990). However, with both caldesmon and calponin potentially interacting with actin, models of smooth muscle thin filaments seem to be becoming somewhat congested. Clearly a more complete understanding of the composition, assembly, and distribution of thin filaments in smooth muscle is required to clarify the roles played by these thin filament-binding proteins. A number of studies analyzing reconstituted and intact thin filaments have begun to address these questions and provide direction for future investigation. Experiments in the laboratories of Shirinsky and Dabrowska may have yielded a key piece of evidence pertinent to our understanding of how caldesmon and calponin interact with actin. In their studies, F-actin complexed with purified calponin was exposed to increasing concentrations of caldesmon; this resulted in binding of the caldesmon and displacement of bound calponin. Analogously, they also found that calponin was able to displace caldesmon bound to actin filaments (Makuch eta]., 1991). Hence, caldesmon and calponin compete for actin binding sites on thin filaments and apparently do not form a complex with each other, as do the subunits of troponin on skeletal muscle thin filaments. This would explain why, in some preparations, calponin seems to interfere with caldesmon inhibition of actomyosin ATPase (S. B. Marston, personal communication) and, in others, caldesmon with calponin inhibition (Sutherland eta]., 1990). Surprisingly, the presence of tropomyosin does not seem to influence calponin inhibition of ATPase (Makuch eta]., 1990; Winder & Walsh, 1990), and its role in calponin function and in calponin binding on thin filaments remains undefined. Even though calponin and caldesmon both bind to actin, it seems unlikely that they are randomly or even sequentially arranged on individual thin filaments in vivo. Caldesmon, at least, is regularly and periodically distributed on relatively intact, isolated, thin filaments in

Cooperative interactions between troponin molecules bound to the cardiac thin filament.

Striated muscle thin filaments contain many troponin molecules, which contact each other indirectly via tropomyosin and actin. Such allosteric interactions between troponin molecules may be responsible for cooperative Ca2+ binding to the regulatory sites of the cardiac thin filament (Tobacman, L. S., and Sawyer, D. S. (1990) J. Biol. Chem. 265, 931-939). To test whether thin filament-bound troponin molecules interact, we studied the competitive binding of troponin and troponin T-troponin I (an inhibitory complex lacking the Ca2+ binding subunit troponin C) to actin-tropomyosin. The relative affinities of these two forms of troponin for the thin filament depended upon their relative concentrations. Under conditions where total binding was saturated, each form binds with greater apparent affinity to sites that have similar neighbors. A theoretical model for competitive binding of two ligands to interacting sites on a linear lattice was developed and fit to the data. Surprisingly, energetically unfavorable interactions occurred between adjacent troponin and troponin T-troponin I molecules not only in the presence of Ca2+, but also in the presence of [ethylenebis(oxyethylenenitrilo)]tetraacetic acid and/or myosin subfragment 1. Removal of Ca2+ strengthened the affinity of troponin for the thin filament less than 50%. These results suggest that, even in the absence of myosin, long range allosteric interactions occur between troponin molecules. The detailed involvement of tropomyosin and actin in these interactions remains to be established.

Co-operative activation of skeletal muscle thin filaments by rigor crossbridges: The effect of troponin C extraction

When Ca 2+ binds to troponin C (TnC), all 26 troponin-tropomyosin (TnTm) complexes of a regulatory strand change in concert from the inactive to the active configuration. To see if the complexes respond similarly when they are activated by rigor crossbridges in the absence of Ca 2+, we determined the slope ( n S) of the bell-shaped pS/tension (pS = −log [MgATP], where S = MgATP 2−) relationship between pS 5, where the tension is maximal, and pS 2.3, where fibers are fully relaxed. In control skinned rabbit psoas fibers the n S value is >4; it progressively decreases with TnC extraction. This decrease in n S with TnC extraction is analogous to the decrease in the slope (Hill coefficient) of the pCa/tension (pCa = −log [Ca 2+]) relationship with extraction. Complete TnC extraction reduces the maximum substrate-induced tension by only 25%; in contrast, it reduces the maximum Ca 2+ induced tension to zero. The effects of TnC extraction on the slope of the pS/tension curve are explained by the assumptions that (1) extracted Tn-Tm complexes no longer change in concert with their neighbors but change independently of them, and (2) co-operative signals cannot cross extracted Tn-Tm complexes. The n S value, therefore, like the n H, is a direct function of the number of contiguous, intact, Tn-Tm complexes in a stretch of a regulatory strand. To describe qualitatively the bi-phasic pS/tension relationship, the mono-phasic pCa/tension relationship, and the effects of TnC extraction on them, we introduce a version of the concerted-transition formalism which includes two activating ligands, Ca 2+ and rigor crossbridges.

Calcium-Induced Movement of Troponin-I Relative to Actin in Skeletal Muscle Thin Filaments

The role of troponin-I (the inhibitory subunit of troponin) in the regulation by Ca2+ of skeletal muscle contraction was investigated with resonance energy transfer and photo cross-linking techniques. The effect of Ca2+ on the proximity of troponin-I to actin in reconstituted rabbit skeletal thin filaments was determined. The distance between the cysteine residue at position 133 (Cys133) of troponin-I and Cys374 of actin increases by approximately 15 angstroms on binding of Ca2+ to troponin-C. Also, troponin-I labeled at Cys133 with benzophenone-4-maleimide could be photo cross-linked to actin in the absence of Ca2+, but not in its presence. These results suggest that troponin-I is attached to actin in the Ca2(+)-free or relaxed state of muscle, and that it detaches from actin on Ca2+ activation of contraction. Thus, troponin-I may function as a Ca2(+)-dependent molecular switch in regulation of skeletal muscle contraction.

Resonance energy transfer between points in a reconstituted skeletal muscle thin filament

The spatial relationships between Lys-61, Cys-374 on actin or SH1 on myosin subfragment-1 (S1) and Cys-190 on tropomyosin or Cys-133 on troponin-I (TnI) in a reconstituted thin filament were studied by fluorescence resonance energy transfer. 5-(2-Iodoacetylaminoethyl)aminonaphthalene 1-sulfonic acid (IAEDANS) attached to Lys-190 on tropomyosin or to Cys-133 on TnI was used as a donor. Fluorescein 5-isothiocyanate (FITC) attached to Lys-61 or 5-(iodoacetoamido)fluorescein (IAF) attached to Cys-374 on actin and 4-dimethylaminophenyl-azophenyl 4'-maleimide (DABMI) attached to SH1 on S1 were used as an acceptor. The transfer efficiency between AEDANS attached to Cys-190 on tropomyosin and FITC attached to Lys-61 on actin was 0.42 in the absence of troponin, 0.46 in the presence of troponin and Ca2+ and 0.55 in the presence of troponin and absence of Ca2+. The corresponding distances between the probes were calculated to be 4.7 nm, 4.6 nm and 4.3 nm respectively, assuming a random orientation factor K2 = 2/3. A large difference in the transfer efficiency from AEDANS attached to Cys-133 on TnI to FITC attached to Lys-61 on actin was observed between in the presence (0.52) and absence (0.70) of Ca2+. The corresponding distances between the probes were calculated to be 4.5 nm in the presence of Ca2+ and 3.9 nm in the absence of Ca2+. The distance between Cys-190 on tropomyosin and Cys-374 on actin was measured to be 5.1 nm and the transfer efficiency (0.35) did not change upon addition of troponin whether Ca2+ is present or not, in agreement with the previous report [Tao, T., Lamkin, M. & Lehrer, S. S. (1983) Biochemistry 22, 3059-3064]. The distance between Cys-133 on TnI and Cys-374 on actin was measured to be 4.4 nm. No detectable change in transfer efficiency (0.58) was observed between values in the presence and absence of Ca2+. These results suggest that a relative movement of the two domains of actin monomer in a reconstituted thin filament occurs in response to a change in Ca2+ concentration. The transfer efficiencies between DABMI attached to SH1 on S1 and AEDANS attached to Cys-190 on tropomyosin or Cys-133 on TnI were too small (less than 2%) for an accurate estimation of the distances, suggesting the distances are longer than 7.3 nm.

Binding of phosphorylase a and b to skeletal muscle thin filament proteins

Phosphorylase plays an important role in energy generation during muscle contraction. We have demonstrated that purified rabbit skeletal muscle phosphorylase a and phosphorylase b bind to rabbit muscle F-actin, F-actin-tropomyosin, F-actin-tropomyosin-troponin, and myofibrils. Neither phosphorylase a nor phosphorylase b binds to myosin. Phosphorylase a and b bind to F-actin with S 0.5 values of 1.5 × 10 −6 and 2.1 × 10 −6 m, respectively. At saturation, 0.035 mol of phosphorylase a and b is bound for every seven G-actin monomers in the F-actin polymer. Using the F-actin-tropomyosin-troponin complex as opposed to F-actin as a binding target, there are five- and threefold increases in the maximal binding capacity for phosphorylase a and phosphorylase b, respectively, without a significant change in the S 0.5 value for either form of the enzyme. A similar stoichiometry and affinity of phosphorylase binding are observed when myofibrils are used as the binding target. Ca 2+ ions and AMP increase the maximal binding capacity for phosphorylase a to myofibrils while ATP decreases the B max. Our study suggests that in skeletal muscle, phosphorylase a and phosphorylase b may interact with the thin filament, and that this binding to thin filament proteins may be controlled by changes in sarcoplasmic concentration of Ca 2+ and ligands of phosphorylase during muscle contraction.

Excitation energy transfer studies of the proximity between tropomyosin and actin in reconstituted skeletal muscle thin filaments.

Excitation energy transfer studies of the proximity between tropomyosin and actin in reconstituted skeletal muscle thin filaments.

The distribution of troponin-like proteins on thin filaments of the bay scallop, aequipecten irradians.

Antibodies reacting either with bay scallop troponin-C-like protein or both bay scallop troponin-I-like protein and a possible scallop troponin-T were used to determine the distribution of the troponin complex on bay scallop striated muscle thin filaments. Both antibodies cause thin filaments to associate laterally and electron microscopy of such aggregates indicates a periodicity of approximately 38 nm, a distribution characteristic of proteins comprising the troponin complex in vertebrate and arthropod striated muscle. These studies therefore provide additional evidence for the presence of thin filament-linked regulation in bay scallops.

Millisecond photo-cross-linking of protein components in vertebrate striated muscle thin filaments.

Troponin I (TnI) was reacted with a photosensitive heterobifunctional reagent, methyl 4-azidobenzimidate (ABI), and then troponin was reconstituted with the ABI-modified TnI. Flash irradiation of the reconstituted troponin resulted in the formation of cross-links between TnI and other components of troponin, troponin C (TnC) and troponin T (TnT), suggesting that TnI is in contact with TnC and TnT when troponin is free in solution. No effect of calcium on the cross-linking could be detected. When the reconstituted troponin was complexed with F-actin-tropomyosin, flash irradiation of the reconstituted then filament yielded the cross-linked products of TnC-TnI, TnI-TnT, and TnI-actin in the presence ans absence of calcium, indicating that TnI is in contact with TnC, TnT, and actin in the thin filament complex irrespective of calcium concentration. No cross-linking could be detected between TnI and tropomyosin. Calcium was found to affect the cross-linking of TnC-TnI and TnI-actin; when TnC was saturated with calcium, the extent of the TnC-TnI cross-linking increased, while that of the TnI-actin cross-linking decreased. Calcium did not affect the TnI-TnT cross-linking.

Calcium ion-regulated thin filaments from vascular smooth muscle.

Myosin and actin competition tests indicated the presence of both thin-filament and myosin-linked Ca2+-regulatory systems in pig aorta and turkey gizzard smooth-muscle actomyosin. A thin-filament preparation was obtained from pig aortas. The thin filaments had no significant ATPase activity [1.1 +/- 2.6 nmol/mg per min (mean +/- S.D.)], but they activated skeletal-muscle myosin ATPase up to 25-fold [500 nmol/mg of myosin per min (mean +/- S.D.)] in the presence of 10(-4) M free Ca2+. At 10(-8) M-Ca2+ the thin filaments activated myosin ATPase activity only one-third as much. Thin-filament activation of myosin ATPase activity increased markedly in the range 10(-6)-10(-5) M-Ca2+ and was half maximal at 2.7 x 10(-6) M (pCa2+ 5.6). The skeletal myosin-aorta-thin-filament mixture gave a biphasic ATPase-rate-versus-ATP-concentration curve at 10(-8) M-Ca2+ similar to the curve obtained with skeletal-muscle thin filaments. Thin filaments bound up to 9.5 mumol of Ca2+/g in the presence of MgATP2-. In the range 0.06-27 microM-Ca2+ binding was hyperbolic with an estimated binding constant of (0.56 +/- 0.07) x 10(6) M-1 (mean +/- S.D.) and maximum binding of 8.0 +/- 0.8 mumol/g (mean +/- S.D.). Significantly less Ca2+ bound in the absence of ATP. The thin filaments contained actin, tropomyosin and several other unidentified proteins. 6 M-Urea/polyacrylamide-gel electrophoresis at pH 8.3 showed proteins that behaved like troponin I and troponin C. This was confirmed by forming interspecific complexes between radioactive skeletal-muscle troponin I and troponin C and the aorta thin-filament proteins. The thin filaments contained at least 1.4 mumol of a troponin C-like protein/g and at least 1.1 mumol of a troponin I-like protein/g.