Figure 1. The Hanson‐Lowy model of F‐actin. Electron micrographs of negatively stained actin filaments (A, B) were the basis of the F‐actin model proposed by Hanson and Lowy (74) shown in (C). In (B), an F‐actin filament is magnified to illustrate individual actin subunits (dash marks) and points where the two strands of the filament crossover each other (arrows). In (C), the Hanson‐Lowy model is illustrated with actin subunits depicted as spheres that associate to form two intertwined “long‐pitch” helical strands crossing over approximately every 36 nm. The arrangement also yields a single‐start short‐pitch “genetic” helix with a repeat of 5.9 nm. Since each successive actin subunit is shifted axially by 27.5 Å (i.e., half staggered relative to the next one) and is rotated 166° to 167° along the short‐pitch helix (i.e., fairly close to a 180° value), the arrangement also yields two right handed long‐pitch helices repeating every 770 Å. Figure adapted, with permission, from (74). Note that up until the late 1980s and early 1990s it was not widely recognized that actin subunits are not spherical in shape, viz., that even at low resolution they look rather closer to “book‐like”‐shaped rectangular solids. Thus, when actin subunits associate to form helices, we now know that no distinct grooves are evident (cf. Figs. 9,10,12,16,17,19).

Figure 2. (A, B) The two‐state on‐off steric regulatory model showing tropomyosin movement on actin filaments as initially proposed by Huxley (102). Versions of thin filament components are drawn schematically and viewed in transverse section, actin (Ac, purple) with tropomyosin's on‐ and off‐positions (Tm, green and red) influencing the S1 myosin head (yellow), blocking binding in low‐Ca²⁺ relaxed muscle while permitting interaction in high‐Ca²⁺ active muscle. Note that the steric‐blocking position of tropomyosin (red) overlaps with the S1‐binding site on actin, in low Ca²⁺. (A) Modeling based on X‐ray diffraction data of intact muscle and EM reconstructions of S1‐decorated F‐actin. Figure adapted, with permission, from (119) based on (102). (B) A corresponding model of the thin filament by Potter and Gergely (185) based on additional in vitro studies suggesting binding interactions between troponin subunits, actin and tropomyosin and the influence of Ca²⁺ on the behavior of troponin‐tropomyosin. At low Ca²⁺ (a), troponin subunits (labeled C, I, T) clasp onto tropomyosin, trapping it over the myosin‐binding site, whereas in (b) Ca²⁺ binding to TnC breaks the actin‐TnI interaction releasing the constraint on tropomyosin. The cartoon shows movement of troponin‐tropomyosin away from the myosin‐binding site on actin following Ca²⁺ binding, so that myosin heads (S1) can now attach to actin. Figure adapted, with permission, from (185), Copyright 1974 American Chemical Society. Hitchcock et al. (84) described a similar set of changing interactions based on related work. Note that the relative azimuthal positioning of tropomyosin and troponin was not known at the time and was drawn arbitrarily in the cartoons [cf. Yang et al. (229)].

Figure 3. (A) Schematic diagram showing the arrangement of the t‐tubule and S.R. network in striated muscle. Striated muscle cells, also called muscle fibers, are composed of myofibrils (#1) containing thick and thin filaments. Myofibrils are surrounded by the S.R. (#2) and its terminal cisternae (#3). The t‐tubules, are extensions of the plasma membrane (#4), also called the sarcolemma, thus, they are part of a separate membrane system from the S.R. The juxtaposition of S.R. terminal cisternae on both sides of the t‐tubule forms a “triad.” Also shown are mitochondria and a basal lamina around the sarcolemma. Figure from (56) with permission. (B) Illustration comparing the control of Ca²⁺ release from the S.R. of skeletal (left) and cardiac muscle (right). In skeletal muscle, an action potential passing from the sarcolemma through the t‐tubule system is detected by dihydropyrindine receptors (DHPR) voltage sensors in the t‐tubule membrane, which directly open RyR1 ryanodine‐Ca²⁺ release channels in the adjacent SR. In cardiac muscle, action potentials open DHPR voltage‐dependent Ca²⁺ channels, and inflow of extracellular Ca²⁺ activates RyR2 Ca²⁺‐release channels. Figure adapted, with permission, from (117).

Figure 4. Ebashi's model of troponin‐tropomyosin on thin filaments. (A) Tropomyosin is shown lying longitudinally along successive actin subunits of the thin filament with troponin distributed periodically on every seventh actin subunit. Figure adapted, with permission, from (41). The troponin complex is now known to be considerably larger and is more asymmetric than depicted in (A). (B) The relative orientation and component organization of troponin and tropomyosin, along with end‐to‐end tropomyosin linkage, is more realistically drawn in the schematic of Flicker et al. (51). Figure adapted, with permission, from (51). T1 and T2 denote the N‐ and C‐terminal domains of TnT that can be separated by chymotryptic cleavage (155).

Figure 5. (A) Immunoelectron microscope localization of troponin. Negatively stained thin filaments emerging from a Z‐disk fragment are shown labeled with antibodies to troponin‐C (white spots) and yielding a 40 nm periodicity. Figure adapted, with permission, from (166). (B) A negatively stained Cohen‐Longley‐type tropomyosin paracrystal (26) with the arrangement of tropomyosin molecules indicated by arrows and giving rise to a 40 nm periodicity. (C) A tropomyosin paracrystal additionally containing troponin. Note the extra band of density due to the presence of troponin bound to tropomyosin and repeating at 40 nm intervals (arrows). Figures in (B, C) adapted, with permission, from (165); cf. reference (25,43,67,160) for similar images. Scale bar = 50 nm.

Figure 6. (A) Low‐angle X‐ray fiber diffraction of skinned rabbit psoas muscle fibers, shown as split “half‐film” records comparing layer lines patterns of relaxed and Ca²⁺‐activated preparations; right panel, calcium absent; left panel, calcium present. To avoid effects of myosin cross‐bridge‐actin interactions for the pattern displayed, muscle fibers were stretched by Poole et al. (183) so that thick and thin filaments in sarcomeres did not overlap. Note the heavy central scattering that dominates the pattern, and the enhancement of the J4, second‐order layer line in calcium, whereas the 59 Å layer line, derived from the F‐actin short‐pitch genetic helix is constant, as was originally noted by Huxley (102). (B, C) Difference patterns generated from the same data as in (A) by subtracting one pattern from the other. (B) In going from Ca²⁺ to EGTA containing solutions there an increase in the strength of the J₂ term. (C) In going from EGTA to Ca²⁺ there is an increase in the J₄ term. Figures taken, with permission, from (183).

Figure 7. Possible angular shifts of tropomyosin (Tm, red, green, pink) around actin (Ac, purple), as first modeled by Haselgrove (77) and by Parry and Squire (173). Here, F‐actin is illustrated in transverse section with actin subunits represented as spheres and tropomyosin as a cylinder. The fourfold symmetry of the filament is most evident when tropomyosin is positioned at a 90° azimuth (green), and least pronounced when tropomyosin is at a 0° azimuth (pink), since in the latter case the twist of the tropomyosin super‐helices (i.e., their coiled coiled‐coil configuration) and the actin helices are in phase. Similarly, the fourfoldedness enhances second‐order layer line intensity in corresponding diffraction diagrams most when tropomyosin lies closest to the 90° azimuth, and as depicted close to the “groove” between actin subunits. Red colored tropomyosin indicates a possible steric blocking location. The modeling suggests that radial distance of tropomyosin (r) to the central axis of the filament varies as tropomyosin repositions toward and away from the filament groove. It now is known that the tropomyosin radius is virtually constant during regulatory transitions, while current evidence indicates that tropomyosin's low‐ to high‐Ca²⁺ azimuthal movement over a relatively flat surface of actin is between 15° to 25° (183). Illustration adapted and modified, with permission, from (173).

Figure 8. Myosin arrowhead structure on F‐actin. (A) Electron micrograph of negatively stained actin filaments “decorated” with myosin S1. The polarity of S1 arrowheads defines the pointed and barbed ends of F‐actin filaments. (B) Model of myosin heads bound to F‐actin based on the first 3D reconstructions of the S1‐decorated F‐actin. Figure adapted, with permission, from (154).

Figure 9. Actin structure solutions. (A) Ribbon representation of the first crystal structure of the actin molecule. Note the distinct domain organization, and the cleft between subdomains containing nucleotide and Ca²⁺; domains and some residues labeled. Figure of PDB ID:1ATN adapted, with permission, from (111). (B,C) Orthogonal views of 5 actin subunits (variously colored) from a high resolution cryo‐EM structure of actin, constructed from PDB ID:3J8A (219). The shape of actin subunits is obviously flat and not spherical as depicted in earlier cartoons. The global shape of the actin subunits can be appreciated by comparing the face on and side views of the red highlighted actin molecule of the 5 subunit structure. Thus, it was known as early as 1990 that the surface over which tropomyosin moves is also likely to be flat. The actin pointed end is facing up, a convention used in all subsequent figures.

Figure 10. The first cryo‐EM reconstructions showing the molecular structure of F‐actin and location of surface binding sites (152). (A) Reconstruction of F‐actin tagged with an undecagold cluster label (gold) linked to the C‐terminal Cys374 residues of actin. (B) The undecagold attachment site is shown as a reference point establishing the actin subdomain organization in the reconstruction [subdomains numbered on one actin subunit, in (A)]. (C, D) Myosin S1 binding sites on actin are shown (yellow). Difference maps between reconstructions of F‐actin decorated with S1 and undecorated F‐actin localized the myosin‐binding sites for S1‐A2 (C) and S1‐A1 (D) (yellow) (myosin S1 density itself is not shown in this figure but is in Fig. 17B). The region defining the strong S1‐binding site on actin subdomain 1 is marked by an asterisk. Essential myosin light chains (either A1 or A2, a.k.a. alkali light chains) are low molecular subunits associated with the lever arm portion of skeletal muscle myosin S1. [Muscle myosin is a hexamer composed of two heavy chains, representing the head and rod of parts of myosin, and two nonphosphorylatable essential light chains and two phosphorylatable regulatory light chains located on the lever arms of myosin and thus positioned between the head and rod domains (142).] S1‐A2 and S1‐A1 bind to actin with some differences (arrows) due to the presence of the different myosin light chain isoforms. (E, F) Tropomyosin and its binding site on S1‐decorated F‐actin containing troponin‐tropomyosin. Note that both tropomyosin and S1 interact on the same actin interface and that tropomyosin lies next to the S1‐binding site on such S1‐decorated F‐actin filaments. Figure adapted, with permission, from (152).

Figure 11. Electron micrographs of negatively stained Ca²⁺ regulated thin filaments showing troponin and tropomyosin densities. (A) Filaments isolated directly from frog cardiac muscles. Note helically arranged tropomyosin cables (black arrows) seen extending along the filament in many micrographs. Figure adapted, with permission, from (128). (B) Reconstituted thin filament here consisting of rabbit skeletal muscle F‐actin and bovine cardiac troponin‐tropomyosin. Note in this micrograph periodic troponin densities that are in register on opposite sides of the filament (marked by white arrow heads) partially masking tropomyosin cables (black arrow). Figure adapted, with permission, from (228). Scale bar = 50 nm.

Figure 12. Ca²⁺‐induced tropomyosin movement on thin filaments revealed by 3D‐reconstruction. (A) Surface representation of EM reconstructions generated from negatively stained Limulus thin filaments following exposure to low‐ and high‐Ca²⁺ conditions. Note the movement of the tropomyosin cable (black arrows) from the outer domain (A_o) toward the inner domain of actin (A_i) that occurs in Ca²⁺. At low‐Ca²⁺, tropomyosin associates with subdomain 1 of actin and bridges over subdomain 2, whereas, at high‐Ca²⁺, its association is with subdomain 3 of actin and it arches over subdomain 4 (subdomains numbered on one actin subunit). (B) “Helical projections” of the same reconstructions made by projecting densities along their helical paths onto a plane perpendicular to the filament axis. Again, note the change in position of the tropomyosin (Tm); the center of the tropomyosin densities are highlighted in red and green for the low‐Ca²⁺ and high‐Ca²⁺ locations. The two projections are also superposed for comparison. Ca²⁺‐induced tropomyosin movement was first demonstrated on thin filaments isolated from striated muscles of the horseshoe crab Limulus and later to occur on filaments obtained from other invertebrate and vertebrate striated muscles. Figure adapted, with permission, from (122). It should be noted that the two inner and two outer domains of neighboring actin subunits shown in projection already delineate what can be considered four long‐pitch “strand‐like” densities (two inner and two outer strands). In the off‐state, tropomyosin is positioned close to the mid‐point between inner and outer domains of actin, perhaps not adding significantly to the filament fourfoldedness. In contrast, in the on‐state when tropomyosin moves closer to the inner domain of actin, the appearance of fourfoldedness is clearly emphasized. Thus, the four‐strandedness defined in the Fourier transforms of thin filaments (see Fig. 13) may not be accurately described by the conventional meaning of the term, viz., two actin and two tropomyosin strands, but rather is due to contributions from four actin components (two inner and two outer domains), augmented differentially by tropomyosin in one or another position.

Figure 13. Fourier transforms of the Limulus filament EMs. The averaged layer line data used to produce the helical reconstructions in Figure 12 is plotted here. (A) low‐Ca²⁺ and (B) high‐Ca²⁺ data. Layer line amplitudes are represented as solid lines and phases as dotted lines. Bessel and layer line orders, (n) and (l), defining the actin‐helical symmetry are given. As in fiber diffraction patterns records of intact muscle (see Fig. 6), the amplitude along the second order J4 layer line (l = 2, n = 4) of the EM data is greater for the filaments associated with the high‐Ca²⁺ than for those at low‐Ca²⁺ (arrows). Figure adapted, with permission, from (122).

Figure 14. Dual effect of tropomyosin and troponin‐tropomyosin on actomyosin subfragment 1 ATPase. Acto‐S1 ATPase assays show that troponin‐tropomyosin regulated thin filaments activate actomyosin ATPase in a Ca²⁺‐dependent manner at all S1 concentrations tested, while control F‐actin and F‐actin‐tropomyosin (no troponin) do not confer Ca²⁺ dependence. The activation of actomyosin S1 ATPase by control F‐actin is linear as a function of added S1, yet sigmoidal when either troponin‐free tropomyosin or troponin‐tropomyosin is present (plateaus at high S1 levels not shown). Thus, relative to control F‐actin values, tropomyosin, in the presence and absence of troponin and Ca²⁺, inhibits actomyosin ATPase at low S1:F‐actin ratios, and stimulates ATPase at high S1:F‐actin ratios. Inhibition and activation by tropomyosin are greatest when troponin is present. The observations of Lehrer and Morris (131,132), like those of Weber and colleagues (12,13,57), demonstrated that cooperative activation of thin filaments is dependent on S1 interactions with tropomyosin on actin. The studies suggested that, in addition to Ca²⁺ involvement in the activation process, myosin itself induces an “open‐state” of the thin filament. Figure adapted, with permission, from (131). This figure was originally published in The Journal of Biological Chemistry by S.S. Lehrer and E.P. Morris in an article “Dual effects of tropomyosin and tropomyosin‐troponin on skeletal muscle subfragment 1 ATPase” J. Biol. Chem. 1982; 257:682‐693 © the American Society for Biochemistry and Molecular Biology.

Figure 15. Schematic diagrams of the McKillop‐Geeves three‐state model of the thin filament. (A) Regulatory units of the thin filament are illustrated by single lines depicting tropomyosin over seven circles representing actin; myosin is drawn as a triangle. Blocked, closed, open states are indicated. The dynamic equilibrium between states is illustrated as an azimuthal translocation of tropomyosin on actin and an isomerization of myosin into a strongly attached configuration. K _B and K _T are Ca²⁺‐sensitive steps. (B) The same transitions represented in different cartoon format. It should be recognized that for the weak state of myosin attachment to actin and also for actin‐bound myosin that has not yet released phosphate (not shown in the figure), any favorable or unfavorable tropomyosin‐myosin interactions are still beyond current knowledge. Moreover, the structural correlates of the model do not explain all aspects of myosin regulatory function, and were derived from experiments not including the presence of ATP. Figure in panel A adapted, with permission, from (147), Copyright 1974 American Chemical Society. Figure in panel B adapted, with permission, from (148).

Figure 16. The effect of troponin I trapping tropomyosin in the blocked B‐state thin‐filament position. Reconstructions of (A) F‐actin control filaments, subdomains numbered, and (B) TnI‐decorated F‐actin‐tropomyosin, where the C‐terminal regulatory fragment of TnI is present on every actin subunit. TnI densities are indicated by arrows. (C) Densities attributable to TnI (purple) seen in (B) trapping tropomyosin (red) in the blocking position on actin. (D) Reconstruction of low‐Ca²⁺ native cardiac thin filaments now containing tropomyosin and the entire troponin complex (hence there is one troponin complex on every seventh actin). Densities representing the troponin core domain consisting of the TnIT arm and the N and C lobes of TnC labeled. Black arrow marks tropomyosin in the blocking state. (E) Difference maps highlighting troponin (gold) made by subtracting the actin‐tropomyosin from the map in (D) and superposing the difference on actin‐tropomyosin. Note the position of density attributable to TnT (open arrow) and to the C‐terminal domain of TnI (double‐sided arrow). In (F), maps in (B) and (E) merged; the TnI density seen in (E) superposes on the corresponding TnI density found in (B). Figure adapted, with permission, from (229) (panels A, B, D‐F) and (55) (panel C).

Figure 17. Docking atomic structures of F‐actin into 3D reconstructions of native thin filaments. This procedure by Vibert et al. (215) provided the first structural evidence for the three‐state model of thin filament regulation. (A‐D) Helical reconstructions shown as a blue wire mesh density envelopes cut to about 6 nm to illustrate one actin subunit [actin subdomains numbered in (A)] and the changing position of tropomyosin (arrows) in response to Ca²⁺ and myosin binding. In each case, a high‐resolution actin subunit has been docked within the reconstruction and is shown as a yellow α‐carbon chain with several clusters of residues highlighted in various colors. (A) Reconstruction of low‐Ca²⁺ thin filaments. Note that charged residues 1 to 4 on the N‐terminus of actin and actin residues 92 to 95 (green) on the edge of the subdomain, which are thought to interact electrostatically with myosin prior to cross‐bridge cycling, are not blocked by the tropomyosin density. In contrast, actin residues involved in strong stereospecific binding to myosin (e.g., 24 to 28, 144 to 148, and 340 to 346) (187) (red) are obstructed by tropomyosin. In (C), a reconstruction of high‐Ca²⁺ thin filaments shows that tropomyosin has repositioned and that most highlighted myosin binding residues on actin are now exposed (green) with the exception of, for example, amino acids 332 to 336 (red) at the junction of actin subdomains 1 and 3, which remain occluded. (B) Reconstruction of S1 decorated thin filaments. Note that tropomyosin (arrows) has moved further away from the blocking position than it does in Ca²⁺ (cf. to C) and that all highlighted residues (magenta) are free to bind myosin. In (D), the reconstruction of S1 decorated thin filaments has been contoured to remove part of the S1 density to provide an outline of the S1‐binding site and a better view of the actin residues involved in S1 binding. The densities shown in the reconstruction of the S1 decorated filament and its corresponding actin‐binding site detected are similar to those shown by Milligan et al. (152). The Lorenz et al. atomic model of F‐actin (140) was used for the docking shown. Figure adapted, with permission, from (215).

Figure 18. Fitting atomic coordinates of actin, tropomyosin, and myosin to single particle reconstructions of reconstituted thin filaments. Filaments consisted of actin, troponin, and tropomyosin. Poole et al. (183) brought together the 3D‐EM of Pirani et al. (181) and the synchrotron X‐ray diffraction data from nonoverlapped muscle fibers, thereby synthesizing ideas about steric blocking into a consistent and comprehensive account. (A, B, C) Structural models of F‐actin (blue, cyan, white) [Holmes et al. (87)] and tropomyosin (red, yellow, green) [Lorenz et al. (140)] were fitted within EM density envelopes of reconstituted filaments (translucent surfaces). Reconstructions of (A) low‐Ca²⁺, (B) high‐Ca²⁺, and (C) S1‐decorated filaments fitted with actin and tropomyosin structures. Note the shifting position of tropomyosin in the three cases and the position of the fitted S1 [Rayment et al. (187,188)] in (C). (D) Tropomyosin in its three average positions superposed for comparison; the locations of tropomyosin agree with earlier results of Vibert et al. (215). (E) The S1 binding configuration in (C) is superposed on the low‐Ca²⁺ B‐state filament model from (A). Note the steric clashes between S1 and tropomyosin. While the superposition in (E) makes apparent that tropomyosin occludes the S1‐binding site on actin, it also shows that tropomyosin at low‐Ca²⁺ keeps the cleft between the upper and lower 50K domains of myosin (red) separated and prevents S1 cleft closure, thus interfering with actin‐myosin ATPase and cross‐bridge cycling. This can be best appreciated by comparing S1 heads and tropomyosin on the left and right side of the figure. Figure adapted, with permission, from (183).

Figure 19. High‐resolution cryo‐EM reconstructions of F‐actin and F‐actin tropomyosin. (A) F‐actin at 3.7 Å resolution [constructed from PDB ID:3J8A (219)]. Note the charged residues on actin [Lys 326, Lys 328, Arg 147, Arg 28 (blue), and Glu 25 (red)] that define the binding path assumed by tropomyosin in (B) (8,138) when unperturbed by troponin or myosin. Also note that residues near to Arg 28 and Glu 25 protrude from the surface of actin to form a barrier on subdomain 1 that limits the azimuthal rotation of tropomyosin toward the extreme outer edge of actin in the blocked B‐state. Similarly, the lateral edge of subdomain 4 bulges out, limiting the extent of azimuthal rotation of tropomyosin in the direction of the inner actin domain in the M‐state. Pro 333 (highlighted in pink) defines the boundary between the B‐ and C‐states, but as yet, there is no high resolution model for the C‐state position of tropomyosin. (B, C) The locations of tropomyosin in blocked (magenta) and open (green) positions, respectively, and superposed in (D). B‐ and M‐state tropomyosin constructed from PDBs given in (138) and (8). (E) 8 Å resolution reconstruction of F‐actin‐tropomyosin decorated with Dictyostelium S1 (8) superposed on (D), defining the binding positions of tropomyosin relative to myosin interaction. Behrmann et al. (8), Sousa et al. (200), and von der Ecken et al. (219) were the first to resolve individual α‐helical chains of the tropomyosin coiled‐coil; however, at 8 Å resolution, tropomyosin's side chains are not defined and hence corresponding residues cannot be assigned explicitly to densities in the maps.

Figure 20. Atomic model of F‐actin‐tropomyosin on F‐actin. Li et al. (138) proposed a residue‐specific model for the actin‐tropomyosin structure by optimizing the electrostatic interaction energies between actin and tropomyosin structures. The model was later refined to include the head‐to‐tail tropomyosin overlapping domain by Orzechowski et al. (169). In (A), likely contacts between tropomyosin pseudorepeats and actin subunits are displayed, and, in (B), a magnified view of tropomyosin pseudorepeat 4 is highlighted, illustrating acidic residues on tropomyosin in close proximity to basic Arg 147, Lys 326, and Lys 328 located on actin at the border between subdomain 1 and 3, while Glu 334, Arg 28, and Asp 25 higher on actin subdomain 1 position to associate with oppositely charged amino acids in tropomyosin (cf. Fig. 19). This pattern of contacts is repeated between all tropomyosin pseudorepeats and successive actin subunits along the filament. The position of tropomyosin in the model of troponin‐free actin‐tropomyosin (138) agrees with that observed for tropomyosin when pinned down on regulated filaments by TnI at low‐Ca²⁺ (183). The model also agrees with predictions of actin‐tropomyosin interaction by Brown and Cohen (16,18) and Barua et al. (7). Figure adapted, with permission, from (35) and is based on (138).